TEXT stringlengths 1 7.72M | SOURCE stringclasses 1
value |
|---|---|
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/357752610Maintenance modeling of technical systems under conditions of uncertaintyPresentation · December 2021CITATIONS0READS301 author:Some of the authors of this publication are also working on these related projects:Robotic system of intelligent internal transport View projectRisk and safety assessment in transport View projectSylwia Werbińska-WojciechowskaWroclaw University of Science and Technology105 PUBLICATIONS 615 CITATIONS SEE PROFILEAll content following this page was uploaded by Sylwia Werbińska-Wojciechowska on 11 January 2022.The user has requested enhancement of the downloaded file.5th International Conference on Mathematical Techniques in Engineering Applications (ICMTEA2021) December 3-4, 2021 Graphic Era Deemed to be University | Graphic Era Hill University Dehradun, Uttarakhand, INDIA Sylwia Werbińska-WojciechowskaFaculty of Mechanical EngineeringDepartment of Technical Systems Operation and MaintenanceWroclaw, Poland2IntroductionClassification of main maintenance strategies –model-oriented approachUncertainty in operation and maintenanceResilience-based maintenance concept –potential-oriented approachConclusionsDiscussion, Q&AAgenda3IntroductionThe main difficulties that need to be overcome:•selection of proper maintenance strategy (PM, CBM, Predictive, RBM, RCM,....),•definitionofmachines’maintenancescheduleandperiodicinspections(inspection intervals, maintenance activities times),•selection of operational parameters that should be monitored continously orperiodically,•the amountof data to be processed and analyzed by the Manufacturer/Maintenance Service Provider,•transferring data (about all monitored parameters) between machinery basicmonitoring systems and used maintenance software (system reliability, softwarequality, interface readability, ….)According to PN-IEC 60300-3-10 standard maintenance isdefined as a combination of activities to retain a component in,or restore it to, a state (specified condition) in which it canperform its designated function.Main question:What? When? At what cost? should be maintained in order to provide a desiredsystem operational effectiveness/availabilityWerbińska-Wojciechowska S.: Technical system maintenance. Delay-time-based modeling, Springer 2019 4MaintenancestrategyselectionFailure prediction/ degradationmodellingMaintenanceplanningSpare partsprovisioningRisk management in maintenanceWarranty and maintenanceSystem designMaintenance performance measurementWerbińska-Wojciechowska S.: Technical system maintenance. Delay-time-based modeling, Springer 2019 The main challenge for the maintenance manager is to structure themaintenance procedures and activities to be undertaken in such a way thatthe strategic objectives associated with them are achieved. (De Almeida etal. 2015)Introduction5Classification of main maintenance strategiesCriterionDichotomous divisionDiagnosingDuring operation (e.g. Condition-based M. -CBM)During machine downtime(e.g. First Time M. – FTM)ForecastingProbabilistic (e.g. Predictive M.)Deterministic (e.g. Shutdown M.)PlanningScheduled (e.g. Time-based M.)Unscheduled (e.g.Breakdown M.)ComplexityLow or medium (e.g. production systems)High or very high (e.g. infrastructure networks)UncertaintyLimited (low or medium)Deep (high or very high)ImpactNon-criticalCritical Model-oriented policy for systems operating under conditions of limited uncertainty6The main classification of maintenance strategiesMaintenance strategyPMCBMINSPECTION MAINTENANCEDTMDecision variablesPM frequency/maintenance scheduleinspection frequency/ maintenance threshold inspection frequency/ inspection schedule inspection frequency, delay time parameterMain objectives minimize costs, maximize availability, maximize throughput Classification of main maintenance strategiesWerbińska-Wojciechowska S.: Technical system maintenance. Delay-time-based modeling, Springer 2019 7The main classification of maintenance strategiesMaintenance strategyPMCBMINSPECTION MAINTENANCEDTMDecision variablesPM frequency/maintenance scheduleinspection frequency/ maintenance threshold inspection frequency/ inspection schedule inspection frequency, delay time parameterMain objectives minimize costs, maximize availability, maximize throughput Classification of main maintenance strategiesWerbińska-Wojciechowska S.: Technical system maintenance. Delay-time-based modeling, Springer 2019 8The main classification of maintenance strategiesMaintenance strategyPMCBMINSPECTION MAINTENANCEDTMDecision variablesPM frequency/maintenance scheduleinspection frequency/ maintenance threshold inspection frequency/ inspection schedule inspection frequency, delay time parameterMain objectives minimize costs, maximize availability, maximize throughput Classification of main maintenance strategiesWerbińska-Wojciechowska S.: Technical system maintenance. Delay-time-based modeling, Springer 2019 9Classification of main maintenance strategiesCriterionDichotomous divisionDiagnosingDuring operation (e.g. Condition-based M. -CBM)During machine downtime(e.g. First Time M. – FTM)ForecastingProbabilistic (e.g. Predictive M.)Deterministic (e.g. Shutdown M.)PlanningScheduled (e.g. Time-based M.)Unscheduled (e.g.Breakdown M.)ComplexityLow or medium (e.g. production systems)High or very high (e.g. infrastructure networks)UncertaintyLimited (low or medium)Deep (high or very high)ImpactNon-criticalCritical What policy for highly complex systems (SoS) operating underconditions of deep uncertainty and critical impact of disruptions??1010Uncertainty and variability are one of the fundamental characteristicsof systems and processes.Model uncertaintyProcessuncertaintyEnvironment uncertaintyUncertainty due to modeling process, physical properties, assumed safety factors, data availability, etc. Uncertainty of the ongoing operational/management process and its associated phases Uncertainty related to external conditions (e.g., pricing, regulations, customer safety/reliability requirements), environmental impact on process/system Uncertainty in operation and maintenance11Uncertainty in operation and maintenanceTheconsequenceofknowledgeimperfectionsistheuncertainty in the maintenance process.The concept of uncertainty is understood here as asituation of having limited knowledge, such as: the order, nature, or state of things is unknown, and theconsequence,extent,ormagnitudeofcircumstances, conditions, or events is unpredictable.Bukowski L., Werbińska-Wojciechowska S.: Using fuzzy logic to support maintenance decisions according to Resilience-Based Maintenance concept, Eksploatacja i Niezawodnosc – Maintenance and Reliability, 2021; 23 (2): 294–307.12UncertaintyAleatoryuncertaintyEpistemicuncertaintyUncertainty in operation and maintenanceShamsi M.H, Ali A., Mangina E., O’Donnell J.,A framework for uncertainty quantification in buildingheat demand simulations using reduced-order grey-box energy models, Applied Energy, 275, 2020,115141.1311000Small amount of dataAverage data volumeNo data BIG DATA30DeepuncertaintyMedium uncertaintyLow uncertaintyDEEP LEARNING – AUTONOMOUS D-MSTATISTIC-BASED DECISION-MAKINGBAYESIAN-INFERENCE - BASED D-MHigh uncertaintyRESILIENCE-BASED DECISION-MAKINGVery high riskareaHigh risk areaMedium risk areaLow risk areaBukowski L. Cognitive Dependability Engineering: Managing Risks in Cyber-Physical-Social Systems under Deep Uncertainty, CRC Press Taylor & Francis Group 2021 Data volumeLack of knowledgeWeak knowledgeModerateknowledgeStrongknowledgeUncertainty in operation and maintenance14Highly complex systems (SoS) operating under conditions of deepuncertainty and critical impact of disruptions – potential*-orientedpolicy implementationSystemTypeSystembehaviorThepurposeofthe systemResponsetodisruptionMaintenancepolicySystemexamplePassivestaticorquasi-staticpreservationof the stateresistanceCorrectiveMaintenancenon-repairablemodulesReactivedynamic;predictablyvariablemaintainingsystemavailabilityrobustnessReliabilityBasedMaintenancetechnicalsystemsActivedynamic;unpredictablyvariableensuringsystemcontinuityresilienceResilienceBasedMaintenancecyber-physicalsystemsPotential-oriented policy - RBMSystem types with maintenance policy recommendations*Potential –someone’s or something’s ability to develop, achieve, or succeed (possibility of being or becoming) (https://dictionary.cambridge.org/dictionary/english/potential)15General definition:Resilience – ability of complex systems to survive disruptive changesand disturbances while preserving essential services.Definition of resilienceBukowski L. Cognitive Dependability Engineering: Managing Risks in Cyber-Physical-Social Systems under Deep Uncertainty, CRC Press Taylor & Francis Group 2021 16Modelling resilience17Maintenance management in an organization requires it toprepare specific ability, which can be generally described asMaintenance Support Capability. We propose to define thisterm as follows:MaintenanceSupportCapabilityistheabilityofanorganization to ensure that physical assets continue tofulfilltheirintendedtaskseffectively,efficiently,andsafely,undergivenexpectedaswellasunexpectedconditions of use and maintenance.Ameasureofanorganization’smaturityintermsofmaintenance is its degree of ability to create and maintainspecific potentials over time to resiliently respond to anyforeseeable and unpredictable operating events.Resilience-based Maintenance (RBM) conceptBukowski L., Werbińska-Wojciechowska S.: Using fuzzy logic to support maintenance decisions according to Resilience-Based Maintenance concept, Eksploatacja i Niezawodnosc – Maintenance and Reliability, 2021; 23 (2): 294–307.18Maintenance Support Potentials (MSP) Potential to respond (PR): knowing what to do and being able toreact correctly to any threats and hazards by activating correctlyplanned and prepared actions, by adjusting the required mode ofoperation,orbyintroducingnewactivities,proceduresorprocesses. Potential to monitor (PM): being able to monitor all signals fromthe internal and external environment that may affect anorganization’s performance in the near- or long-term future. Potential to learn (PL): being able to draw conclusions fromexperience, in particular ‘to learn the right lessons from theright experiences’. Potential to anticipate (PA): knowing what to be expected andpredicting future developments considering particular potentialdisruptions, constraints, and changing operating conditions.Resilience-based Maintenance (RBM) conceptBukowski L., Werbińska-Wojciechowska S.: Using fuzzy logic to support maintenance decisions according to Resilience-Based Maintenance concept, Eksploatacja i Niezawodnosc – Maintenance and Reliability, 2021; 23 (2): 294–307.19A functional diagram of the Maintenance Support System for RBMBukowski L., Werbińska-Wojciechowska S.: Using fuzzy logic to support maintenance decisions according to Resilience-Based Maintenance concept, Eksploatacja i Niezawodnosc – Maintenance and Reliability, 2021; 23 (2): 294–307.Resilience-based Maintenance (RBM) concept20Resilience-based Maintenance (RBM) conceptA new organization's maintenance support potential level assessment ratio (�):�������(3)where:� – organization's maintenance support potential level; � – ith maintenance support potential; � – weight for ith maintenance support potential; n – number of analyzed maintenance support potentials.We may use:-qualitative approaches-quantitative approachesBukowski L., Werbińska-Wojciechowska S.: Using fuzzy logic to support maintenance decisions according to Resilience-Based Maintenance concept, Eksploatacja i Niezawodnosc – Maintenance and Reliability, 2021; 23 (2): 294–307.21Assessment methods and modelsModelling based on fuzzy logicuseQualitative analysis basedonexpertopinionsandweighted point methodResilience-based Maintenance (RBM) concept22Resilience-based Maintenance (RBM) conceptBukowski L., Werbińska-Wojciechowska S.: Using fuzzy logic to support maintenance decisions according to Resilience-Based Maintenance concept, Eksploatacja i Niezawodnosc – Maintenance and Reliability, 2021; 23 (2): 294–307.MSP assessment parametersReadiness levelReadiness to respondReadiness to monitorReadiness to learnReadiness to anticipateProcess regency Responding regencyMonitor regencyLearning regencyForecasting regencyApplication of fuzzy logic – an example for automotive company23Surface view of the fuzzy inference system – an applicationexample for automotive companyBukowski L., Werbińska-Wojciechowska S.: Using fuzzy logic to support maintenance decisions according to Resilience-Based Maintenance concept, Eksploatacja i Niezawodnosc – Maintenance and Reliability, 2021; 23 (2): 294–307.Resilience-based Maintenance (RBM) concept24ConclusionsNew challenges regarding Industry 5.0 In 2021, the European Commission formally called for the Fifth Industrial Revolution Industry 5.0: towards a sustainable, human-centric and resilient European industry. Luxembourg, LU: European Commission, Directorate-General for Research and Innovation; 2021• Sustainable–circular processes (4’s „R” -reuse, repurpose, recycle, reduce),• Human-centric–from technology-driven to society-centric approach,• Resilient–arming production and services against disruptions and crisis. 25Conclusions• Needs a multidimensional approach incorporation:a)Configuring a resilient topologyb)Creating a functionally resilient SoSc) Providing the modelling approachd)Developing maturity model – for maintenance potential levelassessment26Thank you for attentionSylwia Werbińska-Wojciechowska, Department of Technical Systems Operation and Maintenance e-mail: sylwia.werbinska@pwr.edu.pl; tel. +48 71 320 34 27 View publication stats | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/338422343The Influence of Plasticizer Nature and of Processing Mode Upon theCharacteristics of Flexible Poly(vinyl chloride) CompositesArticle in MATERIALE PLASTICE · January 2020DOI: 10.37358/MP.19.4.5275CITATION1READS7926 authors, including:Some of the authors of this publication are also working on these related projects:Nano-Composite Coatings: Processing, Characterization, Properties, and Applications View project"Contingency of CBRN hazards and improvement of national security resources SECURE-NET" PN-III-P1-1.2-PCCDI no.70/2018 View projectComan Alina ElenaThe National Institute for Research & Development in Chemistry and Petrochemis…7 PUBLICATIONS 18 CITATIONS SEE PROFILERaluca GaborNational Institute for Research and Development in Chemistry and Petrochemistry123 PUBLICATIONS 1,053 CITATIONS SEE PROFILECristian-Andi NicolaeNational Research and Development Institute for Chemistry and Petrochemistry -…203 PUBLICATIONS 2,140 CITATIONS SEE PROFILEValentin RaditoiuNational Institute for Research and Development in Chemistry and Petrochemistry155 PUBLICATIONS 1,044 CITATIONS SEE PROFILEAll content following this page was uploaded by Coman Alina Elena on 19 June 2020.The user has requested enhancement of the downloaded file.MATERIALE PLASTICE ♦ 56 ♦ no. 4 ♦ 2019 845 http://www.revmaterialeplastice.ro The Influence of Plasticizer Nature and of Processing Mode Upon the Characteristics of Flexible Poly(vinyl chloride) Composites ALINA ELENA COMAN1,2, AUGUSTA RALUCA GABOR1, CRISTIAN ANDI NICOLAE1, VALENTIN RADITOIU1, GHEORGHE HUBCA2*, TANTA-VERONA IORDACHE1* 1Polymers Department-The National Research & Development Institute for Chemistry and Petrochemistry-ICECHIM, Bucharest, 060021, Romania 2Faculty of Applied Chemistry and Materials Science, Advanced Polymer Materials Group, University Politehnica of Bucharest,1-7 Polizu Str., 011061, 011061, Bucharest, Romania The topic of this paper refers to the influence of the plasticizer and the processing mode upon the characteristics of the poly(vinyl chloride) (PVC) composites. Thereby, in this study two types of industrial plasticizers were used to highlight their influence upon the properties of final composites. The employed lubricant was stearic acid, the most common and cheapest additive used in the industry for cables manufacturing. For economic reasons, calcium carbonate of 2 m size was used as reinforcing agent. Further on, two sets of samples were prepared, targeting the influence of the processing mode upon the properties of final composites. Beside the structure (by FT-infrared), thermal behavior (thermal analyses and differential scanning calorimetry) and mechanical properties (dynamic mechanical analyses, tensile strength and elongation at break) of PVC composites, the limiting oxygen index (LOI) and the overall morphology of the samples were also investigated. Keywords: plasticizer influence, LOI, processing methods PVC is intensively used nowadays, due to its low cost, fire resistance, chemically resistance to acids, salts, bases, fats, and alcohols, making it resistant to the corrosive effects of sewage (which is why it is extensively utilized in manufacturing sewer piping systems), electrical protecting properties, good insulation properties, good tensile strength, ability to use it in rigid, semi flexible or flexible state and because it can be colored in a wide range of colors [1-5]. In industry, the most important issues are the quality of the product and the price [6]. It can be said that both are on the same line. Therefore, in this paper a commercial formulation of flexible PVC composite was modified to obtain a more cost-efficient product with improved properties. To accomplish these objectives, diisononyl phthalate (DINP) and dioctyl terephthalate (DOTP) were used as plasticizers. These two plasticizers are intensively used in the industry, because they are cheap and also because they deliver the required properties for the PVC composites according to their final application. DOTP is a non-phthalate plasticizer, being the diester of terephthalic acid and of branched-chain 2-ethylhexanol. DINP is usually a mixture of chemical compounds consisting of various isononyl esters of phthalic acid [7-8]. Polyvinyl chloride (PVC) is rigid at normal temperature due to intermolecular forces between molecules and short distances between them. When the polymer is heated to a high temperature, the energy of the molecular motions exceeds the energy of the intermolecular forces, and this broadens the molecular distances resulting in PVC softening. Addition of plasticizer prevents the PVC molecules from coming closer to each other. The molecules of the polymer are isolated even at normal temperature and its softening is preserved. This process is called plasticizing [9]. Diisononyl phthalate is categorized as a commodity phthalate ester plasticizer and maintains good performance of composites at low temperature. Dioctyl terephthalate is preferred for volatility resistance, which prevents the health and toxicity issues [10]. Lubricants play a fundamental role in the way in which polyvinyl chloride melts and flows during compounding [11]. Lubricants can be internal, external or combined. External lubricants are used to reduce the friction between the polymer melt and metal surfaces during processing, and internal lubricants are used to reduce frictional forces within the polymer matrix and to lower the effective viscosity. Mixed lubricants have both uses. The employed lubricant in this study was stearic acid, which is a mixed lubricant [12-14]. The calcium carbonate reinforcing material is cheap and also provides good fire-resistance properties to PVC composites. When the compounds are exposed to a flame, CaCl2 is produced and captures the hydrochloric acid emitted during decomposition of the PVC matrix [15-16]. \ In a previous study of the same group the influence of the lubricant upon the properties of PVC composites was investigated [17]. As a result, considering the present literature, this paper provides forward a comparative study on the influence of plasticizers and of the processing mode. Therefore, the study was performed on two variation levels, which implied analyzing two series of PVC composites (with DINP and DOTP) for structure, thermal stability, homogeneity * email: gheorghe_hubca@yahoo.com, Phone: +40744754210; iordachev.icechim@gmail.com, Phone: +40755159896 MATERIALE PLASTICE ♦ 56 ♦ no. 4 ♦ 2019 846 http://www.revmaterialeplastice.ro and fire-resistance. The samples were prepared using different processing modes, i.e. by passing once or by passing twice through the extruder. The literature of PVC composite materials is abundant with new fillers or nanofillers and new plasticizers, but they lead to very high production costs. Therefore, this study also provides an alternative to the eccentric recipes for PVC composites by improving the existing commercial formulations, at lower costs. Experimental part Materials and method For this research the following materials were used: - PVC Ongrovil S5070 from Borsod Chem (K-value 70.4); Diisononyl phthalate (DINP) as plasticizer (LG Chem); Dioctyl terephthalate (DOTP) as plasticizer (Chem Flexx); Ca-Zn salt as heat stabilizer (Baeropan MC 91446 KA/3, Baerlocher); Calcium carbonate particles of 2 m (CaCO3) as reinforcing agent (Omya Calcita); Stearic acid (Stera Chemicals) as lubricant. The test specimens were prepared by melting and mixing PVC with the heat stabilizer, reinforcing agent, plasticizer and lubricant. First of all, the heat stabilizer was mixed with PVC to prevent degradation before processing, afterwards the rest of materials were added in the following order: plasticizer, lubricant and reinforcing agent (formulations given in Table 1). The ingredients were blended in a masticator and heated at 110 °C. After 1 hour, the obtained mass was cooled-down at 50 °C. The next procedure consisted of passing the material through an extruder. The one series of PVC with plasticizer was passed once through the extruder and another series was passed twice. Afterwards, the grains resulted from extrusion were blended into a two-roll mill at 160 °C for 20 min. The temperature of roll milling depends on the final use of end product. For cable isolation, as final application of the employed PVC composites, the temperature is 160 °C. The roll milling smoothed the product by rolling followed by compression at 170 °C to form sheets with dimension 150×150×1 mm3. Table 1 FORMULATIONS OF PVC BASED COMPOSITES Sample code* Plasticizer Processing mode PVC_DINP 1 DINP 1 extruder pass PVC_DOTP 1 DOTP 1 extruder pass PVC_DINP 2 DINP 2 extruder passes PVC_DOTP 2 DOTP 2 extruder passes *All the samples have the following formulation: PVC 100 phr, plasticizer 50 phr, reinforcing agent 80 phr, lubricant 0.4 phr and heat stabilizer 4 phr Infrared spectroscopy analysis (FTIR) Molecular structure of the PVC composites was analyzed using an infrared spectrometer type JASCO FT/IR 6300 Specac Golden Gate ATR with diamond in transmission mode. The measured range was 400-4000 cm-1, 16 scans. Thermogravimetric analysis (TGA) The thermogravimetric analysis was performed using a thermogravimetric analyzer (Q5000IR, TA Instruments) at a heating rate of 10 °C min-1 under a nitrogen atmosphere or air. Samples of approximately 8 - 14 mg were analyzed in the 40-700 °C temperature range. Differential scanning calorimetric analysis (DSC) Differential scanning calorimetry (DSC) experiments were carried out using DSC Q2000 from TA Instruments, working in temperature-modulated mode (MTDSC) with dynamic purge gas (helium 99.99% with flow rate 25 ml min-1, modulated scan). Approximately, 25 - 30 mg samples were weighted and heated from -110 to 155 °C with a heating rate of 20 °C min-1. Dynamic mechanical analysis (DMA) Dynamic mechanical analysis (DMA) experiments were carried out using DMA Q800 from TA Instruments, module DMA Multi-Frequency–Strain, in the tension mode. The thermal behaviour of the PVC composites was obtained through the temperature ramp method. The sample size was 13.5 × (6.4-7) × (0.77-0.79) mm and the heating rate was 5 °C min-1 from – 75 to 105 °C. Tensile strength and elongation at break The tensile strength and elongation at break of the samples were performed on a computer-controlled electronic tensile testing machine (Instron 3382) with a constant speed of 100 mm min-1 at room temperature. Limited oxygen index (LOI) The limited oxygen index (LOI) index was tested using a Stanton Redcroft FTA Flammability Unit instrument. The test specimens were 120×60×2 (L×W×T) mm in dimension. MATERIALE PLASTICE ♦ 56 ♦ no. 4 ♦ 2019 847 http://www.revmaterialeplastice.ro Scanning electronic microscopy (SEM) Scanning electronic microscopy (SEM) was used to study the morphology and homogeneity of the PVC samples. Vega-3 (Tescan, Brno, Czech Republic) apparatus in high vacuum mode with working distances in the range of 4–8 mm and an acceleration voltage of 5 kV was used. Before examination, the samples were coated with a thin layer of gold with a gold coater (Quorum Q 150 RS; Quorum Technologies, Lewes, UK). Results and discussion FTIR analysis Plasticizers provide compound flexibility, which improves viscosity and processability [18]. These types of phthalate plasticizers behave like solvents but they are not chemically bounded to PVC [19]. Figure 1 presents the plasticizers IR spectra, which are very much similar due to similarities of the structure. Figure 2 shows the characteristic bands of PVC, plasticizer, lubricant and CaCO3. The presence of PVC is given by de C-Cl bond appearing around 600 cm−1. The band at around 1720 cm-1 is attributed to the C=O vibrations belonging to the ester plasticizer. The decrease of the band intensity at 1720 cm-1 suggests the plasticizer migration. It can be observed that the compound with DINP had a higher decrease of intensity compared to DOTP-based composites. This means that DINP migrate faster from the compound than DOTP [20]. Fig.1 Infrared spectra of plasticizers Fig. 2 FTIR spectra of CaCO3 and 2 PVC composites For carbocatenary long chain (CH and CH2 bonds) of the plasticizer and of the lubricants are overlapping with those of PVC chain at 2850 and 2916 cm−1. At 1270-1273 cm-1 the vibrations of C-O groups belonging to the lubricant (aliphatic carboxyl) and phthalate plasticizer (aromatic carboxyl) appear [21]. The reinforcing agent displays specific bands around 874 cm-1 attributed to Ca-O vibrations, and the bands from 1418 cm-1 corresponding to C-O group from the carbonate structure [22]. The series of PVC composites passed 2 times through the extruder, presented the same bands as the series passed 1 time through the extruder. Thermogravimetric analysis (TGA) TGA analysis (Table 2) showed that the total mass loss of neat PVC, in the first stage (150 - 380 °C) is about 64 wt.%. In the 40 - 150 °C temperature range no significant mass losses are observed (approximately 0.01 wt.%). The thermal degradation of PVC takes place in two steps, as follows: in the first step (150 - 380 °C), hydrochloric acid is emitted and polyene structures are formed, and in the second step (380- 540 °C) the cyclization of the polyene structures takes place with the formation of aromatic compounds. In the second stage the mass losses are about 27 %. The second stage represents the thermal decomposition of PVC backbone and, therefore, lower molecular weight compounds and smoke are produced [23-25]. MATERIALE PLASTICE ♦ 56 ♦ no. 4 ♦ 2019 848 http://www.revmaterialeplastice.ro For the PVC composites however, an improvement regarding the weight losses in both stages can be observed. After addition of plasticizer, lubricant, stabilizer and reinforcing agent, the mass loss in the first step is about 47-51 wt.% for DINP-based composites and 49-61 wt.% for DOTP-based composites. This mass loss is due to plasticizer desorption [26]. In this temperature range (i.e. 150 – 380 °C), decomposition of the stearic acid lubricant (and of Ca-Zn stabilizer, which is also a metal carboxylate) also begins. In the second step, the composites lose approximately 20 wt. % from the total mass and CaCO3 starts to decompose slowly. Hence, it can be confirmed that this stabilizer is good for this temperatures range as the composites were much more stable at higher temperatures. CaCO3 also minimizes the emissions of hydrochloric acid during the PVC composites degradation by adsorbing hydrochloric acid released during PVC decomposition, leading to the formation of CaCl2. CaCO3 with 2 µm particle sizes can absorb higher amounts of hydrochloric acid [17]. Therefore, it can be stated that calcium carbonate acts as a heat stabilizer by trapping hydrochloric acid. Table 2 THERMOGRAVIMETRIC ANALYSIS FOR RAW MATERIALS USED IN PVC COMPOSITES Sample 40 -150°C 150 - 380°C 380 - 540°C 540 - 640°C Onset Point Residue at 700°C Wt. loss Wt. loss Tmax1 Wt. loss Tmax2 Wt. loss Tmax3 Temp Weight (N2) (Air) % % °C % °C % °C °C % % % Neat PVC 0.01 64.19 291.0 27.85 453.2 0.42 - 274.2 99.99 7.29 0.001 PVC_DINP 1 0.15 51.49 286.4 16.70 458.0 1.81 620.4 253.7 98.98 28.63 19.48 PVC_DINP 2 0.16 46.73 275.7 16.89 459.1 2.38 597.9 262.8 98.95 32.78 19.76 PVC_DOTP 1 0.08 61.31 284.6 14.43 456.8 1.02 622.1 261.4 99.06 21.22 13.54 PVC_DOTP 2 0.08 49.18 274.3 11.38 453.3 1.56 556.9 261.8 99.10 33.86 26.77 From the two series of PVC composites, it was noticed that series 2 (composition two times passed through the extruder) presented higher decomposition temperatures, indicating a better homogeneity of the material; this is highly desired to prevent decomposition of PVC before processing. Hence, TGA analysis, underlined that PVC composites with DINP presented similar thermal stability to those with DOTP, as lubricant, but a much better thermal behavior compared to either neat PVC or the first series of samples passed only once through the extruder. Differential scanning calorimetric analysis (DSC) To evaluate the glass transition temperature (Tg) of the composites, DSC was used. The Tg is very important for the composites in general, to establish the processing temperature without degrading the polymer. The decomposition temperature of neat PVC (around 160 °C) is lower than the melting temperature (260 °C). Therefore, it is mandatory to decrease the Tg in order to prevent degradation before processing [27-28]. In Table 3, the Tg for each PVC composite can be observed. Neat PVC, without additives, has a Tg of approximately 85 ºC. According to Table 3, the lowest Tg values were obtained for the samples that passed two times through the extruder machine. This is due to a better homogenization of the materials, gained when the raw materials were extruded twice. PVC compositions with DINP presented a decrease of Tg with about 3 ºC, compared to those with DOTP. This may be due to a more hydrophilic character of DINP, as suggested by the TGA as well (higher water weight loss). Table 3 DSC RESULTS FOR PVC BASED COMPOSITES 1st Heating PVC - Glass Transition Total heat Flow Onset (°C) Tg (°C) End (°C) Δ Cp [J/(g·°C)] PVC_DINP 1 -56.1 -28.9 -2.4 0.252 PVC_DINP 2 -58.0 -36.4 -2.8 0.275 PVC_DOTP 1 -53.7 -25.6 6.0 0.309 PVC_DOTP 2 -56.6 -33.4 -2.3 0.256 Dynamic mechanical analysis (DMA) One of the most important characteristics for processing the polymeric materials is viscoelasticity. This parameter may be evaluated using a dynamic mechanical analysis instrument, to measure the storage modulus, loss modulus and damping properties. An important index for measuring the stiffness and elasticity of polymeric materials is the storage modulus. Figure 3 shows the storage modulus of PVC composites in DMA temperature scans. It can be seen that composites presented MATERIALE PLASTICE ♦ 56 ♦ no. 4 ♦ 2019 849 http://www.revmaterialeplastice.ro higher storage modulus than neat PVC through the whole temperature scan range. Slight difference can be observed between the samples due to different plasticizers. Composites with DINP showed a higher storage modulus than those with DOTP. The second pass through the extruder is important and it can be observed by the difference in the storage modulus. According to Figure 3 the second series of PVC composites (PVC_DOTP 2 and PVC_DINP 2) registered an increase of the storage modulus and a decrease of the loss modulus. These results confirmed once again that passing the material twice through the extruder leads to more homogenous and compact composites. Tan delta (Figure 4) however, did not present significant differences between composites. The range of this parameter was between 0.39 and 0.4. Concluding, the plasticizer incorporation increased the storage and loss modulus, compared with neat PVC, which explained the higher heat dissipation in the PVC composites compared with neat PVC, confirming that the new formulations of PVC are flexible. Fig. 3 Storage modulus of PVC composites Fig. 4 Tan delta for PVC based composites Tensile strength and elongation at break In practical applications, for evaluating the quality of flexible PVC composites, the most important properties are the mechanical properties such as: tensile strength and elongation at break. It can be observed from Figure 5 a and b that elongation at break and tensile strength of test samples increased when the samples were passed two times on the extruder. From the Figure 5, it can be stated that the mechanical properties made the difference between the two plasticizers. For PVC composites with DINP the increase of tensile strength and elongation at break is approximately 14 %, but for those with DOTP the increase for these properties was about 65 %. Samples with DINP did not attained remarkable mechanical properties compared to neat PVC. Therefore, it can be concluded that DOTP is more efficient when the composites are passed twice through the extruder. The improvement of mechanical characteristics can also be partially attributed to the increase of PVC mobility after passing twice on the extruder [29]. a) b) Fig. 5 Mechanical properties for the two PVC series of composites a) tensile strength, b) elongation at break Limited oxygen index (LOI) The limiting oxygen index represents the percentage of oxygen present to support combustion of the plastic, therefore the higher the LOI the lower the flammability. Any material with LOI value of less than 21 will probably burn in an open-air situation. Values of LOI greater than 21 mean that the composites do not burn immediately in air [30-31]. MATERIALE PLASTICE ♦ 56 ♦ no. 4 ♦ 2019 850 http://www.revmaterialeplastice.ro To register the LOI values, the composite samples were trapped in a vertical glass column fed with a slow stream of oxygen/nitrogen mixture. The samples were ignited with a flame and burned downward. These results are shown in Figure 6. Calcium carbonate cannot be considered a flame-retardant agent, but it acts as a radical scavenger during the ignition and prevents the reaction between PVC molecules and oxygen [16]. The plasticiser plays an important role, as well, in the flame-retardancy of the PVC samples. Usually plasticized composites are more flammable than rigid composite, due to the plasticiser. The highest and lowest LOI were attained for PVC_DINP 2 and PVC_DOTP 1, respectively. Comparing the two series of composites, with DINP or DOTP, it can be observed that the highest value for oxygen index is achieved when the samples are extruded two times. Hence, the sample with DINP (second pass) presented a LOI of 26 % and the one with DOTP (second pass) a LOI of 24.5 %. Fig. 6 LOI values for PVC composites Scanning electronic microscopy (SEM) The final study regarding the PVC samples is the morphology of the composites. As it can be seen from the micrographs in Figure 7, that the samples passed twice through the extruder are more homogenous. Fig. 7. SEM analysis for PVC composites The raw materials are mixed better, and the particle agglomerations are avoided. Samples with one pass through the extruder presented macropores and agglomerations. The plasticiser exhibited an influence on the surface morphology, as well. Samples with DOTP presented a smoother surface compared with DINP samples. This means the compatibility of DOTP with the other raw materials was better assured. Conclusions PVC composites with potential applications in the industry for electrical wires and cables were prepared and characterized. Calcium carbonate with particle size of 2 m as cheap version of flame-retardant filler was used as reinforcing agent. Further on, the influence of plasticizer was studied using two common commercial plasticizers used in the cable industry, i.e. DINP and DOTP. Both plasticizers have proven to decrease the Tg value of PVC and also to deliver adequate properties for the final products. The thermal, mechanical and flame-retardant properties of the PVC composites presented similar improvements when the material was passed twice through the extruder. Yet, a significant MATERIALE PLASTICE ♦ 56 ♦ no. 4 ♦ 2019 851 http://www.revmaterialeplastice.ro increase was observed for the mechanical properties when DOTP was used. Therefore, it can be stated that the optimum formulation of PVC composites for this particular application should be based on DOTP as plasticizer. Plus, the material should definitely be passed through the extruder two times, for better homogenization, as the overall characteristics of the composites have proven to be highly dependent upon the homogenization degree. Acknowledgements: The work has been funded by the Operational Programme Human Capital of the Ministry of European Funds through the Financial Agreement 51668/09.07.2019, SMIS code 124705 and by a grant awarded by UEFISCDI in framework of the PCCDI competition, contract no. 70PCCDI/2018 SECURE-NET. References 1. ZHANG M., BUEKENS A., JIANG X., LI X., Dioxins and polyvinylchloride in combustion and fires, Waste Manag. Res., 33, 2015, p. 630–643. 2. PETROVIC E. K., HAMER L. K, Improving the Healthiness of Sustainable Construction: Example of Polyvinyl Chloride (PVC), Buildings, 8, no. 28, 2018. 3. WILKES C.E., SUMMERS J.W., DANIELS C.A., PVC Handbook. Hanser, Munich, 2005. 4. BRAUN D. J., Polym. Sci. Part A: Polym. Chem., 42, no. 3, 2003. 5. FECHTER R. H., SANDROCK C., KUHNERT I., LABUSCHAGNE J., Modelling and Optimization of a Flexible Poly(vinyl chloride) Compound Formulation for Mine Cables, Journal Vinyl Addit.Technol., 2017. 6. ALTARAZI S., HEMEIMAT R., WAKILEH M., QSOUS R., KHREISAT A., Optimizing Materials Cost and Mechanical Properties of PVC Electrical Cable’s Insulation by Using Mixture Experimental Design Approach, International Scholarly and Scientific Research & Innovation, 7, no. 3, 2013. 7. REINECKE H., NAVARRO R., PÉREZ M., Plasticisers, Encyclopedia of Polymer Science and Technology, John Wiley & Sons, 2011. 8. KRAUSKOPF L.G., GODWIN A., Plasticizers, PVC Handbook, Hanser Publishers, Munich, Hanser Publications, Cincinnati, 2005. 9. CARROLL W. F., JOHNSON R. W., MOORE S. S., PARADIS R. A., Poly(Vinyl Chloride), Applied Plastics Engineering Handbook, Elsevier Inc., 2011. 10. DOTP Plasticizer: Next Move in Plasticizer Industry, accessed in 11.09.2019, http://techsciblog.com/dotp-plasticizer-next-move-plasticizer-industry/. 11. MIRCI, L. E., BORAN, S., Plasticizers and lubricants of adipic ester type with complex structure, Mat Plast., 44, no. 4, 2007, p. 383-392. 12. PEGGY S. HOLSOPPLE, An Examination of DSC as a Method of Characterization of PVC Lubricants, J. Vinyl Technol., 9, no. 4, 1987. 13. RABINOVITCH E. B., LACATUS E., Summers J. W., The Lubrication Mechanism of Calcium Stearate/Paraffin Wax Systems in PVC Compounds, J. Vinyl Technol., 6, no. 3, 1984. 14. WILSON A. P., Lubricant for polyvinyl chloride, US3981838A, 1976. 15. KARAYILDIRIM T., YANIK J., YUKSEL M., SAGLAM M., VASILE C., BOCKHORN H., The effect of some fillers on PVC degradation, J. Anal. Appl. Pyrolysis, 75, 2006, p. 112–119 16. TUEN B. S., HASSAN A., BAKAR A. A., Thermal Properties and Processability of Talc- and Calcium Carbonate-Filled Poly(vinyl Chloride) Hybrid Composites, J. Vinyl Addit. Technol., 2012 17. COMAN A., GABOR A. R., STOIAN S., NICOLAE C. A., RADITOIU V., GIFU I. C., IORDACHE T. V., HUBCA G., New Formulations of Flame-retardant Flexible Polyvinylchloride Composites, Mat. Plast. (Bucharest), 56, no. 3, 2019, p. 568-577 18. HERMAN M., Plasticizers, Encyclopedia of Polymer Science and Technology, 3, John Wiley & Sons, U.K. 19. GAVRILĂ D. E., Studies of Degradation of Plasticized Polyvinyl Chloride (PPVC), Int J Eng Res Appl, 6, no. 1, 2016, 56-63. 20. DEMIR A. P. T., ULUTAN S., Migration of Phthalate and Non-Phthalate Plasticizers Out of Plasticized PVC Films into Air, J. Appl. Polym. Sci., 2012, DOI: 10.1002/APP.38291. 21. UL-HAMID A., SOUFI K. Y., AL-HADHRAMI L. M., SHEMSI A. M., Failure investigation of an underground low voltage XLPE insulated cable, Anti-Corrosion Methods and Materials, 62, 5, 2015, p. 281–287. 22. REIG F. B., ADELANTADO J. V., MOYA MORENO M. C., FTIR quantitative analysis of calcium carbonate (calcite) and silica (quartz) mixtures using the constant ratio method. Application to geological samples, Talanta, 58, no. 4, 2002, p. 811-821. 23. MAYER Z., Thermal Decomposition of Polyvinyl Chloride) and of Its Low-Molecular-Weight Model Compounds, Journal Macromol. Sci., Part C, 10, no. 2, 1974, p. 263-292. 24. SHAH B. L., SHERTUKDE V. V., Effect of Plasticizers on Mechanical, Electrical, Permanence, and Thermal Properties of Poly(vinyl chloride), J. Appl. Polym. Sci., 90, no. 12, 2003, p. 3278-3284. 25. ZHU H.M., JIANG X.G., YAN J.H., CHI Y., CEN K.F., TG-FTIR analysis of PVC thermal degradation and HCl removal, J. Anal. Appl. Pyrolysis, 82, 2008, p. 1–9. 26. CAILEAN, A., SUTIMAN, D., CREPESCU, I., CROITORU, C., MARECI, D, Thermal behaviour prediction of rigid polyvinyl chloride, Mat Plast., 44, no. 4, 2007, p. 294-297. 27 UNAR I. N., SOOMRO S.A., AZIZ S., Effect of various additives on the physical properties of Polyvinylchloride resin, J. Anal. Environ. Chem., 11, no. 2, 2012, p. 44 – 50. 28. MARTIN T. M., YOUNG D.M., Correlation of the glass transition temperature of plasticized PVC using a lattice fluid model, Polym J., 44, 2003, p. 4747–4754. 29. BURDONOV A. E. , BARAKHTENKO V. V., ZELINSKAYA E. V., YUDIN V. E., ELOKHOVSKII V. Y., Dynamic mechanical analysis of composite materials based on polyvinyl chloride and thermal power station waste, Int. Polymer Sci. and Tech., 42, no.7, 2015, p. 53-57. 30. NAKAGAWA Y. A., Note on an Oxygen Index test on Electric Cables, J Fire Sci., 1994; https://doi.org/10.1177/073490419401200603. 31. AL-MOSAWI A. I., MAROSSY K., KÓNYA C., Effect of Plasticizer Percentage on Thermal Properties of Plasticised PVC, Elixir Materials Science, 117, 2018, p. 50509-50511. Manuscript received: 12.11.2019 View publication stats | RESEARCH GATE |
Rural Ceramic Manufacture in PrecolumbianHonduras: The Application of Pétrographie Analysisto the Study of the Chaînes OpératoiresMeaghan Peuramaki-Brown+ABSTRACT. Tbis study presents a prelimi-nary attempt to examine stages witbin thechaîne opératoire of Late Classic (650-900C.E.) ceramic manufacture at the Preco-lumbian site of Rancho del Rio, Honduras.Materials recovered from patio, bousemound, and midden excavations in a ruralsetting, including finished vessel sherdsand potstands, in addition to briquettesmade from local clay sources, are examinedthrough tbin section petrography. Tbisattempt to outline technological chainsand styles allows the classification andunderstanding of behaviour and ultima-tely, cognition, through the identificationof the series of units of actions that bringa material from its natural state to a fabri-cated form.RÉSUMÉ. Cette étude présente une tenta-tive préliminaire d'examiner les étapes dela chaîne opératoire dans la fabricationde céramique au site précolombien (650-900 ap. J.-G.) Rancho del Rio, en Honduras.Des tessons de poterie et des supports àvases recouverts dans les fouilles archéolo-giques d'un patio, d'un monticule domes-tique, et d'un dépotoir dans un contexterurale, en plus de briquettes faites à partirde sources locales d'argile, ont été examinéspar analyse pétrographique. Cette tenta-tive de décrire les chaînes et les modèlestechnologiques permet la classification etla comprébension de comportements, etfinalement, l'approche cognitive, par l'iden-tification de la série des unités d'actions quiapporte un matériel de son état normal àune forme fabriquée.APPLYING MACROSCOPIC TECHNIQUES,. to the analysis of pottery as a meansto understand socio-political identi-ties and boundaries, world systems,and daily activities is commonplacein Precolumbian Mesoamerican stud-ies. Research and the application ofmicrovisual and elemental techniquesto address how and where pottery wasmanufactured were relatively morelimited until approximately 20 yearsago. Key reasons for ignoring such ques-tions and approaches included issues ofpreservation, an inability to efficientlyrecognize manufacture locations andremains, the focus of excavations onmajor site epicentres and elite civic-ceremonial structures to the detrimentof lower strata houselots where muchmanufacture was likely to have occurred(Robin 2002), and a lack of convictionon the part of many archaeologists asto the advantages of archaeometricand geological approaches to the studyof ceramics. Despite numerous ethno-graphic and ethnoarchaeological stud-ies of pottery-making in Mexico andGentral America, literature of the pastcentury reveals additional cause for theomission of archaeometry and geologyfrom these studies including: an absence• Department of ArchaeologyUniversity of Calgary, 2500 University Dr. NW,Calgary, AB T2N 1N4 [mmpeuram@ucalgary.ca]CANADIANJOURNAL OF ARCHAEOLOGY/JOURNAL CANADIEN D'ARCHÉOLOGIE 36: 166-187 (2012)RURAL CERAMIC MANUFACTURE IN PRECOLUMBIAN HONDURAS • 167of questions posed regarding the chaîneo/>érato¿re of manufacture; and the beliefthat archaeological typologies based onmacroscopic characteristics can supplysignificant and abundant informationregarding finished vessels (Wardle1992:11).This paper presents a study in thecharacterization of the micro-natureof ceramics from the Precolumbiansite of Rancho del Rio in northwesternHonduras through the use of thin sec-tion pétrographie analysis to outiine thechaîne opératoire of manufacture. Thisstudy serves as a preliminary attempt tocombine macrovisual and microvisualanalyses of both finished ceramics andmanufacture residue in the discussion ofmanufacture behaviour on a local, com-munity level in Precolumbian Honduras.Identifying Ceramic ManufactureWhy study ceramic manufacture andproduction?' Is it important to under-stand where, how, and by whom vesselswere made? Manufacture, or technol-ogy, is "not simply a body of explicitiyformulated and objectively describedknowledge," but "is one of the socialprocesses by which individuals negoti-ate and define their identities, in termsof gender, age, belief, class, and so on"(Sinclair 2000:196). Individuals learnthe techniques of manufacture as mem-bers of their respective societies, makingthem part of the social processes thatarchaeologists and anthropologists areso keen to understand. As such, outlin-ing the manufacturing processes ofceramic vessels can lead to an under-standing of incorporated social proc-esses and behaviours.Understanding ceramic manufacturethrough the behaviour of producers isan attempt to codify phenomenal orderswithin ceramic variation as opposed tothe ideational orders of typical type-variety systems (Ford and Steward 1954).The adoption of techniques that focuson the understanding of manufactur-ing choices, or chaînes opératoires, allowsthe classification and consideration ofbehaviour and potentially cognition,through the identification of series ofbehavioural units or actions that bringa material from its natural state to afabricated form (Bleed 2001; Cresswell1976; Lemonnier 1986). Such actionsare imbued with cultural and evenindividual influences while at the sametime being restricted by the physio-chemical properties of the raw materials(Arnold 1971; Hassan 1988; Schiffer1976). This approach to material cultureallows archaeologists a closer view ofdynamic individual agency and sharedideas within given social groups of theprehistoric past, through the identifica-tion of patterns and differences in static"things" that are reflective of choices(Lemonnier 1992). The chaîne opératoireapproach serves to outiine the four ele-ments of any technological process—materials, tools, actions, and specificknowledge (Sheets et al. 1975)—andfocuses directiy on the behaviours of arti-sans and the use of discontinuities in themanufacture process to isolate analyticaltaxa in an attempt to place behaviours ina framework of theoretical interrelation-ships (Sheets etal. 1975:378).Considering behavioural units,Arnold (1991:87) distinguishes threeclasses of data that might potentiallyserve to identify Mesoamerican ceramicmanufacture locations and inferredbehaviour from the archaeologicalrecord induding: 1 ) the tools and facili-ties of manufacture, including oftencited "enigmatic firing features"; 2) themistakes and residues of manufacture;and 3) the finished products (see RiceJOURNAL CANADIEN D'ARCHÉOLOGIE 36 (2012)168 • PEURAMAKI-BROWN1996 and Stark 1983 for similar classesof data). This third class of data canprovide a wealth of important informa-tion concerning patterns of distribution,characteristics of consumer populationsand most importantly, information onthe physical manufacture of the items.The ability to recover such diverse datafrom this final category is critical as mostoften it is the third class archaeologistsare left to consider.Any combined analysis of all three ofArnold's classes can therefore potentiallyprovide more in-depth interpretationsregarding the reconstruction of specificprimary and secondary technologi-cal behaviours comprising the chamesopératoires involved in ceramic manufac-ture. A study of these chains thereforerequires multi-step approaches involv-ing both the characterization of finishedproducts and the linking of these itemsto a manufacturing location, associatedtools, and material source on the land-scape.Valle de CacaulapaDuring the summers of 2003 and 2004,archaeological survey, testing, andexcavations were conducted at the Pre-Columbian site of Rancho del Rio in thelower Valle de Cacaulapa of the SantaBarbara District, northwestern Hondu-ras (Figure 1 ). The Valle de Cacaulapa isa narrow valley of the Rio Chamelecónand its tributary, the Rio Cacaulapa, andis located approximately 35 km westof the modern city of San Pedro Sula.Originating in the igneous/volcanic andmetamorphic highlands to the south,the rivers carry water year-round andthe valley itself is situated at the nexusof three geological zones: the Atimalimestone formation to the southeast;Contour lines = 500 m asi50 kmFIGURE i. Map showing location of Rancho del Rio in northwestern Honduras (redrawn fromSmall and Shugar 2004:Figure 1 ).CANADIANJOURNAL OF ARCHAEOLOGY 36 (2012)RURAL CERAMIC MANUFACTURE IN PRECOLUMBIAN HONDURAS • 169igneous rock intrusions and flows to thesouth; and various areas of sand andsiltstone to the north and east (Wells2004:68-69). This unique posirionprovided numerous resources for thePrecolumbian populations of the valley:copper ores, basalt, perlite, limestone,chert, volcanic tuff/ash, and clay-basedsoils. Valley soils are of limestone originand are highly fertile, while other valleysediments are well suited for ceramicproduction (Wells 2004:69).The valley was first extensively sur-veyed in the late 1990s (Urban et al.2000; Wells 2003), at which time 38sites were identified, the largest beingEl Coyote in the southern portion of thevalley. This site covers approximately6 km'-^ and encompasses a minimum of360 platform structures, 28 of whichform the site epicentre. El Coyote wasoccupied from the Middle Preclassic(ca. 600-400 B.C.E.) until the EarlyPostclassic (900-1100 C.E.), with itsstrongest episode of occupation occur-ring in the Late Classic (650-900 C.E.)and Early Postclassic (900-1100 C.E.)periods (Small and Shugar 2004; Urban2007). The rural hinterland sites associ-ated with El Coyote range from hamletscomposed of 22 mounds to mere artifactscatters; Rancho del Rio is includedamong these rural hinterland sites.Rancho del RioThe small site of Rancho del Rio, situ-ated on private farmland immediatelyadjacent to the main highway and aschool yard, was first visited and surveyedby archaeologists in 1997 as part of thelarger valley reconnaissance (Small andShugar 2004). The site consists of atleast seven mounds, ranging in heightfrom under .5 m to 5 m, surrounding alarge central courtyard (Figure 2). Thisconfiguration suggests a small rural set-tlement in the hinterlands of El Coyotelikely inhabited by a loosely connectedgroup of people. The initial investiga-tive season at the site in 2003 aimed tounderstand its physical extent and toconstruct a chronology of its occupa-tion. Test excavations concentrated onthe primary courtyard of the site and thearea to the south of Mound 1 to revealthe relationship of these open areaswith surrounding mounds (Small andShugar 2004). Occupation was found tobe limited to the Late Classic and EarlyPostclassic periods, with the majorityof activity taking place in the former. Avery similar occupation span was notedat many other small hamlet locationsstudied to date in the valley (Urban2007). Excavations in 2004 opened upa large section of the northeast portionof the courtyard (Suboperation B) andMound 7, following discoveries made inthe previous season (Small and Peura-maki-Brown 2004).6\Plaza Excavations¡2004)Midden 2(2003)FIGURE 2. Rectilinear map of Rancho delRío site and location of 2004 excavationsand 200.^5 midden discoveries (redrawn fromSmall and Shugar 2004:Figure 2).JOURNAL CANADIEN D'ARCHÉOLOGIE 36 (2012)170 • PEURAMAKI-BROWNRancho del Rio Ceramic ManufacturingDuring the 2003 testing it was discoveredthat the ancient inhabitants of Ranchodel Rio were possibly engaged in potteryproduction. Large middens, consist-ing primarily of ceramic debris, wereencountered in the courtyard and southof Mound 1. This debris included oppor-tunistic ceramic tools (broken sherdsused as scrapers, shapers, and smoothersin pottery manufacture, similar to thosefound at the site of K'axob in NorthernBelize [Várela etal. 2001:186-187]),vitrified ceramic (occasionally labelledas ceramic "slag"), pieces of clay waste,crushing implements (manos andmetates), countless broken sherds ofthefour paste groups examined in this study,possible temper blocks of a low-gradeschist (discussed below), and potstands.Potstands comprise a distinctive arti-fact class in this region during the LateClassic. They typically conform to a fairlystandard design consisting of a flaringbase surmounted by a straight neck ter-minating in a flat rim, usually fashionedfrom a modified jar rim (Figure 3). Manyare spattered with clay globules towardtheir bases, possible tell-tale signs ofmanufacture. This presence of globules,often multiple layers thick, implies theuse of these artifacts as supportive restsfor vessels while they were still wet andbeing shaped. The baked nature of theclay globules also suggests that the standsheld the shaped vessels during the firingstage, thus serving as a piece of firingfurniture (Urban 2007:44).Similar potstands have been recov-ered at the nearby site of Las Canoasand elsewhere in the valley (Stockett2005:388). Despite excavation of twoLate Classic pottery kilns at La Sierra inthe neighbouring Naco valley, no similarFIGURE 3. Geramic potstand with arrow indicating clay globules on exterior surface (photo-graphed by M. Peuramaki-Brown, 2004).CANADIANJOURNAL OF ARGHAEOLOGV 36 (2012)RURAL CERAMIC MANUFACTURE IN PRECOLUMBIAN HONDURAS • 171artifacts were uneartbed along with thesefiring features (Urban etal. 1997:44).Such a pattern might suggest ceramicswere fashioned at the same time in bothareas, but following different proce-dures within each valley. The potstandstherefore serve a potential direct linkin the chaîne opératoire, from the rawmaterial through the forming stage tothe finished object, as they supply two,and sometimes three, possible identifi-able local fabrics—tbe original pot andthe adhering clay(s). Work with ceramicsherds and potstands recovered from2003 testing and the more extensive2004 excavations at the site strives tocreate a preliminary base from whichinvestigations into the chaînes opératoiresof ceramic manufacture and resultantquestions concerning production mightbe launched.Thin Section PetrographyThe importance of knowledgeof the potter's materials is self-evident. The materials set limitswithin which the potter had towork and the status of the crafthas to be judged witbin tbeselimits. Furthermore, the potter'schoice of materials and the waysin which she used them, togetherwith form and style of decoration,are trademarks—our means, oftenpowerful, of locating centres ofproduction [Shepard 1965: xii].In order to begin the study of ceramicmanufacturing processes it is necessaryto adopt techniques that provide greaterinsight into the physical materials (char-acterization of finished products) thanis provided by traditional macroscopictype-variety systems. Thin section pétro-graphie analysis is a geological techniqueemployed to systematically describe,classify, and identify mineral and rock(Barclay 2001:9; Bishop et al. 1982:285;Gribble and Hall 1992:6; Rice 1987:376;Rye 1981:51-52; Shepard 1965:139; Tite1999:195; Vince 2003). It is derived fromthe broader field of petrology and is con-cerned witb the origin, occurrence, com-position, and history of rocks, incltidingboth chemical and physical-opticalcharacterizations (Rice 1987:376). Thetechnique is not complex but requirestime to prepare thin sections (or, less fre-quently, individual grain mounts [Rice1987:381]), and to perfect identificationskills. Even after many years of applica-tion this technique remains a qualitativeand individual method that varies frompractitioner to practitioner. For a briefhistory of the technique, see referencesin Bishop (1994) and Thompson (1991).The application of pétrographieanalysis to pottery studies is hingedon a view of ceramics as "an anthropo-genic low pressure meta-sediment... anartificial mudstone ... usually contain-ing temper... Typically this material isalso geological, e.g., sand grains, shellsor rock fragments" (Groom 2004;see also Rice 1987:376). Petrographyis applied to ceramic analysis for anumber of potential purposes includ-ing the description and classification offabrics, identification of raw materialsand paste recipes, prediction of rawmaterial source-locations and manufac-ture locales, and technological sttidies(Barclay 2001:10; MacSween 1995:135).Each of these issues cannot always besuccessfully addressed, however, "inregions with a relatively complex igne-ous and metamorphic geology, it isoften possible, in the case of coarse-textured pottery, to suggest the generalarea (or areas) from which the tempercould have originated on the basis ofJOURNAL CANADIEN D'ARCHÉOLOGIE 36 (2012)172 • PEURAMAKI-BROWNpétrographie description of the potteryfabric itself (Tite 1999:196; see alsoVince 1995:121). For an ideal exampleof ceramic pétrographie analysis, seeShepard's (1936) work on the ceramicsfrom Pecos in the southwestern UnitedStates in which she demonstrates thatpottery previously believed to have beenmade locally had actually been traded infrom neighbouring regions.The Pétrographie ProcessPétrographie analysis initially involvesthe preparation of thin sections, a proc-ess involving the removal of a slice ofthe sherd perpendicular to the surfaceusing a tile saw (Barclay 2001:9; Vince2003). If the specimen is found to bebrittie or friable, as is often the case withlow-fired ceramics, it is impregnated withepoxy resin prior to fine grinding andpolishing. The section (roughly 2 cm^)is then glued to a frosted glass slide withepoxy and once dry, is ground to an idealthickness of .03 mm (30 ]im), and sealedwith a thin cover slip using Canadianbalsam sap that allows removal of the slipif desired (Barclay 2001:9-10; Shepard1965:139).^The prepared thin section is thenplaced under a polarizing microscopefor examination. Light is directed tovibrate/travel in a single plane, ratherthan multiple planes as normal by pass-ing through two polarizers (calcite prismreflectors) (Gribble and Hall 1992:1-4;Rice 1987:377-379). When the light ispassed through both prism reflectors(the lower polarizer and the analyzer)the specimen is under a "crossed polar-ized" condition (Rice 1987:377).Since mineral crystals are distin-guished from each other by differentinternal and external morphologies(faces, planes, axes, cleavage points),the transmission of the polarized lightthrough the various mineral crystalsproduces unique images, textures, andcolours that can be interpreted by theanalyst to produce an identification(Gribble and Hall 1992; Mason andBerry 1968:12; Rice 1987:376). Thereare several primary properties examinedas a means of mineral identification:relief, pleochroism, and colouration(under plane polarized light); birefrin-gence, twinning, and isotropism (undercrossed-polarized light); and mineralfeatures including cleavage, fracture,habit (shape), and degradation (Gribbleand Hall 1992:6-15; Rice 1987:378-379;Whitbread 1986:79). Simple chemicalstaining of sections can also be used todifferentiate between minerals of simi-lar optic and pétrographie appearance(e.g., calcite and dolomite) (Barclay2001:9).When dealing with ceramics, pétro-graphie analysis serves to identify natu-rally occurring mineral/rock inclusions(aplastics) in the clay body and may alsolead to the identification of intention-ally added materials known as "tempers"(Barclay 2001:10). This technique servesan ideal middle-ground between mac-rovisual and compositional (elemental)analysis, thus shedding light on suchissues as potting choices, raw materialaccess, and paste recipes. Each aplastictype is identified, using the above men-tioned characteristics, and is countedthrough either a technique known as"point counting" or an area percentageestimation. Research comparing theeffectiveness of both, as well as grainsize estimates, suggest little variationbetween results from each method(Friedman 1958).For the petrofabric descriptionsmade in this study, I expressed aplasticabundance as a percentage of the totalceramic body (due to time constraints).CANADIAN JOURNAL OF ARCHAEOLOGY 36 (2012)RURAL CERAMIC MANUFACTURE IN PRECOLUMBIAN HONDURAS • 173achieved by visually comparing theaplastic grains present in the thin sec-tions with prepared charts illustratingdifferent percentages. Granulometryor textural analysis was performed byobserving grain size, shape, degree ofroundedness and sphericity, degreeof sorting, and particle density foreach aplastic type and determined bycomparison with charts developed bysedimentologists (Figure 4). These char-acteristics can help to further distinguishdifferent petrofabrics represented in aceramic assemblage and may also pro-vide clues concerning the geographicor geologic location of raw materialsources and manufacturing centres aswell as vessel function. Each descriptionconsisted of macroscopic detail (colourbased on the subsurface margins of thesherds, believed to represent the "natu-ral" clay colour [see Rye 1981:119]) anda pétrographie description. Descriptionsare modelled after those of Sunahara(2003).Stages of AnalysisThe following analysis addresses thequestion of pottery manufacture atRancho del Rio during the Late Classic(650-900 C.E.) (for complete analysisresults and interpretations, see Peura-maki-Brown [2004]). Pétrographie anal-ysis was chosen above other analyticaltechniques due to the large number ofaplastics within the pastes. The analysiswas composed of three stages, conform-ing to Arnold's three classes of dataaforementioned.Stage 1: Preliminary Identification (Arnold'sThird Class)Stage 1 consisted of the selection andpettographic description of 20 unslipped(utilitarian) ceramic sherds from fourprominently represented paste groupswithin midden material at the site. TheCacaulapa, San Joaquin, Pueblo Nuevo,and Pitones paste groups were based onthe type-variety-mode system employedby Urban and Schortman for the Valleyof Cacaulapa (Henderson and Beaudry-Corbett 1993). The sample consisted offive jar rim sherds from each of the mac-rovisual paste groups. As differences inpaste can occur between different partsof a vessel (e.g. the base versus the rim),only sherds of the same type (rims) wereselected for the analysis. This ensuredvalid comparisons could be drawn whenthe sherds were observed microscopi-cally•1%10%• • •2%15%5%'••-•' •21»ií50% 75%INCtUSION PERCENTAGEvery well sorted well sortedDEGREE OF SORTINGSUB-ROUNDEDROUNDEDNESSGRITVERY COARSE SANDCOARSE SANDMEDIUM SANDFINE SANDVERY FINE SANDSILT>2.0 mm1.0-2.0 mm05-1.0 mm1/4-1/2 mm1/8-1/4 mm1/16-1/amm<l/16mmFIGURE 4. Granulometr)' visualization charts (redrawn from Groom 2004 based on varioussedimentology source material; no associated figure or page numbers).JOURNAL CANADIEN D'ARCHÉOLOGIE 36 (2012)174 • PEURAMAKI-BROWNThe sherds (including potstandsmentioned below) were systematicallyselected from midden and occupationdebris lots in Suboperation B. These lotswere directly associated with the late LateClassic courtyard surface, which, in turnare associated with the middens idend-fied in 2003. Each sherd was describedmacroscopically, illustrated, photo-graphed, and thin sectioned. This sam-pling strategy was based on suggestionsin Orton (2000) and Schneider (1995)for preliminary pétrographie studies.Time constraints and available fundingfurther restricted the number of samplesanalyzed. Petrofabric descriptions wereconducted based on qualitative analyses:individual aplastic identification; aplasticabundance (as percentage); and granu-lometry (inclusion sphericity, size, anddegree of sorting). Any changes due tofiring were also noted.Stage 2: Potstand Analysis (Arnold'sFirstand Second Classes)Stage 2 consisted of the pétrographieanalysis of clay globules on potstandsfrom Rancho del Rio courtyard debrisand middens. Thin sections were createdfrom five identified potstarid sherds (PI,P2, P3, P4, and P5) with globules repre-senting two macrovisually different resi-dues (large tan-coloured globules and athinner less clumping grey residue) inan attempt to match this manufacturingwaste to one or more of the identifiedpetrofabrics.Stage 3: Clay Sourcing (Arnold's Secondand Third Classes)The final stage consisted of attemptsto identify the environment and/orsources from which the clays and possibletempers used in pottery manufacture atRancho del Rio were obtained. Three claysources were collected, based primarilyon proximity to the site and ease of cur-rent access, and tested through briquettemanufacture and open-pit firing: SourceA consisted of an exposed clay bed on thewest side of the Rio Cacaulapa, southeastof the site; Source B was an exposed claybed from the north side of the Rio Cha-melecón beneath the highway bridgesouth of the site; Source C was an openpit in an escarpment approximately250 m northwest of the site on the oppo-site side of the main highway. The clayfrom this pit is currently mined as washfor bajareque (watde and daub) houses.During Precolumbian times it was possi-bly used in a similar fashion, as well as inceramic production. A fourth briquetteincluded the crushed schist (possibletemper) from courtyard and middentesting and excavations. Thin sectionsof each briquette were examined macro-scopically and petrographically.Results and DiscussionAs is common practice, petrofabrics'are named for their most abundantinclusion (Sunahara 2003:187). As thefive petrofabrics from Rancho del Riohave volcanic ash as their most abun-dant inclusion, I will use this rule foronly three fabrics and will name theremaining two for their next most abun-dant inclusion that distinguishes thesegroups from the others. From the fourtype-variety paste groups representedin the Rancho del Rio ceramic sample,five petrofabric groups were identifiedusing the aforementioned observationcategories: Volcanic Ash 1 (VAl); Vol-canic Ash 2 (VA2) ; Volcanic Ash 3 (VA3) ;Muscovite 1 (Ml); and Muscovite 2 (M2)(Figure 5 and Table 1 ).The clay bodies appear very similarin aplastic content, including volcanicash and many fine particles of muscoviteand polycrystalline quartz, and are likelyCANADIAN JOURNAL OF ARCHAEOLOGY 36 (2012)RURAL GERAMIG MANUFAGTURE IN PREGOLUMBIAN HONDURAS • 175FIGURE 5. Rancho del Rio petrofabrics, plane polarized light lOx magnification (photo-graphed by M. Peuramaki-Brown, 2004).derived from similar environments.They are secondary in nature and likelyof metamorphic and igneous origin,indicated by metamorphic and igneousrock fragments within the clay body(natural inclusions), volcanic ash, andpolycrystalline quartz (MacKenzie andAdams 1994:48,153-155; MacKenzieand Guilford 1980:71). This would seemtypical of river sediments in the area,particularly from the Chamelecón River,which has its headwaters in the igneous/volcanic and metamorphic highlandsto the south. Unfortunately, good geo-logical maps are not currentiy availablefor the valley (the creation of such amap would be a useful future project).Trace amounts of ash in current localday deposits (discussed below) may bea result of the temporal nature of ashdeposits as layers within day sediments.PetrofabricsAlthough the aplastic content appearsto represent a continuum, differencesin ratios of aplastics, grain size, as wellas shape, provide a rationale for theproposed division of the pétrographiegroups. All contain a high percentage ofvolcanic ash, though only VAl appearsto have ash temper. The "crisp" edgesof the ash within the clay body, unlikethe "blended" ash borders seen in VA2,VA3, Ml, and M2, together with thegreat abundance of ash, lends credenceto this premise (for similar observations,see Jones [1991:172]). Volcanic ash is adesirable inclusion because its low levelof thermal expansion and irregular parti-cle shapes allow for stronger bonds withday (Arnold 1991:23-24). During theLate Classic throughout most of lowlandMesoamerica, volcanic ash is a highlyfavoured tempering material (Ford andRose 1995). Therefore it is possible thistemper addition is related to larger world-systems trends. Overall, VAl is extremelyhomogenous when compared with theother petrofabrics in the sample. The rel-atively fine grain size of all inclusions andthe abundance of volcanic ash suggestthat this paste was carefully prepared, pos-sibly by sieving the clay prior to manufac-ture and/or the addition of ash temper.JOURNAL CANADIEN D'ARGHÉOLOGIE 36 (2012)176 • PEURAMAKI-BROWNTABLE I. Comparison of granulometric observationsPetrofabrie AplasticsVolcanic volcanic ashAshlplagioclase^^•mmmm-zi g y p s u mscbistpolyciystallinequartzopaquesbasaltchert^^^^^pp muscovitegrogK^Ç& chloritepseudobone'•" " - " granite/gneissVolcanic volcanic ashAsh2plagioclaseg>'psum''-.'^^^^ scbistpolycrystallinequartzopaquesbasaltchertmtiscovitcgrogcbloritepseudobonegranite/gneissVolcanic volcanic ashAsh 3plagioclasegypsumschistpolycrystallinequartzopaques%391<1<15<1<1<15001<^32313105<1<15103<^3231382Shapeangularsubrounded tosubangularsubangular toangularsubrotinded tosubangularsubangular toangularrotindedsubangularsubangularsubangularangularsubangular toangularangularsubangtilarsubangular toangularsubangularsubangular toangularroundedsubangularsubrounded tostibangularsubangtilarangularangularsubangtilar toangularangtilarsubangtilarsubangular toangularsubangtilarsubangular toangularroundedfor each of the five petrofabrics.Sortingwellpoorlymoderatelypoorlypoorlywellmoderatelymoderatelypoorlywellwellmoderatelypoormoderatelypoorpoormoderatelywellwellmoderately-wellwellwellwellmoderatelymoderatelyinoderatelypoorpoormoderatelyGrain sizefine to meditim sand |Ivery fine to medium sandvery fine to medium sandvery fine to coarse sandvery fine to coarse sand •fine to meditim sandmedium to coarse sandmedium to coarse sandsilt to medium sand ^ji^g;medium to coarse sandmedium to coarsefine to coarse sandvery fine to coarse sandvery fine to medium sandvery fine to very coarse ,isandveiy fine to medium sandvery fine to medium sandmedium to very coarsemedium to coarse sandmedium sandmedium to coarse sandcoarse sand imédium to coarsefine to médium sandvei'y fine to fine sandvery fine to fine sandvery fine to coarse sandvery fine to fine sandvery fine to medium sandCANADIANJOURNAL OE ARCHAEOLOGY 36 (2012)RURAL CERAMIC MANUFACTURE IN PRECOLUMBIAN HONDURAS • 177TABLE I continued.Petrofabric Aplasticsbasaltchertmuscovitegrogchloritepseudobonegranite/gneissMuscovite 1 volcanic ashplagioclasegypsum1,"' 'lî schist*polycrystallinequartz«• at,' opaquesbasaltl^\'" chertmuscovitec hlorite^ i ^ pseudobonegranite/gneissMuscovite 2 volcanic ashplagioclasegypsumschistpolycrystallinequartzopaques^ ^ ¿ basaltcherts K ^ muscovitegrogchloritepseudobone^.h- ' ' •" gi.inite/gneiss%33<103<^251<115157<11152<15<^152110107<1110<1<15<1Shapesubangularsubrounded tosubangularsubangularangularangularsubangular toangularangularangularsubangular toangularsubrounded toangularsubangular toangularroimded tosubangularsubroundedsubroundedsubangular toan0*ularc u JI C^ LA 1 vLlsubangular toangularangularangularsubangular toangularangularangularsubangular toangularsubrounded toangularsubangular toangularsubroundedsubroundedsubroundedsubangular toangularangularangularangularsubangular toangularSortingwellwellwellwellwellwellpoorlypoorlymoderatelypoorlypoorlymoderatelywellpoorlymoderatelymoderatelyvery wellwellwellmoderatelypoormoderatelypoorpoormoderatelywellwellmoderatelymoderatelyvery wellwellwellGrain sizemedium to very coarse *medium to coarse sandfine sandmedium sand^ ^medium to coarse sandmedium to coarsefine to very coarse sandfine to coarse sandfine to coarse sandmedium to very coarsesandvery fine to coarse sandvery fine to medium sandcoarse to very coarse sandfine to coarse sandfine to medium sandfine sand to gritcoarse sandcoarse sand *Ï^^Bmedium to coarsevery fine to medium sandfine to coarse sandfine to coarse sand ,fine to coarse sandvery fine to coarse sandvery fine to medium sandcoarse :coarse to very coarse sandfine sandmedium sandfine to medium sandcoarse sandmedium to coarse |JOURNAL CANADIEN D'ARCHÉOLOGIE 36 (2012)178 • PEURAMAKI-BROWNVA2 and VA3 are similar to VAl intheir high content of volcanic ash andtheir relatively porous body; however,the ash in these petrofabrics is moresimilar to that of Ml and M2. The bor-ders of the ash are blended making itappear to be a natural part of the clay.They also possess a slightly higher per-centage of inclusions (other than ash)when compared with VAl, though lessthan the Muscovite petrofabrics. Theyare slightly coarser than VAl and thelighter colour (7.5 YR 6/3 to 7/2, lightbrown to pinkish gray) of VA2 as com-pared with VAl and VA3 (7.5 YR 6/4 to5 YR 5/4 light brown to reddish brown)may be due to its higher content of mus-covite and polycrystalline quartz. Wlienfired, higher silica contents produce acream-pink to light brown colour (de laFuente 2004:6).While the most abundant inclusion inMl and M2 is volcanic ash, they differ sig-nificantly from their Volcanic Ash petro-fabric counterparts in their very highcontent of muscovite mica, polycrystallinequartz, and micaceous schist. The pres-ence, although small, of chlorite (likelypart of the schist) also distinguishes thesepetrofabrics from the three Volcanic Ashpastes. The abundance of these threeinclusions, as well as their more angularand coarse nattire (characteristic of mate-rials that have been ground up) whencompared with the previous three petro-fabrics suggest their possible addition astemper (Rye 1981:37).During excavations in the courtyard,lumps of easily flaked low-grade meta-morphic micaceous schist were recov-ered in association with other ceramicdebris. Schist blocks were not used inthe architecture of the valley, nor haveknown source outcrops been identifiedto date. The nearby Naco Valley doeshave schist sources and Precolumbianinhabitants used varieties of the stonein architectural construction. As isexplained below, experimentation withthe schist found at the site producedresults similar to those observed in theMuscovite petrofabrics. The possibilityof trade in schist between the two valleysis a subject that deserves future investiga-tion. Finally, the differences in ratios ofmuscovite, polycrystalline quartz, andmicaceous schist between Ml and M2may account for the colour differentia-tion between the two petrofabrics, as itdoes between VA2 and VA3.While grog (crushed, recycledsherds) could add strength to a vessel,due to the angular nature of the crushedparticles, and would have been a readilyavailable material at any ceramic manu-facture location (Jones 1986:20), onlya small amount was present in four ofthe five petrofabrics. These pieces weredistinguished from other argillaceousinclusions by their angular shape, inclu-sions (similar to those of the identifiedpetrofabrics, though the clay bodyappears different), and narrow inter-face between inclusion and clay body(a corona shaped void) (Jones 1986:20;Whitbread 1986). Although it is possiblethat only a small amount of grog waspurposely added to these ceramic fab-rics, perhaps for ritual purposes includ-ing the transference of mana (spiritualforce) as suggested by L. Cecil for somelowland Maya censer pastes (personalcommunication), it is also possible thatthe addition of any grog was uninten-tional. For example, crushed bits ofceramic on work surfaces could havebeen accidentally kneaded into the claybodies during preparation.PotstandsWhen the potstand thin sections wereexamined (Figure 6), the thinner lessCANADIANJOURNAL OF ARCHAEOLOGY 36 (2012)RURAL CERAMIC MANUFACTURE IN PRECOLUMBIAN HONDURAS • 179FIGURE 6. Rancho del Río potstands with arrows indicating globule layers on exteriors, planepolarized light lOx magnification (photographed by M. Peuramaki-Brown, 2004).clumped gray sediment found coat-ing PI, P2, and P3 did not resembleany of the petrofabrics representedin the Rancho del Rio ceramic thinsections. Under the microscope, theresidue appeared to be of silt size grainsinstead of clay. It is possible this siltyresidue was the result of past floodingin areas of the courtyard due to a rise inwater levels in the nearby ChamelecónRiver. The high quantity of calcite inthe silt might be related to the calciteobserved in the clay sources analyzedfrom the valley and linked to the flowof the river through the Atima forma-tion. P4 and P5 were much more typi-cal in appearance for potstands withglobules observed at other sites in thevalley. The similarity of the exteriorglobules with two of the petrofabrics(VA2 and VA3 respectively) repre-sented in the Rancho del Rio ceramicmaterial suggest that manufacture wasindeed occurring at the site; however,it is important to remember these areonly preliminary sttidies and otherpotstands similar to those in this studyshould be analyzed before any firmconclusions are attained.Clay SourcesAlthough none of the clay sources(Figure 7) analyzed were an exact matchto the ceramic petrofabrics, primarilydue to the presence of temper in theceramics and lack of large volcanicash percentages in the sources, somesimilarities were observed. Muscovite,polycrystalline quartz, micaceous schist,plagioclase feldspar, basalt, and gypsumare all rocks and minerals found in boththe Rancho del Rio petrofabrics and thethree clay sources. The lack of ash withinthe sources, except for a small amount inSource B, could be due to the spatial andtemporal distribution of volcanic ashdeposits (correct clay bed not sampled)and addition of volcanic ash as temperto ceramics. If sources from differentgeological times/periods are testedfrom the valley, it is possible volcanic ashdeposits will be located within the sedi-ments. The high content of calcite in allsources (absent in ceramics), hkely dueto river flow through the Atima forma-tion, may also be a temporal character-istic. Overall, the particular mineralsand rocks found within the clay sources(metamorphic and igneous/volcanic),as well as their angularity, suggest a pos-sible match in environmental sourcing.However, there can be no definite con-clusion drawn as analyses of more sourceand sherd samples are required.When some of the easily flaked mus-covite schist (a low-grade metamorphic)JOURNAL CANADIEN D'ARCHÉOLOGIE 36 (2012)180 • PEURAMAKI-BROWNSOURCE ASOURCE BSOURCE C with schist temper1 mmFIGURE 7. Glay sources tested, plane polarized light lOx magnification (photographed byM. Peuramaki-Brown, 2004).that was recovered from courtyardexcavations was crushed and added to atest briquette with Clay Source C, someinteresting observations were made.The overall percentage of muscovite(common in schistose rock), polycrys-talline quartz, and micaceous schistincreased. Also present after the addi-tion of the temper was chlorite, whichis found primarily within the identifiedschist fragments.Numerous observations indicate theaddition of this material as temper byLate Classic potters. First, the presenceof this rock associated with potterysherds, possible ceramic manufacturingtools, and manufacture residue withinthe late facet of the Late Classic court-yard suggest associated use patterns. Theabsence of known outcropping in thevalley and the lack of use of this type ofrock in Late Classic valley architecturesuggest an anomalous use pattern forsuch stone sources. Finally, the increasedpresence of muscovite, polycrystallinequartz, micaceous (muscovite) schist,and chlorite within two of the petrofab-rics from the Rancho del Rio sample andthe angular nature of these inclusionssuggest the addition of schist temper.Why schist was added, given that it hasno known advantage in pottery manu-facture, is uncertain. However, the addi-tional "sparkle" that the muscovite in theschist provides may have been a desiredcharacteristic, exemplified by other highmicaceous content paste groups fromthe valley such as Joya, Monte Redondo,and Minitas (Pat Urban and EdwardSchortman, personal communication2004) and the specular hematite finisheson Coner ceramics at the nearby MayaCANADIANJOURNAL OF ARCHAEOLOGY 36 (2012)RURAL CERAMIG MANUFACTURE IN PREGOLUMBIAN HONDURAS • 181city of Copan (Goodall et al. 2009). Ifwe consider modern uses, muscovite(the predominant mineral in the RDRrecovered schist) is frequently used inwindows for high temperature ovens andas a heat and electrical insulator becauseit is fireproof and withstands extremetemperatures; ideal properties for manyvessel types (Deer et al. 1997:1-55).Finally, the addition of the schist toClay Source C also created a post-firingcolour change (10 YR 7/6 yellow whenwet, 7.5 YR 6/4 light brown post-firing)not observed in tbe briquette made onlyof Clay SourceC (10YR7/6yellow whenwet, 7.5 YR 7/4 pink post-firing). Theaddition of the schist caused a colourchange from a yellow to a light brown,similar to that observed in Ml, M2, andVA2, all of which have more schist, mus-covite, and polycrystalline quartz thanthe other petrofabrics.ConclusionsResults from this study allow the initial"roughing out" of behavioural units andassociated knowledge of the chaînes opéra-toires of ceramic manufacture at Ranchodel Rio. The location and contents ofmiddens associated with a number ofhouse mounds and the central court-yard of the site suggest manufacture ata community level, versus individualhouseholds, perhaps similar to Stark'sethnographic/ ethnoarchaeologicalobservations of "worksbop production"(Stark 1983:160). Such a communitylevel organization of ceramic manufac-ture might also suggest this activity wasintegral to the identity of the overallcommunity, botb economically andsocially, within the larger valley system(Urban 2007:67).Local river-derived clays beds were thelikely source of materials for manufac-ture, suggesting a degree of knowledgeof local environments and sediments.Based on similarities of content charac-teristics between petrofabrics and localsources, as well as ethnographic stud-ies relating the close proximity of claysources to vessel manufacture locations(Arnold 1971, 1985; Stark 1983:164), itis reasonable to assume clays were col-lected near the site with access to claybeds even being controlled by commu-nity residents. Clay Source C is currentlymined as wash for bajareque (wattleand daub) bouses and a cursory pétro-graphie analysis of daub from arcbaeo-logical contexts at tbe site conductedat tbe same time as tbe ceramics of thisstudy shows similar characteristics to thisparticular source.Formation of pots was achieved usingpotstands for at least two of the petro-fabrics, and assisted by the use of oppor-tunistic ceramic tools. The use of thesespecialized ceramic manufacture toolsmay also support a more specializedview of production for the community asopposed to the occasional firing of potsfor individual use. The presence of allmaterials required for paste preparationand vessel shaping within the courtyardarea also suggest unilocal manufacture.Paste recipes appear very similar andwere possibly prepared on hard multi-purpose surfaces based on the occa-sional grog inclusions. Although similar,some recipes vary in terms of the degreeof sifting and temper additions, basedon observations of degrees of sorting ofaplastics and non-local materials foundin excavations and petrograpbic observa-tions. These differences may be linkedto individual potter's styles, desiredvessel performance, or appearance.Regardless, it does suggest cognition ofintended manufacture outcomes.The presence of particular petro-fabrics analyzed in this study at otherJOURNAL CANADIEN D'ARCHÉOLOGIE 36 (2012)182 • PEURAMAKI-BROWNsites in tbe valley may suggest excbangefrom this manufacture locale; however,ceramic manufacture is demonstratedfor many other sites in the valley.Complementary pétrographie data fromother sites is required to further thisdiseussion of individual eommunity dif-ferenees and roles. If fabries are found tobe manufaetured at other rural hamlets,this may suggest that pottei^y manufae-ture was important to individual eommu-nity identity at this time, as opposed tomore overarebing politieal entities thatmay control manufacture and link theiridentities to more specialized forms andiconography (or shared iconographyon widespread forms), as has been sug-gested for tbe neighbouring Naco Valley(Scbortman et al. 2001; Wells 2003). Tbiswould suggest two completely differentbut neighbouring strategies for politicalorganization and manipulation in tbisregion (Urban 2007:67). A comparisonof pastes and manufacturing tecbniquesfrom different sites might highlight avariety of manufacture and productionscenarios, including the sharing orguarding of recipes between communi-ties. Future work should also attemptto address any possible correlationsbetween petrofabric pastes, macrovisualtype-variety groups, and vessel form.Researcb at tbe site of Rancbo del Rioin tbe Valle de Cacaulapa, northwesternHonduras has the potential to uncovercritical information concerning thechaîne opératoire of ceramic manufactureand production in this corner of theMesoamerican world. Beginning such astudy with the analysis of the final stageof manufacture is the most logical start-ing point. Although traditional type-vari-ety systems of ceramic organization doprovide useful information for the studyof ceramics in general, more in-depthpetrological information is requiredfor the study of ceramic manufacture.Information derived from petrographycan complement other forms of ceramictypologies and compositional analysesand might serve to answer questions pre-viously unaddressed by other typologicalclassifications.This study covered three stages ofpétrographie analysis on eeramies andelays from the site and area of Ranehodel Rio: 1) the ereation of petrofabriegroups based on 20 eeramie sherds;2) tbe comparison of clay globules onmanufacture potstands witb the identi-fied petrofabrics; and 3) the compari-son of the petrofabrics witb tbree claysources from the valley. This three-stageapproach allowed a preliminary glimpseat the various steps of the chaîne opératoireof manufacture of particular ceramictypes at Raneho del Rio through theidentifieation and analysis of Arnold'sthree elasses of data. The results encour-age future investigation into ceramicmanufacture and production in this areaof Mesoamerica where "typical" ceramicmanufacture signatures are recovered(debris, tools, etc.) in addition to "atypi-cal" remains (shaping and firing furni-ture) . Further questions may address thetrade of raw materials, such as the schisttemper identified in this study, used inpottery manufacture in the valley, andthe degree of production representedin the archaeological record, usingpetrography to examine the standard-ization of vessel form with petrofabricas well as tbe standardization of tooland potstand forms. Additional studieswould also be encouraged in tbe com-parison of Raneho del Rio petrofabriesand elay sourees petrographieally withthose of other sites in the valley and inneighbouring valleys. The results wouldfurther eontribute to our understandingof inter- and intra-valley relationsbipsCANADIANJOURNAL OF ARCHAEOLOGY 36 (2012)RURAL CERAMIC MANUFACTURE IN PRECOLUMBIAN HONDURAS • 183regarding ceramic production in thisarea, and possibly within Mesoamericain general.Acknowledgements. Thank you to my supervi-sory committee at The Institute of Archaeol-ogy, University College London, in particularDrs. Bill Sillar and Elizabeth Graham. TheInstituto Hondureno de Antropología eHistoria granted permission for the exportof samples and field excavations. Dr. DavidSmall (Lehigh University) and Dr. AaronShugar (The Smithsonian Institution) co-directed the Rancho del Rio project (2004)and encouraged this research topic. Fundingwas provided by Lehigh University and theInstitute of Archaeology, University CollegeLondon. Drs. Edward Schortman and Patri-cia Urhan (Kenyon College) were invaluablesources of information and guidance. RobIxer and Simon Groom instructed me inpétrographie methods and provided initialguidance. Thank you to Maria, Glenn, Cait-lin, Hannah, and Shawn. Thank you to LeslieG. Cecil, Ellen S. Moriarty, and one anony-mous reviewer for their helpful comments.Any mistakes and/or inconsistencies withinthis work are those of the author.Notes1. Rice (1996:167) distinguishes betweenmanufacture, "the actual act of fab-ricating ceramics", and production,"the social and economic organiza-tional arrangements within whichpottery manufacture is carried out."2. Initial wet tile sawing, epoxy impreg-nation, and slide mounting wasconducted by the author in the labo-ratories of the Institute of Archaeol-ogy, UCL. Final sample grinding tothe .03 mm desired thickness andpolishing was conducted by trainedgeological technicians.3. The term petrofabric is used todescribe ceramic samples that sharesimilar characteristics when viewedunder a pétrographie microscope(Mason and Tite 1994:20; Sunahara2003:187).References CitedArnold, Dean E.1971 Ethnomineralogy of Ticul,Yucatan Potters: Etics and Emics.American Antiquity 36:20-40.1985 Ceramic Theory and Cultural Pro-cess. Cambridge University Press, Cam-bridge, England.Arnold, Philip J.1991 Domestic Ceramic Production andSpatial Organization: A Mexican CaseStudy in Ethnoarchaeology. CambridgeUniversity Press, Cambridge, England.Barclay, Katherine2001 Scientific Analysis of Archaeologi-cal Ceramics: A Handbook of Resources.Oxbow Books, Oxford, England.Bishop, Ronald L.1994 Pre-Columbian Pottery:Research in the Maya Region. InArchaeometry of Pre-Columbian Sites andArtifacts, edited by David A. Scott andPeter Meyers, pp. 15-65. The CettyConservation Institute, Los Angeles,California.Bishop, Ronald L., Robert L. Rands,and George R. Holley1982 Ceramic Compositional Analy-sis in Archaeological Perspective. InAdvances in Archaeological Method andTheory, Vol. 5, edited by Michael B.Schiffer, pp.275-330. Academic Press,New York.Bleed, Peter2001 Trees or Chains, Links orBranches: Conceptual Alternatives forConsideration of Stone Tool Produc-tion and Other Sequential Activities.JOURNAL CANADIEN D'ARCHÉOLOGIE 36 (2012)184 • PEURAMAKI-BROWNJournal of Archaeological Method andTheory 8:101-127.Cresswell, Robert1976 Avant-Propos. Techniques etCulture 1:5—6.de la Fuente, Guillermo A.2004 Technological Characteriza-tion of Inka and Pre-Inka Pottery: ACeramic Petrology Approach. The OldPotter's Almanack 12 ( 1 ) : 1-14.Deer, William A., Robert A. Howie, andJ. Zussman1997 Rock-Forming Minerals, Vol. 3A:Micas. The Geological Society of Lon-don, London, England.Ford, Anabel, and William I. Rose1995 Volcanic Ash in Ancient MayaCeramics of the Limestone Lowlands:Implications for Prehistoric VolcanicActivity in the Guatemala Highlands.Journal of Volcanology and GeothermalResearch 66:\A9-\63.Ford, James A., and Julian Steward1954 On the Concept of Types. Ameri-can Anthropologist 56:42-57.Friedman, Gerald M.1958 Determination of Sieve-Size Dis-tribution from Thin-Section Data forSedimentary Petrological Studies. TheJournal of Geology 66:394-416.Goodall, R. A., J. Hall, R. Viel, andP. M. Fredericks2009 A Spectroscopic Investigationof Pigment and Ceramic Samplesfrom Copan, Hondurjis. Archaeometry51:95-109.Gribble, Colin D., and Allan J. Hall1992 Optical Mineralogy: Principles andPractice. UCL Press, London, England.Groom, Simon2004 Introducing Ceramic Petrog-raphy 1: Approaching Ceramics asMetasediments, Handout, Practical 1,25 February 2004. Institute of Archae-ology, UCL, London, England.Hassan, Fekri A.1988 Prolegomena to a Grammati-cal Theory of Lithic Artifacts. WorldArchaeology 19:281-296.Henderson, John S., and MarilynBeaudry-Corbett (editors)1993 Pottery of Prehistoric Honduras:Regional Classification and Analysis.Monograph Vol. 35. Institute ofArchaeology, University of California,Los Angeles, California.Jones, Lea D.1986 Loivland Maya Pottery: The Placeof Petrological Analysis. BAR Interna-tional Series 288. British Archaeologi-cal Reports, Oxford, England.1991 Tempering Trends in LowlandMayan Pottery. In Recent Develop-ments in Ceramic Petrology, edited byAndrew Middleton and Ian Free-stone, pp. 165-182. Occasional PaperNo. 81. British Museum Press, Lon-don, England.Lemonnier, Pierre1986 The Study of Material CultureToday: Toward an Anthropology ofTechnical Systems. yoMT7iö;/ of Anthropo-logical Archaeology 5:147-186.1992 Elements for an Anthropology ofTechnology. Anthropological PapersNo. 88. Museum of Anthropology,University of Michigan, Ann Arbor,Michigan.MacKenzie, W.S., and A. E. Adams1994 A Colour Atlas of Rocks and Miner-als in Thin Section. Manson Publishing,London, England.MacKenzie, W.S., and C. Guilford1980 Atlas of Rock-Forming Minerals inThin Section. Longman Group Lim-ited, Essex, England.CANADIANJOURNAL OF ARCHAEOLOGY 36 (2012)RURAL GERAMIG MANUFAGTURE IN PREGOLUMBIAN HONDURAS • 185MacSween, Ann1995 Fabric Analysis of Coarse WareAssemblages: Examples from ThreeSites in the Northern Isles of Scot-land. In The Aim of Laboratory Analysesof Ceramics in Archaeology, edited byAnders Lindahl and Ole Stilborg,pp. 127-138. KVHAA Konferenser,Vol. 34, Almqvist and Wiksell Interna-tional, Stockholm, Sweden.Mason, Brian, and Leonard G. Berry1968 Elements of Mineralogy.W. H. Freeman, San Francisco, Cali-fornia.Mason, R.B., and M.S. Tite1994 The Beginnings of StonepasteTechnology. Archaeometry, 36:77-91.Orton, Clive2000 Sampling in Archaeology. Cam-bridge University Press, Cambridge,England.Peuramaki-Brown, Meaghan2004 The Chaîne Opératoire ofCeramic Manufacture and Production:Preliminary Analysis through CeramicPetrography at Rancho del Rio, Vallede Cacaulapa, Santa Barbara District,Honduras. Unpublished Master's the-sis. University College London, Insti-tute of Archaeology, London, England.Rice, Prudence M.1987 Pottery Analysis: A Sourcebook.University of Chicago Press, Chicago,Illinois.1996 Recent Ceramic Analysis: 2.Composition, Production, and The-ory. Journal of Archaeological Research4:165-202.Robin, Cynthia2002 Outside of Houses: The Prac-tices of Everyday Life at Chan Nôo-hol, Belize. Journal of Social Archaeology2:245-268.Rye, Owen S.1981 Pottery Technology: Principles andReconstruction. Manuals on Archaeol-ogy Vol. 4. Taraxacum, Washington,D.C.Schiffer, Michael B.1976 Behavioral Archaeology. AcademicPress, New York.Schneider, Gerwulf1995 A Short Note on Project Plan-ning and Sampling for LaboratoryAnalysis of Archaeological Ceram-ics. In The Aim of Laboratory Analysesof Ceramics in Archaeology, edited byAnders Lindahl and Ole Stilborg,pp. 23-27. KVHAA Konferenser,Vol. 34, Almqvist and Wiksell Interna-tional, Stockholm, Sweden.Schortman, Edward M., Patricia A.Urban, and Marne Ausec2001 Politics with Style: Identity For-mation in Prehispanic SoutheasternMesoamerica. American Anthropologist103:312-330.Sheets, Payson D., B.W. Anthony,David A. Breternitz, David S. Brose,B.K. Chatterjee, Carl B. Compton,Scott Cook, Richard G. Forbis,Alexander Gallus, Thomas R. Hester,Fumiko Ikawa-Smith, Maxine R.Kleindienst, Carl-Axel Moberg,Hansjürgen Müller-Beck, Jon Muller,J.-Ph. Rigaud, Michael W. Spence,Robin Torrence, F. Van Noten,Marcus C. Winter, and John Witthoft1975 Behavioral Analysis and theStructure of a Prehistoric Industry.Current Anthropology 16:369-391.Shepard, Anna O.1936 The Technology of Pecos Pot-tery. In The Pottery of Pecos: Vol. II, byAlfred V. Kidder and Anna O. Shepa-rd, pp. 389-587. Papers of the PhillipsAcademy Southwestern ExpeditionJOURNAL CANADIEN D'ARGHÉOLOGIE 36 (2012)186 • PEURAMAKI-BROWNNo. 7. Yale University Press, NewHaven, Connecticut.1965 Ceramics for the Archaeologist. 5thed. Publication 609, Carnegie Institu-tion of Washington, Washington, D.C.Sinclair, Anthony2000 Constellations of Knowledge:Human Agency and Material Affor-dance in Lithic Technology. In Agencyin Archaeology, edited by Marcia-AnneDobres and John Robb, pp. 196-212.Routledge, New York.Small, David, and Meagban Peuramaki-Brown2004 Excavaciones Arqueológicas enRancho del Rio, Santa Barbara. Paperpresented at the VIII Seminario deAntropologia Hondurena, Teguci-galpa, Honduras.Small, David, and Aaron Shugar2004 National Science Founda-tion Grant Proposal: Rancho delRio 2004. Electronic document,http://www.lehigh.edu/~inarch/,accessed May 15, 2004.Stark, Barbara L.1983 Archaeological Identificationof Pottery Production Locations:Fthnoarchaeological and Archaeologi-cal Data in Mesoamerica. In DecodingPrehistoric Ceramics, edited by Ben A.Nelson, pp. 158-194. Southern IllinoisUniversity Press, Carbondale, Illinois.Stockett, Miranda K.2005 Approacbing Social Practicetbrough Access Analysis at Las Cano-as, Honduras. Latin American Antiquity16:385-407.Sunahara, Kay2003 Pétrographie Analysis of Ceram-ics from the Belize Valley. Unpub-lished Ph.D. dissertation. Departmentof Anthropology, McMaster University,Hamilton, Ontario.Thompson, Raymond H.1991 Shepard, Kidder, and Carn-egie. In The Ceramic Legacy of Anna O.Shepard, edited by Ronald L. Bishopand Frederick W. Lange, pp. 11-41.University Press of Colorado, Niwot,Colorado.Tite, Michael S.1999 Pottery Production, Distribu-tion, and Consumption: Tbe Contri-bution of the Physical Sciences. Jour-nal of Archaeological Method and Theory6:181-233.Urban, Patricia A.2007 Rural Production in NorthwestHonduras: The 2004 Season of theLower Cacaulapa Valley Archaeolo^'^'^^^^cal Project. Report submitted to theFoundation for the Advancement ofMesoamerican Studies. Electronicdocument, http://www.famsi.org/reports/03030/03030Urban01 .pdf,accessed June 2011.Urban, Patricia A., Edward M.Schortman, and Marne Ausec2000 Tbe Proyecto Valle de Cacaula-pa, Northwestern Honduras: Archaeo-logical Investigations ConductedDuring the 1999 Field Season. Manu-script on file. Instituto Hondureno deAntropología e Historia, Tegucigalpa,Honduras.Urban, Patricia A., Christian Wells, andMarne Ausec1997 The Fires Without and the FiresWithin: Evidence for Ceramic Produc-tion Facilities at the Late Classic Siteof La Sierra, Northwestern Honduras,and in its Environs. In The Prehistoryand History of Ceramic Kilns, edited byPrudence Rice, pp. 173-194. Ceram-CANADIANJOURNALOE ARCHAEOLOGY 36 (2012)RURAL CERAMIC MANUFACTURE IN PRECOLUMBIAN HONDURAS • 187ics and Civilization, Vol. 7, AmericanCeramic Society, Columbus, Obio.Várela, Sandra L. López, Patricia A.McAnany, and Kimberly A. Berry2001 Ceramics Technology at LateClassic K'axob, Belize. Journal of FieldArchaeology 28:177-191.Vince, Alan1995 Integrating Ceramic Petrol-ogy and Urban Archaeology: A CaseStudy. In The Aim of Laboratory Analysesof Ceramics in Archaeology, edited byAnders Lindabl and Ole Stilborg,pp. 121-126. KVHAAKonferenser,Vol. 34, Almqvist and Wiksell Interna-tional, Stockbolm, Sweden.2003 Ceramic Petrology: An Introduction.Electronic document, bttp://www.postex.demon.co.uk/petrology.htm,accessed May 15, 2004.Wardle, Peter1992 Earlier Prehistoric Pottery Produc-tion and Ceramic Petrology in Britain.BAR British International Series225. British Archaeological Reports,Oxford, England.Wells, Christian2003 Artisans, Chiefs, and Feasts:Classic Period Social Dynamics atEl Coyote, Honduras. UnpublishedPh.D. dissertation. Department ofAnthropology, Arizona State Univer-sity, Tempe, Arizona.2004 Investigating Activity Patterns inPrehispanic Plazas: Weak Acid-Extrac-tion ICP-AES Analysis of Anthrosols atClassic Period El Coyote, Northwest-em Honduras. Archaeometry 46:67-84.Whitbread, I.K.1986 The Characterisation of Argil-laceous Inclusions in Ceramic ThinSections. Archaeometry, 28:79-88.Manuscript received June 8, 2011;Final revisions December 12, 2011.JOURNAL CANADIEN D'ARGHÉOLOGIE 36 (2012)Copyright of Canadian Journal of Archaeology is the property of Canadian Archaeological Association and itscontent may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder'sexpress written permission. However, users may print, download, or email articles for individual use. | RESEARCH GATE |
https://doi.org/10.1177/14732254211004764Youth Justice 1 –17© The Author(s) 2021Article reuse guidelines: sagepub.com/journals-permissionsDOI: 10.1177/14732254211004764journals.sagepub.com/home/yjjSentencing explanations provided via judicial remarks made within the English magistrates’ youth court: Towards a better global understandingMax LowensteinDepartment of Humanities and Law, Bournemouth University, UKAbstractThis article qualitatively explores the English judicial approach towards sentencing explanations via remarks made within the magistrates’ youth court. First, the extent of their correlation with the three known purposes behind sentencing explanations is considered within a wider introductory discussion. Second, judicial interviews provide new insights regarding the extent of their alignment with the introductory discussion by indicating degrees of correlation. Third, the English judicial approach towards sentencing explanations and the degrees of correlation are concluded upon. Finally, recommendations are made to assist in a better understanding of sentencing explanations globally, particularly in jurisdictions where their publication has increased.KeywordsYouth, sentencing, explanation, remarks, criminal, judiciary, England, globalIntroductionHailed for promoting young adults’ (18–25-year-olds’) sentencing transparency (Howard League for Penal Reform, 2017), English young adult sentencing explanations have expanded, from their increasing written publication online since 2014 and their increas-ing video publication since 2020 under the Crown Court (Recording and Broadcasting) Order 2020.1 Under Section 51(4), the Coronavirus Act 2020 encourages live courtroom links to assist English young adult sentencing processes and judicial role play, which have been disrupted by COVID-19 – but not for children (10–17-year-olds). Meanwhile, Corresponding author:Max Lowenstein, Bournemouth University, Fern Barrow, Poole, Dorset, BH12 5BB, UK. Email: mlowenstein@bournemouth.ac.uk1004764 YJJ0010.1177/14732254211004764Youth JusticeLowensteinresearch-article2021Original Article2 Youth Justice 00(0)English youth sentencing explanations provided to children are not published and remain limited to the private arena of the magistrates’ youth court, where judicial remarks are delivered face-to-face to children, their parents and youth workers, bucking the increas-ing publication trend of sentencing explanations for young adults. To better understand this, this article attempts to answer the following focal question: What is the English judicial approach towards youth sentencing explanations made within the magistrates’ youth court? To answer this question, the three theoretically defined purposes behind them (Weijers, 2004: 30) – related to the relevant legal obligations, ideological perspec-tives, and understandings of childhood and moral judgements – are drawn on to provide a wider introductory discussion. Second, building on this, the English judicial approach towards youth sentencing explanations made within the magistrates’ youth court is explored further through normative interview data findings. Third, the introductory dis-cussion and findings are concluded upon and the next steps towards a better global understanding are considered.Legal obligationsInternational and domestic children’s standards place some legal obligations on English judges as they try to understand the human behaviours exhibited by children when sen-tencing in the magistrates’ youth court (Weijers, 2004: 30). The international standards, despite being ‘incomplete’, ‘are the common language of youth justice’ (Kilkelly, 2008: 191). English judges should ensure that their sentencing communications provide: ‘(1) child-friendly information, (2) child participation in proceedings and (3) child-friendly remedies’ (Liefaard, 2019: 216). The effective implementation of sentencing communica-tions at sentencing depends on nuanced interpretations between the child and the adult judge, testing judicial communication skills. While understanding the former matters, unlike the latter, children do not create and are not obligated to improve their sentencing communications, justifying the judicial focus here. Globally, judicial interpretations of these standards and their effective implementation face ‘many challenges’ (Kilkelly and Liefaard, 2019: 617). In England, out of court, youth offending service managers have ‘asserted their own priorities and objectives in seeking to deliver effective services’ (Smith and Gray, 2019: 567), while, in court, child-friendly justice and enhanced child participa-tion at sentencing require improvement in relation to ‘what children want’ (Daly and Rap, 2019: 15) and recognition of their ‘gendered, cultural and material inequalities’ (Phoenix, 2018: 18).The relevant international legal obligations outlined in the United Nations Convention on the Rights of the Child are legally binding, albeit not on the courts. Under Article 37(b), English judges are obligated to protect a child’s liberty following their ‘arrest, detention or imprisonment’, ensuring custodial sentences remain as a ‘last resort and for the shortest appropriate period of time’. Furthermore, under Article 40(1), English judges are obligated to treat children ‘in a manner consistent’ with promoting their ‘dignity and worth’, recog-nising their positive ‘reintegration’ and ‘constructive role’ within society, while, under Article 40(4), English judges are obligated to apply a variety of youth sentencing options that are ‘appropriate to [the child’s] well-being and proportionate both to their Lowenstein 3circumstances and the offence’ (United Nations, 1989). Child-friendly justice under Rule 14 of the Beijing Rules guides English judges to manage courtroom procedures in a way that promotes a child’s ‘free expression’ and ‘participation’ (United Nations, 1985), while a ‘child-centred orientation’ under Rule 3 of the Riyadh Guidelines (United Nations, 1990) recommends a child’s ‘constructive reintegration’ into society, reaffirming the global com-mitments of the Convention on the Rights of the Child (Goldson, 2019: 210–211).Domestic legal obligations when sentencing children are placed on English judges externally via legislation and sentencing guidelines, and internally via judicial precedent in case law and judicial training publications. Externally, the principal legislative basis comes from both Section 44 of the Children and Young Persons Act 1933, which requires judges to ‘preserve the welfare of young offenders’, and Section 37(1) of the Crime and Disorder Act 1998, which requires judges to ‘prevent re-offending’. The Sentencing Council for England and Wales issues offence-specific sentencing guidelines, which English judges ‘must’ follow, unless it is contrary to ‘the interests of justice’ under Section 125(1) of the Coroners and Justice Act 2009. The sentencing steps include, first, the offence category, which ranges from high to low seriousness based on the young offend-er’s culpability and offence harm (Sentencing Guidelines Council, 2004: 3). Second, the starting points and category range provide a non-exhaustive list of typical aggravating and mitigating factors (Sentencing Council, 2017: 15–16). The final step is the sentencing explanation, where the judge ‘must’ give clear reasons for their sentence ‘in ordinary lan-guage and in general terms’ under Section 52 of the Sentencing Act 2020.Internally, English judges are guided by case-law precedents set in higher courts and judicial training publications issued by the Judicial College, Magistrates Association and Ministry of Justice. Most youth offending is of low to medium seriousness and is sentenced by a panel of three lay magistrates, with one as chairperson, in the youth courts (Ministry of Justice, 2020: 14). These magistrates receive specific youth court training to be ‘active listeners’ who adopt ‘age-appropriate language’ and promote ‘individual responsibilities in preventing further offending behaviour’ (Judicial College, 2017: 106, 104). The magistrates’ court sentencing guidelines, issued online by the Sentencing Council, are child-centred (Judicial College, 2020: 46) and welfare-focused (Sentencing Council, 2017: 5–6), and ensure that parental responsibilities are supported (Ministry of Justice, 2007: 4–5). To summarise, the legal obligations of child-friendly justice, participation, constructive societal reintegration and welfare appear to be important.Ideological perspectivesAdopting simplified language within English youth sentencing explanations (Weijers, 2004: 30) can be related to the ideological perspectives of judicial legitimacy and experi-ence. Judicial legitimacy at sentencing has attracted metaphorical, reformative and socio-legal control critiques (Carlen, 1976: 48). With regard to the latter, primarily sociological understandings of the law (Weber, 1978: 311–312) have in recent times increasingly tried to connect various perspectives of social theory with legal practice (Wandall, 2008: 10–13). For Luhmann (1993), judicial legitimacy requires judges to justify the law 4 Youth Justice 00(0)appropriately, which judges attempt to do in their sentencing explanations, the impact of which can be a measure of ‘functional success’ (Cotterrell, 2006: 24), while Weber (1968: 874–875) views judicial legitimacy in sentencing processes that culminate in sentencing explanations as being ‘the product or the technical means of a compromise of interests’.Judicial attempts to demystify their sentencing requirements in Habermas’s ‘lifeworld’, where Luhmann’s (1993) functional systems involve everyday social experiences, result in the development of ‘background knowledge’ (Habermas, 1996: 23). Judges, in creating and then providing their sentencing explanations to young offenders, self-legitimise and objectively rationalise their efforts based on their own subjective sentencing experiences. Habermas (1996) further argues that the legal regulation of contemporary society requires the ‘lifeworld’ to be legitimate. He appears to suggest that judicial consensus and con-formity-building contribute to the everyday social experience of judges, which he labels communicative rationality. Communicative rationality defines law and morality as dis-tinct and is itself justified by judicially controlled legal processes (Cotterrell, 2006: 26). To summarise, the ideological perspectives that consider judicial legitimacy and judges’ sentencing experience appear to be important.Understandings of childhood and moral judgementsJudicial difficulties in understanding children in England can be exacerbated by ‘marginal participation’ with other courtroom actors (Weijers, 2004: 23). Procedural delays can dis-rupt trial proceedings and lead to frustrated children at sentencing; some have reported that their lives have been ‘unnecessarily disrupted’ (Taylor, 2016: 30). Children’s ‘active involvement in the trial process’ can be inherently limited (Bevan, 2016: 5) and they may report confusion as to ‘who to trust, what to do, how to do it, when and why?’ (Howard League for Penal Reform, 2018: 62). Judicial engagement with children, their parents and youth services can determine ‘how young people’s experiences and views are heard, and what is heard’ (Drake et al., 2014: 38). It is important to overcome these judicial difficul-ties when setting moral goals at sentencing to help children reform (Weijers, 2004: 30).Judicial attempts at moral goal-setting for reform assume individual child autonomy, and it ‘benefits the person who will experience it, as a way of helping him to gain moral knowledge, if he chooses to listen’ (Hampton, 1984: 214). Judicial moral messages can involve ‘a concrete moral goal, which punishment should be designed to accomplish, and that goal includes the benefiting of the criminal himself’ (Hampton, 1984: 215). If they regularly appear in the form of judicially led sentencing explanations, they may legiti-mately be dismissed as a ‘moral limit to the law’ (Edmundson, 1990: 505). However, for moral education theorists, judges would then risk appearing ‘indifferent to the people they punish’ (Hampton 1984: 219).According to Haidt’s (2003) theoretical exposition of universal moral foundations, judicial moral goal-setting (a) can be related to characteristic emotions (b) and punish-ment outcomes (c). Applying Haidt’s exposition of universal moral foundations to judges setting moral goals for reform includes discussions with the child regarding the harm(s) caused (a), which is related to characteristic emotions (b) and, if associated with genuine remorse, includes ‘intentions and often actions’ (Maslen, 2015: 5–6; for a list of remorse Lowenstein 5emotions, Maslen specifically refers to Smith, 2002: 98–99). Moral foundations (a) and characteristic emotions (b) then relate to the punishment given (c), which may be a Parenting Order that has been judicially drafted (c) (Canton, 2015: 67). To summarise, the understandings of childhood and moral judgements which appear to be important are the judicial difficulties in understanding children and moral goal-setting to help children reform.Sample and designThe research started in 2013 with a literature review and informal discussions with magis-trates already known to the researcher, which informed a qualitative interview guide. The proposed interview guide, qualitative methodology and safeguarding of the sample’s data were then ethically reviewed and approved at the institutional level. Funding to cover travel and institutional transcription costs was acquired in 2014. Judicial access, which required a formal interview at the Royal Courts of Justice, London, received approval from the then Senior Presiding Judge, Lord Justice Gross, in 2014. Informal telephone and/or email discussions with magistrate court managers, as well as individual magistrates, assisted the researcher’s dissemination of the interview guide. These discussions facilitated the magistrates’ understanding of the scope of the research and helped promote the informed consent of the sample before agreeing to be interviewed. The researcher flexibly engaged with the busy judicial sample over four years from September 2014 to March 2018. As an experienced qualitative interviewer of the English judiciary, there was an acute awareness of the potential time, cost and judicial access difficulties (Lowenstein, 2013: 33).Furthermore, this reliance on the engagement of the sample over time meant that the researcher could not ensure that the sample was representative of the wider population. To boost the sample size, it was necessary to take time to arrange and complete all 24 inter-views at the participants’ convenience. However, while this increased the risk of delays to data gathering, it increased the opportunity for judicial engagement with the researcher. The researcher also spent time in developing greater familiarity with all of the interview-ees via phone calls or emails before the interviews took place. This helped the researcher’s attempts to explain the interview guide and gain the trust and respect of the sample (Steinke, 2004: 185). The researcher reassured the sample that all of the face-to-face inter-views would, first and foremost, be focused on and always led by the sample (Harding, 2019: 75). By giving the sample the space and time to lead and be in control of their own interviews, the breadth and depth of the interviews was maximised (King et al., 2010/2019: 66). Each interview lasted 45 minutes on average, which appeared to reflect a high degree of judicial interest in engaging with the interview guide.Respecting the anonymity and confidentiality of the judicial sample, the qualitative interview guide only recorded their gender, age and length of bench time served. This was gathered from each judicial interviewee in order merely to define the overall sample char-acteristics. A quantitative analysis of these characteristics could not be expected to yield any detailed information regarding youth sentencing explanation practices. This signifi-cant limitation further justified why a qualitative method and analysis were considered necessary and adopted as the primary mode of analysis (Kleining and Witt, 2001: 6).6 Youth Justice 00(0)Overall, in terms of gender, there were 14 males and 10 females, 13 of whom were regular chairpersons. This meant that a relatively equal gender balance was present. Overall, in terms of age, perhaps unsurprisingly all of the 24 judges interviewed were between 41 and 70. By age bracket, 5 were aged 41–50, 9 were aged 51–60 and 10 were aged 61–70, suggesting that they were predominantly at a senior stage in their sentencing careers. The seniority of the English judicial volunteers was also reflected in their overall length of bench experience, where 11 had served for 7–12 years, 9 had served for 13–18 years and 4 had served for 19–24 years.The eight youth court areas were chosen, first, as they were all geographically proxi-mate, saving time and costs, as well as being reasonably accessible and familiar to the researcher. Second, the court areas provided a helpful snapshot of youth court magistrates serving in the south of England. Third, they could each be distinguished by using a quan-titative approach in terms of rural and urban development by gathering the publicly avail-able population statistics. The eight magistrates’ courts selected served the areas around Folkestone (Kent), Brighton (East Sussex), Worthing (West Sussex), Portsmouth and Southampton (Hampshire), Bournemouth (Dorset), Exeter (Devon) and Truro (Cornwall). However, beyond these eight selected areas, magistrates serving Guildford (Surrey) also volunteered to become involved. Overall, the researcher noted that the 24 youth court magistrates who sentenced youth offenders were largely based in urban areas, which were surrounded by rural areas that they also served (Pateman, 2011: 7).Method and analysisAs other qualitative researchers will know, there is no standard method for analysing the complex responses captured during interviews. Each magistrate received the interview guide and consent form via post or email, which required their formal agreement before an interview could take place. The interview guide contained the questions and methodol-ogy. Although the opportunity was there before the interviews for the magistrates to debate the meaning of the interview guide, no detailed explanation from the researcher was requested before the interviews. Prior to the interviews, all 24 magistrates confirmed that they fully understood the interview guide and the methodology applied, and wel-comed that they were open to their interpretation. This meant that, at interview, they were free to, and did, spontaneously lead their interview. This approach allowed the judicial interviewees to debate their own perceptions freely and to control their own interview (Foddy, 1993: 22).The qualitative analysis of the judicial perceptions gathered at interview encompassed three stages. First, the researcher, in reading the full interview transcripts, identified simi-lar and repeated meanings in the judicial responses. This helped to condense the volume of the qualitative data. Identifying common experiences, feelings and values helped to reveal shared thematic understandings (King et al., 2010/2019: 299). The similar shared perceptions of the sample formed the basis of the researcher’s understanding, which was supported by the selection of explanatory direct quotations from the sample.Second, the level of analysis turned to ad hoc meaning-generation, which by its own flexible nature allowed the researcher to understand the deeper data semantics. This Lowenstein 7approach recognised that an interpretation of meaning and language required recognition of the academic perspective adopted. In this case, it was socio-legal, which incorporated the interviewer’s subjective interpretations of what could be positive or negative, similar or different. The researcher’s perspective could then be interpreted for, and later validated by, a readership of judges, academics, legal professionals, politicians and the public. The ad hoc meaning-generation approach involved an attempt by the researcher to work out metaphors that captured the meaning of the full transcripts (Kvale, 1996: 203–204).Third, the researcher recognised and embraced their own subjective bias when explain-ing the interview questions to the interviewees. This also included their subjective inter-pretations during the qualitative analysis of the data gathered. This significantly reduced the objective validity of the data examined. However, this could be justified as an alterna-tive and innovative way to develop an in-depth understanding of the judicial perceptions gathered, by treating them as narrative interview data (Kvale, 1996: 236).FindingsLegal obligationsThe findings have been structured thematically, as in the preceding wider introductory discussion, to discuss the relevant legal obligations, ideological perspectives, and under-standings of childhood and moral judgements. First, judicial alignment to the legal obliga-tions of child-friendly justice, participation, constructive societal reintegration and welfare was discussed with the judicial sample. Each of these legal obligations was spontaneously recognised as relevant – particularly welfare – by the whole sample. When asked to iden-tify which legal sources referred to them most, the whole sample similarly agreed that it was domestic sources, and that they had the most influence, with sentencing guidelines having the highest relevance, followed by judicial training publications and, lastly, legis-lation. As one of the magistrates summarised:We must safeguard the well-being of young people first and foremost. We want them to feel included during proceedings and at sentencing, by discussing the ways in which they need to reform. This is written into our sentencing guidelines, which guide us and which we are mandated to follow.For the three magistrates who referred to judicial training publications as most relevant for their legal obligations, this was based on the detailed guidance provided by the ‘Youth court bench book’ (Judicial College, 2017) and ‘Equal treatment bench book’ (Judicial College, 2020). Child-friendly justice and participation with children were encouraged within their specialist youth court training when understanding behaviour. As one of the three magistrates explained:Our youth court training is extensive. We receive many case examples that contain youth health and well-being issues, which we discuss together in training. Their life experiences, from mental health, limited communication abilities, drug and alcohol abuse to suffering from bullying or poor and neglectful parenting, are very relevant to us.8 Youth Justice 00(0)The importance of constructive societal reintegration and safeguarding child welfare was discussed. When protecting a child’s health and well-being, 22 magistrates spontaneously shared the view that the young offender, their parents or primary caregiver, as close sup-porters, mattered most, followed by youth services interventions as detailed in their pre-sentence reports. This was based on who the magistrates most perceived young offenders had regular involvement with and their expected greater influence on the young offenders’ day-to-day life. As one of these 22 magistrates concluded:Establishing those who are closest and thereby have influence is crucial to protecting their welfare and providing them with the encouragement required to improve their life. Parents will nearly always be closest as they have raised them and have known them the longest.However, while 22 magistrates provided similarly positive accounts of how proactively questioning and listening to the young offender and their close supporters assisted their understanding of the offender’s health and well-being needs, for two magistrates, the lim-ited assistance of close supporters could hinder their constructive societal reintegration and welfare. For them, negative accounts were spontaneously shared of persistent serious offending and strained family relationships, which harmed the young offender’s welfare and hindered their reform, despite the best efforts of their youth worker to support them. As one of these magistrates reflected:We make a point of asking young offenders if they need anything to better support them, which improves our understanding. We closely work with their youth worker and parents. However, where there is persistent offending and few close family bonds, perhaps due to local authority care, from experience we know it will be much harder for us to help them and, more importantly, for them to help themselves.The judicially perceived extent of expected child reform from their youth sentencing explanations further revealed the importance of child-friendly justice, participation, con-structive societal reintegration and welfare. To initially help the magistrates frame their discussion of the extent of the expected reform, the whole sample completed a Likert attitudinal scale before the interviews, ranging from a very high to a very low expected reform impact from their sentencing explanations. Of the 24 magistrates, 5 responded with very high, 12 with high, 7 with moderate and none with low or very low reform impact. For the 17 magistrates who felt they had a very high to high reform impact, they spontaneously related child-friendly discussions more with first-time offenders. They described first-time offenders similarly, who they felt were more likely to engage with them through receptive body language. This was attributed to maintaining eye contact and listening, which suggested there was genuine remorse and engagement with their sentenc-ing explanations. They similarly described how they facilitated reform through positive and individually tailored child-friendly discussions, ensuring the offender’s participation and mapping their path to constructive societal reintegration while safeguarding their wel-fare needs. As one of these 17 magistrates described:To achieve meaningful engagement and promote their future reform, we tend to ask personal questions, such as ‘What do you need to reform?’ and ‘Who is around most to help you to Lowenstein 9improve your life?’ To help them to think more about their reform, we can make suggestions, such as getting psychological support, developing their confidence by going to school or college, and trying to remove themselves from bad influences such as peer pressure, often exerted in criminal gangs.The remaining seven magistrates, who felt that they had a moderate reform impact, spon-taneously related child-friendly discussions more with persistent young offenders, who had frequent previous convictions that had increased in seriousness. They similarly described that this was evident from the facts of the case, with escalating seriousness fea-turing high levels of offender culpability and harm. Such persistent young offenders were negatively perceived as less likely to participate with them, through unreceptive body language, which was commonly attributed to limited eye contact and poor listening skills. These magistrates agreed that they were more likely to encounter failure in their attempts to map a path for the constructive societal reintegration of persistent young offenders with complex welfare needs. As one explained:As I encourage reform, I have found that they have less and less impact as the custody cusp is reached. With such offenders, there is little eye contact and little care for those around them or the victims they have harmed. I will include their youth worker and their pre-sentence reports revealing the reasons for their reoffending, current health and well-being. I will try to include their parents, family and friends who support them.Ideological perspectivesSecond, the relevant ideological perspectives, which consider judicial legitimacy and their sentencing experience, were discussed. In these discussions, the whole sample spon-taneously agreed that they, first, ensured young offenders had a clear understanding regarding their sentence requirements, as a matter of judicial legitimacy, and, second, recognised the limited ability of children to fully understand them, in their sentencing experience. As one of the magistrates remarked:I don’t think you will get much disagreement on their primary function. We are duty-bound to adopt simplified language that they are familiar with. We check their understanding by asking questions like ‘Do you fully understand this?’ ‘Would you like me to explain that more?’To successfully enhance their judicial legitimacy, 21 magistrates spontaneously referred most to their frequently positive, inclusive and successful sentencing explanations. The magistrates directed them mostly towards the young offender and, to a lesser extent, their close supporters – that is, their parents and youth worker in the courtroom. They similarly described, from their sentencing experience, the importance of protecting each young offender’s welfare with the assistance of youth services and their parents. As one of the magistrates concluded:It is crucial that every young person fully engages and understands their sentence. We constructively engage with them, so they are crystal clear, and involve their family and youth worker in our discussions about their well-being. Their parents and youth worker can help 10 Youth Justice 00(0)support them beyond the courtroom, but it is the young person themself who is most responsible for adhering to the requirements of our order.However, for the remaining three magistrates, they spontaneously referred most to infre-quent neutral, limited or unsuccessful sentencing explanations that hampered their judi-cial legitimacy. The magistrates had similarly directed them mostly towards the young offender’s youth worker and parents, or, to a lesser extent, the child in their sentencing experience. Further researcher prompting was required to ascertain why. They also referred to challenging circumstances where the young offender had a very limited under-standing of their sentence requirements. Due to this, further intervention and guidance from their close supporters within the courtroom were required to ensure their full under-standing of the sentencing requirements and protect their future welfare. As one of these magistrates revealed:Parents and youth workers can greatly assist us with some children who are really struggling to understand their sentence and their own health issues over time. These children need significant therapeutic intervention. Sadly, if they don’t interact well with it, when appointments with their youth offending team become sporadic and parental supervision is poor, they are much less likely to complete their sentence.How each judge predominantly expressed their sentencing explanations was discussed. The whole sample responded with ambiguity regarding what their principal mode of expression could mean. Interviewer prompting was therefore necessary to further define it as either positive or negative words and actions. A Likert attitudinal scale was pre-sented, which ranged from a predominantly positive to negative expression of their sen-tencing explanations. Fifteen magistrates responded with a predominantly positive, nine with a moderately positive, and none with a neutral, moderately negative or negative mode of expression. The whole sample shared a similar consensus that their mode of expression in words and actions was overwhelmingly positive in their sentencing experi-ence. As one of the magistrates explained:It really has to be positive, doesn’t it? We try to softly encourage by our tone of voice and by looking at the young person to better understand the consequences of their actions in order to begin changing themselves. To achieve this, they must be honest with us about themselves and thoughtfully process the hurt their offending has caused to their life and those closest to them.The reasons why the whole sample similarly described their mode of expression in words as either positive or moderately positive were focused on. For 21 magistrates, their posi-tively expressed words mostly coalesced around providing simplified phrases which suc-cinctly explained their sentencing requirements, thereby enhancing their judicial legitimacy. As one remarked:To be clear, you must get to the point quickly with no legal jargon to successfully connect with them. We largely focus on what we feel they will need to know, such as what they will do and when.Lowenstein 11The other three magistrates similarly felt that their words mostly coalesced around utter-ing reassuring phrases that explained their sentencing requirements thoroughly, thereby enhancing their judicial legitimacy. As one of them commented:When I speak, I choose my words very carefully, as each word can be understood differently. Taking my time does slow me down, but sometimes they say they understand when they don’t out of embarrassment. That is why encouraging them and anyone else with them to ask questions helps give them greater ownership of their sentence.The reasons why the whole sample described their mode of expression in actions as either positive or moderately positive were focused on. The whole sample similarly related posi-tive actions in their past sentencing experience, where their pace and tone of voice, as well as body language, had assisted them. For 22 magistrates, their positive actions were mostly static with them commonly adopting a slow pace, a calm and measured tone of voice, and open body language, while maintaining eye contact to engage with a receptive young offender. As one of them explained:Over many years, I have learnt that, as a panel, we should avoid showing any agitation or lack of patience with them in both our voice and body language. To remain positive and inclusive, I take my time, do not raise my voice and look at who I am talking with to ensure we are developing a meaningful connection with them, their parents and youth worker.For the other two magistrates, their positive actions remained mostly fluid and commonly involved adopting a moderate pace, an assertive tone of voice and closed body language, while limiting their eye contact to adapt to an unreceptive young offender. As one of these magistrates described:Looking back, I have noticed that repeat youth offenders at sentencing are more likely to look down with crossed arms or legs. We ask them to uncross them and look up, but sometimes they just won’t listen and don’t want to be there. In those circumstances, we are left with no alternative than to simply reel off the requirements of our order, leaving the details for their parents and youth worker to go through with them later when they are in a better frame of mind to receive them.Understandings of childhood and moral judgementsThird, the relevant understandings of childhood and moral judgements, from judicial dif-ficulties in understanding children to moral goal-setting for reform, were discussed. The whole sample spontaneously agreed that, with regard to the former, what mattered most was asking questions that explored each child’s vulnerabilities, utilising their sentencing experience. The child vulnerabilities that were most often recognised by the magistrates included physical and mental health disabilities, developmental maturity and communica-tion difficulties, as well as past child abuse or trauma. As one of the magistrates explained:We ask questions to facilitate open discussions about the difficulties they are facing in life. They may not be attending school and have learning difficulties affecting that. They may have increasingly common mental health issues such as anxiety and depression. They may have 12 Youth Justice 00(0)suffered past trauma and abuse that can lead to self-harm and periods of isolation. Being aware of these complex difficulties is crucial for us and we are reliant upon their legal counsel, parents and youth worker for that important information.The whole sample was asked whether morality featured in their youth sentencing explana-tions, and there was agreement by all of the judges that it did. Further researcher prompt-ing was required to help the judges explore the morality within their youth sentencing explanations. The relevant morality depended, first, on relating to the case facts and set-ting moral goals for reform, which the magistrates commonly felt improved each young offender’s understanding of their own culpability. Second, it depended on relating to those perspectives that each young offender was closest to, amplifying discussions around offence harm and developing genuine remorse for the harm caused. When setting moral goals for reform through their offence-harm discussions, the magistrates similarly dis-cussed both morally right and wrong values, behaviours and associated emotions with the young offenders, their parents and youth worker. As one of the magistrates identified:Of course, there is morality in what we do. In reshaping their morals, we discuss wrongs and rights. If they have assaulted someone, for example, in that moment of aggression they have hurt someone else, which is wrong. We want them to develop a sense of regret about being aggressive and better manage their anger by avoiding situations that make them anxious or upset leading to violence, which is right.How best to address judicial difficulties in understanding children and set moral goals for reform attracted two different schools of thought amongst the sample as to which court-room perspective(s) mattered most. For 18 magistrates, their focus mostly centred on recognising each young offender’s unique perspective when setting moral goals for reform. They commonly described how they, first, attempted to ascertain the young offender’s existing moral values; second, attempted to correct any immorality detected; and, third, guided them towards setting their own moral goals for reform. As one of these magistrates discussed:I mainly seek to explore what actions the young person can do to make amends for what they have done and change their moral code. So, for example, where they have been antisocial, perhaps causing damage to other people’s property, they can agree to be more considerate and thoughtful in future and voluntarily fix or compensate them. If they are involved in knife or drugs crime, they can agree that carrying a knife or taking drugs is very dangerous for them and those around them.For the remaining six magistrates, their focus mostly centred on recognising the perspec-tives of the young offender’s close supporters when setting moral goals for reform. They commonly described how they, first, gathered the moral perspectives of the parents and youth worker; second, used these perspectives to help correct any immorality they detected in the young offender; and, third, included them in setting moral goals for reform. The magistrates similarly felt that this inclusive approach assisted them most in developing within the young offender a greater awareness and recognition of the harm their offending had caused. As one of these magistrates explained:Lowenstein 13We try to keep their parents and youth worker fully involved by asking them what they think about the offending. They help us as we suggest ways in which the young person can improve their behaviour. If they are stealing, their parents and youth worker can encourage them to stop being dishonest and become trustworthy. Coming from someone close to them, that can be really effective in recalibrating their moral compass.Exploring how the judges amplified offence-harm discussions and developed genuine remorse for harm caused when setting moral goals required further prompting from the interviewer because offence harm itself relates to varying levels of offence seriousness in England. The interviewer therefore suggested, and the whole sample agreed to, a common sentencing context of both an early guilty plea and a low to medium seriousness offence focus for their offence-harm discussions. There were two different schools of thought within the sample as to which courtroom perspective attracted their principal focus.For 18 magistrates, their offence-harm discussions when moral goal-setting were mostly directed at the individual young offender, utilising their sentencing experience. Within the courtroom, they similarly discussed how they, first and foremost, encouraged genuine remorse for the harm caused from the young offender. For these magistrates, genuine remorse was commonly expressed through the young offender’s words and changes in their body language as they expressed feelings of regret for the harm caused. As one of the magistrates remarked:When we first discuss the harm that they have caused and their regret for it, our focus is on them. Their body language can be poor, looking down a lot and keeping their hands in their pockets. We ask them: ‘How do you feel about your offending?’ ‘Why are you doing it?’ ‘What do you regret most about it?’ By engaging with them directly, their body language often improves, they look and sit up, and their shame for what they have done seems to be expressed with more sincerity.For the other six magistrates, their offence-harm discussions when setting moral goals were mostly directed at the young offender’s close supporters – typically their parents and youth worker – utilising their sentencing experience. Their most successful moral goal-setting attempts commonly involved thoughts and feelings about the harm caused to their close sup-porters in order to help encourage genuine remorse in the young offender. For these magis-trates, too, genuine remorse was commonly expressed through words, changes in body language and feelings of regret for the harm caused. As one of these magistrates discussed:What is genuine remorse? Looking back, I have found that discussing the hurt caused to their nearest and dearest family members can really help young people empathise with others and feel sorry for their offending, if they don’t already do so. In one case, only after his mother had spoken about the hardship his offending had caused her as a single parent did his demeanour improve. His eye contact increased, he sat up with his arms uncrossed and his apology really appeared to be much more heartfelt.ConclusionsThis article has posed and then attempted to answer the following focal question: What is the English judicial approach towards youth sentencing explanations made within the 14 Youth Justice 00(0)magistrates’ youth court? In the sample and design, the randomness of the opportunity sampling adopted was limited by seeking the most committed and engaged judicial vol-unteers. Furthermore, in the findings, the judicial interviews could only provide insights into the magistrates’ perceptions of their sentencing remarks, rather than the actual prac-tice of sentencing explanations. Overall, the high degree of judicial concordance in the findings indicates the relevance of the wider introductory discussion, where the three theoretically defined purposes behind sentencing explanations have some relevance (Weijers, 2004: 30). However, ascertaining whether most English judges follow a similar predefined script or operate within a narrow, perhaps erroneous framework of understand-ing children’s offending when explaining their sentences requires more data. Should their judicial publication ever occur in England, or expand across offences, as has been the case for young adults, this will become more attainable.First, the findings indicate that the relevant legal obligations of child-friendly justice, participation, constructive societal reintegration and, particularly, welfare appear to be important, as English judges try to understand the human behaviours exhibited by chil-dren in their courtrooms. Judicial understandings of each child’s health and well-being needs appear to be important, but thus far academics have been reluctant to provide pre-scriptive guidance for their youth sentencing remarks in England due to the ‘characteris-tics of the individual child and the complexity and circumstances of each case’ (Hollingsworth and Stalford, 2020: 4–5). Determining whether creating and sharing pre-scriptive sentencing explanation guidance would promote judicial understandings of each child, and protect children’s distinct participatory rights to be included and human rights to be treated equally at sentencing (Lam, 2013: 151), would benefit from more empirical data-gathering, but this could be developed.Second, the findings indicate that the relevant ideological perspectives, which consider judicial legitimacy and their sentencing experience, appear to be important, as English judges use simplified language in their sentencing explanations for children. Accessing magistrates’ sentencing experiences of language simplification to develop prescriptive sentencing explanation guidance capturing best practice would require a ‘specific judicial investment in tailoring judgments for children, responding to children’s uniquely and uni-versally disenfranchised position within the legal process’ (Stalford and Hollingsworth, 2020: 1057). The data presented here suggests that this may be enhanced by greater judi-cial attempts to provide positive sentencing explanations that include parents and youth workers.Third, the findings indicate that the relevant understandings of childhood and moral judgements that appear to be important are English judges’ difficulties with understand-ing children and their setting of moral goals to assist each child’s reform. With regard to the former, each child’s vulnerabilities were judicially recognised, including their ‘chronic inability to successfully navigate particular sorts of interaction’ (Johnson, 1993: 76, referring to Habermas’s communicative action concept). With regard to the latter, moral goals to help children reform involved offence-harm discussions with judges, who sought to develop genuine remorse as part of their moral education attempts both within and beyond the youth courtroom. Developing prescriptive sentencing expla-nation guidance incorporating offence-specific moral goals for children could improve Lowenstein 15transparency and be part of ‘the hallmarks of a good judgment more generally’ (Stalford and Hollingsworth, 2020: 1058).To improve our understanding of youth sentencing explanations globally, particularly in jurisdictions where their publication has been expanding, the next steps involve ‘finer-grained (international and intra-national/sub-national) critical inquiry’ (Goldson, 2019: 209). Such inquiries may be qualitative surveys or interviews from the judicial perspec-tive to better understand the theoretically defined purposes behind them, with courtroom observations and case transcripts acting as alternatives to explore the judicial lens (Bouhours and Daly, 2007). They could consider the impact of youth sentencing explana-tions in and beyond the courtroom from the perspectives of the child, their parents, legal counsel and youth worker to test whether youth sentencing explanations actually provide ‘clarity, caring, integrity and are being well informed’ (Hollingsworth, 2020: 6). Their impact beyond the courtroom can involve children’s relationships and interactions with their parents and youth workers, and may include their life development via their future ‘careers and trajectories’ (Evans, 2006: 148). Improving our understanding of their pur-pose and impact globally should prompt comparative inquiries across different criminal justice systems, despite the ‘risks of ethnocentrism and relativism’ (Nelken, 2010: 18–19) and uncritical ‘criminological tourism’ (Smith, 1995/2014: 219). Moving towards a better global understanding of youth sentencing explanations, jurisdictions where the publica-tion of judicial remarks has been increasing are likely to prove most fruitful when enhanc-ing our juvenile justice conversations in this emerging research field.FundingThe Socio-Legal Studies Association Small Research Grant Scheme supported this research in 2014–2015.ORCID iDMax Lowenstein https://orcid.org/0000-0003-2648-8959Note1. See the Courts and Tribunals Judiciary website at https://www.judiciary.ukReferencesBevan M (2016) Effective participation in the youth court. Howard League What Is Justice? Working Papers 21/2016. Available at: https://howardleague.org/wp-content/uploads/2016/04/HLWP_21_16.pdf (accessed April 2020).Bouhours B and Daly K (2007) Youth sex offenders in court: An analysis of judicial sentencing remarks. Punishment and Society 9(4): 371–394.Canton R (2015) Crime, punishment and the moral emotions: Righteous minds and their attitudes towards punishment. Punishment and Society 17(1): 54–72.Carlen P (1976) The staging of magistrates’ justice. British Journal of Criminology 16(1): 48–55.Children and Young Persons Act 1933. Available at: https://www.legislation.gov.uk/Coronavirus Act 2020. Available at: https://www.legislation.gov.uk/Coroners and Justice Act 2009. Available at: https://www.legislation.gov.uk/Cotterrell R (2006) Law, Culture and Society: Legal Ideas in the Mirror of Social Theory. Aldershot: Ashgate.Crime and Disorder Act 1998. Available at: https://www.legislation.gov.uk/Crown Court (Recording and Broadcasting) Order 2020. Available at: https://www.legislation.gov.uk/16 Youth Justice 00(0)Daly A and Rap S (2019) Children’s participation in youth justice and civil court proceedings: Have states made sufficient progress? In: Kilkelly U and Liefaard T (eds) International Human Rights of Children. Singapore: Springer, 299–319.Drake D, Fergusson R and Briggs D (2014) Hearing new voices: Re-viewing youth justice policy through practitioners’ relationships with young people. Youth Justice 14(1): 22–39.Edmundson WA (1990) Liberalism, legal decisionmaking, and morality ‘as such’. Oxford Journal of Legal Studies 10(4): 505–521.Evans J (2006) Welfare versus punishment? The careers of young offenders with a background in public care. PhD Thesis, Cardiff University, UK. Available at: https://orca.cf.ac.uk/54336/1/U584249.pdf (accessed September 2019).Foddy W (1993) Constructing Questions for Interviews. Cambridge: Cambridge University Press.Goldson B (2019) Juvenile Justice in Europe: Past, Present and Future. London: Routledge.Habermas J (1996) Between Facts and Norms: Contributions to a Discourse Theory of Law and Democracy. Cambridge: Polity Press.Haidt J (2003) Elevation and the positive psychology of morality. In: Keyes CL and Haidt J (eds) Flourishing: Positive Psychology and the Life Well-Lived. Washington, DC: American Psychological Association, 275–289.Hampton J (1984) The moral education theory of punishment. Philosophy and Public Affairs 13(3): 208–238.Harding J (2019) Qualitative Data Analysis: From Start to Finish. London: SAGE.Hollingsworth K (2020) Sentencing remarks for children: A new approach. Newcastle Law School Research Briefing no. 14. Available at: https://www.ncl.ac.uk/media/wwwnclacuk/newcastleuniversitylawschool/files/JANUARY%202020%20Research%20briefing%20sentencing.pdf (accessed February 2020).Hollingsworth K and Stalford H (2020) Judgements for children. Newcastle Law School Research Brief-ing no. 18. Available at: https://a10-vip-wwwprod.ncl.ac.uk/media/wwwnclacuk/newcastleuniversity-lawschool/files/Briefing%20(KH)%20-%20Judgments%20for%20Children.pdfHoward League for Penal Reform (2017) Judging maturity: Exploring the role of maturity in the sentencing of young adults. Available at: https://www.t2a.org.uk/wp-content/uploads/2017/07/JudgingMaturity.How-ardLeague.pdf (accessed September 2019).Howard League for Penal Reform (2018) Children and Sentencing: A Guide for Adults Supporting Children Facing Sentence in the Criminal Courts in England and Wales. London: Howard League for Penal Reform. Available at: https://howardleague.org/wp-content/uploads/2018/03/D_ADULT_Guide.pdf (accessed April 2019).Johnson J (1993) Is talk really cheap? Prompting conversation between critical theory and rational choice. American Political Science Review 87(1): 74–86.Judicial College (2017) Youth court bench book. Available at: https://www.sentencingcouncil.org.uk/wp-content/uploads/youth-court-bench-book-august-2017.pdf (accessed January 2019).Judicial College (2020) Equal treatment bench book. Available at: https://www.sentencingcouncil.org.uk/wp-content/uploads/Equal-Treatment-Bench-Book.pdf (accessed March 2020).Kilkelly U (2008) Youth justice and children’s rights: Measuring compliance with international standards. Youth Justice 8(3): 187–192.Kilkelly U and Liefaard T (2019) International children’s rights: Reflections on a complex, dynamic, and rela-tively young area of law. In: Kilkelly U and Liefaard T (eds) International Human Rights of Children. Singapore: Springer, 617–627.King N, Horrocks C and Brooks J (2010/2019) Interviews in Qualitative Research. 2nd ed. London: SAGE.Kleining G and Witt H (2001) Discovery as basic methodology of qualitative and quantitative research. Forum: Qualitative Social Research 2(1): Article 16.Kvale S (1996) Interviews: An Introduction to Qualitative Research Interviewing. London: SAGE.Lam C (2013) Childhood, Philosophy and Open Society. Dordrecht: Springer.Liefaard T (2019) Access to justice for children: Towards a specific research and implementation agenda. International Journal of Children’s Rights 27(2): 195–227.Luhmann N (1993) Das Recht der Gesellschaft. Frankfurt am Main: Surhkamp Verlag.Lowenstein M (2013) Towards an understanding of judicial denunciation – Relating theory to practice by comparing the perceptions of English and Danish lower Court judges when sentencing minor theft offenders. Criminology and Criminal Justice 13(1): 21–35. Lowenstein 17Maslen H (2015) Remorse, Penal Theory and Sentencing. Oxford: Hart.Ministry of Justice (2007) Parenting contracts and orders guidance. Available at: https://dera.ioe.ac.uk/7949/7/parenting-contracts.pdf (accessed May 2020).Ministry of Justice (2020) Youth justice statistics 2018/19: England and Wales. Available at: https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/862078/youth-jus-tice-statistics-bulletin-march-2019.pdf (accessed April 2020).Nelken D (2010) Comparative Criminal Justice: Making Sense of Difference. London: SAGE.Pateman T (2011) Rural and urban areas: Comparing lives using rural/urban classifications. Regional Trends 43, Office for National Statistics. Available at: https://www.ons.gov.uk/ons/rel/regional-trends/regional-trends/no–43–2011-edition/rural-and-urban-areas–comparing-lives-using-rural-urban-classifications.pdf (accessed December 2018).Phoenix J (2018) A child-friendly youth justice? In: Bateman T, Goodfellow P, Little R, et al (eds) Child-Friendly Youth Justice? London: National Association of Youth Justice, 15–19. Available at: https://thenayj.org.uk/wp-content/uploads/2019/02/NAYJ-Child-friendly-youth-justice-May-18.pdf (accessed September 2020).Sentencing Act 2020. Available at: https://www.legislation.gov.uk/Sentencing Council (2017) Sentencing children and young people: Overarching principles and offence specific guidelines for sexual offences and robbery. Definitive guideline. Available at: https://www.sentencingcouncil.org.uk/wp-content/uploads/Sentencing-Children-and-young-people-Definitive-Guide_FINAL_WEB.pdf (accessed February 2019).Sentencing Guidelines Council (2004) Overarching principles: Seriousness. Available at: https://www.sen-tencingcouncil.org.uk/wp-content/uploads/Seriousness-guideline.pdf (accessed January 2019).Smith A (2002) The Theory of Moral Sentiments. Cambridge: Cambridge University Press.Smith R (1995/2014) Youth Justice: Ideas, Policy, Practice. 3rd ed. London: Routledge.Smith R and Gray P (2019) The changing shape of youth justice: Models of practice. Criminology and Crimi-nal Justice 19(5): 554–571.Stalford H and Hollingsworth K (2020) ‘This case is about you and your future’: Towards judgments for children. Modern Law Review 83(5): 1030–1058.Steinke I (2004) Quality criteria in qualitative research. In: Flick U, Von Kardoff E and Steinke I (eds) A Companion to Qualitative Research. London: SAGE, 184–190.Taylor C (2016) Review of the Youth Justice System in England and Wales. London: Ministry of Justice. Available at: https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/577105/youth-justice-review-final-report-print.pdf (accessed March 2019).United Nations (1985) United Nations standard minimum rules for the administration of juvenile justice (‘The Beijing Rules’). Available at: https://www.ohchr.org/documents/professionalinterest/beijingrules.pdf United Nations (1989) Convention on the Rights of the Child. Available at: https://www.ohchr.org/en/profes-sionalinterest/pages/crc.aspx United Nations (1990) United Nations guidelines for the prevention of juvenile delinquency (The Riyadh Guidelines). Available at: https://www.ohchr.org/en/ProfessionalInterest/Pages/PreventionOfJuvenile-Delinquency.aspxWandall RH (2008) Decisions to Imprison: Court Decision-Making Inside and Outside the Law. Aldershot: Ashgate.Weber M (1968) Economy and Society: An Outline of Interpretive Sociology. Berkeley: University of Cali-fornia Press.Weber M (1978) Economy and Society: An Outline of Interpretive Sociology, vol. 1. Berkeley: University of California Press.Weijers I (2004) Requirements for communication in the courtroom: A comparative perspective on the Youth Court in England/Wales and The Netherlands. Youth Justice 4(1): 22–31.Author biographyDr Max Lowenstein is a Senior Lecturer in Law at Bournemouth University, UK. His research interests encompass sentencing law, comparative law, criminology and criminal justice. | RESEARCH GATE |
Simulation and educationComparison of different methods of more effectivechest compressions during cardiopulmonaryresuscitation (CPR) in the dental chairTakashi Hitosugi a,*, Norimasa Awata b, Yoichiro Miki c, Masanori Tsukamoto d,Takeshi Yokoyama aAbstractIntroduction: When performing cardiopulmonary resuscitation (CPR) on a patient who has suffered a cardiopulmonary arrest during dental treat-ment, few dental chairs have sufficient stability to perform effective chest compressions. We previously proposed a method of stabilizing the backrestof a dental chair using a support stool. As a result, we confirmed that the vertical displacement of the backrest could be significantly reduced. In thepresent study, we verified the effectiveness of the stool stabilization method using several dental chairs (flat and curved) with significantly differentbackrest shapes.Methods: Vertical displacement of the backrests of dental chairs was recorded. Data were obtained at three different stool positions (without a stool,under the chest at the level that participants were performing manual chest compressions, and under the shoulders). Reduction displacement ratioswere calculated to evaluate the effectiveness of the stool positions.Results: The method significantly reduced the vertical displacement of the backrest for all types. When the curvature of the backrest was large, thereduction in vertical displacement was 40% when the stool was placed under the chest at the level of manual chest compressions and 65% whenplaced underneath the shoulder. In the case of a flat dental chair, this reduction was 90% when using a stool in either position, compared to no stool.Conclusion: When we need to perform CPR on a patient in the dental chair, placing a stool under the shoulders allows effective manual chestcompression by firmly supporting the backrest of a dental chair of any shape.Keywords: Cardiopulmonary resuscitation (CPR), Manual chest compression, Dental chair, Stool position, DentalIntroductionThe positioning of a stool underneath a dental chair to stabilize it inorder to perform an effective manual chest compression site hasbeen reported previously.1,2 This technique is intended to providegreater stability by reclining the dental chair and placing a stool underthe backrest so that it is in contact with the back of the chair. It wasadopted in the 2015 and 2021 European Resuscitation Council(ERC) guidelines.3,4 However, we found that the support effect var-ied depending on the position of the stool. For dental chairs with aheavily curved back outline, a stool placed under the shoulder com-pared to under the manual chest compression site significantlyreduced vertical displacement, resulting in a large reduction ratio.2However, that study was based on only one type of dental chair.The purpose of this study was to compare the effectiveness ofthree stool positions (no stool, under the manual chest compression,and under the shoulder) in stabilizing three types of dental chairs withsignificantly different external shapes.https://doi.org/10.1016/j.resplu.2022.100286Received 5 June 2022; Received in revised form 25 July 2022; Accepted 26 July 2022Available online xxxx2666-5204/� 2022 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).Abbreviations: CPR, Cardiopulmonary resuscitation, ERC, European Resuscitation Council, AHA, American Heart Association* Corresponding author.E-mail address: hitosugi.takashi.724@m.kyushu-u.ac.jp (T. Hitosugi).R E S U S C I T A T I O NP L U S 1 1 ( 2 0 2 2 ) 1 0 0 2 8 6Available online at www.sciencedirect.comResuscitation Plusjournal homepage: www.elsevier.com/locate/resuscitation-plusMethodsEthical statementsThis was a manikin study and there were no human participants. TheInstitutional Review Board at Kyushu university confirmed that noethical approvals were required.Study design and settingThree dental chairs (#1: EOMaⅡ�; GC, Tokyo, Japan, #2: EOM-PLUS SS�; GC, Tokyo, Japan, #3: EOM P�; GC, Tokyo, Japan)were used in this study. #1 exterior has a severely curved backrestexterior, #2 and #3 have a flatter shape.The study procedure was performed according to a previouslyestablished method in which the CPR manikin (Resusci Anne� TorsoBasic version 2011; Laerdal Medical AS) was positioned on thereclined dental chair with the upper end of the manikin torso alignedwith the top edge of the backrest in Fig. 11,2. The superior surface ofthe backrest under the lower half of the manikin’s sternum was posi-tioned horizontally using a leveling instrument. The hand position formanual chest compressions was the center of the chest. A metalindicator (point P) was secured to the inferior surface of the dentalchair directly under the area for manual chest compressions andmade parallel to the floor with a level gauge. The distance of pointP relative to the inferior surface of the backrest remained fixed forthe duration of the study. Displacement of point P was capturedusing video recordings and measured while each health care provi-der performed manual chest compressions on the resuscitation man-ikin. Manual chest compressions depth was kept between 5.1 to6.0 cm during the study. The actual manual chest compressionsdepth was evaluated with the manikin‘s Skill-Reporter system whichhas a green light that indicates manual chest compressions depths of5.1 to 6.0 cm. Any compressions outside of that range (i.e., withoutthe green light) were excluded. When compression depths werewithin 5.1 to 6.0 cm, the vertical displacement of the backrest fromits initial position was recorded and included for analysis. Video datawere transferred to a computer, and the backrest vertical displace-ment measurements were determined using the simultaneously cap-tured ruler for reference. The stabilizing stool placed under thebackrest of the dental chair for this study was round with a hard seat-ing surface (diameter 30 cm; height 45 cm; FB-01ALLBK, Fuji BoekiCo., ltd.). The superficial edge of the stool’s seat was set to verticallycontact the backrest either just under the area for manual chest com-pressions or under the shoulders.OutcomesThree American Heart Association-trained BLS providers (A: 47-year-old male, 175 cm, 93 kg; B: 44-year-old male, 177 cm, 60 kg;C: 44-year-old female, 157 cm, 50 kg.) were tested at their conve-Fig. 1 – Manikin setup and positioning of the stabilizing stool. The upper end of the manikin torso was aligned withthe top edge of the backrest (A; red line). The superior surface of the backrest under the lower half of the manikinsternum was positioned horizontally using a leveling instrument. The edge of the stool’s seating surface was set totouch the backrest vertically under the area for manual chest compressions (A; blue line). The stool was set to touchthe backrest vertically under the shoulders (B; green line). The center of the manikin’s chest (C; red ellipse) was thehand position during chest compressions. Measuring vertical displacement and chest compression depth. A metalindicator (D: point P) was made parallel to the floor with a leveling gauge and secured to the inferior surface of thedental chair directly under the area for manual chest compressions next to a fixed vertical-measurementinstrument. Chest compression depth was assessed using the manikin’s Skill-Reporter system (E) with greenlights indicating chest compression depths of 5.1–6.0 cm and red lights for 3.8–5.0 cm. The distance of point Prelative to the inferior surface of the backrest remained fixed (F).2R E S U S C I T A T I O NP L U S 1 1 ( 2 0 2 2 ) 1 0 0 2 8 6nience. Each study participant individually performed 10 contiguousrounds of manual chest compressions (20 compressions per round;200 compressions total) at a pace of 100 compressions per minute insynchrony with a metronome for each of the 3 stool configurations(i.e., under manual chest compressions area, under shoulders, orno stool). Participants rested sufficiently after each round to avoidfatigue. A total of 600 manual chest compressions 5.1 to 6.0 cm indepth were recorded per participant in each dental chair. Each par-ticipant had a total of 1,800 manual chest compressions recordedat three different dental chairs.Analyses and statisticsThe programming language R (version 4.0.2; The Comprehensive RArchive Network, USA) was used for statistical analysis. Displace-ment measurements (maximum distance of Point P from baselineduring MCC) by three participants were analyzed separately at eachof the three respective positional configurations. The change in ver-tical displacement for each of the two stool positions (under manualchest compressions or under the shoulders) compared with no stoolposition (baseline) was calculated using the following equation (1).reduction ratio ¼ 1 �displacement with stooldisplacement without stoolð1ÞFig. 2 – Vertical Displacement Measurements. The physiques of the Provider’s body size (A, B, C) who performedmanual chest compression differed greatly. The thick lines represent median values, the boxes representinterquartile ranges, over and underlines represent date ranges, and circles represent outliers. In dental chairs#2 and #3, there were significant differences in stool positions because between under the shoulders and under thechest. In contrast #1 made little difference. (Please note that the horizontal axis does not start with zero only in thealpha figure).R E S U S C I T A T I O NP L U S 1 1 ( 2 0 2 2 ) 1 0 0 2 8 63Data sets were analyzed using the Shapiro-Wilk test to determinethe normality of distribution. The non-parametric data sets were thenanalyzed using the Wilcoxon rank sum test to determine statisticalsignificance (P < 0.001).ResultsThe vertical displacements of the dental chair backrest induced bymanual chest compressions were assessed with and without theplace of a stool (placed under manual chest compressions or underthe shoulders). A total of 5400 (1800 per participant) manual chestcompressions were recorded, but 32 were excluded due to anunclear recording or inappropriate compression depth. The verticaldisplacement was significantly reduced, and the reduction rateincreased in all situations when using the stool as a stabilizer(Fig. 2). The model of dental chairs has different characteristics forsupportive effect depending on the position of the stool. Particularlyin chair #1 which has a severely curved backrest exterior, there wasa significant difference in stool positions. As we placed a stool undermanual chest compressions in #1, Compared to no stool, a stoolunderneath the manual chest compression site resulted in a medianreduction ratio in chair displacement of 0.41, 0.42, and 0.38 for thethree BLS providers, and a stool underneath the shoulder resultedin a median reduction ratio in chair displacement of 0.67, 0.65, and0.64 (Fig. 3). In contrast chairs, #2 and #3 had flatter shapes. Theymade little difference between under the shoulders and under man-ual chest compressions. However, these were clearly different fromthe no stool (Fig. 2, 3).Fig. 3 – Calculated Reduction Ratios. Measurements of dental chair backrest vertical displacement during manualchest compressions and calculated reduction ratios. Results are expressed as median (interquartile range).4R E S U S C I T A T I O NP L U S 1 1 ( 2 0 2 2 ) 1 0 0 2 8 6DiscussionThis study has shown that manual chest compressions can be effec-tively performed by placing a stool under the shoulder for any shapeof the dental chair when we need to perform CPR on a patient in adental chair.When a patient suffers a cardiopulmonary arrest during dentaltreatment, many practitioners are uniformly upset and impatient. Thisis because they are not skilled in emergency life-saving proceduresand have difficulty calmly determining what to do next and reactingsmoothly; the ERC and AHA guidelines emphasize the importanceof minimizing interruptions during manual chest compressions. If car-diac arrest is suspected, manual chest compressions should be ini-tiated as soon as possible in a stable location.5,6 However, if it isdifficult to quickly and safely move a patient seated in a dental chairto the floor, CPR for a patient seated in a dental chair may be lesseffective than manual chest compressions if the vertical displace-ment of the backrest is large. Additionally, the backrest may not beadequately supported for effective resuscitation.7–10 Many moderndental chairs are ergonomically designed with a curved backrest thatgradually flattens upward to the patient’s shoulders; the externalshape of the chair backrest is on the lower bodyside (e.g., abdomenand lower back), rather than on the area where manual chest com-pressions are performed, is often more pronouncedly curved, andfor stabilization adversely affect the point of contact with the stool,reducing stability (Fig. 4). The exterior of many current dental chairsis characterized by an upward more flattened profile. If the stabilizingstool is positioned below the shoulders, it provides a larger, widercontact area and is more easily stabilized. The back of the dentalchair is further stabilized when there is a distance between thepatient’s seating position and the stabilizing stool.The video image results showed little displacement of the stool inresponse to the external force of the chest compressions. When per-forming this method of CPR in practice, we determined that no otherperson was needed to stabilize the stool position.LimitationsThis study had some limitations. First, only three types of dentalchairs were included in this study; however, many types of dentalchairs are in use worldwide. Second, the usefulness of other typesof stabilizers has not yet been tested, and the stools were only setin specific positions. Additional studies should be conducted to eval-uate other positions of the stabilizing stools for maximum effective-ness. Third, this study did not consider the effect of the dentalFig. 4 – Differences in stool position for manual chest compressions (Red arrow). A: The stool was placed under thechest compression site. B: The stool was placed under the shoulders. When the backrest of the dental chair wascurved in shape as shown in B, placing the stool on the shoulder side reduced the vertical displacement.R E S U S C I T A T I O NP L U S 1 1 ( 2 0 2 2 ) 1 0 0 2 8 65chair cushion or the use of a hard backboard during manual chestcompressions.11,12 Next, we need to demonstrate the clinical effec-tiveness of the method. For example, we should study subjects toimprove the success rate of chest compression depth and reducefatigue from chest compressions. Finally, because this study wasconducted using a mannequin model, caution should be exercisedin applying the results to human patients. However, no previous stud-ies have demonstrated that vertical displacement is significantlyreduced using this technique in several types of dental chairs. Theplacement of a stool to support the dental chair is a simple and effec-tive way to increase the effectiveness of manual chest compressionsand should be utilized when performing CPR on a patient sitting in adental chair.ConclusionThe present study demonstrates that our method is more stable foreffective manual chest compressions in any dental chair if the stoolis positioned under the shoulders rather than the manual chest com-pressions. These results suggest that dentists need to prepare inadvance an appropriate method (stool position) according to theexternal shape of the dental chair backrest to perform effective man-ual chest compressions and that it is very effective.Declaration of Competing InterestThe authors declare no competing interests.CRediT authorship contribution statementTakashi Hitosugi: Conceptualization, Methodology, Writing – origi-nal draft, Writing – review & editing. Norimasa Awata: Data cura-tion. Yoichiro Miki: Software, Validation, Investigation, Writing –review & editing. Masanori Tsukamoto: Supervision. TakeshiYokoyama: Conceptualization, Methodology, Writing – review &editing, Project administration, Investigation, Supervision.AcknowledgmentsNot applicable.Availability of data and materialsThe data sets during and/or analyzed the current study availablefrom the corresponding author upon reasonable request.Informed consentFor this type of study, formal consent is not required.Author detailsaSection of Dental Anesthesiology, Division of Maxillofacial Diag-nostic & Surgical Sciences, Faculty of Dental Science, KyushuUniversity, 3-1-1 Maidashi, Higashi-ku, Fukuoka City, Fukuoka 812-8582, Japan bSmile Dental Clinic, 1-1-5 Kamimoto-cho-nishi, Chuo-ku, Osaka City, Osaka 542-0062, Japan cSchool of InterdisciplinaryScience and Innovation, Faculty of Arts and Science, KyushuUniversity, 744 Motooka, Nishi-ku, Fukuoka-shi, Fukuoka 812-8582, JapandDepartment of Dental Anesthesiology, Kyushu Uni-versity Hospital, 3-1-1 Maidashi, Higashi-ku, Fukuoka City, Fukuoka812-8582, JapanR E F E R E N C E S1. Awata N, Hitosugi T, Miki Y, et al. Usefulness of a stool to stabilizedental chairs for cardiopulmonary resuscitation (CPR). BMC EmergMed 2019;19:46. https://doi.org/10.1186/s12873-019-0258-x.2. Hitosugi T, Awata N, Miki Y, et al. A Comparison of two stoolpositions for stabilizing a dental chair during CPR. Anesth Prog 2022.In press.3. Truhla´rˇ A, Deakin CD, Soar J, et al. Cardiac arrest in specialcircumstances section Collaborators, European ResuscitationCouncil Guidelines for Resuscitation 2015: Section 4. Cardiac arrestin special circumstances. Resuscitation 2015;95:148–201.4. Lott C, Truhla´rˇ A, Alfonzo A, et al. ERC Special CircumstancesWriting Group Collaborators, European Resuscitation CouncilGuidelines 2021: Cardiac arrest in special circumstances.Resuscitation 2021;161:152–219.5. Monsieurs KG, Nolan JP, Bossaert LL, et al. ERC Guidelines 2015Writing Group. European Resuscitation Council Guidelines forResuscitation 2015 Section 1. Executive summary. Resuscitation2015;95:1–80.6. Kleinman ME, Goldberger ZD, Rea T, et al. 2017 American HeartAssociation Focused Update on Adult Basic Life Support andCardiopulmonary Resuscitation Quality: An Update to the AmericanHeart Association Guidelines for Cardiopulmonary Resuscitation andEmergency Cardiovascular Care. Circulation 2018;137:7–13.7. Chi CH, Tsou JY, Su FC. Effects of rescuer position on thekinematics of cardiopulmonary resuscitation (CPR) and the force ofdelivered compressions. Resuscitation 2008;76:69–75.8. Chi C, Tsou J, Su F. Effects of compression-to-ventilation ratio oncompression force and rescuer fatigue during cardiopulmonaryresuscitation. Am J Emerg Med 2010;28:1016–23.9 Tomlinson AE, Nysaether J, Kramer-Johansen J, et al. Compressionforce-depth relationship during out-of-hospital cardiopulmonaryresuscitation. Resuscitation 2007;72:364–70.10. Hasegawa T, Daikoku R, Saito S, Saito S. Relationship betweenweight of rescuer and quality of chest compression duringcardiopulmonary resuscitation. J Physiol Anthropol 2014;33:16–23.11. Perkins GD, Smith CM, Augre C, et al. Effects of a backboard, bedheight, and operator position on compression depth during simulatedresuscitation. Intensive Care Med 2006;32:1632–5.12. Olasveengen TM, Mancini ME, Perkins GD, Collaborators ABLS,Support ABL, et al. International Consensus on CardiopulmonaryResuscitation and Emergency Cardiovascular Care Science WithTreatment Recommendations. Circulation 2020;2020:41–91.6R E S U S C I T A T I O NP L U S 1 1 ( 2 0 2 2 ) 1 0 0 2 8 6 | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/360298624PSYCHOSOMATIC RELATIONSHIPS IN THE MUSCULOSKELETAL SYSTEM ANDTHEIR IMPORTANCE FOR WELLNESSConference Paper · May 2022CITATIONS0READS141 author:Some of the authors of this publication are also working on these related projects:Body IMage View projectPhysical Culture from an Interdisciplinary Perspective (Frontiers in Psychology, Impact Factor: 4.232; CiteScore: 4.0; Q2) View projectDaniela StackeováVŠTVS Palestra s.r.o.18 PUBLICATIONS 52 CITATIONS SEE PROFILEAll content following this page was uploaded by Daniela Stackeová on 01 May 2022.The user has requested enhancement of the downloaded file.Vysoká škola tělesné výchovy a sportu PALESTRA, spol. s r.o. 7. BIENÁLE MEZINÁRODNÍ KONFERENCE WELLNESS, ZDRAVÍ A KVALITA ŽIVOTA PRAHA 4. 11. 2021 The College of Physical Education and Sport PALESTRA, L.t. d 7th BIENNALE INTERNATIONAL SCIENTIFIC CONFERENCE WELLNESS, HEALTH AND QUALITY OF LIFE PRAGUE 04/11/2021 Vysoká škola tělesné výchovy a sportu PALESTRA, spol. s r.o. 7. BIENÁLE MEZINÁRODNÍ KONFERENCE WELLNESS, ZDRAVÍ A KVALITA ŽIVOTA 4. 11. 2021 The College of Physical Education and Sport PALESTRA, L.t. d 7th BIENNALE INTERNATIONAL SCIENTIFIC CONFERENCE WELLNESS, HEALTH AND QUALITY OF LIFE 04/11/2021 TEMATICKÉ OKRUHY, CÍL: Diagnostika a funkční testování ve wellness. Multidimenzionální přístup ke kvalitě života a podpoře zdraví. Sebehodnocení a wellness plán. Bio-psycho-sociální rovnováha. Cílem mezinárodní konference je pomoci překlenout propast mezi výzkumem a aplikací výstupů. Celkovým účelem je usnadnit sdílení znalostí a praktických transformací z mezinárodního hlediska. THEMATIC AREAS, AIM: Diagnostics and function testing in wellness. Multidimensional approach to quality of life and health support. Self-assessment and wellness plan. Bio-psycho-social balance. The aim of the international conference is to help bridge the gap between research and application outputs. An overall purpose is to facilitate the sharing of knowledge and practical transformations from an international perspective. PRAHA Sborník abstraktů 7. bienále mezinárodní konference „Welness, zdraví a kvalita života“ Proceedings of abstracts of 7th Biennale of the International Scientific Conference “Wellness, health and Quality of Life“ Editoři/Editors: prof. PaedDr. Milada Krejčí, CSc. Mgr. Kamila Bulířová Alena Malcová © Vysoká škola tělesné výchovy a sportu PALESTRA, spol. s r.o. © The College of Physical Education and Sport PALESTRA, L.t. d PRAHA 2021 PRAGUE 2021 ISBN 978-80-87723-95-1 (online) 17 environment of Karlovy Vary on the mood states (tension, anger, fatigue, depression, confusion, vigour) of the participants. The questionnaires were completed by 36 participants (25 women, 11 men) aged between 20 and 65 years (⌀ 45 years). Based on the statistical evaluation, positive changes were observed in all investigated mood states. First of all, significantly increased vitality and, on the contrary, decreased fatigue (p-value <0.001) were confirmed. The training also had a positive effect on the anger and depressive states, with a significant reduction in both cases (p-value <0.05). This study results have confirmed that Nordic walking training in the natural environment of Karlovy Vary forests has a positive effect on the mood states of the participants and their mental health. These results will be the basis for follow-up research investigating the impact of physical activity in spa environments on the mental health of visitors and residents of spa towns. Keywords: Mental health; POMS; nordic walking; therapeutic landscape; Karlovy Vary; spa. PSYCHOSOMATIC RELATIONSHIPS IN THE MUSCULOSKELETAL SYSTEM AND THEIR IMPORTANCE FOR WELLNESS DANIELA STACKEOVÁ Abstract The locomotor system is directly affected by our emotional state, both muscle tension and its distribution, and locomotor functions. Knowledge of these psychosomatic relationships is important for all professions that work with the locomotor system of an individual, whether clinical, such as physiotherapy or non-clinical, including wellness specialists. We present an overview of them here. Increase and redistribution of muscle tension in persons who are under stress (different reaction in postural and phasic muscles) potentially leads to impairment of the present muscle imbalances, in the long run to a development of painful condition associated with muscle hypertension and strain of fibrous structures. Emotions associated with stress are accompanied by a specific reaction in the postural system, including impairment of balance functions and ultimately decrease in quality of motor functions (fluency of movement, its effectiveness, motor coordination, etc.). The neurophysiological basis of this reaction is the influence of the limbic system on the control of motor functions. Significant is also the change in respiratory function due to stress. A change in the type and depth of breathing is accompanied by a change in the involvement of the respiratory muscles (especially involvement of the upper scalupa fixators supporting the impairment of the functional state of the musculoskeletal system). By targeted intervention, whether of a passive nature, such as massage, or active nature, such as various methods of kinesiotherapy, we can not only eliminate the impact of stress on the musculoskeletal 18 system, but also target the psychic and support the harmony of the individual. However, a necessary condition is erudition in this area, including knowledge of the presented psychosomatic relationships. Keywords: Psychosomatics; psychosomatic relationships; wellness; kinesiotherapy. PHYSICAL AND SPORTS EDUCATION AS A MEANS OF FORMING A RELATIONSHIP TO LIFELONG PHYSICAL ACTIVITY AND HEALTH ELENA BENDÍKOVÁ Abstract The aim of the study presents and points out the importance of including physical and sport education in movement regime of pupils as a mean involved in forming the relationship to lifelong physical activity. The research sample consisted of 𝝨n=726 pupils in Slovakia, of which 378 girls in age (x) 17.5±1.4 years (body height (x) 166.4±3.6 cm, body weight (x) 61.9±4.8 kg) and 348 boys in age (x) 17.9±1.2 years (body height (x) 178.9±7.9 cm, body weight (x) 63.2±5.6 kg). In terms of data collection methods in the period 2020-2021, an interrogation method was used, based on pedagogical practice. Our findings show that physical and sport education is perceived in both genders with a positive attitude, but different interests with a significant difference (p <0.01) to the detriment of girls. Demands for the quality of the subject content is (p <0.01) higher in girls. Girls prefer within the PA subject of aesthetic focus, without direct contact (p <0.01), while boys prefer dynamic and conditioning PA (p <0.01). In girls and boys, this subject belongs to those that significantly (p <0.01) form a relationship to lifelong PA as well as to health (p <0.01). In the overall movement regime of pupils, there are significant (p <0.01) differences between the genders to the detriment of girls in terms of bulk of performed PA, frequency, and intensity (p <0.01). Which is related (p <0.01) also to a worse perception of health and health problems (p <0.01). At the same time, we significantly (p <0.01) confirm that physical and sport education is one of the transformational means involved in performing PA in daily movement regime of pupils (even in the pandemic period). This project was supported by VEGA 1/0519/19. Keywords: Health; interest; physical aktivity; physical and sport education; pupil. COVID 19 - UP CALL TO FOCUS ON WELLNESS MILADA KREJČÍ, LUDMILA VACEK Abstract Covid-19 pandemic represents an exceptional experience in the whole world and still is going through touched everybody. Never before has health and wellness declined so fast. It has also been a hard lesson about general health issues and health prevention as it became evident that healthier individuals with stronger immune systems were and View publication stats | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/335063376Serosolver: an open source tool to infer epidemiological and immunologicaldynamics from serological dataPreprint · August 2019DOI: 10.1101/730069CITATIONS0READS1696 authors, including:Some of the authors of this publication are also working on these related projects:Estimation of influenza vaccine effectiveness from observational studies View projectSequencing the supply chain for food safety and public health View projectJames A HayUniversity of Oxford35 PUBLICATIONS 2,801 CITATIONS SEE PROFILEKylie AinslieImperial College London77 PUBLICATIONS 7,630 CITATIONS SEE PROFILEJustin LesslerJohns Hopkins Bloomberg School of Public Health445 PUBLICATIONS 23,511 CITATIONS SEE PROFILEAll content following this page was uploaded by James A Hay on 12 August 2019.The user has requested enhancement of the downloaded file.Serosolver: an open source tool to infer epidemiological andimmunological dynamics from serological dataJames A. Hay1, Amanda Minter2, Kylie Ainslie1, Justin Lessler3, Adam J. Kucharski2†,Steven Riley1*†1 MRC Centre for Global Infectious Disease Analysis, Department of Infectious DiseaseEpidemiology, School of Public Health, Imperial College London, London, UnitedKingdom2 Centre for the Mathematical Modelling of Infectious Diseases, London School ofHygiene & Tropical Medicine, London, United Kingdom3 Department of Epidemiology, Johns Hopkins Bloomberg School of Public Health,Baltimore, Maryland, United States of America* s.riley@imperial.ac.uk† contributed equallyAbstractWe present a flexible, open source R package designed to obtain additional biologicaland epidemiological insights from commonly available serological datasets. Analysis ofserological responses against pathogens with multiple strains such as influenza pose aspecific statistical challenge because observed antibody responses measured inserological assays depend both on unobserved prior infections and the resultingcross-reactive antibody dynamics that these infections generate. We provide a generalmodelling framework to jointly infer these two typically confounded biological processesusing antibody titres against current and historical strains. We do this by linking latentinfection dynamics with a mechanistic model of antibody dynamics that generatesexpected antibody titres over time. This makes it possible to use observations ofantibodies in serological assays to infer an individual’s infection history as well as theAugust 8, 20191/30.CC-BY-NC 4.0 International licenseIt is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint (which. http://dx.doi.org/10.1101/730069doi: bioRxiv preprint first posted online Aug. 8, 2019; parameters of the antibody process model. Our aim is to provide a flexible inferencepackage that can be applied to a range of datasets studying different viruses overdifferent timescales. We present two case studies to illustrate how our model can inferkey immunological parameters, such as antibody titre boosting, waning andcross-reaction, and well as latent epidemiological processes such as attack rates andage-stratified infection risk.Introduction1Serological assays measure the interaction of a virus with the antibody repertoire of an2individual host [1]. Originally developed in the mid-20th Century, assays based on3haemagglutination inhibition (HI) and viral neutralization (VN) are still widely used4and are highly repeatable within the same lab [2]. These assays can be setup relatively5easily when viral culture systems are in place and require no specialist kits. Usually,6serum is diluted in 2- or 4-fold steps. Limiting dilutions with higher titres indicate a7stronger antibody response, whereas titres below the limit of detection indicate the8absence of a significant response. In influenza, ‘lower than 1:10’ is often the minimum9reading and dilutions of 1:1024 or higher indicate strong antibody responses. The10longevity of antibodies such as IgG make serological assays a key tool in epidemiological11surveillance, particularly where virological assays are not possible and symptoms are12non-specific or non-existent [3–6]. When only a single sample is available for an13individual, a threshold titre is often used as evidence of prior exposure or protection or14both, for example the commonly used threshold of 1:40 for influenza [7,8].15Paired blood samples and serological assays using known circulating strains can be16used to estimate exposure within a specific period of time. Samples taken before and17after an influenza season for which the main circulating strain is known can therefore be18used to infer attack rates [9–11]. Samples are usually processed as a pair to limit the19impact of between batch variability in testing. A ≥ 4-fold rise in titre against the20circulating strain (homologous titre) between the pre- and post-exposure samples is21typically assumed to be evidence of influenza infection. Because there is a degree of22subjectivity in the characterization of a sample being a limiting dilution, a ≥ 4-fold23difference, within a 2-fold dilution scheme, is deemed to be more robust against human24August 8, 20192/30.CC-BY-NC 4.0 International licenseIt is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint (which. http://dx.doi.org/10.1101/730069doi: bioRxiv preprint first posted online Aug. 8, 2019; error than a ≥ 2-fold difference in assessing the presence of haemaglutination (for HI) or25cell death (for VN) in each well of the assay plate [12,13]. However, a Bayesian analysis26of titre rise data has suggested that the somewhat arbitrary fourfold rise hides a27substantial number of lesser true titre rises that may represent missed infections [14].28Individual-level differences in age, infection history, time between exposure and29measurement, and virus-specific effects likely all play a role in generating sub-fourfold30titre rises [15–18].31Cross-reactivity complicates the interpretation of serological results when an32individual may have been exposed to two or more antigenically related viruses. Two33pathogens are considered antigenically related if exposure to one generates a34cross-reactive antibody response to the other in a serological assay. For example,35antibody responses against one dengue virus serotype can cross-react with another [19],36as well as other flaviviruses such as Zika virus [20,21]. Moreover, sequential lineages of37individual influenza A subtypes cross-react with their precursors and progeny [22]. One38popular use of HI assays is to assess the cross-reactivity between current influenza A39sub-types. Naive ferrets are inoculated with one of a panel of current reference strains40to produce virus-specific serum. HI titres are then measured for potentially novel41viruses using stocks of these reference ‘antisera’ [23].42Recently, there have been a number of initiatives to refine the analyses of commonly43available serological data. Antigenic cartography was developed to reduce complex44tables of HI readings for novel viruses and reference antisera to two dimensional space,45visualised as an ‘antigenic map’ [23–25]. An individual’s entire antibody repertoire46against an antigenically variable pathogen can be then projected as a surface over these47antigenic maps, with the height of the surface at any specific point indicating the48expected titre for that individual against a strain at that location in the map [26].49These ‘antibody landscapes’ can be used to generate biological insight by investigating50how antibody profiles develop over an individual’s life [27]. Further, compartmental51transmission models can be defined with explicit strata for each serological assay result52and used to test hypothesis about the interplay of social mixing and pre-existing53immunity [28]. These approaches retain much of the information present in the54magnitude of an assay measurement that may be lost when using seroconversion and55seropositivity thresholds.56August 8, 20193/30.CC-BY-NC 4.0 International licenseIt is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint (which. http://dx.doi.org/10.1101/730069doi: bioRxiv preprint first posted online Aug. 8, 2019; Here, we present the R package serosolver, which is the latest version of a code57base developed specifically to increase the epidemiological insight available from58serological assays [27,29]. Serosolver takes assay results from one or more serum59samples for an individual, which may have been tested against one or more related viral60strains, and infers a history of infections for that individual that is consistent with the61observed titres. It can jointly estimate the process parameters for the antibody kinetic62process by simultaneously inferring infection histories for many people. We use a63Bayesian approach and obtain correlated samples from the posterior densities for64infection histories and process parameters. The required assumptions for some priors65are straightforward and may incorporate previously observed immunological phenomena.66Prior assumptions for infection histories and the process that generates them can also67be incorporated, but can require additional justification, as we shall discuss.68The basic inference challenge can be summarised as follows. For a given set of69serological data (Y , which may include assay measurements against one or more70strains), we wish to obtain the joint posterior distribution of the process parameters (θ),71individual infection histories (Z) and temporal probability of infection in the population72(φ). This posterior is proportional to three components: (i) the observation and73antibody process models f(Yi|Zi, θ), which give the likelihood of observing a set of74titres Yi for each individual i at serum sampling times (ti), given infection history Zi75and process parameters θ; (ii) the transmission level P(Zi,j|φj), which gives the76probability of individual i having an infection with the strain circulating in time period77j, given population infection probability φj; and (iii) the prior level, giving the prior78probability for the process parameters, P(θ) and the prior probability of any infection79at each time period j, P(φj). This results in the following expression:80P(Z, φ, θ|Y ) ∝n�i=1� �k∈tif(Yi,k|Zi, θ)jmax�j=jminP(Zi,j|φj)P(φj)�P(θ)(1)First we outline how this expression is flexibly implemented in the serosolver81package, then we show how the package can be applied to cross-sectional and82longitudinal influenza data from China and Hong Kong to infer key epidemiological and83immunological values.84August 8, 20194/30.CC-BY-NC 4.0 International licenseIt is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint (which. http://dx.doi.org/10.1101/730069doi: bioRxiv preprint first posted online Aug. 8, 2019; Design and Implementation85Antibody process model86For a given individual infection history and set of biological parameters, the antibody87process model generates a set of expected log titres for that individual against all88possible test strains. Following previous work [27], the expected log titre individual i89has against the strain that circulated at time j when observed at time k is defined as a90linear combination of the contribution of antibody responses from each prior infection:91Xi,j,k =�m∈Zis(Zi, m) [µ1d1(j, m) + µ2w(m, k) d2(j, m)](2)The model components are defined by:921. Long-term boosting. This is defined by a parameter µ1, equivalent to the expected93persistent rise in titre against a homologous strain following primary infection.942. Short-term boosting. The transient component of the antibody dynamics is95defined by µ2w(m, k) = µ2 max{0, 1 − ωtm}, where µ2 is the boost in titre, ω is a96waning parameter to be fitted, and tm = k − m is the time since infection with97strain m.983. Long-term cross-reactive antibody response from related strains. We assume the99level of cross-reaction between a test strain j and infecting strain m ∈ Zi100decreases linearly with antigenic distance (see Data section below for definition).101The cross-reaction function is d1(j, m) = max{0, 1 − σ1δm,j}, where δm,j is102antigenic distance between strains j and m, and σ1 is a fitted parameter.1034. Short-term cross-reactive antibody response. Similar to the long-term response,104except this can wane over time. Cross-reactivity between a test strain j and105infecting strain m is defined as d2(j, m) = max{0, 1 − σ2δm,j}1065. Antigenic seniority by suppression. This results in lower titres from later107infections in comparison to earlier ones. In the model, this works by scaling the108titre contribution by a factor s(Zi, m) = max{0, 1 − τ(Nm − 1)}, where Nm is the109infection number (i.e., primary infection is 1, secondary is 2) and τ is a fitted110August 8, 20195/30.CC-BY-NC 4.0 International licenseIt is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint (which. http://dx.doi.org/10.1101/730069doi: bioRxiv preprint first posted online Aug. 8, 2019; parameter.111The antibody process model can be reduced to simpler models by setting certain112parameter values equal to 0. For example, a model without antigenic seniority can be113created by setting τ = 0 or a model with only waning responses by setting µ1 = 0.114In addition, serosolver can been extended to include more complex antibody115kinetics, as described in Supplementary Material 2. We note that the additional116immunological phenomena described in Supplementary Material 2 are not exhaustive,117and additional mechanisms may easily be implemented by making minor modifications118to the package code.119Antigenic distance120The antibody process model described in Equation 2 includes parameters which121describe short- and long-term cross-reactive antibody processes. These processes depend122on a metric of antigenic distance between each pair of strains [23]. In the model, the123antigenic distance δm,j between strains m and j is therefore defined by a matrix of124pairwise distances. Serosolver can accommodate antigenically varying strains (all δm,j125are specified) or a single homologous strain (all δm,j = 0). The extent to which strains126are antigenically distinct or similar can be described using the distance matrix.127Observation model128The expected titre Xi,j,k defined in Equation 2 feeds into the observation model, which129converts the continuously valued model predicted titre into a discrete observed titre.130The distribution of the observed titre consists of a normally distributed random variable131g(s) with mean Xi,j,k and variance ε, which is then censored to account for132integer-valued log titres in the assay. Hence the probability of observing an empirical133titre at time k within the limits of a particular assay Yi,j,k ∈ {0, ..., Ymax} given134expected titre Xi,j,k is,135P(Yi,j,k|Xi,j,k) = f(Yi,j,k|Zi, θ) =� Yi,j,k+1Yi,j,kg(s)ds if Yi,j,k ∈ {1, Ymax − 1} ;� 1−∞ g(s)ds if Yi,j,k = 0 ;� ∞Ymax g(s)ds if Yi,j,k = Ymax .(3)August 8, 20196/30.CC-BY-NC 4.0 International licenseIt is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint (which. http://dx.doi.org/10.1101/730069doi: bioRxiv preprint first posted online Aug. 8, 2019; Serosolver includes an additional option to include strain-specific measurement bias,136which may arise through strain-specific differences in assay reactivity [26,30–32].137Specifically, an additional observation error is added to the predicted log antibody titres;138this measurement error can be different for each individuals strain or can be specified139for a group (or cluster) of strains. The predicted titre X′i,j,k taking into account140strain-specific measurement bias is given as:141X′i,j,k = Xi,j,k + χj(4)Where χj is the measurement offset for strain j. The hierarchical form of the142measurement bias term may also be specified by the user: χj may be estimated as an143independent parameter for each j; may be assumed to come from a hierarchical144distribution χj ∼ N(¯χ, σ2χ); and may be fixed for particular strains/groups e.g., fixing145¯χ = 0 or χjmax = 0.146Infection history model147Serosolver tracks each individual’s infection history as a binary vector of latent states148indicating the presence (1) or absence (0) of infection, where each element of the vector149represents a time period during which individuals could be infected. The set of infection150histories for the sample population are therefore described by a binary matrix, Z, where151each row represents an individual, i, and each column represents a time, j, at which an152individual could be infected once. The probability of the infection history matrix, P(Z)153is given by,154P(Z) =n�i=1jmax�j=jmin�P(Zij|φj)P(φj)�.(5)Each infection event (Zi,j) is the outcome of a single Bernoulli trial, with probability155P(Zi,j|φj) = φZi,jj(1 − φj)Zi,j. The choice of the prior distribution for the probability of156infection, P(φj), is discussed below and in further detail in Supplementary Material 1.157The time resolution of infection times may be set by the user depending on the data;158frequent sampling times affords greater time resolutions (e.g., months), whereas less159frequent sampling may be better suited to cruder time resolutions (e.g., years).160August 8, 20197/30.CC-BY-NC 4.0 International licenseIt is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint (which. http://dx.doi.org/10.1101/730069doi: bioRxiv preprint first posted online Aug. 8, 2019; The infection history posterior can be used to calculate a key epidemiological161measure of interest: the population attack rate over time. Attack rates can be inferred162through combining inferred infection histories post-hoc to estimate the proportion of at163risk individuals (those that were alive and in the sample) that were infected in a given164time period. Summing the columns of the infection history matrix gives the total165number of infections for a given time period, whereas summing the rows give the total166number of lifetime infections for an individual. To ensure biological plausibility,167individual infection histories are constrained to prevent infections before an individual is168born and after the last time at which a serum sample was taken. A key feature of the169package is that the user is given control over the prior assumptions for the infection170history and the probability of infection in each time unit (months, years etc).171Application to influenza A/H3N2172The initial development of serosolver focused on influenza A/H3N2, which has173circulated in human populations since 1968 and has undergone substantial antigenic174evolution over this time [23,32–34]. Figure 1 illustrates how our analytical approach175applies to influenza A/H3N2. In this case, we make the assumption that the antigenic176distance between strains can be described by a two-dimensional distance, with strains177moving through the space over time. The expected log antibody titre for a given178individual against a specific strain at a specific time can therefore be predicted using179this antigenic distance map, the antibody process model described by Equation 2, and180the individual level infection histories. Finally, the observed log antibody titres can be181used to infer individual level infection histories and antibody process parameters based182on time of sampling and the observation model.183Data184The serosolver package requires two datasets as inputs. The first is an antigenic map,185which defines the two-dimensional location of viruses that circulated at each time point186during the period of interest, and hence can be used to calculate the pairwise antigenic187distance between any two viruses (i.e., δm,j in the antibody process model, for strains m188and j). The model automatically sets the potential period during which individuals189August 8, 20198/30.CC-BY-NC 4.0 International licenseIt is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint (which. http://dx.doi.org/10.1101/730069doi: bioRxiv preprint first posted online Aug. 8, 2019; log antibody titrelog HI titreAntigenic distanceAntigenic distanceTimeFig 1. Conceptual overview of the analytical approach used in serosolver,as applied to influenza A/H3N2. Top panel: antigenic map for influenzaA/H3N2 using coordinates from [23], with different viruses coloured by year of isolation.Solid points show centroids across all strains isolated in a given year, hollow points showindividual strains. Dashed line shows an antigenic summary path, generated by fitting asmoothing spline through the observed isolates. Points further apart in space are lesscross-reactive. Middle panel: conceptual illustration of the antibody kinetics model.An individual is infected with the orange virus, which results in boosting and waning ofhomologous antibody titres. In parallel, antibodies that cross react with viruses atdifferent points in antigenic space also boost and wane (green and blue viruses). Theindividual is later infected by the green virus, which leads to further boosting andwaning of antibodies. Bottom panel: HI titres measured from serum samples taken atdifferent times capture different parts of the homologous and cross reactive antibodykinetics. Different sampling strategies will represent different subsets of thesemeasurements e.g., a cross-sectional study might inform a single subplot, whereas alongitudinal study might inform just the orange bars from each of the three subplots.Clearly a sampling strategy with multiple serum samples and many viruses tested persample will provide the most information.August 8, 20199/30.CC-BY-NC 4.0 International licenseIt is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint (which. http://dx.doi.org/10.1101/730069doi: bioRxiv preprint first posted online Aug. 8, 2019; could have been infected based on the earliest and latest circulating strains in the190antigenic map.191The second dataset consists of individual-level log titres against one or more viruses192defined in the antigenic map. Each titre measurement is accompanied by a sampling193time k (i.e., when the serum sample was collected) and strain circulation time j (i.e.,194when the strain was originally isolated).195Inference196Prior assumptions197Inference in serosolver is fully Bayesian, which means priors must defined for all198model parameters and infection histories. The priors on the antibody process199parameters are uniform by default, but users may create their own prior function, which200may be based on previous analyses. For example, constrained estimates for the short201term antibody waning parameters may be used to specify strong beta or Gaussian202priors on some of the antibody kinetics parameters for analyses where serum samples203may be poorly suited to inform such short term effects.204Priors on the infection histories require more consideration, as the prior also205captures any assumptions regarding the infection generating process. Because the206number of potential infection times and strains can be vast, the contribution of the207infection history prior must be well characterised to avoid any unforeseen bias during208inference. The prior assumption on the functional form of φ, whether individual209infection risks are independent at a given time j, and whether an individual’s risk of210infection depends on infection outcomes at previous times can have important211implications for the prior on key infection history summary metrics, such as the attack212rate in a given time period and the lifetime number of infections for an individual.213Although the literature for Bayesian variable selection presents a number of214potential options, infection states are influenced by epidemiological and immunological215structures that are not well characterised by standard prior assumptions (i.e., highly216dispersed attack rates and variation in individual-level susceptibility) [35]. We therefore217provide the user with flexibility in the assumed infection history and attack rate priors,218with different prior assumptions each bringing their own biases and rationale.219August 8, 201910/30.CC-BY-NC 4.0 International licenseIt is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint (which. http://dx.doi.org/10.1101/730069doi: bioRxiv preprint first posted online Aug. 8, 2019; Serosolver includes four infection history prior options. We summarise these priors in220the main text, though an extensive discussion is provided in Supplementary Material 1.221Hyper-prior on the probability of infection over time, version 1: Under222this prior, the probability of infection is given by φj. The infection generating process is:223φj∼f(j)(6)Zi,j∼Bernoulli(φj)(7)where f is a user specified function describing the prior distribution on φ, P(φj). By224default, f is the uniform distribution, φj ∼ unif(0, 1), though it may be set to225incorporate information related to transmission such as seasonality or changes in social226behaviour.227Beta prior on the probability of infection over time, version 2: As in prior2281, this prior assumes that individuals are under a common infection process during a229given window of time. However, by placing a beta prior with parameters α and β and230integrating over values for φ, each φ need not be estimated explicitly. We have found231that this improves convergence of the model fitting framework. The infection generating232process is:233φj∼Beta(α, β)(8)Zi,j∼Bernoulli(φj)(9)The probability of infection in a given time period is independent of other time234periods, but dependent on the infection status of other individuals in the population at235that time. The prior on the per-capita attack rate is therefore a beta distribution, and236the prior on the lifetime number of infections for any individual follows a binomial237distribution.238Beta-binomial prior on the total number of infections during an239individual’s life, version 3: Unlike priors 1 and 2, this prior assumes that an240individual’s risk of infection at a given time is independent of all other individuals.241Rather, a prior is placed on the total number of infections that an individual is expected242to experience over the course of their life. This is the prior used in our previous243August 8, 201911/30.CC-BY-NC 4.0 International licenseIt is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint (which. http://dx.doi.org/10.1101/730069doi: bioRxiv preprint first posted online Aug. 8, 2019; work [27]. The infection generating process is assumed to be:244pi∼Beta(α, β)(10)Zi,j∼Bernoulli(pi)(11)The prior on the per-capita attack rate across all individuals therefore follows a245binomial distribution, and the prior on the lifetime number of infections for any246individual follows a beta-binomial distribution, with parameters α and β that can be set247by the user.248Beta prior on the probability of any infection, version 4: In the final prior249version, infection states are assumed to be independently and identically distributed250with respect to both time and individual under the following infection generating251process:252φ∼Beta(α, β)(12)Zi,j∼Bernoulli(φ)(13)This assumption places a beta-binomial prior on both the number of infections at a253given time j (the attack rate) and the number of lifetime infections experienced by254individual i.255Markov Chain Monte Carlo256Serosolver uses a custom, adaptive Markov Chain Monte Carlo (MCMC) framework257to sample from the joint posterior distribution of θ and Z conditional on the antibody258titre data (Equation 1). The package jointly estimates θ and Z using a259Metropolis-Hastings algorithm, alternating between sampling values for θ and Z. The260MCMC framework automatically tunes the proposal step size for θ, and changes the261number of individuals sampled for Z to achieve a specified acceptance rate. Given that262MCMC sampling of binary variables is a challenging problem [35,36], serosolver263includes additional custom proposal steps included for Z to improve chain mixing. The264full sampling algorithm for Z is described in Supplementary Material 1. Briefly, the265August 8, 201912/30.CC-BY-NC 4.0 International licenseIt is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint (which. http://dx.doi.org/10.1101/730069doi: bioRxiv preprint first posted online Aug. 8, 2019; algorithm uses a random-scan Metropolis-within-Gibbs proposal on infection histories to266either propose new infection states or swap the times of existing infection states. These267steps were developed to improve MCMC mixing when the infection states in adjacent268time periods may be highly correlated. Where automated tuning is insufficient to269achieve good mixing, all of the parameters controlling the proposal algorithm are270exposed to the user to be changed manually from their default values.271MCMC diagnostics272To ensure reliable MCMC model fitting, thorough convergence diagnostics must be273calculated to ensure that separate MCMC chains have converged on the same274distribution, are not trapped in local modes and provide estimates of the posterior275distribution with sufficient sample size. Serosolver includes functions to test these276criteria in two broad categories: (i) visual assessment of convergence and goodness of fit;277(ii) metrics of convergence checking between-chain agreement, auto-correlation and278effective sample size. Alongside existing tools in the coda and bayestools279packages [37,38], these functions include: MCMC trace and density plots for antibody280kinetics parameters; MCMC trace and density plots for inferred attack rates over time;281MCMC trace and density plots for inferred infection histories; model predicted titres282plotted against observed titres; and inferred attack rates over time. MCMC chain283outputs are written to disk during the fitting procedure, and the chain outputs are284compatible with the coda and bayesplot R packages. The full posterior distribution of285infection states as augmented data is therefore easily recoverable for further analysis, for286example regression analysis of numbers of infections during some period of time.287Implementation288In serosolver, model inputs and assumptions may be changed depending on the289serological data and hypotheses under consideration. For example, in some cases the290user may be most interested in short-term, fine-scale (e.g., weekly or monthly) dynamics291of infection; in other situations, long-term annual dynamics may be of interest.292Furthermore, although much of the development of this package came from analysis of293influenza A/H3N2 dynamics, these concepts and inputs are easily adaptable to294August 8, 201913/30.CC-BY-NC 4.0 International licenseIt is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint (which. http://dx.doi.org/10.1101/730069doi: bioRxiv preprint first posted online Aug. 8, 2019; antigenically stable pathogens by specifying the input antigenic map.295The package work flow is divided into a number of distinct stages, which handle the296data and parameter inputs, simulation, inference, posterior diagnostics, and analysis297(Fig 2) We developed the package to rely on only a few function calls for each of these298stages, but with ample room for customisation and flexibility at each stage.299To set up the model, users only need to provide: a data frame describing the model300parameters (they can also change a flag to fix or estimate any of the parameters); a301data frame with the antibody titre data in long format; and an antigenic map describing302the antigenic relationship between each strain. Examples of a typical data cleaning303workflow are provided in Supplementary Material 4.304Serosolver allows users to create their own likelihood and prior functions on top of305those provided by default, requiring only that they return a vector of likelihoods (one306per individual), and accept arguments for a vector of parameters (matching those307defined in the general serosolver model) and the infection history matrix. Users can308specify which prior assumption about infection histories is used, as specified above. In309addition to the range of inbuilt options, the modular workflow of serosolver means310that custom extensions tailored to particular problems should be readily achievable with311only minor modifications to the code. In particular, alternative antibody kinetics312models that capture pathogen-specific immunology and alternative assumptions about313the infection history generating process.314It is essential to run multiple chains to assess mixing properties and potential bias in315any MCMC analysis. Furthermore, model comparison and sensitivity analyses are a316common output of model fitting analysis. It is simple to use serosolver with a parallel317back-end, either through a computing cluster or locally with packages such as318doParallel [39] to generate multiple chains in parallel. The accompanying vignettes319(Supplementary Material 3 and Supplementary Material 4) demonstrate how multiple320chains may be run in parallel locally, but we note that much of our own work with321serosolver is done using a high performance cluster.322August 8, 201914/30.CC-BY-NC 4.0 International licenseIt is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint (which. http://dx.doi.org/10.1101/730069doi: bioRxiv preprint first posted online Aug. 8, 2019; Fig 2. Inputs and outputs for the serosolver R package. The package has twosets of inputs to define data and parameters. These feed into the process model thatcan either be used to simulate data by itself, or combined with observed data andMCMC to obtain three posterior outputs: individual-level infection histories, populationprobability of infection, and biological parameters. Once these posteriors have beenobtained, serosolver can run MCMC diagnostics and plot key immunological andepidemiological processesResults323We present two case studies to highlight the range of insights that serosolver can324generate from serological samples. These cover two types of study designs commonly325used to examine epidemiological and immunological dynamics using serological data,326which can be thought of as subsets of the observations shown in Fig 1, bottom panel.327The first is a serological survey testing individuals against a single homologous strain,328which can reveal short-term epidemic dynamics, analogous to observing each of the bars329of a single colour from Fig 1. We use data from a longitudinal study conducted in Hong330Kong between 2009 and 2011 to estimate short-term antibody kinetics parameters331against A/H1N1pdm09 in a population with no prior immunity. The second type of332study design involves testing samples against a panel of previously circulating strains,333which can provide insights into historical patterns of infection, analogous to observing334all of the bars within a single serum sample from Fig 1. To illustrate this application,335we apply the package to cross-sectional samples tested against a panel of historical336A/H3N2 influenza strains to infer infection histories and antibody kinetics.337Case Study 1338The first case study uses data from a cohort study in Hong Kong during and after the3392009 A/H1N1pdm09 outbreak [40]. With repeat serological samples tested against a340August 8, 201915/30.CC-BY-NC 4.0 International licenseIt is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint (which. http://dx.doi.org/10.1101/730069doi: bioRxiv preprint first posted online Aug. 8, 2019; given virus, serosolver can reconstruct the unobserved infection dynamics from341measured titres collected several months apart. It is also possible to examine these342infection dynamics stratified by available demographic variables, such as vaccination343status (Fig 3A) and age (Fig 3B). Finally, we can estimate biological parameters344shaping the short-term antibody response (Fig 3C).345We were able to estimate quarterly exposure rates, which could include either346infection or vaccination. The inferred peaks in exposure rates are consistent with the347observed two waves of the 2009 pandemic. We investigated the impact of vaccination348status and age on inferred exposure rates. We found differences in exposure rates in349vaccinated individuals compared to unvaccinated individuals, with higher overall350exposure rates in vaccinated individuals. Intuitively, we would expect infection rates to351be lower in vaccinated individuals; however, the converse suggests that vaccination352causes boosts in antibody titres that are being inferred as infections. Thus, an353individual’s vaccination status is an important consideration when using serological data354to infer infection history. Additionally, we observed clear differences in age-stratified355exposure rates with exposure rates highest among adults and children, and lowest356among the elderly, confirming previous findings of age-stratified exposure rates during357the 2009 pandemic [41]. Finally, we aimed to characterise the short-term immune358response following infection by estimating short-term antibody kinetics parameters. We359found that there is a strong short term boost (16-fold rise) in antibody titre following360infection which wanes by 3% every 4 months.361To assess whether data contain enough information to reliably estimate the infection362histories and biological process parameters, serosolver can be used to run a363simulation recovery study. For example, if data of the same structure as the364A/H1N1pdm09 outbreak in Hong Kong are generated using plausible parameter365values [27], it is possible to re-infer these parameters (Fig 4B) alongside the366individual-level infection histories (Fig 4C) and overall probabilities of infection367(Fig 4A). However, depending on the sampling frequency, number of tested strains and368number of repeat measurements, there are varying levels of information to estimate369these quantities. When antibody titre data is sparse, the priors placed on either the370antibody parameters, infection histories or probability of infection parameters will have371a greater effect on the estimation performance. We therefore recommend routine372August 8, 201916/30.CC-BY-NC 4.0 International licenseIt is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint (which. http://dx.doi.org/10.1101/730069doi: bioRxiv preprint first posted online Aug. 8, 2019; Fig 3. Influenza A/H1N1pdm09 infection dynamics in Hong Kong cohort.A: Exposure rates in unvaccinated individuals. Red line shows median estimate fromserosolver, with 95% credible intervals (CI); black line shows reported A/H1N1pdm09isolates. B: Age-specific infection rates in unvaccinated individuals. Lines show medianestimates from serosolver for each age group (red: <19, green: 19-64, blue: >64) with95% CI. C: Posterior densities of process parameter estimates. Dashed vertical linesrepresent 2.5th, 50th, and 97.5th percentiles.implementation of simulation recovery on new data to ensure that the most suitable373model is being applied to the data available.374Case Study 2375The second case study considers cross-sectional serological samples collected in southern376China in 2009, which were tested against nine historical influenza A/H3N2 strains that377circulated between 1968 and 2008 [29,42]. Serosolver can be used to reconstruct378August 8, 201917/30.CC-BY-NC 4.0 International licenseIt is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint (which. http://dx.doi.org/10.1101/730069doi: bioRxiv preprint first posted online Aug. 8, 2019; Fig 4. Simulation-recovery of parameter and infection estimates usingsimulated single strain longitudinal data in same format as the Hong Kongdataset. A: Model estimated attack rates versus ‘true’ attack rates. Solid line showsestimated attack rate with 50% and 95% credible intervals (CI); green dashed line showstrue attack rates. B: ‘True’ process parameters used for simulation compared toestimated posterior densities. Black solid vertical lines indicate true parameter values;dashed vertical lines represent 2.5th, 50th, and 97.5th percentiles. C: Model predictedtitres and inferred infections compared to observed titres and known infections. Blackpoints indicate observed titres; black lines indicate posterior median model predictedtitres; green shading shows 50% and 95% CI on model predicted latent titres; dashedvertical lines indicate the timings of true infections; blue shading indicates posteriorprobability of infection.several features of the epidemiological and immunological dynamics. First, Fig 5A379shows substantial variation in the inferred historical attack rates of A/H3N2, with clear380periods of high incidence interspersed by periods of very low incidence (range of381posterior medians: 3.63% to 95.2%). In these analysis, we used a weakly informative382prior on the annual attack rate with a mode of 15% with prior version 2. Our posterior383estimates were very similar to this, with a median inferred attack rate of 14.6%,384suggesting either agreement between the data and prior or a lack of information in the385data. We also identified clear age-specific patterns of infection. Fig 5D shows the386median number of infections per 10 years alive stratified by age at the time of exposure.387August 8, 201918/30.CC-BY-NC 4.0 International licenseIt is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint (which. http://dx.doi.org/10.1101/730069doi: bioRxiv preprint first posted online Aug. 8, 2019; These results agree with previous analyses that individuals are infected, or at least388experience antibody boosting, less frequently as they get older [27]. Fig 5E shows the389proportion of individuals infected at least once by a virus from each of the 14 antigenic390clusters considered here stratified by age at the time of exposure. Inference of long-term391biological parameters suggested that individuals experience a long-term antibody boost392mu1 of 2.24 log units (posterior median, 95% CI: 1.95-2.51), corresponding to393approximately a 4-fold boost to long term homologous titres that wanes with antigenic394distance (long term cross reaction σ1 = 0.105 posterior median, 95% CI: 0.0962-0.113)395and decreases with each successive exposure (antigenic seniority parameter, τ = 0.0310396posterior median, 95% CI: 0.0210-0.0415).397As with the first case study, simulation recovery was used to validate the ability of398serosolver to correctly infer underlying processes from a given dataset (discussed in399detail in Supplementary Material 4).400Fig 5. Influenza A/H3N2 dynamics in southern China. A: Inferred historicalattack rates. Shaded regions show 50% and 95% credible intervals( CI), black line showsposterior median, dashed green line shows maximum posterior probability estimate; B:Example latent titre trajectory (dark grey region, light grey region and black line show50% CI, 95% CI and posterior median estimates respectively) against observed titres(black dots) of inferred or one individual. D: Frequency of infection by age group. C:Posterior densities for the inferred antibody kinetics parameters. 95% CI and posteriormedians shown as dashed lines. E: Per cluster attack rates in <20 and ≥20 age groups.Clusters with darker shading circulated for longer before succession.August 8, 201919/30.CC-BY-NC 4.0 International licenseIt is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint (which. http://dx.doi.org/10.1101/730069doi: bioRxiv preprint first posted online Aug. 8, 2019; Computational performance401Serosolver uses a C++ back-end with substantial optimisation to scale the model to402large data sets and high infection time resolutions with reasonable run times. Table 1403displays the mean run time of 5 MCMC chains fitting the serosolver model to404serological data of different dimensions. In the most complex scenario, which involves405fitting the model to 164,000 antibody titre measurements and inferring the infection406state of 1000 individuals at 164 different time points (164,000 infection states), effective407sample sizes >200 are achievable for both the antibody process parameters and attack408rate estimates in <12 hours. For smaller scale analysis (e.g., 100 individuals, <5000409titres), high effective sample sizes and well-mixed chains are easily generated within 30410minutes.411August 8, 201920/30.CC-BY-NC 4.0 International licenseIt is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint (which. http://dx.doi.org/10.1101/730069doi: bioRxiv preprint first posted online Aug. 8, 2019; Availability and Future Directions412Serosolver provides a general inference framework to estimate epidemiological and413immunological dynamics from serological data. The open source package is available414from GitHub (https://github.com/seroanalytics/serosolver), with detailed415accompanying vignettes covering the main implementation and case studies we describe416here. The aim of this package is to provide an open source, modifiable framework to fit417antibody kinetics models that also require inference of unobserved infections. Disparate418serosurveys measuring antibody titres over time are often underpinned by comparable419dynamics, and we therefore felt that a unifying tool to enable quick reproduction and420direct comparison of analyses across different datasets would be a useful addition to the421literature.422As well as the stand-alone applications we have illustrated in the case studies above,423serosolver could easily link with traditional epidemiological analysis. The results424presented here are not intended to be exhaustive analyses, but rather to demonstrate425the utility and range of insights that can be generated from serological data. In426particular, the posterior latent individual-level infection histories and titre trajectories427could act as inputs into regression models. For example, serosolver outputs could be428combined with syndromic or lab-confirmation data to examine the relationship between429susceptibility and titre at time of infection [43]. These methods could also apply to430other pathogens; a similar model structure has recently been used to examine latent431titres for dengue [44].432Moreover, serosolver can incorporate prior knowledge on time of exposure either433from surveillance data or, if relevant, temporal climate variables. In the case studies434presented, we used relatively simple priors for the probability of infection. However,435more complex temporal priors could be imposed by having a different prior distribution436for the probability of at each time point (i.e., different value of α and β) to account for437seasonality in transmission dynamics. In the future, we hope to extend serosolver to438include non-linear feedback between past exposures and future risk, by embedding an439epidemic model as well as the probability of infection [45]. In theory, this package could440be used to generate an ongoing database of inferred immunological parameters, allowing441estimates to be updated and combined between to better estimate attack rates and442August 8, 201921/30.CC-BY-NC 4.0 International licenseIt is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint (which. http://dx.doi.org/10.1101/730069doi: bioRxiv preprint first posted online Aug. 8, 2019; infection histories in less data-rich cohorts.443Serosolver could also be used to inform the design of serological sample collection444and testing. Given potential logistical or budgetary restrictions on analysis of stored445sera or collection of new samples, serosolver could be used to simulate different study446designs and show how accurately these designs could recover the main parameters of447interest.448At present, serosolver focuses on inference for a single exposure type. However, for449viruses like influenza and dengue, individuals may be exposed to multiple subtypes or450serotypes in the same season. Exposure to one antigen may cross react with another451antigen providing protection against antigens an individual has not been directly452exposed to. For example, infection with influenza A/H1N1 may provide cross-reactive453protection against other group 1 viruses, and A/H3N2 against group 2 viruses [46].454Additionally, the incorporation of multiple exposures can facilitate the inclusion of455vaccine exposure. In influenza, where vaccination is recommended annually, exposure to456vaccination is an important piece of the immunological life course puzzle of an457individual [47]. In its current form, serosolver can estimate differences between458exposures by being fit independently to different subtypes. It can also fit models459separately to vaccinated or unvaccinated populations to estimate how serological460dynamics vary between these groups. Although this is a useful first approximation,461future versions of serosolver will include potential for multiple exposure types during462the same season so that any interactions can be modelled explicitly.463There is increasing evidence that serological titre data contain substantial additional464information about infection and immunity dynamics, which are not captured by simple465four-fold rise metrics [14,44,48,49] Furthermore, in multi-strain pathogen systems,466evidence is mounting that individual-level heterogeneity in unobserved exposure467histories is a key driver of susceptibility to infection and disease [26,47,50,51].468Serosolver provides a generic framework to extract this information from commonly469collected data. As serological data become increasingly available, it will be important to470develop modern analytical methods and tools that account for known biological and471epidemiological processes that may confound or bias inference [49,52–54].472August 8, 201922/30.CC-BY-NC 4.0 International licenseIt is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint (which. http://dx.doi.org/10.1101/730069doi: bioRxiv preprint first posted online Aug. 8, 2019; Supporting information473Supplementary Material 1.Full description and discussion of the474infection history priors and their implications for inference.475Supplementary Material 2.Additional immunological mechanisms and476how to modify code to incorporate alternative antibody kinetics.477Supplementary Material 3.Case study 1 vignette with all code required478for model fitting, figure generation and simulation recovery.479Supplementary Material 4.Case study 2 vignette with all code required480for model fitting, figure generation and simulation recovery.481Acknowledgments482References1. Hirst GK. The quantitative determination of influenza virus and antibodies bymeans of red cell agglutination. J Exp Med. 1942;75(1):49–64.2. Laurie KL, Engelhardt OG, Wood J, Heath A, Katz JM, Peiris M, et al. Aninternational laboratory comparison of influenza microneutralisation assayprotocols for A (H1N1) pdm09, A (H3N2) and A (H5N1) influenza A viruses byCONSISE. Clin Vaccine Immunol. 2015; p. CVI–00278.3. Cutts FT, Hanson M. Seroepidemiology: an underused tool for designing andmonitoring vaccination programmes in low- and middle-income countries.Tropical Medicine & International Health. 2016;21(9):1086–1098.doi:10.1111/tmi.12737.4. Amanna IJ, Carlson NE, Slifka MK. Duration of humoral immunity to commonviral and vaccine antigens. The New England journal of medicine.2007;357(19):1903–1915. doi:10.1542/peds.2008-2139LLLL.August 8, 201923/30.CC-BY-NC 4.0 International licenseIt is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint (which. http://dx.doi.org/10.1101/730069doi: bioRxiv preprint first posted online Aug. 8, 2019; 5. Salje H, Cauchemez S, Alera MT, Rodriguez-Barraquer I, Thaisomboonsuk B,Srikiatkhachorn A, et al. Reconstruction of 60 Years of ChikungunyaEpidemiology in the Philippines Demonstrates Episodic and Focal Transmission.Journal of Infectious Diseases. 2016;213(4):604–610. doi:10.1093/infdis/jiv470.6. Cardoso CWC, Kikuti M, Prates AAPPB, Paploski IIAD, Tauro LBL, SilvaMMO, et al. Unrecognized Emergence of Chikungunya Virus during a Zika VirusOutbreak in Salvador, Brazil. PLOS Neglected Tropical Diseases.2017;11(1):e0005334. doi:10.1371/journal.pntd.0005334.7. Hobson D, Curry RL, Beare AS, Ward-Gardner A. The role of serumhaemagglutination-inhibiting antibody in protection against challenge infectionwith influenza A2 and B viruses. J Hyg. 1972;70(4):767–777.8. Coudeville L, Bailleux F, Riche B, Megas F, Andre P, Ecochard R. Relationshipbetween haemagglutination-inhibiting antibody titres and clinical protectionagainst influenza: development and application of a bayesian random-effectsmodel. BMC medical research methodology. 2010;10:18.doi:10.1186/1471-2288-10-18.9. Wu JT, Ho A, Ma ESK, Lee CK, Chu DKW, Ho PL, et al. Estimating InfectionAttack Rates and Severity in Real Time during an Influenza Pandemic: Analysisof Serial Cross-Sectional Serologic Surveillance Data. PLoS Medicine.2011;8(10):e1001103. doi:10.1371/journal.pmed.1001103.10. Wu JT, Leung K, Perera RAPM, Chu DKW, Lee CK, Hung IFN, et al. InferringInfluenza Infection Attack Rate from Seroprevalence Data. PLoS Pathogens.2014;10(4):e1004054. doi:10.1371/journal.ppat.1004054.11. Zhao X, Siegel K, Chen MIC, Cook AR. Rethinking thresholds for serologicalevidence of influenza virus infection. Influenza and Other Respiratory Viruses.2017;11(3):202–210. doi:10.1111/irv.12452.12. Katz JM, Hancock K, Xu X. Serologic assays for influenza surveillance, diagnosisand vaccine evaluation. Expert Review of Anti-infective Therapy.2011;9(6):669–683. doi:10.1586/eri.11.51.August 8, 201924/30.CC-BY-NC 4.0 International licenseIt is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint (which. http://dx.doi.org/10.1101/730069doi: bioRxiv preprint first posted online Aug. 8, 2019; 13. Wood JM, Gaines-Das RE, Taylor J, Chakraverty P. Comparison of influenzaserological techniques by international collaborative study. Vaccine.1994;12(2):167–74.14. Cauchemez S, Horby P, Fox A, Mai LQ, Thanh LT, Thai PQ, et al. Influenzainfection rates, measurement errors and the interpretation of paired serology.PLoS Pathog. 2012;8(12):e1003061.15. Miller E, Hoschler K, Hardelid P, Stanford E, Andrews N, Zambon M. Incidenceof 2009 pandemic influenza A H1N1 infection in England: a cross-sectionalserological study. Lancet. 2010;375(9720):1100–1108.doi:10.1016/S0140-6736(09)62126-7.16. Freeman G, Pereram RAPM, Ngan E, Fang VJ, Cauchemez S, Ip DKM, et al.Quantifying homologous and heterologous antibody titre rises after influenzavirus infection. Epidemiology and Infection. 2016;144(11):2306–2316.doi:10.1017/S0950268816000583.17. Barr IG, Russell C, Besselaar TG, Cox NJ, Daniels RS, Donis R, et al. WHOrecommendations for the viruses used in the 2013ˆa€“2014 Northern Hemisphereinfluenza vaccine: Epidemiology, antigenic and genetic characteristics of influenzaA(H1N1)pdm09, A(H3N2) and B influenza viruses collected from October 2012to January 2013. Vaccine. 2014;32(37):4713–4725.doi:10.1016/J.VACCINE.2014.02.014.18. Huang QS, Bandaranayake D, Wood T, Newbern EC, Seeds R, Ralston J, et al.Risk Factors and Attack Rates of Seasonal Influenza Infection: Results of theSouthern Hemisphere Influenza and Vaccine Effectiveness Research andSurveillance (SHIVERS) Seroepidemiologic Cohort Study. The Journal ofInfectious Diseases. 2019;219(3):347–357. doi:10.1093/infdis/jiy443.19. Katzelnick L, Fonville J, Gromowski G, Arriaga J, Green A, et al. Dengue virusescluster antigenically but not as discrete serotypes. Science.2015;349(6254):1338–1343.20. Priyamvada L, Quicke KM, Hudson WH, Onlamoon N, Sewatanon J, EdupugantiS, et al. Human antibody responses after dengue virus infection are highlyAugust 8, 201925/30.CC-BY-NC 4.0 International licenseIt is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint (which. http://dx.doi.org/10.1101/730069doi: bioRxiv preprint first posted online Aug. 8, 2019; cross-reactive to Zika virus. Proceedings of the National Academy of Sciences.2016;113(28):7852–7857. doi:10.1073/pnas.1607931113.21. Montoya M, Collins M, Dejnirattisai W, Katzelnick LC, Puerta-Guardo H, JadiR, et al. Longitudinal Analysis of Antibody Cross-neutralization Following ZikaVirus and Dengue Virus Infection in Asia and the Americas. The Journal ofInfectious Diseases. 2018;218(4):536–545. doi:10.1093/infdis/jiy164.22. Nachbagauer R, Choi A, Hirsh A, Margine I, Iida S, Barrera A, et al. Definingthe antibody cross-reactome directed against the influenza virus surfaceglycoproteins. Nature Immunology. 2017;18(4):464–473. doi:10.1038/ni.3684.23. Smith DJ, Lapedes AS, de Jong JC, Bestebroer TM, Rimmelzwaan GF,Osterhaus ADME, et al. Mapping the antigenic and genetic evolution of influenzavirus. Science. 2004;305(5682):371–376.24. Lapedes A, Farber R. The Geometry of Shape Space: Application to Influenza.Journal of Theoretical Biology. 2001;212(1):57–69. doi:10.1006/jtbi.2001.2347.25. Cai Z, Zhang T, Wan XF, Dhillon I, Butler E. A Computational Framework forInfluenza Antigenic Cartography. PLoS Computational Biology.2010;6(10):e1000949. doi:10.1371/journal.pcbi.1000949.26. Fonville JM, Wilks SH, James SL, Fox A, Ventresca M, Aban M, et al. Antibodylandscapes after influenza virus infection or vaccination. Science.2014;346(6212):996–1000.27. Kucharski AJ, Lessler J, Cummings DAT, Riley S. Timescales of influenzaA/H3N2 antibody dynamics. PLoS Biol. 2018;16(8):e2004974.28. Yuan HY, Baguelin M, Kwok KO, Arinaminpathy N, van Leeuwen E, Riley S.The impact of stratified immunity on the transmission dynamics of influenza.Epidemics. 2017;.29. Kucharski AJ, Lessler J, Read JM, Zhu H, Jiang CQ, Guan Y, et al. Estimatingthe life course of influenza A(H3N2) antibody responses from cross-sectional data.PLoS Biol. 2015;13(3):e1002082.August 8, 201926/30.CC-BY-NC 4.0 International licenseIt is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint (which. http://dx.doi.org/10.1101/730069doi: bioRxiv preprint first posted online Aug. 8, 2019; 30. Hensley SE, Das SR, Bailey AL, Schmidt LM, Hickman HD, Jayaraman A, et al.Hemagglutinin receptor binding avidity drives influenza A virus antigenic drift.Science (New York, NY). 2009;326(5953):734–6. doi:10.1126/science.1178258.31. Li Y, Bostick DL, Sullivan CB, Myers JL, Griesemer SB, Stgeorge K, et al. Singlehemagglutinin mutations that alter both antigenicity and receptor binding avidityinfluence influenza virus antigenic clustering. Journal of virology.2013;87(17):9904–10. doi:10.1128/JVI.01023-13.32. Bedford T, Suchard MA, Lemey P, Dudas G, Gregory V, Hay AJ, et al.Integrating influenza antigenic dynamics with molecular evolution. eLife.2014;2014(3):e01914. doi:10.7554/eLife.01914.33. Fitch WM, Bush RM, Bender CA, Cox NJ. Long term trends in the evolution ofH(3) HA1 human influenza type A. Proceedings of the National Academy ofSciences of the United States of America. 1997;94(15):7712–8.doi:10.1073/PNAS.94.15.7712.34. Russell CA, Jones TC, Barr IG, Cox NJ, Garten RJ, Gregory V, et al. TheGlobal Circulation of Seasonal Influenza A (H3N2) Viruses. Science.2008;320(5874).35. O’Hara RB, Sillanp¨a¨a MJ. A review of Bayesian variable selection methods: what,how and which. Bayesian Analysis. 2009;4(1):85–117. doi:10.1214/09-BA403.36. George EI, McCulloch RE. Approaches for Bayesian Variable Selection. StatisticaSinica. 1997;7:339–373. doi:10.2307/24306083.37. Plummer M, Best N, Cowles K, Vines K. CODA: Convergence Diagnosis andOutput Analysis for MCMC. R News. 2006;6(1):7–11.38. Gabry J. bayesplot: Plotting for Bayesian models; 2017. Available from:http://mc-stan.org/.39. Foreach Parallel Adaptor for the ’parallel’ Package [R package doParallel version1.0.14];.August 8, 201927/30.CC-BY-NC 4.0 International licenseIt is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint (which. http://dx.doi.org/10.1101/730069doi: bioRxiv preprint first posted online Aug. 8, 2019; 40. Kwok KO, Riley S, Perera RAPM, Wei VWI, Wu P, Wei L, et al. Relativeincidence and individual-level severity of seasonal influenza A H3N2 comparedwith 2009 pandemic H1N1. BMC Infectious Diseases. 2017;17(1):337.doi:10.1186/s12879-017-2432-7.41. Riley S, Kwok KO, Wu KM, Ning DY, Cowling BJ, Wu JT, et al.Epidemiological characteristics of 2009 (H1N1) pandemic influenza based onpaired sera from a longitudinal community cohort study. PLOS Medicine.2011;8(6):e1000442. doi:10.1371/journal.pmed.1000442.42. Lessler J, Riley S, Read JM, Wang S, Zhu H, Smith GJD, et al. Evidence forantigenic seniority in influenza A (H3N2) antibody responses in southern China.PLoS Pathogens. 2012;8(7):e1002802. doi:10.1371/journal.ppat.1002802.43. Lau MSY, Cowling BJ, Cook AR, Riley S. Inferring influenza dynamics andcontrol in households. Proc Natl Acad Sci U S A. 2015;112(29):9094–9099.44. Salje H, Cummings DA, Rodriguez-Barraquer I, Katzelnick LC, Lessler J,Klungthong C, et al. Reconstruction of antibody dynamics and infection historiesto evaluate dengue risk. Nature. 2018;557(7707):719.45. Ranjeva S, Subramanian R, Fang VJ, Leung GM, Ip DKM, Perera RAPM, et al.Age-specific differences in the dynamics of protective immunity to influenza.Nature Communications. 2019;10(1):1660. doi:10.1038/s41467-019-09652-6.46. Gostic KM, Ambrose M, Worobey M, Lloyd-Smith JO. Potent protection againstH5N1 and H7N9 influenza via childhood hemagglutinin imprinting. Science.2016;354(6313):722–726.47. Cobey S, Hensley SE. Immune history and influenza virus susceptibility. Currentopinion in virology. 2017;22:105–111.48. Pepin KM, Kay SL, Golas BD, Shriner SS, Gilbert AT, Miller RS, et al. Inferringinfection hazard in wildlife populations by linking data across individual andpopulation scales. Ecology Letters. 2017;doi:10.1111/ele.12732.August 8, 201928/30.CC-BY-NC 4.0 International licenseIt is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint (which. http://dx.doi.org/10.1101/730069doi: bioRxiv preprint first posted online Aug. 8, 2019; 49. White MT, Griffin JT, Akpogheneta O, Conway DJ, Koram KA, Riley EM, et al.Dynamics of the antibody response to Plasmodium falciparum infection inAfrican children. J Infect Dis. 2014;210(7):1115–1122.50. Guthmiller JJ, Wilson PC. Harnessing immune history to combat influenzaviruses. Current Opinion in Immunology. 2018;53:187–195.doi:10.1016/J.COI.2018.05.010.51. Andrews SF, Huang Y, Kaur K, Popova LI, Ho IY, Pauli NT, et al. Immunehistory profoundly affects broadly protective B cell responses to influenza. SciTransl Med. 2015;7(316):316ra192–316ra192. doi:10.1126/scitranslmed.aad0522.52. Cauchemez S, Hoze N, Cousien A, Nikolay B, Ten bosch Q. How Modelling CanEnhance the Analysis of Imperfect Epidemic Data. Trends in Parasitology.2019;35(5):369–379. doi:10.1016/J.PT.2019.01.009.53. de Lusignan S, Borrow R, Tripathy M, Linley E, Zambon M, Hoschler K, et al.Serological surveillance of influenza in an English sentinel network: pilot studyprotocol. BMJ open. 2019;9(3):e024285. doi:10.1136/bmjopen-2018-024285.54. Metcalf CJE, Farrar J, Cutts FT, Basta NE, Graham AL, Lessler J, et al. Use ofserological surveys to generate key insights into the changing global landscape ofinfectious disease. The Lancet. 2016;388(10045):728–730.doi:10.1016/S0140-6736(16)30164-7.August 8, 201929/30.CC-BY-NC 4.0 International licenseIt is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint (which. http://dx.doi.org/10.1101/730069doi: bioRxiv preprint first posted online Aug. 8, 2019; Table 1. Comparison of run time and posterior sampling efficiency across a range of serosurvey designs.Case studyNumber ofindividualsMeanrun-time(min-utes)Number of ob-served titresNumberoftime segmentsθ ESSZ ESSθ ESS perminuteZ ESS perminute11006.4140041020503015978411006.88400899841000145596011007.87400168714650111591150013.820004102054500073.939500150014.82000810002580067.71740150016.8200016849384050.52291100022.740004981104000043.2457001100026.9400089872200036.78171100031.4400016913454029.1145210012.28004120302990167245210018.84100411360198072.4105210019800821190240062.4126210034.28200821070166031.348.62100378001641980247053.566.92500384000411730186045.448.9250051.24000821630250031.848.8250072.32050041160065122.19.012100073.6800041158091021.412.4250078.640001641550242019.730.8210087.81640016484621009.6323.92100090.48000821550205017.222.7250015041000829254786.173.1921000153410004115305559.993.63210001828000164125022706.8912.52100032782000829262132.830.652500346820001645538371.62.42210006741640001643104160.460.618August 8, 201930/30.CC-BY-NC 4.0 International licenseIt is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint (which. http://dx.doi.org/10.1101/730069doi: bioRxiv preprint first posted online Aug. 8, 2019; View publication stats | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/357282579The Effectiveness of the Blended Learning in Enhancing EFL Learning andCollaborationArticle in World Journal of English Language · December 2021DOI: 10.5430/wjel.v12n1p92CITATIONS5READS1,1901 author:Some of the authors of this publication are also working on these related projects:I was awarded the Medal of Scientific Creativity for an article and presentation of Sudanese fashion at the first scientific conference on fashion and textures and theircomplications in the civilization of Wadi al-Rafidain View projectAssessing English Language Speaking proficiency among Sudanese Secondary Schools' Students View projectFaiza Abdalla ElhussienMajmaah University and Omdurman Islamic University Sudan26 PUBLICATIONS 34 CITATIONS SEE PROFILEAll content following this page was uploaded by Faiza Abdalla Elhussien on 29 December 2021.The user has requested enhancement of the downloaded file.http://wjel.sciedupress.com World Journal of English Language Vol. 12, No. 1; 2022 Published by Sciedu Press 92 ISSN 1925-0703 E-ISSN 1925-0711 The Effectiveness of the Blended Learning in Enhancing EFL Learning and Collaboration Faiza Abdalla Elhussien Mohamed1,2 1 Zulfi College of Education, Majmaah University, Saudi Arabia 2 Omdurman Islamic University, Sudan Correspondence: Faiza Abdalla Elhussien Mohamed, Saudi Arabia, P.O. Box: 1221 Zulfi 11932, Saudi Arabia. Received: November 9, 2021 Accepted: December 16, 2021 Online Published: December 23, 2021 doi:10.5430/wjel.v12n1p92 URL: https://doi.org/10.5430/wjel.v12n1p92 Abstract This descriptive-analytical study investigates undergraduates' perceptions and reflections toward adopting the Blended-Learning system in the university instructions. For this goal, the study employed two tools to collect data. Firstly, a) (n=38) pre-service teachers wrote reflection essays, and b) Google forms closed-ended five scales' questionnaire investigated the academic and interaction indicators, each consisting of thirty items, distributed among (n=110) pre-service teachers. The participants were majoring in English as a Foreign language at multi-stages, Saudi Arabia, 2020/2021. The study employed the content and the SPSS analysis. The questionnaire's results showed the undergraduates' positive perceptions toward combining online and face-to-face learning and how this environment improved their learning outcomes, created a collaborative community, fostered openness for sharing, asking, expressing, and getting talk-worthy ideas. The essays' content analysis reflected the undergraduates' experiences and how they enjoyed transitioning between online and face-to-face learning; they were satisfied by the ability to track their scores. These results created a continuous feedback loop correlated with their academic progress through various activities. However, they encountered a few challenges in online classes like a) missed face-to-face warm environment, b) lost attention, and c) missed instructors' nonverbal cues. Generalized, blended learning facilitated learners' knowledge by reducing education costs, distance, efforts, and time without reducing the students' benefits. The study recommends that the BL instructors have to show some enthusiasm and inspiration. The study proposes future research. Keywords: Blended-learning (BL), blended learning assessment, traditional vs. blended learning, pre-service teachers, and collaboration 1. Introduction Technology has gained an essential part in creating people's intelligent life that revealed through recent research intriguing findings. It highlights all human aspects of life from business, communication, science, and social interaction worldwide. Education, in general, is one of these dimensions, especially in teaching and learning a foreign language to EFL students, online education, in particular, has not only changed the geography of distance education but has significantly impacted advanced education as a whole across the globe. According to Tomlison & Whittaker (2013), blended learning aims to recognize the maximum mix of course delivery to maintain the most effective language learning outcomes. Education institutions have profited from advanced education, the commercialization of education, and the increasing demand for online education (Allen & Seaman, 2011). Blended learning is a transformational concept that intends to sustain collaborative learning; constructive learning and computer-assisted learning both offline learning and online learning (Lalima & Dangwal, 2017) The world advancement is currently witnessing computers, communication technology, and smartphones with internet power a decade ago, besides the help of some learning management systems such as Desire2Learn, Blackboard, Moodle. The essential change in future e-learning environments depends on removing the administrative portions of an LMS to create an e-learning environment that mainly focuses on learning (Jobe, 2011). The appropriate program to be developed is blended learning based on these learning models. Blended learning's tools facilitate what teachers do in the traditional classrooms by the aid of the various types of platforms give a chance to create content, schedule classes, records lessons; it can also provide social interaction facilities as virtual discussion rooms (Siregar et al., 2019) http://wjel.sciedupress.com World Journal of English Language Vol. 12, No. 1; 2022 Published by Sciedu Press 93 ISSN 1925-0703 E-ISSN 1925-0711 Some studies, however, have explored how pre-service teachers can reflect their experiences' growth and development. Exploring these experiences is potentially significant because the researcher observed that some of them are unwilling and unenthusiastic to study through blended learning rather than the traditional learning, saying, "Why to try something new since what I am doing is working well," as claimed by (Tucker, 2019). In this qualitative and quantitative study, the researcher sought to address these gaps by addressing three questions that guided the study: 1. What were the Blended learning benefits on the English-major pre-service teachers' academic performance? 2. How did English-major pre-service teachers evaluate the Blended learning environment's effectiveness on their engagement and interaction? 3. What were the obstacles did the English- major pre-service teachers face in the blended learning courses? The Significance The outcomes from this study will be of significant benefits to the following: Colleges Data given will provide colleges with information about the impact of blended learning on the students' academic progress. These results enable them to accelerate the adoption of more blended learning courses among their educational programs, solve the obstacles faced by the participants and reduce the students' dropping out rates. Instructors Not only do college or students have access to more collaboration through blended learning, as claimed by Schaber et al. (2010), but teachers also need to collaborate frequently with their colleagues to maintain and improve their practice in a blended learning environment. Moreover, the results will help instructors evaluate the quality of the blended learning's roles in advancing their students' academic performance and attitudes to develop the strategies of instructions. Students The study results will encourage and create space for students to engage deeply in collaborative blended works and to foster new ways for enhancing their knowledge, skills. They can develop and improve their learning responsibilities, gain more digital competence systems through blended systems various tools, and develop better practice using the blended learning tools for mastering, managing, and engaging. The Objectives The research aimed to: 1. Explore and assess pre-service teachers' perceptions and reflections toward the effectiveness of the blended-learning on their academics' performance. 2. Investigate pre-service teachers' knowledge about the facilities, forms of instruction, and blended learning interaction to support their learning. 3. Explore the benefits of the BL environment and its roles in increasing the participants' interactions. 4. Identify the fundamental challenges faced by the participants during the blended classes, and suggest solutions 2. Background Blended Learning Blended learning is one of the current directions in education. It supports detailed information about a successful program's strategy, design, and implementation (Michael, Staker & Clayton (2014). Tucker (2019, p. 2) defined blended learning as a combination of online learning with learning in the physical classroom. According to Lim et al. (2017, p. 263), shifting to blended learning requires a significant shift in learning and teaching practice, redefining the role of the teaching staff members and students and scrutinizing the responsibility for learning. Bowyer (2017, P.16) claimed that blended learning adoption improves the course outcomes of the students' retention, pass-rate, improve attainment autonomy, and responsibility towards their education. Kitchenham (2011, p. xiv) explained why hybrid learning has proven to be the most popular choice for students at institutions because of its role in blending some aspects of technology with face-to-face learning. Tucker et al. (2016) defined blended learning as a new formal educational program with elements that help students control their time, place, and pace. Innes, Norman & http://wjel.sciedupress.com World Journal of English Language Vol. 12, No. 1; 2022 Published by Sciedu Press 94 ISSN 1925-0703 E-ISSN 1925-0711 Garrison. (2013, p. 25) stated that the blended learning environment considers a social interaction that sets the environmental conditions in higher education for open cohesion, communication, and interpersonal connection. Learning Assessment Blended learning assessment helps teachers build comfortable environments to motivate learners to interact and allow teachers to observe students' progress. Moreover, to modify instructions to maximize learning outcomes. According to Koc, Lui & Wachira (2015, p 188,) and OECD (2016, P. 9), software, such as Management Learning Systems (MLS), scaffolds blended learning in its job, which defines as "software that offers various ways for assessment such as rubric, quizzes, exams, and discussion. It can assess using: Traditional vs. Blended Learning Traditional teachers lead, instruct, control the learning environment, and use the traditional grading system in the conventional learning process. While in the blended learning system, the students are its center and involved in paced learning (Tucker (2019, p.5). The aims of transitioning from traditional education to blended learning, as stated by (Tucker et al.,2016, p. 67), are to personalize learning from a teacher- as- driver to student- driver. The students drive their knowledge and become primary stakeholders in their education. Teachers help students set their learning self-discipline and goals, seek valuable resources, and build pathways to create learning outcomes. Pre-Service Teachers Pre-service teachers are university undergraduates enrolled in a teacher education program. It enables them to get involved in practical teaching experience under the guidance of a senior teacher. In this experience, the pre-service teachers are progressively trained to be responsible for managing classroom. They are prepared for future professional-level teaching positions Marquardt (2011).. Collaboration According to Garrison (2005), the objective of the collaboration is to build a community in which learners are entirely engaged in collaboratively constructing and meaningful knowledge. Studies show that collaboration in learning helps students work together creatively, away from traditional education learning experience through work with one another (Glassmeyer, Dibbs, & Jensen, 2011). Liu (2007) indicates that an association occurs betwixt the sense of learning community and the perceived learning outcomes and quality. There is abundant confirmation that blended learning innovates strategies consolidating cognitive and practical competence. It consists of studies indicated that the undergraduates’ motivation was developed, displayed deeper understanding and more remembering when asked to resolve problems, and built knowledge from their experiences both in and outside the classroom (Hancock et al., 2010) Ja'ashan (2015) argues the problem that annoys Bisha University level eight male and female students towards Blended English language courses. The survey results revealed that blended learning is as active as face-to-face learning in developing and improving knowledge and skill. On the other hand, the current research results opposed the results of Nazarnkoa (2015) concerning traditional vs. blending learning in which the researcher distributed two electronic Monkey surveys among sixty- two undergraduates' students. They compared the new features of the blended learning course in its new format with a traditional one; most of the students' responses (96%) are favorable toward the course's blended format with few unfavorable impressions as it wastes time and is easy to cheat. In contrast to the above study, Baeer (2018) conducted a case study concerning Blended Learning to improve (n=60) of Al Guds Open University students. The findings showed that blended learning affects positively on the undergraduates' language skills and autonomous learning and motivation. As a result, the study recommends strengthening blended learning to enrich students' knowledge, language skills, and proficiency. Albiladi & Khlod (2019) from Saudi Arabia carried out a study review, discuss, and examine some research that integrates blended learning in EFL and ESL contexts and the educational and interaction merit of this procedure. The study showed that blended learning can capably develop language competence, scaffold the language learning environment, and support students' inspiration to learn the language. Fathelrahman's (2018) study intended to review and document literature and practices related to the different ways to collect feedback from students. It follows the socio-technical approach to management systems and reflective model during distance education courses; the findings indicated that reflecting on the students' feedback could improve the teaching quality of distance learning. Moreover, Okaz (2015, p.1) discussed the pros and cons of integrating blended learning if the instructors incorporate http://wjel.sciedupress.com World Journal of English Language Vol. 12, No. 1; 2022 Published by Sciedu Press 95 ISSN 1925-0703 E-ISSN 1925-0711 blended learning in higher education and provide some hands-on experience and classroom activities to create student-centered and collaborative learning environment. Furthermore, Akbarvo, Gonen, and Haken (2018, p. 2) investigated a hundred and sixty-two students' attitudes towards blended learning and related concepts as a paperless and traditional classroom. The students preferred blended learning to face-to-face, but they chose to have exams in paper and pen form. They trust digital learning materials and link their English language proficiency with combined education. Wei et al., (2017) explored the effectiveness of mixed learning methods on the experimental and control groups of a hundred and four undergraduates' performance and press, they found that the undergraduates who taught through blended learning got the topmost scores than students taught with the conventional methods, but no noteworthy variation in students' academic press between both methods. Many studies examined the Blended Learning approach's impact on EFL students' educational goals and the degree of satisfaction. The present study explores and assesses blended learning tools and environments that influence the participants" academic performance concerning Miller's (2014, p. 19) question about integrating technology in learning: Does technology degrade or improve understanding or have no natural effect at all? The majority of responses revealed that blended learning positively helped the participants to involve in different coactively or independent learning environments, understand, and be confident enough to address and solve the problems encountered in their academic progress. The difference between this study and the previous studies was that the last one investigated the learners' perceptions of integrating blended learning in the learning process using closed-ended tools. In contrast, this study gets the participants' reflection through uncontrolled devices (essays) to reach results and answer the questions about the benefits and obstacles of Blended learning and its environment on their academic and collaboration performance3. 3. Research Methodology This case study uses a mixed-method to explore the blended learning benefits and challenges from the pre-service teachers' perceptions and reflections. The study designed tools to suit the present study scope, objectives and answer the study questions; based on this understanding, the methodology outlined as follows: Participants Two groups (n=148) of English- major pre-service teachers at the English language Education program, Majmaah /University, KSA. were involved in the study. They were divided into a) purposive sampling group of thirty-eight (n= 38), at level five, b) hundred-ten (n=110) from varied levels enrolled in many blended learning course. Study Tools The study employed two tools: Tool 1: A Quantitative Tool An electronic closed-ended questionnaire, as suggested by Tang and Chaw (2013, p. 83), consisted of two indicators: a) academic indicators (cognitive and knowledge) of twenty statements, and b) the environmental indicators (engagement, collaboration, and interaction), of thirteen statements. Tool 2: Qualitative Tool A purposive group (n=38) wrote reflection essays about The role of Virtual Classes on the Students' Academics' Success" to express their opinions and knowledge about the blended learning benefits and challenges. (Taylor, Bogdan, & De Vault (2016, p. 10) Data Collection Procedures The researcher sends a google forms questionnaire's link through the blackboard announcement to the mixed-level English- major pre-service teachers (received only 110 responses) within four weeks. Thirty-eight submit their performative essays to the Blackboard dropbox within two weeks. The complete study preparation and implementation were carried out during COVID-19 Lock-down in the academic year 2020- 2021. The Validity and Reliability The juries proved the questionnaire's overall accuracy. In terms of validity. They asserted that it covered all the research questions in a systematically relevant fashion way, and stated a consistency between the theoretical concepts and the proposed measurement tools. Statistically, as claimed by the traditional test theorists, the most test's validity is the square of the reliability; hence, the http://wjel.sciedupress.com World Journal of English Language Vol. 12, No. 1; 2022 Published by Sciedu Press 96 ISSN 1925-0703 E-ISSN 1925-0711 subjective validity value of the questionnaire statements is as follow: Table 1. The Questionnaire statements' validity and reliability No Dimensions No. of Statements Coefficient of alpha-Cronbach Validity 1 The first Question 20 0.933 0.870 2 The second Question 13 0.926 0.857 Source: Prepared by the researcher from data (2020) The above table 1 indicated that the constancy (Alpha-Cronbach coefficient) of the statements that answered the first question was (0.933), while the second question reached (0.926), which will give the same findings if it uses again in the same condition. Moreover, the validity of the questionnaire's academic and environmental indicators is (0.87%) and (0.85) respectively, which means that the questionnaire is valid for measuring the students’ perceptions towards the blended learning benefits and obstacles. 4. Results The results obtained from the quantitative and qualitative data were organized under the academic and environmental indicators as follow: 4.1 Quantitative Results from the Questionnaire 4.1.1 Academic indicators' Results The following Tables 2 and 3 show the frequency distribution and percentage of the first indicator's statements that answered the study's first question: What were the blended learning benefits on the English-major pre-service teachers' academic performance? Table 2. Descriptive Statistics for the academic indicators' statements from (1-10) Statements Strongly agree Agree Neutral Disagree Strongly disagree 1. Blended learning eased my access to the course materials. F 25 51 25 7 2 % 22.7 46.4 22.7 6.4 1.8 2. Blended learning, supported my intellectual, technical capacity. F 22 53 29 5 1 % 20 48.4 26.4 4.5 0.9 3. Online classrooms were as active as traditional ones. F 11 45 36 13 5 % 10 40.9 32.7 11.8 4.5 4. I got the technical support I need during BL courses. F 15 52 33 8 2 % 13.6 47.3 30 7.3 1.8 5. Blended learning improved my academic achievement, and I scored high marks. F 21 51 29 6 3 % 19.1 46.4 26.4 5.5 2.7 6. Blended learning helped me download and upload the courses materials. F 21 46 36 7 - % 19.1 41.8 32.7 6.4 - 7. Blended courses were the only option that fitted my schedule. F 13 39 42 14 2 % 11.8 35.5 38.2 12.7 1.8 8. It was challenging to take BL courses. F 10 38 42 14 6 % 9.1 34.5 38.2 12.7 5.5 9. Blended learning courses improved my four language skills reading, writing, listening, and speaking. F 14 40 36 15 5 % 12.7 36.4 32.7 13.6 4.5 10. I hope the university adds more English BL courses. F 19 39 34 16 2 % 17.3 35.5 30.9 14.5 1.8 Table 2 shows the percentages of the first indicators’ statements: 1. The first statement: (69.1%) of the respondents agreed with it, and (22.7%) were neutral. In contrast, (8.2%) respondents disagreed with this statement. 2. The second statement: (68.4%) respondents agree with it, and (26.4%) were neutral. In contrast, (5.4%) respondents disagreed with this statement. 3. The third statement: (50.9%) of the respondents agree with it, and (32.7%) were neutral. In comparison, (16.3%) respondents disagreed with this statement. 4. The fourth statement: (60.9%) respondents agree with it, and (30.0%) were neutral. In comparison, (6.4%) respondents disagreed with this statement. http://wjel.sciedupress.com World Journal of English Language Vol. 12, No. 1; 2022 Published by Sciedu Press 97 ISSN 1925-0703 E-ISSN 1925-0711 5. The fifth statement: (65.5%) respondents agree with it, and (26.4%) were neutral. In contrast, (8.2%) respondents disagreed with this statement. 6. The sixth statement: (60.9%) respondents agree with it, and (32.7%) were neutral. In contrast, (6.4%) respondents disagreed with this statement. 7. The seventh statement: (47.2%) respondents agreed with it, and (38.2%) were neutral. Simultaneously, (14.5%) respondents disagree and strongly disagree with this statement. 8. The eighth statement: (43.6%) of the respondents agree with it, and (38.2%) were neutral. In contrast, (18.2%) respondents disagreed with this statement. 9. The ninth statement: (49.1%) of the respondents agree with it, and (32.7%) were neutral. In comparison, (18.1%) respondents disagreed with this statement. 10. The tenth statement: (52.8%) of the respondents agree with it, and (30.9%) are neutral. In contrast, (16.3%) respondents disagreed with this statement. Table 3. Descriptive Statistics for the academic indicators’ statements from (11- 20) Statements Strongly agree Agree Neutral Disagree Strongly disagree 11. Blended learning instant feedback helped me in correcting assignments. F 16 49 35 9 1 % 14.5 44.5 31.8 8.2 0.9 12. I was adequately introduced to the complete courses’ instructions. F 14 49 36 9 2 % 12.7 44.5 32.7 8.2 1.8 13. Sometimes it was difficult for me to attend face-to-face classes. F 9 35 37 23 6 % 8.2 31.8 33.6 20.9 5.5 14. Blended learning assisted and assessed my inputs, unlike the traditional classroom. F 15 39 38 16 2 % 13.6 35.5 34.5 14.5 1.8 15. Adopting blended learning to learn a foreign language is not a good idea. F 10 43 38 16 3 % 9.1 39.1 34.5 14.5 2.7 16. Blended Learning eased my learning. F 22 44 37 5 2 % 20 40 33.6 4.5 1.8 17. I feel comfort in Face-to-face interaction F 27 41 35 5 2 % 24.5 37.3 31.8 4.5 1.8 18. Blended learning helped me to be more productive and more successful F 12 49 33 13 3 % 10.9 44.5 30 11.8 2.7 19. In blended learning classes, teachers could easily follow my progress F 12 44 41 9 4 % 10.9 40 37.3 8.2 3.6 20. Blended learning courses' materials were available to download on smartphones, tablets, and desktops whenever I need them F 24 46 31 6 3 % 21.8 41.8 28.2 5.5 2.7 Q1:What were the blended learning benefits on the English-major pre-service teachers' academic performance? F 332 893 703 216 56 % 15.1 40.6 32.0 9.8 2.5 Table 3 showed the percentages of the academic indicators’ statements: 11. The eleventh Statement: (59.0%) respondents agreed with it, and (31.8%) were neutral. Nevertheless, simultaneously, (9.1%) respondents disagreed with this statement. 12. The twelfth Statement: (57.2%) of the respondents agreed with it, and (32.7%) were neutral. In contra (16.3%) respondents disagreed with the statement. 13. The thirteenth Statement: (40.0%) of the respondents agree with it, and (33.6%) were neutral. In contrast, (26.4%) respondents disagreed with this statement. 14. The fourteenth Statement: (49.1%) respondents agreed with it, and (34.5%) were neutral. In comparison, (16.3%) respondents disagreed with this statement. 15. The fifteenth Statement: (38.2%) respondents agreed with it, and (34.5%) were neutral. In contrast, (17.2%) respondents disagreed with this statement. 16. The sixteenth Statement: (60.0%) respondents agreed with it, and (33.6%) were neutral. In comparison, (6.4%) respondents disagreed with this statement. 17. The seventeenth Statement: (61.8%) respondents agreed with it, and (31.8%) were neutral. In comparison, (6.4%) respondents disagreed with this statement. http://wjel.sciedupress.com World Journal of English Language Vol. 12, No. 1; 2022 Published by Sciedu Press 98 ISSN 1925-0703 E-ISSN 1925-0711 18. The eighteenth Statement: (55.4%) respondents agreed with it, and (30.0%) were neutral. Simultaneously, (14.6%) respondents disagreed with this statement. 19. Nineteenth Statement: (50.9%) respondents agreed with it, and (37.3%) were neutral. In comparison, (11.8%) respondents disagreed with this statement. 20. Twentieth Statement: (63.6%) respondents agreed with it, and (28.2%) were neutral. In contrast, (8.2%) respondents disagreed with this statement. To conclude, regarding the participants' responses to the first question: What were the Blended learning benefits on the English-major pre-service teachers' academic performance? (55.7%) respondents strongly agree with it, and (32.0%) were neutral. In contrast, (12.3%) respondents disagreed. 4.1.2 Environmental Indicators’ Results The following are the frequency distribution and percentage of the first indicator’s statements that answered the study's second question: How did English-major pre-service teachers evaluate the Blended learning environment's effectiveness in their engagement and interaction? Table 4. Descriptive Statistics for Environmental Indicators Statements Strongly agree Agree Neutral Disagree Strongly disagree 1. I was highly motivated through blended learning environment. F 20 0 38 10 2 % 18.8 36.4 34.5 9.1 1.8 2. I like teachers’ non-verbal expressions like eye contact, facial expressions, and gestures through face-to-face classes. I feel closer. F 25 47 29 9 - % 22.7 42.7 26.4 8.2 - 3. Blended learning helped me and my classmates to regularly evaluate each other. F 12 44 38 14 2 % 10.9 40.0 34.5 12.7 1.8 4. Engagement, motivation, and interaction through BL learning were the factors that enhanced my success. F 12 52 33 11 2 % 10.9 47.3 30 10 1.8 5. Online classes were less annoying than the traditional ones. F 22 39 28 19 2 % 20 35.5 25.5 17.3 1.8 6. Online classroom activities made me feel less involved. F 18 43 3 17 1 % 16.4 39.1 28.4 15.5 0.9 7. BL trained me in a new interactive learning methodology. F 15 46 40 9 - % 13.6 41.8 36.4 8.2 - Table 4 explains the percentage of the second hypothesis (Environmental Indicators) as: 1. The first statement: (54.6%) of the respondents agree with it, and (34.5%) are neutral. In contrast, (0.9%) of the respondents disagree with the first statement. 2. The second statement: (65.4%) of the respondents agree with it, and (26.4%) are neutral. In contrast, (8.2%) of the respondents disagree with the second statement. 3. The third statement: (50.9%) of the respondents agree with it, and (34.5%) were neutral. In contrast, (14.5%) respondents disagreed with this statement. 4. The fourth statement: (58.2%) respondents agreed with it, and (30.0%) were neutral. In contrast, only (11.8%) respondents disagree with this statement. 5. The fifth statement: (55.5%) of the respondents agreed with it, and (25.5%) were neutral, whereas (19.0%) respondents disagreed with this statement. 6. The sixth Statement: (55.5%) of the respondents agreed with it, and 28.2%) were neutral. In contrast, (16.4%) respondents disagreed with this statement. 7. The seventh Statement: (55.4%) respondents agreed with it, and (36.4%) were neutral. In contrast, (8.2%) respondents disagreed with this statement http://wjel.sciedupress.com World Journal of English Language Vol. 12, No. 1; 2022 Published by Sciedu Press 99 ISSN 1925-0703 E-ISSN 1925-0711 Table 5. Descriptive statics for Environmental Indicators Statements Strongly agree Agree Neutral Disagree Strongly disagree 8. I felt isolated during online courses, unlike face-to-face ones. F 18 46 27 15 4 % 16.4 41.8 24.5 13.6 3.6 9. Blended learning eased my connection during my off days. F 20 49 28 12 1 % 18.2 44.5 25.5 10.9 0.9 10. Blended learning provided me with feedback from my peers and teachers on my performance. F 17 44 36 10 3 % 15.5 40.0 32.7 9.1 2.7 11. Blended learning chat rooms encouraged me to discuss topics covered previously in the traditional classes. F 17 48 33 12 - % 15.5 43.6 30.0 10.9 - 12. Blended learning interactive tools enhanced my learning input and output. F 13 49 37 9 2 % 11.8 44.5 33.6 8.2 1.8 13. I would to take blended learning courses in the future. F 19 44 31 12 4 % 17.3 40.0 28.2 10.9 3.6 Q2: How do English-major pre-service teachers evaluate the Blended learning environment's effectiveness in their engagement and interaction? F 228 551 401 159 23 % 16.7 40.5 29.4 11.7 1.7 Table 5 showed the following: 8. The Eighth Statement: (58.2%) of the respondents agree with it, and (24.5%) were neutral. In contrast, (17.2%) respondents disagreed with this statement. 9. The Ninth Statement: (62.7%) respondents agreed with it, and (25.5%) were neutral. In contrast, (11.8%) respondents disagreed with this statement. 10. The Tenth Statement: (55.5%) respondents and agreed with it, and (32.7%) were neutral. In comparison, (11.8%) respondents disagreed with this statement. 11. The Eleventh statement: (59.1%) respondents agreed with it, and (30.0%) were neutral. However, at the same time, (10.9%) respondents disagreed with this statement. 12. The Twelfth Statement: (56.3%) of the respondents agree with it, and (33.6%) were neutral. In comparison, (10.1%) of the respondents disagreed with this statement. 13. The Thirteenth Statement: (57.3%) of the respondents agreed with it, and (28.2%) were neutral. In contrast, (14.5%) of the respondents disagreed with this statement. To conclude, the second question asked about the environmental Indicators: (How did English-major pre-service teachers evaluate the Blended learning environment's effectiveness in their engagement and interaction? (57.2%) respondents agreed with it, and (29.4%) were neutral. In contrast, (13.4%) respondents disagreed. 4.2 Qualitative Results from the Written Essays Responses presented below were generated from the informative essays' contents analysis written by thirty-eight of the purposive sample group, entitled "The virtual classes and the student's academic success. Below is a summary of the students' opinions that answered the study's three questions. 4.2.1 Results Answering the First Question The following are the participants' written reflections that answered the study's first question: What were the Blended learning benefits on the English-major pre-service teachers' academic performance? The Purposive Group Reflection on the Blended Learning Benefits: The Blended learning a)saved the continuity of the academic year 2020-2021 without taking the risk of COVID-19's infection, and guarantee the students' academic success, b) has powerful tools not found in the physical space, and it paved the way to the students' success, c) it made the courses flexible, easy, fast, and enjoyable, d) it enabled a better understanding through online activities and participation, e) allowed instant feedback and instant grading, f) diverted learning methods, g) videos, texts, projects, presentations, discussions, helped us to absorb the content easily, h) facilitated obtaining the courses' materials, i) it eased the course works' evaluation through different fair online methods, j) the online tests' grades were fairer than papers- pens' tests, k) helped the students to be more productive by managing http://wjel.sciedupress.com World Journal of English Language Vol. 12, No. 1; 2022 Published by Sciedu Press 100 ISSN 1925-0703 E-ISSN 1925-0711 and allocating extra time to do other extra urgent tasks, l) it was an evidence of a good indication for running a better education in our region, m) It eased the formation of virtual groups' discussions. Overall, the participants shared positive opinions on their academic progress towards blended learning. 4.2.2 Results Answering the Second Question The participants' written reflections answered the second question: How do English-major pre-service teachers evaluate the Blended learning environment's effectiveness in their engagement and interaction? The responses show that the blended learning environment was: a) feasible in saving time and money spent preparing the traditional courses' projects, assignments, and daily university costs. b) It allowed reviewing the recorded lectures or taking exams anywhere. c) it replaced the papers' pens' tasks with phones, tablets, or computers. d) It reduced the geographical distance that helped manage cost and time freely e) had an observable change in the students' bad health habits by buying the university's expensive junk foods and beverages. f) enhanced the collaboration and communication between the students and teachers. g) reached a vast number of students at the different zones, h) protected shy students, and encouraged them to ask questions and engage in online discussions. i) allowed the students, who had a part-time job, to attend classes from their workplace, j) enabled them to highlight the essential points and study more than usual. Overall, the students' agreed that blended learning's environment had significant impacts on increasing the number of engagements and interactions, which positively correlated with their academic performance. 4.2.3 Results Answering the Third Question Concerning the Obstacles The following is a summary generated from the participants' written reflections toward the blended learning challenges, which answer the study's third question: What obstacles did the English- major pre-service teachers face in the blended learning courses? a) the blended learning virtual classes lacked supervision and consultation, b) it required computers and a stable internet connection, and some students do not have, c) it could not control over students' total attendance; and finally, d) the physical absence of the teachers made it difficult for some students to concentrate well. All responses focused on the instructors' physical absence, which led to a lack of concentration and actual attendance among some irresponsible students. In some cases, as reflected by a few participants, they were experienced the followings a) missed the energetic and warm climate of face-to-face courses, b) missed instructor's body language and nonverbal signals, c) the physical absence of their educators drove the participants to lose consideration., finally, d) some online courses lack administration, eagerness, and motivation. Study Limitations The researcher believes no study is perfect. Therefore, this study tried to avoid sampling errors that led to shortcomings and designed the qualitative and quantitative methodology in a simple random sampling that covered a wide range of respondents. These influenced positively the interpretation of the results of the study. 5. Discussion The present study explored and assessed the students' perceptions and reflections toward blended learning's useful, easy, clear, feasible tools and supported environments on their academic success, interactions, and engagement. The questionnaire and the essays written by the purposive sample answered the study's three questions. The results indicated that: 1. The blended learning courses facilitated encouraged them to download materials, upload assignments, carry out online examinations, and encourage the participants to engage in discussions, asking or answering questions their instructors and classmates raised. 2. Blended learning environment developed their EFL four skills' interaction. 3. Helped them move beyond the traditional learning methods and take part in the learning processes 4. It paved the way for them to acquire the skills needed for their future career. See appendixes' links (1) and essays' samples. Moreover, the quantitative and qualitative results show that blended learning is a beneficial approach to diversifying teaching methods and positively affects learning inputs and outputs. Furthermore, the study agreed with (Napier & Smith, 2009) that the blended learning helped the participants be http://wjel.sciedupress.com World Journal of English Language Vol. 12, No. 1; 2022 Published by Sciedu Press 101 ISSN 1925-0703 E-ISSN 1925-0711 independent learners and achieve higher grades. They showed positive and vigorous agreements on most blended learning academic and environment indicators. For instance, students feel more comfortable interacting with classmates during virtual classes than face-to-face ones. In addition, the instant feedback inside the virtual courses helped them to correct their mistakes. The findings agree, support, and go in one line with Millar's theory (2014), Nazarenkoa (2015), Albiladi & Khlood (2019), and Bakeer (2018) that most of the participants were favorable the blended learning format. However, the study contrasts with some responses that believed combined learning wastes time, is easy to cheat in exams, and creates social isolation. The results of the current study contrast with Ja'ashan's study (2015) from Saudi Arabia, in which the students saw traditional learning as active as blended learning. Whereas the survey agreed with the results reached by Fathelrahman (2018) from King Faisal University, Alhasa, KSA, that feedback from the students could improve the teaching quality of distance learning, which leads to students' better understanding, the thing that employed by the current study. Concerning the obstacles, most participants agreed upon item four in the electronic survey, which attracted the researcher's attention; it rated (74.3%) regarding the blended learning environment's impact on interaction and collaboration. Unlike the traditional classroom's warm climate, the participants stated they encountered emotional deprivation during virtual courses; they missed the non-verbal expressions, body language, eye contact, and non-verbal message exchange with our teachers during the face-to-face classes. They also complained about the unstable internet connection in their geographical area. Some stated that the absence of teachers' control made some students lose concentration during lectures. Others indicated that they chose the BL course because it was the only one that fitted their schedule. Few students assumed that the virtual classes lacked supervision to control students' actual attendance. To conclude, the results answered the research questions that blended learning met their academic performance expectations, success, built their EFL skills, eased engagement, individual reflection, or group presentations. It was less annoying, facilitating the interaction. Chat rooms encouraged them to incorporate materials from the discussion. 6. Conclusion The blended learning system increased the participants' academic performance through useful, easy, flexible, clear, and feasible tools. Its comfortable environment increased the participants' interaction, engagement, collaboration, and communication skills. They also became more self-regulated, self-dependents, friendly, and confident. In addition, it enables them to build strong relationships with their classmates and instructors. The researcher observed a significant satisfaction among the participants toward the blended learning tools, which enabled them to record sessions for later reviewing, usefulness, clarity and, feasibility that encouraged long-term learning. At the same time, there was a bit, negative emotional feeling regarding their instructors' absolute physical absence and the lack of warm face-to-face interaction. Finally, integrating blended learning in the academic program enables the university to reach the most distinguished and prominent ranks. Implication and Gaps Filling Investigating the impact of blended learning and its environment on English-major pre-service teachers' collaboration and academic performance puts the study among the research that highlights the roles of technology in advancing knowledge and empowering learners' mental, cognitive, and social forms. Furthermore, the study sought to collaborate with the previous studies by conducting more future research to explore the different viewpoints of blended learning users, which motivated the current research to move the existing investigation forward. Few studies have examined, described, and assessed the significant impact of blended learning on students' academic performance using questionnaires supported by written essays. The participants informatively expressed their attitudes freely toward the effects of integrating blended learning to enhance teaching and learning delivery. 7. Recommendations The researcher recommends the blended learning instructors to: a) support, inspire and encourage the students to feel friendly; b) create a dynamic environment, give more elements of the course, and monitoring it too; c) encourage the students to be open up by spending few minutes at the beginning of a blended learning class, to engage in ice-breaking subjects to create a warm and inviting atmosphere. The study suggests investigating the correlation between the students' psychological and emotional status with their http://wjel.sciedupress.com World Journal of English Language Vol. 12, No. 1; 2022 Published by Sciedu Press 102 ISSN 1925-0703 E-ISSN 1925-0711 academic progress for future research. Concerning the obstacles, most participants agreed upon one item in the electronic survey, which attracted the researcher's attention, the fourth item. It rated (74.3%) the Blended learning environment's impact on interaction and collaboration. The participants stated that they encountered emotional separation during virtual courses, unlike the traditional classroom's warm climate. In addition, they added, "We missed the non-verbal expressions, body language, eye contact, and non-verbal message exchange with our teachers during the face-to-face classes." They faced an unstable internet connection in their geographical area. Some stated that the absence of teachers' control made some students lose concentration during lectures. Others indicated that they chose blended learning because it was the only one that fitted their schedule. Acknowledgement The author extends her appreciation to the Deanship of Scientific Research at Majmaah University for supporting this work research under grant No. 94/1439. Thankfulness goes to my brother Eng. Abdulwahab Elhussein for his precious technical support and to the participants who willingly participated in this research. References Akbarov, A., Gonen, K., & Aydogan, H. (2018). Students’ Attitudes toward Blended Learning in EFL Contexts. Acta Didactica, Napocensia, 11(1). https://doi.org/10.24193/adn.11.1.5 Albiladi, W. S., & Khlood, K. A. (2019). Blended Learning in English Teaching and Learning: A Review of the Current Literature. Journal of Language Teaching and Research, 10(2), 232-238. https://doi.org/10.17507/jltr.1002.03 Allen, I. E., & Seaman, J. (2011). Going the distance: Online education in the United States. The Online Learning Consortium. Retrieved from http://sloanconsortium.org/publications/survey/ going_distance_2011 Bakeer, A. M. (2018). Students' Attitude towards Implementing Blended Learning in Teaching English in Higher Education Institutions: A Case of Al-Quds Open University. International Journal of Humanities and Social Science, 8(6). https://doi.org/10.30845/ijhss.v8n6a15 Bakerson, M., Trottier, T., & Mansfield, M. (2015). The Value of Embedded Formative Assessment: An Integral Process in Online Courses. Bowyer, J. (2017). Blending Learning: Bringing the Elements Together. Research Matters: A Cambridge Assessment Publications. Retrieved from http://www.cambridge assessment.org.uk/research-matter Fathelrahman, A. (2018). Using Reflection for Improving Distance Learning Course Delivery: A Case of Teaching Management Information Systems Course. The Journal of Open, Distance, and e-Learning. https://doi.org/10.1080/02680513.2018.1508338 Garrison, D. R. (2005). Learning Collaboration Principles. Paper presented for the Sloan-C Summer Workshop, Victoria, BC, Canada, 2005. https://doi.org/10.5539/elt.v8n9p40 Grassmeyer, D., Dibbs, R., & Jensen, T. (2011). Determining utility of formative assessment through virtual community: Perspectives of online graduate students. The Quarterly Review of Distance Education, 12(1), 23-35. Hancock, T., Smith, S., Timpte, C., & Wunder, J. (2010). PALs: Fostering Student Engagement and Interactive Learning. Journal of Higher Education Outreach and Engagement, 14(4), 37. Innes, M. C., Norman D. V., & Garrison, R. D. (2013). Teaching in Blended Learning Environments: Creating and Sustaining ..Pub. AU Press, Athabasca University. Ja’ashan, M., & N., H., M (2015), Perceptions and Attitude towards Blended Learning for English Language Courses. A case Study of Students at the University of Bisha. English Language Teaching, 8(9). Jobe, W. (2011). HIVE (Hypermedia Interactive Virtual Environment) – Utilizing HTML5 and instructional design to improve online learning. Paper presented at the Society for Information Technology & Teacher Education International Conference 2011. Retrieved from http://www.editlib.org/p/36820 Koc, S., Liu, X., & Wachira, P. (2015). Assessment in Online and Blended Learning Environment. Information Age Publishing. Inc. The USA. http://wjel.sciedupress.com World Journal of English Language Vol. 12, No. 1; 2022 Published by Sciedu Press 103 ISSN 1925-0703 E-ISSN 1925-0711 Lalima & Dangwal, K. L. (2017), Blended Learning: An Innovative Approach. Universal Journal of Educational Research, 5(1), 129-136. https://doi.org/10.13189/ujer.2017.050116 Lim, C. P., & Wang, L. B. (Eds.) (2016). Blended Learning for Quality Higher Education: Selected Case Studies on Implementation from Asia-Pacific. Paris, France: UNESCO. Liu, X. J., Magjuka, R., Bonk, C. J., & Lee, S. H. (2007). Does a sense of community matter? An examination of participants’ perceptions on building learning communities in online courses. The Quarterly Review of Distance Education, 8(1), 9-24.565-9. Paris, France. Marquardt, S. K. (2011). (Re)telling: A narrative inquiry into pre-service TESOL teachers’ study abroad experiences (Unpublished doctoral dissertation). Michigan State University, Ann Arbor, MI. ERIC Number: ED533232, ISBN: ISBN-978-1-1248-5886-9. Michael, B. H., Staker, H., & Clayton, N. C (2014), Blended: Using Disruptive Innovation to Improve Schools 1st Edition. Library of Congress, USA. Miller, M. D. (2014). Minds Online: Teaching Effectively with Technology. Harvard University Press. London, England. https://doi.org/10.4159/harvard.9780674735996 Napier, N. P., & Smith, S. (2009). Assessing Blended Learning: Student Outcomes and Perceptions. Presented at Americas Conference on Information Systems, San Francisco, California. Nazarenkoa, A. L. (2015). Blended Learning vs. Traditional Learning: What Works? (A Case Study Research). Procedia - Social and Behavioral Sciences, 200(2015), 77-82. https://doi.org/10.1016/j.sbspro.2015.08.018 OECD. (2016). Innovating Education and Educating for Innovation: The Power of Digital Technologies and Skills. OECD Publishing, aris. https://doi.org/10.1787/9789264265097-en Okaz, A. (2015). Integrating Blended Learning in Higher Education: Procida: Social and Behavioral Science. https://doi.org/10.1016/j.sbspro.2015.04.086 Siregar, N., Siregar, T. M., & Siregar, B. H. (2019). Blended learning in students' views. IOP Conf. Series: Journal of Physics: Conference Series. https://doi.org/10.1088/1742-6596/1188/1/012099 Tang, C. M., & Chaw, L. Y. (2013). Readiness for Blended Learning: Understanding Attitude of University Students. International Journal of Cyber Society Education, 6(2), 79-100. https://doi.org/10.7903/ijcse.1086 Tomlinson, B., & Whittaker, C. (Eds.). (2013). Ed, Blended Learning in English, Language Teaching: Course Design and Implementation. British Council, 2013 Brand and Design/D057. 10 Spring Garden London SW1A 2BN, UK. Tucker, C. R., Wycoff, T., Green, J. T. (2016). Blended Learning in Action: A Practical Guide Toward Sustainable Change (Corwin Teaching Essentials). SAGE –Crown Press. Delhi, India. Tucker, K. R. (2019). Power Up Blended Learning: A Professional Learning Infrastructure to Support Sustainable Change (Corwin Teaching Essentials). 1st Edition. SAGE Production, USA. Tylor, J. S., Bogdan, R., & M., DeVault, L. (2016). Introduction to Qualitative Research Methods- a Guidebook and Resource 4th ed. Published by John Wiley & Sons, Inc., Hoboken, New Jersey. Wei, Y., Shi, Y., Yang, H. H., & Liu, J. (2017). Blended Learning versus Traditional Learning: A Study on Students’ Learning Achievements and Academic Press”. A Conference Paper. https://doi.org/10.1109/ISET.2017.57 Appendix (1) The Survey link: https://docs.google.com/forms/d/1sXQMXmCyTG78CzFYIejtt9GqBT_wEdLwgzzK_E-uBfk/edit Copyrights Copyright for this article is retained by the author(s), with first publication rights granted to the journal. This is an open-access article distributed under the terms and conditions of the Creative Commons Attribution license (http://creativecommons.org/licenses/by/4.0/). View publication stats | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/337089289Impact of a Modified Fenton Process on the Degradation of a ComponentLeached from Microplastics in Bottom SedimentsArticle in Catalysts · November 2019DOI: 10.3390/catal9110932CITATIONS6READS1273 authors:Some of the authors of this publication are also working on these related projects:Visegrad and Ukraine Dialogues on Climate Change and Sustainable Development View projectMałgorzata KidaRzeszów University of Technology39 PUBLICATIONS 296 CITATIONS SEE PROFILESabina ZiembowiczRzeszów University of Technology29 PUBLICATIONS 215 CITATIONS SEE PROFILEPiotr KoszelnikRzeszów University of Technology94 PUBLICATIONS 681 CITATIONS SEE PROFILEAll content following this page was uploaded by Piotr Koszelnik on 07 November 2019.The user has requested enhancement of the downloaded file. Catalysts 2019, 9, 932; doi:10.3390/catal9110932 www.mdpi.com/journal/catalysts Article Impact of a Modified Fenton Process on the Degradation of a Component Leached from Microplastics in Bottom Sediments Małgorzata Kida *, Sabina Ziembowicz and Piotr Koszelnik Department of Chemistry and Environmental Engineering, Faculty of Civil and Environmental Engineering and Architecture Rzeszów University of Technology, al. Powstańców Warszawy 6, 35-959 Rzeszów, Poland; s.ksiazek@prz.edu.pl (S.Z.); pkoszel@prz.edu.pl (P.K.) * Correspondence: mkida@prz.edu.pl; Tel: +48 17 743 24 07 Received: 7 October 2019; Accepted: 5 November 2019; Published: 7 November 2019 Abstract: This paper describes work to assess the possibility of a modified Fenton process being used to remove the hard-to-degrade plasticizer di(2-ethylhexyl) phthalate (DEHP) from the bottom sediments of a reservoir. The modifications in question entail iron(II) ions being replaced by iron(III), as well as facilitation of the process using a chelating agent. Analysis further revolved around the impact of such factors as amounts of reagents, reaction of the environment, initial contents of the contaminant, and the presence of other “competing” contaminants also of a hard-to-decompose nature. As the maximum efficiency of DEHP removal obtained did not exceed 30%, the low susceptibility to degradation is made clear, as is the need for earlier desorption of the contaminant from the matrix. The effect of the modified Fenton process on the content of organic matter and dissolved organic carbon was also considered, as was the tendency to cause selected metals and plant nutrients to leach from bottom sediments. Keywords: di(2-ethylhexyl) phthalate; modified Fenton process; bottom sediments; persistent organic pollutants; 1. Introduction In recent years, special attention has been paid to researching and deploying so-called advanced oxidation processes (AOPs). While these are varied, a common feature is their generation of hydroxyl radicals (HO.) capable of reacting with almost all organic compounds. Due to the high efficiency they achieve in degrading most organic pollutants, oxidation methods are now seen increasingly as the most promising alternative methods of purification or decontamination to be set against so-called conventional methods [1–6]. The Fenton process is one of the advanced methods of oxidation, whereby a reaction between hydrogen peroxide (H2O2) and catalyst Fe2+ generates the aforesaid hydroxyl radicals, which manifest marked reactivity and strong oxidizing properties arising out of their high oxidizing potential. The radicals react non-selectively with organic impurities, oxidizing them to such intermediates as alcohols or carboxylic acids, and then to water and carbon dioxide [4]. While the classic Fenton process already has many advantages—not least high efficiency, a lack of exacting requirements where equipment is concerned, wide availability of the reagents, and a lack of harmful by-products given the capacity to degrade organic compounds to H2O and CO2—an array of further modifications has arisen, to increase effectiveness and reduce costs still further. These modifications include the use of alternative catalysts and sources of hydrogen peroxide, a more heterogeneous process, and deployment in combination with other processes [7,8]. Catalysts 2019, 9, 932 2 of 10 In the work described here, the aim was to modify the Fenton process to achieve optimal removal of di(2-ethylhexyl) phthalate (DEHP) from bottom sediments. The modification we arrived at used Fe3+ ions instead of Fe2+ in the reaction environment, with the result being a steadier formation of hydroxyl radicals that raised the level of process efficiency. Degradation of organic compounds in the presence of Fe3+ ions was slower than with Fe2+, and this was seen as advisable in bottom-sediment matrix and soil, in line with the need for earlier desorption of pollutants into the aqueous phase [7]. Di(2-ethylhexyl) phthalate is an organic compound from the group of phthalic acid esters (PAEs), which are regarded as hazardous, given their capacity to cause genetic aberrations, affect reproduction and further development negatively, and give rise to endocrine disorders [9–11]. DEHP is the most common compound of the phthalate group to pollute the natural environment, and its low level of solubility in water combines with its unfortunate stability in the environment to ensure accumulation in bottom sediments, where the content may even reach 322 mg/kg d.w. Its high content in the aquatic environment is associated with leaching from plastics, and in particular from microplastics, where it acts as a plasticizer [9,12]. This phenomenon accounts for the urgency of the search for means of degradation, including the modified Fenton process detailed here. The purpose of the work detailed here was thus to apply a modification of the Fenton process in removing high contents of DEHP from bottom sediments. Research encompassed the possibility of the reaction being conducted in a natural reaction environment with no need for acidification (bearing in mind the negative effects on bottom sediments or soil), in the presence of the chelating agent sodium pyrophosphate Na4P2O7 (PS). The latter represents the group of inorganic chelating substances that differ from their organic counterparts in competing less for access to hydroxyl radicals. Analysis concerned the impact of the process applied on the contents of soil organic matter (SOM) and dissolved organic carbon (DOC). The possibility of various components in bottom sediments (including plant nutrients) being leached was also investigated, as was competition for access to hydroxyl radicals in the presence of other polycyclic aromatic hydrocarbon (PAH) pollutants and organochlorine pesticides. The work also sought to check the suitability of the process developed in the presence of various initial contents of DEHP. 2. Results and Discussion 2.1. Impact of Reagent Dose on DEHP Degradation Studies of reagent dose in relation to DEHP distribution in a solid matrix involved DEHP:H2O2:Fe3+ molar ratios of 1:1:1, 1:5:1, 1:28:1, 1:100:1, and 1:1:5, at constant DEHP content equal to 0.13 mM/kg d.w., for times equal to 1, 2, 4, 12, and 24 h, at pH = 3. The results of these tests are presented in Figure 1. Figure 1. Impact of reagent doses in the di(2-ethylhexyl) phthalate (DEHP):H2O2:Fe3+ molar ratio on the efficiency of DEHP removal at pH 3. Ct denotes the DEHP content in bottom sediments after time t, and C0 is the initial content of di(2-ethylhexyl) phthalate in bottom sediments. Catalysts 2019, 9, 932 3 of 10 The impact of hydrogen peroxide dose on the efficiency of DEHP oxidation was tested across a range of oxidant quantities from 0.13 to 13 mM/kg d.w. The maximum observed efficiency of di(2-ethylhexyl) phthalate removal following the introduction of catalytic Fe3+ ions into the reaction environment was 25.49% for t = 24 h and a DEHP:H2O2:Fe3+ molar ratio of 1:1:1. Larger amounts of H2O2 in relation to iron(III) ions did not result in fuller degradation of the contaminant. Furthermore, a relative excess of Fe3+ (with 1:5 H2O2:Fe3+) was associated with effectiveness at just 21.07% on average. The modified Fenton process proved most effective during the first four hours of reaction. Likewise, in the case of their H2O2/Fe3+ process, Chiou et al. (2006) [11] achieved a 46% reduction in amounts of dibutyl phthalate (DBP) via a reaction involving Fe3+ at 0.36 mM/dm3, H2O2 at 0.032 mM/min·dm3, t = 60 min, and pH = 3. This degree of DBP removal (C0 = 5 mg/dm3) was attributed to the presence of HO. and HO2. formed as H2O2 and Fe3+ interacted (Fe3+ + H2O2 → Fe2+ + HO2. + H +, followed by Fe2+ + H2O2 → Fe3+ + HO. + OH-). Also using an Fe3+ catalyst, Barbusiński (2004) [7] and Jorfi et al. (2011) [13] asserted that reaction rate was limited by Fe2+ formation. HO. radicals are generated in a two-stage process, via a slow reaction between Fe3+ and H2O2, and a subsequent rapid reaction between generated Fe2+ and H2O2. The H2O2-mediated elimination of hard-to-decompose substances exemplified by the non-ionic surfactant Triton X-114 was studied by Ledakowicz et al. (2001) [14], among others. After 26 hours, they achieved a reduction of 100 mg/dm3 of this substance at a level of just 4%, in the case of a 0.15 mmol/dm3 concentration of hydrogen peroxide. However, a sixfold increase in the amount of oxidant raised this level of removal to 24%. In research by Goi and Viisimaa (2015) [15], hydrogen peroxide (at a soil:H2O2 ratio of 1:0.0005 w/w) was used to degrade selected polychlorinated biphenyls (PCBs) present at 52 g/kg d.w. The limited level of removal (at just 10% for t = 60 min) in this case too, was probably due to rapid decomposition of the oxidant. For their part, Zhang et al. (2013) [16] applied a Fenton process to reduce levels of petroleum hydrocarbons (PHCs) in oily sludge. The best level of removal they were able to achieve was 13.8%. The degradation of organic compounds was slower in the presence of Fe3+ as opposed to Fe2+, and —where pollutants are present in bottom sediments and soil—it is advisable that they be subject to earlier desorption into the water phase, followed by rapid decomposition of hydrogen peroxide to oxygen and water [17,18]. 2.2. Impact of pH on DEHP Degradation The process was investigated in a reaction environment with pH values of 3, 5, 7.95, or 10, where the dose of hydrogen peroxide was equal to 0.13 mM/kg d.w., and the amount of catalyst was set at 0.13 mM/kg d.w., with t = 1, 2, 4, 12, and 24 h (Figure 2a). Figure 2. Impact on the decomposition of DEHP using a modified Fenton process due to: (a) pH (with H2O2:Fe3+ at 1:1); (b) a sodium pyrophosphate (PS) chelating agent (with H2O2:Fe3+:PS at 1:1:1). Catalysts 2019, 9, 932 4 of 10 Oxidation of the DEHP in bottom sediments proved least effective where the tested reaction environment was alkaline. Slightly higher values for removal (16.27%–22.10% over 1–24 h time intervals) were obtained for pH = 5. However, across the pH range analyzed, the best effects obtained were at pH = 3, with 25.49% removal on average achieved after 24 hours. However, the reaction in fact proved effective over the first four hours of the process only, with no significant further change thereafter. The process taking place in medium with a natural (unmodified) reaction was found to be comparable to that taking place in an acid reaction medium. Compared with pH = 3 (depending on the duration of the reaction), differences in the efficiency of removal of DEHP were just 0.1%–4.96%. This is an encouraging finding in the light of the more typical situation, whereby proven higher efficiency of removal of impurities at pH = 3 motivates soil acidification in support of the Fenton process, in spite of the negative impact on soil properties and quality. According to Bokare and Choi (2014) [19], at pH = 6, most Fe3+ ions are in the form of Fe(OH)3 and Fe(OH)2+. The presence of a chelating agent allows for retention of the Fe3+ ions formed in the course of the Fenton reaction, or introduced in a dissolved form at a higher pH. Under such conditions, the effects are usually less favorable than in an acidic environment [7]. This compares with our own research, in which only slight differences were noted in the degradation of DEHP achieved in the natural or acidified reaction environments (Figure 2b). By adding sodium pyrophosphate (at 0.13 mM/kg d.w.) as a chelating agent in the modified Fenton process (at pH = 7.95), it proved possible to obtain an efficiency higher by 1.69%–3.89% than at pH = 7.95, and lower by a maximum of 3.04% compared with pH = 3. The chelating agent stabilized the hydrogen peroxide, enhanced desorption of the entrapped pollutant, and solubilized part of the iron from the bottom sediments. The presumed mechanism of action of the chelating substance in the process is shown in Figure 3. Figure 3. Mechanism of action of PS in the modified Fenton process [developed on the basis of Zhao et al. (2018) [20]]. On the one hand, sodium pyrophosphate can inhibit iron precipitation by forming stable complexes with it, while on the other it can accelerate the Fe3+/Fe2+ cycle, with the effect that more hydroxyl radicals are produced. Under their optimal process conditions (i.e., 300 mM of H2O2, 30 mM of Fe3+, pH = 3, t = 6 h), Jorfi et al. (2013) [13] achieved 99% removal of pyrene from soil (C0 = 100 mg/kg). At pH 7, the corresponding figure was 93% in the presence of sodium pyrophosphate, which proved the most useful of the chelating agents they analyzed (alongside ethylenediaminetetraacetic acid (EDTA), sodium citrate, humic acids, and fulvic acids). In turn, in PAH-contaminated soils of neutral pH studied by Gan et al. (2012) [21], sodium pyrophosphate as a chelating agent was also found to enhance a modified Fenton (H2O2/Fe3+) method of removal. The authors analyzed the usefulness of five chelating substances before homing in on sodium pyrophosphate as the most effective at removing selected substances of the PAH group. According to those authors, inorganic chelating substances differ from organic ones in competing less for access to hydroxyl radicals, while the overall carbon content in soil does not increase, and the phosphate ions can represent an additional source of soil nutrients. A positive effect of the chelating agent was Catalysts 2019, 9, 932 5 of 10 also confirmed by Pardo et al. (2014) [22], Qin et al. (2015) [23], and Wang et al. (2017) [24], among others. In situ chemical oxidation (ISCO) is a powerful technology for soil remediation. However, one of the main drawbacks of the Fenton process lies in the instability of H2O2 when in contact with soil. In addition, organic pollutants can be entrapped in soil organic matter so strongly that the efficiency of ISCO technology is impaired. However, an H2O2 stabilization effect is also noted where sodium pyrophosphate is applied. Fenton processes are relatively inexpensive and easy to operate without any further energy requirements. It is therefore reasonable to combine a modified Fenton reaction with other technologies to improve radical-leading oxidation. The content of organic matter decreased gradually as the reaction proceeded, from an average value of 7.86% to 6.82%. The value of DOC in the first hour of the process almost doubled, while subsequently there was a drop back to almost the initial level (Figure 4). Figure 4. Impact of the H2O2/Fe3+ process (H2O2:Fe3+ 1:1, pH = 3) on (a) soil organic matter (SOM) depending on t; (b) dissolved organic carbon (DOC) depending on t. The use of a modified Fenton process resulted in a fourfold increase in the value noted for DOC (to 1.2 mg/g) in trials run by Cheng et al. (2016) [25] seeking to remove atrazine (at 617.5 mg/kg) from soil. Those authors noted that oxidation with Fenton’s reagent resulted in a slight decrease in SOM content, usually by about 10% of the initial value, as the humic acids which are the main components of SOM prove resistant to oxidation. In turn, in research by Zhao et al. (2018) [20], a DOC content four times higher was also obtained as DEHP was being removed from bottom sediments in an H2O2/Fe3+ system. Organic matter content decreased slightly—from 8.51% on average to around 7.70%. 2.3. Impact of Initial DEHP Content and the Presence of Other Impurities on DEHP Degradation The impact of the initial content of DEHP on the efficiency of removal from bottom sediments was as presented in Figure 5. C0 was analyzed across the 10–100 mg/kg d.w. range, while reaction environment conditions were pH = 3, DEHP:H2O2:Fe3+ 1:1:1, and reaction time in the range 1–24 h. Catalysts 2019, 9, 932 6 of 10 Figure 5. Influence of initial content (C0 = 10–100 mg/kg d.w.) on DEHP decomposition via a modified Fenton process (H2O2:Fe3+ 1:1, pH 3). Work starting with C0 = 10 mg/kg d.w. demonstrated a decrease in DEHP content to an average of 7.07 mg/kg d.w. after 24 hours (the removal efficiency was 28.5%). In the first two hours, efficacy for C0 = 20 mg/kg d.w. was comparable with the higher DEHP contents tested (70 and 100 mg/kg d.w.) and did not exceed 14.05% on average. Changes in amounts of di(2-ethylhexyl) phthalate decomposed via the modified Fenton process for C0 = 70 mg/kg d.w. and C0 = 100 mg/kg d.w. depended on reaction time, but were anyway only in the 11.01%–13.73% and 8.73%–13.32% ranges respectively. The research analyzed high initial contents of the pollutant, showing how efficiency was greater where the initial content of DEHP was lower. A further important factor found to determine the efficiency of DEHP removal from bottom sediments via the above process was the presence of selected PAH-group contaminants at 3.2 mg/kg d.w., and of pesticides at 4 mg/kg d.w. (for C0 = 10 and 50 mg/kg d.w.) (Figure 6). The efficiency of removal of di(2-ethylhexyl) phthalate in the presence of other impurities was lower by a maximum of 16.32% for C0 = 10 mg/kg d.w. (t = 1 h) (Figure 6a). Where the initial DEHP content was higher (at C0 = 50 mg/kg d.w.), no significant differences in obtained values were observed (Figure 6b). In the presence of additional pollutants, and with a lower initial content of C0 = 10 mg/kg d.w., hard-to-degrade compounds are found to limit access of DEHP to hydroxyl radicals. The presence of other substances, especially in comparable amounts, thus contributes to increased competition for HO· radicals, with a consequent reduction in removal efficiency. Figure 6. Impact of other hard-to-degrade impurities in the H2O2/ Fe3+ process (H2O2:Fe3+ 1:1, pH = 3) on the DEHP distribution for: (a) C0 = 10 mg/kg d.w.; (b) C0 = 50 mg/kg d.w. PAH: polycyclic aromatic hydrocarbon. The susceptibility to degradation by Fenton’s reagent of PAHs present in soil, sewage sludge, and bottom sediments was investigated by Flotron et al. (2005) [26]. Benzo(a)pyrene (BaP) was oxidized more effectively than fluoranthene, while the process was ineffective with benzo(b)fluoranthene, given its ease of absorption. The level of decomposition of these substances was found to be affected by the initial content of PAHs, as was also the case for the research presented here. Across the analyzed range, the modified Fenton process did not allow for a high degree of DEHP removal from bottom sediments. However, effectiveness was enhanced slightly in comparison with the classic Fenton reagent, as was likewise confirmed by research from Yap et al. (2011) [27], among others. 2.4. Impact of the Process on the Leaching of Selected Metals and Plant Nutrients Various chemical elements (and especially heavy metals) deposited in bottom sediments constitute permanent and non-decomposable environmental pollution, posing a great secondary risk to water. This reflects the fact that, as many forms of these elements are mobile, transfers deeper or towards the sediment surface are possible, with the threats reflecting ecotoxic and phytotoxic effects Catalysts 2019, 9, 932 7 of 10 [28]. In the research detailed here, it did prove possible to observe leaching of bottom-sediment components, including pollutant lead, copper, zinc, nickel, and aluminum, as well as key plant nutrients magnesium, potassium, and calcium (Table 1). Table 1. Leaching elements from sediments during the modified Fenton process (pH = 3, t = 1 h). Element Ni Pb Cu Zn Al Ca Mg K Leaching after process [mg/g d.w.] 0.031 0.004 0.005 0.014 0.594 0.080 0.015 2.105 Blank samples [mg/g d.w.] 0.0001 0.002 0 0.0004 0.002 0.068 0.010 1.754 A change of bottom-sediment pH to slightly acidic or acidic helped mobilize plant-available heavy metal forms. However, lead and copper were shown to release from bottom sediments to the most limited extent, while Mg was the least-mobilized nutrient, and K the most mobilized. As plants’ absorption of calcium, magnesium, and potassium is reduced in low-pH conditions, the use of a chelating agent seems to be a favorable measure if it permits the process to occur at a natural reaction. The modified Fenton process had a much greater impact on the leaching of Pb, Cu, Zn, Ni, and Al as compared with plant nutrients. The factor determining the degree of elution plant nutrients from the matrix was acidification of the reaction environments. 3. Materials and methods 3.1. Reagents Di(2-ethylhexyl) phthalate (DEHP) and di(2-ethylhexyl) phthalate-3,4,5,6-d4 (DEHP-3,4,5,6-d4) were purchased from Sigma-Aldrich (Darmstadt, Germany). Standard solutions of the 16 PAHs (polycyclic aromatic hydrocarbons) were purchased from Reagecon (Shannon, Ireland), while a standard solution of the organochlorine pesticides was obtained from Sigma-Aldrich. Analytical-grade n-hexane, methanol, acetone, NaOH, HCl, H2O2 solution (30%), Na4P2O7·10H2O, and Fe(NO3)3·9H2O were obtained from POCH (Gliwice, Poland). Cellulose acetate membrane filters of 0.20 and 0.45 µm pore size and a syringe filter of 0.22 µm pore size were both purchased from LaboPlus (Warsaw, Poland). 3.2. Experimental Procedures The bottom sediments used in the research were collected from Rzeszow Reservoir in Poland, and dried in air and then at 105 °C to constant weight, before being ground up and passed through a 1.0-mm sieve. This sediment washed with acetone was allotted to 1 g samples, then placed in glass reaction vessels. DEHP was then introduced in the form of a solution in acetone, in amounts ranging between 10 and 100 mg/kg d.w. The obtainment of homogeneous samples further entailed shaking for one hour, in advance of a 24-hour set-aside period during which the solvent evaporated. To check for effects of the presence of other hard-to-degrade contaminants, substances from the PAH group (at 3.2 mg/kg d.w.) and organochlorine pesticides (at 4 mg/kg d.w.) were added to samples. Distilled water was introduced into the contaminated sediments to obtain a suspension (1 g sample + 3 mL H2O) before DEHP oxidation proceeded, using a 30% solution of hydrogen peroxide plus an Fe(NO3)3·9H2O catalyst. Reactions progressed at room temperature, but with solution pH varied by adding HCl and NaOH, with different durations of oxidation in the range 1–24 h, and with different initial contents of both DEHP itself, as well as other barely degradable contaminants potentially “in competition” with it. Separate steps of the work aimed to determine what doses of hydrogen peroxide and catalyst were optimal in relation to the DEHP present in bottom sediments. At this stage we also studied the impact on the removal of DEHP exerted by addition to the reaction mixture (prior to the introduction of the catalyst, though following Fe3+) of the chelating agent sodium pyrophosphate (Na4P2O7·10H2O). Reaction initiation entailed the addition to the sample of the appropriate amount of 30% hydrogen peroxide. Catalysts 2019, 9, 932 8 of 10 Elution of Ni, Pb, Al, Cu, and Zn from bottom sediments as a result of the applied processes was also evaluated. The samples were mineralized in the presence of 10 cm3 HNO3 in a MARS 6 microwave mineralizer (CEM, Matthews, USA). The obtained aqueous extracts were then filtered through a membrane filter of 0.20 µm pore size. A GBC Quantima E 1330 ICP-OES optical emission spectrometer (GBC, Melbourne, Australia) was used in analysis. For the determinations of calcium, magnesium, and potassium ions, aqueous extracts obtained were filtered through 0.22-µm syringe filters (LaboPlus, Warsaw, Poland) and then purified on OnGuard II filters with the aim of removing organic substances. Analysis for the presence of calcium, magnesium, and potassium ions was carried out using a DIONEX DC ICS—5000 ion chromatograph (Thermo Scientific, Waltham, USA). 3.3. Analytical Methods of DEHP Determination Quantitative determination of DEHP entailed microwave-assisted extraction in the presence of methanol, as well as the addition of an internal standard in the form of DEHP-3,4,5,6-d4. The samples were then filtered, and the organic and aqueous layers separated. The extract was dried over anhydrous sodium sulfate, concentrated to 1 mL volume, and subjected to chromatographic analysis. Quantitative DEHP determinations were made by capillary gas chromatography, using a gas chromatograph coupled with a mass detector (Thermo Scientific, Waltham, USA) (GC-MS). 4. Conclusions It can be concluded that: • The modified Fenton process described here requires the selection of optimal reagent doses if high-efficiency di(2-ethylhexyl) phthalate removal is to be achieved, as excesses of either iron ions or hydrogen peroxide can result in a hydroxyl scavenging capable of inhibiting the process; • The process removing di(2-ethylhexyl) phthalate is dependent on reaction-mixture pH, with physico-chemical properties of the analyzed bottom-sediment matrix determining the effectiveness of removal at a given environmental reaction; • The process of DEHP removal is hindered by the presence of other hard-to-degrade impurities competing for access to highly reactive hydroxyl radicals, given their presence in amounts comparable with those of the analyzed substance; • A chelating agent present in a properly-selected dose allows for the removal of DEHP in circumstances of non-modified pH, at a level of efficiency similar to that otherwise only obtainable under acidified conditions; • The use of the modified Fenton process results in the leaching of both plant nutrients and metals from bottom sediments, probably on account of the reduced pH; • As the modified Fenton reaction process failed to achieve sufficiently effective removal of DEHP from bottom sediments even under the most favorable conditions noted, prior desorption of bottom-sediment contaminants has to be advocated, along with further research to look for other solutions that include combined processes. Author Contributions: Conceptualization, M.K., S.Z. and P.K; Formal analysis, M.K. and S.Z.; Investigation, M.K. and S.Z.; Methodology, M.K., S.Z. and P.K.; Project administration, P.K.; Writing – original draft, M.K. and S.Z.; Writing – review & editing, P.K. Funding: This research received external funding from the Polish Ministry of Science and Higher Education. Conflicts of Interest: The authors declare no conflict of interest. References Catalysts 2019, 9, 932 9 of 10 1. Kudlek, E.; Dudziak, M. Degradation pathways of pentachlorophenol and benzo(a)pyrene during heterogeneous photocatalysis. Water Sci. Technol. 2018, 77, 2407‒2414. doi:10.2166/wst.2018.192. 2. Kida, M.; Ziembowicz, S.; Koszelnik, P. Study on the suitability of using low-frequency ultrasonic field for removing di(2-ethylhexyl) phthalate from bottom sediments. Sep. Purif. Technol. 2020, 233, 116010. doi:10.1016/j.seppur.2019.116010. 3. Pochwat, K.; Kida, M.; Ziembowicz, S.; Koszelnik, P. Odours in Sewerage—A Description of Emissions and of Technical Abatement Measures. Environments 2019, 6, 89. doi:10.3390/environments6080089. 4. Kozak, J.; Włodarczyk-Makuła, M. Comparison of the PAHs degradation effectiveness using CaO2 or H2O2 under the photo-Fenton reaction. Desalin. Water Treat. 2018, 134, 57–64. doi:10.5004/dwt.2018.22708. 5. Pochwat, K.; Iličić, K. A simplified dimensioning method for high-efficiency retention tanks. In Proceedings E3S Web of Conferences; EDP Sciences, France: 2018; Volume 45, p. 00065. 6. Gajewska, M.; Kopeć, Ł.; Obarska-Pempkowiak, H. Operation of small wastewater treatment facilities in a scattered settlement. Rocz. Ochr. Srod. 2011, 13, 207–225. 7. Barbusiński, K. Intensyfikacja Procesu Oczyszczania Ścieków i Stabilizacji Osadów Nadmiernych z Wykorzystaniem Odczynnika Fentona; Zeszyty Naukowe: Inżynieria Środowiska/Politechnika Śląska; Wydawnictwo Politechniki Śląskiej: Gliwice, Polska, 2004; p. 50. 8. Ziembowicz, S.; Kida, M.; Koszelnik, P. The use of alternative catalysts in processes of the chemical degradation of di-n-butyl phthalate in aqueous solutions. Chemosphere 2019, 237, 124450. doi:10.1016/j.chemosphere.2019.124450. 9. Zolfaghari, M.; Drogui, P.; Seyhi, B.; Brar, S.K.; Buelna, G.; Dubé, R. Occurrence, fate and effects of Di (2-ethylhexyl) phthalate in wastewater treatment plants: A review. Environ. Pollut. 2014, 194, 281–293. 10. Czarny, K.; Szczukocki, D.; Krawczyk, B.; Zieliński, M.; Miękoś, E.; Gadzała-Kopciuch, R. The impact of estrogens on aquatic organisms and methods for their determination. Crit. Rev. Environ. Sci. Technol. 2017, 47, 909‒963. 11. Chiou, C.S.; Chen, Y.H.; Chang, C.T.; Chang, C.Y.; Shie, J.L.; Li, Y.S. Photochemical mineralization of di-n-butyl phthalate with H2O2/Fe3+. J. Hazard. Mater. 2006, 135, 344–349. 12. Wang, F.; Xia, X.; Sha, Y. Distribution of phthalic acid esters in Wuhan section of the Yangtze River, China. J. Hazard. Mater. 2008, 154, 317–324. 13. Jorfi, S.; Rezaee, A.; Moheb–ali, G.A.; alah Jaafarzadeh, N. Pyrene removal from contaminated soils by modified Fenton oxidation using iron nano particles. J. Environ. Health Sci. Eng. 2013, 11, 17. 14. Ledakowicz, S.; Olejnik, D.; Perkowski, J.; Żegota, H. The use of in-depth oxidation processes to break down the non-ionic Triton X-114 surfactant. Przem. Chem. 2001, 80, 453–459. 15. Goi, A.; Viisimaa, M. Integration of ozonation and sonication with hydrogen peroxide and persulfate oxidation for polychlorinated biphenyls-contaminated soil treatment. J. Environ. Chem. Eng. 2015, 3, 2839–2847. 16. Zhang, J.; Li, J.; Thring, R.; Liu, L. Application of ultrasound and Fenton's reaction process for the treatment of oily sludge. Proced. Environ. Sci. 2013, 18, 686–693. 17. Yim, B.; Yoo, Y.; Maeda, Y. Sonolysis of alkylphenols in aqueous solution with Fe (II) and Fe (III). Chemosphere 2003, 50, 1015–1023. 18. Ziembowicz, S.; Kida, M.; Koszelnik, P. The impact of selected parameters on the formation of hydrogen peroxide by sonochemical process. Sep. Purif. Technol. 2018, 204, 149–153. 19. Bokare, A.D.; Choi, W. Review of iron-free Fenton-like systems for activating H2O2 in advanced oxidation processes. J. Hazard. Mater. 2014, 275, 121–135. 20. Zhao, M.; Cheng, M.; Zeng, G.; Zhang, C. Degradation of di(2-ethylhexyl) phthalate in sediment by a surfactant-enhanced Fenton-like process. Chemosphere 2018, 198, 327–333. 21. Gan, S.; Ng, H.K. Modified Fenton oxidation of polycyclic aromatic hydrocarbon (PAH)-contaminated soils and the potential of bioremediation as post–treatment. Sci. Total Environ. 2012, 419, 240–249. 22. Pardo, F.; Rosas, J.M.; Santos, A.; Romero, A. Remediation of a biodiesel blend-contaminated soil by using a modified Fenton process. Environ. Sci. Pollut. Res. 2014, 21, 12198–12207. 23. Qin, Y.; Song, F.; Ai, Z.; Zhang, P.; Zhang, L. Protocatechuic acid promoted alachlor degradation in Fe (III)/H2O2 Fenton system. Environ. Sci. Technol. 2015, 49, 7948–7956. 24. Wang, H.; Liang, H.; Gao, D.W. Occurrence and risk assessment of phthalate esters (PAEs) in agricultural soils of the Sanjiang Plain, northeast China. Environ. Sci. Pollut. Res. 2017, 24, 19723–19732. Catalysts 2019, 9, 932 10 of 10 25. Cheng, M.; Zeng, G.; Huang, D.; Lai, C.; Xu, P.; Zhang, C.; Liu, Y.; Wan, J.; Gong, X.; Zhu, Y. Degradation of atrazine by a novel Fenton-like process and assessment the influence on the treated soil. J. Hazard. Mater. 2016, 312, 184–191. 26. Flotron, V.; Delteil, C.; Padellec, Y.; Camel, V. Removal of sorbed polycyclic aromatic hydrocarbons from soil, sludge and sediment samples using the Fenton’s reagent process. Chemosphere 2005, 59, 1427–1437. 27. Yap, C.L.; Gan, S.; Ng, H.K. Fenton based remediation of polycyclic aromatic hydrocarbons-contaminated soils. Chemosphere 2011, 83, 1414–1430. 28. Bogacki, J.; Naumczyk, J. Ex-situ physicochemical methods for the treatment of heavy metal bottom sediments. Gaz Woda i Technika Sanitarna 2009, 11, 3–42. © 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/). View publication stats | RESEARCH GATE |
ISIJ International, Vol. 61 (2021), No. 5© 2021 ISIJ1459ISIJ International, Vol. 61 (2021), No. 5, pp. 1459–1468https://doi.org/10.2355/isijinternational.ISIJINT-2020-564* Corresponding author: E-mail: zuohaibin@ustb.edu.cn© 2021 The Iron and Steel Institute of Japan. This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs license (https://creativecommons.org/licenses/by-nc-nd/4.0/).CCBYNCND1. IntroductionIn recent years, the demand for iron ore has increased with the rapid development of the iron and steel indus-try, and the problem of iron ore shortage is prominent. As a potential resource, there are abundant reserves of medium-high phosphorus iron ore in the world, and effec-tive exploitation and utilization of these phosphate ores is an important guarantee for the sustainable development of the iron and steel industry.1–3) However, the main forms of phosphorus in phosphorous ore are monazite and apatite, which are closely surrounded by gangue. The ore has high inclusion, small particle size and complex chemical com-position, which makes it very difficult to concentrate.4–6) The phosphate ore enters the blast furnace, and most of the phosphorus reduced in the smelting process goes into the hot metal. The introduction of hot metal with high phosphorus content into the converter aggravates the task of phosphorus removal in the steelmaking process, and seriously restricts the recycling of steel slag resources.7) Eventually leads to the vicious circle of phosphorus in the Gasification Behavior of Phosphorus during Pre-reduction Sintering of Medium-high Phosphorus Iron OreYanbiao CHEN and Haibin ZUO*State Key Laboratory of Advanced Metallurgy, University of Science and Technology Beijing, Beijing, 100083 PR China.(Received on September 9, 2020; accepted on December 9, 2020)The effects of dephosphorization agents (SiO2, Na2CO3 and Al2O3) and reduction time on phosphorus gasification during pre-reduction sintering gasification dephosphorization were studied. The results indicate that the dephosphorization rate of apatite by carbothermal reduction is 23%, the dephosphorization rate of adding SiO2 and Al2O3 is 25%, and the optimal dephosphorization rate is 36% by adding SiO2 and Na2CO3. The dephosphorization rate increased from 13% to 31% with the reduction time increased from 20 min to 60 min. Thermodynamic analysis shows that the promotion effect of Na2CO3 on apatite reduction is stron-ger than that of Al2O3, and the thermodynamic conditions of apatite reduction are optimized to the greatest extent by adding SiO2 and Na2CO3, with the starting temperature of carbothermal reduction of apatite decreased from 1 464°C to 746°C. P and C have the same variation and change trend, and the migration of phosphorus ito metallic iron is related to the carburizing reaction. The good heat storage of the sinter layer in pre-reduction sintering increases the time for maintaining the apatite reaction temperature and promotes the occurrence of dephosphorization reaction, however part of the phosphorus gas is absorbed by liquid metallic iron to form stable iron phosphorus compounds,which leads to the decrease of dephos-phorization rate. Besides addition of dephosphorization agent, restricting migration of phosphorus into iron phase is the key point to increase gasification dephosphorization efficiency in sintering process.KEY WORDS: pre-reduction sintering; dephosphorization agent; carbothermic reduction; phosphorus gasification; reduction time.whole process of iron and steel production.8–10) Therefore, phosphorus-containing charge must be dephosphorized before smelting in the blast furnace.11–13)In order to solve the problem of “phosphorus damage” in iron and steel production process, domestic and foreign researchers have carried out experimental study on iron ore dephosphorization by using a variety of methods. Traditional dephosphorization methods such as beneficiation,14,15) chem-ical treating,16,17) gas-based reduction,18,19) direct reduction20) and microbial leaching,21) due to their high energy consump-tion, high pollution, high equipment requirements, high cost and poor dephosphorizing effect, these methods could not meet the demand of industrial production.22,23) Although the partial removal of phosphorus can be achieved by iron ore powder sintering and agglomeration process, however, in order to ensure the strength and reducibility of sinter in the common sintering process, the carbon content of raw materials must be controlled at about 5%. The reducibility of sintering atmosphere becomes worse, and it is difficult to achieve efficient removal of phosphorus. The realization of dephosphorization of medium and high phosphorus iron ore has become a difficult problem concerned by researchers at home and abroad.24,25)ISIJ International, Vol. 61 (2021), No. 5© 2021 ISIJ1460In order to reduce the carbon dioxide emission and energy consumption of blast furnace, JFE put forward a new pre-reduction sinter process with the characteristics of high carbon content, strong atmosphere reducibility and high temperature negative pressure operation based on the exist-ing sintering process,26,27) and the thickness of the sintered layer is larger than that of the ordinary sintering process. These measures can effectively improve the thermodynamic and kinetic conditions of phosphorus reduction, which is very beneficial to the reductive dephosphorization of the sinter process.28,29) In sintering process, with the addition of suitable dephosphorization agent, the stable phosphorus-containing gas generated by the reduction of phosphorite is discharged with sinter flue gas to achieve the purpose of gasification and dephosphorization.It is of great significance for the efficient utilization of medium-high phosphorus iron ore resources and reducing the raw material cost of iron and steel enterprises.In this work, the effects of different dephosphorization agents and reduction time on the gasification of phosphorus in pre-reduction sintering process were studied systemati-cally. The experiment was carried out through roasting of pressed tablets in a vacuum tube furnace. Thermodynamic calculation of possible reactions is carried out by using FactSage6.1 software, and the phase transformation and distribution of phosphorus in the process of roasting reduction were analyzed by XRD, SEM-EDS and EPMA. The mechanism of sintering gasification dephosphoriza-tion and the migration law of phosphorus in iron phase are expounded, which could provide theoretical basis for further study of pre-reduction sintering gasification dephos-phorization process.2. Experimental2.1. Experimental Material2.1.1. Medium-high Phosphorus Iron ConcentrateThe raw material used in this experiment is Bayan Obo iron concentrate, and the main chemical composition is shown in Table 1.2.1.2. CokeCoke is used as reducing agent, and its composition is shown in Table 2.2.1.3. Dephosphorization AgentThe main dephosphorization agents used in this experiment are SiO2, Na2CO3 and Al2O3, which are all analytically pure.2.2. Experimental MethodThe main purpose of this paper is to study the effect of different dephosphorization agents on phosphorus gasifica-tion during the process of pre-reduction sintering gasifica-tion dephosphorization. Phosphorus in Bayan Obo iron concentrate mainly exists in form of calcium phosphate. In order to clearly show the changing trend of phosphorus in the reduction process when XRD, SEM-EDS and EPMA analysis were employed and improve the accuracy of experimental results, 0.5% calcium phosphate was added into the iron concentrate to enhance the reaction related to phosphorus. In this experiment, the effect of dephos-phorization agents (SiO2, Na2CO3, Al2O3) on phosphorus gasification in the sintering process was studied, and the ratios of 3%SiO2, 3%SiO2+1%Na2CO3, 3%SiO2+3%Al2O3 and 3%SiO2+3%Al2O3+1%Na2CO3 in the experiment were determined according to calculation result of the FactSage (Equilib module), respectively. The above different propor-tions of dephosphorization agents were mixed and tableted with 20% coke and iron concentrate. The tablets were made under constant pressure of 5 MPa and dried. The tablet was roasted in a vacuum tube furnace with protection of argon at pressure of 0.9 atm. In this work, the prepared tablets were heated from room temperature to reaction temperature of 1 200°C at a rate of 15°C/min, holding for 60 min. Then, the roasted tablets were cooled to room temperature along with the furnace and taken out for the following tests. The effect of sintering time on phosphorus gasification was studied, and the reduction was carried out at an interval of 10 minutes within the range of 20–60 min. Figure 1 is the schematic diagram of the experimental device.Phosphorus content in roasting products was tested by chemical analysis (The bismath phosphomolybdate biue photometric method for the determination of phosphorus content). The thermodynamic relation of dephosphorization reaction equation ΔrG0-T is calculated by FactSage6.1 soft-ware to explain the reduction and gasification mechanism of apatite. The reduced samples were treated by X-ray diffrac-tometer to determine the phase composition and structure of the sample. SEM-EDS and EPMA were used to detect the element composition distribution of the reduced samples.The phosphorus content of the material before and after reduction was detected by multi-point sampling, and the Table 2. The chemical composition analysis of coke (mass frac-tion, mass%).Fixed carbonAshVolatile matterSSiO2Al2O386.8412.211.210.966.772.67Table 1. Main chemical composition of iron concentrates (mass fraction, mass%).Fetotal FeO MgO CaO SiO2FSPAl2O3 Na2O K2O63.00 27.00.831.585.280.52 1.80 0.080.50.230.14Fig. 1. Schematic of the experimental apparatus.ISIJ International, Vol. 61 (2021), No. 5© 2021 ISIJ1461dephosphorization rate was calculated according to the fol-lowing formula. � �������1100%100PMPM1In the formula: P0 and P1 are the content of phospho-rus in the tablet before and after reduction, %. M0 and M1 are the mass of tablets before and after reduction, %. η-dephosphorization rate, %.3. Results and Discussion3.1. Effect of Dephosphorization Agent on Sintering Dephosphorization3.1.1. Effect of Dephosphorization Agent on Dephosphori-zation RateFigure 2 shows the change of dephosphorization rate with different dephosphorizing agents (1-C,2-C + SiO 2,3-C + SiO 2+ Na 2CO 3,4-C + SiO 2+ Al 2O 3,5-C+SiO2+Al2O3+Na2CO3), with the tablets heated to 1 200°C and maintained for 60 min. The lowest dephospho-rization rate is 23% with only carbon added. The dephospho-rization rate increased in varying degrees with addition of different dephosphorization agents. The dephosphorization rate increased to 29% with the addition of SiO2. The high-est dephosphorization rate of 36% was achieved by adding SiO2 and Na2CO3. The dephosphorization rate was reduced to 31% with the addition of SiO2, Al2O3 and Na2CO3. The dephosphorization rate is 25% by adding SiO2 and Al2O3, Adding SiO2 and Al2O3 has a lower dephosphorization rate than adding SiO2 and Na2CO3. The results show that the Al2O3 has weaker reduction effect on apatite than Na2CO3, the gasification dephosphorization effect is the best when SiO2 and Na2CO3 are selected as dephosphorization agents.3.1.2. Thermodynamic Analysis of Carbothermal Reduc-tion of Apatite with Different Dephosphorization AgentsIn order to clarify the carbothermal reduction mechanism of apatite with different dephosphorization agents, the ther-modynamics of possible reduction reactions were discussed in detail. The ΔrG0-T of the relevant reactions were calcu-lated by FactSage software, and the change of the standard Gibbs free energy under different temperatures were linearly fitted. The thermodynamic relationship of each reaction is shown in Figs. 3–5 and Table 3.Fig. 2. Effect of dephosphorization agent on dephosphorizing rate of reduction gasification (1-C, 2-C+SiO2, 3-SiO2+Na2CO3, 4-C+SiO2+Al2O3, 5-C+SiO2+Al2O3+Na2CO3). (Online ver-sion in color.)Fig. 4. The relationship between standard Gibbs free energy change and reduction reaction temperature (Eqs. (4) and (5)–(8)). (Online version in color.)Fig. 5. The relationship between standard Gibbs free energy change and reduction reaction temperature (Eqs. (4) and (9)–(12)). (Online version in color.)Fig. 3. The relationship between standard Gibbs free energy change and reduction reaction temperature (Eqs. (1)–(4)). (Online version in color.)ISIJ International, Vol. 61 (2021), No. 5© 2021 ISIJ1462The effect of adding SiO2 on the carbothermal reduction of apatite is shown in Fig. 3. In the case of only carbon, the starting temperature of apatite reduction is 1 464°C, and phosphorus is mainly volatilized in the form of P2 gas at high temperature. In the process of reduction, iron oxide gives priority to apatite reduction. On the one hand, the reduction temperature of iron oxide is lower, and the step-by-step reduction begins at 702°C. On the other hand, the reduction of iron oxides is mainly the indirect reduc-tion of CO, even if the amount of carbon is small, the good kinetic conditions of gas-solid reaction are still beneficial to the reduction of iron oxides. The reduction of apatite is mainly solid-solid phase direct reduction with carbon as reducing agent,before reaching the required temperature of reduction, carbon is mainly consumed by the reduction of iron oxides, the reduction reaction of carbon and apatite can not be carried out fully.Pre-reduction sintering dephos-phorization is difficult to achieve. The reduction starting temperature decreased with the addition of SiO2, and the lowest reduction temperature decreased to 1 208°C with the increase of SiO2 ratio. Ca POCCOCaOPrGTlnKP34 221005531 488 9241017179 086 4().������� ��//.T �122 3 ............. (1) Ca POSiOCCOCa Si OPrGTlnKP34 2232722002551258 705 1016() ��������� ��151395 8122 2. /.T .......... (2) Ca POSiOCCOCa SiOPrGTln34 22242301 5551 51264 546989()..��������KTP0152 098 4119� ��. / .... (3) Ca POSiOCCOCaSiOPrGTlnKP34 223240035531209 9601002() �����������145 532 8120 5. /.T ........... (4)The effect of adding Na2CO3 and SiO2 on the carbother-mal reduction of apatite is shown in Fig. 4. The minimum reduction temperature of apatite can be reduced to 746°C by adding SiO2 and Na2CO3 at the same time. The phosphorus-containing minerals are reduced to produce phosphorus gas at a lower temperature, which fully meets the temperature range of dephosphorization in the pre-reduction sintering process. Ca POCSiONa CONa Ca Si OPCOrG34 2223236162506671353 501() ���������� ��1384162 797 8166 50TlnKTP. /. ............ (5)Table 3. Evolution of equilibrium constant and standard Gibbs free energy versus temperature of the reactions (1)–(12).Reactiont/(°C)7008009001 0001 1001 2001 3001 4001 500(1)ΔrG10(KJ)77767557447237026916765−37KP00.8750.9030.9260.9450.9600.9730.9850.9941.003(2)ΔrG20(KJ)54844634424314140−62−164−265KP00.9100.9350.9550.9710.9850.9961.0061.0141.021(3)ΔrG30(KJ)57247337427617778−21−120−219KP00.9070.9310.9510.9680.9810.9921.0021.0101.018(4)ΔrG40(KJ)5094083082081088−93−193−293KP00.9170.9410.9600.9750.9880.9991.0091.0171.023(5)ΔrG50(KJ)385246108−30−169−307−446−584−722KP00.9360.9640.9861.0041.0191.0311.0421.0511.060(6)ΔrG60(KJ)143−170−483−796−1 108−1 421−1 734−2 046−2 359KP00.9761.0261.0671.1011.1291.1531.1741.1921.208(7)ΔrG70(KJ)699405111−183−477−771−1 065−1 359−1 653Kp00.8870.9410.9851.0221.0531.0801.1041.1241.142(8)ΔrG80(KJ)2419−223−455−687−919−1 152−1 384−1 616KP00.9600.9991.0301.0571.0781.0961.1131.1261.139(9)ΔrG90(KJ)46336125915755−47−149−251−353KP00.9240.9470.9660.9810.9941.0051.0141.0221.028(10)ΔrG100(KJ)1 036829621414207−0.4970−208−415−622KP00.8370.8820.9200.9510.9771.0001.0191.0371.051(11)ΔrG110(KJ)47337828218691−5−100−196−292KP00.9220.9450.9630.9780.9901.0001.0091.0171.023(12)ΔrG120(KJ)355230106−19−144−269−394−518−643KP00.9410.9660.9861.0021.0161.0271.0371.0461.053ISIJ International, Vol. 61 (2021), No. 5© 2021 ISIJ1463 Ca POCSiONa CONa Ca Si OPCOrG34 2223433926011963172 331() ��������4093127280 419 6376 10�� ��TlnKTP. /. .......... (6) Ca POCSiONa CONa Ca Si OPCOrG34 22232239270139332162 75() ��������5 9102 939331 478353 50�� ��TlnKTP/. .......... (7) Ca POCSiONa CONa CaSi OPCOrG34 2223251228081533111865() ��������7232 321224 407 4279 20�� ��TlnKTP. /. .......... (8)As shown in Fig. 5, the starting temperature of apatite reduction with the addition of SiO2 and Al2O3 is reduced to 1 154°C, but this high temperature is not conducive for sin-tering dephosphorization reaction. The starting temperature of apatite reduction with addition of three dephosphorization agents of SiO2, Na2CO3 and Al2O3 is 985°C. The results show that Al2O3 has weaker reduction effect on apatite than Na2CO3. The formation of low melting point slag phase such as NaAlSiO4 increases the amount of liquid phase, improves the fluidity of the system, improves the kinetic conditions of mass transfer and promotes the reduction of apatite. but the high-melting point substance of CaAl2Si2O8 deteriorates the fluidity of liquid phase which make the liquid phase generated in the cooling crystallization process makes phosphate rock-bearing materials dissolve in gangue phase, and lowers the dephosphorization rate. Ca POCSiOAl OCaAl SiOPCOrG34 222326290533351176 607()()���������� ��1020141521 2122 70TlnKTP. /.............. (9) Ca POCSiOAl OCa Al SiOPCOrG34 2223227210010333102 48()()��������71032 073299146 4249 30�� ��TlnKTP. /. .......... (10) Ca POCSiOAl OCa Al Si OPCOrG34 222332312211053101142()()��������368956137 402 91150�� ��TlnKTP. /............ (11) Ca POCAl OSiONa COCaAl Si OPCONaAlS34 2232232282648372() ��������iOPrGTlnKTP4212001228 881 1248147 808 6150 1���� ���. /. ...... (12)Table 3 summarizes the main thermodynamic data for the above reaction Eqs. (1)–(12). Addition of dephosphorization agents improved the thermodynamic conditions of apatite carbothermal reduction in varying degrees, and the effects of SiO2 and Na2CO3 are the best.3.1.3. Phase Transition of Reduction Roasting Products under Different Dephosphorization AgentsThe elemental composition of roasted products with dif-ferent dephosphorization agents was analyzed by XRD, as shown in Fig. 6 and Table 4. Most iron oxide in roasting products are completely reduced to metallic iron, and was carburized to form Fe3C. With addition of SiO2, the inter-mediate product Fe2SiO4 was formed and it was difficult to reduce,In addition, the diffraction peak of CaSiO3 was detected, indicating that apatite was reduced, but no FexP was detected, indicating that the reduction of apatite was less at this time. With addition of SiO2 and Na2CO3, apatite was greatly reduced, and part of the generated phosphorus gas entered into the metallic iron to form FexP, which lead to the decrease of dephosphorization rate. At the same time, the formation of NaAlSiO4 was observed, which was caused by the reaction between SiO2, Al2O3 and Na2CO3. The peak intensity of metallic iron decreased with the addition of SiO2 and Al2O3, which changed the reduction pathway of iron oxide, and Feo preferentially reacted with gangue Al2O3 and SiO2 to produce FeAl2O4 and Fe2SiO4 compounds. The formation of FexP and nepheline (NaAlSiO4) was detected by adding three kinds of dephosphorization agents SiO2, Na2CO3 and Al2O3,which meant that apatite was reduced. The results are consistent with the above thermodynamic calculations.Fig. 6. (a) XRD patterns of roasted samples with different dephosphorization agents and (b) partially enlarged view of the zone with the peaks of D and M (A–Fe, B–Fe2SiO4, C–CaSiO3, D–Fe3C, E–Ca2SiO4, F–Na2CaSi5O12, H–NaAlSiO4, I–FeAl2O4, L–CaAl2O4, M–FeXP). (Online version in color.)ISIJ International, Vol. 61 (2021), No. 5© 2021 ISIJ1464The Fe–P binary phase diagram is shown in Fig. 7, with the change of phosphorus content in iron phase, phosphorus in FeXP exists in the form of γ iron and α iron or solid solu-tion of Fe3P, Fe2P and FeP. With the increase of phosphorus content, the eutectic mixture of α Fe–Fe3P (melting point is 1 048°C) and Fe2P–FeP (melting point is 1 262°C) is forme, At 1 200°C, the maximum solubility of P in γ-Fe is about 0.56%.30)3.1.4. Distribution Characteristics of Phosphorus in Reduction Roasting Products with Different Dephosphorization AgentsThe element distribution of roasting products with differ-ent dephosphorization agents was analyzed by SEM-EDS and EPMA, as shown in Figs. 8, 9 and 10.Figure 8 shows the SEM image and EDS spectrum of the roasted product with the addition of SiO2 and Na2CO3. As shown in Fig. 8(a), in the case of carbon only, the overlap-ping distribution of P and Ca could be seen from the EDS spectrum and most of the apatite was not reduced. In Fig. 8(b), with SiO2 added, most of P and Ca are overlapped and wrapped in gangue minerals, and a small part of P is overlapped with iron, and it indicated that a small amount of apatite was reduced. This is mainly because the phosphorite in the ore is wrapped by gangue components, which hinders the contact with reducing agent.As shown in Fig. 8(c), when SiO2 and Na2CO3 are added, the overlapping area of P and Ca is small, while the overlap-ping area of Fe and P is obviously increased, and in this case most apatite is reduced. In addition, the non-overlap of Fe and Ca indicates that phosphorus enters the iron phase in the form of elemental rather than calcium phosphate.On the one hand, the addition of SiO2 reduces the reduc-tion temperature of apatite. On the other hand, the reduc-tion of FeO is the most difficult in reduction process of iron oxide and FeO preferentially reacts with SiO2 to form Fe2SiO4. The reduction temperature of Fe2SiO4 is higher than the escape temperature of P2 gas, which can wrap iron before P2 gas is discharged, thus inhibit the migration of P to iron phase to a certain extent,In addition, with the increase of SiO2 addition, Fe2SiO4 can combine with SiO2 to form Fe2SiO4–SiO2, with low melting point, which is beneficial to the improvement of mass transfer kinetic conditions pro-mote dephosphorization.The addition of Na2CO3 reduces the reduction temperature, and at the same time decomposes into Na2O at high temperature reacts reacts with the gangue components such as SiO2 and Al2O3 around the phospho-rite, which breaks the wrapping of phosphorous minerals by gangue in iron concentrate. However, some Na2O will react with apatite to form Na3PO4, which is more difficult to reduce than apatite.beside with the carburizing reaction in the reduction process, metallic iron begins to melt to form liquid iron, which has higher solubility than solid iron and shows a strong ability to absorb phosphorus, then part of P2 gas enters the iron phase.Figure 9 shows the SEM image and EDS spectrum of the roasted product with the addition of SiO2, Al2O3 and Na2CO3.In Fig. 9(a), when SiO2 and Al2O3 is added, the results show that Ca, P and Al are overlapped in the calcined products, and most of apatite exists in the slag phase and is not reduced. As shown in Fig. 9(b), with the simultane-ous addition of SiO2, Al2O3 and Na2CO3, it is obviously observed that most of P overlaps with Fe and a small part of P overlaps with Ca and Al, which means apatite is reduced. Compared with no addition of Na2CO3, apatite reduction is more complete, which indicates that Na2CO3 can promote apatite reduction better than Al2O3.With addition of SiO2 and Na2CO3, apatite is reduced completely with part of P evaporating into gas phase and the other part binding with metal iron. Therefore, it is necessary to quantitatively analyze carbon and phosphorus entering the metallic iron. Figure 10 shows the analysis of the distribution of C, Fe and P elements in the roasted products by EPMA. Without adding SiO2 and Na2CO3, there is no phosphorus in the metallic iron particles. Phosphorus mainly exists in slag phase as shown in Fig. 10(a1). With the addition of SiO2 and Na2CO3, phosphorus is uniformly distributed in the iron phase, and phosphorus and Fe are closely combined as shown in Fig. 10(b1). It is found that the carbon and phosphorus in the iron phase have the same Fig. 7. Fe–P binary phase diagram.Table 4. Composition of roasted samples with different dephosphorization agents in XRD analysis.SamplesComposition of calcined samples in XRDABCDEFHILMDephosph-orization agentC√√√C+SiO2√√√√√C+SiO2+Na2CO3√√√√√√√√C+SiO2+Al2O3√√√√√√√C+SiO2+Al2O3+Na2CO3√√√√√√√√ISIJ International, Vol. 61 (2021), No. 5© 2021 ISIJ1465distribution and change trend by line scanning as shown in Figs. 10(a2), 10(b2). The elements in the iron phase are quantitatively detected by EPMA, as shown in Table 5. With the addition of SiO2 and Na2CO3, the contents of carbon and phosphorus in iron phase increased from 4.657% to 4.774%, 0.235% to 0.321% (Point D, E), respectively. Compared with carbon direct reduction, the contents of car-bon and phosphorus in iron phase increased from 3.659% to 4.774%, 0.160% to 0.321%, respectively (Point C, E). The results show that the addition of SiO2 and Na2CO3 promotes the reduction of apatite, but part of the reduced phosphorus gas enters the iron phase.meanwhile the phosphorus content in iron phase increases with the increase of carbon content. It is found that the distribution of phosphorus in the iron phase is related to the carburizing reaction.3.2. Effect of Sintering Time on Phosphorus Removal of SinterCompared with conventional sintering, the increase of carbon content in pre-reduction sintering makes the tem-perature curve of material layer rise faster and cool down slower. The heat transfer rate of material layer increases rapidly and has good performance in heat storage which would then offer sufficient heat and increase the maintain-ing time for apatite reduction. Therefore, it is particularly important to study the effect of reduction time on sintering gasification dephosphorization. Figure 11 shows the change of gasification dephosphorization rate under different reduc-tion time when SiO2 and Na2CO3 are selected as dephospho-rization agents and heated to 1 200°C. With the reduction time extended from 20 min to 60 min, the dephosphorization rate increased from 12% to 36%, and the iron metallization increased from 88% to 96%.Figure 12 shows the SEM images and EDS spectrum of the reduction products at different reduction time (20 min, 40 min, 60 min). In Fig. 12(a) at the reduction time of 20 min, most of P overlaps with Ca exists in the gangue phase, at which time a small amount of phosphorus migrates into the iron phase, and there is iron oxide that is not completely reduced. In Fig. 12(b) at the reduction time of 40 min, the Fig. 8. Representative SEM-EDS analysis of reduced samples: (a) C; (b) C+SiO2; (c) C+SiO2+Na2CO3. (Online version in color.)ISIJ International, Vol. 61 (2021), No. 5© 2021 ISIJ1466Fig. 9. Representative SEM-EDS analysis of reduced samples: (a) C+SiO2+Al2O3; (b) C+SiO2+Al2O3+Na2CO3. (Online version in color.)Fig. 10. Representative EPMA analysis of reduced samples: (a) C; (b) C+SiO2+Na2CO3. (Online version in color.)ISIJ International, Vol. 61 (2021), No. 5© 2021 ISIJ1467phosphorus in metallic iron increased obviously showing a disordered microstructure. The phosphorus in metallic iron is uniformly distributed with the reduction time to 60 min, as shown in Fig. 12(c). The SEM-EDS analysis shows that the content of phosphorus in metallic iron increases obvi-ously with the increase of reduction time. This is mainly because the reduction of iron oxide is mainly carried out at the initial stage of reduction, and the low carbon content in iron mainly exists in the solid phase. With the increase of reduction time, the carbon content in the iron phase increases and the liquid iron increases, which reduces the resistance of phosphorus into the iron phase and increases the phosphorus content in the iron phase. In conventional sintering process, due to rapid cooling down of the upper material layer the dephosphorization reaction could not process fully, and this could lead to lowering of dephosphorization rate.To sum up, combined with the characteristics of high carbon content,strong reduction atmosphere around carbon particles and negative pressure operation of the system in the pre-reduction sintering process, the thermodynamic and kinetic conditions of apatite reduction were improved by adding Na2CO3 and SiO2 as dephosphorization agents, the phosphorous gas produced by the reduction of apatite is dis-charged with the sintering waste gas to achieve the purpose of sintering gasification and dephosphorization. However, on the one hand, the Carburizing reaction is improved with high carbon content and high metallization rate. On the other hand, the reduction temperature of iron oxide is lower than that of apatite. Before the reduction of apatite, the metal iron produced by reduction roasting may wrap the unreduced apatite layer, thus accelerating the entry of phosphorus into the iron phase and reducing the gasification Fig. 11. Effect of sintering time on dephosphorization and iron metallization of sinter. (Online version in color.)Fig. 12. Representative SEM-EDS analysis of reduced samples: (a) 20 min; (b) 40 min; (3) 60 min. (Online version in color.)Table 5. Determine the element content in the metal iron of the reduced sample in Fig. 11 by EPMA.ElementC (wt%)P (wt%)Fe (wt%)C3.6590.16096.181D4.6570.23595.108E4.7740.32194.905ISIJ International, Vol. 61 (2021), No. 5© 2021 ISIJ1468dephosphorization rate. Therefore, in the future work, the method of restraining the migration of phosphorus to iron phase during sintering should be further studied to improve the gasification dephosphorization rate of pre-reduction sintering.4. Conclusions(1) According to the effect of dephosphorization agents SiO2, Na2CO3 and Al2O3 on the gasification removal of phosphorus in the sintering process, the addition of SiO2 and Na2CO3 decreased the starting temperature of apatite carbothermal reduction, inhibited the entry of phosphorus into the iron phase, and the dephosphorization effect was the best.(2) During the sintering process, part of the produced phosphorus is discharged with waste gas, and part of produced phosphorus enters into metallic iron, and the absorption of phosphorus by metallic iron is related to the carburizing reaction.(3) With increase of reduction time, the reduction of apatite is fully carried out and the dephosphorization rate is increased. The high carbon content and good heat storage of pre-reduction sintering could increase the temperature keeping time for the apatite reaction, which is beneficial to dephosphorization.AcknowledgementsThe authors gratefully acknowledge the Project sup-ported by National Natural Science Foundation of China (U1960205).REFERENCES1) W. Jie, W. Zhijun and C. Mingjin: Steel Res. Int., 82 (2011), 494.2) Y. Sun, Q. Zhang, Y. Han, P. Gao and G. Li: JOM, 70 (2018), 144.3) Y. Zhang, Q. Xue, G. Wang and J. Wang: ISIJ Int., 58 (2018), 2219.4) Z. H. Li, Y. X. Han, Y. S. Sun and P. Gao: J. Northeast. Univ. (Nat. Sci.), 36 (2015), 1757.5) H. Han, D. Duan, X. Wang and S. Chen: Metall. Mater. Trans. B, 45 (2014), 1634.6) X. Zhang, P. Cao, H. Zhou and D. Wang: Nonferrous Met. (Min. Sect.), 70 (2018), 1.7) Z. Tian, B. Li, X. Zhang and Z. Jiang: J. Iron Steel Res. Int., 16 (2009), 6. https://doi.org/10.1016/S1006-706X(09)60036-48) S. H. Song, J. Wu, L. Q. Weng and T. H. Xi: Mater. Sci. Eng. A, 520 (2009), 97.9) B. Lv, F. C. Zhang, M. Li, R. J. Hou, L. H. Qian and T. S. Wang: Mater. Sci. Eng. A, 527 (2010), 5648.10) M. Miyata, T. Tamura and Y. Higuchi: ISIJ Int., 57 (2017), 1751.11) N. Oyama, T. Higuchi, S. Machida, H. Sato and K. Takeda: ISIJ Int., 49 (2009), 650.12) J. Zhao, Z. Chen, H. Zuo, J. Wang and Q. Xue: Metals, 8 (2018), 1003.13) K. Ionkov, S. Gaydardzhiev, A. Correa de Araujo, D. Bastin and M. Lacoste: Miner. Eng., 46–47 (2013), 119.14) S. Bai, S. Wen, D. Liu, W. Zhang and Q. Cao: ISIJ Int., 52 (2012), 1757.15) A. P. L. Nunes, C. L. L. Pinto, G. E. S. Valadão and P. R. de Magalhães Viana: Miner. Eng., 39 (2012), 206.16) C. Lv, S. Wen, K. Yang and S. Bai: Steel Res. Int., 88 (2017), 1600314. https://doi.org/10.1002/srin.20160031417) C. Y. Cheng, V. N. Misra, J. Clough and R. Muni: Miner. Eng., 12 (1999), 1083.18) H. Q. Tang, Z. C. Guo and Z. L. Zhao: J. Iron Steel Res. Int., 17 (2010), 1.19) A. L. Petelin, L. A. Polulyakh, D. B. Makeev and V. Y. Dashevskii: Russ. Metall., 2018 (2018), 1.20) E. Matinde and M. Hino: ISIJ Int., 51 (2011), 220.21) C. N. Anyakwo and O. W. Obot: J. Miner. Mater. Charact. Eng., 9 (2010), 1131.22) C. Du, X. Gao, S. Ueda and S. Kitamura: ISIJ Int., 58 (2018), 860.23) M. Altiner: Can. Metall. Q., 58 (2019), 445.24) J. Yin, X. Lv, C. Bai, G. Qiu, S. Ma and B. Xie: ISIJ Int., 52 (2012), 1579.25) C. Xiao, X. Wu and R. Chi: Appl. Biochem. Biotechnol., 176 (2015), 518.26) H. Sato, S. Machida, K. Nushiro, K. Ichikawa, M. Sato, T. Ariyama and K. Takeda: Tetsu-to-Hagané, 92 (2006), 815 (in Japanese).27) T. Ariyama and M. Sato: ISIJ Int., 46 (2006), 1736.28) K. I. Ohno, H. Konishi, T. Watanabe, S. Ishihara, S. Natsui, T. Maeda and K. Kunitomo: ISIJ Int., 60 (2020), 1520.29) C. Lan, Q. Lü, S. Zhang and M. Jiang: J. Iron Steel Res., 30 (2018), 943.30) Z. M. Cao, K. P. Wang, Z. Y. Qiao and G. W. Du: Acta Phys. Chim. Sin., 28 (2012), 37. | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/350855106The Volunteers of Desistance (VoD): a concrete French link betweenrestorative justice and desistance (eng.version, journal: Cahiers de la Sécuritéet de la Justice, Revue de l'Ins...Article · April 2021CITATIONS0READS3682 authors, including:Some of the authors of this publication are also working on these related projects:Psychocriminological Investigation of Crime Scene (PICS) View projectTIM-E : Modèle de l'Identité Temporelle et Entretiens Temporels / Temporal Identity Model & Experiencing and Temporal Interviewings View projectErwan DieuARCA173 PUBLICATIONS 1,072 CITATIONS SEE PROFILEAll content following this page was uploaded by Erwan Dieu on 14 April 2021.The user has requested enhancement of the downloaded file.Ont participé à ce numéro :Laura ASCONE, Olivier CHOVAUX, Pierre-Alain CLÉMENT, Alexis DEPRAU, Erwan DIEU, Jessica FILIPPI, Cécile GODÉ, Johanne GOJKOVIC-LETTE, Hinda HEDHILI-AZEMA, Grégory HOUILLON, Jean Fabrice LEBRATY, Laurent MERCHAT, Marie NICOLAS-GRÉCIANO, Laurène RENAUT, Jordan VAZQUEZCAHIERS de lasécurité et de la justiceHors-sérieRevue de l’Institut des hautes études du ministère de l’IntérieurDe la théorie de la prévention à ses applications numériques, la trajectoire d’une idée humaniste© Cahiers de la sécurité et de la justice - IHEMISommaire2 Éditorial – Éric FreysselinardDossier3 Introduction Pierre-Alain Clément5 La mutation doctrinale de la prévention comme politique publique criminelle dans l'œuvre de Marc Ancel Hinda Hedhili-Azema11 The doctrinal mutation of the concept of prevention as public criminal policy in the work of Marc Ancel Hinda Hedhili-Azema17 Le Programme de parrainage de désistance (P.P.D.) : une mise en lien concrète française de la justice restaurative et de la désistance Erwan Dieu, Laurent Merchat27 The Volunteers of Desistance (VoD): a concrete French link between restorative justice and desistance Erwan Dieu, Laurent Merchat37 La justice restaurative des mineurs en France : entre tendance maximaliste et minimaliste Jessica Filippi48 Restorative justice for juveniles in France: maximalist and minimalist tendencies Jessica Filippi59 Prévention et répression de la pratique du drone civil : un équilibre à trouver Johanne Gojkovic-Lette, Grégory Houillon67 Prevention and repression of civil drone use: finding a balance Johanne Gojkovic-Lette, Grégory Houillon 75 Stop-djihadisme ou comment déjouer la radicalisation : examen critique d'un plan de prévention par l'information Laurène Renaut, Laura Ascone87 “Stop-djihadisme”, or how to thwart radicalization: a critical examination of a plan of prevention through information Laurène Renaut, Laura Ascone99 L’intelligence artificielle : nouvel outil au service de la prévention de la récidive ? Marie Nicolas-Gréciano 108 Artificial Intelligence: a new tool to prevent reoffending? Marie Nicolas-Gréciano VVaria116 Le football amateur, « terrain » ou « terreau » de la radicalisation religieuse ? L’exemple de la ligue des Hauts-de-France (2015-2018) Olivier Chovaux125 Are amateur soccer fields a breeding ground for religious radicalization? Case study of the Hauts-de-France League (2015-2018) Olivier Chovaux134 Renseignement, délinquance et violences urbaines Alexis Deprau 141 I ntelligence, delinquency and urban violence Alexis Deprau148 Environnement big data et prise de décision : l'étape de contre-la-montre du Tour de France 2017 Jordan Vazquez, Cécile Godé, Jean Fabrice Lebraty160 Big data environments and decision-making: the time trial stage of the 2017 Tour de France Jordan Vazquez, Cécile Godé, Jean Fabrice LebratyNuméro spécial des Cahiers de la sécurité et de la justice Hors-série "Cahier de recherche"Directeur de la publication : Éric FreysselinardRédacteur en chef : Manuel Palacio Coordinateur du numéro : Pierre-Alain Clément, avec l’appui de Nacer LALAMTraduction : Iris FellousResponsables de la communication : Simon Nataf, Claire BrisouxConception graphique : Laetitia Bégot© Direction de l’information légale et administrative, Paris, 2021Conditions de publication : Les Cahiers de la sécurité et de la justice publient des articles, des comptes rendus de colloques ou de séminaires et des notes bibliographiques relatifs aux différents aspects nationaux et comparés de la sécurité et de ses acteurs. Les offres de contribution sont à proposer à la rédaction pour évaluation. Les manuscrits soumis ne sont pas retournés à leurs auteurs. Toute correspondance est à adresser à l’IHEMI à la rédaction de la revue. Tél. : +33 (0)1 76 64 89 00Premier trimestre 2021 – Hors sérieCAHIERS DE LA SÉCURITÉ ET DE LA JUSTICE IHEMI École militaire - Case 3975700 Paris 07 SPTél : +33 (0)1 76 64 89 00www.ihemi.fr© Cahiers de la sécurité et de la justice - IHEMIe premier numéro des Cahiers de la sécurité et de la justice publié par l'IHEMI s’inscrit dans une longue et riche histoire de construction de passerelle entre chercheurs et décideurs sur les questions de sécurité et de justice. Cette construction relève d’une double démarche. En premier lieu, une démarche de diffusion des connaissances sous une forme accessible, propice à son appropriation par les praticiens. Et en deuxième lieu, une démarche de documentation et de conceptualisation par les praticiens de leur travail, un retour d'expérience essentiel pour les chercheurs. C’est à cette fin que l’I’IHEMI accueille dans ce numéro spécial des Cahiers de la sécurité et de la justice des contributions rédigées et évaluées scientifiquement par des chercheurs du monde de la sécurité et de la justice.En conservant les Cahiers, l’IHEMI reven-dique sa filiation avec les instituts qui l'ont précédé et la continuation de son rôle de passeur. En effet, les Cahiers de la sécurité et de la justice, lancés en 2006, étaient la revue phare de ses prédécesseurs, à savoir l'Institut national des hautes études de la sécurité et de la justice (INHESJ, 2010-2020) et l'Institut national des hautes études de la sécurité (INHES, 2004-2009). Elle a elle-même succédé aux Cahiers de la sécurité intérieure (1990-2006), lancée par l’orga-nisme originel, l'Institut des hautes études de la sécurité intérieure (IHESI, 1989-2004).Créés comme une revue de sociologie de la police, les Cahiers de la sécurité intérieure se donnaient comme objectif spécifique de diffuser en France les résultats des travaux de la sociologie nord-américaine de la police pour stimuler d'une part la construction d'une sociologie française de la police alors balbutiante et d'autre part celle d'une toute nouvelle direction de la formation de la police. À cet égard, la rubrique « Les fondamentaux de la sécurité » proposait dans chaque numéro la publication commentée d'un « classique » de sociologie anglophone.S’adaptant d'une part à la structuration d'une sociologie de la police et d'autre part à l'institutionnalisation d'échanges entre théoriciens et praticiens, les Cahiers de la sécurité et de la justice ont changé de format pour devenir hybrides : ni totalement revue scientifique, ni purement revue professionnelle. Cette position intermédiaire est source de richesse, par les discussions entre univers culturels différents mais unis par leur sujet, par le décloisonnement et le transfert des connaissances.La création de l’IHEMI s’inscrit dans cette évolution et en revendique la continuité. Ce nouvel institut, toujours chargé de former et de mener des recherches au contact des opérationnels et avec indépendance, a aussi vocation à être un incubateur de prospective sur les questions de sécurité. En ciblant ici le thème spécifique de la prévention, ce numéro spécial répond à cette vocation, tout en répondant à des préoccupations universitaires et pratiques actuelles.À travers ce numéro spécial, nous vous vous souhaitons donc la bienvenue dans ce nouvel institut, au service de la connaissance et des politiques publiques éclairées nCÉditorialÉric FREYSSELINARDDirecteur de l’Institut des hautes études du ministère de l’Intérieur2 I ÉDITORIALÉric Freysselinard est agrégé d’espagnol, diplômé de Sciences Po Paris et de l’ENA. Après avoir enseigné l’espagnol pendant plusieurs années, il a occupé de nombreux postes en administration centrale, en inspection générale, en cabinet ministériel et en préfecture. Préfet délégué à l’égalité des chances dans l’Essonne, préfet de département à Vesoul et Carcassonne, il a mené une mission de près de cinq ans comme directeur des stages de l’ENA et a été préfet de Meurthe-et-Moselle pendant près de trois ans ans avant d’être nommé directeur de l’IHEMI le 3 octobre 2020.Éric Freysselinard a publié une dizaine d’ouvrages d’espagnol, de catalan et d’histoire contemporaine. Éric FREYSSELINARD© Cahiers de la sécurité et de la justice - IHEMICahiers de la sécurité et de la justice – Hors-série INTRODUCTION I DOSSIER I 3 ermée au printemps 2019, l’idée d’un numéro spécial des Cahiers de la sécurité et de la justice composé d’articles arbitrés (peer-reviewed) a émergé comme une nouvelle modalité du rôle de passerelle entre théoriciens et praticiens que joue cette revue. Ainsi, ce numéro spécial se distingue car il entend valoriser la recherche sur ces sujets, dans un format court et accessible facilitant son usage par les institutionnels. Pour en renforcer la diffusion, il publie, en ligne et gratuitement, en français et en anglais, des contributions venant des multiples sciences sociales, avec une ouverture sur les sciences formelles et appliquées.Ce positionnement s’inscrit dans le contexte d’une absence de reconnaissance de la criminologie comme discipline, en particulier sans section dédiée au CNU. Certes, les revues ne manquent pas qui traitent des questions de sécurité et justice, et elles connaissent quelques décennies d’ancienneté 1. Rassemblant des disciplines proches (sociologie, science politique, droit), elles font désormais figure de référence. Il faut attendre le tournant des années 2000 pour observer une seconde vague de création de revues dédiées aux questions de sécurité et justice 2. Prises ensemble, elles se caractérisent par une volonté de s’extraire des structures existantes : positionnement pluridisciplinaire, traduction vers anglais, publication en continu, diffusion parfois exclusivement numérique. Après un nouveau hiatus d’une dizaine d’année, deux revues thématiques ont émergé à la fin des années 2010, avec des approches très différentes 3.En observant plus en détail ces revues, on constate que, malgré les obstacles épistémologiques et professionnels, l’interdisciplinarité est promue au sein des sciences humaines, qui s’étend parfois à la médecine légale (Criminalistique) ou les arts (Violence, Cultures et Conflits depuis 2005). De plus, la norme unilingue recule, plusieurs proposant des résumés en anglais (Cultures et Conflits ; Revue interdisciplinaire d’études juridiques), des articles en anglais ou en espagnol (Criminocorpus ; Civitas Europa) ou des traductions d’articles de l’anglais (Droit et Société).Cette histoire s’accompagne également d’une évolution de l’économie de l’édition scientifique. La plupart de ces revues fonctionnent avec un système de barrière d’accès mobile. Dans les années 2000, deux tendances divergentes naissent. D’une part, des revues plus orientées vers un public praticien et publiées par des éditeurs privés spécialisés comme Dalloz (Cahiers de la justice) ou ESKA (Sécurité globale, Criminalistique). D’autre part, des revues adossées à des institutions de GIntroduction Pierre-Alain CLÉMENTChargé de recherche à l’IHEMI, coordinateur du numéro spécial(1) Voir : les Archives de politique criminelle (1975), Déviance et Société (1977), la Revue interdisciplinaire d’études juridiques (1978) Droit et Société (1985) et dans une moindre mesure Cultures et Conflits (1990) et Champ de Mars (1996).(2) Voir : Histoire de la justice (2000), Champ pénal (2004), Criminocorpus (2005) et les Cahiers de la justice (2006)(3) Voir : Criminalistique (2018) et Violence (2019).Pierre-Alain Clément est chargé de recherche à l’Institut national des hautes études de la sécurité et de la justice et secrétaire adjoint de Regards géopolitiques, revue du Conseil québécois d’études géopolitiques. Il est l’auteur de plusieurs publications sur le terrorisme et l’antiterrorisme, dont une monographie (G.I. contre jihad, PUQ, 2010), des articles (« Définition du terrorisme », Le Banquet, 2010 ; « Le terrorisme est une violence politique comme les autres », Études internationales, 2014), mais aussi des chapitres d’ouvrages arbitrés (« Le contre-terrorisme à l’ère multipolaire », 2013). Il a également publié des articles arbitrés sur la culture populaire (« La signification du politique dans le rap », Cultures et Conflits, 2015 ; avec Barthélémy Courmont « When geopolitics meets the game industry », Hemispheres, 2014.) (pierre-alain.clement@ihemi.fr.)Pierre-Alain CLÉMENT © Cahiers de la sécurité et de la justice - IHEMI4 I DOSSIER I INTRODUCTIONrecherche optant pour un format numérique et gratuit (Champ pénal, Criminocorpus, Violence), qui correspond aux objectifs français et européens de science ouverte, qui inclut le libre accès (open access).Finalement, pourquoi le thème de la prévention ? Son intérêt est reconnu de longue date dans le monde anglophone, trois revues portant sur le sujet étant répertoriées par l’American society of criminology 4. Le thème connaît un intérêt soutenu puisque Palgrave Macmillan a lancé en 2008 une collection dédiée au sujet 5. Cet intérêt pour la prévention est encore plus marqué en dehors des sciences sociales puisque ce thème est déjà fortement traité par d’autres disciplines, et plus particulièrement les sciences médicales 6 et les sciences de l’ingénieur 7. En France, En France cependant, le thème reste très peu exploité spécifiquement, si ce n’est le magazine à la croisée des mondes médical et industriel Prévention au travail (2012) et l’organisateur d’évènements Préventica (1997).L’intérêt pour la prévention est donc légitime. Il procède d’une logique d’action publique reconnue, en particulier hors de France où la notion est moins construite en opposition avec la notion de répression. Prévention de la récidive, prévention de la délinquance, prévention de la radicalisation : ces thèmes mobilisent régulièrement les décideurs et administrateurs des politiques publiques. Les efforts de prospective, tels que ceux que l’IHEMI se donne pour mission d’accomplir, peuvent également être lus comme une facette de la prévention des scénarios indésirables.C’est dans la logique de stimulation et d’animation d’une communauté de recherche spécifiquement axée sur la prévention que ce numéro spécial s’inscrit, en tenant compte des évolutions du champ de l’édition scientifique au sens large et sur les questions de sécurité et de justice en particulier. Il a été pensé, dans son appel à contributions, pour inclure l’ensemble des enjeux de prévention : délinquance, radicalisation, récidive, gestion des risques, préparation de la résilience, etc.À cet égard, les propositions répondant à l’appel à contributions montrent la diversité des questionnements actuels sur la prévention des risques et menaces. Partant de la conviction que la politique pénale pourrait avoir d’autres ressorts que la punition, Marc Ancel a défendu l’idée de réadaptation sociale du délinquant (article de Hinda Hedhili-Azema), ouvrant la voie à la justice réparatrice face à une justice traditionnellement rétributive (article d’Erwan Dieu), y compris à l’égard des mineurs (article de Jessica Filippi). Pourtant, avec l’émergence de nouveaux risques et menaces, liés à la technologie (article de Johanne Gojkovic-Lette et Grégory Houillon), la prévention, elle aussi adoptant des outils numériques (article de Laurène Renaut et Laura Ascone), pose de multiples questions sur l’anticipation du crime, d’ordre éthique et juridique entre autres (article de Marie Nicolas-Gréciano).Dans la section varia, les contributions s’attaquent à des sujets tout aussi cruciaux. Se pose d’abord la délicate question, encore peu documentée rigoureusement, du lien éventuel entre sport amateur et radicalisation (article d’Olivier Chovaux). De même, est ici interrogée la réalité d’une hybridation, maintes fois évoquées et toujours débattue, entre délinquance et radicalité (article d’Alexis Deprau). Enfin, la question de l’usage des nouvelles technologies, cette fois au service de la protection civile, est examinée en détail à partir d’un évènement sportif majeur (article de Jordan Vazquez, Cécile Godé et Jean-Fabrice Lebraty).Enfin, ce numéro n’aurait pu voir le jour sans la vingtaine de chercheurs et chercheuses, en poste dans divers pays de la francophonie, qui s’est amicalement prêtée à l’exercice de la révision anonyme et bénévole afin de sélectionner les propositions de contributions et d’améliorer les articles retenus. Qu’ils et elles en soient ici chaleureusement remerciés n(4) Voir : Journal of prevention and intervention in the community (1976), Crime prevention and community safety (1999) et le Journal of Scandinavian studies in criminology and crime prevention (2000).(5) Intitulée « Crime prevention and security management ».(6) Accident analysis and prevention (1969) et Prevention science (2000).(7) Journal of loss prevention in the process industries (1988) et Disaster prevention and management : an international journal (1992).© Cahiers de la sécurité et de la justice - IHEMI DOSSIER I 5Cahiers de la sécurité et de la justice – Hors série n°1La mutation doctrinale de la prévention comme politique publique criminelle dans l’œuvre de Marc AncelHinda HEDHILI-AZEMAa prévention pénale est une notion ancienne et nouvelle qui connaît par sa polysémie historique une forme d’instabilité doctrinale. Cela signifie qu’elle a un sens changeant et parfois contradictoire dans l’écriture des penseurs du droit, ceci dans le temps long. Cette instabilité fondatrice n’est pas pour autant négative, puisqu’elle fait de la prévention une notion modulable qui varie selon la société. L’intention préventive doit donc être comprise dans sa subtilité pour assurer une efficacité des politiques pénales. Le doute que suscite la prévention provient pour sûr de l’héritage délicat du xixe siècle. On n’évoque d’ailleurs que très rarement le sens premier de la prévention dans le débat public 1, tant sa définition historique est enfouie. Poser ce sens fondamental permet pourtant de clarifier ses divers degrés de Marc Ancel est souvent présenté au sein des facultés de droit françaises comme un criminaliste et pénaliste contemporain, ayant œuvré pour la dépénalisation. À l’international, il demeure l’un des fondateurs d’un système de sanction moderne prenant en compte la nature humaine des individus. Ce mouvement de dépénalisation s’inscrit dans le cadre du mouvement de la défense sociale nouvelle créé par Marc Ancel lui-même en 1954. Ce courant a été pourtant mal compris et ce malgré la clarté des textes de Marc Ancel. Il définit dans ses ouvrages une nouvelle façon de traiter le problème criminel. Il propose d’abord de dépasser l’étude juridique pour mettre en place une étude scientifique au service du procès. Il critique pour ce faire la répression légaliste, dans laquelle le juge obéit à la lettre de la loi. Il rénove la sanction en procédant à une enquête de personnalité. Sans jamais nier le droit, Ancel considère que le juge doit étudier la personnalité du criminel ou du délinquant et peut mettre en place une peine évolutive. Il consacre surtout ses efforts à humaniser une action pénale autrefois liée à une « peine-châtiment ». Son œuvre est une source d’inspiration pour une politique criminelle juste et humaniste.LHinda HEDHILI-AZEMAMaître de conférences en histoire du droit et des institutions, membre Centre aquitain d’histoire du droit (C.A.H.D.) et de l’Institut de recherche Montesquieu (I.R.M.). Spécialiste de la doctrine pénale des XIXe et XXe siècles.(1) Les contemporains prennent la prévention pénale davantage dans le sens d’une simple prévention de la récidive. Cette optique est aussi relayée par l’État, voir Les 12 recommandations pour une nouvelle politique contre la récidive, 20 février 2013. L’état actuel de la question de la prévention ne peut être abordé dans cet article historique. On dénote simplement une dédoctrinalisation du terme, pris comme un terme plutôt technique ou un objectif du parcours post-pénitentiaire.© Cahiers de la sécurité et de la justice - IHEMI6 I DOSSIER La mutation doctrinale de la prévention comme politique publique criminelle dans l’œuvre de Marc Ancel – Hinda Hedhili-Azemacompréhension. Cette définition se veut un outil pour celui qui souhaite penser la peine à travers son évolution historique et scientifique. La prévention n’y est pas un détail technique parmi d’autres, mais bien un principe général du droit qui a vocation à régir un système pénal. À première vue simpliste, l’idée historique préventive relève en réalité d’un enchevêtrement complexe de plusieurs intérêts et puissances concurrentes : judiciaire, politique et administrative. La prévention organise justement dans l’œuvre d’Ancel un espace de collaboration institutionnelle fécond. Si elle n’est pas une tâche complémentaire, l’action préventive est conçue comme un véritable logiciel social. Tantôt décriée comme liberticide, tantôt adulée comme non pénale, la prévention ne se plie pas aux cadres traditionnels des notions juridiques. Et pour cause : le principe de prévention est une réponse politique aux situations critiques de son temps. Il n’est donc pas étonnant que la plupart des lecteurs avertis du droit soient déroutés par l’approche historique, qui aborde la prévention non pas comme une forme technique, mais comme un idéal. Si l’on doit figer la prévention dans une définition de départ pour en faciliter l’étude, on dira que la prévention est une politique de contrôle du phénomène criminel, destinée à empêcher la réalisation du crime.Au regard de ses usages historiques, la prévention est plus que ça, elle est une idée-force qui se destine à évincer la peine châtiment 2. Le basculement que permet la pensée préventive française n’est pas véritablement perceptible pour nous contemporains, puisque nous n’avons pas connu, pour la plupart des lecteurs d’Ancel, la révolution humaniste de notre vivant ; il s’agit de l’action de lutte contre la dérive pénale. L’ère préventive est ce moment dans le monde du xxe siècle où la peine se plie au principe général de justice humaine. L’observateur contemporain a oublié que la peine est l’un des derniers domaines du droit à être sortie de l’archaïsme. Le courant de la prévention est acteur de ce basculement. Enfermer ce courant dans une technique juridique le viderait de sa signification première, philosophique. Il nous faut pourtant ici préciser que les propos qui suivront tenteront de démontrer que l’enjeu de la politique préventive française a été de faire glisser la notion de punition vers la notion de mesure pénale. Cet élan n’est pas isolé, il s’intègre dans un mouvement mondial qui prescrit le progrès du droit par la comparaison des législations. L’action préventive demeure d’ailleurs un principe toujours relayé par la doctrine européenne contemporaine et la majorité des gouvernements occidentaux. Si on se concentre pourtant sur l’école française, la prévention prend en compte l’ensemble des mesures qui permettent de régler les problèmes criminels et de déviance de la société, avec cette indication qu’elle n’oppose pas frontalement le groupe et l’individu criminel ou délinquant. Elle réalise plutôt une forme de paix sociale. Elle est par là une science sociale complète et une science toujours en faveur du respect de la personne. La prévention contemporaine intriguerait d’ailleurs plus d’un criminaliste de l’ancien droit puisqu’elle franchit les frontières de la punition pour faire de la pénalité une politique polymorphe et pluridisciplinaire au service de l’intérêt individuel et de la protection sociale. Faire acte de prévention signifie donc gérer le phénomène criminel de manière moins agressive. Il n’en a pas toujours été ainsi dans la littérature criminelle ancienne. Il va sans dire que la prévention était en fait un levier du pouvoir politique 3. Le monarque possédait dans le cadre de sa majestas (Ancel, 1956 : 936) le pouvoir d’évoquer des affaires importantes et de les traiter par lui-même comme questions contentieuses épineuses. La prévention est le pouvoir politique de celui qui possède la fonction de dire le droit, autrement dit la puissance de juridiction. Dans une version plus technique, on détermine la prévention ancienne comme la possibilité pour une autorité de statuer sur un cas 4. Elle va de pair avec cette capacité et cette qualité du magistrat à attirer auprès de lui une affaire pour la juger en équité. De source multiple, notamment romaine et ecclésiastique, prévenir est un devoir moral de juger justement et plus largement une compétence de toutes les magistratures qu’elles soient politiques ou institutionnelles. La doctrine préventive est longtemps restée la désignation de l’action du juge, comme défenseur d’une justice criminelle équitable. Sa mutation en politique publique est une œuvre contemporaine, attribuée en France au renommé professeur Marc Ancel 5.Dans cette nouvelle configuration, prévenir le crime ne signifie plus la désignation d’un juge pour entendre une affaire ou encore une action précise de la loi, mais bien un programme scientifique de socialisation. La prévention est plus qu’une compétence, elle devient un procédé de résolution du problème criminel. C’est en ce sens qu’elle est politique. Dans l’œuvre du professeur Ancel, la prévention ne sert pas uniquement à juger, mais à comprendre un phénomène complexe. Ancel et ses acolytes 6 effectuèrent (2) Pour comprendre cette idée, nous vous renvoyons à notre ouvrage à paraître, intitulé Doctrine pénitentiaire française et méridionale.(3) Ceci est à tempérer considérant la concurrence des différentes justices du royaume : ecclésiastique, municipale, seigneuriale.(4) Portalis sur les bonnes lois et la nature des lois capables de prévenir les inégalités.(5) Marc Ancel a une renommée internationale, si bien qu’au moment de sa disparition en 1990, le VIIIe congrès des Nations unies sur la prévention du crime et le traitement des délinquants organisa un hommage solennel avec un représentant de chaque continent prenant part au congrès. Simone Rozes relève ses qualités remarquables de magistrat (Rozes, 1991).(6) Parmi ces derniers : Amor, Cannat, Clément, Charpentier, Gorphe, Heuyer, Herzog, Pinatel. © Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 7 Cahiers de la sécurité et de la justice – Hors-série donc une rupture avec le langage criminel du siècle précédent et proclameront solennellement une nouvelle conception de la sanction sociale. Ce système fondé sur la science criminelle moderne opéra le virage doctrinal de l’après-guerre 7. Le postulat d’Ancel ne faisant cependant pas l’unanimité (Rappaport, 1956 ; Dreyfus, 2010 : 17), il fut défendu par des structures, des écrits, des personnalités, qui croient en une défense de la société qui garantit sa préservation et non sa dissolution. La vocation sociétale de l’œuvre d’Ancel permet de promouvoir une conception unique de la prévention. Il faudra venir en aide à une frange de la société et rendre plus responsable la communauté des citoyens. À ce titre, on ne prévient pas pour rendre la société laxiste, mais bien pour diversifier le mode de sanction et faire de la répression punitive un mode marginal de la politique criminelle. La prévention ne s’oppose ainsi pas à la répression, mais lui est préférable et doit lui être préférée. L’état d’esprit est celui d’une meilleure observation 8, sans voile ni aveuglement. La société est capable d’observer et d’agir, sans réaction. Elle peut éviter les écueils d’un rigorisme pénal : le repli, la vengeance, la xénophobie, la stigmatisation et le châtiment. Si nous devions qualifier le sens contemporain de la prévention « ancelienne », nous dirions qu’elle désigne une politique dans laquelle la personne juridique criminelle et délinquante est placée en sécurité par la communauté des hommes pour être resocialisée. Par effet boule de neige, la protection sociale en est assurée. Le pragmatisme de la prévention emporte avec lui le souvenir d’un dogme répressif rigide. L’« État » de prévention est un « État » de justice nouveau. Il informe la personne criminelle qu’elle a rompu ses obligations et l’engage dans une voie de réintégration non négociable. Un acte de précaution, car il protège l’individu et la société. Le pacte préventif se situe entre justice et sécurité.La qualification de la défense sociale ancelienne comme nouvelle politique criminelle : au-delà du crimeL’œuvre de Marc Ancel témoigne de sa hauteur de vue et de sa justesse d’esprit. La société doit-elle nécessairement punir ? Selon lui, la réponse est certainement positive. La voie ouverte par Marc Ancel se répand sur de multiples territoires et entraîne la plume de plusieurs auteurs qui lui sont affiliés. Il sera en ce sens considéré par ses pairs comme le créateur de la pénalité moderne de l’après-guerre, tant par sa pensée que par son influence sur la codification pénale. Né le 14 juillet 1902, magistrat, professeur, membre de l’Académie des sciences morales et politiques, il est l’instigateur d’un courant doctrinal majeur. Son travail s’inscrit dans le cadre d’un mouvement collectif 9 (Hedhili-Azema, 2016), et sa prise de position demeure particulièrement importante. À cette époque s’opère dans le débat doctrinal européen une différenciation entre prévention du crime, c’est-à-dire solution non répressive du phénomène criminel, et lutte préventive contre la délinquance, présentée comme un problème différent. Ancel embrasse cette distinction et réintègre intelligemment l’héritage doctrinal passé. Il s’affilie aux idées modernistes qui plaçaient dès le xviiie siècle la réponse au crime comme un devoir de toute société organisée avec respect du contrat social 10. Il reprend à son compte la valeur scientifique des études réformatrices et morales pénitentiaires (Hedhili-Azema, 2014) et purge les théories positivistes antérieures de leurs vices. Ancel formalise par ses propositions une politique annoncée dès le début de la Troisième République, à travers des mesures qui débordent largement de l’action judiciaire et de la froideur légale. Confier la prévention des crimes à la seule action de l’arithmétique pénale, c’est amputer la politique criminelle d’un levier immense qui se situe dans les zones extérieures de la technique légale. Car il symbolise l’intention de son auteur, l’ouvrage de Marc Ancel de 1953 dérange dès sa parution.Sa pensée s’inscrit en effet dans une approche assez singulière, celle de la plus grande vigilance à l’égard du respect de la dignité humaine. L’idée d’une peine qui pourrait porter atteinte à la personne physique semble un postulat ancien désormais dénigré par la majorité de la jeune garde pénale. Mais en ce temps-là, cette position était d’autant plus singulière que les années antérieures au second conflit mondial faisaient état d’une présence accrue de la peine corporelle forcée, d’un système pénitentiaire assimilé à une mise à mort lente du détenu (Hedhili-Azema, 2019) et d’une peine de mort largement plébiscitée par l’opinion. Marc Ancel, alors âgé de 52 ans, (7) Marc Ancel est marqué par ces questions d’après-guerre, en particulier la défense de la liberté individuelle réprimée par les régimes totalitaires, mais aussi la protection de la dignité humaine et du respect de la personne. Selon Levasseur, il va clairement s’associer à Paul Amor et à Jean Chazal en 1944, notamment dans la réforme de la législation de l’enfance délinquante par l’ordonnance du 2 février 1945. (8) L’ancienne défense sociale notamment celle de Gramatica part du postulat selon lequel une approche pluridisciplinaire avec une antisocialité du délinquant. Marc Ancel fait différer son sens puisqu’il ajoute une protection de l’être humain et une garantie des droits fondamentaux. (9) On compte Amor, Rolland, Chazal, Cannat, Herzog, Vienne, Heuyer, Roumajon, Colin, Michard, Sélosse, Yvonne Marx.(10) La société se doit de protéger ses composants mais elle a le pouvoir de protéger son unité, par la violence légitime.© Cahiers de la sécurité et de la justice - IHEMI8 I DOSSIER La mutation doctrinale de la prévention comme politique publique criminelle dans l’œuvre de Marc Ancel – Hinda Hedhili-Azemaécrivait à cette époque sous l’égide du Centre d’étude de défense sociale de l’Institut de droit comparé. Il va opérer un renversement de cette école. Son parcours de juriste le justifie, il est un praticien remarquable à plusieurs niveaux : son courage politique de résistant, sa rigueur scientifique et sa qualité de pensée. Ses écrits n’ont de cesse dès lors d’expliquer la trajectoire à prendre pour détourner la doctrine de la violence répressive. Formé à la fois aux lettres et au droit, il connaît bien le terrain de l’arsenal des peines. Il entre dans la magistrature en 1929 et s’intéresse comme chercheur aux pratiques pénales. Comme tous les penseurs de la pénalité, Ancel possède un profil de comparatiste qui le mène toujours à revoir le droit comme une enquête sur le territoire mondial, en vue de collecter les progrès à diffuser. Grâce à cette expérience pratique, Ancel revendique la valeur sociale de la politique criminelle. Cette dernière aura désormais pour but la possible réadaptation offerte aux hommes-citoyens. Une telle approche vient préciser que la politique criminelle est la compréhension du phénomène criminel et non une punition. Elle procède aussi à une seconde amputation à la nature de la peine classique en minimisant la fonction d’intimidation. Marc Ancel pense qu’il faut creuser, fouiller l’humain, la personnalité criminelle et délinquante et rejeter une répression dure, systématique et rationnelle de certains courants 11. Ce qui nous intéresse dans la genèse de la philosophie pénale de Marc Ancel, c’est la dimension positive qu’il lui attribue. Bien que s’inscrivant dans le cadre de la légalité, sa pensée place au-dessus du droit pénal une politique globale qui vise à protéger l’ordre social et l’être humain (Ancel, 1954 : 130).La symbolique historique de la prévention ancelienne comme forme de déjudiciarisation : au-delà du jugementLa prévention comme mesure de déjudiciarisation de l’action contre le crime doit bien être liée à la conception même de la défense sociale nouvelle. Elle contribue au mouvement d’ouverture méthodologique, mais aussi au moyen de réformer l’action du juge. Juger ce n’est pas condamner. Ancel présente la déjudiciarisation dans la seconde partie de son ouvrage publié en 1953 comme une des voies de transformation de l’ancienne prévention. Elle est qualifiée de voie positive. Ce processus débute par un élargissement des champs criminels qui induit que la politique active ne peut se réduire à l’élan strictement judiciaire. Une ouverture doit intervenir par des instruments issus des autres sciences du crime : la police, la biologie, la psychologie, la sociologie ou encore la médecine. Les anciennes disciplines dites auxiliaires du droit pénal sont hissées au même rang que l’art de la justice criminelle. Elles sont des outils et des sciences propres à exister à côté de la science normative. Pour Ancel, le droit pénal cesse d’apparaître une seule voie de régulation en soi. Ancel ne réfute jamais la nécessité de norme pénale et de la science du droit, il récuse le légalisme dans lequel s’enferme le juge depuis la Révolution. Le juge n’est plus en ce sens l’esclave de la loi, il en est sa conscience. Le droit pénal et ses représentants devront se plier à une nouvelle équité. Plus précisément, le droit, ses théories, ses notions, ses fondements ne sont justifiés que s’ils concourent à la finalité sociale que se donnent les citoyens : la paix, la concorde, la fraternité, la liberté. Or la fin de l’« a-priorisme » juridique préconisé par Ancel n’est pas une position antijuridique, au contraire elle est un retour au vrai droit. Le criminaliste doit avoir un point de vue technique certes mais aussi efficace. Il ne doit pas sacrifier son observation à la simple lettre de la loi. Ancel préconise donc une attitude nouvelle du juriste qu’il décrit comme réaliste. Cette voie contribuera à une transformation plus profonde du sens de la peine. Pour Ancel, il s’agit en vérité de considérer que la justice pénale ne produit pas nécessairement une fonction d’expiation, de vengeance ou de rétribution. Il faut dès lors vider la sanction de ses fonctions archaïques et modérer la fonction punitive. Ceci non pas dans le sens d’une absence de sanction pénale, mais dans le sens de supprimer l’idée d’un châtiment. Il ne faut plus punir, mais traiter la délinquance. La modération de la répression n’est pas la déclaration de l’irresponsabilité criminelle comme il a pu être répandu par ses adversaires mais bien celle d’une responsabilité autrement sanctionnée. La nouvelle force de la politique criminelle est de traiter la question de manière scientifique : comprendre la situation personnelle du sujet, ses probabilités de relèvement et ses possibilités morales et psychiques pour contribuer à la croissance sociale. L’appréhension nouvelle rappelle les critères de la science pénitentiaire ancienne (Ancel, 1959). La sanction pénale rénove au passage une neutralité doctrinale qui aspire à restaurer la place de l’individu dans la société. La politique préventive tend donc à une réorientation générale de l’intention politique vers un traitement social de la personne et une action générale de protection de la société par le traitement individuel. Les moyens permettent une précaution de jugement, en déclenchant l’étude de la personnalité du délinquant. Le changement de ton est un point d’orgue de la nouvelle pensée pénale contemporaine, qui rappelle la voie de la (11) Ces courants seront explicités dans un ouvrage à venir.© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 9 Cahiers de la sécurité et de la justice – Hors-série doctrine pénitentiaire de la Monarchie de Juillet. Ancel parle d’un critère fonctionnel qui détermine l’état d’esprit de la nouvelle science des sanctions : le réalisme de la lutte contre le crime et son efficacité à l’égard des personnes concernées par les mesures.La définition matérielle de la prévention ancelienne comme forme de dépénalisation : au-delà de la peineLe contenu matériel de la prévention va se manifester par le traitement du cas par le juge à l’aide des autres acteurs de la politique criminelle : les experts criminalistes, criminologues, les institutions policières, pénitentiaires et sociales (Hedhili-Azema, 2015). La première expression préventive est tout naturellement une aide à la science du palais 12. En ce sens, l’allure du procès est modifiée. Le juge se voit sortir de la mécanique de la loi. Il ne jugera pas uniquement l’acte qualifié d’infraction, mais ira au-delà du simple critère objectif, sur des éléments subjectifs liés à la personnalité de l’auteur de l’acte (Gramatica, 1963). Cela impliquera pour lui de dépasser la simple lecture des faits et de porter à sa connaissance une observation individuelle. Apparaissent dans les lignes de la nouvelle défense des cadres précis d’expertise : la constitution biologique, la réaction psychologique, la situation sociale exigeant une collaboration de la part des experts. La mutation ne doit cependant pas affecter les droits afférents à la protection des prévenus. La dépénalisation offre de surcroît des innovations dans le prononcé de sanction, puisqu’elles ne distinguent plus peines et mesures de sûreté (Chazal, 1953). Ancel évoque très tôt l’idée que peine et mesure de sûreté cessent de s’opposer. La sanction est un nouveau langage engagé, qui décrit une évolution criminelle possible et un choix évolutif 13. Elle sera guidée par l’efficacité du moment, curative, éducative ou neutralisante. Le juge aura bien la possibilité d’une progressivité de la sanction, d’une flexibilité expliquée par des motifs personnels, mais Ancel réfute la classification des sanctions abstraites juridiques. Finalement, peu importe la qualification première de la sanction, c’est la coloration de la sanction qui compte et non sa dénomination juridique. Ancel prônera naturellement l’utilisation plus libre de la sanction pouvant évoluer sur le temps. Il réclamera d’ailleurs la sentence indéterminée comme véritable levier de régulation de la politique criminelle. Il rejettera l’illusion d’une fixité de la peine alors même que la sentence ne dépend que de l’évolution de l’enquête sociale. L’essence de la lutte criminelle préventive réside dans un sens profondément social qui cumule la précaution du juge et celle des institutions pénitentiaires. Il s’agit d’une collaboration institutionnelle. Ce sens profond reflète le but préventif. Ne pas punir, mais résorber l’a-socialisation, car elle ramène l’auteur dans une vie libre. On pourrait commenter Ancel et voir ici la liberté comme l’exercice neuf de la citoyenneté, et la conscience comme la prise d’une observation plus claire sur soi-même. Ancel opère surtout une réfutation importante de la pénalité ancienne, celle qui pénalise et détruit la société et les hommes, celle qui avait pour but l’élimination de l’être. La prévention réside dans la réfutation d’une conception erronée de la répression comme moyen d’infliger une souffrance au délinquant ou une destruction massive sur des motifs pénaux.La prévention est sans conteste ce mouvement humaniste, qui prône l’éducation des citoyens, la mentalité de l’opinion vers une meilleure justice, dans laquelle l’homme ne décide pas de la destruction d’un autre homme. Il y a bien sûr le rejet d’un rigorisme pénal propre aux sociétés traditionnelles, mais il y a encore une forme d’idéal du bien. La modération (Hedhili-Azema, 2016) ressort comme le principe fédérateur de Marc Ancel, et ici ce dernier poursuit un héritage ancien 14. Est engendrée naturellement par l’intention préventive de l’auteur, une communauté en responsabilité vis-à-vis de l’individu en perdition. Sur le plan politique, le modèle ancelien engendre automatiquement un universalisme qui favorise la transposition de la prévention dans les législations étrangères. Pour construire ce modèle universel, Ancel doit proclamer des droits nouveaux de la société et même de l’homme. Si l’homme a manqué à ses obligations sociales, il ne peut être traité par le système préventif sans considération de ses droits naturels et sociaux. Dans cette conception moderne de la prévention, les acteurs œuvrent à rendre possible une politique consacrée à ce droit de resocialisation qui faisait défaut dans le système ancien. Ceci se rattache sans nul doute à un motif naturel dans lequel le droit de sanction de la société aboutit à la restauration de droits subjectifs : liberté, sécurité, propriété et travail. L’étude ancelienne respecte la personne criminelle ou délinquante, sans jamais la rendre irresponsable de ses actes. Cette humanisation rappelle que l’homme social est libre de tomber dans le crime, mais aussi libre d’en sortir. Il affecte à la politique criminelle une idée qui redonne un sens à une technique viciée. La politique criminelle (12) La science du palais désigne la production du corps judiciaire, magistrature, avocats et avoués afin de commenter le droit.(13) Voir le IIIe congrès international de défense sociale, Anvers 20-24 avril 1954.(14) À cette époque, l’écriture d’Ancel rejette tout de même un concept de ses pairs, le châtiment, comme centre du système criminel.© Cahiers de la sécurité et de la justice - IHEMI10 I DOSSIER La mutation doctrinale de la prévention comme politique publique criminelle dans l’œuvre de Marc Ancel – Hinda Hedhili-Azemad’Ancel sera animée par un engagement de nature saine, une orientation intellectuelle qui dirige l’action juridique vers un bien faire. L’idéal ancelien n’est pas un irréalisme pénal utopique, mais bien un non-rigorisme. Cet idéal demande une compréhension subtile que nous avons tenté d’offrir ici au lecteur. Il nous apparaît nécessaire de préciser que l’étude ancelienne a parfois été mal interprétée et réduite à tort au laxisme pénal. À l’heure du repli nationaliste européen et international, nourri par un sentiment d’insécurité, la pensée ancelienne peut permettre aux décisionnaires d’avoir un modèle, contre les dérives autoritaires et les populismes ambiants. À ce titre, on peut considérer Marc Ancel comme l’instigateur d’une œuvre fondamentale. Ses pensées n’étant pas, à leur époque, majoritaires ni d’ailleurs plébiscitées, il lui est apparu nécessaire de les écrire pour laisser une trace de son intention doctrinale. Il apparaît aujourd’hui nécessaire de les restituer de manière claire afin d’enrichir le débat sur les politiques criminelles actuelles nRéférencesAncel, Marc, 1954, la Défense sociale nouvelle, un mouvement de politique criminelle humaniste, Cujas, Paris.Ancel, Marc, 1956, « Politique criminelle et psychologie judiciaire dans la détermination de la sanction pénale », Revue de science criminelle et de droit pénal comparé, n°2, avril-juin.Ancel, Marc, 1959, Peine et rééducation dans l’évolution du droit pénal, congrès des prisons, 16-17 avril.Chazal, Jean, 1953, « Mesures de rééducation et peines devant les tribunaux pour enfants », Revue de sciences criminelles et de droit comparé, n°4, octobre-décembre.Dreyfus, Bruno, 2010, Regard contemporain sur la défense sociale nouvelle de Marc Ancel, L’Harmattan, Paris.Gramatica, Filippo, 1963, Principes de défense sociale, Cujas, Paris.Hedhili-Azema, Hinda, 2014, Sciences et pratiques pénitentiaires en France, xixe et xxe siècles, L’Harmattan, Paris.Hedhili-Azema, Hinda, 2015, « Politiques pénitentiaires et criminologie en Europe », dans MBanzoulou, Paul (dir.), Criminologie et pratiques pénitentiaires, une voie vers la professionnalisation des acteurs, Presses de l’ÉNAP.Hedhili-Azema, Hinda, 2016, « La réforme pénitentiaire impulsée par Paul Amor », L’administration pénitentiaire, 1945, 1975, 2015, Naissance des réformes, problématiques, actualités, dans Kensay, Annie (dir.), Actes de la Journée d’études internationales, Collection Travaux et Documents, Direction de l’administration pénitentiaire.Hedhili-Azema, Hinda, 2016, « L’idée de proportion dans les débats sur la réforme pénale de 1830 », dans Mengès-Le-Pape, Christine (dir.), La justice entre théologie et droit, Colloque international, CTHDIP, Presses de l’université de Toulouse 1.Hedhili-Azema, Hinda, 2016, « Le suivi des personnes placées sous-main de justice dans la réforme Amor », dans Mbanzoulou, Paul (dir.), L’efficacité du suivi des personnes placées sous-main de justice, CIRAP.Hedhili-Azema, Hinda, 2019, Droit disciplinaire pénitentiaire de la Belle Époque, Paris, L’Harmattan.Levasseur, Georges, 1991, « L’influence de Marc Ancel sur la législation répressive française contemporaine », Revue de science criminelle et de droit pénal comparé, n°2, janvier-mars.Rappaport, Emil-Stanislaw, 1956, « En lisant la Défense sociale nouvelle (impression, méditations et conclusions) », Revue de science criminelle et de droit pénal comparé, n°2, avril-juin.Rozes, Simone, 1991, « Le président Marc Ancel et sa carrière judiciaire », Revue de science criminelle et de droit pénal comparé, n°1, janvier-mars.© Cahiers de la sécurité et de la justice - IHEMICahiers de la sécurité et de la justice – Hors-série DOSSIER I 11The doctrinal mutation of the concept of prevention as public criminal policy in the work of Marc AncelHinda HEDHILI-AZEMAenal prevention is a notion both old and new, which has been subjected to a kind of doctrinal instability, due to its historical polysemy. This is to say that its meaning has changed, and can sometimes have contradictory significations in the writings of legal theorists over large periods of time. However, this founding instability isn’t necessarily negative, as it makes prevention into an adjustable notion that evolves depending on society. The preventive intent must therefore be understood in its subtlety in order to ensure efficiency of penal policies. The doubt associated with the concept of prevention certainly derives from the delicate legacy of the 19th century. Indeed its historical definition has been buried so deep, that the original meaning of prevention is hardly ever mentioned in public discourse. 1 And yet, establishing this fundamental meaning would allow us to clarify its multiple layers of interpretation. This definition is meant to be a tool for who wishes to understand the concept of penal sentence in its historical and scientific evolution. Prevention does Marc Ancel is often presented in French law schools as a contemporary criminal law scholar who worked in favor of decriminalization. Internationally, he is still remembered as one of the founders of a modern criminal system that takes into account the human nature of individuals. The promotion of decriminalization is part of the New Social Defense movement, created by Marc Ancel himself in 1954. However, that movement has been misunderstood, despite the clarity of Ancel’s writings. In his books, he defined a new way of dealing with the issue of crime. First, he suggested moving beyond the passive legalistic approach and replacing it with scientific examination in order to better serve the purposes of a judicial trial. He thus criticized the narrow, literalist perspective of prosecution whereby the judge follows the letter of the law. He reinvented the notion of punishment by introducing compulsory character investigations: without ever rejecting the law, Ancel considered that the judge must study the personality of the criminal or offender, and can implement sentences that evolve over time. He mostly focused his efforts on humanizing the criminal justice system which used to be excessively tied to the notions of punishment and retribution. His work is a source of inspiration for a just and humanitarian criminal policy.Keywords: Ancel; Amor; Cannat; social defense; New Social Defense; prevention; criminology; safety measure; criminal doctrine; penal humanitarianism; indeterminate punishment; penitentiary system.PHinda HEDHILI-AZEMALecturer in History of Law and Institutions, member of the Aquitaine Law History Center (CAHD) and the Montesquieu Research Institute (IRM). Specialist in 19th and 20th century criminal doctrine.© Cahiers de la sécurité et de la justice - IHEMI12 I DOSSIER The doctrinal mutation of the concept of prevention as public criminal policy in the work of Marc Ancel – Hinda Hedhili-Azemanot feature in that history as just one of many technical details, but rather as a general principle of the law which is intended to govern the penal system. As simplistic as it may seem at first glance, the historical idea of prevention really falls under the complex and intricate intertwinement of multiple interests and competing powers: judicial, political and administrative. And in the work of Ancel, prevention is precisely what articulates a space for fruitful institutional cooperation. Preventive action is conceived not as a complementary task, but as a true social framework. Sometimes decried as a threat to freedom, sometimes praised as non-penal, prevention does not submit to the traditional norms of legal notions. And with good reason: the principle of prevention is a political response to the critical situations of the times. It is therefore not surprising that most readers well-versed in legal matters might be taken aback by the historical approach, as it designates under the term of prevention not a technical form, but an ideal. If one must solidify a definition of prevention, as a starting point to facilitate its analysis, one could say that prevention is a policy of control of the criminal phenomenon, intended to impede the commission of crime.But in view of its historical usage, prevention is much more than that: it is a key concept that aims to get rid of the retributive notion of the penal sentence seen as punishment. 2 The shift that was made possible by the advent of French preventive thought is not truly perceptible to us now, as most of the readers of Ancel did not witness firsthand the humanitarian revolution in their lifetime; that revolution consisted in the fight against penal excesses. The preventive era was that 20th century moment in the world when the penal sentence yielded to the general principle of human justice. Contemporary observers seem to have forgotten that the penal sentence was one of the last areas of the law to amend its archaic approach. The prevention movement was instrumental in catalyzing this shift. Restricting this movement to a judicial technicality would deprive it of its original, philosophical meaning. However, we must specify here that the following remarks are an attempt to demonstrate that the purpose of French preventive policy was to operate a shift from punishment to penal measure. That aspiration was not isolated, it was part of a global movement to advance progress in the law through a comparative approach of legislation. Preventive action remains a central principle, still expressed by contemporary European doctrine and by most Western governments. Nevertheless if we focus on the French school of thought, prevention takes into account all the measures that allow to solve society’s crime and deviance problems, with the specification that it does not pit the group against the criminal or delinquent individual. Instead, it aims at keeping a form of social peace. In that way, it is a complete social science, always promoting respect for the individual. Contemporary prevention practices would have certainly intrigued many a criminal law scholar of the pre-revolutionary Old Law era, as they move past the bounds of mere punishment, in order to make penal law a polymorphic and pluridisciplinary policy that serves the interests of individuals and social protection. Exercising prevention thus means dealing with the phenomenon of crime in a less aggressive manner. But it hasn’t always been that way in early criminal literature. Needless to say, at the time, prevention was really a lever of political power. 3 The monarch’s majestas (Ancel, 1956 : 936) involved the power to have a say in important cases and to handle them himself as thorny, contentious issues. Prevention is the political power of he who possesses the function of speaking the law, that is to say, the potency of jurisdiction. In a more technical version, in the Old Law regime, prevention was defined as the possibility for an authority to rule on a case. 4 It goes hand in hand with the magistrate’s capability and qualification to claim a case as his own and draw it to himself to judge it fairly. In multiple sources, especially from Roman law and ecclesiastical law, preventing is the moral duty to judge justly, and more broadly, refers to the competence of any judicial authority whether political or institutional. For a long time, the preventive doctrine remained simply the designation of the action of the judge as a defender of equitable criminal justice. Its mutation into a matter of public policy is a recent achievement which, in France, is attributed to the renowned Professor Marc Ancel. 5(1) Nowadays, contemporary legal scholars give penal prevention the simple meaning of prevention of recidivism. That perspective is also shared by the French State (see Les 12 recommandations pour une nouvelle politique contre la récidive [12 recommendations for a new policy against recidivism], February 20th, 2013). But the current treatment of the issue of prevention shall not be dealt with in this historical article. It is simply worth noting that prevention is less and less seen as an element of doctrine, and more and more understood as a technicality or a goal of the post-sentence process. (2) To better understand that idea, we refer you to our upcoming book entitled Doctrine pénitentiaire française et méridionale, to be published.(3) This should be nuanced considering the existence of three different competing justice systems within the kingdom: ecclesiastical, municipal and seigniorial.(4) See the work of Portalis regarding good laws and the nature of laws capable of preventing inequalities.(5) Marc Ancel is internationally recognized, so much so that when he passed away in 1990, a ceremony to pay a solemn tribute to his memory was organized during the 8th United Nations Congress on the Prevention of Crime and the Treatment of Offenders, with representatives of each of the continents that participated in the Congress. Simone Rozes points out his remarkable qualities as a magistrate (Rozes, 1991).© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 13 Cahiers de la sécurité et de la justice – Hors-série In this new configuration, preventing crime no longer means a judge’s designation to hear a case, or a specific action of the law, but rather a scientific program of socialization. Prevention is more than a competence, it becomes a technique to resolve the issue of crime. This is what makes it political. In the work of Professor Ancel, preventing doesn’t aim only at judging, but also at understanding a complex phenomenon. Ancel and his acolytes 6 broke away from the previous century’s vocabulary around crime, and solemnly proclaimed a new conception of social sanctions. It was that system, founded on modern criminal science, which operated the post-war doctrinal shift. 7 Ancel’s postulate, however, was not accepted unanimously (Rappaport, 1956; Dreyfus, 2010 : 17). It was defended by institutions, writings, and personalities who believed in a defense of society that guarantees its preservation rather than its dissolution. The social purpose of Ancel’s work allowed to promote a unique conception of prevention. It aimed to reach out to help those on the fringes of society, and to make the community of citizens more responsible. In that regard, prevention does not intend to make society more lenient, but rather to diversify the modes of sanction and make punitive retribution a marginal form of criminal policy. Hence prevention isn’t opposed to retribution, but is preferable to it and must be chosen over it. The mindset is about observing better, 8 seeing without a veil, without blinding oneself. Society is able to observe, and then to act rather than reacting. It can avoid the pitfalls of penal rigorism: isolation, revenge, xenophobia, stigmatization and punishment. If we were to define the contemporary meaning of “Ancelian” prevention, we would say that it refers to a policy whereby criminal or delinquent individuals, as legal persons, are placed in a safe environment by the human community in order to be resocialized. Social protection is guaranteed as a result, by virtue of a snowball effect. The pragmatism of prevention thus turns rigid punishment dogmas into no more than a distant memory. The “State” of prevention is a new “State” of justice. It informs the criminal individuals that they have breached their obligations and commits them to a non-negotiable journey towards rehabilitation. It is an act of precaution, as it protects both the individual and society. The preventive contract lies at the intersection of justice and security.Qualifying Ancelian social defense as a new criminal policy: beyond crime The work of Marc Ancel is evidence of his visionary perspective and sharp mind. Must society inevitably punish? According to him, the answer is most certainly affirmative. The path opened by Marc Ancel extended over multiple territories and was followed by several authors affiliated to him. In that regard, he was considered by his peers as the founder of post-war modern penal law, for his thought as well as for his influence on penal codification. Born on July 14th, 1902, this magistrate, professor and member of the Academy of Moral and Political Sciences was the instigator of a major doctrinal current. His work reflects a collective movement 9 (Hedhili-Azema, 2016), and to this day, his stance remains significant. At that time, a distinction was operated in the European doctrinal debate between crime prevention, i.e. non-retributive solutions to the criminal phenomenon, and preventive action against delinquency, which was seen as a different issue. Ancel embraced this distinction and cleverly recovered the doctrinal heritage of the past. He adopted the modernist ideas that characterized, as early as the 18th century, the response to crime as a duty of any organized society committed to the respect of the social contract. 10 He reaffirmed the scientific value of the moral reformation carceral studies (Hedhili-Azema, 2014), and purged the earlier positivist theories of their flaws. In his propositions Ancel formulated a policy that was announced as early as the beginning of the Third Republic, through measures that went way beyond the scope of judicial action and the coldness of the law. Entrusting the task of preventing crime to the sole action of penal arithmetic would mean (6) Among them Amor, Cannat, Clément, Charpentier, Gorphe, Heuyer, Herzog and Pinatel. (7) These post-war issues left a strong impression on Marc Ancel, in particular the defense of individual liberties repressed by totalitarian regimes, but also the protection of human dignity and the respect owed to every person. Levasseur states that Ancel clearly began to associate with Paul Amor and Jean Chazal in 1944, especially in devising the reform of the legislation on juvenile delinquency by introducing the ruling of February 2nd, 1945.(8) The earlier theory of social defense, especially the work of Gramatica, hinged on the postulate of the delinquent’s antisociality and adopted a pluridisciplinary approach. Marc Ancel gave social defense a new meaning as he introduced in it the protection of the human person and the guarantee of fundamental rights.(9) Among its champions are Amor, Rolland, Chazal, Cannat, Herzog, Vienne, Heuyer, Roumajon, Colin, Michard, Sélosse and Yvonne Marx.(10) Society has a duty to protect its components but it also has the power to protect its unity through legitimate violence. © Cahiers de la sécurité et de la justice - IHEMI14 I DOSSIER The doctrinal mutation of the concept of prevention as public criminal policy in the work of Marc Ancel – Hinda Hedhili-Azemaamputating criminal policies of a great lever which is to be found outside of the bounds of legal technicality. Because it symbolized his author’s intentions, Marc Ancel’s book, published in 1953, caused a stir upon its release.Indeed his thought adopted quite a singular approach, that of a greater attention given to the respect of human dignity. The idea of a sentence that could cause harm to the person’s physical integrity now seems an old, outdated notion, and is rejected by most current penal theorists. But in that era, Ancel’s stance was all the more unique that in the years prior to the Second World War, there was an increase in forced corporal punishments, the carceral system was assimilated to a slow death sentence for the inmate (Hedhili-Azema, 2019), and the death penalty was widely favored by public opinion. Marc Ancel, who was 52 years old at the time, was then writing under the auspices of the Center of Social Defense Studies of the Comparative Law Institute. His work turned that school upside down. His record as a jurist justified it, and he was a remarkable practitioner on several levels: his political courage as a member of the Resistance, his scientific precision and the quality of his thought. His writing never ceased to describe the steps that needed to be taken in order to divert the doctrine away from retributive violence. Trained both in literature and law, he was well-versed in the matter of the penal arsenal. He was admitted to magistracy in 1929, and as a researcher, his work focused on penal practices. Like all penal law theorists, Ancel had a comparatist background, which led him to always see law as the study of global legislation, performing an investigation across the world’s territory in order to establish a collection of progressive improvements deemed worthy of being spread. Equipped with this practical experience, Ancel argued for the social value of criminal policies. Their goal, from then on, would be to offer humans-citizens the possibility of rehabilitation. Such an approach implies that criminal policies are the comprehension of the criminal phenomenon rather than its punishment. It also proceeds to surgically remove another crucial element from the traditionally understood nature of the sanction, by minimizing its function of intimidation. According to Marc Ancel, it is necessary to dig deep, to probe thoroughly into the human mind, into the personality of the criminal or offender, and reject the harsh, systematic and rationalized retribution promoted by some currents. 11 What is of interest to us in the genesis of Marc Ancel’s penal philosophy is the positive dimension he attributed to it. While it remained within the framework of legality, it placed above penal law a general policy intended to protect social order and the human being. (Ancel, 1954 : 130)The historic symbolism of Ancelian prevention as a tool for diversion programs: beyond sentencingPrevention as a tool for diversion programs against crime must be linked to the very conception of the new social dense. It contributes to the methodological opening movement, but also to the means of reforming the judge’s action. Judging is not sentencing. Ancel presents diversion in the second part of his work published in 1953 as one of the transformation paths of ancient prevention. It is considered a positive path. This process starts with a widening of criminal fields, which infers that active policy cannot limit itself to strictly judicial action. An opening must occur through instruments originating from other criminal sciences: police, biology, psychology, sociology or even medicine. Old disciplines, so-called ancillary to the penal code are raised to the same level as the art of criminal justice. They are tools and sciences worthy of existing besides normative science. For Ancel, penal law ceases to appear as the only regulating path in itself. Ancel never denies the need for penal norm and legal science, he objects to legalism in which judges have been locking themselves up since the Revolution. In this vision, judges are no longer slaves to the law, they are its conscience. Penal law and its representatives will have to abide by a new equity. More precisely, the law, its theories, its notions, its foundations are justifies only if they concur to the social goal defined by citizens: peace, concord, fraternity, freedom. Now, the end of judicial “a-priorism” is not an anti-judicial position, on the contrary, it is a return to real law. The criminalist must have a technical viewpoint, indeed, but also an efficient one. He must not sacrifice his observation to the letter of the law. Ancel therefore advocates a new attitude for the jurist, which he describes as realistic. This path will contribute to a deeper transformation of the meaning of the penalty. For Ancel, It is in fact necessary to consider that penal justice does not necessarily produce an expiation function, revenge or retribution. It is therefore necessary to empty the sanction of its archaic functions and mitigate the punitive function, not leading to the absence of a penal sanction but to the suppression of the idea of castigation. The objective is no longer to punish but to treat delinquency. Mitigating repression does not mean declaring that the criminal is not responsible, as was alleged by his adversaries, but that this responsibility must be sanctioned in a different way. The new strength of criminal policy is to treat the matter in a scientific way: understanding the personal situation of (11) These currents will be described in an upcoming work.© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 15 Cahiers de la sécurité et de la justice – Hors-série the subject, his potential for recovering and his moral and psychological capabilities to contribute to social growth.This new understanding recalls the criteria of ancient penitentiary science (Ancel, 1959). The penal sanction at the same time renews a doctrinal neutrality which aspires to restore the place of the individual in society. Preventive policy therefore aims at generally reorienting political intention towards the social treatment of the person and a general protection of society through the individual treatment. The means allow for judgement cautiousness, by triggering the delinquent personality’s study. The change in tone is the culmination of the new contemporary penal thinking, which recalls the path of the penitentiary doctrine of the July Monarchy. Ancel evokes a functional criterion which determines the mindset of the new sanction science: the realism of the fight against crime and its efficiency with the persons concerned by the measures.The material definition of Ancelian prevention as a form of depenalization: beyond the penalty The material content of prevention will lie in the treatment of the case by the judge with the help of other players of the penal policy: crime experts, criminologists, police, penitentiary and social institutions (Hedhili-Azema, 2015). The first preventive expression is naturally helping the court’s science 12. In this sense, the appearance of the trial is modified. The judge exits from the legal mechanism. He will judge not only the offense but will go beyond the simple objective criterion, to subjective elements linked to the personality of the perpetrator (Gramatica, 1963). It will imply for him to go beyond simply reading the facts and examine individual observations. In the lines of the new defense, precise expertise frameworks appear: the biological constitution, the psychological reaction, the social situation which requires the collaboration of experts. This mutation must not however affect the rights relative to the protection of the accused. Depenalization, furthermore, offers innovations in the statement of the sanction, since there is no longer a distinction between penalties and security measures (Chazal, 1953). Very early on, Ancel evoked the idea that penalty and security measure stop being opposed to each other. The sanction is a new and committed language, which describes a possible criminal evolution and a scalable choice. 13 It will be guided by the efficiency of the moment, intended to cure, educate or neutralize. The judge will have the possibility of progressivity in the sanction and flexibility justified by personal motives, but Ancel rejects the classification of abstract legal sanctions. Finally, no matter what the first qualification of the sanction is, it is the spirit of the sanction which matters, not its judicial denomination. Ancel will naturally advocate using the sanction more freely as it may evolve over a period of time. He will demand that an undetermined sentence be used as leverage to regulate criminal policy. He will reject the illusion of a fixed sentence when the sentence only depends on the evolution of the social investigation. The essence of preventive fight against crime lies in a profoundly social sense which combines the judge’s precaution and that of penitentiary institutions. It is an institutional collaboration. This profound sense reflects the preventive goal. Not to punish but eliminate a-socialization as it brings back the perpetrator to a free life. You could comment Ancel and see here freedom as the novel exercise of citizenship, and conscience as a clearer observation of oneself. Ancel mostly objects to ancient-style penalty, which penalizes and destroys society and men, whose goal was to eliminate the being. Prevention lies in rejecting the erroneous conception of repression as a means to inflict pain to the delinquent or massive destruction on penal grounds. Prevention is undoubtedly this humanistic movement, which advocates for the education of citizens, public opinion and mindset yearning for a better justice in which a man does not decide to destroy another man. There is of course the rejection of penal rigorism proper to traditional societies, but there is still a sort of ideal good. Moderation (Hedhili-Azema, 2016) appears to be Marc Ancel’s federating principle, and here he is in line with an ancient legacy. 14 The preventive intention of the author naturally generates a community of responsibility toward the lost individual. From a political viewpoint, the Ancelian model automatically engenders a universalism which favors the transposition of prevention into foreign legislations. To build this universal model, Ancel must proclaim new rights for society and even for man. If man has failed in his social natural and social rights. In this modern conception of prevention, players strive to make possible a policy devoted to the right to resocialization which was missing in the old system. This certainly has to do with a natural justification in which the right of sanction of society leads to the restoration of (12) The court’s science is the production of: the judicial body, magistrates, lawyers and solicitors in order to comment the law. (13) Voir le IIIe congrès international de défense sociale, Anvers 20-24 avril 1954.(14) At that time, Ancel’s writing still rejected punishment, a concept favored by his peers, as the focal point of the criminal system.© Cahiers de la sécurité et de la justice - IHEMI16 I DOSSIER The doctrinal mutation of the concept of prevention as public criminal policy in the work of Marc Ancel – Hinda Hedhili-Azemasubjective rights: freedom, security, property and work. The Ancelian study respects criminals and delinquents, without ever failing to make them accountable for their actions. This humanization recalls that the social man is free to fall into crime but also free to exit from it. It assigns criminal policy an idea which restores meaning to a flawed technique. Ancel’s criminal policy will be animated by a healthy commitment, an intellectual orientation which leads judicial action towards doing good. The Ancelian ideal is not a utopian penal lack of realism, but indeed non-rigorism. This ideal requires a subtle understanding that we have tried to offer the reader. It seems necessary to mention that the Ancelian study has sometimes been misinterpreted and wrongly reduced to penal laxity. At a time when nations are tempted by withdrawal from the rest of the world, nourished by a feeling of insecurity, the Ancelian thought can enable decision-makers to have a model against authoritarian downward spirals and rampant populist movements. In this regard, we can consider Marc Ancel as the instigator of a fundamental body of work. As his thoughts were not, in his days, mainstream or widely accepted, it seemed necessary for him to write them down to leave a trace of his doctrinal intention. It seems today necessary to restore them in a clear way in order to enrich the debate about current criminal policies nReferencesAncel, Marc, 1954, La Défense sociale nouvelle, un mouvement de politique criminelle humaniste, Cujas, Paris.Ancel, Marc, 1956, « Politique criminelle et psychologie judiciaire dans la détermination de la sanction pénale », Revue de science criminelle et de droit pénal comparé, n°2, April-June.Ancel, Marc, 1959, Peine et rééducation dans l’évolution du droit pénal, congrès des prisons, April 16th-17th.Chazal, Jean, 1953, « Mesures de rééducation et peines devant les tribunaux pour enfants », Revue de sciences criminelles et de droit comparé, n°4, October-December.Dreyfus, Bruno, 2010, Regard contemporain sur la défense sociale nouvelle de Marc Ancel, L’Harmattan, Paris.Gramatica, Filippo, 1963, Principes de défense sociale, Cujas, Paris.Hedhili-Azema, Hinda, 2014, Sciences et pratiques pénitentiaires en France, xixe et xxe siècles, L’Harmattan, Paris.Hedhili-Azema, Hinda, 2015, « Politiques pénitentiaires et criminologie en Europe », in MBanzoulou, Paul (Ed.), Criminologie et pratiques pénitentiaires, une voie vers la professionnalisation des acteurs, Presses de l’ÉNAP.Hedhili-Azema, Hinda, 2016, « La réforme pénitentiaire impulsée par Paul Amor », L’administration pénitentiaire, 1945, 1975, 2015, Naissance des réformes, problématiques, actualités, in Kensay, Annie (Ed.), Actes de la Journée d’études internationales, Collection Travaux et Documents, Direction de l’administration pénitentiaire.Hedhili-Azema, Hinda, 2016, « L’idée de proportion dans les débats sur la réforme pénale de 1830 », in Mengès-Le-Pape, Christine (Ed.), La justice entre théologie et droit, International Conference, CTHDIP, Presses de l’Université de Toulouse 1.Hedhili-Azema, Hinda, 2016, « Le suivi des personnes placées sous-main de justice dans la réforme Amor », in Mbanzoulou, Paul (Ed.), L’efficacité du suivi des personnes placées sous main de justice, CIRAP.Hedhili-Azema, Hinda, 2019, Droit disciplinaire pénitentiaire de la Belle Époque, Paris, L’Harmattan.Levasseur, Georges, 1991, « L’influence de Marc Ancel sur la législation répressive française contemporaine », Revue de science criminelle et de droit pénal comparé, n°2, January-March.Rappaport, Emil-Stanislaw, 1956, « En lisant la Défense sociale nouvelle (impression, méditations et conclusions) », Revue de science criminelle et de droit pénal comparé, n°2, April-June.Rozes, Simone, 1991, « Le président Marc Ancel et sa carrière judiciaire », Revue de science criminelle et de droit pénal comparé, n°1, January-March.© Cahiers de la sécurité et de la justice - IHEMICahiers de la sécurité et de la justice – Hors-série DOSSIER I 17Le Programme de parrainage de désistance (P.P.D.) : une mise en lien concrète française de la justice restaurative et de la désistanceErwan DIEU, Laurent MERCHATContexte du Programme de parrainage de désistance (P.P.D.)a puissance publique a pour mission de maintenir l’ordre et de prévenir son trouble, notamment les atteintes à la sécurité des personnes et des biens. Depuis près de 20 ans, au cœur des préoccupations, avec plus ou moins d’acuité médiatique, la délinquance n’a cessé d’interroger quant à l’efficience de la réponse apportée. Les informations statistiques 1 du Ministère de la Justice ont mis en évidence que l’infraction est trop largement commise par une personne ayant déjà été condamnée. Si les études chiffrées notent une variabilité selon la nature de l’infraction et selon le critère retenu (récidive légale, réitération, antécédent judiciaire) force est de constater le faible pourcentage de primo-délinquants (moins de 38% de la population pénale 2). La prévention de la récidive constitue depuis 2008 la mission principale des Services Pénitentiaires d’Insertion et de Probation (SPIP), services déconcentrés Depuis la loi du 15 août 2014 et la circulaire de mars 2017, la justice restaurative est une mesure qui a vocation à s’appliquer concrètement sur le terrain correctionnel français. Toutefois, les dispositifs restauratifs habituels (p. ex. médiation, groupe, rencontre) ne sont pas toujours mobilisables dans les conditions actuelles et exigent des adaptations. Le Programme du parrainage de désistance (P.P.D.) est une innovation du Service Pénitentiaire d’Insertion et de Probation (SPIP) de la Drôme (France) avec le soutien du Service de Criminologie ARCA. Il constitue une déclinaison des Cercles de Soutien et de Responsabilité (C.S.R.). Il ouvre la discussion sur le lien « Restauratif-Désistance-Réhabilitatif ». Ce programme original présenté aux journées internationales du good lives model (GLM) (Tournai, Belgique) sera ici exposé, discuté au regard des modèles criminologiques et des effets observés sur les participants.Mots clés : désistance ; justice restaurative ; prévention de la récidive ; Programme de parrainage de désistance (P.P.D.) ; réhabilitationLErwan DIEUCriminologue, praticien accrédité EMDR-Europe et formé officiellement à la supervision EMDR, docteur en psychologie et directeur général du Service de Criminologie ARCA. Il est référent de Justice restaurative auprès du Ministère de la Justice, membre du board international du « Good Lives Model », fondateur du modèle de l’identité temporelle TIM-E, implanté dans les services correctionnels de différents pays sur diverses problématiques et dans la prise-en-charge du psychotraumatisme. Il est coordonnateur du programme européen 3C2D sur le traitement des problématiques de radicalisation.Laurent MERCHATDirecteur pénitentiaire d’insertion et de probation (DPIP) au Service Pénitentiaire d’Insertion et de Probation (SPIP) de Roanne et anciennement Conseiller (CPIP) de la Drôme (France). Formé à la justice restaurative dès 2015, s’inspirant des cercles de soutien et de responsabilité, il propose avec Doriane Serrières, directrice, un programme susceptible de répondre aux besoins criminogènes d’un plus grand nombre : « le parrainage de désistance ».(1) Infostat Justice n°50, 68, 88, 108, et 1227.(2) Infostat Justice n° 127, graphique 1 : 63% des condamnés ont des antécédents judiciaires dans les 8 années précédentes - Champ exploré : condamnés pour crimes, délits et C5 entre 2004 et 2011.© Cahiers de la sécurité et de la justice - IHEMI18 I DOSSIER Le Programme de parrainage de désistance (P.P.D.) : une mise en lien concrète française de la justice restaurative et de la désistance – Erwan Dieu, Laurent Merchatde l’administration pénitentiaire. En 2012, elle est portée sur le devant de la scène politique et médiatique comme « un enjeu majeur pour la société et la justice » au cours d’une « conférence de consensus sur l’efficacité des réponses pénales afin de mieux prévenir la récidive ». La loi du 15 août 2014 relative à l’individualisation des peines et renforçant l’efficacité des sanctions pénales a introduit la justice restaurative (J.R.) dans le Code de procédure pénale en son article 10-1 (Zehr, 1990). La circulaire d’application du 15 mars 2017 définit la J.R. « comme un modèle de justice complémentaire du procès pénal, qui consiste à restaurer le lien social endommagé par l’infraction, à travers la mise en œuvre de différentes mesures associant la victime, l’auteur et la société ».Au sein des mesures pénales de justice restaurative, les Cercles de soutien et de responsabilité (C.S.R.) sont les seuls programmes à établir un consensus scientifique international ayant un impact direct sur la diminution du risque de récidive (Wilson et al., 2010). Au sein des C.S.R., un auteur retournant à la vie libre et présentant un haut risque de récidive est entouré par des membres de la communauté. Une fois par semaine pendant un an, le membre principal, les bénévoles et l’animateur se réunissent et font cercle. Ce dernier modèle peut être considéré comme dérogatoire à la J.R. puisque ne faisant pas intervenir les victimes. Fort de ce constat, en 2016, le SPIP de la Drôme, réunissant plusieurs personnels formés à la J.R., et considérant sa mission principale de prévention de la récidive, choisissait d’investir plus particulièrement les C.S.R. Actuellement, ce service assure sur son territoire un volet de suivis chiffré aux alentours de 2 000 personnes condamnées, qu’elles soient physiquement hébergées au Centre Pénitentiaire (environ 450) ou placées sous main de justice en milieu dit ouvert (plus de 1 500). La mise en place d’un tel dispositif suppose beaucoup de temps professionnel. Or, la charge de travail individuel étant déjà conséquente, comment concentrer autant d’énergie pour quelques-uns, si « dangereux » soient-ils ? Ainsi, sans exclure la mise en place de C.S.R., la direction du SPIP inscrivait son service dans une démarche inspiré de la J.R. dénommé « le parrainage de désistance ».Le Programme du parrainage de désistance (P.P.D. ou VoD pour volunteers of desistance) a débuté en décembre 2016 au SPIP Drôme-Ardèche jusqu’à la restitution de son évaluation auprès des participants en mars 2018. Il a été évalué en cours de programme par le Service de recherche en criminologie appliquée ARCA 3 sur demande du SPIP. Le parrainage de désistance (Dieu et Merchat, 2018) est un programme qui met en place les conditions pour l’instauration de relations sociales, soutenantes, sincères et authentiques, entre des personnes bénévoles membres de la communauté et des personnes placées sous main de justice (PPSMJ). La rencontre est fondée sur la liberté et la responsabilité de chacun, c’est à dire, sur le volontariat des participants, en l’absence de contrepartie. Tous les participants se réunissent une fois par mois et font cercle. Il est alors possible que filleul et parrain se choisissent dans une relation de parrainage dont ils déterminent la durée (4, 6 ou 8 mois). À l’heure actuelle, s’agissant d’un programme à entrée et sortie constante, on dénombre environ 14 filleuls et 14 bénévoles engagés actifs dans le dispositif (présents à la réunion mensuelle 7 fois sur 10 pour les premiers et 8 fois sur 10 pour les seconds). En juin 2018, après 18 mois d’existence, 25 condamnés se sont impliqués dans le programme, 20 contrats de parrainage ont été signés.Théorisation du programme : un retour au contrat social par la criminologie positiveL’infraction est une atteinte portée au contrat social. Il convient donc non seulement de la sanctionner mais aussi et surtout de faire en sorte que le fait infractionnel ne se reproduise plus. Or, force est de constater que lorsqu’une infraction est commise, le procès pénal ne suffit pas pour extraire définitivement le risque infractionnel du tissu social, et, par voie de conséquence, pour rétablir confiance dans le lien social. Le procès permet de reconnaître un auteur dans sa responsabilité et une victime dans son préjudice. La Justice octroie au premier une peine et à la seconde des dommages et intérêts, à charge pour chacun d’eux d’assumer ensuite les répercussions que l’infraction a fait naître dans leur vie. Le programme propose une déclinaison des principes et valeurs de la justice restaurative (Zehr, 2002) pour répondre à ce constat. Il est original car il s’affirme comme un dispositif susceptible de soutenir un grand nombre de personnes confiées par mandat judiciaire, quelle que soit la gravité des infractions commises. Nombre d’entre elles souffrent, en effet, dans leur processus de changement pour une sortie de la délinquance, d’isolement - car n’ayant pas toujours les points d’appui nécessaires - et de difficultés à s’inscrire dans la construction et la réalisation de leur projet – car ne disposant pas toujours des habiletés ou de la position de responsabilité requis pour un parcours d’insertion.(3) Association créée en 2008 dont le service est conventionné par le Service de l’accès au droit et à la justice et de l’aide aux victimes (SADJAV) du ministère de la Justice concernant la justice restaurative et reconnue dans la circulaire de mars 2017 comme Service français de formation officielle de justice restaurative.© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 19 Cahiers de la sécurité et de la justice – Hors-série Le Programme de parrainage de désistance, par son nom même et ses objectifs, revendique des liens avec les théories de la désistance, mais aussi avec le good lives model (GLM) et le modèle de l’identité temporelle (TIM-E) (Dieu, 2018a). Voici les théories de la désistance pouvant être en lien avec l’approche choisie du P.P.D. :– Selon Laub et Sampson (1993, 2001), le processus de désistance est à considérer en fonction des « points tournants » (p. ex. mariage, parentalité) qui disposent l’individu dans une trajectoire menant in fine à l’arrêt des comportements d’infraction.– L’analyse phénoménologique de Maruna (2001) propose une lecture psychologique subjective du processus de désistance en deux strates. La désistance primaire caractérise la rencontre du sujet avec une période sans infraction, tandis que la désistance secondaire concerne le réaménagement identitaire du sujet via les rôles sociaux qu’il investit.– La transformation cognitive (Giordano et al., 2002) envisage que les sujets puissent entrer dans une reconsidération de leurs convictions permettant de mettre à distance leurs comportements délictueux, tandis que la transformation affective (Giordano et al., 2007) suppose l’apprentissage d’émotions au contact d’interactions positives.– Paternoster et al., (2015) prennent appui sur la théorie du choix rationnel selon laquelle un délinquant passe à l’acte à la suite d’un calcul coût / bénéfice qui lui semble propice à l’infraction. Une trajectoire ponctuée de plusieurs infractions bascule le calcul global dans le négatif et suscite chez le sujet une « peur de Soi » (self-fear), de ce qu’il est devenu via cette trajectoire (Paternoster et Bushway, 2009).Le P.P.D. entretient aussi des liens étroits avec le good lives model (Ward et Brown, 2004), ainsi que d’autres approches océaniennes (p. ex. le multifactor offender readiness model de Ward et al., 2004). Le good lives model, comme le PDD, est un modèle attentif au bien-être du sujet et à son développement personnel qui vise la personne dans sa globalité. À travers le P.P.D., le sujet traverse un processus d’acquisition de capacités internes et externes pour envisager une vie davantage prosociale et épanouissante. Le modèle de l’identité temporelle TIM-E (Dieu, 2019) est un modèle de compréhension criminologique orienté sur les perspectives futures des auteurs d’infraction. L’objectif de l’accompagnement se focalise sur les ressentis temporels vécus par les sujets et les répercussions sur l’identité, entre sentiment d’unicité et mouvements perpétuels. L’individu ressentira la « présence » de son identité désistante, initiant et facilitant dans le temps son processus de renoncement à l’infraction. Si l’intervention professionnelle ne peut prétendre avoir prise sur « l’âge, au sens de maturité », ou sur « les événements positifs de la vie (ex. : rencontre amoureuse, naissance d’un enfant) », il en est tout autre en ce qui concerne les deux autres facteurs de désistance 4 : « le renforcement du capital humain (ex. : capacités de communication ou de gestion des émotions) » et « le développement du capital social (ex. : intégration dans des relations et réseaux sociaux non délinquants ou développement de compétences personnelles et sociales, insertion professionnelle) ».Mise en application du Programme de parrainage de désistance (P.P.D.)Le dispositif P.P.D. : un Cercle de soutien et de responsabilité entièrement réadaptéLe dispositif vise le renforcement du capital humain et le développement du capital social. Au terme de l’évaluation, ces objectifs se confirment non seulement comme des pré-conditions au changement vers la non-récidive à terme mais aussi comme vecteurs d’insertion (travail, formation, mise en place des droits sociaux…), et ce, sans que cette insertion ne soit portée dans ses outils comme objectif. Selon le principe de réceptivité (Bonta et Andrews, 2007), un programme efficient doit être adapté aux capacités de la personne. Le P.P.D. est un remaniement des C.S.R., tout en s’appuyant sur l’évolution des méthodologies d’intervention des SPIP (Lefebvre et al., 2018). Il témoigne d’une appropriation professionnelle de l’évolution de la mission de l’Administration Pénitentiaire.Voici les points communs et les différences entre le P.P.D. et le C.S.R. (Dieu, 2018a) :Les points communs entre C.S.R. et P.P.D. :– Personne en situation de fragilité, parcours infractionnel marqué.(4) Conférence de consensus p. 122 : FICHE 4 Les facteurs de risque, de protection et de désistance© Cahiers de la sécurité et de la justice - IHEMI20 I DOSSIER Le Programme de parrainage de désistance (P.P.D.) : une mise en lien concrète française de la justice restaurative et de la désistance – Erwan Dieu, Laurent Merchat– Volontariat de la part du bénévole et du bénéficiaire (« membre principal »).– Formation des bénévoles aux principes des C.S.R., de la justice restaurative, de la motivation au changement et de la désistance et de la prévention de la récidive. Réunion en dehors des locaux professionnels ainsi que des lieux intimes des personnes (lieux publics). Grande disponibilité offerte à la PPSMJ, rencontres possibles en plus des bases initiales du contrat (échange hebdomadaire).– SPIP dans certains cas de C.S.R. en France.Les spécificités du Programme de parrainage de désistance (P.P.D.)– Tout type d’infraction, aucune évaluation du risque de récidive.– 4, 6, 8 mois, renouvelable, devoir d’assister aux réunions mensuelles même sans parrainage. 1 bénévole « parrain » pour 1 membre principal « filleul », ouvrant sur un soutien individuel au quotidien. – Rencontre mensuelle entre l’ensemble des bénévoles (parrains ou non) et des PPSMJ (filleuls ou non), dans une salle extérieure au SPIP. Les deux conseiller pénitentiaire d’insertion et de probation (CPIP) coordonnent le cadre du groupe, notamment chaque mois.– Animation directe par les CPIP du SPIP.Intérêts du P.P.D. vis-à-vis du C.S.R. (pour le SPIP et la PPSMJ)– Flux constant dans le recrutement des PPSMJ, avec une diversité de profils dans les échanges et une ouverture à plusieurs types de partenaires.– Sécurité autour de la personne, ouverture au groupe même sans signature (Cercle secondaire à réunion mensuelle).– Liberté dans le choix mutuel entre parrain et filleul, l’affinité devenant le meilleur suivi au quotidien.– S’inscrit dans une optique de désistance, reliant la question de l’arrêt des activités infractionnelles au maintien des actions prosociales et de la réinsertion dans la société. – Deux CPIP du SPIP, avec évaluation externe du programme.Mise en action du Programme de parrainage de désistance (P.P.D.)Le P.P.D. se présente aux PPSMJ à l’aide d’un prospectus intitulé Adopter un parrain ou une marraine de désistance. Les étapes du changement telles que simplifiées permettent d’emblée une évaluation collaborative claire et explicite. Étape 1 : « La motivation : je veux changer mais je me sens encore empêtré dans des comportements qui me freinent ». Étape 2 : « Le plan d’action : j’ai décidé de changer et j’essaye d’agir en conséquence ». Étape 3 : « La consolidation : j’ai changé mais je dois rester vigilant car j’ai encore des points de fragilité ». Ces trois étapes s’inscrivent dans la notion de réceptivité spécifique (Bonta et Andrews, 2007) et la théorie de la disposition au changement (Prochaska et Norcross, 2010) entre un stade antérieur dit de pré-contemplation – dans lequel le désir de changement n’apparaît pas vraiment comme point d’appui – et un stade postérieur marquant la sortie définitive de la délinquance – avec inscription satisfaisante dans le contrat social. Par le terme « adopter », la personne est d’emblée invitée à une participation active : la relation n’est pas à sens unique, choisir un parrain c’est prendre sa part de responsabilité dans la relation, c’est se soucier de l’autre dans la réciprocité. Après un premier engagement oral de confidentialité auprès du CPIP animateur, la PPSMJ entre dans le cercle avec un badge blanc « prise de contact ». La réunion mensuelle a lieu en extérieur le premier samedi matin de chaque mois. Bénévoles et PPSMJ se réunissent alors en un grand cercle. Chaque participant est solidaire du groupe, de son fonctionnement, et veille à ce que chacun puisse s’exprimer dans le respect de règles de communication du P.P.D. S’en suit un entretien avec le CPIP animateur au cours duquel la charte du parrainage de désistance qui décrit la place et l’engagement de chacun est lue et explicitée. La personne peut alors faire le choix de s’engager ou non dans le programme, tout en sachant que sa participation n’aura pas d’incidence sur le traitement judiciaire futur. Si elle signe avec la direction du SPIP la convention d’engagement réciproque elle reviendra dans le cercle muni d’un badge indiquant son prénom et sa qualité de filleul. À une exception près, toutes les PPSMJ qui ont pris contact avec le dispositif se sont engagés. Les badges colorés (selon la contractualisation) constituent le signe visuel que les participants sont alors tous reliés dans le dispositif, chacun ayant sa place. Lorsque la rencontre interpersonnelle amène au choix mutuel de s’engager dans un lien de parrainage, un contrat est signé dans le groupe pour une durée renouvelable de 4, 6 ou 8 mois. Chacun témoignera dans le cercle de cette relation, car tout ce qui est vécu dans cette relation peut être dit dans le groupe, même si tout n’a pas besoin d’être partagé.© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 21 Cahiers de la sécurité et de la justice – Hors-série Population accueillie : profils des PPSMJ « filleuls de désistance » et des Bénévoles « parrains de désistance »L’analyse des casiers judiciaires des 18 PPSMJ engagés dans le dispositif au moment de l’évaluation témoigne en effet de parcours délinquant enraciné. Plus de 50% de l’effectif a été condamné au moins 5 fois. On remarque que 30% de l’effectif a été condamné plus de 10 fois (jusqu’à 29 condamnations), tandis que seulement 1 filleul sur 10 peut être qualifié de primo-délinquant. 70% des filleuls ont déjà été écroués. 50% ont déjà été incarcérés et 50% ont connu ou connaissent l’aménagement de peine. Seulement 20% ont bénéficié d’aménagement de peine sans passer par « la case prison ». La moitié de l’effectif est écroué pendant sa participation au programme (6 en placement sous surveillance électronique, 1 en placement extérieur) ou sort tout juste de prison (2 sortants depuis moins d’un mois). Le nombre de condamnations est lié à la durée de la période infractionnelle qui va de 1 à 19 ans, avec une moyenne entre 5 et 6 ans. Vingt-trois types d’infraction ont été commises par l’effectif de filleuls, majoritairement des vols, violences et infractions routières. On remarque que sur les 45% de l’effectif responsables de vols, les trois quarts ont aussi commis des violences, que les 61% des condamnés pour violence l’ont aussi été pour infraction à la législation des stupéfiants ou conduite en état d’ivresse. 39% des filleuls ont commis au moins une infraction routière (aucun de ces derniers n’a été condamné que pour infraction routière). Les deux personnes ayant été condamnés pour homicide involontaire sont concernés par la consommation de stupéfiant, tandis que celui qui a été condamné pour agression sexuelle est impliqué pour conduite en état d’ivresse. Les condamnés engagés dans le dispositif sont des hommes (une seule exception). Les deux tiers ont moins de 30 ans, avec un écart d’âge entre 21 et 58 ans. Près de la moitié des filleuls vivent auprès de leur famille, il s’agit majoritairement des moins de 30 ans.Les bénévoles impliqués dans le dispositif sont majoritairement des hommes. L’âge moyen est de 50 ans, avec une répartition des âges entre 26 et 78 ans. Ils sont issus de toutes catégories socioprofessionnelles ce qui contribue à la représentativité de la société. Les deux tiers sont en position d’activité (dont 2 en recherche d’emploi), ce qui permet un modèle plus rapproché sur les exigences liées à l’emploi. Les deux tiers vivent en situation maritale. La moitié d’entre eux témoignent d’engagement associatif par ailleurs. Soulignons qu’un parrain peut avoir été condamné, avoir été suivi par la justice, autrement dit être sorti de la délinquance. Dès lors, on ne se soucie pas de savoir qui est concerné. Et lorsqu’un filleul termine son suivi judiciaire après un beau parcours partagé et lorsqu’il change de place pour revêtir la carte parrain, personne ne s’étonne. Le dispositif n’a pas vocation à transformer les filleuls en parrains, mais cela est possible, preuve que la délinquance est un chemin de traverse, qu’elle peut être circonscrite à une période de vie, qu’elle peut être dépassée.Retours concernant le P.P.D. : des résultats encourageants en termes d’engagement et de satisfactionEngagement, attentes et satisfactions liées au P.P.D.Le programme a été évalué sur sa période d’existence de 15 mois, de novembre 2016 à février 2018, date de l’évaluation (Dieu et al., 2020b). Au début du P.P.D., en décembre 2016, 2 PPSMJ étaient engagées pour 12 participants, pour atteindre 13 PPSMJ engagées et 23 participants sept mois plus tard en juillet 2017 et se stabiliser à 10 PPSMJ engagées et 23 participants jusqu’à l’évaluation du programme en février 2018. 12 PPSMJ sont en moyenne engagées pour 19 participants à la suite de la stabilisation du programme en avril 2017. Le nombre moyen de participants aux réunions mensuelles est 19, avec un taux de présence de 70% des filleuls sur la période d’engagement et 80% des parrains. Les contrats de parrainage sont variés, 7 PPSMJ se sont engagées 4 mois, 6 PPSMJ durant 6 mois et 3 PPSMJ durant 8 mois. Notons que trois contrats furent renouvelés 6 mois, dont deux qui étaient des contrats initiaux de 8 mois et portant le total à 14 mois de programme. En termes de motivation au changement, une évolution des stades de disposition au changement de Prochaska et Norcross (2010) fut interrogée par les filleuls et parrains puis reprise par le CPIP référent du programme, et ce tout au long du P.P.D. Si l’on exclut les deux entrées récentes dans le dispositif et une succincte, sur les 15 filleuls dont la mesure était possible, il fut constaté une large évolution de la disposition des PPSMJ engagées dans le programme. Au départ, seulement deux semblaient déjà en action pour 10/15 en fin de parcours. Une seule PPSMJ sur 15 termine son parrainage avec une disposition de changement relativement faible.Plus que la disposition au changement, la notion d’engagement et de disposition au traitement (Ward et al., 2004) est aujourd’hui reconnue comme un élément probant sur la récidive à prendre en considération au sein des programmes réhabilitatifs (McMurran et McCulloch, © Cahiers de la sécurité et de la justice - IHEMI22 I DOSSIER Le Programme de parrainage de désistance (P.P.D.) : une mise en lien concrète française de la justice restaurative et de la désistance – Erwan Dieu, Laurent Merchat2007 ; McMurran et Ward, 2010). À l’entrée du programme selon la méthode d’analyse des motivations et satisfactions restauratives (Dieu, 2018b ; Dieu et al., 2020a), les sujets avaient des attentes plus ou moins fortes, variant en intensité de 54% à 74% pour une moyenne de 54% d’intensité maximale. Ce résultat est donc éloigné de l’intensité des satisfactions restauratives retirées du programme, avec une moyenne de 76% pour une variation allant de 61% à 89%. Les filleuls ont indiqué voir leurs attentes comblées, qu’ils « recommanderai[ent] ce programme à d’autres PPSMJ dans une situation semblable ». Ils ont pu exprimer les conséquences positives du programme, qu’il s’agissait d’un « bon programme », qu’il « favorise les rencontres » et « aide à la réinsertion ». Ce que les parrains ont souligné sous l’angle interne du processus, c’est un programme qui « favorise les échanges et la confiance ». Les satisfactions cognitives et émotionnelles évoquent « apaisement », investissant grandement la « reprise du contrôle de sa vie, meilleure estime et confiance en soi », un « regard différent sur le parcours » en lien avec un « évitement récidive » largement rapporté dans les discours.Une évolution des stades de changement observables depuis les narrations rapportées des Plans de vieSur l’effectif de filleuls préalablement décrit, lors de leur entrée dans le programme : 33% sont identifiés à l’étape 1 du changement et 54% à l’étape 2. Donc, seulement 13% ont déjà parcouru les étapes 1 et 2. Il est possible de remarquer que ce stade initial semble ne pas être corrélé avec la situation pénale préalable. Le fait d’avoir été ou non emprisonné ne présage pas du niveau de désistance, et on constate que l’aménagement de peine dans le parcours ne précise pas non plus le niveau de changement (écroué avec ou sans aménagement de peine : 1/3 au niveau 1, et 2/3 au niveau 2). À la fin du programme, entre 0% et 6% de l’effectif se situe au stade 1, entre 20 et 26% au stade 2, et entre 67 et 80% au stade 3. Si 87% de l’effectif présentait des comportements à risque infractionnel en début de programme, ce n’est le cas qu’entre 20 et 32% de l’effectif au moment de l’évaluation. Cela représente une sortie du risque infractionnel pour 62% à 77% des PPSMJ concernées. On constate que 81% de l’effectif a connu au cours du programme une variation significative de son insertion en termes d’emploi, d’activité ou de projet mis en acte. On notera que cette évolution est particulièrement forte pour 31% de l’effectif.En termes d’évolution de trajectoire, nous observons un inversement des priorités de vie, qui démontre une volonté d’insertion sociale et professionnelle plus appuyée. La sphère relationnelle s’aménage de manière plus cohérente au regard des changements exprimés et souhaités par les PPSMJ mais également dans le cadre du contrat de parrainage et du cadre institutionnel du SPIP. Les souffrances, si elles sont toujours présentes, émotionnellement négatives et intensément ressenties, sont moins présentes au quotidien. Les personnes y accordent moins de temps, s’y exposent moins, d’autres aspects de la vie ayant été priorisés. Nous émettons l’hypothèse d’une meilleure régulation cognitive des éléments de vie suscitant des émotions négatives. Les personnes, tant dans leur vision du présent que dans leur projection du futur, semblent concrètes et réalistes. Il ne s’agit pas d’un plan de vie fantasmé, mais d’une perspective subjective désirée construite à partir des changements en cours dans le présent. Elles sont dans une réflexion non pas de changement radical mais bien de réaménagement de leur quotidien, en vue semblerait-il d’une amélioration de leur qualité de vie (qualité des relations, qualité des investissements sociaux) qui tendrait à construire un bonheur accessible. Les PPSMJ apparaissent compétentes à établir les lignes du plan de vie qu’elles estiment adapté et à discriminer les apports positifs et négatifs de ce qu’elles investissent. Ceci est d’ailleurs tout à fait nouveau puisqu’elles évoquaient leur quotidien dans un ensemble auparavant neutre.Retours concernant le P.P.D. : des résultats exploratoires encourageant dans une cadre de criminologie positiveLe P.P.D. favoriserait le rôle central de l’agency 5 dans l’émergence et le renforcement de certains facteurs de protection et le lien avec les besoins primaires du good lives modelLes facteurs de protection s’envisagent comme des outils essentiels dans l’accompagnement du processus de désistance (Vries Robbé et al., 2015). Ils peuvent être définis d’un point de vue générique comme toute caractéristique (5) L’agency, ou l’agentivité, est un concept qui renvoie à la manière dont les personnes se considèrent comme décisionnaires de leurs actions dans un environnement donné. Il serait observé chez les personnes désistantes une narration évoquant un état d’agentivité dans leur vie, de prise de contrôle, à la différence des personnes récidivistes.© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 23 Cahiers de la sécurité et de la justice – Hors-série propre à une personne, à son milieu ou à sa situation qui « réduit le risque de récidive violente » (Vogel et al., 2011 : 172). Trois modèles coexistent concernant les facteurs de protection (Ward, 2017 ; Vries Robbé et Willis, 2016 ; Serin, Chadwick, Lloyd, 2016 ; MacDonald, 2016) et ils se déclinent en trois sous-types et sont définis par des éléments bien précis : facteurs internes, motivationnels et externes (Guay et Vries Robbé, 2017).Facteurs internes de protection :– meilleure habileté d’adaptation et un développement de la théorie de l’esprit (plus que de l’empathie en soi) qui semblent émerger à travers la qualité de la relation qui se construit dans les liens établis entre les filleuls et parrains. – emphase du facteur de maîtrise de soi avec une amélioration de la régulation cognitive des émotions et une meilleure maîtrise de la violence. En prodiguant des conseils inspirés de leur vie sur la gestion des situations-problèmes et des éléments de rechute, les parrains (et autres filleuls) deviennent des miroirs et des modèles pour la PPSMJ. Ils soutiennent la conscientisation des limites des moyens de réponse et des alternatives cognitivo-comportementales et émotionnelles possibles. Facteurs motivationnels de protection :– la motivation au traitement apparaît dans la mise au travail de certains, notamment dans l’investissement d’un soin psychologique, mais également au regard du traitement des addictions, visible dans la diminution de prise de substances. – l’investissement de la sphère professionnelle (travail) est mentionné comme un levier mais également comme une finalité par les filleuls et les parrains.– les activités de loisirs sont citées au travers des voyages effectués ou envisagés et une meilleure structuration des activités sociales du quotidien.– les échanges entre les filleuls et le soutien des parrains ont un effet sur la situation financière et la gestion financière des PPSMJ qui se voit impactée tant objectivement (p. ex. maintien du travail) que subjectivement (p. ex. valeurs dégagées et convictions entourant l’argent). – l’attitude envers l’autorité semble avoir été modifiée.– les objectifs de vie des PPSMJ sont impactés, avec des visées prosociales.– la médication est visible dans les changements et actions de changement en lien avec l’évolution des Plans de vie concernant le traitement général de la prise de substances.Facteurs externes de protection :– le réseau social a évolué, les amis sont plus investis temporellement et certaines personnes évoquent un éloignement des éléments nocifs pour eux. – la famille (ascendante et descendante) voit une évolution importante dans un investissement plus positif des personnes qui leur accordent plus de temps.– les soins professionnels et le contrôle externe se recoupent au sein du programme, puisque les parrains au quotidien (et les autres filleuls et CPIP référents chaque mois) sont autant des éléments essentiels du contrôle social que des facilitateurs de soins professionnels. L’attention portée sur des valeurs prosociales partagées visant la non récidive et l’insertion légitiment le contrôle social formel et informel et l’évitement de l’isolement et l’exclusion (Wilson, 2007). – l’importance donnée au couple et au conjoint suppose une amélioration et une modification des relations intimes et en partie des conditions de vie.De manière générale, indiquons que plus le cercle est solide plus le travail mené semble profond et orienté à la fois vers les besoins personnels du membre principal (logement, travail, problèmes administratifs) et l’apprentissage de comportements positifs par le biais du modelage social. La relation à l’autre ici en œuvre est reconnue dans la littérature scientifique comme un élément majeur du champ réhabilitatif (Estroff et al., 1994 ; Grubin, 1997 ; Bonta et Andrews, 2007). Le groupe fait bénéficier ses membres d’une large ouverture, d’un espace sans jugement facilitant la confiance en soi. Les conseils et échanges autour de situations concrètes amenant des alternatives comportementales et cognitives favorisent l’expérimentation des solutions proposées dans la vie quotidienne. Le groupe réinterrogera (et le parrain au quotidien) l’individu sur sa situation et la mise en application ou non des conseils prodigués.Nous retrouvons l’hypothèse de Ward (2017) selon lequel il faudrait dissocier deux classes de facteurs de protection : les capacités d’agency et les facteurs contextuels. Le programme est une mise à disposition d’un ensemble de facteurs contextuels positifs internes (dans le cercle) et externes (en dehors du cercle) à l’égard du sujet. Les capacités d’agency se voient investies et stimulées par cet © Cahiers de la sécurité et de la justice - IHEMI24 I DOSSIER Le Programme de parrainage de désistance (P.P.D.) : une mise en lien concrète française de la justice restaurative et de la désistance – Erwan Dieu, Laurent Merchatensemble de facteurs contextuels. Si nous ne pouvons ici prendre en considération les caractéristiques personnelles des sujets ou soutenir que l’interaction de ces facteurs contrecarre le risque de récidive, le processus présenté par le programme encourage la survenue d’effets positifs dans la vie du sujet – conformément aux facteurs facilitateurs, indépendants du niveau de risque de récidive (Farrington, 2016). Le P.P.D. est résolument ancré dans les approches de la criminologie positive (Dieu et al., 2020b). Par sa philosophie humaniste et ses actions il consiste en un accompagnement bienveillant et structuré centré sur le sujet et visant la définition d’une vie nouvelle à l’heure de son engagement dans une communauté. En ce sens, il répond également au principe fondateur du good lives model de promouvoir une vie épanouissante où l’infraction n’a plus sa place (Ward et Gannon, 2006).Le P.P.D. répond en grande partie aux préconisations de la désistanceLe P.P.D. met en exergue les rationalisations et les questionnements identitaires, tant les possibilités de honte réintégrative (Braithwaite, 1989) que les ressentis de « peur de soi » (Paternoster et al., 2015). Avec une vie sociale diverse, le P.P.D. présente des figures d’identification positive pertinentes et soutient la transformation affective des PPSMJ (Giordano et al., 2007). La transformation affective est visible dans l’évolution et l’investissement des sphères affectives. Toutefois la transformation affective est encore de l’ordre des projections et aspirations inspirées des nouvelles interactions. C’est surtout la transformation cognitive des relations affectives qui est en jeu (Giordano et al., 2002) via les remises en question et contradictions assumées par les auteurs vis-à-vis de l’identité passée. Le programme démontre des effets sur la désistance primaire et secondaire (Maruna, 2004) à travers plusieurs éléments apportés dans les dossiers judiciaires (une baisse de la quantité d’occurrence durant le P.P.D.) et narratif sur le plan restauratif, socio-éducatif et psychologique. Les PPSMJ évoquent une amélioration de vie globale.Le P.P.D. répond à plusieurs des éléments soulevés dans les théories de la désistance présentées, comme les éléments pratiques observés par Weaver et McNeill (2010) d’une part et Owers et al., (2011) dans le cadre de la réinsertion post-carcérale Irlandaise :– Le programme favorise aisément les approches informelles entre les professionnels, les parrains et les PPSMJ. Il s’agit du cœur du contrat.– Il respecte par essence l’individualité des sujets, en personnalisant la relation de parrainage et en écoutant de manière bienveillante et basée sur la progression propre à chacun de sa disposition aux changements vers la désistance.– Le programme reconnaît l’importance des contextes sociaux tout en soulignant la centralité des initiatives individuelles et de l’engagement.– Le programme est une disponibilité sociale et communautaire à l’égard de la PPSMJ. Il interagit sur les questionnements éducatifs, professionnels, économiques, sociaux, relationnels, affectifs. Il y apparaît le souhait d’un renforcement du capital social et de l’image de soi symbolique, le changement au niveau social s’opérant ici avec l’apparition d’un nouveau réseau de soutien identificatoire (les autres filleuls, les parrains, les professionnels, la modification du plan de vie).– Effectivement, pour les différentes raisons déjà évoquées précédemment, le programme soutient et valorise les relations positives des sujets ainsi que leur processus de motivation et de disposition au changement.– Le programme administre à la personne un traitement judiciaire particulièrement respectueux de ses droits, besoins et attentes.Une démarche exploratoire encourageante à confirmer et des limites à comblerNombre de PPSMJ ont déjà rencontré des professionnels dans leur parcours de vie, pour autant, leur situation n’a pas forcément évolué comme ils auraient pu le souhaiter. De ce fait, les condamnés n’ont trop souvent que peu d’espoir dans le suivi qui pourrait à nouveau leur être proposé. Leur parcours est fait de nombreuses exclusions venues signer le manque de solidité de l’intérêt qui leur était préalablement porté. Le P.P.D. apporte une expérimentation des rapports sociaux sensiblement différente : la conviction que chacun a droit de trouver au bout du compte une place satisfaisante pour lui et pour les autres dans la société. Le P.P.D. s’intéresse aux besoins personnels au quotidien dans une relation horizontale nouée avec le premier cercle. Membre de la communauté à épanouir, et de fait à prévenir de la rechute, il se voit investi dans le dispositif de la mise en action de compétences d’auto-régulation sur les éléments internes et externes sources d’une infraction (Ward et Gannon, 2006 ; Yates et al., 2010 ; Ward et Beech, 2016). Pour © Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 25 Cahiers de la sécurité et de la justice – Hors-série autant, le P.P.D. montre encore des limites à combler, tant au niveau de sa structuration que de son évaluation. Voici sept améliorations à envisager : i) permuter les animateurs avec les référents ; ii) établir plus de liens avec les psychologues et le suivi CPIP ; iii) renforcer le suivi pour les personnes qui en ont le plus besoin ; iv) préciser les limites des approches informelles à l’aide d’exemples concrets issus de l’expérience ; v) accentuer les projections vers l’avenir ; vi) créer un référentiel commun partageable ; vii) renforcer le processus d’évaluation à l’aide d’outils structurés à l’entrée et la sortie du programme et un suivi de 3 à 5 ans des casiers judiciaires en plus d’une auto-évaluation nRéférencesBonta, James et Donald Andrews, 2007, Modèle d’évaluation et de réadaptation des délinquants fondé sur les principes du risque, des besoins et de la réceptivité, Sécurité Publique et protection civile du Canada.Braithwaite, John, 1989, Crime, shame and reintegration. Cambridge, Cambridge University Press.Dieu, Erwan, 2018a, « Programme de parrainage de désistance (P.P.D.) et Cercle de Soutien et de Responsabilité (C.S.R.) : synthèse des ressemblances et dissemblances pour une justice restaurative adaptée en France », Revue internationale de criminologie et de police technique et scientifique, LXXII (3) : 351-362.Dieu, Erwan, 2018b, Justice restaurative : réflexions psycho-criminologiques d’une Clinique judiciaire. Thèse de doctorat en psychologie, sous la direction d’Astrid Hirschelmann, Université Rennes 2, janvier 2018.Dieu, Erwan, 2019, « Que faire des modèles de la désistance dans l’accompagnement des auteurs d’infraction ? » Revue internationale de criminologie et de police technique et scientifique, LXXII (2) : 170-190.Dieu, Erwan et Laurent Merchat, 2018, « Le Parrainage de désistance : un programme original des Services correctionnels français basé sur la Criminologie positive », communication au « Good lives model (GLM) : Deuxièmes journées internationales francophones », Tournai, Belgique.Dieu, Erwan, Astrid Hirschelmann et Catherine Blatier, 2020a, « Motivé ? Satisfait ? Oui, mais à propos de quoi précisément ? Justice restaurative : tentative d’une taxonomie des attentes et satisfactions des participants », Revue internationale de criminologie et de police technique et scientifique, 83(1): 65-75.Dieu, Erwan, Ronan Palaric et Axel Maillot, 2020b, « Pratique restaurative et processus de désistance identitaire : le Programme de parrainage de désistance (P.P.D.) », Revue internationale de criminologie et de police technique et scientifique, 73(1) : 220-235.Estroff, Sue, Catherine Zimmer, William Lachicotte et Julia Benoit, 1994, « The Influence of Social Networks and Social Support on Violence by Persons With Serious Mental Illness », Hospital and community psychiatry, 45(7): 669-79.Farrington, David, 2016, « Risk, promotive, and protective factors in youth offending: Results from the Cambridge study in delinquent development », Journal of Criminal Justice, 45: 63-70.Giordano, Peggy, Stephen Cernkovich et Jennifer Rudolph, 2002, « Gender, crime and desistance: toward a theory of cognitive transformation », American Journal of Sociology, 107(4), 990-1064Giordano, Peggy, Ryan Schroeder et Stephen Cernkovich, 2007, « Emotions and crime over the life course: a neo-Meadian perspective on criminal continuity and change », American Journal of Sociology, 112(6), 1603-1661.Grubin, Don, 1997, « Predictors of risk in serious sex offenders », British Journal of Psychiatry, 170: 17-21.Guay, Jean-Pierre et Michiel de Vries Robbé, 2017, « L’évaluation des facteurs de protection à l’aide de la SAPROF », Annales Médico-Psychologiques, sous presse.Laub, John et Robert Sampson, 1993, « Turning points in the life course: why change matters to the study of crime », Criminology, 31(3): 301-325.Laub, John et Robert Sampson, 2001, « Understanding desistance from crime », Crime and Justice, 28, 1-69.Lefèbvre, Hans, Erwan Dieu et Esther Issen, 2018, « Les C.S.R. comme lien possible entre les principes RBR et la justice restaurative ? » Revue internationale de criminologie et de police technique et scientifique, LXXI(3) : 334-350.MacDonald, Imogen, 2016, Protective Factors in the Sexual Offending Area: Analysis of the Concept and a Preliminary Mode. Mémoire de master, Victoria University of Wellington, New Zealand.© Cahiers de la sécurité et de la justice - IHEMI26 I DOSSIER Le Programme de parrainage de désistance (P.P.D.) : une mise en lien concrète française de la justice restaurative et de la désistance – Erwan Dieu, Laurent MerchatMaruna, Shadd, 2001, Making Good: How ex-convicts reform and rebuild their lives. Washington, American Psychological Association.Maruna, Shadd, 2004, « Desistance and explanatory style: a new direction in the psychology of reform », Journal of Contemporary Criminal Justice, 20, 184–200.McMurran, Mary et Anna McCulloch, 2007, « Why don›t offenders complete treatment? Prisoners’ reasons for non-completion of a cognitive skills programme », Psychology, Crime & Law, 13(4): 345-54.McMurran, Mary et Ward, T, 2010, « Treatment readiness, treatment engagement and behavior change », Criminal Behaviour and Mental Health, 20(2): 75-85.Owers, Anne, Paul Leighton, Clodach McGrory, Fergus McNeill et Phil Wheatley, 2011, Review of the Northern Ireland Prison Service : Conditions, management and oversight of all prisons. Prison Review Team Final Report (http://cain.ulst.ac.uk/issues/prison/docs/2011-10-24_Owers.pdf).Paternoster, Ray, Ronet Bachman, Shawn Bushway, Erin Kerrison, et Daniel O’Connell, 2015, « Human Agency and Explanations of Criminal Desistance: Arguments for a Rational Choice Theory », Journal of Developmental and Life-Course Criminology, 1(3): 209–235.Paternoster, Ray et Shawn Bushway, 2009, « Desistance and the feared self: toward an identity theory of desistance », Journal of Criminal Law and Criminology, 99, 1103–1156.Prochaska, James et John Norcross, 2010, Systems of Psychotherapy: A Transtheoretical Analysis. Belmont, Brooks and Cole (7e éd.)Serin, Ralph, Nick Chadwick et Caleb Lloyd, 2016, « Dynamic risk and protective factors », Psychology, Crime & Law, 1-2: 151-170.Vogel, Vivienne (de), Michiel de Vries Robbé, Corine De Ruiter et Yvonne Bouman, 2011, « Assessing Protective factors in forensic Psychiatric Practice: introducing the SAPROF », International Journal of Forensic Mental Health, 10(3): 171-177Vries Robbé, Michiel (de), et Gwenda Willis, 2017, « Assessment of protective factors in clinical practice », Aggression and violent behavior, 32 : 55-63, janvier-février.Vries Robbé, Michiel (de), Ruth Mann, Shadd Maruna, et David Thornton, 2015, « An exploration of protective factors supporting desistance from sexual offending », Sexual abuse: a journal of research and treatment, 27: 16-33.Ward, Tony, 2017, « Prediction and agency: the role of protective factors in correctional rehabilitation and desistance », Aggression and violent behavior, 32: 19-28.Ward, Tony et Anthony Beech, 2016, « The Integrated Theory of Sexual Offending–Revised: A Multifield Perspective », in Boer, Douglas, The Wiley Handbook on the Theories, Assessment and Treatment of Sexual Offending. John Wiley and Sons, p.123-137.Ward, Tony et Theresa Gannon, 2006, « Rehabilitation, etiology, and self-regulation: The comprehensive good lives model of treatment for sexual offenders », Aggression and violent behavior, 11(1): 77-94.Ward, Tony, Andrew Day, Kevin Howells et Astrid Birgden, 2004, « The multifactor offender readiness model », Aggression and violent behavior, 9(6): 645-73.Ward, Tony, et Mark Brown, 2004, « The good lives model and conceptual issues in offender rehabilitation », Psychology, Crime and Law, 10(3): 243-57.Weaver, Beth et Fergus McNeill, 2011, « Travelling hopefully: desistance research and probation practice », in Brayford, Jo, Francis Cowe et John Deering (dir.), What else works? Creative work with offenders. Cullompton, Willian.Wilson, Chris, Andrew Bates et Birgit Völlm, 2010, « Circles of Support and Accountability: An Innovative Approach to Manage High-Risk Sex Offenders in the Community », The Open Criminology Journal, 3: 48-57. Wilson, Robin, Franca Cortoni et Monica Vermani, M, 2007, Cercles de soutien et de responsabilité : Reproduction à l’échelle nationale des résultats obtenus. Humber.Yates, Pamela, David Prescott et Tony Ward, 2010, Applying the good lives and self-regulation models to sex offender treatment. A Practical Guide for Clinicians. Brandon, The Safer Society.Zehr, Howard, 1990, Changing lenses. A new focus for crime and justice, Herald Press.Zehr, Howard, 2002, The little book of restorative justice, Good books.© Cahiers de la sécurité et de la justice - IHEMICahiers de la sécurité et de la justice – Hors-série DOSSIER I 27The Volunteers of Desistance (VoD): a concrete French link between restorative justice and desistanceErwan DIEU, Laurent MERCHATContext of the Volunteers of Desistance (VoD) programne mission of public authorities is to enforce public order and prevent disturbances, including damage done to the security of people and property. In the last 20 years, crime has been a major concern, and in spite of the varying intensity in media coverage, the efficiency of the implemented solutions has been constantly questioned. The statistical data 1 of the Ministry of Justice have highlighted that offenses are, much too often, committed by someone who has already been convicted before. The numbers appearing in several different studies vary depending on the nature of the offense, and according to the selected criterion (legal recidivism, reiteration, prior judicial record). However, the percentage of first-time offenders is undeniably low (less than 38% of prison population 2). Prevention of recidivism has been, since 2008, the main mission of the Penitentiary Service of Rehabilitation and Probation (SPIP), which are delegated services of the Penitentiary Since the law of August 2015 and the circular of March 2017, restorative justice is a penal measure meant to be applied concretely in the French correctional services. However, the usual restorative programs (e.g. mediation, group, circle) are not always applicable in the current conditions and require adaptations. The “Volunteers of Desistance” (VoD) is an innovation of the Penitentiary Service of Rehabilitation and Probation (SPIP) of the department of Drôme (France) with the support of Service of Criminology ARCA. It is a new version of the “Circles of Support and Accountability” (CoSA). VoD is a possible practical link between restoration, desistance and rehabilitation. This original program presented at the international congress of good lives model (GLM) (Tournai, Belgium) will be exposed here, and discussed in light of the criminological models and the effects observed on the participants.Key words: desistance; prevention of recidivism; rehabilitation; restorative justice; Volunteers of Desistance (VoD)OErwan DIEUCriminologist, practioner trained and accredited in EMDR psychotherapy (EMDR-Europe), doctor in psychology and director of the Service of Criminology (ARCA). He is specialist of Restorative Justice for the Minister, member of Good Lives Model international board, and he designed TIM-E (temporal identity model), implemented in correctional services of several countries on various issues and the treatment of psychotrauma. He is coordinator for the European program 3C2D on the treatment of radicalization.Laurent MERCHATPenitentiary Director of Rehabilitation and Probation (DPIP) to the Penitentiary Service of Rehabilitation and Probation (SPIP) in Roanne and former Penitentiary Counselor of Rehabilitation and Probation (CPIP) in the department of Drôme (France). Trained to restorative justice since 2015, inspired by the Circles of Support and Accountability (CoSA), he proposed to Director Doriane Serrières a program likely to answer the desistance needs of the many: the “Volunteers of Desistance”.(1) Infostat Justice n° 50, 68, 88, 108, and 1227.(2) Infostat Justice n° 127, graph 1: 63% of people convicted of an offense have a criminal record within the previous 8 years (the data covers people convicted of crimes, infractions and gravest misdemeanors from 2004 to 2011).© Cahiers de la sécurité et de la justice - IHEMI28 I DOSSIER The Volunteers of Desistance (VoD): a concrete French link between restorative justice and desistance – Erwan Dieu, Laurent MerchatAdministration. In 2012, prevention of recidivism was brought to the fore in political and media discourse as “a major challenge for society and the justice system” during a “consensus conference on the efficiency of penal responses to better prevent recidivism”. The law of August 15th, 2014 regarding the individualization of sentences and the maximization of the efficiency of penal punishment introduced restorative justice into the Code of criminal procedure in its article 10-1 (Zehr, 1990). The circular of March 15th, 2017 defines restorative justice as “a model of justice that complements the penal process, and aims to restore the social fabric damaged by the offense, through the execution of various measures involving the victim, the perpetrator and society.”Among the penal measures of restorative justice, the Circles of Support and Accountability (CoSA) are the only programs that have reached international scientific consensus in that they have direct impact on the decrease of the risk of recidivism (Wilson et al., 2010). Within the CoSA, a perpetrator returning to free life and deemed to pose a high risk of recidivism is surrounded by members of the community. Once a week during a year, the main member, the volunteers and the facilitator meet in a circle. Such a model might be considered not to qualify as restorative justice, as it doesn’t involve the victims. In 2016, considering its primary mission of prevention of recidivism, and based on the evidence of their efficiency, the SPIP of the department of Drôme decided to invest in a CoSA program, and gathered a staff of several professionals specifically trained in restorative justice. At the moment, the SPIP of Drôme is responsible for supervising approximately 2,000 convicted people within the territory of its jurisdiction, whether they are physically residing in the Penitentiary (around 450), or serving their sentences in open custody, for example those doing community service, on house arrest or on probation (over 1,500). Implementing such a CoSA program requires a great amount of work hours. Given the already important individual workload, how could it be possible to dedicate that much energy to just a few, however “dangerous” they might be? As a consequence of this reasoning, although the idea of implementing a CoSA program was not completely dismissed, the direction of the SPIP decided to launch an initiative inspired by restorative justice, called the “Volunteers of Desistance”.The Volunteers of Desistance program (VoD, or in French PPD, Programme de parrainage de désistance), began in December 2016 at the SPIP of Drôme-Ardèche, and lasted until the feedback and assessment report meeting attended by the participants in March 2018. It was evaluated during the program by the ARCA criminology service 3 by request of the SPIP. The Volunteers of Desistance (Dieu & Merchat, 2018) is a program that seeks to set up the conditions for the restoration of authentic, sincere and supportive social bonds between volunteer members of the community and persons placed under the supervision of the justice system after release (P.P.S.M.J.s). The meetings are based on the freedom and responsibility of each person, i.e. voluntary participation without any compensation. The participants meet once a month in a circle. Sponsors and sponsorees can then choose each other, in a sponsorship they can determine the duration of (4, 6 or 8 months). Currently, being that new people constantly enter and exit the program, approximately 14 sponsorees and 14 volunteers are active in the process (sponsorees attended on average 7 out of 10 monthly meetings, and sponsors 8 out of 10). In June of 2018, after 18 months of existence, 25 convicts got involved in the program and 20 sponsorship contracts were signed.Theoretical basis of the program: a return to the social contract through positive criminologyCrime is a breach of the social contract. It is therefore necessary not only to punish it, but also, and especially, to make sure it doesn’t happen again. Yet it is evident that when a crime is committed, a penal process is not enough to permanently remove criminal risk from the social fabric, and consequently restore trust in the social cohesion. A trial allows to recognize the perpetrator’s responsibility, and the harm caused to the victim. The justice system confers a sentence on the former, and damages compensation on the latter, and leaves to each of them the task of handling the impact the crime had in their lives. The program is a proposal based on the principles and values of restorative justice (Zehr, 2002) to respond to that issue. It is an original device which has the potential to support a great amount of people, handed to the program by the justice system regardless of the seriousness of the offense they committed. Indeed, many of them suffer from being isolated during the transformation process they go through on their path towards desistance, because they don’t have the support needed. Besides, they also struggle with building and achieving their projects, as they are not always capable and/or responsible enough to carry out their rehabilitation journey. (3) Nonprofit association founded in 2008 whose service is accredited by the Department of access to legal services and support to victims (SADJAV) of the Ministry of Justice with regards to restorative justice, and recognized in the Circular of March 2017 as an official French institution of training and education in restorative justice.© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 29 Cahiers de la sécurité et de la justice – Hors-série The Volunteers of Desistance program, as its goals and its very name indicate, is related to theories of desistance, but also to the good lives model (GLM) and the Temporal Identity Model (TIM-E) (Dieu, 2018a). These are the desistance theories that inspired the VoD’s chosen approach:– According to Laub and Sampson (1993, 2001), the desistance process is to be considered with regards to “turning-points” (e.g. marriage, parenthood) which set the individual on a path that eventually results in the interruption of offending behaviors. – Maruna’s phenomenological analysis (2001) adopts a subjective psychological understanding of the desistance process in two stages. Primary desistance being when the subject enters an offense free period, while secondary desistance relates to the reconfiguration of the subject’s identity through the social roles he or she inhabits. – The theory of cognitive transformation (Giordano et al., 2002) considers that the subjects might begin to question their beliefs which allows them to distance themselves from their delinquent behaviors, while affective transformation (Giordano et al., 2007) implies learning new emotions through positive interactions. – Paternoster et al. (2015) base their analysis on the theory of rational choice, according to which an offender acts out after weighing the costs and benefits of an act, and deeming it favorable. After a certain number of offenses throughout the criminal’s life-course, the global result of the cost-benefits analysis becomes negative, and causes the subject to fear what he has become and to reject his “feared self” (Paternoster & Bushway, 2009).The VoD program is also closely linked to the good lives model (Ward & Brown, 2004), as well as other Oceanian approaches (e.g. Multifactor Offender Readiness Model, Ward et al., 2004). The good lives model, like the VoD program, is a model that takes into account the wellbeing and personal development of the subject, and considers the person as a whole. With the VoD program, the subject goes through a process of acquisition of internal and eternal skills in order to advance towards a more fulfilling and prosocial life. The Temporal Identity Model “TIM-E” (Dieu, 2019) is a model of criminological understanding focused on the future perspectives of offenders. The purpose of the guidance emphasizes the subject’s perception of time and its impact on identity, which is both characterized by a feeling of unity, and constantly shifting. The individual feels the “presence” of his or her desistant identity, which initiates and facilitates over time the process of turning away from crime. Professional intervention cannot claim to have any control over “age, meaning maturity”, or “the positive events of life (e.g. romantic relationship, birth of a child)”. However, it certainly can impact the other two main factors of desistance 4: “the enhancement of human capital (e.g. ability to communicate or handle emotions)” and “the development of social capital (e.g. integration in non-delinquent relationships and social networks, development of personal and social skills, professional rehabilitation)”. Implementation of the Volunteers of Desistance program (VoD)The VoD device: a complete readjustment of the CoSA modelThe device aims to enhance human capital and develop social capital. Upon completion of its evaluation, it was confirmed that these goals are not only pre-conditions to a shift towards the end of reoffending, but are also factors of social integration (finding work, pursuing training or education, gaining access to social benefits...), although such social integration does not even initially appear as an objective of the program. According to the responsivity principle (Bonta & Andrews, 2007), in order to be efficient, a program must be adapted to the person’s abilities. The VoD program is a new, reshuffled version of the CoSA, which also draws on the new intervention methods of the SPIP (Lefebvre, et al. 2018). It is evidence of the fact that the Prison Administration staff is taking on the new missions of the penal Justice system and putting them into practice. Here are the similarities and differences between the VoD and the CoSA of the VoD (Dieu, 2018a):The similarities between CSA and VoD:– Vulnerable people who have an important history of violations. – Both the volunteers and the beneficiaries (“main mem-bers”) are participating willingly, by their own volition. – The volunteers get training on the principles of recidivism prevention, restorative justice, motivation to change, (4) Report of the Consensus Conference on the Prevention of Recidivism, p. 122: Fiche 4 “Les facteurs de risque, de protection et de désistance”, Paris, 2012© Cahiers de la sécurité et de la justice - IHEMI30 I DOSSIER The Volunteers of Desistance (VoD): a concrete French link between restorative justice and desistance – Erwan Dieu, Laurent Merchatdesistance and CoSA. Meetings take place outside of intimate spaces or professional facilities (public spaces). The P.P.S.M.J. is offered a great availability of time, and the possibility of extra meetings with the sponsor beyond the initial meetings provided by the contract (weekly interactions).– The SPIP is also involved in some cases in CoSAinitiatives in France.The specificities of the Volunteers of Desistance program (VoD)– All kinds of offenses are covered, inclusion in the program doesn’t depend on an assessment of the risk of recidivism.– Duration of 4, 6, or 8 months, renewable. Mandatoryattendance in monthly meetings, even without asponsorship contract. For each main member or“sponsoree”, there is one volunteer or “sponsor”, whichallows for day-to-day, individualized support.– Monthly meeting between all the volunteers (whetheror not they are sponsoring someone) and theP.P.S.M.J. (whether or not they are being sponsored bysomeone), in a room outside the SPIP facilities. TwoCPIP (Penitentiary Counselors of Rehabilitation andProbation) coordinate the monthly group meetings.– The interactions are directly facilitated by the SPIP’s CPIP.Advantages of the VoD program in comparison with the CoSA (both for the SPIP and the P.P.S.M.J.)– Constant turnover as the P.P.S.M.J. are recruited intothe program or exit the program, important diversity ofprofiles and interactions between all kinds of partners.– The person’s safety is valued, the group remains open topeople regardless, even if they don’t sign a sponsorshipcontract (secondary circle in the monthly meetings).– Freedom in the mutual choice between sponsor andsponsoree, as getting along well and sharing commoninterests guarantees optimal, ongoing support and guidance.– VoD adopts an approach of desistance which relates thecessation of offending with sustained prosocial behaviorand rehabilitation into society.– The program is coordinated by two CPIP from theSPIP, and evaluated externally by a third-party entity.Putting the Volunteers of Desistance program (VoD) into action)The VoD program is presented to the the P.P.S.M.J. via a flyer entitled Adopt a sponsor for desistance. The steps towards change are simplified so as to allow for an easy and explicit collaborative assessment. Step 1: “Motivation: I want to change but I still feel stuck in behaviors that hold me back” ; step 2: “a Plan for Action: I have decided to change and I try to act accordingly”; step 3: “Consolidation: I have changed, but I have to remain cautious because I still have vulnerabilities”. These three steps fall within the notion of specific responsivity (Bonta & Andrews, 2007) and the theory of readiness to change (Prochaska & Norcross, 2010), between an initial stage called precontemplation, in which the desire for change isn’t really present, and a final stage which marks the permanent termination of delinquency, and satisfying integration within the social contract. The term “adopt” is chosen in order to immediately invite the person to active participation: the relationship is not unilateral, choosing a sponsor means taking one’s part of the responsibility in the relationship, mutually caring about each other and establishing reciprocity. After an initial oral non-disclosure commitment with the CPIP in charge of facilitating the interactions, the P.P.S.M.J. enters the circles with a white “first contact” badge. The monthly meeting takes place outdoors, on the first Saturday of each month. Volunteers and P.P.S.M.J.s meet and form a large circle. Each participant is responsible for the whole group and its smooth operation, and makes sure each person can express him or herself within the bounds of the rules of communication of the VoD program. Then follows a conversation with the CPIP, during which the VoD code of conduct, describing the roles, rights and duties of each participant, is read and explained. The person can then decide whether or not to enter the program, knowing that this choice will not have any impact on the future judicial treatment of his or her case. If the person signs the mutual commitment convention with the SPIP management, he or she will go back to the circle with a badge indicating first name and sponsoree status. All of the P.P.S.M.J. who came in contact with the program decided to enter it, except one. The color-coded badges (specific color depending on the contract) become the visual sign that the participants are all connected within the device, all of them in their proper place. When the interaction between two participants leads to the mutual decision to commit to a sponsorship, a contract is signed within the group, for a renewable duration of 4, 6 or 8 months. Both of them will share their testimony of the relationship within the circle, as anything that happens in the relationship can be said in the group, although not everything needs to be shared. © Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 31 Cahiers de la sécurité et de la justice – Hors-série Demographics of the participants: profiles of the P.P.S.M.J. “sponsorees of desistance” and the volunteers “sponsors of desistance”Analyzing the criminal background of the 18 PPMSJ involved in the program at the time of the evaluation evidences deeply rooted histories of delinquency. Over 50% of the group members have been convicted at least 5 times. 30% of the group members have been convicted over 10 times (up to 29 convictions), while only 1 out of 10 sponsorees can be qualified as a first offender. 70% of the sponsorees have been sentenced to jail time. 50% have actually been incarcerated. 50% have served or are currently serving adjusted sentences. Only 20% of them benefitted from sentence adjustments without ever being incarcerated. Half of the group member were serving an adjusted sentence at the time of their participation in the program (6 were on house arrest with electronic monitoring, 1 was on an outside supervised program similar to parole, probation or work release programs), or had just gotten out of jail (2 had been out for less than a month). The number of convictions is related to the duration of the offending period (from 1 year to 19 years, 5 to 6 years on average). The members of the group of sponsorees committed 23 different types of offenses, mostly thefts, violent assaults and road traffic violations. It is noted that among the 45% of the group who committed thefts, 75% of them also committed violent assaults. And among those who were convicted of violent assaults, 61% were also convicted of narcotic violations or driving under the influence of alcohol. 39% of the sponsorees have committed at least one road traffic violation (but none of them was convicted only of traffic violations). Both of the sponsorees who were convicted of involuntary manslaughter were drug users, while the one who was convicted of sexual assault was involved in alcohol-related incidents (DUI). The P.P.S.M.J. involved in the program were men (only one exception). Two thirds of them were under 30 years old, the age of the participants varied between 21 and 58. Almost half of the sponsorees lived with their families (mostly those under 30).The volunteers involved in the device are also mostly men. Their ages vary between 26 and 78 years old (50 years old on average). They belong to a diversity of socio-economic categories, which makes them representative of society. Two thirds of them are of working age (out of which 2 are unemployed), which allows for a more accurate example of the requirements of the job market. Two thirds of them live with a spouse or partner. Half of them are also involved in other nonprofits or associations. It must be noted that a sponsor might have been convicted in the past, might have been supervised by the justice system, and might have desisted from delinquency. It is irrelevant to know who might be in that case. And when a sponsoree’s parole or supervision period ends after sharing a beautiful journey, and decides to switch roles and don the sponsor badge, it doesn’t come as a surprise. The device is not specifically intended to transform sponsorees into sponsors, but it is a possibility, which demonstrates delinquency is a path one can leave, that it can be limited to a time in one’s life, that it can be overcome. Feedback on the VoD program: promising results in terms of engagement and satisfaction Engagement, expectations and satisfaction with the VoD program The program’s evaluation covered its whole period of existence: 15 months, from November of 2016 to February 2018, date when it was evaluated (Dieu et al. 2020b). At the beginning of the VoD program, in December of 2016, 2 P.P.S.M.J. were involved (out of 16 participants), and seven months later in July of 2017, the program reached 13 involved P.P.S.M.J. out of 23 participants, and was then stabilized at 10 involved P.P.S.M.J. out of 23 participants until the evaluation of the program. From the stabilization of the program in April of 2017, 12 P.P.S.M.J. were involved on average, out of 19 participants. The average number of participants in the monthly meetings was 19. The attendance rate over the commitment period was 70% for the sponsorees, and 80% for the sponsors. The sponsorship contracts vary: 7 P.P.S.M.J. committed for 4 months, 6 of them committed for 6 months, and 3 of them committed for 8 months. Three contracts were renewed for 6 months, two of which were initially 8 months contracts and were thus brought to a total duration of 14 months. In terms of motivation for change, the evolution of the stages of readiness to change described by Prochaska et Norcross (2010) was followed by the sponsorees and the sponsors, and recalled by the CPIP in charge, all through the VoD program. 2 PPMSJ had just recently entered the program at the time of the evaluation, and one did not stay long enough to be taken into account in the statistics. Excluding those three, all 15 P.P.S.M.J. whose results can be measured showed a great evolution in their readiness to change. At the beginning, only 2 of them seemed to have reached the “action” stage, while at the end of their journey, 10 out of 15 had. Only one P.P.S.M.J. out of 15 reached the end of the sponsorship period with a relatively low readiness to change. © Cahiers de la sécurité et de la justice - IHEMI32 I DOSSIER The Volunteers of Desistance (VoD): a concrete French link between restorative justice and desistance – Erwan Dieu, Laurent MerchatMore than readiness to change, the notions of treatment engagement and treatment readiness (Ward et al., 2004) are nowadays considered as a conclusive factor to be taken into consideration in rehabilitation programs (McMurran & McCulloch, 2007; McMurran & Ward, 2010). Upon entry into the program, according to the method of assessment of restorative motivation and satisfaction (Dieu, 2018b; Dieu et al., 2020a), the subjects had relatively strong expectations, varying from 54% to 74%, reaching on average 54% of maximal intensity. Such results are much lower than the intensity of restorative satisfaction derived from the program, which is estimated at 76% on average, and varies from 61% to 89%. The sponsorees have indicated that their expectations were fulfilled, that they “would recommend the program to other P.P.S.M.J. in a similar situation”. They have described the positive impact of the program calling it a “good program”, which “makes it easier to meet people” and “helps rehabilitation”. As for the sponsors, from their internal perspective on the process, they highlighted the fact that the program “favors interactions and builds trust”. The emotional and cognitive satisfactions reported are “feeling at peace”, “taking back control over one’s life, better self-esteem and self-confidence”, a “new outlook on one’s life course”, related to an “avoidance of reoffending” which was frequently mentioned in the declarations.Observing the evolution of the stages of change in the Life Plan narratives Among the groups of sponsorees described above, upon entry into the program: 33% are identified as being at step 1 of change, and 54% at step 2. Therefore only 13% of them had already been through step 1 and 2. It can be noted that the initial stage they are at does not seem to be correlated with their previous penal situation. Having or not having been incarcerated does not foretell the outcome of desistance, nor does having benefited from a sentence adjustment: with or without sentence adjustment, upon entry, one third are at level 1, and two thirds are at level 2. At the end of the program, 0% to 6% of the group are still at stage 1, 20% to 26% are at stage 2, and 67% to 80% are at stage 3. At the beginning of the program, 87% of the group displayed behaviors that exposed them to the risk of reoffending, which was only the case for 20% to 32% of the group at the time of the evaluation. This means 62% to 77% of the affected P.P.S.M.J. moved away from reoffending risk. It also appears that during the course of the program, 81% of the group experienced a significant evolution in their integration in terms of employment, activity or advancement of their projects. That evolution was especially significant for 31% of the group. In terms of evolution of trajectories, an inversion of life priorities is observed, evidencing a stronger motivation towards professional and social integration. The sphere of relationships is reconfigurated in ways that are more consistent with the changes desired and expressed by the P.P.S.M.J., but also in the context of the sponsorship contract, and in the institutional context of the SPIP. Suffering, albeit still present, emotionally negative and felt strongly, is less overwhelming in day-to-day life. The participants dedicate less time to it and expose themselves to it less, as they have given priority to other aspects of their lives. Our hypothesis is that the elements of life that cause negative emotions are now better regulated cognitively. The participants seem more concrete and realistic in their vision of the present as well as their projection of the future. It is no longer about a fantasized life plan, but rather a subjective desired perspective, built upon the changes being made in the present. The idea is not radical transformation but rather a reconfiguration of day-to-day life, in order to achieve better quality of life (quality of relationships, quality of social investment) and build accessible happiness. The P.P.S.M.J. seem to have obtained the necessary skills to draw the outlines of the Life Plan they deem adequate, and to distinguish the positive and negative outcomes of their choices and investment. Which is quite new, as they previously used to talk about their day-to-day lives in undistinguished neutral terms.Feedback on the VoD program: promising exploratory results translated in a positive criminology frameworkThe VoD program could foster the crucial role of agency 5 in the emergence and reinforcement of certain protective factors in relation to the “Good Lives” primary needs Protective factors are considered to be essential tools in supporting the desistance process (De Vries Robbé, et al., 2015). They can be broadly defined as any characteristic of a person, his or her environment or situation which “reduce the likelihood of violent recidivism” (Vogel et al, 2011: 172). There are three coexisting models regarding protective factors (Ward, 2017; De Vries Robbé & Willis, 2016; Serin, Chadwick, Lloyd, 2016; MacDonald, 2016). Protective factors are classified into three subtypes, defined by specific criteria: internal, motivational and external factors (Guay & De Vries Robbé, 2017).© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 33 Cahiers de la sécurité et de la justice – Hors-série Internal protective factors:– a better adaptation ability and a development of the theory of mind (more than empathy in itself) seem to emerge from the quality of the relationship being built through the bond established between sponsorees and sponsors. – emphasis on the self-control factor, better cognitive regulation of emotions and better control of violent impulses. By providing advice based on their own lives on how to handle situations/issues or relapse patterns, the sponsors (and the group of sponsorees) become mirrors and models for the PSMJ. They support the conscious realization of the responses’ limited resources and the possible emotional and cognitive-behavioral alternatives.Motivational protective factors:– motivation for treatment appears as some of the participants begin to work on getting invested in psychological treatment, but also with regards to addiction treatment, which is evidenced by the decrease in substance abuse. investing in the professional sphere (work) is mentioned by sponsorees and sponsors as a lever for change, but also as an end goal.– leisure activities are mentioned, for example traveling (considered or achieved) and a better structuration of social activities in daily life. – the interactions between the sponsorees and the support of the sponsors have an impact on the P.P.S.M.J.s’ economic situation and budget management which is transformed both objectively (e.g. sustained professional activity) and subjectively (e.g. new values and beliefs regarding money). – attitudes towards authority seem to have evolved. – the life goals of the P.P.S.M.J. are impacted, they now include prosocial aspirations.– medication is manifested in the changes and actions for change related to the evolutions in the Life Plans regarding general treatment of substance abuse.External protective factors:– social networks have evolved, the participants invest more time in their friends, and some of them mention putting some distance between them and the people who were a toxic influence in their lives. – another significant evolution is that the participants become invested more positively in family (both ascendant and descendant) and dedicate more time to it – professional care and external control overlap within the program, as the sponsors are, daily (and the other sponsorees and the CPIP in charge, monthly), both essential elements of social control and facilitators of professional care. The attention given to shared prosocial values oriented towards non-reoffending and rehabilitation legitimates the formal and informal social control and the avoidance of isolation and exclusion (Wilson, 2007). – the importance given to romantic relationships and to one’s partner indicate an improvement and a transformation of intimate relationships, and in part, of living conditions. In general, it must be noted that the tighter and stronger the circle, the deeper the work accomplished can be, and the more it can focus on the main member’s personal needs (housing, work, administrative issues) and on the learning of positive behaviors through social modeling. The relation to others, that operates through the program, is recognized in scientific literature as a major element in the field of rehabilitation. (Estroff, et al., 1994; Grubin, 1997; Bonta & Andrews, 2007). The group offers its members a great openness, a judgment-free space which fosters self-confidence. The interactions and advice provided on concrete situations in order to suggest cognitive and behavioral alternatives allows the participants to experiment the proposed solutions in real life. In the following meeting, the group (as well as the sponsor in day-to-day interactions) will ask the individual again about the situation and whether or not the advice provided was put into practice.Ward’s hypothesis (2017) is that there are two distinct categories of protective factors: capacity for agency, and contextual factors. The program makes available to the subject a set of positive contextual factors, both internal (within the circle) and external (outside of the circle). The capacity for agency is then stimulated by these contextual factors and more energy is dedicated to it. Here, it is impossible to take into account the individual (5) Agency is a concept that refers to the way the people consider themselves able to make their own decisions over their actions in a given environment. It seems that people who desist from crime express more agency, more control in their life narratives, than people who reoffend.© Cahiers de la sécurité et de la justice - IHEMI34 I DOSSIER The Volunteers of Desistance (VoD): a concrete French link between restorative justice and desistance – Erwan Dieu, Laurent Merchatcharacteristics of the subjects, or to claim that the interaction of these factors thwarts the risk of reoffending. However, the process offered by the program fosters the onset of positive effects in the life of the subject - according to the facilitating “promotive factors”, which are not dependant on the level of reoffending risk (Farrington, 2016). The VoD program strongly hinges upon positive criminology approaches (Dieu, et al., 2020b). Through its humanistic philosophy and its actions, it consists in a compassionate, benevolent and structured support centered on the subject and aims at defining a new life as the person engages in the community. In that regard, it is also in accordance with the founding principles of the good lives model which promotes a fulfilling life in which there is no longer a place for offending (Ward & Gannon, 2006).The VoD program responds to a great extent to the recommendations of desistance models The VoD program emphasizes processes of rationalizing and questioning identities, possibilities of reintegrative shaming (Braithwaite, 1989) as feelings of Fear of one’s Self (Paternoster, et al., 2015). By diversifying social life, the VoD program introduces relevant positive identification figures and supports the affective transformation of the P.P.S.M.J. (Giordano, Schroeder, Cernkovich, 2007). Affective transformation is evidenced by the evolution and the commitment to affective spheres. However, affective transformation remains at the projection or aspiration level, inspired by the newly emerging interactions. What is at work, most of all, is a cognitive transformation of affective relationships (Giordano, Cernkovich, Rudolph, 2002) through the questioning and admitted contradictions of the perpetrators regarding their past identity. The program demonstrates an impact on primary and secondary desistance (Maruna, 2004) through various elements that appear in the criminal records (the number of offenses diminished during the VoD program) and in the personal narratives on restorative, socio-educational and psychological levels. The P.P.S.M.J. reported a global life improvement.The VoD program responds to various of the leading principles identified as supporting desistance, as the practical elements observed by Weaver and McNeill (2010) and Owers et al. (2011) on post-incarceration rehabilitation in Ireland:– The program easily favors informal approaches between the professionals, the sponsors and the P.P.S.M.J.. They are at the core of the contract.– By nature, it respects the individuality of each subject, by offering personalized sponsorship and compassionate listening, based on each person’s progress and readiness to change towards desistance.– The program recognizes the significance of social contexts all the while highlighting the importance of engagement and individual initiatives.– The program makes available to the P.P.S.M.J. social and community resources. It interacts with relationships, affective, social, economic, professional and educational issues. The participants express through the program their desire to reinforce their social capital and symbolic self-image, and the social change operates through the apparition of a new support network for identification (the rest of the sponsorees, the sponsors, the professionals, the transformation of the life plan). – Indeed, for all the reasons mentioned above, the program supports and values the subjects’ positive and meaningful relationships as well as their motivation and readiness to change processes.– The program provides the people with a particularly respectful and fair judicial treatment of their rights, needs and expectations.A promising exploratory approach to be furthered, and limitations to be addressed Many P.P.S.M.J. have met professionals during their life course, however, it was not always enough to improve their situation in the ways they might have wished it had. In consequence, convicts often feel hopeless about any future support they might be offered. Their trajectories are marked with numerous exclusions which were the sign of the lack of solidity of the attention they have been given thus far. The VoD program brings to them quite a different experience of social relations: the belief that anyone is entitled to eventually finding a satisfying place for oneself and for others within society. The VoD program focuses on the personal needs of day-to-day life, in the horizontal relationship built with the first circle. As a member of the community whose needs deserve to be fulfilled and who, thereby, deserves to be prevented from relapse, the P.P.S.M.J. becomes invested in a device which puts into action skills of self-regulation that operate on the internal and external causes of reoffending (Ward & Gannon, 2006; Yates et al., 2010; Ward & Beech, 2016). However, the VoD still has limitations that must © Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 35 Cahiers de la sécurité et de la justice – Hors-série be addressed, both in terms of structuration and in terms of evaluation. Here are the 7 improvements that should be taken into consideration: i) replace the facilitators with an adviser; ii) establish stronger links with the psychologists and the CPIP follow-up; iii) intensify the support for those who need it most; iv) specify the limitations of informal approaches via concrete experience-based examples; v) increase projections towards the future; vi) create a shareable common reference framework; vii) enhance the evaluation process through structured tools on entry and exit of the program, and a 3 to 5 years follow-up on criminal records, plus an self evaluation nReferencesBonta, James and Donald Andrews, 2007, Modèle d’évaluation et de réadaptation des délinquants fondé sur les principes du risque, des besoins et de la réceptivité, Sécurité Publique et protection civile du Canada.Braithwaite, John, 1989, Crime, shame and reintegration, Cambridge, Cambridge University Press.Dieu, Erwan, 2018a, «Programme de parrainage de désistance (P.P.D.) et Cercle de Soutien et de Responsabilité (C.S.R.): synthèse des ressemblances et dissemblances pour une justice restaurative adaptée en France», Revue internationale de criminologie et de police technique et scientifique, LXXII (3): 351-362.Dieu, Erwan, 2018b, Justice restaurative: réflexions psycho-criminologiques d’une Clinique judiciaire. PhD dissertation in psychology, Université Rennes 2, janvier 2018.Dieu, Erwan, 2019, «Que faire des modèles de la désistance dans l’accompagnement des auteurs d’infraction?» Revue internationale de criminologie et de police technique et scientifique, LXXII (2): 170-190.Dieu, Erwan and Laurent Merchat, 2018, «Le Parrainage de désistance: un programme original des Services correctionnels français basé sur la Criminologie positive», communication to the «Good lives model (GLM): Deuxièmes journées internationales francophones», Tournai, Belgique.Dieu, Erwan, Astrid Hirschelmann and Catherine Blatier, 2020a, « Motivé? Satisfait? Oui, mais à propos de quoi précisément? Justice restaurative: tentative d’une taxonomie des attentes et satisfactions des participants», Revue internationale de criminologie et de police technique et scientifique, 83(1): 65-75.Dieu, Erwan, Ronan Palaric and Axel Maillot, 2020b, «Pratique restaurative et processus de désistance identitaire: le Programme de parrainage de désistance (P.P.D.)», Revue internationale de criminologie et de police technique et scientifique, 73(1) : 220-235.Estroff, Sue, Catherine Zimmer, William Lachicotte and Julia Benoit, 1994, «The Influence of Social Networks and Social Support on Violence by Persons With Serious Mental Illness», Hospital and community psychiatry, 45(7): 669-79.Farrington, David, 2016, «Risk, promotive, and protective factors in youth offending: Results from the Cambridge study in delinquent development», Journal of Criminal Justice, 45: 63-70.Giordano, Peggy, Stephen Cernkovich and Jennifer Rudolph, 2002, «Gender, crime and desistance: toward a theory of cognitive transformation», American Journal of Sociology, 107(4), 990-1064Giordano, Peggy, Ryan Schroeder and Stephen Cernkov-ich, 2007, «Emotions and crime over the life course: a neo-Meadian perspective on criminal continuity and change», American Journal of Sociology, 112(6), 1603-1661.Grubin, Don, 1997, «Predictors of risk in serious sex offenders», British Journal of Psychiatry, 170: 17-21.Guay, Jean-Pierre and Michiel de Vries Robbé, 2017, «L’évaluation des facteurs de protection à l’aide de la SAPROF», Annales Médico-Psychologiques, forthcoming.Laub, John and Robert Sampson, 1993, «Turning points in the life course: why change matters to the study of crime», Criminology, 31(3): 301-325.Laub, John and Robert Sampson, 2001, «Understanding desistance from crime», Crime and Justice, 28, 1-69.Lefèbvre, Hans, Erwan Dieu and Esther Issen, 2018, «Les C.S.R. comme lien possible entre les principes RBR et la justice restaurative?» Revue internationale de criminologie et de police technique et scientifique, LXXI(3): 334-350.MacDonald, Imogen, 2016, Protective Factors in the Sexual Offending Area: Analysis of the Concept and a Preliminary Mode. Unpublished MSc thesis. Victoria University of Wellington, New Zealand.© Cahiers de la sécurité et de la justice - IHEMI36 I DOSSIER The Volunteers of Desistance (VoD): a concrete French link between restorative justice and desistance – Erwan Dieu, Laurent MerchatMaruna, Shadd, 2001, Making Good: How ex-convicts reform and rebuild their lives. Washington, American Psychological Association.Maruna, Shadd, 2004, «Desistance and explanatory style: a new direction in the psychology of reform», Journal of Contemporary Criminal Justice, 20, 184–200.McMurran, Mary and Anna McCulloch, 2007, «Why don’t offenders complete treatment? Prisoners’ reasons for non-completion of a cognitive skills programme», Psychology, Crime & Law, 13(4): 345-54.McMurran, Mary and Ward, T, 2010, «Treatment readiness, treatment engagement and behavior change», Criminal Behaviour and Mental Health, 20(2): 75-85.Owers, Anne, Paul Leighton, Clodach McGrory, Fergus McNeill and Phil Wheatley, 2011, Review of the Northern Ireland Prison Service: Conditions, management and oversight of all prisons. Prison Review Team Final Report (http://cain.ulst.ac.uk/issues/prison/docs/2011-10-24_Owers.pdf).Paternoster, Ray, Ronet Bachman, Shawn Bushway, Erin Kerrison, and Daniel O’Connell, 2015, «Human Agency and Explanations of Criminal Desistance: Arguments for a Rational Choice Theory», Journal of Developmental and Life-Course Criminology, 1(3): 209–235.Paternoster, Ray and Shawn Bushway, 2009, «Desistance and the feared self: toward an identity theory of desistance», Journal of Criminal Law and Criminology, 99, 1103–1156.Prochaska, James and John Norcross, 2010, Systems of Psychotherapy: A Transtheoretical Analysis. Belmont, Brooks and Cole (7e éd.)Serin, Ralph, Nick Chadwick and Caleb Lloyd, 2016, «Dynamic risk and protective factors», Psychology, Crime & Law, 1-2: 151-170.Vogel, Vivienne (de), Michiel de Vries Robbé, Corine De Ruiter and Yvonne Bouman, 2011, «Assessing Protective factors in forensic Psychiatric Practice: introducing the SAPROF», International Journal of Forensic Mental Health, 10(3): 171-177Vries Robbé, Michiel (de), and Gwenda Willis, 2017, «Assessment of protective factors in clinical practice», Aggression and violent behavior, 32: 55-63, janvier-février.Vries Robbé, Michiel (de), Ruth Mann, Shadd Maruna and David Thornton, 2015, «An exploration of protective factors supporting desistance from sexual offending», Sexual abuse: a journal of research and treatment, 27: 16-33.Ward, Tony, 2017, «Prediction and agency: the role of protective factors in correctional rehabilitation and desistance», Aggression and violent behavior, 32: 19-28.Ward, Tony et Anthony Beech, 2016, «The Integrated Theory of Sexual Offending–Revised: A Multifield Perspective», in Boer, Douglas, The Wiley Handbook on the Theories, Assessment and Treatment of Sexual Offending. John Wiley and Sons, p.123-137.Ward, Tony and Theresa Gannon, 2006, “Rehabilitation, etiology, and self-regulation: The comprehensive good lives model of treatment for sexual offenders”, Aggression and violent behavior, 11(1): 77-94.Ward, Tony, Andrew Day, Kevin Howells and Astrid Birgden, 2004, «The multifactor offender readiness model», Aggression and violent behavior, 9(6): 645-73.Ward, Tony and Mark Brown, 2004, «The good lives model and conceptual issues in offender rehabilitation», Psychology, Crime and Law, 10(3): 243-57.Weaver, Beth and Fergus McNeill, 2011, «Travelling hopefully: desistance research and probation practice», in Brayford, Jo, Francis Cowe and John Deering (dir.), What else works? Creative work with offenders. Cullompton, Willian.Wilson, Chris, Andrew Bates and Birgit Völlm, 2010, «Circles of Support and Accountability: An Innovative Approach to Manage High-Risk Sex Offenders in the Community», The Open Criminology Journal, 3: 48-57. Wilson, Robin, Franca Cortoni and Monica Vermani, M, 2007, Cercles de soutien et de responsabilité: Reproduction à l’échelle nationale des résultats obtenus. Humber.Yates, Pamela, David Prescott and Tony Ward, 2010, Applying the good lives and self-regulation models to sex offender treatment. A Practical Guide for Clinicians. Brandon, The Safer Society.Zehr, Howard, 1990, Changing lenses. A new focus for crime and justice, Herald Press.Zehr, Howard, 2002, The little book of restorative justice, Good books.© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 37 Cahiers de la sécurité et de la justice – Hors-série La justice restaurative des mineurs en France : entre tendance maximaliste et minimalisteJessica FILIPPIa justice pénale des mineurs a été, pendant près de 70 ans, régie par l’ordonnance n° 45-174 du 2 février 1945 relative à l’enfance délinquante. Modifiée par une quarantaine de réformes, elle manquait de cohérence, de lisibilité et était difficilement applicable pour les professionnels de terrain. En 2008, la commission Varinard a proposé un projet de Code de justice pénale des mineurs. Puis, lors du gouvernement suivant, Christiane Taubira, puis Jean Jacques Urvoas ont travaillé sur un projet de codification (Projet de loi 2015,2016). En novembre 2017, Nicole Belloubet a relancé le sujet et, le 11 décembre 2018, l’Assemblée nationale a voté le projet de loi de programmation et de réforme de la justice (L’Épée-Boulanger, 2018). Ce texte a été amendé par le gouvernement pour y introduire un article l’habilitant à réformer par voie d’ordonnance le cadre légal de la justice des mineurs défini par l’ordonnance du 2 février 1945. La procédure d’ordonnance serait acceptable sous réserve que les modifications de fond ne viennent pas aggraver la situation pénale des adolescents concernés. Nombreuses ont été ensuite les contributions des juristes et des sociologues, portant sur les conséquences et les risques de cette réforme dans la prise en charge de la délinquance des mineurs (Couster, 2019 ; Januel, 2019 ; L’Épée-Boulanger, 2018). Parmi ces publications, aucune n’a fait état des articles L. 13-4 sur la justice restaurative et L. 112-8 2° sur la médiation (Code de la justice pénale, 2019). Suite à la récente introduction de la justice restaurative dans le régime de droit commun et l’article 10-1 du Code de procédure pénale, une circulaire relative à sa mise en Expérimentée tout d’abord dans le champ des majeurs à la maison centrale de Poissy en 2010 (Beauhaire et Davau, 2019) puis consacrée dans le cadre d’un régime de droit commun en 2014, la justice restaurative tarde encore à trouver sa place tant dans le système de justice pénale que dans les pratiques de la protection judiciaire de la jeunesse (milieu ouvert, Juvisy-sur-Orge, 2016). Si, la circulaire de mars 2017 a donné une place à la justice restaurative des mineurs, l’environnement judiciaire et le cadre légal limitent son développement. Il faut espérer que le Code de la justice pénale des mineurs adopté le 13 septembre 2019 offre à la justice restaurative, les moyens de s’épanouir.Mot clés : Justice restaurative, mineurs, France, Code de justice pénale des mineurs, réforme del’ordonnance, maximaliste, minimaliste, expérimentation.LJessica FILIPPIChercheuse en criminologie à l’École Nationale de Protection judiciaire de la jeunesse (É.N.P.J.J.) et chercheuse associée en criminologie au centre de recherche Pénalité, Sécurité et Déviances de l’Université Libre de Bruxelles ULB067.© Cahiers de la sécurité et de la justice - IHEMI38 I DOSSIER La justice restaurative des mineurs en France : entre tendance maximaliste et minimaliste – Jessica Filippioeuvre auprès des publics majeurs comme mineurs a été publiée le 15 mars 2017. En plus de préciser le cadre normatif de la justice restaurative, elle a pour objectifs de clarifier l’articulation entre le système de justice pénale et la justice restaurative en définissant ses principes et conditions de mise en œuvre. Dans le même temps, le modèle des mineurs et les transformations qu’il connaît conduisent la justice restaurative à se développer dans une certaine mesure (I). La manière dont elle a été légiférée semble chercher un équilibre entre tendances minimalistes et maximalistes. Ces tendances se constatent également dans les pratiques expérimentales des professionnels de la protection judiciaire de la jeunesse (II). Si la justice restaurative dépasse le champ des expérimentations et prend une place dans le modèle des mineurs, il faut lui offrir les moyens de se développer pour intégrer pleinement l’arsenal du juge (III).L’organisation de la justice pénale des mineurs un modèle protectionnel sous logique gestionnairePrincipes de la justice pénale des mineursLe 13 septembre 2019, le Code de la justice pénale applicable aux mineurs (C.J.P.M.) a été publié au Journal officiel pour entrer en vigueur le 31 mars 2021, en raison de la crise sanitaire 1. Dans ce code, il ne s’agit pas uniquement de textes relevant de la procédure pénale, mais aussi d’éléments de droit pénal et de modalités d’interventions à l’encontre des mineurs délinquants. Le code repose sur le même socle que l’ordonnance du 2 février 1945. Il est régi par trois principes : tout d’abord, celui de soutenir l’esprit de protection qui dirige l’intervention au bénéfice des mineurs et affirmer la primauté de l’éducatif sur le répressif (Legrand et Rouet, 2019), ensuite, la spécialisation des juridictions au travers des procédures adaptées et des professionnels qui travaillent auprès des mineurs et pour finir, l’atténuation de la responsabilité en fonction de l’âge. Le Code présente deux modifications majeures, celle d’une présomption de discernement, dès 13 ans 2 et le remplacement de la phase d’instruction (sauf pour les crimes et les délits graves) par la procédure « de mise à l’épreuve éducative ». Les services principaux qui viennent en appui à la justice des mineurs sont : l’Aide sociale à l’enfance (ASE) 3, les services publics de la protection judiciaire de la jeunesse (P.J.J.) et les secteurs associatifs habilités par l’État (S.A.H.). La Direction de la protection judiciaire de la jeunesse (D.P.J.J.) coordonne la justice des mineurs depuis le décret du 9 juillet 2008 relatif à l’organisation du ministère de la Justice. Elle est chargée « de l’ensemble des questions intéressant la justice des mineurs et de la concertation entre les institutions intervenant à ce titre ». La P.J.J. est essentiellement composée d’assistants sociaux, d’éducateurs, de psychologues. Elle a pour rôle d’assurer l’exécution des mandats des magistrats et de veiller à la réhabilitation des mineurs délinquants. Cependant par les réformes successives passées et actuelles, les principes susmentionnés s’effritent, du fait de l’accélération des temps judiciaires (Bastard et al., 2015) et la multiplication des modalités d’intervention.Prise en charge socio-éducative et temporalité de la protection judiciaire de la jeunesseCes vingt dernières années, face à la nécessité de répondre à la délinquance des mineurs, des changements se sont opérés dans le système de justice pénale des mineurs par le biais d’une logique d’efficacité. Se sont développés des moyens pour accélérer les procédures judiciaires mais également des pratiques s’éloignant de la pensée sociale et éducative (Sallée, 2016 : 84).Pour rendre la justice effective et efficace, il ne suffit plus de rendre une décision juridique équitable, mais de la rendre dans un délai raisonnable. Cela se traduit par une progression de la présence du parquet, qui se traduit par l’adoption d’une multitude de mesures alternatives aux poursuites, de procédures rapides, de comparution à délais rapprochés. Actuellement, la politique pénale des parquets repose sur une réponse pénale quasi-systématique. La circulaire du 15 octobre 1991 élargit le rôle du parquet et insiste sur la rapidité des réponses pénales. La circulaire du 2 octobre 1992 recommande la systématicité et la rapidité des réponses pénales. Celles du 8 juin 1998 et du 6 novembre 1998 visent à apporter une réponse systématique, rapide et lisible à chaque acte de délinquance, quel qu’il soit.(1) Article 25, Loi n°°2020-734 du 17 juin 2020 relative à diverses dispositions liées à la crise sanitaire, à d’autres mesures urgentes ainsi qu’au retrait du Royaume-Uni de l’Union européenne, J.O.R.F., 18 juin 2020.(2) Article L.11-1 C.J.P.M.(3) L’ASE est chargée de la protection et de l’éducation des mineurs en dangers.© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 39 Cahiers de la sécurité et de la justice – Hors-série L’accélération de la réponse pénale est également visible au niveau du jugement des mineurs (Bitton et Rosenczveig, 2011 : 24-31). Si récemment le Code de la justice pénale des mineurs a souhaité « simplifier » la procédure applicable aux jeunes délinquants en instituant un mode de poursuite unique, cette volonté suppose une accélération des temps judiciaires (Bastard et al., 2015). Une volonté qui n’en demeure pas moins ancienne et dont l’instauration a été progressive (le projet de la loi LOPPSI 2 4, la comparution à délai rapproché 5, le jugement à délai rapproché 6, la présentation immédiate devant la juridiction pour mineur 7, la convocation par officier de police judiciaire aux fins de jugement devant le tribunal pour enfants 8). Aujourd’hui, cette accélération se poursuit avec la procédure « de mise à l’épreuve éducative » passant d’un délai de jugement de 18 mois à 12 mois séquencés en trois temps : un jugement de culpabilité au bout de 3 mois, suivi d’une mise à l’épreuve de 9 mois maximum et à l’issue de celle-ci, un jugement sur la sanction.La substitution du temps long de l’action judiciaire par une réponse immédiate et une injonction d’intervention éducative plus courte a des conséquences dans l’organisation des services de la P.J.J. et la prise en charge des mineurs (Vicour, 2006 : 425-455). Une véritable culture organisationnelle se développe, de nombreux dispositifs se déploient autour d’un système du contrôle (augmentation du nombre de mesures coercitives et de surveillances, création de vingt Centres éducatifs fermés supplémentaires). Une pensée néolibérale de la justice prend forme, rompant avec la philosophie de la réhabilitation (Bailleau, 2008 : 418).Si des pratiques restauratives ont existé avant la « mesure de justice restaurative » de 2014, leurs mises en œuvre actuelles sont colorées par les réformes judiciaires successives et les politiques gestionnaires. De la sorte, nombreux sont les services contraints de suivre une cadence rapide dans la mise en œuvre des mesures éducatives. L’approche des éducateurs vis-à-vis des mineurs ainsi que leurs prises en charge sont transformées. Plus particulièrement, la mesure de réparation pénale, si elle s’inscrit dans le courant maximaliste 9 de la justice restaurative par son objectif de réparer, son approche est détournée du fait des réformes successives, des temporalités courtes d’intervention et des politiques gestionnaires. Ainsi, l’auteur mineur d’une infraction est appréhendé selon l’acte commis et, accompagné de son éducateur, effectue une activité de « réparation ». Il arrive que pour une même catégorie d’infraction, plusieurs auteurs se retrouvent dans une même activité. La réparation est davantage centrée sur l’infraction que sur ses conséquences et répercussions. De plus, elle ne comprend pas l’ensemble des dimensions du préjudice causé par l’infraction : le préjudice matériel, les souffrances psychologiques et relationnelles causées à la victime, et à la communauté par le délit commis et le préjudice social que l’auteur se cause à lui-même (Filippi, 2015). Si la mesure de réparation pénale peut s’inscrire théoriquement dans le courant maximaliste de la justice restaurative, sa pratique interroge cette appartenance.Une consécration légale de la justice restaurative dans la justice des mineursDéfinitions et courants de la justice restaurativeIl est erroné de penser qu’il existe une définition (Lemonne, 2002 : 413) ou une approche unique (Miers, 2007 : 447-467) de la justice restaurative. Nommée sous (4) Le Conseil constitutionnel a censuré une partie de la loi « ne garantissait pas que le tribunal aurait disposé d’informations récentes sur la personnalité du mineur » qui lui permettrait de veiller à son « relèvement éducatif et moral ». V. Déc. n° 2011- 625 DC, Cons 32 à 34, AJDA 2011. 1097.(5) Loi n° 95-125 du 8 février 1995 relative à l’organisation des juridictions et à la procédure civile pénale et administrative, J.O.R.F., 9 février 1995 ; Loi n° 96-585 du 1er juillet 1996 portant modification de l’ordonnance n° 45-174 du 2 février 1945 relative à l’enfance délinquante, J.O.R.F., 2 juillet 1996. La loi n° 96-585, dite loi Toubon introduit les procédures de convocation par officier de police judiciaire aux fins de jugement et de comparution à délai rapproché. Loi n° 2011-1940 du 26 décembre 2011 visant à instaurer un service citoyen pour les mineurs délinquants.(6) Loi n° 2002-1138 du 9 septembre 2002, d’orientation et de programmation pour la justice, J.O.R.F., 10 septembre 2002.(7) La loi n° 2007-297 du 5 mars 2007 relative à la prévention de la délinquance, J.O.R.F., 6 mars 2007« […] a modifié le deuxième alinéa de l’article 5 de l’ordonnance du 2 février 1945 afin de donner à la procédure de « jugement à délai rapproché » la dénomination de « présentation immédiate devant la juridiction pour mineurs » ; que ce changement terminologique n’appelle par lui-même aucune critique de constitutionnalité » ; V. Décision n° 2007-553 DC du 03 mars 2007, J.O.R.F., 7 mars 2007, p. 4356 ; Loi n° 2011-939 du 10 août 2011 sur la participation des citoyens au fonctionnement de la justice pénale et le jugement des mineurs, J.O.R.F., 11 août 2011.(8) Il s’agit d’une sorte de citation directe supprimant le recours à l’instruction par le juge des enfants « La première circulaire d’application de la loi du 10 Août 2011 », dans Dalloz Actualité, Le quotidien du droit, le 14 septembre 2011(9) Voir section II.A pour une discussion des conceptions maximaliste et minimaliste.© Cahiers de la sécurité et de la justice - IHEMI40 I DOSSIER La justice restaurative des mineurs en France : entre tendance maximaliste et minimaliste – Jessica Filippiune multitude de références en francophonie, « justice restaurative » pour la France, « justice restauratrice » pour les Belges, « justice réparatrice » pour les Suisses et les Canadiens à laquelle s’ajoute même la dénomination de « justice transformatrice », la justice restaurative connaît en France ces dernières années une montée en popularité (expérimentations auprès de publics majeurs et mineurs, médiatisation, consécration légale). Si la justice restaurative est sujette à de nombreuses définitions, cette flexibilité amène des discussions sur ce qui relève ou non de la justice restaurative en France.Dans la littérature scientifique, les définitions de la justice restaurative s’articulent autour de son processus, de sa finalité ou encore des deux à la fois.Selon Tony Marshall « la justice restaurative est un processus par lequel toutes les parties impliquées dans une infraction spécifique se réunissent pour résoudre collectivement la façon de faire face aux séquelles de l’infraction et à ses implications pour l’avenir » (1996, 21, 1999, 36). Définis en fonction de leur processus, la négociation ou le consensualisme deviennent des critères de la justice restaurative, une approche qui appartient au mouvement minimaliste ou puriste. Une conception qui accorde peu d’importance à la finalité de la justice restaurative, à savoir la réparation voire la restauration des torts subis. Pour Bazemore et Walgrave, défenseurs de l’approche maximaliste, la justice restaurative ne peut se réduire à son processus et vise la restauration des participants touchés par l’infraction. C’est une « manière de faire justice, orientée prioritairement vers la restauration des souffrances et dommages causés par un délit » (Bazemore et Walgrave 1999). Même si elle peut ne concerner que l’auteur, il s’agit d’une approche qui privilégie, dans un cadre imposé formel ou informel, ou lors d’un processus de négociation, la réparation des préjudices vécus à l’occasion d’une infraction.Pour Howard Zehr « la justice restaurative est un processus visant à impliquer, dans la mesure du possible, ceux qui ont un intérêt dans une infraction spécifique pour identifier collectivement et traiter les préjudices, les besoins et les obligations afin de guérir et de redresser au mieux la situation » (Zehr, 2002 : 37, 2012 : 98). Ainsi, au-delà des définitions, les sources de tensions entre ces deux modèles s’articulent autour du cadre d’intervention de la justice restaurative et sa place au sein du système de justice pénale. Les puristes estiment que la justice restaurative et la contrainte sont opposées et ne peuvent coexister puisque cela entraverait le processus de restauration des participants. Les maximalistes quant à eux, voient un rôle restaurateur dans la punition et la contrainte. En effet, leur approche « suppose la mise en œuvre d’obligations coercitives judiciaires en vue d’une réparation (partielle) visée ». (Walgrave et Zinsstag, 2014 : 33). Soutenant que la contrainte peut faire partie de la justice restaurative, Walgrave (2000 : 422) déclare : « Nous [les maximalistes] acceptons la contrainte lorsque le caractère volontaire n’est pas atteint et quand il est jugé nécessaire de réagir à l’infraction ». Ainsi, les sanctions qui ont une visée réparatrice, les sanctions restauratives, font partie de la justice restaurative (Van Fraechem et Walgrave, 2006).Pour finir, les courants maximaliste et puriste de la justice restaurative s’opposent dans leur institutionnalisation au sein du système de justice pénale. D’un côté, il y a les théoriciens puristes qui considèrent que la justice restaurative peut se développer au sein du système de justice pénale sans affaiblir les règles de procédures. L’institutionnalisation de la justice restaurative est alors vue comme extérieure au système de justice pénale (London, 2003 : 177). De l’autre côté, les maximalistes estiment que la justice restaurative peut s’aligner avec le système de justice pénale, l’intégrer et le transformer. Ils critiquent les défenseurs de l’approche puriste qui restreint la justice restaurative : la justice pénale traite les situations quand la rencontre volontaire n’est pas faisable (Dignan, 2002, 2003 ; Von Hirsch, Ashworth, et Shearing, 2003).L’inscription juridique de la justice restaurative dans la procédure pénale applicable aux mineursDepuis le 1er octobre 2014, la loi n° 2014-896 du 15 août 2014, relative à l’individualisation des peines et renforçant l’efficacité des sanctions pénales est entrée en vigueur (Loi, 2014). Les mesures de justice restaurative ont ainsi trouvé une place dans l’article 10-1 du Code de procédure pénale. L’adoption de ce texte de loi est le témoignage d’une volonté générale de sortir des réponses pénales habituelles accordées, souvent jugées comme peu effectives, peu participatives, peu responsabilisantes, voire trop punitives, et prenant peu en considération la victime. Les mesures adoptées s’inscrivent dans le mouvement maximaliste de la justice restaurative. Les mesures sont présentées au travers de leur finalité, orientée vers la réparation des conséquences du délit : « constitue une mesure de justice restaurative toute mesure permettant […] de participer activement à la résolution des difficultés résultant de l’infraction, et notamment à la réparation des préjudices de toute nature résultant de sa commission ».Trois ans après la promulgation de cette loi, la circulaire du 15 mars 2017, concernant la mise en œuvre de ces mesures ad hoc à tous les stades de la procédure judiciaire, © Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 41 Cahiers de la sécurité et de la justice – Hors-série est entrée en vigueur. Intitulée « mise en œuvre de la justice restaurative applicable immédiatement suite aux articles 10-1 […] », la circulaire a pour objectif d’expliquer et de développer comment la justice restaurative peut être appliquée dans les pratiques professionnelles (magistrat, avocat, éducateur, travailleur social et psychologue) et auprès des majeurs et des mineurs. Dans les premières pages, la justice restaurative est présentée de nouveau selon le courant maximaliste théorisé par Lode Walgrave. Elle est « un processus complémentaire du procès pénal qui consiste à restaurer le lien social […]. Elle est conçue pour appréhender l’ensemble des répercussions personnelles, familiales et sociales liées à la commission des faits […] ». Cette définition associe à la justice restaurative les mesures judiciaires éducatives telles que la médiation pénale et la réparation pénale.Si la justice restaurative est présentée selon sa finalité, dans la circulaire, elle est également considérée au sein du système de justice pénale selon son processus « volontaire et confidentiel ». De plus, la circulaire annonce que l’autorité judiciaire exerce un contrôle d’opportunité de la mesure. Ce contrôle interroge cependant l’incidence qu’il peut avoir sur la procédure et les décisions judiciaires. Le contrôle du magistrat s’organise a priori seulement sur les éléments de légalité de la mesure (information des participants à la justice restaurative, reconnaissance des faits, consentement à participer à la mesure). Ensuite, la circulaire dispose que la mesure doit être confidentielle, et que cette confidentialité ne peut être rompue qu’à deux conditions : « si les parties le souhaitent et excepté les cas où un intérêt supérieur lié à la nécessité de prévenir ou de réprimer des infractions [le] justifie ». Par cette confidentialité, « l’autonomie de la justice restaurative par rapport à la procédure pénale est assurée ». Au regard de ces éléments, on peut affirmer que les mesures de justice restaurative sont alors inscrites « à la marge » du système de justice pénale. Quand bien même leur complémentarité est évoquée, leur autonomie est tout autant valorisée. La justice restaurative laisse le système de justice pénale hors de la discussion et se limite à une complémentarité dans ses effets (Walgrave, 1999 : 12). La circulaire adopte donc une approche minimaliste en avançant que le « caractère innovant de cette mesure repose sur son autonomie vis-à-vis de la procédure pénale. Elle constitue une voie offerte aux parties, facultative, et sans conséquence sur le déroulement de la procédure judiciaire qui s’exerce en parallèle » (Circulaire, 2017). Le texte confirme l’« imperméabilité » de la mesure de justice restaurative et rappelle son autonomie en ce qu’elle n’a « pas d’incidence sur la décision d’engager des poursuites ou de classer » ni n’a « d’effet sur l’octroi d’éventuels dommages-intérêts dus à la partie civile ». Sa place au sein de la procédure pénale est périphérique. Ce choix d’organisation de la justice restaurative selon l’approche minimaliste s’expliquerait par une volonté de se détacher des temporalités judiciaires et des modalités d’intervention de la justice pénale des mineurs (comme la réparation pénale) qui affecteraient le processus.Si le discours de la circulaire s’inscrivait dans une approche maximaliste, la procédure pénale des mineurs pourrait, tout en préservant la confidentialité des échanges, envisager l’intégration de la justice restaurative en son sein. Alors que dans la circulaire, le magistrat à l’origine de la proposition de la mesure de justice restaurative observe un contrôle de légalité vis-à-vis de la mesure, il serait peut-être nécessaire, dans l’approche maximaliste, de l’informer de son déroulement. Une vigilance pourrait être exercée concernant l’intérêt du renvoi au magistrat. La philosophie de la justice des mineurs, dans une volonté protectionnelle éducative et de responsabilisation, pourrait envisager l’établissement d’un rapport succinct en justice restaurative sur les engagements des participants et sa transmission au magistrat. Il veillerait à ce que ces derniers n’aillent pas à l’encontre de l’intérêt du mineur (protection) et assurent son relèvement éducatif (éducation/responsabilisation) 10. Le magistrat en charge du contrôle de légalité (Circulaire 2017) poursuivrait ce rôle à la fin de la mesure de justice restaurative. Ayant connaissance de la participation volontaire d’un mineur à une mesure, par exemple si celui-ci fait part au cours de son audience, de sa participation à un tel processus, il est légitime de penser que cette information l’influencerait. L’idée serait que la participation ou non à un processus de justice restaurative ne doit pas jouer en défaveur du mineur (buts d’éducation et de protection). Dans cette perspective maximaliste, la justice restaurative prolongerait l’objectif réparateur jusque dans la réaction judiciaire. Aussi, si l’approche maximaliste est consacrée dans les premières lignes de l’article 10-1 du Code de procédure pénale ainsi que dans les éléments d’introduction de la circulaire, elle mérite de s’étendre à tout le système de justice pénale des mineurs. Cependant, il demeure un risque : que cette considération conduise à « coller » aux temporalités du judiciaire. Enfin, les mesures de justice restaurative peuvent être mises en œuvre à toutes les étapes de la procédure judiciaire y compris lorsqu’une peine est exécutée. Les exemples de mesures de justice restaurative présentés à la fin de la circulaire témoignent, là encore, d’une approche maximaliste. Ainsi sont présentées, les rencontres condamnés-victimes (R.C.V.) et les rencontres (10) Se pose également la question du suivi de la mesure de justice restaurative et que faut-il faire si celle-ci n’est pas effectuée.© Cahiers de la sécurité et de la justice - IHEMI42 I DOSSIER La justice restaurative des mineurs en France : entre tendance maximaliste et minimaliste – Jessica Filippidétenus-victimes (R.D.V.), la médiation restaurative et la conférence restaurative. La circulaire ouvre la perspective de la justice restaurative vers d’autres mesures dont le cercle restauratif, les cercles de soutien et de responsabilité (C.S.R.), ou les cercles d’accompagnement et de ressources (C.A.R.) 11. Les mesures ne sont donc pas envisagées seulement comme des rencontres ou des échanges entre victimes et délinquants (Walgrave, 1999 : 13), mais aussi, comme des processus auxquels auteurs et communauté (Filippi, 2018) peuvent participer 12.Les expérimentations de la justice restaurative à l’égard des mineurs 13Les expérimentations conduites depuis 2016 à la P.J.J. ont abouti à peu de médiations directes entre auteurs et victimes. Parmi les pratiques citées dans la circulaire, les médiations restauratives indirectes, les conférences restauratives ou encore les groupes de paroles sont mis en œuvre auprès des mineurs. Le déroulement de ces mesures est lui aussi variable. Selon les services et les partenariats, la justice restaurative peut soit s’inscrire dans une mesure libre sans forme prédéfinie et évoluant au gré des participants, soit dans une standardisation de la pratique (une mesure identifiée, avec un nombre donné d’entretiens de préparation et de rencontres). Depuis 2018 et jusqu’en 2020 neuf directions interrégionales (D.I.R.) de la protection judiciaire de la jeunesse expérimentent la justice restaurative à destination des mineurs. L’établissement des conventions de partenariat avec les différents acteurs du système de justice pénale permet de définir les rôles et les places de chacun dans l’orientation et la mise en œuvre de la justice restaurative. Si certaines DIR ont des conventions avec différents partenaires (D.P.J.J., Éducation nationale, S.P.I.P. 14, S.A.V. 15, barreaux), d’autres n’en ont pas encore signé. En outre, les mesures de justice restaurative demeurent encore méconnues des professionnels de justice (forces de l’ordre, avocats, magistrats).À la P.J.J., le développement de la justice restaurative s’inscrit « à la marge » et en parallèle de la procédure pénale. À la marge, puisque sa mise en œuvre n’est envisagée qu’à la condition que le mineur fasse l’objet d’une mesure au pénal. Ainsi, seuls les mineurs inscrits dans le système de justice pénale avec une prise en charge éducative accèdent à la mesure de justice restaurative. Les mineurs suivis au civil ou dont l’affaire aboutit à un classement sans suite dans une procédure pénale ne pourraient pas se voir proposer cette mesure. Cette proposition de la justice restaurative dans un cadre contraint s’inscrit dans un contexte où ce type de justice n’existe pas en tant que mesure unique. Cette modalité de proposition contribue aussi à répondre aux besoins de respecter la confidentialité de la justice restaurative et son imperméabilité par rapport à la procédure judiciaire. Cette approche semble par ailleurs être liée à la perception que les magistrats ont de la justice restaurative en ce qu’elle prend du temps, qu’elle les dépossède de leur autorité et qu’elle n’est pas une solution de remplacement à effet rapide. De surcroît, si une mesure éducative est effectuée en même temps, nul besoin de connaître la mesure de justice restaurative : les rapports de la mesure éducative éclaireront le magistrat sur l’évolution et l’implication du mineur.La culture professionnelle (éducation, protection) et l’environnement de travail (rationalité managériale, temps court d’intervention et disponibilités des services locaux) limitent l’appropriation et le développement de la justice restaurative par les acteurs de justice et les justiciables. En ce qui concerne l’information et l’orientation de la mesure, il est observé sur les neuf D.I.R., que les forces de l’ordre ne pensent pas à informer les justiciables de la possibilité de participer à la justice restaurative. Pour les magistrats du parquet et du siège, des pratiques hétéroclites s’observent. Certains informent et orientent à tout va la justice restaurative et laissent une large appropriation de la mesure par les services (P.J.J., S.A.H., S.A.V., ou service dédié), d’autres contrôlent les dossiers pour lesquels la justice restaurative est la plus pertinente, d’autres encore proposent et imposent une forme et pour finir, dans une moindre mesure, certains bloquent toute (11) Ces derniers mobilisent deux cercles autour de l’auteur (membre principal) à la fin de sa peine ou avant sa libération, en soutien à un processus de réinsertion. Ces cercles incluent des membres de la communauté (bénévoles et professionnels) et des coordinateurs qui aident à établir des liens entre chacun des membres.(12) Systèmes de réparation communautaire, commissions de réparation, dédommagement direct des victimes, systèmes de soutien aux victimes et témoins, cercles de soutien aux victimes, cours de sensibilisation aux victimes, éducation des prisonniers ou des délinquants, tribunaux de résolution de problèmes, C.S.R., cérémonies de réinsertion des délinquants, et les projets impliquant des délinquants et leurs familles ou d’autres victimes d’actes criminels.(13) L’ensemble des données récolté (2016-2019) et présenté dans cette section est une synthèse issue de groupes de travail entre les professionnels de la P.J.J., les différentes institutions de formations et d’accompagnement des pratiques : Association de recherche de criminologie appliquée (A.R.C.A.), Citoyen et Justice (C.J.), École nationale de Protection judiciaire de la jeunesse (É.N.P.J.J.), Institut français pour la justice restaurative (I.F.J.R.)(14) Service pénitentiaire d’insertion et de probation(15) Service d’aide aux victimes.© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 43 Cahiers de la sécurité et de la justice – Hors-série mesure de justice restaurative plus spécifiquement pour la phase pré-sentencielle (risque de conflit d’intérêts, secret de l’instruction, interférences entre les participants).Dans la mise en œuvre de la justice restaurative par les services, on observe une multitude de pratiques : soit la justice restaurative se déroule au sein d’un service identifié (P.J.J. ou S.A.V.), soit elle est mise en œuvre dans un lieu se détachant de la connotation judiciaire (services dédiés à la justice restaurative ou mairie par exemple). Elle est animée soit par un binôme de professionnels formés, issus de professions différentes (S.P.I.P., S.A.V., S.A.H., P.J.J. et service de justice restaurative) soit par un binôme de professionnels d’un même service. Quand bien même certains viendraient d’horizons différentes, ils co-animent et préparent ensemble les justiciables à participer à la justice restaurative. Pour d’autres binômes professionnels, l’information et la préparation se déroulent « chacun de leur côté ». Un professionnel formé s’occupe de l’information et de la préparation de l’auteur, de ses représentants légaux, voire de ses proches, tandis qu’un autre est en charge dans un autre lieu, de l’information et de la préparation de la victime.Dans les services de la P.J.J., lors de la prise en charge éducative, les éducateurs informent les mineurs de leur droit de participer à la justice restaurative. Ils repèrent les situations susceptibles de partir en justice restaurative (infractions, sur l’un des territoires l’âge minimum pour bénéficier d’une mesure de justice restaurative est fixé à 16 ans) et transmettent les dossiers aux professionnels de justice restaurative. Sur d’autres territoires, les éducateurs formés et en binômes proposent et mettent en œuvre la justice restaurative pour les mineurs (dont ils n’ont pas le suivi éducatif). Cependant, l’administration de la justice pénale influence également l’appropriation de la mesure par ces services. En effet, l’accélération des temps judiciaires, les logiques managériales des services et les charges de travail des professionnels de la P.J.J. les amènent soit à mettre en œuvre la justice restaurative a minima (car chronophage) soit à la déléguer à d’autres services. À l’inverse, certains professionnels disposent d’une décharge de travail qui leur permet de se consacrer à la mise en oeuvre de la justice restaurative, soit au sein du service éducatif où ils exercent, soit dans l’établissement d’un des partenaires.Nous l’aurons compris, « la justice restaurative est loin d’être un ensemble complet de pratiques basées sur une théorie juridique toute faite. C’est plutôt un mouvement et un terrain d’expérimentation et de recherche reposant sur un idéal intuitif de justice dans un idéal vague de société » (Walgrave, 1999 : 19). Dans le C.J.P.M., plusieurs places sont accordées à la justice restaurative. Elle est située à l’extérieur (L. 13-4) et à l’intérieur (L. 112-8, 2°) du système de justice pénale. Si la partie législative du Code est stabilisée, la recommandation européenne publiée en octobre 2018 peut donner un éclairage quant à la partie réglementaire qui accompagnera ces textes. Dans cette contribution, l’analyse se centrera sur l’article L. 13-4 du code 16.La justice restaurative dans le Code de la justice pénale des mineurs à la lumière de la recommandation CM/Rec (2018)8 du comité des ministres du Conseil de l’EuropeSi jusqu’à présent il n’existait pas d’article dédié à la justice restaurative dans l’ordonnance du 2 février 1945, l’article L. 13-4 du C.J.P.M. propose la justice restaurative aux victimes et auteurs mineurs. Cet article répond dans une certaine mesure, aux directives et recommandations européennes invitant les États à intégrer la justice restaurative dans leur législation interne. En tout début du code, dans le titre unique, la justice restaurative à destination des mineurs y trouve une place, démontrant ainsi l’importance accordée à cette philosophie dans le modèle protectionnel des mineurs. Selon le texte, « il peut être proposé à la victime et à l’auteur de l’infraction de recourir à la justice restaurative, conformément à l’article 10-1 du Code de procédure pénale, à l’occasion de toute procédure concernant un mineur et à tous les stades de celle-ci, y compris lors de l’exécution de la peine, sous réserve que les faits aient été reconnus. La justice restaurative ne peut être mise en œuvre que si le degré de maturité et la capacité de discernement du mineur le permettent, et après avoir recueilli le consentement des représentants légaux ». Cette formulation est une adaptation pour les mineurs de l’article 10-1 du Code de procédure pénale. La Recommandation européenne publiée en octobre 2018 (CM/Rec (2018)8) relative à la justice restaurative (16) L’article L. 112-8, 2° sera l’objet d’une analyse dans le cadre d’un ouvrage collectif. V. Filippi Jessica « Le parcours de la justice restaurative en droit pénal des mineurs », La réforme de la justice pénale des mineurs : quelles spécificité(s) ? Dalloz, T&C, 2021. Cette contribution abordera l’intérêt que présente l’approche intégrée d’une justice restaurative pour les mineurs à tous les stades de la procédure judiciaire. Ce chapitre sera aussi l’occasion d’évoquer les adaptations à adopter dans la rédaction de la partie réglementaire du C.J.P.M.© Cahiers de la sécurité et de la justice - IHEMI44 I DOSSIER La justice restaurative des mineurs en France : entre tendance maximaliste et minimaliste – Jessica Filippien matière pénale peut nous donner un éclairage sur les orientations pratiques à envisager dans le C.J.P.M.Dans son article 1er, la recommandation « a pour but d’encourager les États membres à développer la justice restaurative et à y recourir au sein de leur système de justice pénale. Elle met en avant des normes relatives à l’utilisation de la justice restaurative dans le cadre de la procédure pénale et cherche à défendre les droits des participants, ainsi qu’à optimiser l’efficacité du processus pour répondre à leurs besoins. » (CM/Rec (2018)8, article 1). La recommandation confirme la volonté de développer une approche maximaliste de la justice restaurative. L’article 8 met en avant les différentes formes de justice restaurative : « Les pratiques qui ne prévoient pas de dialogue entre les victimes et les auteurs d’infractions peuvent avoir un caractère réparateur si elles se conforment étroitement aux principes fondamentaux de la justice restaurative […]. Les principes et approches restauratives peuvent aussi être appliqués au sein du système de justice pénale, en dehors d’une procédure pénale. » (CM/Rec (2018)8, article 8). Cette formulation reprend les propos théoriques et empiriques des chercheurs criminologues en justice restaurative (Walgrave, 1994, 1995 ; Bazemore et Umbreit, 1995 ; Wright, 1996, Van Ness et Heetderks Strong, 1997 ; Bazemore et Walgrave, 1998). L’article 59 confirme cette approche maximaliste de la justice restaurative, invitant à en étendre la dimension pratique : « Alors que la justice restaurative est typiquement caractérisée par un dialogue entre les parties, de nombreuses interventions qui n’incluent pas de dialogue entre la victime et l’auteur de l’infraction peuvent être conçues et utilisées, en appliquant les principes de la justice restaurative. » (CM/Rec (2018)8, article 59).À la lumière de cette recommandation, l’article L. 13-4 du C.J.P.M. intéressant la justice restaurative pourrait considérer que les victimes d’une infraction, les auteurs ou les responsables d’un préjudice peuvent faire la demande ou se voir proposer de participer à la justice restaurative, conformément à l’article 10-1 du Code de procédure pénale.Cette approche confirmerait d’une part que la justice restaurative est un droit pour tous, une manière d’accéder à la justice et d’y participer activement, et d’autre part soulignerait que l’infraction a des répercussions non seulement sur sa victime directe, mais aussi sur les membres d’une communauté et toute autre victime indirecte. Les participants seraient perçus d’une manière plus élargie. Les policiers, les enseignants, les travailleurs sociaux, les proches des auteurs, des victimes et les représentants de la « communauté » pourraient être inclus parmi les professionnels et les publics concernés. Cette lecture permettrait d’envisager toutes les formes de la justice restaurative et non seulement les rencontres entre condamnés et victimes ou entre détenus et victimes, la médiation directe ou indirecte et la concertation restaurative en groupe. L’inclusion d’un groupe plus large peut aider à éduquer, à renforcer le capital social et à développer d’autres capacités de l’auteur susceptibles de contribuer à la prévention ou à la gestion de la criminalité et des conflits à l’avenir.La notion de préjudice, même si elle peut renvoyer « au droit civil de l’indemnisation » (Cario, 2019 : 87), contribue à nous détacher de l’infraction. Si la circulaire du 15 mars 2017, n’envisage ni la résolution des conséquences de l’infraction ni ses modalités de réparation, la notion de préjudice pourrait y remédier dans une certaine mesure. Si cette possibilité était offerte, la justice restaurative pourrait être considérée à la fois en tant que mesure parallèle au système de justice pénale, indépendante de ses résultats, mais aussi en tant que mesure intégrée au système de justice pénale pouvant avoir des effets sur la réparation du préjudice et des suites à donner à l’affaire 17. La justice restaurative serait mise en oeuvre à tous les stades de la procédure judiciaire, pour tout type d’infraction et à toute personne qui en exprime le besoin sous réserve d’une reconnaissance minimale des faits (Circulaire, 2017). Le recours à la justice restaurative ne nécessite pas une décision judiciaire quant à la culpabilité ; un aveu de responsabilité ne présuppose ni ne nécessite la constatation d’une culpabilité légale. Pour finir, l’évaluation du discernement ou de la maturité n’est pas une condition de la proposition de la mesure mais seulement de sa mise en œuvre, et cette évaluation appartient aux professionnels mettant en œuvre la justice restaurative.Un mineur avec des capacités moindres pourrait participer à la justice restaurative puisqu’en principe, elle s’adapte aux besoins des individus. Si un mineur est en difficulté et s’il le souhaite, il peut être représenté et accompagné. En outre, l’établissement d’un âge d’une présomption de discernement à 13 ans signifierait qu’en matière pénale, un mineur en dessous de cet âge ne pourrait pas bénéficier de justice restaurative. Si tel était le cas, la justice restaurative doit pouvoir être présentée et proposée dans d’autres instances que la sphère judiciaire au pénal. Des pratiques (17) Cette proposition contribuerait d’une part à intégrer l’article L. 112-8 dans la justice restaurative. D’autre part, elle développerait la médiation à tous les stades de la procédure judiciaire y compris au niveau du parquet. À ce jour, la médiation est envisagée dans le cadre d’une mesure éducative judiciaire. Dans la continuité de la circulaire du 15 mars 2017, il est vraisemblable que les dommages et intérêts ne soient pas considérés dans la résolution du conflit contrairement aux travaux de réforme amorcés en 2015 (Projet de loi 2015, 2016).© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 45 Cahiers de la sécurité et de la justice – Hors-série de justice restaurative trouvent leur place dans des établissements scolaires du primaire et du secondaire pour des altercations, des faits de violence et des vols. Aussi, si le conflit, quand bien même de nature infractionnelle, relève d’un traitement de la justice par le civil, la justice restaurative pourrait aussi s’appliquer dans cet espace.Ces propositions invitent donc à considérer l’article L. 13-4, comme une introduction à la justice restaurative dans le C.J.P.M pour envisager ensuite ses déclinaisons possibles à tous les stades de la procédure judiciaire. Mais à ce jour, il transparait que deux mouvements de la justice restaurative « s’opposent » au sein du C.J.P.M, un maximaliste et un minimaliste. ConclusionLes avancées législatives en France quant à la consécration de la justice restaurative en matière pénale prennent au sérieux cette approche. Le chemin reste encore à parcourir pour que la justice restaurative trouve une place pleine et entière au sein du système de justice pénale. Les réformes multiples et en cours, l’accélération du temps judiciaire, les politiques gestionnaires des services, la culture professionnelle (éducation, protection) sont quelques obstacles au développement des principes de la justice restaurative. En novembre 2019, les 30 ans de la Convention internationale des droits de l’enfant étaient célébrés. La justice restaurative, comme processus d’empowerment (O’Mahony et Doak 2017), permet le plein développement des droits capacitaires des enfants (Filippi, 2018). S’il est vrai que dans les effets, la justice restaurative et la justice pénale sont complémentaires, il est nécessaire de songer à une complémentarité des deux modèles au cœur de la réponse socio-judiciaire. Elle aurait d’autant plus de sens pour les mineurs auteurs d’infraction. Une telle combinaison serait davantage respectueuse de leurs droits, de leur personne en cours de construction. En plus, d’un cadre légal, des espaces de formations, d’échanges, de recherches (ARCA, C.J., É.N.P.J.J., I.F.J.R.) et d’accompagnement (supervision) des professionnels sont indispensables pour installer ces pratiques professionnelles innovantes dans le paysage judiciaire nRéférencesOuvrages et articles Beauhaire, Alexandra et Inès Davau, 2019, « Retour sur la conférence «justice restaurative» », INHESJ / ONDRP / A2S.Bailleau, Francis, 2008, « L’exceptionnalité française. Les raisons et les conditions de la disparition programmée de l’ordonnance pénale du 2 février 1945 », dans Droit et Société, n° 69-70, p. 399-438.Bastard, Benoit, David Delvaux, Christian Mouhanna et Frédéric Schoenaers, 2015, « Vitesse ou précipitation ? La question du temps dans le traitement des affaires pénales en France et en Belgique », dans Droit et société, n°90, p. 271-286Bazemore, Gordon et Lode Walgrave (dir.), 1999, Restorative Juvenile Justice : Repairing the Harm of Youth Crime, Monsey, Criminal Justice Press, 408 p.Bazemore, Gordon et Mark Umbreit, 1995, “Rethinking the sanctioning function in Juvenile Court : Retributive or restorative responses to youth Crime”, Crime and delinquency, vol. 41, p. 296-316.Bazemore, Gordon et Lode Walgrave, 1998, Exploring restorative justice for juveniles, Monsey, Criminal justice press.Bitton, Josine et Jean-Pierre Rosenczveig, 2011, « Quelle justice pour les mineurs. La justice au tribunal pour enfants de Bobigny : modernité ou régression ? », dans la Revue Après-Demain, n°19, p. 24-31 Cario, Robert, 2004, La justice restaurative, Paris, L’Harmattan.Cario, Robert, 2007, « La justice restaurative : vers un nouveau modèle de justice pénale ? ». AJ Pénal, Dalloz, p. 372-375.Cario, Robert, 2010, Justice restaurative, Principes et promesses, Paris, L’Harmattan, 2e éd.Cario, Robert, 2019, « Les apports de la recommandation (18)8 du Conseil de l’Europe dans la consolidation de la justice restaurative en France », AJ Pénal, Dalloz, 87.Christie, Nils, 2005, Au bout de nos peines. Bruxelles, Ed. Larcier.Coignac, Anaïs, 2019, « Justice restaurative un dispositif encore trop peu utilisé », Dalloz actualité, le Quotidien du droit [En ligne] Accessible à l’adresse https://www.dalloz-actualite.fr/dossier/justice-restaurative-reparation-yeux-dans-yeux#.XRS_cVJG7j8 [Consulté le 29 juillet 2019].© Cahiers de la sécurité et de la justice - IHEMI46 I DOSSIER La justice restaurative des mineurs en France : entre tendance maximaliste et minimaliste – Jessica FilippiCouster, Thomas, 2019, « Réforme de la justice des mineurs : la Chancellerie lance une consultation de trois mois », Dalloz actualité, le Quotidien du droit [En ligne], Accessible à l’adresse https://www.dalloz-actualite.fr/flash/reforme-de-justice-des-mineurs-chancellerie-lance-une-consultation-de-trois-mois#.X4mWmdUzbIU [Consulté le 16 octobre 2020].Cunneen, Chris et Carolyn Hoyle, 2010, Debating restorative justice, Oxford, Hart Publishing.Dachy, Aurore, 2013, L’application de la concertation restauratrice en groupe en Fédération Wallonie Bruxelles, Ministre de la Jeunesse et de l’Aide à la Jeunesse de la Fédération Wallonie-Bruxelles, [En ligne] Accessible à l’adresse https://www.huytebroeck.be/IMG/pdf/Rapport_CRG_avril_2013.pdf [Consulté le 29 juillet 2019].Dignan, Jim, 2002, Restorative justice and the law : the case for an integrated, systemic approach, dans Walgrave Lode, Ed.), Restorative Justice and the Law, Willan Publishing, Cullompton, p. 168-190.Dignan, Jim, 2003, Toward a systemic model of restorative justice, dans Von hirsh Andreas, Roberts Julian, Bottoms Anthony, Roach Kent, Schiff Mara. (Eds.), Restorative justice and criminal justice. Competing or reconciliable paradigms ?, Hart Pub., p. 135-156.Filippi, Jessica, 2015, Droit pénal des mineurs et justice restaurative. Approche comparée franco-belge, Thèse, octobre 2015.Filippi, Jessica, 2018, « Restorative justice for young offenders : an analysis of the French circular confronting at European and national legal perspective », dans UCL Journal of Law and Jurisprudence , vol. 7 , n° 1, p. 142-162.Hannem, Stacey, 2013, « Experiences in reconciling risk management and restorative justice, How circles of support and accountability work restoratively in the risk society », dans International journal of offender therapy and comparative criminology, n°57, p. 269-288.Januel, Pierre, 2019, « Justice des mineurs : des pistes parlementaires pour la réforme », Dalloz Actualité, Le quotidien du droit, [En ligne] Accessible à l’adresse https://www.dalloz-actualite.fr/flash/justice-des-mineurs-des-pistes-parlementaires-pour-reforme#.XROD9FJG7j8 [Consulté le 29 juillet 2019].L’Épée-Boulanger, Audrey, 2018, « Réformer la justice des mineurs par ordonnances ? Les prérequis pour une réforme de l’ordonnance de 1945 », Dalloz actualité, Le quotidien du droit, [En ligne] Accessible à l’adresse https://www.dalloz-actualite.fr/node/reformer-justice-des-mineurs-par-ordonnances-prerequis-pour-une-reforme-de-l-ordonnance-de-1945#.XROR2FJG7j8 [Consulté le 29 juillet 2019].Legrand, Sophie et Lucille Rouet, 2019, « Juge de la protection, de l’éducation mais aussi de la détention des mineurs, les paradoxes de la fonction de juge des enfants », dans Enfances et Psy, n°83, p. 22-29.Lemonne, Anne, 2002, « A propos de la 5e conférence internationale sur la justice restauratrice. Accord ou contradiction au sein d’un mouvement en expansion », Revue de droit pénal comparé, p. 411-428.London, Ross, 2003, «The restoration of truth : bringing restorative justice from the margins to the mainstream », Criminal justice studies, vol.16, n°3, p. 175-195.O’Mahony, David et Jonathan Doak, 2017, Reimagining restorative justice : agency and accountability in the criminal process, Oxford, coll. Hart Publishing.Marshall,. Tony, 1996, « The evolution of restorative justice in Britain », European Journal on Criminal Policy and Research, vol. 4, n°4, p. 21-43.Marshall, Tony, 1999, Restorative justice, an overview, Home Office, Research development and statistics directorate.Mary, Philippe et Dominique Defraene, 2007, Sanctions et mesures dans la communauté. Etat critique de la question en Belgique, Bruxelles, Fondation Roi Baudouin.Mbanzoulou, Paul, 2012, La médiation pénale, L’Harmattan, coll. Sciences criminelles.Miers, David, 2007, « The international development of restorative justice », dans Jonstone, Gerry et Daniel Van Ness (dir.), Handbook of Restorative Justice. Cullompton et Portland, Willan Publishing.Ruggiero, Vincenzo, 2010, Penal Abolitionism, Oxford, Oxford University Press.Sallée, Nicolas, 2016, Éduquer sous contrainte, une sociologie de la justice des mineurs, Éd. EHESS, 84 p.Sawin, Jennifer Larson et Howard Zehr, 2007, « The ideas of engagement and empowerment », dans Jonstone, Gerry et Daniel Van Ness (dir.), Handbook of Restorative justice, Cullompton et Portland, Willan Publishing, p. 41-58.© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 47 Cahiers de la sécurité et de la justice – Hors-série Strimelle, Véronique, 2007, « La justice restaurative : une innovation du pénal ? », Champ pénal/Penal field [En ligne], Accessible à l’adresse https://journals.openedition.org/champpenal/912 [Consulté le 29 juillet 2019].Van Ness, Daniel et Karen Heetderks Strong, 1997, Restoring Justice, Cincinnati, Anderson.Von Hirsch, Andrew, Andrew Ashworth et Clifford Shearing, 2003, « Specifying aims and Limits for restorative justice : a ‘making amend’ model ? », dans Von Hirsh Andrew, Andrew Ashworth, Anthony Bottoms, Kent Roach et Mara Schiff (dir.), Restorative justice and Criminal Justice, Cometing or reconciliable paradigms, Hart Pusblishing, p. 21-41.Vanfraechem, Inge et Lode Walgrave, « Les conférences de groupe familial », Les cahiers de la Justice, Revue semestrielle de l’É.N.M., Dalloz, 2006, n° 1, p. 153-155.Vigour, Cécile, 2006, « Justice : l’introduction d’une rationalité managériale comme euphémisation des enjeux politiques », Droit et société, n° 63-64, p. 425-455.Walgrave, Lode, 1993, « Beyond rehabilitation : in search of a constructive alternative in the judicial response to juvenile crime », dans European Journal on Criminal Policy and Research, vol. 2, n°2, p. 57-75.Walgrave, Lode, 1999, « La justice restaurative : à la recherche d’une théorie et d’un programme », Criminologie, 32(1), p. 7-29.Walgrave, Lode, 2000, « How Pure Can a Maximalist Approach to Restorative Justice Remain? Or Can a Purist Model of Restorative Justice Become Maximalist », Contemporary Justice Review, vol. 1, n.4, p.415-432.Walgrave, Lode, 2008, « Examining restorative justice practice », dans Walgrave, Lode, Restorative justice, self-interest and responsible citizenship, Cullompton, Willan publishing, p. 101-109, traduit dans Gailly, Philippe, 2011, La justice restauratrice, Ed. Larcier Coll. Crimen, Bruxelles, p. 375-389.Walgrave, Lode et Zinsstag, Estelle, 2014, « Justice des mineurs et justice restaurative, Une intégration possible et nécessaire », les Cahiers Dynamiques, n°59, p. 32-40.Wright, Martin, 1996, Justice for victims and offenders, Winchester, Waterside Press.Zehr, Howard, 1990, Changing lenses : A new focus for crime and justice. Scottsdale, Herald Press.Zehr, Howard, 2002, The little book of restorative justice, Good books.Zehr, Howard, 2012, La justice restaurative. Pour sortir des impasses de la logique punitive, Labor et fides.Textes et réglementationCode de la justice pénale, Ordonnance n° 2019-950 du 11 septembre 2019 portant partie législative du Code de la justice pénale des mineurs, J.O.R.F. n° 0213 du 13 septembre 2019, [En ligne] Accessible à l’adresse https://www.legifrance.gouv.fr/jorf/id/JORFTEXT000039085102/ [Consulté le 19 octobre 2020].Circulaire du 15 mars 2017 relative à la Mise en œuvre de la justice restaurative applicable immédiatement suite aux article 10-1, 10-2 et 707 du Code de procédure pénale, [En ligne] Accessible à http ://www.justice.gouv.fr/_telechargement/Circulaire_justice_restaurative_signee_JJU_15.03.2017.pdf [Consulté le 29 juillet 2019].Loi n° 2014-896, J.O.R.F. n° 0189, du 15 août 2014 relative à l’individualisation des peines et au renforcement de l’efficacité des sanctions pénales, J.O., 17 août 2014, page 13647.Nation Unies, Office on Drugs and Crime, 2006, ‘Handbook on restorative justice programmes’ Criminal justice Handbook series, [En ligne] Accessible à l’adresse https ://www.unodc.org/pdf/criminal_justice/Handbook_on_Restorative_Justice_Programmes.pdf [Consulté le 29 juillet 2019].Projet de loi relatif à la Justice pénale des enfants et des adolescents (version du troisième trimestre 2015), 2016, dans Journal du droit des jeunes, n°351-352, p. 98-119.Recommandation CM/Rec(2018)8 du Co-mité des Ministres aux États membres relative à la justice restaurative en matière pénale, adoptée par le Comité des Ministres le 3 octobre 2018, [En ligne], Accessible à l’adresse https://search.coe.int/cm/Pages/re-sult_details.aspx?ObjectId=09000016808e35f4 [Consulté le 19 octobre 2020].© Cahiers de la sécurité et de la justice - IHEMI48 I DOSSIERRestorative justice for juveniles in France: maximalist and minimalist tendencies — Jessica FilippiRestorative justice for juveniles in France: maximalist and minimalist tendenciesJessica FILIPPIuvenile criminal justice was governed for nearly 70 years by the ruling n°45-174 of February 2nd, 1945 regarding juvenile delinquency. After being amended by about forty reforms, it was lacking in consistency and legibility, and had become difficult to apply for professionals on the ground. In 2008, the Varinard commission presented a proposition for a Code of juvenile criminal justice. Under the following governments, Minister of Justice Christiane Taubira and her successor Jean-Jacques Urvoas worked on a codification project (Projet de loi 2015, 2016). In November 2017, then Minister of Justice Nicole Belloubet raised the issue again, and on December 11th, 2018, the National Assembly voted on a draft justice reform and budget planning bill. The bill was amended by the Minister of Justice to introduce an article which enables the government to reform by decrees the ruling of February 2nd, 1945. Many jurists and sociologists then reacted to examine the consequences and risks of that reform in the management of juvenile criminality (Couster, 2019; Januel, 2019; L’Épée-Boulanger, 2018), but none of these publications mentioned the articles L. 13-4 on restorative justice and L. 112-8 2° on mediation (Code of criminal justice, 2019). Following the recent introduction of restorative justice in the fabric of ordinary law and the article 10-1 of the Code of criminal procedure, a circular regarding its implementation with adult as well as juvenile populations was published on March 15th, 2017. In addition to specifying the normative framework of restorative justice, it aims to clarify the articulation between the criminal justice system and restorative Experimentations first took place in the adult justice system at the Poissy detention center in 2010 (Beauhaire et Davau, 2019), then consecrated in the framework of ordinary law in 2014, restorative justice is slowly finding its place in the criminal justice system as well as in the practices of the Judicial Protection of Youth (open custody, Juvisy-sur-Orge, 2016). While the circular of March 2017 opened the possibility of restorative justice for juveniles, the judicial environment and the legal framework limit its development. It is to be hoped that the new Code of juvenile criminal justice adopted on September 13th, 2019, will allow restorative justice to thriveKeywords: Restorative justice, juveniles, France, Code of juvenile criminal justice, amendment to the ruling, maximalist, minimalist, experiment.JJessica FILIPPIJessica Filippi is a criminology researcher at the National School of Judiciary Protection of Youth (ENPJJ) and an associate criminology researcher at the research center on Penality, Security and Deviance of the Université Libre de Bruxelles ULB067.© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 49 Cahiers de la sécurité et de la justice – Hors-série justice by defining its principles and the conditions of its implementation. However, the juvenile justice model and its current transformations favor the development of restorative justice to a certain extent (I). The way it has been introduced in the law seems to seek a balance between minimalist and maximalist tendencies. These tendencies can also be observed in the experimental practices of professionals within the Judiciary Protection of Youth (II). While restorative justice is moving beyond experiments and taking its place in the juvenile justice system, it needs to be afforded the means to develop and to be fully integrated among the tools in every judge’s arsenal (III). The organization of juvenile criminal justice: a protectional model under a managerial logic The principles of juvenile criminal justice On September 13th, 2019, the Code of criminal justice applicable to juveniles (CJPM) was published in the Official Journal, and is set to come into effect on March 31th, 2021, because of the COVID crisis 1. This code doesn’t only consist of texts pertaining to criminal procedure, but also elements of criminal law and precisions on the ways to intervene with regards to juvenile delinquents. The code is based on the same foundation as the ruling of February 2nd, 1945. It is governed by three principles: first, promoting the spirit of protection which must underlie any intervention directed at juveniles, and affirming the primacy of education over repression (Legrand and Rouet, 2019); then, developing specialized jurisdiction courts through adapted procedures and professionals to work with juveniles; and finally, mitigating responsibility according to age. The Code presents two major changes: the presumption of discernment starting at 13 years old 2, and the replacement of the preliminary inquiry phase (except for serious crimes and offences) by a procedure of “educational probation”. The main services involved in juvenile justice system are: the Child Welfare Department (ASE) 3, the public service of Judicial Protection of Youth (PJJ), and the civil society organizations accredited by the public sector (SAH). The Department of the Judiciary Protection of Youth (DPJJ) has been coordinating the juvenile justice system since the July 9th, 2008 order on the organization of the ministry of Justice came into effect. It is in charge of “all the issues concerning the juvenile justice system and the cooperation between the institutions taking part in it.” The PJJ is essentially composed of social workers, educators, psychologists, and is responsible for guaranteeing the execution of the judges’ mandates and to ensure the rehabilitation of the juvenile offenders. However, due to the successive reforms both past and present, the aforementioned principles are progressively crumbling, because of the acceleration of judicial time (Bastard et al., 2015) and the multiplication of modes of intervention. Socio-educational interventions and timeframe within the Judicial Protection of Youth In the last twenty years, in the face of the need to address juvenile delinquency, changes have been made in the juvenile criminal justice system, grounded in an efficiency rationale. New means were developed to accelerate judicial procedures, but this also came with practices of intervention that stray away from the original educational and social purposes (Sallée, 2016 : 84). To make justice efficient and effective, it is no longer enough to make a fair legal decision, it must also be taken within a reasonable timeframe. This concern is manifested in an increased role given to the Attorney General, translated by the adoption of speedy procedures, earlier court appearances, and multiple alternative measures to avoid prosecution. Currently, the penal policy of public prosecutors relies on a nearly systematic penal response. The role of public prosecution, which was expanded by the circular of October 15th, 1991, emphasizes the speediness of penal response. The circular of October 2nd, 1992 recommends systematic and swift penal responses. Those of June 8th, 1998 and November 6th, 1998 aim to provide a systematic, quick and legible response to every single offence, of any kind. The acceleration of the penal response is also observed in the sentencing of juveniles (Bitton and Rosenczveig, 2011 : 24-31). Recently, the Code of criminal justice applicable to juveniles attempted to “simplify” the procedure applicable to young delinquents by establishing (1) Article 25, law n°°2020-734 of June 17th, 2020 regarding various provisions related to the health crisis and other emergencies measures and the withdrawal of the United Kingdom from the European Union, JORF (Official Journal of the French Republic), 18 juin 2020.(2) Article L.11-1 C.J.P.M.(3) The Child Welfare Department (ASE) is in charge of the protection and education of at-risk minors.© Cahiers de la sécurité et de la justice - IHEMI50 I DOSSIER Restorative justice for juveniles in France: maximalist and minimalist tendencies — Jessica Filippia single mode of prosecution, but this intent implies an acceleration of judicial time (Bastard et al., 2015). However, such policies are far from new, and they were introduced gradually (first came the LOPPSI 2 law 4, then the early court appearance 5, early sentencing 6, immediate arraignment before a juvenile jurisdiction court 7, and summons issued by a judicial police officer to appear before the juvenile court. 8) Nowadays, this acceleration continues, with the new procedure of “educational probation”, which shortened the time limit for sentencing from 18 to 12 months, articulated in three stages: a verdict on guilt or innocence within 3 months, followed by a maximum of 9 months of pre-sentencing probation, after which a sentencing decision is taken. The replacement of the slow temporality of judicial intervention by immediate responses and the call to hasten educational interventions has consequences on the organization of the PJJ’s services and the handling of juveniles (Vicour, 2006 : 425-455). A true organizational culture is progressively developed, and numerous devices are deployed within a control-based system (increase in surveillance and coercion measures, inauguration of twenty additional Custodial Educational Facilities). A neoliberal conception of justice is being built, breaking away from the philosophy of rehabilitation (Bailleau, 2008 : 418).While restorative practices already existed before the “restorative justice measure” introduced in 2014, their current implementation is affected by the successive judicial reforms and the managerial policies. Hence many services are forced to speed up the pace of the execution of educational measures, which impacts the treatment and handling of offenders. More specifically, regarding the “measure of penal reparation”: while its purported objective of reparation falls within the maximalist current 9 of restorative justice, this approach has been derailed due to the successive reforms, the short intervention timeframes and the managerial policies. Thus, a juvenile offender is apprehended depending on the offence committed and, with the help of an educator, carries out a “reparation” activity. Offenders are often grouped together in the same activity with others who committed similar offences. Reparation is focused more on the offence itself than on its consequences and repercussions. Besides, it does not include all the dimensions of the damage caused by the offence: the material damage, the psychological and relational suffering caused to the victim and to the community, and the social harm the offender brings upon him or herself (Filippi, 2015). While theoretically, the “measure of penal reparation” falls within the maximalist approach of restorative justice, the way it is put into practice casts doubt on whether it really is so.The legal consecration of restorative justice into the juvenile justice system Definitions and currents of restorative justiceIt would be a mistake to think that there exists a single definition (Lemonne, 2002 : 413) or a single approach (Miers, 2007: 447-467) to restorative justice. Given multiple names in the French-speaking world (“justice restaurative” in France, “justice restauratrice” in Belgium, “justice réparatrice” in Switzerland and Canada, where it is also known under the designation of “justice transformatrice”), restorative justice has known increasing popularity in France in recent years: experiments (4) Some of its provisions were rejected by the Constitutional Council as they “did not guarantee that the court would be in possession of sufficient information on the juveniles’ character” that would allow to attend to their “moral and educational rehabilitation”. See Ruling n° 2011- 625 DC, 32nd to 34th Whereas clauses, AJDA 2011. 1097(5) Law n° 95-125 of February 8th, 1995 regarding the organization of jurisdictions and civil, criminal and administrative procedure, JORF, February 9th, 1995; Law n° 96-585 of July 1st, 1996 modifying the ruling n° 45-174 of February 2nd, 1945 regarding juvenile delinquency, JORF, July 2nd, 1996. The law n° 96-585, also known as Toubon law, introduced the procedures of summonses issued by judicial police officers for early court appearances and sentencing. Law n° 2011-1940 of December 26th, 2011 aiming to establish mandated community service for juvenile delinquents. (6) Law n° 2002-1138 of September 9th, 2002 laying out guidelines and budget plan for Justice policies, JORF, September 10th, 2002.(7) The law n° 2007-297 of March 5th, 2007 regarding the prevention of delinquency, JORF, March 6th, 2007 “[…] modified the second paragraph of article 5 of the ruling of February 2nd, 1945 in order to rename the ‘early sentencing’ procedure under the new appellation of ‘immediate arraignment before a juvenile jurisdiction court’; [...] this change in terminology does not imply in itself a challenge of constitutionality”; see Ruling n° 2007-553 DC of March 3rd, 2007, JORF, March 7th, 2007, p. 4356; Law n° 2011-939 of August 10th, 2011 on the participation of citizens to the operation of criminal justice and juvenile justice, JORF, August 11th, 2011.(8) This procedure consists of a sort of direct citation which removes the phase of preliminary inquiry formerly performed by a juvenile judge. « La première circulaire d’application de la loi du 10 Août 2011 », in Dalloz Actualité, Le quotidien du droit, September 14th, 2011(9) See section II.A for in-depth discussion of the maximalist and minimalist conceptions. © Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 51 Cahiers de la sécurité et de la justice – Hors-série conducted with juvenile and adult populations, media coverage, legal consecration in the judicial framework, etc. But as there are many definitions of restorative justice, such flexibility brings about debate on what qualifies as restorative justice or not in the French context.In the scientific literature, the definitions of restorative justice focus on its process, its purpose, or both. According to Tony Marshall, restorative justice is “a process whereby parties with a stake in a specific offence come together to resolve collectively how to deal with the aftermath of the offence and its implications for the future.” (1996, 21, 1999, 36). Understood as processes, negotiation and consensualism become criteria for restorative justice, in the minimalist or purist approach. In such a conception, little importance is given to the end goal of restorative justice, i.e. the reparation or restoration of the damage sustained. According to Bazemore and Walgrave, proponents of the maximalist approach, restorative justice cannot be reduced to its process, and aims to restore its participants affected by the offence. It is a way of doing justice “that is primarily oriented towards [...] restoring the harm that has been caused by a crime” (Bazemore and Walgrave, 1999). Although it may only involve the offender, whether in a coerced, formal or informal context, or during a negotiation process, this approach favors the reparation of the damage occasioned by an offence. According to Howard Zehr, restorative justice is “a process to involve, to the extent possible, those who have a stake in a specific offence and to collectively identify and address harms, needs, and obligation, in order to heal and put things as right as possible.” (Zehr, 2002 : 37, 2012 : 98). Hence, beyond the issue of the definition, the source of tension between these two models hinges upon the intervention framework of restorative justice, and its place between the criminal justice system. The purists consider restorative justice and coercion to be incompatible: they cannot coexist, as that would hinder the participants’ restoration process. As for maximalists, they believe punishment and constraint can fulfill a restorative role. Indeed, their approach “implies the execution of coercive judicial obligations in order to achieve (partially) the desired reparation.” (Walgrave and Zinsstag, 2014 : 33). Arguing that coercion can be part of restorative justice Walgrave (2000: 422) states: “We [maximalists] opt for accepting coercion when voluntariness is not achieved and when it is deemed necessary to respond to the offence”. Thus sanctions that aim to repair, restorative sanctions, are part of restorative justice (Van Fraechem and Walgrave, 2006). Lastly, the maximalist and purist currents of restorative justice are distinguished from each other in their relation to institutionalization within the criminal justice system. On the one hand, the purist thinkers consider that restorative justice can be developed within the criminal justice system, without weakening the rules of procedure. The institutionalization of restorative justice is thus envisioned to occur outside of the criminal justice system (London, 2003 : 177). On the other hand, maximalists believe that restorative justice can be aligned with the criminal justice system, and even integrated with it in order to transform it. They critique the proponents of the purist approach, which limits the practice of restorative justice: the criminal justice system deals with the situations in which voluntary encounters are not possible (Dignan, 2002, 2003; Von Hirsch Ashworth and Shearing, 2003).The legal introduction of restorative justice in the criminal procedure applicable to juvenilesSince October 1st, 2014, the law n°2014-896 of August 15th 2014 regarding the individualization of sentences and the maximization of the efficiency of penal punishment came into effect (Law, 2014), and “measures” of restorative justice have hence been introduced into the article 10-1 of the Code of criminal procedure. The adoption of that legislation demonstrates a global will to turn away from the usual penal responses, often deemed inefficient, not participatory enough, not focused enough on accountability, not taking the victims into consideration enough, and even too punitive. The measures adopted fall within a maximalist approach to restorative justice. They are presented with an emphasis on their end goal, geared towards the reparation of the consequences of the offence: “any measure allowing [...] to participate actively in the resolution of the difficulties resulting from the offence, and especially in the reparation of the damage of any kind that may have resulted from its commission, constitutes a measure of restorative justice.” Three years after the promulgation of that law, the circular of March 15th, 2017, regarding the implementation of these “ad hoc” measures at every stage of the judicial procedure, came into effect. Entitled “implementation of restorative justice, applicable immediately according to the article 10-1 [...]”, this circular aims to further explain how restorative justice can be applied to the professional practices of judges, lawyers, educators, social workers and psychologists, with adults as well as with juveniles. In the first introductive pages, restorative justice is, again, defined according to the maximalist conception as theorized by Lode Walgrave. It is thus presented as “a process, complementary to the criminal trial, which consists in restoring the social fabric [...]. © Cahiers de la sécurité et de la justice - IHEMI52 I DOSSIER Restorative justice for juveniles in France: maximalist and minimalist tendencies — Jessica FilippiIt is conceived to comprehend all the personal, familial and social repercussions related to the commission of the facts [...].” This definition associates educational judicial measures, such as penal mediation and penal reparation, with restorative justice. While in the circular, restorative justice is defined according to its purpose, its implementation within the criminal justice system is also envisioned as a “voluntary and confidential” process. Additionally, the circular provides that the judicial authority shall exert a suitability control of the measure. This control raises questions as to the impact it may have on the procedure and the judicial decisions. In theory, the judge will only have control over the elements of legality of the measure (information provided to the participants in the restorative justice process, acknowledgement of the facts, consent to participate). Aside from that, the circular provides that the measure shall be confidential, and that confidentiality can only be broken on two conditions: “unless agreed upon by all concerned parties, and except in the cases where a superior interest related to the necessity to prevent or repress violations would justify [it].” This confidentiality “guarantees that restorative justice remains independent from the criminal procedure.” In this regard, one can therefore argue that the measures of restorative justice are established “at the margin” of the criminal justice system. Although they are presented as complementary to the judicial process, their autonomy is also valued. Restorative justice leaves the criminal justice system out of the discussion, and is limited to complementing its effects (Walgrave, 1999 : 12). The circular thus adopts a minimalist approach, stating that “the innovative character of the measure hinges on its autonomy from the criminal procedure. It constitutes an optional route which is made available to the parties, and has no consequence on the development of the judicial procedure which unfolds alongside it” (Circular, 2017). The text confirms the “impermeable” separation between the measure of restorative justice and the judicial process (Circular, 2017), and recalls the independence of the measure of restorative justice in that it “shall not impact the decision whether to press charges or dismiss a case” nor “impact the decision whether to award damages to the plaintiff.” Its place in the criminal procedure is peripheral. This choice to organize restorative justice according to the minimalist approach can be explained by the wish to break away from the judicial timeframe and the modes of intervention of the juvenile criminal justice system (such as penal reparation) which could affect the process.If the circular had adopted a maximalist approach, the juvenile criminal procedure could have considered the possibility of integrating restorative justice within its practices, while still protecting the confidentiality of interactions. While according to the circular, the judge who emits the proposal of a measure of restorative justice is limited to a simple legality check, in a maximalist approach, it would be necessary to inform said judge of the development of the measure. Special attention could be given to what elements of feedback may or may not be provided to the judge. In keeping with the philosophy of juvenile justice and its objectives of protecting, educating and promoting accountability, the possibility of establishing a brief report on the commitments taken by the participants in restorative justice and transmitting it to the judge could be envisioned. The judge would make sure that said commitments do not go against the interest of the juveniles (protection) and that they contribute to their educational rehabilitation (education/accountability). 10 The judge in charge of the legality check (Circular, 2017) would keep this role at the end of the restorative justice measure. The judge, knowing that a juvenile is voluntarily participating in such a measure (for example if the juvenile mentions during a court hearing that he or she is participating in such a process) would probably be influenced by this information. The idea would be that participating or not in a process of restorative justice should not be detrimental to the juvenile (education and protection goals). In such a maximalist perspective, restorative justice would pursue its objective of restoration even in the judicial response. Thus, while the maximalist approach is consecrated in the first lines of the article 10-1 of the Code of criminal procedure as well as in the introductory elements of the circular, it deserves to be extended to the whole juvenile criminal justice system. However, one risk would remain: that this would then lead it to “stick” to the judicial timeframe. Lastly, the measures of restorative justice can be implemented at any stage of the judicial procedure, including during the execution of a sentence. The examples of possible measures of restorative justice presented at the end of the circular demonstrate, once again, a maximalist approach: encounters between convicts and victims (RCV), encounters between inmates and victims (RDV), restorative mediation and restorative conferences. The circular opens up the perspective of restorative justice, suggesting other measures such as restorative circles, Circles of Support and Accountability (known as CoSA in English, CSR in French) or Circles of Guidance and Resources (CAR). 11 The measures are therefore not envisioned only as encounters or interactions (10) One can also wonder about the follow-up on the measure of restorative justice, and about what should be done if the participants do not go through with it. © Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 53 Cahiers de la sécurité et de la justice – Hors-série between victims and perpetrators (Walgrave, 1999 : 13), but also as processes in which the offenders and the community (Filippi, 2018) can participate.12The experiments of restorative justice directed towards juveniles 13The experiments that have been conducted since 2016 by the PJJ have scarcely resulted in direct mediation between offenders and victims. Out of the possible practices of intervention mentioned in the circular, the measures that have been implemented with juveniles are mostly indirect restorative mediation, restorative conferences, and support groups. The specific development of these measures has also been varied. Depending on the services and on the partners, restorative justice is either carried out as a free, open framework, with no predefined form, that evolves according to the wishes of the participants, or as a standardized practice (a determined measure, with a given number of preparatory interviews and a given number of encounters). Since 2018 and until 2020, nine Interregional Departments (DIR) of the Judicial Protection of Youth have been experimenting restorative justice for juveniles. The partnership contracts established with the different actors of the criminal justice system allow to define the role and place of each of them in the orientation and implementation of restorative justice. While some of the DIR have already set up partnerships with various institutions (DPJJ, national school system, SPIP 14, SAV15, bar associations), others still haven’t signed these agreements. Besides, many professionals in the justice system (law enforcement, lawyers, judges) remain unaware of the restorative justice measures.At the PJJ, the development of restorative justice remains “at the margin”, in parallel to the judicial procedure. At the margin, because its implementation is only considered on the condition that the youth be subjected to a penal measure. Thus, only the juveniles handled by the criminal justice system and its educational services can have access to a measure of restorative justice. Juveniles who are going through civil proceedings or whose cases were dismissed without further action in a penal procedure cannot be offered such a measure. This proposal to implement restorative justice in a coercive context is part of a larger context where this kind of justice does not exist as a unique measure. This mode of action also responds to the necessity to respect the confidentiality of restorative justice, and to preserve its impermeable separation from the judicial procedure. Such an approach also seems to be related to the judges’ perception of restorative justice: they deem it time-consuming, they feel that it deprives them of their authority and that it is not a quick and viable alternative solution. Furthermore, if an educational measure is being performed alongside it, there is no need to know what is happening in restorative justice: reports of the educational measure are already illuminating the judge regarding the evolution and involvement of the juvenile. The professional culture (education, protection) and the work environment (managerial rationality, short intervention timeframes, availability of the local services) limit the use and the development of restorative justice by the actors of the justice system and by the citizens. When it comes to information and orientation, it is observed in all nine DIR, that law enforcement officials often don’t think about informing the citizens about the possibility of participating in restorative justice. As for judges and prosecutors, their practices are varied. Some of them inform and guide towards restorative justice left and right, and let the services in charge (PJJ, SAH, SAV, specialized services) use and interpret the measure broadly, while others control for which cases restorative justice would be most relevant. Others still propose or impose a format, and lastly, some (but they are few), veto all measures of restorative justice, especially in the pre-sentencing phase, for fear of the risks of conflicts of interest, infringements of confidentiality with regards to the investigation, or interference between participants. In the implementation of restorative justice by the services, a multitude of practices can also be observed: either restorative justice is carried out within an identified (11) The latter mobilize two circles around the offender (main member) after the end of the sentence or shortly before release, in order to support the rehabilitation process. These circles include members of the community (volunteers and professionals) and coordinators who help establish bonds between the professionals and the community members of the circle. (12) Community restoration systems, reparation commissions, direct compensation to victims, victims and witnesses support services, victim support circles, victim awareness courses, education of inmates and offenders, problem-solving courts, circles of support and accountability, offender rehabilitation ceremonies, and projects involving delinquents and their families or other victims of crime.(13) The data collected (2016-2019) and presented in this section is an overview from the collective work of professionals in the PJJ and the various institutions that provide training and support the practice: the Association for Research in Applied Criminology (ARCA), the association Citizen and Justice (CJ), the National School of Judiciary Protection of Youth (ENPJJ), and the French Institute for Restorative Justice (IFJR).(14) Penitentiary services of rehabilitation and probation.(15) Victim support service.© Cahiers de la sécurité et de la justice - IHEMI54 I DOSSIER Restorative justice for juveniles in France: maximalist and minimalist tendencies — Jessica Filippiservice’s premises (PJJ, SAV), or it is implemented in a neutral space, detached from the connotations of the justice system (specialized restorative justice service or city hall, for example). It is either facilitated by a team of two trained professionals from two different professions (SPIP, SAV, SAH, PJJ or Restorative Justice services), or by a team of two professionals from the same service. Although they may come from different backgrounds, some of the professionals co-host the activities and prepare the participants for the restorative justice measure together as a team, whereas others conduct the information and preparation separately: while one of the two trained professionals attends to the information and preparation of the offenders and their legal representatives or relatives, whereas the other takes on the information and preparation of the victim in a different venue. In the PJJ’s services, as part of the educational management of juveniles, the educators inform the youths of their right to participate in restorative justice, and guide them towards such measures. They detect which cases are susceptible to benefit from restorative justice (offences, in one of the regions, the age threshold to benefit from restorative justice measure is 16), and refer the cases to the restorative justice specialists. In some regions, it is the trained educators themselves, in teams of two, who propose and implement restorative justice measures for juveniles (who are not under their educational supervision). Nonetheless, the management of the criminal justice system also impacts the capacity of these services to adopt the measure. Indeed, the acceleration of judicial time, the managerial logic of the services and the workload of the PJJ professionals lead them to either implement restorative justice to a lesser extent (because it is time-consuming), or to delegate it to other services. On the opposite, some professionals are granted a reduction of their workload in order to dedicate themselves to the implementation of restorative justice, which they carry out either at the premises of their own educational services, or at the location of one of the partner institutions. It has been made clear that “restorative justice is far from being a complete set of practises based on a ready-made theory of justice. Instead it is a movement and a testing ground for experimentation and research that rests on an intuitive ideal of justice in a vague ideal of society.” (Walgrave, 1999 : 19). In the CJPM, restorative justice is reserved several places. It is situated both outside (article L. 13-4) and inside (L. 112-8, 2°) the criminal justice system. If the legislative part of the code is stabilized, the EU recommendation published in October 2018 can give insights as to the regulatory part which will accompany the law. In this article, we’re focusing on article L. 13-4 of the code16.It is to be hoped that the article on restorative justice that appears in the Code of juvenile criminal justice will remain in the text throughout the debates, and will be modified to some extent in accordance with the.Restorative justice in the Code of juvenile criminal justice, in light of the EU recommendation CM/Rec(2018)8 of the Committee of Ministers of the Council of Europe Until now, there was no article dedicated to restorative justice within the ruling of February 2nd, 1945, but the new article L. 13-4 of the CJPM offers the possibility of restorative justice to underage victims and offenders. That article falls, to some extent, within the scope of the EU guidelines and recommendations that seek to invite member states to incorporate restorative justice into their own legislation. Situated at the very beginning of the code, in a unique title, restorative justice aimed at juveniles is mentioned, which demonstrates the importance given to that philosophy in the protection model of the juvenile justice system. The text states that “the recourse to restorative justice can be offered to the victim and the perpetrator of the offence, in accordance with the article 10-1 of the Code of criminal procedure, in the context of any procedure concerning a juvenile, and at any stage of said procedure, including during the execution of the sentence, on the condition that the facts have been admitted. Restorative justice can be implemented only if the level of maturity and capacity for discernment of the juvenile allow for it, and after obtaining consent of the legal representatives.”This wording is an adaptation of the article 10-1 of the Code of criminal procedure. The EU recommendation (16) Article L. 112-8, 2° will be discussed in the collective book : Filippi, Jessica, “Le parcours de la justice restaurative en droit pénal des mineurs”, La réforme de la justice pénale des mineurs : quelles spécificité(s) ? Dalloz, T&C, 2021. This contribution will deal with the utility of an integrated approach to restorative justice for minors at every step of the judicial procedure. This chapter will also evoke the adaptations to bring to the regulatory part of the CJPM.© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 55 Cahiers de la sécurité et de la justice – Hors-série concerning restorative justice in criminal matters (CM/Rec(2018)8), published in October 2018, can give us insights as to what practical orientations to follow in the CJPM.Its first article explains this recommendation “aims to encourage member States to develop and use restorative justice with respect to their criminal justice systems. It promotes standards for the use of restorative justice in the context of the criminal procedure, and seeks to safeguard participants’ rights and maximise the effectiveness of the process in meeting participants’ needs.” (CM/Rec (2018)8, article 1). This EU recommendation demonstrates the wish to develop a maximalist approach to restorative justice. Article 8 highlights the various forms of restorative justice: “practices which do not involve a dialogue between victims and offenders may still be designed and delivered in a manner which adheres closely to the basic principles of restorative justice [...]. Restorative principles and approaches may also be applied within the criminal justice system, outside of the criminal procedure.” (CM/Rec (2018)8, article 8). This wording is based on the empirical and theoretical research produced by criminologists specialized in restorative justice (Walgrave, 1994, 1995; Bazemore, Umbreit, 1995; Wright, 1996, Van Ness and Heetderks Strong, 1997; Bazemore, Walgrave, 1998). Article 59 confirms this maximalist approach of restorative justice, as it promotes an extension of its practical conception: “While restorative justice is typically characterised by a dialogue between the parties, many interventions which do not involve dialogue between the victim and offender may be designed and delivered in a manner which adheres closely to restorative justice principles.” (CM/Rec (2018)8, article 59).In light of this recommendation, the article L. 13-4 of the CJPM concerning restorative justice could consider that the victims of an offence and the perpetrators or persons responsible for the harm caused can request or be offered to participate in restorative justice, in accordance with the article 10-1 of the Code of criminal procedure.This approach would, first, confirm that restorative justice is a right for all, a way to access justice and participate actively in it, and second, highlight that the offence impacts not only its direct victim but also the members of the community, as well as any other indirect victims. The participants would thus be considered more broadly. Police officers, teachers, social workers, relatives of the offenders and the victims, and representatives of the “community” could be included among the relevant professionals and possible participants. This formulation would allow to envision all forms of restorative justice, and not only the encounters between convicts and victims or inmates and victims, direct or indirect mediation and restorative group conferencing. Involving a wider group can help to educate, to strengthen social capital, and encourage the offender to develop more skills likely to contribute to the prevention or management of crime and future conflicts. Speaking in terms of harm, damage and prejudice, although it might carry connotations “to the notion of compensation in civil lawsuits” (Cario, 2019 : 87), contributes to detaching the discourse from the notion of offence. If the Circular of March 15th, 2017 does not mention the resolution of the consequences of the offence nor its modes of reparation, the notion of harm could be part of the solution. With this possibility, restorative justice could be both an additional measure to the criminal justice system, independent of its results, and a measure integral to the criminal justice system, having a possible impact on the followings of the lawsuit17. Restorative justice would be implemented at every step of the judicial procedure, for any type of offence and to anyone who may need it, on condition that the basic facts be acknowledged (Circular, 2017). Resorting to restorative justice does not require a judicial decision regarding guilt or innocence, and acknowledging responsibility does not require or equate to a legal admission of guilt. Lastly, the assessment of discernment or maturity is not a prior condition for proposing the measure of restorative justice, only a condition for carrying it out, and this assessment is incumbent on the professional implementing restorative justice.A juvenile with lesser capacities could still participate in restorative justice as, in theory, it is supposed to adapt to the needs of individuals. If a juvenile is in difficulty and wishes to be represented and supported, that possibility exists. Besides, establishing a fixed age of discernment presumption at 13 years old would mean that, in criminal justice matters, a juvenile below that age would not be able to benefit from restorative justice. If such were the case, it would be necessary for restorative justice to be made available and offered in other institutions than the criminal justice sphere. Practices of restorative justice take place in primary and junior high schools to handle altercations, incidents of violence and theft. Furthermore, (17) This proposition would first contribute to integrate article L. 112-8 into restorative justice. Second it would develop mediation in every step of the judicial procedure, including during prosecution. To date, mediation is considered as part of a judicial educational measure. Following the Circular of March 15th, 2017, it is likely damages cannot be considered in the resolution of the conflict, contrarily to the what the debates around the 2015 reform suggested (Projet de loi 2015, 2016).© Cahiers de la sécurité et de la justice - IHEMI56 I DOSSIER Restorative justice for juveniles in France: maximalist and minimalist tendencies — Jessica Filippiif conflicts, including when they derive from an offence, are dealt with by the justice system through civil lawsuits, restorative justice could also be implemented in that context.These propositions suggest to see article L. 13-4 only as an introduction to restorative justice in the CJPM to develop its applications to every step of the judicial procedure. But to this day, one can only observe that two trends are confronting regarding restorative justice in the code, one maximalist and the other minimalist.ConclusionThe legislative progress made in France for the introduction of restorative justice in criminal matters takes this approach seriously. There is still a long way to go before restorative justice is fully integrated within the criminal justice system. The numerous reforms underway, the acceleration of judicial time, the managerial politics of the services and the professional culture (education, protection) are a few of the obstacles that hinder the development of restorative justice principles. In November 2019 was the 30th anniversary of the Convention on the Rights of the Child (CRC). Restorative justice, as an empowering process (O’Mahony et Doak 2017), allows for a full development of the capability rights of children (Filippi, 2018). While in their effects, restorative justice and criminal justice are complementary, it is necessary to consider the complementarity of the two models within the socio-judicial response. This would make much more sense for juvenile offenders. Such a combination would be more respectful of their rights, of their self under construction. In addition to a legal framework, it is absolutely necessary to cultivate spaces for training, dialogue and research (ARCA, CJ, ENPJJ, IFJR) as well as support and supervision of professionals, in order to establish these innovative professional practices in the judicial fabric nReferencesBooks and articles Beauhaire, Alexandra and Inès Davau, 2019, “Retour sur la conférence “justice restaurative””, INHESJ – ONDRP – A2S.Bailleau, Francis, 2008, “L’exceptionnalité française. Les raisons et les conditions de la disparition programmée de l’ordonnance pénale du 2 février 1945”, in Droit et Société, n° 69-70, pp. 399-438.Bastard, Benoit, David Delvaux, Christian Mouhanna and Frédéric Schoenaers, 2015, « Vitesse ou précipitation ? La question du temps dans le traitement des affaires pénales en France et en Belgique », in Droit et société, n°90, p. 271-286Bazemore, Gordon and Lode Walgrave (Eds.), 1999, Restorative Juvenile Justice : Repairing the Harm of Youth Crime, Monsey, Criminal Justice Press, 408 p.Bazemore, Gordon and Mark Umbreit, 1995, “Rethinking the sanctioning function in Juvenile Court: Retributive or restorative responses to youth Crime”, Crime and Delinquency, vol. 41, pp. 296-316.Bazemore, Gordon and Lode Walgrave, 1998, Exploring restorative justice for juveniles, Monsey, Criminal justice press.Bitton, Josine and Jean-Pierre Rosenczveig, 2011, “Quelle justice pour les mineurs. La justice au tribunal pour enfants de Bobigny : modernité ou régression ?”, in Après-Demain, n° 19, pp. 24-31 Cario, Robert, 2004, La justice restaurative, Paris, L’Harmattan.Cario, Robert, 2007, « La justice restaurative : vers un nouveau modèle de justice pénale ? », AJ Pénal, Dalloz, p. 372-375.Cario, Robert, 2010, Justice restaurative, Principes et promesses, Paris, L’Harmattan, 2e éd.Cario, Robert, 2019, « Les apports de la recommandation (18)8 du Conseil de l’Europe dans la consolidation de la justice restaurative en France », AJ Pénal, Dalloz, 87.Christie, Nils, 2005, Au bout de nos peines, Brussels, Éd. Larcier.Coignac, Anaïs, 2019, « Justice restaurative un dispositif encore trop peu utilisé », Dalloz actualité, le Quotidien du droit [En ligne] Accessible à l’adresse https://www.dalloz-actualite.fr/dossier/justice-restaurative-reparation-yeux-dans-yeux#.XRS_cVJG7j8 [Accessed July 29th, 2019].© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 57 Cahiers de la sécurité et de la justice – Hors-série Couster, Thomas, 2019, « Réforme de la justice des mineurs : la Chancellerie lance une consultation de trois mois », Dalloz actualité, le Quotidien du droit [En ligne], Accessible à l’adresse https://www.dalloz-actualite.fr/flash/reforme-de-justice-des-mineurs-chancellerie-lance-une-consultation-de-trois-mois#.X4mWmdUzbIU [Consulté le 16 octobre 2020].Cunneen, Chris and Carolyn Hoyle, 2010, Debating restorative justice, Oxford, Hart Publishing.Dachy, Aurore, 2013, L’application de la concertation restauratrice en groupe en Fédération Wallonie Bruxelles, Ministry of Youth and Assistance to Youth of the Wallonia-Brussels Federation, [Online] Available at https://www.huytebroeck.be/IMG/pdf/Rapport_CRG_avril_2013.pdf [Accessed July 29th, 2019].Dignan, Jim, 2002, Restorative justice and the law: the case for an integrated, systemic approach, in Walgrave Lode, (Ed.), Restorative Justice and the Law, Cullompton, Willan Publishing, pp. 168-190.Dignan, Jim, 2003, Toward a systemic model of restorative justice, in Von Hirsch Andreas, Roberts Julian, Bottoms Anthony, Roach Kent, Schiff Mara (Eds.), Restorative justice and criminal justice. Competing or reconcilable paradigms?, Hart Pub., pp. 135-156.Filippi, Jessica, 2015, Droit pénal des mineurs et justice restaurative. Approche comparée franco-belge, PhD dissertation, October 2015.Filippi, Jessica, 2018, “Restorative justice for young offenders: an analysis of the French circular confronting at European and national legal perspective”, in UCL Journal of Law and Jurisprudence , vol. 7 , n° 1, pp. 142-162.Hannem, Stacey, 2013, “Experiences in reconciling risk management and restorative justice, How circles of support and accountability work restoratively in the risk society”, in International journal of offender therapy and comparative criminology, n° 57, pp. 269-288.Januel, Pierre, 2019, “Justice des mineurs : des pistes parlementaires pour la réforme”, Dalloz Actualité, Le quotidien du droit, [Online] Available at https://www.dalloz-actualite.fr/flash/justice-des-mineurs-des-pistes-parlementaires-pour-reforme#.XROD9FJG7j8 [Accessed July 29th, 2019].L’Épée-Boulanger, Audrey, 2018, “Réformer la justice des mineurs par ordonnances ? Les prérequis pour une réforme de l’ordonnance de 1945”, Dalloz actualité, Le quotidien du droit, [Online] Available at https://www.dalloz-actualite.fr/node/reformer-justice-des-mineurs-par-ordonnances-prerequis-pour-une-reforme-de-l-ordonnance-de-1945#.XROR2FJG7j8 [Accessed July 29th, 2019].Legrand, Sophie and Lucille Rouet, 2019, “Juge de la protection, de l’éducation mais aussi de la détention des mineurs, les paradoxes de la fonction de juge des enfants”, in Enfances et Psy, n°83, p. 22-29.Lemonne, Anne, 2002, “A propos de la 5e conférence internationale sur la justice restauratrice. Accord ou contradiction au sein d’un mouvement en expansion”, Revue de droit pénal comparé, pp. 411-428.London, Ross, 2003, “The restoration of truth: bringing restorative justice from the margins to the mainstream”, Criminal Justice Studies, vol. 16, n° 3, pp. 175-195.O’Mahony, David and Jonathan Doak, 2017, Reimagining restorative justice : agency and accountability in the criminal process, Oxford, coll. Hart Publishing.Marshall, Tony, 1996, “The evolution of restorative justice in Britain”, European Journal on Criminal Policy and Research, vol. 4, n° 4, pp. 21-43.Marshall, Tony, 1999, Restorative justice, an overview, UK Home Office, Research Development and Statistics Directorate.Mary, Philippe and Dominique de Fraene, 2007, Sanctions et mesures dans la communauté. État critique de la question en Belgique, Brussels, Fondation Roi Baudouin.Mbanzoulou, Paul, 2012, La médiation pénale, L’Harmattan, coll. Sciences criminelles.Miers, David, 2007, “The international development of restorative justice”, in Johnstone, Gerry and Daniel Van Ness (Eds.), Handbook of Restorative Justice, Cullompton and Portland, Willan Publishing.Ruggiero, Vincenzo, 2010, Penal Abolitionism, Oxford, Oxford University Press.Sallée, Nicolas, 2016, Éduquer sous contrainte, une sociologie de la justice des mineurs, Éd. EHESS, 84 p.Sawin, Jennifer Larson and Howard Zehr, 2007, “The ideas of engagement and empowerment”, in Johnstone, Gerry and Daniel Van Ness (Eds.), Handbook of Restorative justice, Cullompton and Portland, Willan Publishing, pp. 41-58.© Cahiers de la sécurité et de la justice - IHEMI58 I DOSSIER Restorative justice for juveniles in France: maximalist and minimalist tendencies — Jessica FilippiStrimelle, Véronique, 2007, “La justice restaurative : une innovation du pénal ?”, Champ pénal/Penal field [Online], Available athttps://journals.openedition.org/champpe-nal/912 [Accessed July 29th, 2019].Van Ness, Daniel and Karen Heetderks Strong, 1997, Restoring Justice, Cincinnati, Anderson.Von Hirsch, Andrew, Andrew Ashworth and Clifford Shearing, 2003, “Specifying aims and Limits for restorative justice: a ‘making amend’ model?”, dans Von Hirsh Andrew, Andrew Ashworth, Anthony Bottoms, Kent Roach and Mara Schiff (dir.), Restorative justice and Criminal Justice, Cometing or reconciliable paradigms, Hart Pusblishing, p. 21-41.Vanfraechem, Inge and Lode Walgrave, “Les conférences de groupe familial”, Les cahiers de la Justice, Revue semestrielle de l’É.N.M., Dalloz, 2006, n° 1, pp. 153-155.Vigour, Cécile, 2006, “Justice : l’introduction d’une rationalité managériale comme euphémisation des enjeux politiques”, Droit et société, n° 63-64, pp. 425-455.Walgrave, Lode, 2008, “Examining restorative justice practice”, in Walgrave, Lode, Restorative justice, self-interest and responsible citizenship, Cullompton, Willan publishing, pp. 101-109, translated in Gailly, Philippe, 2011, La justice restauratrice, Éd. Larcier Coll. Crimen, Brussels, pp. 375-389.Walgrave, Lode, 1993, “Beyond rehabilitation: in search of a constructive alternative in the judicial response to juvenile crime”, in European Journal on Criminal Policy and Research, vol. 2, n° 2, pp. 57-75.Walgrave, Lode, 1999, “La justice restaurative : à la recherche d’une théorie et d’un programme”, Criminologie, 32(1), pp. 7-29.Walgrave, Lode, 2000, “How Pure Can a Maximalist Approach to Restorative Justice Remain? Or Can a Purist Model of Restorative Justice Become Maximalist”, Contemporary Justice Review, vol. 1, n° 4, p.415-432.Walgrave, Lode, 2008, “Examining restorative justice practice”, in Walgrave, Lode, Restorative justice, self-interest and responsible citizenship, Cullompton, Willan publishing, p. 101-109, translated in Gailly, Philippe, 2011, La justice restauratrice, Ed. Larcier Coll. Crimen, Bruxelles, p. 375-389.Walgrave, Lode and Estelle Zinsstag, 2014, “Justice des mineurs et justice restaurative, Une intégration possible et nécessaire”, Les Cahiers Dynamiques, n° 59, pp. 32-40.Wright, Martin, 1996, Justice for victims and offenders, Winchester, Waterside Press.Zehr, Howard, 1990, Changing lenses: A new focus for crime and justice. Scottsdale, Herald Press.Zehr, Howard, 2002, The little book of restorative justice, Good books.Zehr, Howard, 2012, La justice restaurative. Pour sortir des impasses de la logique punitive, Labor et fides.Laws and regulations Code of Criminal Justice, Ruling n° 2019-950 of September 11th, 2019 ratifying into law parts of the Code of Juvenile Criminal Justice, JO n° 0213, September 13th, 2019, [Online] Available at https://www.legifrance.gouv.fr/affichTexte.do?cidTexte=JORFTEXT000039085102&categorieLien=id [Accessed July 29th, 2019].Circular of March 15th, 2017 regarding the implementation of restorative justice, applicable immediately according to the articles 10-1, 10-2 and 707 of the Code of Criminal Procedure, [Online] Available at http ://www.justice.gouv.fr/_telechargement/Circulaire_justice_restaurative_signee_JJU_15.03.2017.pdf [Accessed July 29th, 2019].Law n° 2014-896 of August 15th, 2014 regarding the individualization of sentences and the maximization of the efficiency of penal punishment, JO n° 0189, August 17th, 2014, page 13647.United Nations, Office on Drugs and Crime, 2006, “Handbook on restorative justice programmes”, Criminal justice Handbook series, [Online] Available at https ://www.unodc.org/pdf/criminal_justice/Handbook_on_Restorative_Justice_Programmes.pdf [Accessed July 29th, 2019].Law proposition regarding children and juvenile criminal justice (3rd trimester 2015 version), 2016, in Journal du droit des jeunes, n° 351-352, p. 98-119.Recommendation CM/Rec(2018)8 of the Committee of Ministers to member States concerning restorative justice in criminal matters, adopted by the Committee of Ministers on 3 October 2018, [Online], Available at https://search.coe.int/cm/Pages/result_details.aspx?ObjectId=09000016808e35f4 [Accessed October 19th, 2020].© Cahiers de la sécurité et de la justice - IHEMICahiers de la sécurité et de la justice – Hors-série DOSSIER I 59Prévention et répression de la pratique du drone civil : un équilibre à trouver 1Johanne GOJKOVIC-LETTE, Grégory HOUILLONannée 2018 a vu entrer en vigueur toute une nouvelle partie de la réglementation relative aux drones à usage civil 2. Le drone, aussi appelé unmanned aerial vehicles (UAV) / remotely-piloted aircraft – bien plus qu’un jouet et au-delà de l’aéromodélisme 3, limité quant à lui à un cadre précis – constitue bien une catégorie d’aéronef au sens du Code de l’aviation civile 4. S’il est une pratique ancienne (l’anglicisme « drone » provient d’un avion-cible automatisé « bourdonnant » – droning – utilisé outre-Manche dès les années 1930), La nouvelle réglementation sur les drones fait évoluer le statut juridique du télépilote. Elle distingue le drone professionnel du drone de loisir et elle vient renforcer les obligations préventives à la charge de ce dernier, avec une obligation de formation le faisant évoluer vers une professionnalisation. En matière répressive, de nouvelles infractions sont apparues. Si cela a pour objet de prévenir efficacement les incidents, les accidents et les risques encourus par sa pratique, le volet répressif de la réforme se fonde toutefois sur des sources dispersées et difficiles d’accès pour le télépilote de loisirs. Cela a pour conséquence de faire peser sur cette activité un risque pénal fort et encore largement sous-estimé par les praticiens.Mot clés : drone ; aviation civile ; aéronef ; télépilote ; exploitant ; sûreté ; sécurité ; prévention ; répression ; professionnalisation ; Code pénal ; Code des transports ; ordre public ; libertés fondamentales ; incident ; accident ; infraction ; risque pénal ; sécurité juridique ; complexification ; risque ; imprudence ; négligence ; données.L'Johanne GOJKOVIC-LETTEColonel, commandant la section de recherches de la gendarmerie des transports aériens. Diplômé de l’École spéciale militaire de Saint-Cyr et de l’École de guerre. Il a servi dans plusieurs sections de recherches et à la sous-direction de la police judiciaire de la direction générale de la gendarmerie nationale. Il préside le groupe « Aviation générale et Stupéfiants » du Groupe Pompidou du Conseil de l’Europe.Grégory HOUILLONMaître de conférences en droit public à l’université de Poitiers, commandant de réserve à la Gendarmerie des transports aériens (GTA). Il s’est spécialisé en droit américain (American Society for Public Administration) et en droit public économique. Ses travaux actuels portent sur la pénalisation et l’encadrement juridique de la vie politique et sur les aspects juridiques des questions aéronautiques civiles. Il a publié deux ouvrages : le Lobbying en droit public (Bruylant, 2012) puis Lobbying. Du déni au défi (Éd. Littéraires, 2017). Il a en outre publié une trentaine d’articles dans les revues juridiques et aéronautiques et contribué à de nombreux ouvrages collectifs.(1) Les opinions exprimées dans cet article n’engagent que les auteurs et ne représentent pas forcément les opinions des institutions dans lesquelles ils servent.(2) Décret n° 2018-375 du 18 mai 2018 relatif à la formation exigée des télépilotes d’aéronefs sans personne à bord à des fins de loisir, J.O.R.F., 20 mai, n° 33 ; loi n° 2016-1428 du 24 octobre 2016 relative au renforcement de la sécurité de l’usage des drones civils, J.O.R.F., n° 249, 25 oct., n° 1.(3) Arrêté « aéronefs » du 17 décembre 2015 relatif à la conception des aéronefs civils qui circulent sans personne à bord, aux conditions de leur emploi et aux capacités requises des personnes qui les utilisent, modifié par l’arrêté du 18 mai 2018, J.O.R.F., n° 129, 7 juin, n° 33, Art. 3, 1. Définissant l’aéromodélisme comme utilisé « à des fins de loisirs et de compétition ». La loi du 24 octobre 2016 a veillé à protéger l’aéromodélisme, pratique traditionnelle, exclusivement limitée au loisir ou à la compétition et antérieure à l’apparition du drone, en prévoyant une exemption aux obligations mis à la charge des télépilotes de drones, pour les aéronefs sans personne à bord qui « sont opérés dans un cadre agréé et dans les zones identifiées à cet effet » (D.S.A.C., 2017).(4) Art. L. 6100-1 C. Transp. et art. R. 133-1-2 C. Av. Civ.© Cahiers de la sécurité et de la justice - IHEMI60 I DOSSIER Prévention et répression de la pratique du drone civil : un équilibre à trouver – Johanne Gojkovic-Lette, Grégory Houillonson économie est actuellement en plein essor. Depuis 2012, la réglementation du drone civil utilise une summa divisio et distingue donc le drone à usage professionnel 5 du drone a usage de loisir. Le drone professionnel représente un marché de 10 milliards d’euros en Europe et de 180 millions d’euros en France à l’horizon 2020 avec 400 000 pièces vendues en 2016, pour 7 millions de drones de loisir en circulation (Cormier, 2018). Ce secteur constitue donc économiquement un réservoir de croissance en Europe (EASA, 2016b), qu’il convient de préserver et d’accompagner au mieux et non de « tuer dans l’œuf », et ce d’autant plus que la France peut s’affirmer comme l’un des leaders de cet écosystème. Le drone, juridiquement qualifié d’« aéronef circulant sans personne à bord »6, offre en effet un potentiel important en termes d’utilisations et de services pour les forces de sécurité (Hanicotte, 2014 : 6 ; Sandviket Jumbert, 2015 : 139 ; Pupin, 2018). Mais, au-delà des apports de cette nouvelle technologie, cette pratique croissante doit impérativement se concilier avec l’ordre public et la réglementation aéronautique. Le drone permet en effet le survol de zones qui seraient plus complexes à parcourir avec un aéronef classique et en outre d’acquérir de grandes quantités de données – pan important de son usage professionnel – ce qui pose aussi la question de la captation illicite d’informations ou de données privées7 ainsi que de l’atteinte au respect de la vie privée8, infractions facilitées par l’usage du drone et qui demeurent difficiles à sanctionner dans ce cadre (Geffray, 2015). Le drone professionnel est même devenu un outil dans l’exercice de la répression dans des zones très difficiles à contrôler, comme dans l’utilisation d’« aéronefs pilotés à distance (APAD) » équipés d’un dispositif de surveillance (Zema. 2019). Afin de préserver son apport technologique et limiter au maximum les risques, le législateur a conjugué les mesures préventives et répressives. D’une part, la prévention a été largement renforcée depuis les précédentes réglementations9 et créée désormais des obligations à la charge des télépilotes de loisir, qui s’apparentent à une certaine professionnalisation de la pratique (I). D’autre part, le législateur est venu aussi en renforcer la répression en créant de nouvelles infractions propres s’ajoutant aux infractions aéronautiques existantes, renforçant ce que le télépilote peut percevoir comme un risque pénal (II).La prévention par une professionnalisation accrue des pratiquesLe volet préventif de la réglementation du drone civil vise à sécuriser les nouvelles pratiques croissantes, notamment non professionnelles, afin d’éviter risques, incidents et accidents (A). Mais cette volonté de prévention, louable, instaure néanmoins un régime lourd d’obligations au télépilote de loisir en termes de formation ce qui le rapproche indéniablement du professionnel (B).Les nécessités de l’ordre public et la sécurisation des pratiquesLes incidents impliquant des drones sont croissants à mesure de l’expansion de la pratique. Ainsi, ce sont 2100 incidents qui ont été signalés entre 2010 et 2016 (EASA, 2016a : 6 10). Aussi, les atteintes à la sécurité intérieure sont régulières : survols réguliers de zones sensibles ou interdites de survol 11, de secteurs ou d’opérateurs d’importance vitale 12 (zones militaires 13, agglomérations, sites Seveso ou centrales), par exemple par des ONG 14 ; repérages préalables à une évasion 15 ; introduction d’objet non autorisé en milieu pénitentiaire 16; perturbation du travail (5) L’arrêté « espaces » du 17 décembre 2015 relatif à l’utilisation de l’espace aérien par les aéronefs qui circulent sans personne à bord, J.O.R.F., n° 298, 24 déc., p. 23890, modifié par l’arrêté du 30 mars 2017, J.O.R.F. n° 84, 8 avr., n° 5., Annexe 3, § 1.3.(6) La convention de Chicago du 7 décembre 1944, relative à l’aviation civile internationale, art. 8, fait référence aux « aéronefs sans pilote ».(7) Art. 226-1 et 226-2 C. pén.(8) Article 9, al. 1er C. Civ. Voir, sur cette problématique, Archambault et Mazouz (2016) et Hanicotte (2014).(9) Arrêté « aéronefs », préc.; arrêté « espaces », préc., consolidés.(10) Voir aussi le site internet dronologue.fr, rubrique Sécurité.(11) Cas du survol du Fort de Brégançon le 6 août 2018, neutralisé par brouillage d’onde.(12) L’annexe de l’arrêté du 2 juin 2006, fixant la liste des secteurs d’activités d’importance vitale et désignant les ministres coordonnateurs desdits secteurs, J.O.R.F., n° 129, 4 juin, p. 8502, modifié par l’arrêté du 2 juillet 2008, J.O.R.F., n° 156, 5 juil., p. 10823, définit ainsi 12 secteurs et 200 opérateurs.(13) Cas d’un drone qui s’est endommagé dans la cour de la caserne Thiry à Nancy le 26 juin 2018, ou encore du survol du quartier de gendarmerie Deflandre à Dijon le 31 juillet 2016.(14) Les cas de survols de sites militaires, nucléaires, industriels, urbains ou aéroportuaires explosent : 18 en octobre 2014, dont six centrales nucléaires survolées simultanément dans la seule nuit du 31 octobre 2014 ; 20 en novembre 2014 ; 11 en janvier 2015, 9 en mars 2015 (SGDSN, 2015 : 43). Certains survols ont en outre été fortement médiatisés, portant atteinte à la crédibilité de l’action de l’État (pour les derniers d’entre eux : Chinon, 18 juin 2017 et un drone volontairement écrasé par une ONG contre l’enceinte de la centrale du Bugey (Ain) le 3 juillet 2018).(15) Le ministère de la Justice a évoqué un repérage préalable par drones lors de l’évasion médiatique de Rédoine Faïd de la prison de Réau le 1er juillet 2018.© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 61 Cahiers de la sécurité et de la justice – Hors-série des services de recherche et de secours par un télépilote curieux 17 ; captation d’images ou d’informations 18 ; chute dommageable pour les tiers au sol ; pénétration dans des zones de sûreté ; atteinte à l’intégrité d’aéronefs ou perturbation de la circulation aérienne 19 ; sans compter le risque potentiel de détournement à des fins de projectile ou même d’attentat 20. En ce qui concerne les survols non autorisés, la jurisprudence du juge répressif est aujourd’hui établie sur ce point 21.Face à ces risques, la loi du 24 octobre 2016 est venue instaurer une obligation d’enregistrement par voie élec-tronique 22 auprès de l’autorité de contrôle de la circula-tion aérienne de tout drone civil d’une masse supérieure à 800 g, ainsi qu’une obligation d’immatriculation pour les drones à usage professionnel 23 d’une masse supérieure à 25 kg24, avec obligation d’information sur la nationalité 25. Le décret du 11 octobre 2018 et l’arrêté du 19 octobre 2018 sont venus préciser les modalités d’enregistrement qui seront obligatoires au 26 décembre 2018, addition-nées des informations enregistrées (type, constructeur, modèle, n° de série, équipement – capteur, calculateur, caméra – informations personnelles du propriétaire ainsi que l’accès et les corrections à ces informations, obligation de rendre le numéro d’enregistrement visible et accessible sur l’aéronef) 26. On le voit, le seuil de 800 g est détermi-nant des obligations qui seront à la charge du télépilote de loisir, créant ainsi une nouvelle infraction pour défaut d’enregistrement,déjà applicable aux aéronefs en général circulant sans certificats d’immatriculation 27. S’il a été très discuté car il permettait d’exclure la plupart des télépilotes de loisir d’un formalisme excessif, il sera peut être amené à évoluer : les institutions européennes réfléchissent à un nouvel indicateur qui pourrait faire baisser ce seuil 28.Les enjeux d’ordre public et de défense nationale se heurtent ainsi aux libertés fondamentales (liberté d’aller et venir, droit de propriété) qui sont présentes dans l’exercice du télépilotage (Hanicotte, 2014). Il a fallu en effet les concilier avec les restrictions de police nécessaires à la sûreté de l’espace aérien, des installations civiles et militaires et à la sécurité des personnes et des biens. L’actuelle réglementation a opté pour une conciliation conduisant, dans une optique de sécurisation, à une pratique toujours plus professionnalisée du drone tout en réprimant davantage ; omettant peut-être l’intelligibilité nécessaire à la bonne application de telles règles techniques. Au-delà de ces obligations déclaratives, les drones de plus de 800 g déclarés à partir du 1er juillet 2018 29 doivent désormais être équipés d’un dispositif électronique ou (16) Art. 434-35 à 36 du C. Pén. ; Voir aussi « Un drone parvient à s’introduire dans la cour d’une prison de Valence », le Monde, 16 août 2017 ; Normand, J.-M. « Les drones se bousculent au-dessus des prisons », le Monde, 16 mai 2018. (17) TGI Bayonne, Ord. 14 juillet 2014, n° parquet 1410000004, n° minute 864/2014 : vol perturbant un hélicoptère de secours chargé d’hélitreuiller les victimes d’un naufrage en mer (quatre mois d’emprisonnement avec sursis pour l’auteur et son complice ; contravention de 38 euros et confiscation de l’appareil) (18) Art. 226-1 et 226-2 C. pén.(19) Cas de l’Airbus A320 d’Air France Barcelone-Paris qui a croisé le 19 février 2016 un drone à 1600 m d’altitude en approche de l’aéroport de Roissy-CDG, passant à quelques mètres de son aile gauche.(20) Attentats d’Erbil, septembre 2016 ; Mossoul, novembre 2016 et contre le président Vénézuélien Nicolas Maduro le 4 août 2018. À partir de 1 kg, un drone est en mesure d’emporter une grenade légère (SGDSN, 2015 : 31). Le 15 septembre 2013, un drone a survolé la foule, avant d’atterrir à quelques mètres de la chancelière Angela Merkel. Le 14 octobre 2014, un match de football entre la Serbie et l’Albanie a été interrompu à la suite du survol du stade de Belgrade par un drone transportant un drapeau pro-albanais. En avril 2015, un drone transportant du sable radioactif, en quantité insuffisante pour présenter un danger, s’est posé sur le toit de la résidence officielle du Premier ministre japonais Shinzo Abe, à Tokyo.(21) Sont sanctionnées l’absence d’autorisation du vol, la non-conformité de l’aéronef aux règles de sécurité et le non-respect des conditions de navigabilité (Lepage, 2014) : TGI Paris, Ord. 20 février 2014, n° parquet 14051000548, n° minute 369 ; TGI Nancy, Ord. 20 mai 2014, n° parquet 1406000068, n° minute 173/14 ; TGI Paris, Ord. 2 octobre 2014, n° parquet 14275000307, n° minute 1682 ; TGI Paris, Ord. 13 mars 2015, n° parquet 15058000357, n° minute 49, TGI Bayonne, Ord. 14 juillet 2014, préc., et CA Bourges, 8 juin 2017 (dispense de peine pour le survol de la centrale de Belleville-sur-Loire). (22) Art. L. 6111-1, II, al. 2 C. Transp.(23) En ce qui concerne la pratique du drone professionnel, la D.S.A.C. distingue le télépilote (arrêté « aéronef » consolidé », art. 2, 2) de l’exploitant. Ce dernier est la personne (société, association, entreprise individuelle, particulier) responsable de l’activité, alors que le télépilote est la personne qui contrôle l’aéronef et réalise le vol pour le compte de l’exploitant. Dans l’hypothèse d’un exploitant unipersonnel, l’exploitant et le télépilote sont alors confondus (D.S.A.C., 2017 : 6).(24) L. 6111-1, I et II, al. 1er C. Transp., et réprimé par les art. L. 6142-4 à 6 C. Transp.(25) L. 6111-2 C. Transp.(26) Décret n° 2018-882 du 11 octobre 2018 relatif à l’enregistrement des aéronefs civils circulant sans personne à bord, J.O.R.F., n° 237, 13 oct. n° 41 ; arrêté du 19 octobre 2018 relatif à l’enregistrement des aéronefs civils circulant sans personne à bord, J.O.R.F. n° 248, 26 oct. 2018, n° 46.(27) Art. L. 6142-5 C. Transp. (pour l’infraction générale). Voir, plus particulièrement pour le drone : art. L. 6111-1, II, al. 2 C. Transp., préc.; art. D. 124-1 C. Av. Civ. et arrêté du 19 octobre 2018, préc.(28) Voir, Infra.(29) Art. 4, III de la loi 2016-1428 du 24 octobre 2016, préc. (seuil de 800 g). Pour les drones de plus de 800 g, l’entrée en vigueur est repoussée au 1er janvier 2019.© Cahiers de la sécurité et de la justice - IHEMI62 I DOSSIER Prévention et répression de la pratique du drone civil : un équilibre à trouver – Johanne Gojkovic-Lette, Grégory Houillonnumérique et de signalement sonore destiné à la sécurité de tiers. Toute dérogation à ces règles (survol, équipement, etc.) sera à notifier au préfet territorialement compétent. À ces nombreuses obligations s’ajoute l’insertion dans le Code de la consommation de l’article L. 425-1 mettant à la charge des fabricants (d’aéronefs mais aussi de pièces détachées) une obligation d’information destinée à l’acheteur afin de prévenir les incidents 30, à travers une « notice » contenant les règles et principes d’utilisation. Si cette obligation d’information permet incontestablement un progrès dans la responsabilisation des fabricants, sa portée peut toutefois se révéler très limitée quant à son pouvoir préventif. Elle contribue en effet à transférer de fait cette responsabilité sur l’utilisateur final à travers une nouvelle contrainte de lecture qui n’est pas systématique.Mais si cette première législation avait vocation à mettre en œuvre les premières mesures de prévention, elle ne pouvait faire l’impasse sur un certain niveau de formation.Une prévention par une professionnalisation progressiveEn aéronautique, un adage affirme que « le pilotage privé est un loisir qui requiert une attitude de professionnel » (Brucker, 2013 : 6) et cette formule s’applique également pour le (télé)pilotage. En conséquence, depuis juillet 2018, en matière de drone, la distinction entre profession et loisir tend à s’estomper davantage, sans toutefois se confondre. Jusqu’à présent, seuls les professionnels étaient soumis à des obligations de déclaration d’activité (en sus de la déclaration préfectorale à opérer par activité particulière et pour obtenir des dérogations aux interdictions de survol), de formation (examen théorique d’une licence U.L.M., planeur, avion), de rédaction d’un manuel d’activité particulière auxquelles peuvent s’ajouter la « déclaration de niveau de compétences » (D.N.C.) de l’exploitant et une attestation de compétence pour les drones de plus de 25 kg (avec 100 heures de vol en tant que commandant de bord). Désormais, depuis le décret du 18 mai 2018, même les drones de loisir de plus de 800 g – jusque-là limités à dix règles émises par la direction générale de l’aviation civile (D.G.A.C.) à titre informatif 31 ainsi qu’une obligation préalable de vérification des zones interdites de survol sur le site Géoportail 32 – sont soumis à une obligation de formation 33 se composant d’un questionnaire en ligne portant sur les connaissances théoriques précises à valider, donnant lieu, en cas de réussite, à l’inscription au registre des télépilotes tenu par la D.G.A.C. La réussite de cette formation permet désormais l’obtention, toujours selon le décret du 18 mai 2018 34, d’une « attestation de suivi de formation » 35. Le contenu et les modalités de cette formation (programme détaillé des connaissances théoriques à acquérir au cours de la formation), qui seront dispensés par une fédération d’aéromodélisme reconnue au plan national ou une fédération multisport incluant l’aéromodélisme agréée par le ministre 36, sont en outre précisés 37 par l’arrêté du 12 octobre 2018 38 entrant en application le 26 décembre 2018. Cet arrêté précise aussi les conditions d’âge liées à l’exercice d’une fonction de télépilotage ; la durée de validité de l’attestation de suivi de formation ainsi que les documents nécessaires au télépilote lorsqu’il utilise son drone 39. Pour ce faire, et en vertu de l’arrêté du 19 octobre 2018 40, la D.G.A.C. a créé la plate-forme internet d’enregistrement et de formation Fox-AlphaTango 41. La professionnalisation devient donc presque systématique pour les drones de loisir dès lors qu’ils dépassent 800 g, poids rapidement atteint avec une charge utile. Par ailleurs, s’ajoute à ces dispositions un aspect déterminant de la prévention en aéronautique : celui du « retour d’expérience » ou REX. L’amélioration de la qualité et de la sécurité des opérations des drones civils repose fortement sur l’implication des pilotes, des exploitants et des constructeurs à la suite d’accidents ou incidents ayant un impact sur la sécurité de l’appareil lors de son utilisation. Pour les protagonistes, le REX va consister à rendre compte de l’évènement survenu afin d’en permettre une analyse et d’en informer la D.S.A.C. (30) Art. L. 425-1 C. Consommation. (31) Document D.G.A.C. « Usage d’un drone de loisir ».(32) https://www.geoportail.gouv.fr/donnees/restrictions-pour-drones-de-loisir(33) L. 6214-2 C. Transp.(34) Décret n° 2018-375, préc.(35) D. 136-8, al. 2 C. Av. Civ.(36) D. 136-10 C. Av. Civ.(37) D. 136-11 C. Av. Civ.(38) Arrêté du 12 octobre 2018 relatif à la formation exigée des télépilotes qui utilisent des aéronefs civils circulant sans personne à bord à des fins de loisir, J.O.R.F., n° 248 du 26 oct. 2018, n° 45.(39) D. 136-11 C. Av. Civ., préc.(40) Arrêté du 19 octobre 2018, préc., art. 2.(41) https://fox-alphatango.aviation-civile.gouv.fr/© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 63 Cahiers de la sécurité et de la justice – Hors-série ainsi que le titulaire de l’attestation de conception de type (fabricant) via un formulaire de notification 42. Le REX est déterminant dans l’aspect préventif en ce qu’il permet d’accompagner l’appropriation d’une culture de sécurité en accord avec la réglementation, notamment l’arrêté du 17 décembre 2015 43 dans son annexe III, article 3.5.6. Ainsi, le compte-rendu de tout évènement qui aura mis ou aurait pu compromettre la sécurité des tiers devra être transmis à la D.S.A.C.. De même, le constructeur ou fabricant devra lui aussi être informé de toute défaillance de sécurité. Seront ainsi, par exemple, notifiés à la D.S.A.C. l’approche d’un groupe de personnes par suite d’un défaut d’attention du pilote ; l’approche d’un autre aéronef en vol sans défaillance spécifique ; une destruction à l’atterrissage à la suite d’un comportement inattendu ; la pénétration dans un espace interdit sans défaillance spécifique, etc. D’autres évènements seront notifiés à la fois à la D.S.A.C. et au constructeur, citons par exemple le dépassement du seuil des 150 m du sol par suite d’une perte des commandes, le survol d’un territoire non prévu dans l’étude de sécurité ou encore un écrasement. Une analyse doit impérativement être jointe au REX, elle contribue à son aspect préventif vis-à-vis de la sécurité des vols (D.G.A.C., 2016 : 10).Cet aspect préventif est en outre renforcé par une notion spécifique au secteur aéronautique : celle de la « culture juste » issue des règlements européens 44. Cette notion de culture juste permet le renforcement préventif de la sécurité et vise à engendrer les conditions favorables à la notification des praticiens. En somme, elle part de l’idée qu’un climat de confiance permet de renforcer davantage la sécurité et la sûreté. Cette notion de « culture juste », si chère au milieu aérien, vise spécifiquement à développer cet acte réflexe à savoir, rendre compte des événements qui ont posé, ou auraient pu poser, des problèmes sécuritaires. La culture juste pousse au retour d’expérience et à l’amélioration collective des procédures. Ainsi, l’article L. 6223-2 du Code des transports prévoit qu’ « aucune sanction administrative, disciplinaire ou professionnelle ne peut être infligée à la personne qui a rendu compte d’un événement […], qu’elle ait été ou non impliquée dans cet événement, sauf si elle s’est elle-même rendue coupable d’un manquement délibéré ou répété aux règles de sécurité ».Ainsi, s’il faut saluer l’aspect préventif pour la sécurité publique de ces nouvelles mesures, il en va différemment de ses aspects répressifs. Le cadre juridique relatif à l’utilisation du drone pourrait en effet se voir davantage rationalisé.L’accroissement de la répression et du risque pénal pour le télépiloteLes mesures répressives du régime juridique du drone se complexifient progressivement, et apparaissent, qui plus est, très dispersées, nuisant à leur lisibilité et à leur accessibilité (A). Une telle complexité et les contraintes qui en découlent contribuent ainsi à faire peser un risque pénal fort sur l’opérateur en fragilisant, au nom du primat de l’ordre public, la sécurité juridique des diverses pratiques (B).Complexification et difficile accès à la règle pénaleLa réglementation, malgré un projet européen d’harmonisation entrant en vigueur en juillet 2020 et destiné à se substituer progressivement aux législations sur une période de trois ans 45, n’est pas encore uniforme en l’état et se voit dispersée tant dans la hiérarchie des normes (loi du 24 octobre 2016 relative au renforcement de la sécurité de l’usage des drones civils, arrêté du 17 décembre 2015 relatif à la conception des drones et à l’utilisation de l’espace aérien 46, arrêté du 12 octobre 2018 fixant la liste des zones interdites à la prise de vue 47 ; décrets du 18 mai 2018 relatifs à la formation des télépilotes de loisir 48 et au seuil de masse 49 ; décret du 19 avril 2019 relatif (42) https://www.ecologique-solidaire.gouv.fr/drones-usages-professionnels(43) Préc.(44) Règlement (UE) n° 376/2014 du 3 avril 2014 concernant les comptes rendus, l’analyse et le suivi d’événements dans l’aviation civile, modifiant le règlement (UE) n ° 996/2010, art. 2, (12).(45) Règlement délégué (UE) 2019/945 de la Commission du 12 mars 2019 relatif aux systèmes d’aéronefs sans équipage à bord et aux exploitants issus de pays tiers, de systèmes d’aéronefs sans équipage à bord et Règlement d’exécution (UE) 2019/947 de la Commission du 24 mai 2019 concernant les règles et procédures applicables à l’exploitation d’aéronefs sans équipage à bord ; J.O.R.F.CE, 11 juin 2019, n° L 152.(46) Préc.(47) Arrêté du 27 janvier 2017 fixant la liste des zones interdites à la prise de vue aérienne par appareil photographique, cinématographique ou tout autre capteur, J.O.R.F., n° 25, 29 janv., texte n° 1.(48) Préc.(49) Décret n° 2018-374 du 18 mai 2018 relatif aux seuils de masse prévus par la loi n° 2016-1428 du 24 octobre 2016 relative au renforcement de la sécurité de l’usage des drones civils, J.O.R.F., n° 115, 20 mai, texte n° 32.© Cahiers de la sécurité et de la justice - IHEMI64 I DOSSIER Prévention et répression de la pratique du drone civil : un équilibre à trouver – Johanne Gojkovic-Lette, Grégory Houillonà la notice d’information relative à l’usage des aéronefs circulant sans personne à bord 50) que dans la pluralité des sources (Code des transports, Code civil – droit au respect de la vie privée, droit à l’image et responsabilité civile – et surtout Code pénal). La pratique du drone se retrouve ainsi, matériellement, au carrefour de plusieurs branches du droit, ce qui ne facilite pas l’unité de ses règles et sa lisibilité.Plus encore, cette réglementation est amenée à évoluer pour les utilisateurs de drones, tant professionnels que de loisir. Le droit pénal des drones est en perpétuel changement, rendant son accessibilité un peu plus difficile pour le non-juriste. D’abord, le décret du 19 avril 2019 ajoute depuis juillet 2019, une nouvelle contravention pénale de 3e classe concernant la vente de pièces détachées de drone non assortie de la notice d’information rendue obligatoire par l’article 425-1 du Code de la consommation. Cette notice préventive indique les conditions d’utilisation des aéronefs circulant sans personne à bord, les règles d’utilisation de l’espace aérien et de sécurité ainsi que les règles et principes de respect de la vie privée. Ensuite, le futur règlement européen en cours d’adoption indique que les critères d’enregistrement devraient encore changer : le critère en poids serait remplacé par un critère exprimé en énergie cinétique. Ainsi, si en l’état actuel du droit, l’enregistrement et la formation sont obligatoires pour les drones de plus de 800 g, ils risquent de le devenir pour les appareils « qui, en cas d’impact sur une personne, peuvent transférer de l’énergie au-delà de 80 joules ». Or, si ce critère change de nature et prend en compte le niveau de dégâts produit lors d’une collision à pleine vitesse, il revient également à abaisser sérieusement le seuil d’enregistrement en termes de poids. En effet, les 80 joules correspondent à des drones légèrement plus lourds que 250 g. Aussi, les utilisateurs doivent encore s’attendre à une augmentation de leurs obligations. Plus largement, de tels changements, complications et aggravations des obligations pesant sur les télépilotes, conjugués à la pluralité de sources, ne vont pas renforcer la sécurité juridique du secteur (Balat et Mazouz, 2019 : 411). L’ordre public et la sécurité matérielle prévalant sur la sécurité juridique, il deviendra vraisemblablement nécessaire pour la D.G.A.C. d’opérer une compilation plus homogène de l’ensemble des règles éparses applicables et destinées à l’information des opérateurs, qui fera office de « code de la route » pour drone, et ce, de façon régulière 51. De telles initiatives, certes plus globales au secteur 52, existent déjà et la D.S.A.C. publie régulièrement un guide technique complet pour la pratique du drone, incluant notamment les aspects réglementaires (D.S.A.C., 2017). Un recueil juridique dédié manque toutefois encore à cette activité complexe. Ce sera d’autant plus nécessaire que ces réglementations sont instables et amenées à évoluer avec le temps. Dans une optique de professionnalisation progressive du télépilote au-dessus d’un certain poids, amené à baisser de surcroît, une telle initiative lui permettra de maintenir les connaissances réglementaires acquises lors de sa formation théorique. Elle permettrait de compenser le risque d’une sanction pénale qui pèse fortement sur le télépilote afin de concilier les nécessités impérieuses de l’ordre public aérien avec l’exercice d’une liberté individuelle dans une marge et une mesure « qui ne nuit pas à autrui », comme le définit la Déclaration de 1789 53.Risque pénal et insécurité juridiqueEn effet, au-delà de l’infraction propre aux télépilotes prévue aux articles L. 6232-12 et 13 du Code des transports, les drones relèvent aussi – il ne faut pas l’oublier – des infractions pénales de droit commun, prévues dans ce même code pour tous les aéronefs civils : violation des règles de sécurité 54 ; entrave à la navigation et circulation aérienne 55. Outre ces infractions, s’ajoutent celles prévues directement par le Code pénal ainsi que les autres sources du Code des transports : atteinte physique par imprudence, négligence, maladresse et inattention 56 ; mise en danger de la vie d’autrui en l’exposant à un risque immédiat de mort ou blessure de nature à entraîner une mutilation ou une infirmité permanente 57 ; vol de données et piratage 58 ; violation, captation d’une zone (50) Décret n° 2019-348 du 19 avril 2019 relatif à la notice d’information relative à l’usage des aéronefs circulant sans personne à bord, entrant en vigueur le 1er juillet 2019, J.O.R.F. n° 95, 21 avr. 2019, texte n° 34.(51) EASA, Proposition de création de règles communes pour l’utilisation opérationnelle des drones en Europe, n° A-NPA 2015-10, sept. 2015 (33 propositions).(52) L’exemple du RADIC est particulièrement éloquent : D.G.A.C. (2017), Recueil des arrêtés, décisions, instructions et circulaires relatifs à l’aviation civile (RADIC 2018), T. I, Ministère de la transition écologique et solidaire.(53) Art. 4 de la Déclaration des droits de l’Homme et du citoyen de 1789.(54) Art. L. 6232-2 C. Transp.(55) Art. L. 6372-4 C. Transp.(56) Art. 222-19 C. pén.(57) Art. 223-1 C. pén.(58) Art. 323-1 C. pén.(59) Art. D. 133-10 C. Av. Civ.(60) Art. 413-9 et s. C. pén.© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 65 Cahiers de la sécurité et de la justice – Hors-série sensible 59 ou d’informations relevant du secret de la défense nationale 60 et atteinte à la vie privée 61 ou encore utilisation d’une fréquence non autorisée 62. Autant dire que pèse sur le télépilote, mais aussi sur l’exploitant (s’il est différent) et potentiellement sur le fabricant et le distributeur (sans compter l’hypothèse de la complicité) un nombre d’infractions particulièrement important. Un tel cadre est immanquablement générateur d’un risque pénal toujours croissant, puisque la loi du 24 octobre 2016 était venue ajouter une infraction spécifique à travers deux nouvelles dispositions au corpus répressif de droit commun (obligation d’enregistrement et obligation d’immatriculation, voir I.A 63). Qui plus est, la présence, au-delà du Code pénal, d’infractions de nature pénale – dont l’une d’entre elles propre au drone 64 – au sein du Code des transports ne vient pas favoriser leur visibilité pour l’opérateur. Cette réglementation, complexe et technique, pose ainsi une question d’accessibilité et d’intelligibilité de la loi, objectif à valeur constitutionnelle pourtant censé guider le législateur 65.En définitive, les activités du drone civil gagnent en sécurité matérielle avec des pratiques menées dans un cadre strict et conformes à l’ordre public, mais perdent en sécurité juridique dès lors que la moindre pratique peut générer un doute permanent pour l’opérateur quant au risque pénal potentiellement encouru et à la complexité des règles devenues applicables. Cela vaut également lorsque la formation certifiée de l’opérateur prend en compte la réglementation pénale, en raison de son inhérente complexité.Le curseur de l’équilibre entre ordre public aérien et pratique raisonnée reste donc encore à trouver dans ce secteur tout aussi innovant qu’évolutif – inévitable face à la pénalisation progressive d’un secteur technique, couplée à une réglementation aussi complexe n(61) Art. 226-1 et 2 C. pén.(62) Art. L. 39-1 C. P.et T(63) Loi n° 2016-1428 du 24 octobre 2016, préc., art. 5, insérant ces deux articles au sein d’une section 6 « aéronefs circulant sans personne à bord » dans le Ch. 2, du Titre III, Livre II de la 6e partie du C. Transp.(64) Art. L. 6232-12 et 13, préc.(65) Décision n° 99-421 DC du 16 décembre 1999, Rec. p. 136.RéférencesOuvrages et articlesArchambault, Laurent et Alicia Mazouz, 2015, « L’envol des drones civils : Appréhension par le droit français d’une pratique émergente », Annals of Air and space Law, vol. XL.Archambault, Laurent et Alicia Mazouz, 2016, « Quel horizon juridique pour les drones civils ? », Gaz. Pal., 21 juin 2016, p.19.Balat, Nicolas et Alicia Mazouz, 2019, « Les drones civils et la sécurité juridique », in Siguoirt, Laurent (dir.), Transports et sécurité, Lexis Nexis, p. 411.Cormier, Henri, 2018, « Drones. Vers un renforcement législatif », Aviation civile n° 383.Geffray, Edouard, 2015, « Quel cadre juridique pour les drones aériens civils ? », in CESE, 2015, « Les drones aériens civils : opportunités et risques », colloque du 28 mai 2015.Gojkovic, Johanne et Grégory Houillon, 2019, « La pratique du drone, entre sécurisation et répression », AJ-Pénal, n° 3, p.135.Hanicotte, Robert, 2014, « Une nouvelle catégorie d’OVNI juridique : les drones », Gazette du Palais, n° 317, p.6.Lepage, Agathe, 2014, « Première rencontre du droit pénal avec un drone », Lexis Nexis, Comm. Com. électr., n° 7/8, comm. n° 65.Pupin, Gaëlle, 2018, « Drones : les gendarmes investissent la troisième dimension », GendInfo, (19 mars).Sandvik, Kristin Bergtora et Jumbert Maria Gabrielsen, 2015, « Les drones humanitaires », RIS, n° 98, p. 139 ; Zema, Alexis, 2019, « En France, des drones effectuent désormais des contrôles routiers », Le Figaro, (17 mai).© Cahiers de la sécurité et de la justice - IHEMI66 I DOSSIER Prévention et répression de la pratique du drone civil : un équilibre à trouver – Johanne Gojkovic-Lette, Grégory HouillonComptes-rendus et rapports officielsCommission européenne, communication du 8 avril 2014, Une nouvelle ère de l’aviation. Ouvrir le marché de l’aviation à l’utilisation civile de systèmes d’aéronefs télépilotés, d’une manière sûre et durable, COM(2014) 207.Parlement européen, 2015, rapport Sur l’utilisation sure d’aéronefs autopilotés (RPAS) plus connus sous le nom de véhicules aériens sans pilotes (UAV) dans le domaine de l’aviation civile, n° 2014/2243/INI.SGDSN, 2015, l’Essor des drones aériens civils en France : enjeux et réponses possible de l’État, rapport du Gouvernement au Parlement, SGDSN 20 octobre 2015;EASA, 2015, Proposition de création de règles communes pour l’utilisation opérationnelle des drones en Europe, n° A-NPA 2015-10.EASA, 2016a, « Drone collision » Task Force, Final Report, 4 octobre 2016.EASA, 2016b, High Level Conference on « Drones as a leverage for jobs and new business opportunities », Varsovie, 24 novembre 2016.D.G.A.C., 2016, Incidents de drones – notification et suivi, mai 2016.D.S.A.C., 2017, Aéromodélisme : modèles réduits et drones de loisir, 1ère éd., V. 2, (10 janvier).D.S.A.C., 2017, Aéronefs circulant sans personne à bord : activités particulières, 1ère éd., V. 3, (28 août).Textes et réglementationConvention de Chicago du 7 décembre 1944, relative à l’aviation civile internationale. Règlement (UE) n° 376/2014 du 3 avril 2014 concernant les comptes rendus, l’analyse et le suivi d’événements dans l’aviation civile. Arrêté » aéronef » du 17 décembre 2015 relatif à la conception des aéronefs civils qui circulent sans personne à bord, aux conditions de leur emploi et aux capacités requises des personnes qui les utilisent, J.O.R.F., n° 129.Arrêté « espaces » du 17 décembre 2015 relatif à l’utilisation de l’espace aérien par les aéronefs qui circulent sans personne à bord, J.O.R.F., n° 298, p. 23890Loi n° 2016-1428 du 24 octobre 2016 relative au renforcement de la sécurité de l’usage des drones civils, J.O.R.F., n° 249.Arrêté du 27 janvier 2017 fixant la liste des zones interdites à la prise de vue aérienne par appareil photographique, cinématographique ou tout autre capteur, J.O.R.F., n° 25.Décret n° 2018-374 du 18 mai 2018 relatif aux seuils de masse prévus par la loi n° 2016-1428 du 24 octobre 2016, J.O.R.F., n° 115.Décret n° 2018-375 du 18 mai 2018 relatif à la formation exigée des télépilotes d’aéronefs sans personne à bord à des fins de loisir, J.O.R.F., 20 mai, n° 33. Décret n° 2018-882 du 11 octobre 2018 relatif à l’enregistrement des aéronefs civils circulant sans personne à bord, J.O.R.F., n° 237. Arrêté du 12 octobre 2018 relatif à la formation exigée des télépilotes qui utilisent des aéronefs civils circulant sans personne à bord à des fins de loisir, J.O.R.F., n° 248 du 26 oct. 2018, n° 45.Arrêté du 19 octobre 2018 relatif à l’enregistrement des aéronefs civils circulant sans personne à bord, J.O.R.F. n° 248, 26 oct. 2018, n° 46.Décret n° 2019-348 du 19 avril 2019 relatif à la notice d’information relative à l’usage des aéronefs circulant sans personne à bord, J.O.R.F. n° 95.© Cahiers de la sécurité et de la justice - IHEMICahiers de la sécurité et de la justice – Hors-série DOSSIER I 67Prevention and repression of civil drone use: finding a balance 1Johanne GOJKOVIC-LETTE, Grégory HOUILLONn 2018 was implemented a whole new set of regulations related to the use of civil drones 2. Drones, also known as unmanned aerial vehicles (UAV) or remotely-piloted aircraft systems (RPAS) – much more than toys or gadgets, and far beyond plane model-building 3, which is limited to a specific legal framework – constitute an aircraft category as defined by the Code of Civil Aviation 4. Although drones are nothing new (the term was first used to refer to a buzzing – droning – pilotless target aircraft used in the UK as early as in the 1930s), the drone economy is on the rise. The new regulations on drones have transformed the legal status of drone pilots. They created a distinction between professional use and recreational use, increased legal obligations, and made training compulsory, pushing towards professionalization. New offenses have been created. Although the new law aims at effectively preventing incidents and accidents and reducing the risks related to drone use, its penal dimension originates from scattered sources that are difficult to comprehend for recreational users. The activity is thus burdened with heavy legal liabilities, which are still vastly underestimated by users.Keywords: drone; civil aviation; aeronautical regulations; aircraft; remote pilot; operator; security; safety; prevention; repression; professionalization; Penal Code; Code of Transports; transport regulations; law enforcement (public order); fundamental rights and liberties; incident; accident - crash; offense; legal risk; legal uncertainty; complexification; hazard; imprudence; negligence; data.IJohanne GOJKOVIC-LETTEColonel, commanding the search section of the Air Transport Gendarmerie. Graduated from the École spéciale militaire de Saint-Cyr and the École de guerre. He served in several search sections and in the judicial police department of the Directorate-General of the National Gendarmerie. He presides the “General aviation and drugs” workshop of the Pompidou Group – Council of Europe.Grégory HOUILLONLecturer in public law at the University of Poitiers and reserve commandant (major) of the Air Transport Gendarmerie. He specializes in United States law (American Society for Public Administration) and in economic public law. His works focus on the judicialization and judicial contours of the political life and on the judicial aspects regarding civil aeronautical activities. He published Lobbying en droit public (Bruylant, 2012) and Lobbying. Du déni au défi (Éd. Littéraires, 2017).-(1) The views expressed in this article are solely those of the authors and do not necessarily represent the views of the institution in which they serve.(2) Decree number 2018-375 of May 18th, 2018 regarding compulsory training for remote pilots of aircrafts circulating without anybody on board for recreational purposes, JORFRF, May 20th, number 33; law number 2016-1428 of October 24th, 2016 regarding the reinforcement of safety in civil drone use, JORFRF, number 249, October 25th, number 1.(3) “Aircraft” ruling of December 17th, 2015 regarding the conception of unmanned civil aircrafts circulating without anybody on board, the conditions of their utilization, and the abilities required from the persons who use them, modified by the ruling of May 18th, JORF, number 129, June 7th, number 33, art. 3, 1. which defines plane model-building as used “with leisure and competition purposes”. The law of October 24th, 2016 ensured the protection of plane model-building, a traditional practice which is exclusively limited to leisure or competition and dates back to before the emergence of drones, by providing an exemption from the obligations imposed on drone remote pilots, for aircrafts without anyone on board that are “operated in an authorised context and in designated areas specifically approved for that purpose” (D.S.A.C., 2017).(4) Art. L. 6100-1 Code of Transports, and art. R. 133-1-2 Code of Civil Aviation.© Cahiers de la sécurité et de la justice - IHEMI68 I DOSSIER Prevention and repression of civil drone use: finding a balance – Johanne Gojkovic-Lette, Grégory HouillonSince 2012, the regulations of civil drone use have been based on a primary distinction between professional 5 and recreational drone uses. Professional drones will represent a 10 billion euro market in Europe and a 180 million euro market in France by 2020: in 2016, 400 000 items were sold, and 7 million recreational drones are currently in circulation (Cormier, 2018). This sector carries a great potential for economic growth in Europe (EASA, 2016b), which must be protected and supported in the best possible way rather than “nipped in the bud”, especially since France might very well establish itself as a leader in that ecosystem. Drones, legally defined as “aircrafts circulating without anybody on board” 6, offer indeed considerable possibilities in terms of uses and services for security forces (Hanicotte, 2014 : 6 ; Sandviket Jumbert, 2015 : 139 ; Pupin, 2018). But beyond the progress brought forward by this new technology, it is imperative that the emerging practice must be conciliated with public order and aeronautical regulations. Indeed, drones can fly over zones over which a classical aircraft wouldn’t be able to fly so easily. Besides, they are able to collect a great amount of data, which constitutes an important part of its professional use, but also raises the question of the unlawful obtaining of personal information 7 and other infringements or violations of privacy 8. Committing such offenses is made easier by drone use, which also makes them more difficult to prosecute and punish (Geffray, 2015). Professional drones have become a helpful tool in law enforcement and repression in zones where control is made difficult by limited access, as evidenced by the use of “remotely piloted aircraft system” (RPAS in English, APAD in French) equipped with surveillance devices (Zema. 2019). In order to preserve the potential for technological progress while simultaneously limiting the risks to a minimum, lawmakers have devised both preventive and repressive measures. On the one hand, prevention has been considerably reinforced in comparison with the previous regulations 9. Recreational remote pilots are henceforth bound by obligations which amount to a kind of professionalization (I). On the other hand, lawmakers have also reinforced the repressive component of the legislation, as they created specific new offenses in addition to the existing aeronautical violations, thereby increasing what remote pilots might perceive as a penal risk (II).Prevention of risk through the increased professionalization of practicesThe preventive aspect of the new regulations on civil drones aims at improving the security of the increasing, partly non professional new activity, in order to avoid risks, incidents and accidents (A). But the intention to prevent risks, which is commendable, establishes nonetheless a heavy training requirement for recreational remote pilots, which undeniably borders on professionalization (B).The necessities of public order and improving the safety of practicesIncidents involving drones have in fact been increasing along with the expansion of the activity. Thus 2100 incidents were reported between 2010 and 2016 (EASA, 2016a : 6 10). And violations of homeland security are frequent: flights over sensitive zones or no-fly zones 11, over sectors or operators of “vital importance” 12 (military areas 13, cities, “Seveso” industrial sites containing dangerous substances, or nuclear power plants), sometimes overflown by drones piloted by NGO activists 14; reconnaissance flights over (5) “Spaces” ruling of December 17th, 2015 regarding the use of airspace by aircrafts circulating without anybody on board, JORF, number 298, December 24th, page 23890, modified by the ruling of March 30th, 2017, JORF, number 84, April 8th, number 5, annex 3, § 1.3. (6) Article 8 of the December 7th, 1944 Chicago Convention on International Civil Aviation mentions “pilotless aircrafts”.(7) Art. 226-1 and 226-2 Penal Code.(8) Art. 9, paragraph 1 Civil Code. On that issue, see also Archambault et Mazouz (2016) et Hanicotte (2014).(9) “Aircraft” ruling, ibid.; “spaces” ruling, ibid., consolidated.(10) See also the website dronologue.fr, under the “Sécurité” section.(11) Case of the drone flight over the Fort de Brégançon (official holiday residence of the French presidency) on August 6th, 2018, neutralized by a jammer through wave interference. (12) The annex to the ruling of June 2nd, 2006, establishing the list of sectors of activity of vital importance and designating the ministers coordinating said sectors, JORF, number 129, June 4th, page 8502, modified by the ruling of July 2nd, 2008, JORF, number 156, July 5th, page 10823, identifies 12 sectors and 200 operators.(13) See the case of the drone that crashed in the courtyard of the Caserne Thiry military base in Nancy on June 26th, 2018, or the case of the drone flight over the Deflandre gendarmery headquarters on Dijon on July 31st, 2016.(14) The cases of drone flights over military, nuclear, industrial, urban or airport sites are skyrocketing: 18 in October of 2014, among which 6 drone flights over nuclear power plants in the single night of October 31st, 2014; 20 in November of 2014; 11 in January of 2015, 9 in March of 2015 (SGDSN, 2015: 43). Some of these drone flights received considerable media coverage, which undermines the credibility of the State. The most recent ones: a drone flew over the Chinon nuclear plant on June 18th, 2017, and another one was voluntarily crashed by an NGO against the Bugey nuclear plant (in the Ain department) on July 3rd, 2018.© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 69 Cahiers de la sécurité et de la justice – Hors-série prisons or jails previous to escape attempts 15; introduction of unauthorized items in penitentiaries 16; nosey drone pilots disturbing the work of first responders 17; unlawful recording of images or collection of information 18; crashes causing harm to people on the ground; trespassing into security restricted areas; attacks on the integrity of other aircrafts or disturbances of air traffic 19; not to mention the potential risk of misuses of drones employed as projectiles, or even of terrorist attacks 20. Regarding unauthorised flights, judges have established a broad repressive legal precedence21.To face these risks, the law of October 24th, 2016 established the obligation to register electronically22 any civil drone over 800g (1,7lbs) ) with air traffic control authorities, as well as the obligation for professional drones23 over 25kg (55lbs)24 to carry an identification plate, including information on the owner’s citizenship25. The decree of October 11th, 2018 and the ruling of October 19th, 2018 set the specifics of the registration, which will become compulsory by December 26th, 2018, including the information to be registered (type, manufacturer, model, serial number, equipment – sensor, calculator, camera –, personal information of the owner), as well as the conditions of access and corrections to this information, and the obligation to make the registration number visible and accessible on the aircraft26. As we can see, the 800g threshold determines the obligations recreational remote pilots are subjected to, and creates a new offense for drones: failure to register, which already applied to any other aircraft circulating without a registration certificate27. The 800g threshold was a controversial point, since it allowed to keep most recreational remote pilots away from excessive procedures, and it might very well be lowered in the future, as EU institutions are currently working on new criteria which might set the line lower28.Issues of public order and national defense have thus been confronted with the fundamental liberties (freedom of movement, property rights) exercised within the practice of remote piloting (Hanicotte, 2014). These fundamental liberties had to be conciliated with the police restrictions necessary to the safety of airspace, the security of civil and military infrastructure, and the protection of people and property. The current regulations settled for a middle ground which, in order to improve safety, led to an increased professionalization of drone use, combined with more repression. In doing so, they might have somewhat neglected the intelligibility necessary to the (15) The ministry of Justice mentioned possible “reconnaissance” drone flights previous to Redoïne Faïd’s notorious escape from the prison of Réau on July 1st, 2018.(16) Art. 434-35 to 36 Penal Code; see also “Un drone parvient à s’introduire dans la cour d’une prison de Valence”, le Monde, August 16th, 2017; Normand, J.-M. (2018) “Les drones se bousculent au-dessus des prisons”, le Monde, May 16th. (17) TGI Court of Bayonne, ruling of July 14th, 2014, prosecution number 1410000004, minute number 864/2014: a drone flight disturbed a rescue helicopter tasked with winching shipwreck victims from a sinking boat (four months suspended prison sentence for the offender and his accomplice; 38 euros fine and confiscation of the device). (18) Art. 226-1 and 226-2 Penal Code.(19) Case of the Airbus A320 Barcelona-Paris Air France flight that crossed paths with a drone at 1600 meters (1 mile) of altitude as it was approaching the Roissy-CDG airport on February 19th, 2016; the drone passed at just a few feet’s distance from its left wing.(20) Attacks in Erbil, September 2016; Mosul, November 2016; and against Venezuelan President Nicolas Maduro on August 4th, 2018 (SGDSN, 2015: 31). On September 15th, 2013, a drone flew over a crowd and landed a few meters away from German Chancellor Angela Merkel. On October 14th, 2014, a soccer game between Serbia and Albania was interrupted after a drone carrying an Albanian flag flew over the Belgrade stadium. In April of 2015, a drone carrying radioactive sand, in quantities insufficient to pose a risk, landed on the roof of Japanese Prime Minister Shinzo Abe’s official residence in Tokyo.(21) Flying without authorization, non-conformity of the aircraft with security rules, and the failure to comply with conditions of airworthiness are all penalized (Lepage. 2014): TGI Court of Paris, ruling of February 20th, 2014, prosecution number 14051000548, minute number 369; TGI Court of Nancy, ruling of May 20th, 2014, prosecution number 1406000068, minute number 173/14; TGI Court of Paris, ruling of October 2nd, 2014, prosecution number 14275000307, minute number 1682; TGI Court of Paris, ruling of March 13th, 2015, prosecution number 15058000357, minute number 49; TGI Court of Bayonne, ruling of July 14th, 2014, ibid.; and Court of Appeal of Bourges, ruling of June 8th, 2017 (exemption of punishment for the flight over the nuclear power plant of Belleville-sur-Loire). (22) Art. L. 6111-1, II, paragraph 2 Code of Transports.(23) Regarding professional drone use, the Directorate of Security for Civil Aviation (D.S.A.C.) distinguishes the remote pilot (“aircraft” ruling, consolidated, art. 2, 2) from the operator, defined as the person (company, association, private corporation, individual) responsible for the activity. The remote pilot is the person who controls the aircraft and performs the flight for the operator. In cases where the operator is a single person, operator and remote pilot are one and the same (D.S.A.C., 2017: 6).(24) L. 6111-1, I and II, paragraph 1 Code of Transports, penalized by art. L. 6142-4 to 6 Code of Transports.(25) Law 6111-2 Code of Transports.(26) Decree number 2018-882 of October 11th, 2018 regarding the registration of civil aircrafts circulating without anybody on board, JORF, number 237, October 13th, number 41; ruling of October 19th, 2018 regarding the registration of civil aircrafts circulating without anybody on board, JORF number 248, October 26th, 2018, number 46.(27) Art. L. 6142-5 Code of Transports (the general offense applying to all aircrafts). Regarding the specific offense for drones, see: art. L. 6111-1, II, paragraph 2 Code of Transports, ibid.; art. D. 124-1 Code of Civil Aviation, and decree of 19 October 2018, ibid.(28) See infra.© Cahiers de la sécurité et de la justice - IHEMI70 I DOSSIER Prevention and repression of civil drone use: finding a balance – Johanne Gojkovic-Lette, Grégory Houillonenforcement of such technical rules. Apart from these registration obligations, drones over 800g registered after July 1st, 2018 29 must from now on be equipped with an electronic or numerical sound-emitting device designed to ensure the safety of passersby. Any failure to comply with the rules (unauthorized flights, nonconformity of equipment, etc.) will be liable to prosecution within the competent territorial jurisdiction. In addition to these many obligations, a new article L. 425-1 was introduced into the Code of Consumption, which requires manufacturers (of aircrafts as well as spare parts) to provide the buyer with incident prevention information 30 via a user’s manual or leaflet. Although this obligation of information undeniably allows for a progress in fostering manufacturers’ responsibility, its impact in terms of preventive power is very limited. Indeed, it actually ends up shifting that responsibility onto the final users, requiring them to do some additional reading which they might not be used to.However, if this initial legislation was intended to implement the first measures of prevention, it couldn’t fail to include a certain amount of training as well.Prevention through progressive professionalizationIn the world of aeronautics, a well-known saying goes: “being an amateur pilot is a hobby that requires a professional attitude” (Brucker, 2013 : 6), and this also applies to remote pilots. As a consequence, since July of 2018, the distinction between professional and recreational drone use has somewhat faded, although the difference remains. Until then, only professionals were bound by obligations: they had to declare their activity (in addition to the declarations of specific activities to the police and the requests for derogations to fly over forbidden areas), to undergo training (theoretical exam of the ultralight aviation, glider or airplane license), and to draft a specific activity manual. To these, may have been added extra obligations such as the “declaration of ability level” (in French, C.N.D.) of the operator, and a declaration of abilities for drones over 25kg (with 100 flight hours as captain). But since the decree of May 18th, 2018, even recreational drones over 800g which, until then, were only bound by 10 simple rules emitted for information only by the Directorate General for Civil Aviation (D.G.A.C.)31 and by the obligation to look into the no-fly zones on the Géoportail website32, are now also obligated to undergo training33. Said training consists of an online questionnaire validating precise theoretical knowledge, after which, if passed successfully, remote pilots are registered into the list kept by the D.G.A.C. Still since the decree of May 18th, 201834, pilots who successfully complete that training also get a “training certificate”35. The training will be taught by nationally recognized federations of airplane model building, or by multisports federations including airplane model building and approved by the minister36. The content and the specifics of the training (detailed curriculum of theoretical knowledge to be acquired) are defined37 by the ruling of October 12th, 201838 which came into effect on December 26th, 2018. This ruling also specifies the age conditions required to work as a remote pilot; the validity period of the training certificate, as well as the documents that remote pilots must carry while they are operating a drone39. To that effect, and by virtue of the ruling of October 19th, 201840, the D.G.A.C. created the Fox-AlphaTango online platform41 for registration and training purposes. Professionalization is thus almost systematic for recreational drones over 800g, a threshold which is easily reached when the drone has any payload.Besides these measures, a crucial element of risk prevention in aeronautics is the experience feedback process known as “postmortem documentation”, “lessons learned”, or in French, REX or RETEX. Bettering the quality and security of civil drone operations hinges crucially on (29) Art. 4, III of law number 2016-1428 of October 24th, 2016, ibid. (the 800g limit). For drones over 800g, the regulation was postponed on January 1st, 2019.(30) Art. L. 425-1 Code of Consumption.(31) Document published by the D.G.A.C. “Usage d’un drone de loisir”.(32) https://www.geoportail.gouv.fr/donnees/restrictions-pour-drones-de-loisir (33) L. 6214-2 Code of Transports.(34) Decree number 2018-375, ibid.(35) D. 136-8, paragraph 2 Code of Civil Aviation.(36) D. 136-10 Code of Civil Aviation.(37) D. 136-11 Code of Civil Aviation.(38) Ruling of October 12th, 2018 regarding the training required for remote pilots who use civil aircrafts circulating without anybody on board for recreational purposes, JORF, number 248 of October 26th, 2018, number 45.(39) D. 136-11 Code of Civil Aviation, ibid.(40) Ruling of October 19th, 2018, ibid., art. 2. (41) https://fox-alphatango.aviation-civile.gouv.fr/© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 71 Cahiers de la sécurité et de la justice – Hors-série the involvement of pilots, operators and manufacturers following any incident or accident impacting safety that might occur during their use. Experience feedback means the people involved must report the event in order to allow for its analisis. Both the Directorate of Security for Civil Aviation (D.S.A.C.) and the manufacturer (holder of the conception certificate) should be notified through a form 42. Experience feedback is instrumental for risk prevention as it makes users accountable and fosters a culture of security, in accordance with legislation such as the ruling of December 17th, 2015 43 annex III, article 3.5.6. Thus, the report of any event that affected or could have affected the safety of others must be communicated to the D.S.A.C., and the maker or manufacturer must also be informed of any security flaw or malfunction. For example, the D.S.A.C. must be notified if the drone approaches a group of people in consequence of a distraction of the pilot; or if the drone approaches another flying aircraft in the absence of any specific technical failure; or if the drone causes damage at landing following unexpected behavior; or if the drone enters a forbidden area in the absence of any specific technical failure, etc. In other cases, the incident should be reported both to the D.S.A.C. and the manufacturer, for example if the drone goes past the 150 meters altitude limit following a loss of control, if the drone flies over a territory which hadn’t been planned in the security study, or in the event of a crash. An analysis of the event must be attached to the experience feedback report, as it contributes to the prevention of future risks (D.G.A.C., 2016: 10).This preventive aspect is furthermore strengthened by a notion which is particularly popular within the field of aeronautics: the concept of “Just Culture” as defined by EU regulations 44. The notion of Just Culture promotes a preventive reinforcement of safety, and aims at generating favorable conditions for users to feel comfortable about reporting incidents. Essentially, Just Culture is based on the idea that building a climate of trust is more favorable to fostering security and safety. This notion of Just Culture, so dear to the aeronautical sector, specifically aims at encouraging users to systematically report any event that posed or could have posed security issues. It promotes experience feedback (REX) and the collective improvement of procedures. Thus the article L. 6223-2 of the Code of Transports provides that “no administrative, disciplinary or professional punishment may be inflicted upon a person who reported an event [..], regardless of whether the person was or wasn’t involved in the event, except if the person is guilty of deliberately or repeatedly failing to comply with the safety rules”.As we’ve seen, the preventive aspect of these new measures for public safety is praiseworthy. The repressive aspect, however, is a different matter. Indeed, the legal framework around drones could use some simplification and rationalization.The increasing repression and legal risk faced by remote pilots The repressive measures included in the new drone regulations are increasingly complex and, furthermore, too scattered, which hinders their legibility and accessibility (A). This complexity and the restrictions it entails contribute to burdening operators with a heavy legal risk, and fragilizing, in the name of public order, the legal certainty of drone practices (B).Complexification and inaccessibility of the legal frameworkA project to homogenize drone regulations across the European Union will come into effect in July 2020, and is scheduled to progressively replace local legislations over a three-year period 45. However, the current regulations are still far from unified: they are scattered both in terms of hierarchy of norms (law of October 24th, 2016 regarding the reinforcement of safety in civil drone use; rulings of December 17th, 2015 regarding the conception of drones and their use of airspace 46; ruling of October 12th, 2018 establishing the list of zones where aerial photography and video recording are forbidden 47; decrees of May 18th, 2018 regarding compulsory training for remote pilots of (42) https://www.ecologique-solidaire.gouv.fr/drones-usages-professionnels(43) Ibid.(44) Regulation (EU) number 376/2014 of April 3rd, 2014 on the reporting, analisis and follow-up of occurrences in civil aviation, amending the regulation (EU) number 996/2010, art. 2 (12).(45) Commission delegated regulation (EU) 2019/945 of March 12th, 2019 on unmanned aircraft systems and on third-country operators of unmanned aircraft systems and Commission implementing regulation (EU) 2019/947 of May 24th, 2019 on the rules and procedures for the operation of unmanned aircrafts; OJEU, June 11th, 2019, number L 152.(46) Ibid.(47) Ruling of January 27th, 2017 establishing the list of zones where aerial image recording is forbidden, whether with a photographic or cinematographic device or any other recording device, JORF, number 25, January 29th, text number 1.© Cahiers de la sécurité et de la justice - IHEMI72 I DOSSIER Prevention and repression of civil drone use: finding a balance – Johanne Gojkovic-Lette, Grégory Houillonrecreational drones48 and regarding the weight criterion49; decree of April 19th, 2019 regarding the information leaflet on the use of aircrafts circulating without anybody on board50) and plurality of sources (Code of Transports, Civil Code – privacy rights, image reproduction rights and civil liability – and especially Penal Code). Drone use is thus materially situated at the junction between several laws and rights, which hinders the readability and consistency of regulations.Furthermore, regulations are bound to evolve for drone users, both professional and recreational. Penal laws on drone use are constantly changing, which makes them even more inaccessible to non-jurists. For example, the recent decree of April 19th, 2019, that just came into effect in July 2019, introduces a new third class penal violation: the sale of spare parts of drones unaccompanied by the information leaflet made compulsory by article 425-1 of the Code of Consumption. This preventive leaflet contains information on the conditions of use of aircrafts circulating without anybody on board, as well as the rules regarding the use of airspace, safety, and the respect of privacy. In the near future, depending on the development of the upcoming EU regulations currently being adopted, the registration criteria is expected to change again, shifting from a weight criterion to a new criterion based on kinetic energy. Thus, if in the current state of the law, registration and training are compulsory for drones over 800g, they will probably become compulsory for devices “which, in case of an impact or collision with a person, can transfer energy beyond 80 joules”. Not only is this criterion of a different nature, as it focuses on the level of damage done in the event of an impact at full speed, it also drastically lowers the weight threshold for registration. Indeed, based on those 80 joules of kinetic energy, the obligation to register would apply to any drone slightly over 250g. Thus, drone users should expect an extension of their obligations. Generally speaking, such changes, complexifications and increases in the obligations borne by remote pilots, in conjunction with the numerous scattered sources of these regulations, tend to further weaken the legal certainty of the field (Balat et Mazouz, 2019 : 411). Being that public order and material security have priority over legal certainty, the D.G.A.C. will probably have to provide operators with a homogenous guide compiling all the scattered applicable rules, which will serve as a “traffic code” for drones. The need for such a guide has often been suggested 51. Such initiatives already exist, but they cover the whole field of civil aviation 52. And the D.S.A.C. frequently publishes a complete guide of drone use, which includes the topic of regulations (D.S.A.C., 2017). However, a specific compilation exclusively dedicated to legal aspects is still lacking to simplify such a complex activity. It would be all the more necessary that these regulations are quite unstable and bound to evolve over time. In the perspective of the progressive professionalization of remote pilots of drones above a certain weight (which, besides, is expected to be lowered), such a guide would allow them to maintain the mandatory knowledge acquired during their initial theoretical training. It would also help to mitigate the risk of legal punishment which has been increasingly heavy on remote prilots, in order to conciliate the imperative necessities of aeronautical public order with the exercise of individual liberty, within the boundaries of “doing anything that does not harm others”, as defined by the Declaration of the Rights of Man and of the Citizen of 1789 53.Penal risk and legal uncertaintyIndeed, besides the violations pertaining specifically to remote pilots defined by articles L. 6232-12 and 13 of the Code of Transports, it must be remembered that the other penal offenses established by this same Code for all civil aircrafts also apply to drones: violation of safety rules 54; obstruction of air traffic and navigation 55. And in addition to these offenses, are also applicable those established by the Penal Code itself, as well as other sources: physical harm caused by imprudence, negligence, clumsiness or inattention 56; endangering the life of others (48) Ibid.(49) Decree number 2018-374 of May 18th, 2018 regarding the weight thresholds provided by law number 2016-1428 of October 24th, 2016 regarding the reinforcement of safety in civil drone use, JORF, number 115, May 20th, text number 32.(50) Decree number 2019-348 of April 19th, 2019 regarding the information leaflet on the use of aircrafts circulating without anybody on board, which came into effect on July 1st, 2019, JORF number 95, April 21st, 2019, text number 34.(51) EASA, Proposal to create common rules for operating drones in Europe, number A-NPA 2015-10, September 2015 (33 proposals).(52) A striking example is the RADIC, published by the D.G.A.C. (2017), Recueil des arrêtés, décisions, instructions et circulaires relatifs à l’aviation civile (RADIC 2018), Volume I, Ministry for the Ecological and Inclusive Transition.(53) Art. 4 of the Declaration of the Rights of Man and of the Citizen of 1789.(54) Art. L. 6232-2 Code of Transports.(55) Art. L. 6372-4 Code of Transports.(56) Art. 222-19 Penal Code.© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 73 Cahiers de la sécurité et de la justice – Hors-série exposing them to an immediate risk of death or injury likely to result in mutilation or permanent disability 57; data theft, hacking and piracy 58; trespassing and recording in a sensitive area 59, recording of confidential information protected under “national security” secrecy 60, violation of privacy 61, or even use of unauthorized frequency 62. It is fair to say that this great number of offenses heavily burden remote pilots, but also operators (when they are not the same person), and also potentially manufacturers and distributors, not to mention hypothetical accomplices. Such a context cannot fail to generate an increasing legal risk, as the law of October 24th, 2016 simply added a specific offense via two new provisions, to the already existing general corpus of repressive laws 63. Furthermore, the presence of penal offenses (one of which is specific to drones 64) outside of the Penal Code and within the Code of Transports makes them less accessible and visible to drone operators. These complex and technical regulations challenge the intelligibility of the law, a constitutional principle which, then again, is supposed to constantly guide legislators 65.As we can see, civil drone activities have been gaining material safety, as they are practiced according to a strict framework in compliance with public order, but losing legal safety, as the slightest use of a drone can result, for the operators, in a state of constant doubt and uncertainty regarding the penal risk they might be exposed to, made even more unclear by the complexity of applicable regulations. The certified training operators go through, although it includes information of penal regulations, isn’t enough to compensate the level of uncertainty entailed by their inherent complexity.There is a fine line between enforcing aeronautical public order and preserving the freedom of users within reasonable practices, and the balance remains to be found, in such an innovative and evolutive field. But finding that balance is an unavoidable necessity for a technical sector facing a progressive increase in penalization, paired with such complex legislation n(57) Art. 223-1 Penal Code.(58) Art. 323-1 Penal Code.(59) Art. D. 133-10 Code of Civil Aviation.(60) Art. 413-9 and following articles, Penal Code.(61) Art. 226-1 and 2 Penal Code.(62) Art. L. 39-1 Code of Postal services and Telecommunications.(63) Law number 2016-1428 of October 24th, 2016, ibid., art. 5, which inserts two articles within the section 6 entitled “aircrafts circulating without anybody on board” in Chapter 2 of Title III, Book II of the 6th part of the Code of Transports.(64) Art. L. 6232-12 and 13, ibid.(65) Decision number 99-421 DC of December 16th, 1999, Constitutional Council, Records, p. 136.ReferencesBooks and articlesArchambault, Laurent and Alicia Mazouz, 2015, « L’envol des drones civils : Appréhension par le droit français d’une pratique émergente », Annals of Air and space Law, vol. XL.Archambault, Laurent and Alicia Mazouz, 2016, « Quel horizon juridique pour les drones civils ? », Gaz. Pal., 21 juin 2016, p.19.Balat, Nicolas and Alicia Mazouz, 2019, « Les drones civils et la sécurité juridique », in Siguoirt, Laurent (ed.), Transports et sécurité, Lexis Nexis, p. 411.Cormier, Henri, 2018, « Drones. Vers un renforcement législatif », Aviation civile, Nr 383.Geffray, Edouard, 2015, « Quel cadre juridique pour les drones aériens civils ? », in CESE, 2015, « Les drones aériens civils : opportunités et risques », colloque du 28 mai 2015.Gojkovic, Johanne and Grégory Houillon, 2019, « La pratique du drone, entre sécurisation et répression », AJ-Pénal, Nr 3, p.135.Hanicotte, Robert, 2014, « Une nouvelle catégorie d’OVNI juridique : les drones », Gazette du Palais, Nr 317, p.6.Lepage, Agathe, 2014, « Première rencontre du droit pénal avec un drone », Lexis Nexis, Comm. Com. électr., Nr 7/8, comm. Nr 65.Pupin, Gaëlle, 2018, « Drones : les gendarmes investissent la troisième dimension », GendInfo, (19 mars).© Cahiers de la sécurité et de la justice - IHEMI74 I DOSSIER Prevention and repression of civil drone use: finding a balance – Johanne Gojkovic-Lette, Grégory HouillonSandvik, Kristin Bergtora and Jumbert Maria Gabrielsen, 2015, « Les drones humanitaires », RIS, Nr 98, p. 139 ; Zema, Alexis, 2019, « En France, des drones effectuent désormais des contrôles routiers », Le Figaro, (May 17th).Official reportsEuropean Commission, Communication of April 8th, 2014, A new era for aviation. Opening the aviation market to the civil use of remotely piloted aircraft systems in a safe and sustainable manner, COM (2014) 207.European Parliament, 2015, Report on safe use of remotely piloted aircraft systems (RPAS), commonly known as unmanned aerial vehicles (UAVs), in the field of civil aviation, number 2014/2243/INI.SGDSN, 2015, l’Essor des drones aériens civils en France : enjeux et réponses possible de l’État, report from the Government to the Parliament, SGDSN, October 20th, 2015.EASA, 2015, Proposal to create common rules for operating drones in Europe, number A-NPA 2015-10.EASA, 2016a, “Drone collision” Task Force, Final Report, October 4th, 2016.EASA, 2016b, High Level Conference on “Drones as a Leverage for Jobs and New Business Opportunities”, Warsaw, November 24th, 2016. D.G.A.C., 2016, Incidents de drones – notification et suivi, May 2016.D.S.A.C., 2017, Aéromodélisme : modèles réduits et drones de loisir, 1st edition, version 2 (January 10th).D.S.A.C., 2017, Aéronefs circulant sans personne à bord : activités particulières, 1st edition, version 3 (August 28th).Laws, decrees and other regulationsDecember 7th, 1944 Chicago Convention on International Civil Aviation.Regulation (EU) number 376/2014 of April 3rd, 2014 on the reporting, analisis and follow-up of occurrences in civil aviation.“Aircraft” ruling of December 17th, 2015 regarding the conception of unmanned civil aircrafts circulating without anybody on board, the conditions of their utilization, and the abilities required from the persons who use them, JORF (Official Journal of the French Republic), number 129.“Spaces” ruling of December 17th, 2015 regarding the use of airspace by aircrafts circulating without anybody on board, JORF, number 298, p. 23890Law number 2016-1428 of October 24th, 2016 regarding the reinforcement of safety in civil drone use, JORF, number 249.Ruling of January 27th, 2017 establishing the list of zones where aerial image recording is forbidden, whether with a photographic or cinematographic device or any other recording device, JORF, number 25.Decree number 2018-374 of May 18th, 2018 regarding the weight thresholds provided by law number 2016-1428 of October 24th, 2016, JORF, number 115.Decree number 2018-375 of May 18th, 2018 regarding compulsory training for remote pilots of aircrafts circulating without anybody on board for recreational purposes, JORF, May 20th, number 33. Decree number 2018-882 of October 11th, 2018 regarding the registration of civil aircrafts circulating without anybody on board, JORF, number 237.Ruling of October 12th, 2018 regarding the training required for remote pilots who use civil aircrafts circulating without anybody on board for recreational purposes, JORF, number 248 of October 26th, 2018, number 45.Ruling of October 19th, 2018 regarding the registration of civil aircrafts circulating without anybody on board, JORF number 248, October 26th, 2018, number 46.Décret Nr 2019-348 du 19 avril 2019 regarding la notice d’information relative à l’usage des aéronefs circulant sans personne à bord, JORF number 95.Decree number 2019-348 of April 19th, 2019 regarding the information leaflet on the use of aircrafts circulating without anybody on board, JORF number 95.© Cahiers de la sécurité et de la justice - IHEMICahiers de la sécurité et de la justice – Hors-série DOSSIER I 75Stop-djihadisme ou comment déjouer la radicalisation : examen critique d’un plan de prévention par l’informationLaurène RENAUT, Laura ASCONEuite aux attentats de Charlie Hebdo en janvier 2015 et face au nombre grandissant de jeunes adhérant à l’idéologie promue par l’Organisation de l’État islamique (O.É.I.), le gouvernement français a mis en place des plans successifs de prévention, adoptés sous l’égide du Comité interministériel de prévention de la délinquance et de la radicalisation (C.I.P.D.R.) 1, visant à contrer le phénomène de radicalisation. Un premier pas a d’ailleurs été franchi avec la réalisation du site Stop-djihadisme, qui se donne trois objectifs principaux :– comprendre la menace terroriste et décrypter la propagande djihadiste– sensibiliser et mobiliser la communauté nationale pour lutter ensemble contre ce phénomène– agir pour protéger la population et le territoire français.Dans cette perspective, trois supports différents ont été mis au service de ces grands axes de communication : des articles, des clips vidéo et un numéro Cette contribution, qui s’inscrit dans le champ de l’analyse du discours appliquée à des questions de sécurité ou de justice (et en particulier à leur dimension préventive), propose une analyse critique du dispositif Stop-djihadisme. Lancé en janvier 2015 par le gouvernement, ce plan de prévention spécifique à la radicalisation vise à prévenir ce phénomène par l’information auprès de divers publics (jeunes en voie de radicalisation et leur entourage, éducateurs spécialisés, acteurs associatifs et ensemble de la population française). Dans cette perspective, nous examinerons d’abord les stratégies argumentatives et rhétoriques des clips Ils te disent et Toujours le choix puis nous présenterons les points forts et les limites, d’un point de vue discursif, de ces campagnes. Enfin, nous élargirons la réflexion sur la nécessité de produire des contre-discours au discours djihadiste en proposant une ouverture vers une forme de récit susceptible de constituer un discours alternatif pertinent : le témoignage, aussi bien d’ex-djihadistes « repentis » que des victimes directes ou indirectes du terrorisme.Mots-clés : contre-discours ; témoignage ; analyse du discours ; justice restaurative.SLarène RENAUTDiplômée du CELSA (sciences de l’information et de la communication) et doctorante (sciences du langage) à l’Université Paris-Seine. Sa thèse porte sur la construction de la radicalité djihadiste en ligne à partir d’un corpus de profils radicalisés, d’échanges en ligne, de témoignages et de jugements.Laura ASCONETitulaire d’un master (linguistique anglaise) à l’Université Paris−Diderot et d’un doctorat de l’Université Paris Seine avec une thèse sur l’expression des émotions sur les réseaux sociaux, sur la propagande djihadiste et sur ses contre-discours. Elle effectue un post-doctorat à l’Université de Lorraine dans le cadre d’un projet A.N.R. sur les discours de haine contre les migrants.(1) Le C.I.P.D.R. a pour missions l’animation, la coordination et le soutien financier des politiques de prévention de la délinquance et de la radicalisation.© Cahiers de la sécurité et de la justice - IHEMI76 I DOSSIER Stop-djihadisme ou comment déjouer la radicalisation : examen critique d’un plan de prévention par l’information – Laurène Renaut, Laura Asconevert. Les articles, de nature informative et explicative, considérés par le gouvernement comme des outils de prévention, s’articulent autour de cinq rubriques : « la radicalisation », « le terrorisme djihadiste », « la lutte contre le terrorisme et la radicalisation », « que faire ? » et « en cas d’attaque ». En d’autres termes, si l’objectif du discours djihadiste est d’inciter le destinataire à agir, les articles pédagogiques (composant une grande partie du contre-discours institutionnel français), visent plutôt à l’informer (Ascone, 2018). Toutefois, la rubrique « que faire » et, plus particulièrement, l’article « Devenir acteur de la lutte contre la radicalisation et le terrorisme » se présentent comme une tentative de mettre le citoyen au centre de l’action et de la lutte contre la radicalisation violente. Cette volonté se traduit également par le numéro vert, mis en place pour le signalement de tout individu radicalisé ainsi que pour « soutenir les familles des personnes radicalisées ». Enfin, se distinguent sur le site gouvernemental, deux clips : Ils te disent et Toujours le choix, ce dernier étant articulé autour de deux parcours (celui de Mehdi ou d’Emma) et présentant aussi une section dédiée aux témoignages, « Ils l’ont vécu ». Cette rubrique fait principalement référence aux proches de jeunes radicalisés même si une vidéo met aussi en scène un repenti dont l’objectif est de dissuader d’autres jeunes à rejoindre les rangs de l’O.É.I. Si les articles et le numéro vert s’adressent au grand public ainsi qu’aux proches d’individus radicalisés, les vidéos produites par le Service d’Information du Gouvernement (S.I.G.) visent, quant à elles, un public radicalisé.Ainsi, l’ensemble de ces éléments montre bien que les instances gouvernementales ont investi internet comme un nouveau lieu d’affrontement symbolique contre l’O.É.I. L’objectif : monter une parade aux messages de haine. Le combat contre le terrorisme a donc lieu non seulement en Syrie et Irak mais aussi sur le terrain disputé d’internet dans une optique de prévention qui passe avant tout par la sensibilisation et la détection.Or qui dit contre-discours dit production d’un discours qui s’établit par définition « dans un antagonisme explicite » (Angenot, 1989 : 1) avec un autre discours dont il convient d’ailleurs de comprendre les ressorts pour mieux saisir les spécificités du discours produit en réaction (Auboussier, 2015). Aussi, précisons que nous avons décidé ici d’étudier le contre-discours institutionnel français en relation avec le discours djihadiste de Dar al-Islam, la revue francophone diffusée par l’O.É.I. et, plus particulièrement, par son département communication et médias le Al Hayat Media Center. Depuis 2014, ce centre médiatique diffuse principalement des revues : Dabiq qui est publiée en anglais et qui compte quinze numéros, Dar al-Islam dont les dix numéros sont rédigés en français, et Rumiyah 2, déclinée en huit langues (anglais, turc, ouïghour, pachtoune, russe, allemand, indonésien et français) et qui compte treize numéros. Al Hayat Media Center publie également des vidéos et enregistrements de chants (nachids) visant à promouvoir l’idéologie djihadiste. Cependant, notre étude pour ce présent article a exclusivement porté sur les revues et, plus spécifiquement, Dar al-Islam, afin d’y étudier un discours s’adressant spécifiquement à un public radicalisé francophone, d’une part, et accessible de manière anonyme, une mesure de précaution non négligeable pour un chercheur non rattaché à des structures institutionnelles, d’autre part 3. Bien que cette revue de propagande soit diffusée sur internet, précisons qu’elle présente toutes les caractéristiques d’une revue imprimée (on y retrouve un sommaire, des entretiens, des reportages photo ou des articles sur les opérations menées par l’O.É.I.) et que ses dix numéros (229 762 mots) constituent donc un réservoir important de données. Ce choix, tout en présentant des limites bien sûr, nous a aussi permis d’analyser deux discours officiels, le site Stop-djihadisme d’un côté, à travers l’étude de deux clips 4 sur lesquels nous focaliserons notre attention, et Dar al-Islam de l’autre.C’est dans ce contexte que notre contribution, laquelle s’inscrit dans le champ de l’analyse du discours appliquée à des questions de sécurité et de justice, propose une analyse critique du dispositif Stop-djihadisme, présenté par le gouvernement comme un plan de prévention visant à prévenir la radicalisation par l’information. Plus particulièrement, cette étude présente une analyse des stratégies argumentatives et rhétoriques déployées dans les vidéos examinées. À cet égard, nous entendons rhétorique comme « l’art de persuader par le discours » (Reboul, 1991 : 4) afin de déclencher une action (Plantin, 2011 : 17-18), l’objectif étant « de provoquer ou d’accroître (2) Précisons que Rumiyah remplace les revues Dabiq (en anglais), Dar al-Islam (en français), Istok (en russe) et Konstantiniyye (en turc), dont la publication a été assurée jusqu’à fin 2016.(3) Précisons que le corpus de revues étudiées a été reccueilli en 2016, date à laquelle les supports de propagande djihadiste étaient encore accessibles (notamment en PDF pour les revues sur le site jihadology.net) sans recourir à un enregistrement avec des codes d’accès spécifiques comme c’est le cas aujourd’hui. (4) Ils te disent (http://www.Stop-djihadisme.gouv.fr/stopdjihadisme-retour-2-ans-lutte-contre-propagande-djihadiste) et Toujours le choix (http://www.toujourslechoix.fr/). Précisons que le clip Toujours le choix se décline sur deux parcours, celui d’Emma et celui de Mehdi qui font donc l’objet de deux vidéos.© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 77 Cahiers de la sécurité et de la justice – Hors-série l’adhésion des esprits aux thèses qu’on présente à leur assentiment » (Perelman et Olbrechts-Tyteca, 1988 : 59 ; voir aussi Amossy, 2008).Dans cette optique, nous examinerons d’abord les stratégies argumentatives des clips diffusés sur cette plateforme. S’agissant de contenus plurisémiotiques, nous mènerons donc des analyses sémio-discursives (alliant texte et image) et adopterons une approche qualitative pour examiner ces vidéos. Nous présenterons ensuite les points forts et les limites de ces campagnes d’un point de vue discursif. Enfin, à partir de l’analyse de la rubrique « Ils l’ont vécu » (sur le site Stop-djihadisme), nous élargirons la réflexion sur la nécessité de produire une forme de récit susceptible de constituer un discours alternatif pertinent : nous nous pencherons donc sur les caractéristiques et le rôle du témoignage, aussi bien d’ex-djihadistes repentis que des victimes directes ou indirectes du terrorisme.Stop-djihadisme : les objectifs d’une plateforme numériqueDiscréditer l’autreLe premier clip réalisé par le S.I.G., Ils te disent, dénonce la nature mensongère du discours djihadiste en confrontant les promesses des recruteurs de l’O.É.I. aux crimes commis par ses membres (Figure 1).Cette campagne gouvernementale repose sur des armes non militaires (mots, images et informations) qui s’inscrivent dans une guerre de représentations, centrée sur l’autre. Cette dernière s’appuie en effet sur le triptyque discréditer-culpabiliser-renverser favorisant une forme de « guerre symbolique, psychologique et subversive » (Del Valle, 2014 : 3) en parallèle des opérations militaires. Dans ce contexte, discréditer l’autre aux yeux des jeunes sensibles à l’idéologie djihadiste passe ici par une entreprise de diabolisation dont l’objectif est de déshumaniser l’ennemi. La prédiction « tu découvriras l’enfer sur terre » fait ainsi référence à un lieu mortifère tout en renvoyant au mal absolu, c’est-à-dire l’O.É.I.Or, si les dénominations pour désigner les Occidentaux foisonnent dans les supports de propagande djihadiste, avec le terme « mécréants » (449 occurrences dans les dix numéros de Dar al-Islam) et « apostats » (194 occurrences), le mal indicible dénoncé par cette campagne gouvernementale est pronominalisé : les hommes de l’O.É.I. sont en effet désignés par « ils » à six reprises, une façon de ne jamais nommer l’opposant directement. La répétition de ce pronom construit ainsi une opposition binaire (« ils » versus « nous ») qui vise à rejeter ce « ils » et s’adresser directement aux jeunes d’où le recours au tutoiement. Néanmoins, si l’innommable n’est pas nommé, il est constamment imagé dans le clip à travers ses actes de barbarie. En effet, cet autre qui allégorise la mort et la tromperie est montré comme l’auteur de massacres via des images de guerre, de corps crucifiés et de décapitation. Toutefois précisons bien que si le S.I.G. active des codes similaires aux films de propagande djihadiste en utilisant des images officielles de l’O.É.I., la vidéo Ils te disent distingue bien deux camps en présence : celui des ennemis qui choisissent la barbarie et ses détracteurs qui la rejettent, en ne la montrant que partiellement. Cette mise en scène paradoxale de la violence, présente mais euphémisée, permet ainsi un effet de dévoilement : le clip affirme que ce ne sont pas seulement les cibles revendiquées par l’O.É.I. (l’armée de Bachar El-Assad) qui sont victimes des massacres mais aussi les civils.En plus d’être diabolisé, l’autre s’avère aussi tourné en dérision par des stéréotypes 5 qui construit une vision caricaturale du recruteur djiahdiste. À cet égard, l’élément principalement visé par cette forme de discrédit n’est autre Figure 1 : Exemple d’une séquence de la vidéo Ils te disent(5) Le terme « stéréotype » réfère à des « croyances partagées concernant les attributs personnels d’un groupe humain » (Leyens et al., 1996 : 11).© Cahiers de la sécurité et de la justice - IHEMI78 I DOSSIER Stop-djihadisme ou comment déjouer la radicalisation : examen critique d’un plan de prévention par l’information – Laurène Renaut, Laura Asconeque le langage : d’un point de vue orthographique, les représentations des échanges écrits sur Facebook entre un jeune radicalisé et son recruteur véhiculent en effet l’idée d’un abaissement du niveau de langue, en particulier dans la vidéo Ils te disent. Dans ce clip, après avoir consulté des contenus djihadistes sur Facebook, le protagoniste reçoit un message de la part d’un recruteur (Figure 2).Même si les choix graphiques miment ici une orthographe basée sur l’oralité, visant à produire un effet de réel et donc à mettre en scène la langue « des jeunes », les effets de ces choix pourraient amener certains récepteurs à catégoriser socialement les personnages représentés, voire à assimiler ce parler au « langage des cités ». Cette possible interprétation semble d’ailleurs renforcée d’un point de vue phonétique par le parler des enrôleurs qui présente tous les indices prosodiques de l’accent dit « de banlieue », audible dans le clip Toujours le choix.En outre, le caractère récitatif du discours d’embrigadement reproduit dans ce dernier clip (« Il faut aider les musulmans qui souffrent. […] On va t’aider, tu seras plus seule ») et les choix du montage qui ajoute cette séquence juste après la scène de rencontre avec la jeune Emma, pourraient contribuer à rendre simpliste les ficelles de ce système de recrutement. En effet, ce qui dans la vidéo est présenté comme un basculement rapide pour un jeune en voie de radicalisation, est en réalité un processus dont la durée et la trajectoire dépendent d’une pluralité de facteurs, psychologiques, culturels, sociaux, etc. (Moniquet, 2015). De même, on sait que « cette conversion ne survient pas comme un coup de tonnerre dans un ciel serein, mais prend la forme d’un lent processus » (Haddad, 2015 : 110).Enfin, se greffe à ces stratégies, le ressort du remord, autrement dit la tentative de susciter une forme de culpabilité, notamment visible à travers le slogan de la campagne Toujours le choix : « Se radicaliser, c’est détruire sa vie, sa famille et celle des autres. » Cet énoncé présente en effet une fonction pathémique en faisant référence à la souffrance des proches. Culpabiliser s’appuie donc ici sur une logique de victimisation où les figures de victimes sont plurielles : le jeune radicalisé, son entourage et plus largement les français comme victimes des attentats terroristes. Toutefois, dans le contre-discours institutionnel, la culpabilisation est causée par l’action violente au nom du djihad tandis que, dans la propagande de l’O.É.I., elle découle au contraire de l’inaction face aux bombardements et aux humiliations endurés par les musulmans : « Vas-tu laisser le mécréant dormir tranquille dans sa maison pendant que les femmes des musulmans et leurs enfants tremblent de peur, effrayés par le bruit des avions croisés au-dessus de leur tête nuit et jour. » (Dar al-Islam 2, 2015 : 6). L’objectif, dans les deux cas, est de mettre à jour les actes immoraux perpétrés par l’ennemi et déclencher des émotions (peur, honte, colère, etc.) visant à provoquer des comportements collectifs.Orchestrer la dissonanceLe dispositif du clip Ils te disent repose sur la mise en scène d’un récit polyphonique : deux voix discordantes, celle des djihadistes, mise à distance par un recours au discours direct (« Ils te disent : « viens fonder une famille avec un de nos héros ». ») et la voix institutionnelle introduite par « en réalité », se superposent pour révéler la nature mensongère du discours de l’O.É.I. De cette manière, la communication gouvernementale orchestre la dissonance afin de bouleverser l’univers de croyances des jeunes radicalisés et leur appréhension du réel.Pour ce faire, cette campagne recourt à des inversions sémantiques qui consistent à retourner contre l’ennemi ses références pour déconstruire les fantasmes autour de l’idéologie djihadiste. Ainsi, le glissement du sens philosophique du terme vérité, abondamment utilisé Figure 2 : Exemple de premier contact que le recruteur établit avec sa cible sur Facebook (capture d’écran du clip Ils te disent)© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 79 Cahiers de la sécurité et de la justice – Hors-série par l’O.É.I. 6 et associé à sa valeur universelle dans « la vérité est ici », à une forme plurielle de ce terme dans « comme seules vérités tu découvriras l’horreur et la tromperie » (Figure 3), participe au détournement des significations. La figure de l’humanitaire (« Viens aider les enfants syriens »), est remplacée par celle du criminel de guerre (« Tu seras complice de massacres »). De même, la logique sacrificielle conduisant au paradis pour le martyr, « Sacrifie-toi à nos côtés, tu défendras une juste cause », se transforme en un avertissement : « Tu découvriras l’enfer sur terre ». Le plan gouvernemental de prévention de la radicalisation par l’information repose donc sur une dénonciation des éléments de langage djihadiste à travers un processus de renversement du sens. En outre, on assiste à une inversion des représentations du jeune radicalisé lui-même, le mythe du héros laissant place à une figure de victime manipulée par le groupe djihadiste.Du rêve à la réalité : tel est le glissement principal qui régit la campagne Ils te disent. D’un point de vue iconographique d’abord, le montage saccadé alterne images en couleurs, pour les promesses des djihadistes, et images en noir et blanc pour la réalité.Concernant la bande sonore, une musique « religieuse » d’Omar Omsen 7, traditionnellement utilisée dans la propagande musicale de l’O.É.I. pour exalter la foi de ses recrues dans le combat, est ici volontairement exploitée par le S.I.G. Or, ces incantations mélodiques sont entrecoupées par des grésillements qui donnent l’impression d’un réglage à la recherche d’une bonne fréquence, comme pour échapper à la parole officielle de l’O.É.I. Pleurs de femmes et cris d’enfants ponctuent également ce clip qui abonde d’images de civils maltraités en terre de califat. Cette valorisation du contraste entre fiction et réalité tend à confronter un public radicalisé à des informations incohérentes avec ses croyances et à réinstaurer chez lui le doute.Dissuader d’agirSi le contre-discours porté par le gouvernement est présenté par Manuel Valls comme une tentative bienveillante « de ramener ces jeunes […] vers la France qui les aime » 8, il n’est cependant pas exempt d’un rapport de force. Le clip Toujours le choix ne présente d’ailleurs rien d’autre qu’un véritable bras de fer entre les autorités et les individus suceptibles de se retourner contre leur pays : ce n’est pas un hasard si l’on y voit un jeune Mehdi arrêté et condamné à la prison après avoir fomenté un attentat en France. La menace est ici explicite (provenant de l’État lui-même), et coercitive, visant à contraindre les sympathisants de l’O.É.I. au respect de la loi via le spectre d’une privation de liberté.Pour autant, le gouvernement se contente surtout d’avertir ou prévenir le jeune radicalisé des risques qu’il encourt (mort violente dans le clip Ils te disent ou répudiation pour Emma dans Toujours le choix) et s’attache ainsi à mettre en oeuvre un ethos 9 d’autorité qui tend à faire du conseil une consigne par la convocation d’un univers punitif. À la loi d’Allah dans le discours djihadiste s’oppose donc la loi de la République mise au service d’une stratégie de persuasion par la peur qui permet d’agir sur le comportement (Witte et Allen, 2000). Précisons que si le contre-discours institutionnel fait de cette peur l’épicentre de la rhétorique des émotions qu’il développe, c’est pour mieux dissuader le jeune d’agir, à l’inverse du discours djihadiste qui s’en sert comme tremplin à l’action.(6) On compte 111 occurrences du terme « vérité », uniquement au singulier, dans la revue Dar al-Islam.(7) L’auteur des premières vidéos djihadistes en français.(8) Discours de Manuel Valls du 18 novembre 2016.(9) La notion d’ethos désigne l’image discursive de soi, c’est-à-dire la présentation de soi dans le discours (Amossy, 2010).Figure 3 : Exemple d’une séquence de la vidéo Ils te disent qui montre la stratégie iconographique adoptée par le S.I.G.© Cahiers de la sécurité et de la justice - IHEMI80 I DOSSIER Stop-djihadisme ou comment déjouer la radicalisation : examen critique d’un plan de prévention par l’information – Laurène Renaut, Laura AsconePoints forts et limitesS’adressant à divers publics à travers des supports variés (articles ou vidéos), le dispositif Stop-djihadisme s’inscrit dans un large plan de prévention contre la radicalisation en se plaçant sur le terrain de l’information et de la communication. Ayant suscité, depuis sa mise en place, des réactions controversées au sein du monde académique et professionnel, nous proposons ici une réflexion critique, résultant des analyses sémio-discursives menées précédemment.Le choix de représenter la violenceLe choix d’utiliser un site web, comme outil de diffusion d’un contre-discours institutionnel, a permis au gouvernement d’articuler son plan de prévention autour de supports variés : articles, vidéos, et numéro vert censé favoriser la détection d’individus radicalisés. Depuis son ouverture le 29 avril 2014 jusquau 30 juin 2017 (Benbassa et Troendlé, 2017 : 15), le Centre national d’assistance et de prévention de la radicalisation (CNAPR) a d’ailleurs reçu 51 429 appels pour un total de 5 723 signalements 10 ; ces derniers étant transmis à des services spécialisés chargés d’évaluer et de confirmer ou non le degré de dangerosité d’une personne.Cette hétérogénéité des contenus constitue, selon nous, l’un des points forts de la plateforme Stop-djihadisme. Rendant possible la vulgarisation et le partage de connaissances accessibles sur le phénomène complexe de la radicalisation, ces campagnes parviennent à atteindre divers publics et susciter des réactions : à cet égard, la vidéo Ils te disent a enregistré près de 2 millions de vues en seulement deux jours. Et si ce clip a suscité des débats, c’est notamment en raison de la présence d’images violentes, lesquelles ont provoqué des avis contrastés.En effet, nombreuses sont les recrues de l’O.É.I., ayant rejoint la Syrie ou l’Irak, qui reconnaissent à posteriori avoir été fascinées par les vidéos de décapitation diffusées par Al Hayat Media Center et plus largement par la violence. C’est le cas de ce jeune réfugié afghan de 25 ans, interpellé en 2017 par la police antiterroriste, et qui, accroc à des images violentes, avait téléchargé 2000 vidéos produites par l’État Islamique 11. Un constat que confirme la sociologue des médias Hasna Hussein, lorsqu’elle affirme que ces vidéos, en alternant des images de combats et de paysages exotiques grandioses déclenchent « une forme d’excitation sensorielle chez les internautes » (Hussein, 2019). N’oublions pas que le courant salafiste djihadiste légitime d’ailleurs le recours à la lutte armée en convoquant la théorie du « choc des civilisations » et en utilisant l’actualité proche-orientale, afin de justifier le combat contre l’Occident qui aurait déclaré la guerre à l’islam. Par conséquent, une vidéo qui se focalise sur la violence, laquelle représente une fin en soi dans la communauté djihadiste (Crettiez et Ainine, 2017) prend le risque d’accroître cette fascination chez un public radicalisé.Par ailleurs, si la stratégie de déshumanisation de l’ennemi mise en œuvre dans ces clips peut se révéler pertinente chez certains individus en voie d’initiation à l’idéologie djihadiste, et dont le raisonnement cognitif est encore ouvert à d’autres discours, elle pourrait s’avérer improductive auprès de sympathisants déjà bien engagés dans ce processus et favorables à l’action menée au nom du djihad armé. En effet, ainsi que le souligne Bouzar et Martin (2016), les retours d’expérience montrent que le temps consacré au changement comportemental d’un jeune radicalisé est lié à son niveau de radicalité : plus le diagnostic est précoce, plus le suivi sera court. À l’inverse, plus le jeune est avancé dans le processus, plus le phénomène de déshumanisation de l’autre sera abouti : d’ailleurs, dans la propagande djihadiste, la figure de l’occidental est le terreau d’un déferlement d’images d’une violence extrême visant à justifier et encourager la guerre à son encontre 12. Cette forme de déshumanisation facilite ainsi l’établissement d’une distance psychologique nécessaire pour déclencher la violence contre un groupe (Semelin, 2005), mais aussi rendre inaudible la parole de l’ennemi, ici le gouvernement français.Enfin, relevons un paradoxe qu’il nous semble important de prendre en compte dans la production de contre-discours : « en effet si la victimisation n’est pas loin de concerner chaque prise de parole dans la djihadosphère, notons pourtant un mépris affiché de ses membres pour la figure de victime ; « il fait sa victime » ou « c’est une grosse victime » étant des insultes courantes à l’encontre des ennemis de la communauté » (Renaut, 2019 : (10) Chiffres arrêtés au 30 juin 2017 (Benbassa et Troendlé, 2017 : 15).(11) Site de Lyonmag (21/09/2019) : https://www.lyonmag.com/article/103293/villeurbanne-le-refugie-afghan-etait-devenu-accro-aux-videos-de-decapitations-de-daesh.(12) Si l’on prend l’exemple du nachid « Avance, avance », l’ennemi y est diabolisé. Tour à tour « soldats du diables » ou « criminel », cet autre occidental est à anéantir car source de destruction selon les lois d’une vision victimaire et d’une logique duelle (c’est l’autre ou c’est moi) : « Sois tu les tues, soit ils te tuent ». Le registre est guerrier et l’autre se décline sous la forme d’une masse anonyme comme le souligne la formule anaphorique : « Tue-les ».© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 81 Cahiers de la sécurité et de la justice – Hors-série 28). Dès lors, il convient d’interroger les éléments de langage officiels du clip Ils te disent et les représentations ambivalentes qu’ils véhiculent du jeune djihadiste, entre acteur de la radicalisation et surtout victime d’un discours d’embrigadement. Auprès d’une cible rétive à la figure de victime, le clip Toujours le choix a quant à lui l’avantage de mettre en avant la notion de responsabilité personnelle.Un processus de labellisation institutionnelle à interrogerComme nous l’avons analysé précédemment, la mise en scène du décalage entre les espoirs suscités par la propagande de l’O.É.I. et la réalité de la vie à Raqqa est au cœur de la stratégie communicationnelle du S.I.G. (Figure 4).La mise en scène de ce décalage a pour objectif de bouleverser le système de croyances du destinataire. Toutefois, pour réintroduire le doute chez un public radicalisé, il est important d’anticiper la façon dont la vidéo sera perçue par cette cible en fonction de sa lecture spécifique du monde. Or, enfermés dans une vision complotiste de leur environnement, les jeunes susceptibles d’être touchés par cette idéologie risquent de percevoir ce clip comme une énième forme de manipulation émanant d’un gouvernement impie. À cet égard, le processus de labellisation institutionnelle de ces clips est un élément qu’il convient en lui-même d’interroger : en effet, le choix d’un site gouvernemental prend le risque d’être étiqueté comme un discours mensonger ou manipulateur, donc d’alimenter la vision conspirationniste de l’O.É.I. et renforcer ainsi l’adhésion à leur système de croyance initial. C’est d’ailleurs le danger que pointe Mourad Benchellali, ex-djihadiste sollicité par Christian Gravel (anciennement directeur du S.I.G.), pour collaborer à la création de vidéos, et qui aurait refusé car « le support en lui-même, [lui] posait problème » 13. De ce point problématique découlerait, selon lui, une absence d’intérêt pour cette campagne de la part du public cible, comme il le relate dans une interview pour Libération : « J’ai eu l’occasion de donner la parole à des jeunes de Vénissieux à ce sujet. Tout d’abord, ils ne connaissaient pas le site Internet, alors qu’ils sont bien, a priori, le public visé. Ensuite, je leur ai demandé s’ils avaient désormais envie de le visiter. Encore une fois, ils ont dit non. Ils étaient convaincus qu’ils y trouveraient davantage de la propagande que de l’information, selon leurs mots. » 14 A cet égard, le système discursif binaire, adopté ici pour opposer le mythe djihadiste à la réalité, paraît aisément renversable. On peut ainsi inverser les propos de la Figure 2 : « Ils te disent : « Comme seules vérités, tu découvriras l’horreur et la tromperie ». En réalité, tu vis dans un monde de mécréants impurs, la vérité est ici. »Précisons, toutefois, que si le gouvernement essaye de déconstruire les engagements que l’O.É.I. prend devant ses sympathisants, c’est justement parce que la promesse et la récompense sont au cœur du discours djihadiste. Une belle maison, un bon salaire, plusieurs femmes, le pardon d’Allah figurent ainsi parmi les promesses les plus attrayantes. Cet emploi stratégique de la récompense a pour objectif d’amener le jeune radicalisé à agir au nom de l’idéologie djihadiste. Le départ en Syrie lui permettra de mener une vie sous le sceau de l’aventure et de la gloire, tandis que la mort en martyr lui garantira l’accès au paradis. À l’inverse, le contre-discours institutionnel dissuade d’agir : il s’agit de ne pas rejoindre l’O.É.I. ou de ne pas commettre d’attentats sur le sol français. De même, les articles publiés sur le site Stop-djihadisme, s’adressant au grand public plus qu’aux jeunes radicalisés, montrent que l’action contre le terrorisme djihadiste doit être entreprise par l’État et non par le citoyen (Ascone, 2018).(13) Interview de Mourad Benchellali par Libération datant du 16 février 2016 : https://www.liberation.fr/france/2016/02/08/rien-ne-bouge-sur-les-contre-discours-a-la-radicalisation_1431915.(14) Ibid.Figure 4 : Exemple d’une séquence de la vidéo Ils te disent qui montre le décalage entre les promesses de l’O.É.I. et la réalité© Cahiers de la sécurité et de la justice - IHEMI82 I DOSSIER Stop-djihadisme ou comment déjouer la radicalisation : examen critique d’un plan de prévention par l’information – Laurène Renaut, Laura AsconeC’est pourquoi, si ces campagnes n’excluent pas la notion de liberté comme le révèle l’intitulé, Toujours le choix, le choix se réduit ici surtout à un renoncement (ne pas partir en Syrie) sans véritables alternatives pour combler un désir d’agir. C’est dans cette absence de propositions que se situe peut-être la principale limite de ce discours en plus de son support lui-même, un site gouvernemental, potentiellement suspect aux yeux du public cible.Ces différents points nous conduisent à questionner les cibles de ce contre-discours qui, s’il n’a peut-être pas vocation à concerner les personnes les plus embrigadées, s’adresse plus largement aux familles, aux éducateurs et aux acteurs associatifs à sensibiliser. Conçu, en outre, pour occuper l’espace face aux discours de haine sur internet, il est aussi une réplique communicationnelle adressée aux hommes de l’O.É.I. ainsi qu’à l’ensemble des Français dans un contexte de guerre.Bilan et perspectivesLa campagne gouvernementale présente donc des points forts tels que l’utilisation de différents supports, la vulgarisation de recherches académiques et la volonté d’atteindre des publics variés, mais elle présente aussi des limites. À cet égard, même si en janvier 2015, le clip anti-djihad Ils te disent bat des records d’audience avec près de deux millions de vues, l’impact de cette campagne reste toutefois difficile à évaluer. Le S.I.G. le reconnait d’ailleurs : il n’a pas les moyens scientifiques de savoir qui consulte le site Stop-djihadisme, donc s’il touche un public radicalisé et quelle en est sa perception. À ce titre, l’étude des profils radicalisés sur Facebook, tend à montrer que les retours sur la campagne Stop-djihadisme sont peu nombreux dans la djihadosphère et ne font pas l’objet de véritables débats (Renaut, 2019). Quand le clip Ils te disent est partagé, la tonalité associée est plutôt satirique : le contenu des vidéos et la figure du recruteur djihadiste est notamment tournée en dérision par certains jeunes qui revendiquent, sur les réseaux sociaux, leur attachement à l’idéologie promue par l’O.É.I.Nous avançons ainsi l’hypothèse qu’il existe une certaine défiance vis-à-vis de la source qui diffuse ce contre-discours. N’oublions pas que les individus adhérant à l’idéologie djihadiste rejettent la démocratie et les institutions françaises. C’est pour cette raison que nous ne formulerons pas des recommandations pour la production d’un contre-discours institutionnel à proprement parler. Nous proposerons plutôt, à partir des enjeux soulevés dans notre analyse, d’explorer d’autres formes de discours alternatifs qui nous paraissent pertinentes dans le cadre de la lutte contre la radicalisation.Du contre-discours au discours alternatif : le rôle du témoignageBien que ce type de discours apparaisse en retrait dans le site Stop-djihadisme, le témoignage constitue un contre-discours potentiel qui mériterait selon nous d’être développé et valorisé. À travers une analyse des spécificités de ce discours, nous verrons dans quelle mesure il pourrait fournir des bases pour une stratégie discursive alternative. À cet égard, nous entendons « discours alternatif » au sens d’un contre-discours qui ne se construit pas en opposition à un autre discours, c’est-à-dire un discours autre qui ne se dirait pas contre-discours. Ici, la forme spécifique du témoignage, où la conflictualité est intérieure et non dirigée contre, s’inscrit bien dans cette définition.L’argumentation par le récit de vie : une « arme de persuasion massive » 15Le témoignage renvoie au fait de raconter, de livrer un récit confirmant la véracité de ce que l’on a vu, entendu, perçu ou vécu. Ce récit a la spécificité de partager plus que d’asséner un point de vue dominant : en d’autres termes « il fait signe au lieu de mettre la main sur » (Combet-Galland, 1991) et déclenche un travail chez le récepteur ; élément qu’il nous semble pertinent de prendre en compte s’agissant du public qui nous intéresse dans le cadre de cet article. Par ailleurs, si l’argumentation ne s’impose pas dans le récit, elle se profile toutefois derrière la narration : le témoignage peut ainsi constituer un argument fort qui repose essentiellement sur la personne de l’orateur 16, la crédibilité que l’on accorde au témoin influençant toujours celle que l’on accorde au témoignage (Perelman et Olbrechts-Tyteca, 1988 : 420).(15) Expression de l’ex-djihadiste David Vallat, extraite du documentaire « Jihad : faut-il croire les repentis ? » dans l’émission Complément d’enquêtes sur France 2.(16) Nous avons recours à la terminologie de Perelman et Olbrechts-Tyteca (1988 : 9) : « Aussi, pour des raisons de commodité technique, et pour ne jamais perdre de vue ce rôle essentiel de l’auditoire, quand nous utiliserons les termes « discours », « orateur », et « auditoire », nous comprendrons par là l’argumentation, celui qui la présente et ceux auxquels elle s’adresse, sans nous arrêter au fait qu’il s’agit d’une présentation par la parole ou par l’écrit, sans distinguer discours en forme et expression fragmentaire de la pensée. »© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 83 Cahiers de la sécurité et de la justice – Hors-série Or, l’une des difficultés qui semble se poser en amont de la production des contre-discours, réside précisément dans la mise en place d’une relation de confiance avec leur émetteur, afin de créer les conditions favorables d’une remise en question chez un public radicalisé. Ce point interroge donc le choix des acteurs à intégrer au processus de désengagement au-delà même des stratégies rhétoriques à mettre en place pour instiller le doute dans l’esprit d’une personne radicalisée. C’est d’ailleurs l’enjeu majeur que soulève Mourad Benchallali lorsqu’il évoque son engagement dans la prévention de ce phénomène : « Il faut savoir de quoi on parle. Or, qui mieux qu’un repenti du jihad peut dire ce qu’est la réalité de vivre sur une terre de jihad, ce que ça implique ? ». Si un discours de vérité passe par un vécu, certains obstacles entravent l’intervention des repentis dans les écoles ou dans les prisons en raison de leur casier judiciaire.Néanmoins, le témoignage, en tant que discours argumentatif 17, semble se distinguer par sa puissance émotionnelle en convoquant trois procédés pathémiques (Lausberg, 1960 : 257) :– Premièrement, l’orateur se montre ému, se dit ému et communique par empathie son émotion à l’auditoire.– Deuxièmement, l’orateur inclut d’autres individus venant prouver la vérité référentielle du témoignage livré (parents, époux, compagne, amis susceptibles de démentir le récit en cas d’inexactitude).– Enfin, l’orateur évoque par définition des évènements émouvants.Dans ce contexte, la forme du témoignage, quelque soit le statut des « témoins » (repentis ou victimes du terrorisme), pourrait constituer un outil pertinent dans la prise en charge préventive de la radicalisation voire auprès d’individus incarcérés. Ce n’est d’ailleurs pas un hasard si le genre du témoignage a été exploité dans la propagande de l’O.É.I., notamment dans la revue Dar al-Islam sous la forme d’interviews rédigées par des femmes qui relatent leur bonheur d’avoir atteint la terre du califat : « Louange à Allah qui m’a facilité la route, je n’ai rencontré aucune difficulté. […] Je ressens un soulagement d’avoir accompli cette obligation al-hamdoulillah ». 18Le témoignage sous le prisme de Stop-djihadisme : « Ils l’ont vécu »Le gouvernement a saisi l’intérêt du témoignage en lui donnant une place sur le site Stop-djihadisme, et plus particulièrement dans la campagne Toujours le choix, en créant la section « Ils l’ont vécu », où la parole est donnée à des repentis ou aux proches d’individus partis combattre sur zones. Toutefois, il est évident que le S.I.G. privilégie ici les témoignages des familles aux témoignages des individus radicalisés eux-mêmes. Dans le cadre de cette campagne, seul David Vallat raconte son expérience tandis que Mourad Benchellali a décliné la proposition de figurer dans ces vidéos.Néanmoins, outre la problématique récurrente d’une source institutionnelle, ces témoignages ont le mérite de privilégier une accroche directe avec l’auditoire, tant par le tutoiement régulièrement employé par David Vallat, « on te ment », que le regard face caméra de ce dernier. Par ailleurs, les éléments non verbaux (comme les mains crispées de l’ex-petite amie d’un jeune homme), y compris dans le cas où le visage des témoignants n’est pas filmé, contribuent à nourrir la force pathémique de ces capsules.Le récit se focalise sur le jeune parti en Syrie, privilégiant ainsi la troisième personne du singulier, et échappe à toute logique de confrontation ou d’opposition (« ils » contre « nous »). L’objectif est bien de mettre le jeune radicalisé au centre du discours par le biais de son entourage : « il se sentait rejeté, il pensait qu’il ne pourrait pas vivre son islam en France », « pour lui, il était sur le droit chemin, c’était sans retour, plus personne ne pouvait le retenir ». Il s’agit moins de dissuader par le recours à l’avertissement ou à la menace que de livrer un récit de vie authentique en proposant un autre regard sur le phénomène, celui du vécu.Rappelons que le choix de donner la parole à des repentis a d’ailleurs été recommandé par Dounia Bouzar, anciennement mandatée par le gouvernement pour mettre en place des stratégies de « déradicalisation », selon qui les témoignages d’ex-djihadistes sont susceptibles de faire écho aux préoccupations d’individus radicalisés.(17) Précisons que l’entreprise de persuasion à l’œuvre dans ces ouvrages est le plus souvent indirecte ou inavouée, nous autorisant donc à les catégoriser comme des discours à dimension plus qu’à visée argumentative (Amossy, 2008) : le pouvoir persuasif de la narration réside ici dans la volonté du repenti d’imposer une cohérence à son récit et de structurer les circonstances dans une temporalité ou causalité afin d’en maîtriser le sens.(18) Extrait de l’article « Interview de l’épouse de notre frère Aboû Basîr Abdoullâh Al-Ifriqî ». Précisons que l’article n’est pas signé et donc anonymisé.© Cahiers de la sécurité et de la justice - IHEMI84 I DOSSIER Stop-djihadisme ou comment déjouer la radicalisation : examen critique d’un plan de prévention par l’information – Laurène Renaut, Laura AsconeVers des implications pratiques : pistes de recherche et d’applicationSuite aux attentats de 2015, des repentis du djihad ont exprimé leur volonté de s’engager dans la prévention de la radicalisation. Si les enjeux de ces témoignages sont multiples, nous évoquerons ici leur dimension didactique, visant à prévenir des dangers du djihadisme : « Vingt ans après avoir quitté le tunnel mortifère d’un martyr radicalisé, je retrouve l’autre sens du martyr, celui du témoignage. Je dois désormais raconter ce que j’ai vécu […]. Pour expliquer de l’intérieur le mécanisme de la radicalisation » (Vallat, 2016 : 5). Expliquer pour mieux comprendre ce phénomène et donc mieux le contrer : « Je cherche à comprendre comment tout cela est arrivé pour avertir ceux qui passent par là des pièges qui se referment quand il est trop tard » (idem : 7). Ces pièges, recensés par les repentis, sont décrits comme les symptômes d’un processus destructeur qui mène à la dépersonnalisation :– Est ainsi d’abord évoqué un sentiment de révolte intérieure synonyme d’une volonté d’engagement, point sur lequel de nombreux jeunes engagés sur un chemin radical pourront se reconnaitre : « La Yougoslavie se déchire. […] je me sens inconsciemment une âme de soldat. […] Au moment où je reprends les études, je me retrouve dans un dilemme déchirant. » (Vallat, 2016 : 22).– Puis, une fois l’engagement amorcé, tous les témoignages font état d’une impression d’écartèlement ou de tiraillement intérieur qui peut se lire entre les doutes restants et la ligne de conduite dictée par le groupe djihadiste. C’est ce que relate Laura Pasoni. à propos de son mari qu’elle rejoint en Syrie : « Dans ma tête, un lancinant bras de fer s’est installé entre […] l’obligation de vivre en Terre sainte et l’envie de faire demi-tour. Deux petites voix se battent au fond de mon âme. » (Pasoni et Lorsignol, 2016 : 74).– Enfin, le tiraillement semble se muer en un sentiment d’étrangeté à soi-même et d’irréalité. Laura Pasoni se sent ainsi devenir observatrice de sa propre personne : « comme un automate, j’avance tout de même vers la sortie » (ibid. : 12) ; quand Mourad Benchellali parle de lui comme un autre : « je pars, mais j’ai l’impression que ce n’est pas vraiment moi qui pars » (Benchellali, 2006 : 25). Par ces descriptions d’un monde intérieur en crise, les témoignages de repentis visent à pointer l’objectif de déshumanisation qui anime le discours de l’O.É.I. : « mon cerveau fonctionne en mode automatique » (Pasoni et Lorsignol, 2016 : 75).C’est donc dans une perspective d’éclairage par une parole vécue que le récit des repentis trouve sa place : mis au service du décryptage des mécanismes de l’embrigadement, il semble pouvoir offrir des clés de compréhension donc de déconstruction du discours djihadiste. Quant à la bonne foi du témoin, condition première de l’acceptabilité du témoignage, elle est assurée par les co-auteurs des ouvrages autobiographiques des repentis, qui jouent ici le rôle de garant de leur conversion républicaine. Ce travail de déconversion passe sans doute par l’indispensable glissement du remord, qui enferme dans le passé, au regret, tourné vers l’avenir et donc vers l’action. Or, l’action réside là encore dans la voie du témoignage : « J’espère avoir l’occasion de raconter mon histoire dans les écoles. Ce serait pour moi une manière de me racheter et de faire ce que j’ai toujours voulu : aider les autres. » (Pasoli et Lorsignol, 2016 : 194).Aider et s’investir dans la lutte contre la radicalisation, c’est peut-être aussi ce qui anime certaines victimes du terrorisme ou proches de victimes, à l’image de Georges Salines (père de Lola Salines, décédée au Bataclan) et ex-président de l’association 13onze15 Fraternité et Vérité. Ce dernier nous confiait récemment dans un entretien, sa volonté de continuer à s’engager dans la prévention de ce phénomène en effectuant, en plus des activités en milieu scolaire qu’il réalise déjà actuellement, des interventions auprès d’individus incarcérés en raison de leurs liens avec le djihadisme. Et à la question de savoir si ce souhait est partagé par d’autres familles de victimes, Georges Salines répond que « oui absolument, beaucoup de familles, même si elles ne sont pas encore majoritaires parmi les proches des victimes ».Si cette porte de la justice restaurative 19 n’est pas encore ouverte pour les faits de terrorisme, son esprit reste, selon les mots du magistrat et essayiste Denis Salas, « à l’horizon de nos attentes » (Walgrave et Salas, 2015), et en particulier des attentes des victimes, avec la perspective notamment de rencontres dites restauratives ou réparatrices avec des détenus jihadistes. Ces dernières ne pouvant pas avoir lieu sans la participation volontaire des détenus, on pourrait néanmoins envisager des rencontres entre différents (19) La justice restaurative est définie comme « une optique sur la façon de rendre justice après l’occurrence d’un délit orientée prioritairement sur la réparation des dommages individuels, relationnels et sociaux causés par ce délit » (Walgrave, 2008 : 21).(20) Ce groupe se compose de Robert Cario, Géraldine Blin (DFSPIP, SPIP 95), Ouisa Kies (sociologue CADIS/EHESS, association Dialogues citoyens), Alexandre Cartier (CPIP, SPIP 95) et Jérémy Bridier (juriste, Association française des victimes du terrorisme). Voir le site : http://www.justicerestaurative.org/news/radicalisation-terrorisme-et-mesures-de-justice-restaurative/.© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 85 Cahiers de la sécurité et de la justice – Hors-série types de victimes, primaires ou secondaires (famille ou entourage restreint), et même avec des victimes plus indirectes comme les parents ou proches des auteurs du crime.Si de telles rencontres ont un impact pour les parties prenantes immédiates, elles pourraient donc s’étendre à des détenus volontaires pour se mettre en place en prison. En effet, en France, la justice restaurative, entrée dans le code de procédure pénale par la loi du 15 août 2014, ne s’applique pas encore aux faits de terrorisme, même si un groupe de travail s’est formé au sein du service pénitentiaire d’insertion et de probation (SPIP) du Val d’Oise 20, pour développer un programme de justice restaurative appliqué au domaine de la radicalisation.ConclusionSuite aux attentats qui ont frappé la France en 2015, le S.I.G. a lancé la campagne gouvernementale Stop-djihadisme afin de contrer et de prévenir la radicalisation djihadiste par l’information. Cette contribution, qui s’inscrit dans le champ de l’analyse du discours appliquée à des questions de sécurité et de justice, nous a permis d’examiner les stratégies argumentatives et rhétoriques adoptées dans les clips Ils te disent et Toujours le choix. En discréditant l’ennemi djihadiste et en orchestrant la dissonance, le gouvernement vise à dévoiler les atrocités perpétrées par l’O.É.I., afin de dissuader le destinataire d’adhérer à l’idéologie djihadiste et, dans certains cas, à agir au nom de cette idéologie. Cependant, si la mise en scène des massacres dans la vidéo Ils te disent peut dissuader certains individus d’approuver l’action djihadiste, elle prend aussi le risque de fasciner d’autres jeunes attirés par la violence.Précisons que cet article, ne prétendant pas à l’exhaustivité, présente des limites tant de par son objet d’étude que par la nature des recommandations envisagées. En effet, nous faisons d’abord le choix de nous focaliser sur des discours officiels, ici le contre-discours produit par le gouvernement français d’une part, et nous avons conscience de ne pas pouvoir livrer dans cet article une analyse détaillée du discours djihadiste en lui-même d’autre part : même si nous prenons bien évidemment en compte la variété des supports de propagande produits par l’O.É.I., les vidéos et les nachids ne constituent pas le cœur de ce travail.Par ailleurs, en proposant une synthèse de nos recherches respectives dans le domaine de la radicalisation djihadiste, nous livrons ici davantage des pistes de réflexion que des véritables plans d’action. Il serait en outre intéressant de mener une étude comparative des contre-discours produits par l’ensemble des acteurs sociaux (en prenant notamment en compte le travail des associations) dans le monde francophone mais aussi au-delà bien entendu (Royaume-Uni, Canada, États-Unis et Australie).Néanmoins, l’analyse des points forts et des limites de cette campagne gouvernementale nous a conduites à élargir la réflexion sur une forme de récit susceptible de constituer un discours alternatif : le témoignage. Bien que le site Stop-djihadisme présente une section dédiée aux récits de repentis ou de proches de jeunes radicalisés, cette forme de discours apparaît en retrait par rapport aux autres supports constituant cette campagne gouvernementale.L’étude menée sur les spécificités discursives ainsi que sur le potentiel argumentatif du témoignage nous amène à le considérer comme une forme de discours alternatif à la haine comme à la division. Par conséquent, il nous paraît pertinent qu’il puisse trouver sa place au cœur des stratégies déployée pour combattre le phénomène de radicalisation djihadiste nRéférencesAmossy, Ruth, 2008, « Argumentation et Analyse du discours : perspectives théoriques et découpages disciplinaires », Argumentation et Analyse du discours, (1).Amossy, Ruth, 2010, La présentation de soi : Ethos et identité verbale, Paris : PUF, coll. « Interrogation philosophique ».Angenot, Marc, 1989, « Hégémonie, dissidence et contre-discours : réflexions sur les périphéries du discours social en 1889 », Études littéraires, 22(2), p. 11–24.Ascone, Laura, 2018, « La radicalisation à travers l’expression des émotions sur internet », Thèse soutenue le 22 novembre 2018 à l’Université de Cergy-Pontoise.Auboussier, Julien, 2015, « Présentation », Semen, 39. [En ligne : http://journals.openedition.org/semen/10463].Benbassa, Esther et Catherine Trœndlé, Rapport final de la mission d’information sur le désendoctrinement, le désembrigadement et la réinsertion des djihadistes en France et en Europe, Paris, Sénat, 2017.© Cahiers de la sécurité et de la justice - IHEMI86 I DOSSIER Stop-djihadisme ou comment déjouer la radicalisation : examen critique d’un plan de prévention par l’information – Laurène Renaut, Laura AsconeBenchellali, Mourad, 2006, Voyage vers l’enfer. Éditions Robert Laffont.Bouzar, Dounia, et Marie Martin, 2016, « Méthode expérimentale de déradicalisation : quelles stratégies émotionnelles et cognitives ? », Pouvoirs, (3), p. 83-96.Combet-Galland, Corina, 1991, « Quand le récit devient communication : Dialogue avec un reccueil sur la narration », Revue de philosophie et de théologie, n°123, 213-220.Crettiez, Xavier et Bilel Ainine, 2017, Soldats de Dieu, paroles de djihadistes incarcérés, Editions de l’Aube, Fondation Jean-Jaurès.Del Valle, Alexandre, 2009, « Guerre des représentations et virus sémantiques », Géoéconomie, 51(4), p. 119-145.Haddad, Gérard, 2015, Dans la main droite de Dieu : psychanalyse du fanatisme, Paris : Premier parallèle.Kasiki, Sophie et Pauline Guéna, 2016, Dans la nuit de Daesh : confession d’une repentie, Robert Laffont.Vallat, David, 2016, Terreur de jeunesse, le témoignage d’un ex-djihadiste, Paris, Calmann-Lévy.Lausberg, Heinrich, 1960, Handbuch der literarischen Rhetorik, Munich, Max Hueber.Leyens, Jacques-Philippe, Vincent Yzerbyt et Georges Schadron, 1996, Stéréotypes et cognition sociale, Bruxelles, Mardaga.Moniquet, Claude, 2015, Néo-djihadistes : Ils sont parmi nous, qui sont-ils ? Comment les combattre ?, Jourdan.Pasoni, Laura et Catherine Lorsignol, 2016, Au cœur de Daesh avec mon fils, Paris, La Boîte à Pandore.Perelman, Chaïm et Lucie Olbrechts-Tyteca, 1988, (5e éd.). Traité de l’argumentation. Bruxelles, Éditions de l’Université libre de Bruxelles.Plantin, Christian, 2011, Les bonnes raisons des émotions. Peter Lang Publishing Group.Reboul, Olivier, 1991, Introduction à la rhétorique : théorie et pratique, Paris, Presses universitaires de France.Renaut, Laurène, 2019, « Radicalisation djihadiste et discours victimaire sur les réseaux sociaux : de la victime au bourreau », Argumentation et Analyse du Discours, 23 [En ligne : http://journals.openedition.org/aad/3870].Renaut, Laurène et Laura Ascone, 2019, « Contre-discours au discours de haine djihadiste. De l’expression de la conflictualité à la fabrique du doute », Semen, n°47.Salas, Denis, 2018, La Foule innocente, Paris, Desclée de Brouwer.Semelin, Jacques, 2005, Purifier et détruire. Usages politiques des massacres et génocides, Paris, Seuil.Walgrave, Lode, 2008, Restorative Justice, Self-Interest and Responsible Citizenship (Manuscrit), Cullompton, Willan Publishing.Walgrave, Lode et Denis Salas, 2015, « Le terrorisme intérieur. Un défi pour la justice restaurative », Les Cahiers de la Justice, 3(3), p. 423-438.Witte, Kim et Mike Allen, 2000, « A meta-analysis of fear appeals: Implications for effective public health campaigns », Health education & behavior, 27(5), p. 591-615.© Cahiers de la sécurité et de la justice - IHEMICahiers de la sécurité et de la justice – Hors-série DOSSIER I 87“Stop-djihadisme”, or how to thwart radicalization: a critical examination of a plan of prevention through informationLaurène RENAUT, Laura ASCONEfter the Charlie Hebdo attacks in January of 2015, and given the growing number of youths adopting th ideology promoted by the Islamic State Organization (ISO), the French government has implemented successive prevention plans intended to counter the radicalization phenomenon, which were adopted under the authority of the Interministerial Committee for the Prevention of Delinquency and Radicalization (CIPDR) 1. One of the first steps was achieved with the creation of the Stop-djihadisme website, with three main goals: – understanding the terrorist threat and deciphering jihadist propaganda – raising awareness and mobilizing the national communityto unite against that phenomenon– acting to protect the French population and territory.In that perspective, three different formats have been mobilized to fulfil these major communication purposes: articles, videos and a toll-free number. The articles are informative and explanatory, and are considered by the government to be prevention tools. They are organized into five sections: “radicalization”, “jihadist terrorism”, “the fight against terrorism and radicalization”, “what to do?” and “in case This research, which is part of the field of linguistics applied to security or justice issues (and specifically to their preventive dimension), offers a critical analysis of the Stop-djihadisme platform. This prevention plan, launched by the French government in January 2015, aims to prevent jihadist radicalization through information targeted at radicalized youths, their families, educators, associations and the general French public. Educating the population on the radicalization process is at the heart of this government campaign. This study first investigates the argumentative and rhetorical strategies adopted in videos entitled Ils te disent (“They tell you”) and Toujours le choix (“Always a choice”), then examines the strengths and the limitations of this counter-narrative from a discursive perspective. In conclusion, it develops the hypothesis that testimonies (of both ex-jihadists and victims of terrorism) would constitute a potentially relevant alternative discourse to counter jihadist hate speech.Keywords: counter-narrative; testimony; discourse analysis; restorative justice.ALarène RENAUTLaurène Renault is a CELSA graduate in Information and Communication Sciences, and a PhD student in Linguistics at Université Paris-Seine. Her dissertation research focuses on the construction of jihadist radicality online, drawing on a corpus of radicalized profiles, online discussions and conversations, testimonies and court rulings. Laura ASCONELaura Ascone graduated with a Master’s degree in English linguistics from Université Paris-Diderot, and a PhD from Université Paris-Seine. Her doctoral dissertation dealt with the expression of emotions in social networks, jihadist propaganda and its counter-narratives. She is now conducting post-doctoral research at Université de Lorraine for an ANR project (National Agency for Research) on hate speech against migrants. (1) The CIPDR is in charge of animating, coordinating and supporting financially the delinquency and radicalization prevention policies.© Cahiers de la sécurité et de la justice - IHEMI88 I DOSSIER “Stop-djihadisme”, or how to thwart radicalization: a critical examination of a plan of prevention through information – Laurène Renaut, Laura Asconeof an attack”. In other words, while the goal of jihadist discourse is to incite the receivers to act, the educational articles which make up a great part of French institutional counter-narratives, rather seeks to inform them (Ascone, 2018). However, the “what to do?” section, and especially the article entitled “Becoming an actor of the fight against radicalization and terrorism”, appear as an attempt to place the citizen at the center of the action and the fight against violent radicalization. This intention is also reflected in the toll-free number, established to facilitate the reporting of radicalized individuals, as well as to “support the families of radicalized persons”. Lastly, two video campaigns stand out on the government website, entitled Ils te disent (“They tell you”) and Toujours le choix (“Always a choice”). The latter is divided in two stories, that of Mehdi and that of Emma, and also introduces a section dedicated to testimonies, Ils l’ont vécu (“They lived it”). It mostly refers to the families of radicalized youths, although one video also shows a reformed ex-jihadist’s cautionary tale, aimed at dissuading other youths from joining the ranks of the ISO. Whereas the articles and the toll-free number are directed to the general public and to the relatives of radicalized individuals, the videos produced by the Government Information Service (SIG) specifically target a radicalized audience. Thus, these elements demonstrate that government institutions have taken to the Internet as a new space of symbolic confrontation against the ISO. The goal: to deflect hate speech. Therefore the fight against terrorism is not only set in Syria and Iraq, but also on the disputed territory that is the Internet, through a prevention strategy mostly executed through detection and awareness campaigns.The phrase counter-narrative implies that the discourse is established, by definition, “in explicit antagonism” (Angenot, 1989, p. 1) to another discourse. It is then necessary to understand the devices and mechanisms of the initial narrative, in order to better grasp the specific characteristics of the narrative produced in reaction (Auboussier, 2015). It should also be noted that in this research, we chose to analyze the French institutional counter-narrative in relation with the jihadist discourse of Dar al-Islam, the French language magazine published by the ISO, specifically by its media branch Al-Hayat Media Center. Since 2014, it has mostly been publishing magazines: Dabiq, published in English, totals 15 issues; the 10 issues of Dar al-Islam were published in French; and Rumiyah 2, translated into eight languages (English, Turkish, Uyghur, Pashto, Russian, German, Indonesian and French), has had 13 issues. Al-Hayat Media Center also publishes videos and audio recordings of chants (nasheeds) that aim to promote the jihadist ideology. However, the present study only deals with the magazines, and especially, with Dar al-Islam. This choice was made in order to analyze discourse which firstly, was specifically directed at French-speaking radicalized audiences, and secondly, was accessible anonymously – a necessary precaution for a researcher who isn’t connected to institutional structures. 3 Although this propaganda magazine is published on the Internet, it presents all the characteristics of a printed magazine: it has a table of contents, interviews, photo reports or articles about the ISO’s operations… Its 10 issues (229 762 words) constitute an important corpus of data. In spite of its obvious limitations, this choice also allowed us to study two official discourses: on the one hand, the Stop-djihadisme website, through the analysis of the video campaigns 4 on which we will focus our attention, and on the other hand, the Dar al-Islam magazine. In that context, this contribution, within the field of discourse analysis applied to security and justice issues, offers a critical examination of the Stop-djihadisme platform, presented by the government as a prevention plan aimed at thwarting radicalization through information. More specifically, this research presents an analysis of the argumentative and rhetorical strategies deployed in the videos selected. In this regard, rhetoric should be understood as “the art of persuading through discourse” (Reboul, 1991, p. 4) with the purpose of triggering action (Plantin, 2011, pp. 17-18) in order to “provoke or increase adhesion in the minds of those to whom propositions are presented for agreement” (Perelman and Olbrechts-Tyteca, 1988, p. 59; see also Amossy, 2008). With that in mind, in this paper we will first examine the argumentative strategies of the videos published (2) In fact, Rumiyah came as a replacement for Dabiq (in English), Dar al-Islam (in French), Istok (in Russian) and Konstantiniyye (in Turkish), which were published until the end of 2016.(3) The corpus of magazines studied here was gathered in 2016, when the jihadist propaganda material was still accessible (for example as PDF files on the jihadology.net website) without having to register with a specific username and password as is now the case.(4) Ils te disent (http://www.Stop-djihadisme.gouv.fr/stopdjihadisme-retour-2-ans-lutte-contre-propagande-djihadiste) and Toujours le choix (http://www.toujourslechoix.fr/). The Toujours le choix campaign follows the stories of two characters, Emma and Mehdi, and therefore consists of two videos. © Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 89 Cahiers de la sécurité et de la justice – Hors-série on the platform. Given the plurisemiotic nature of the material, we will carry out a semiodiscursive analysis (combining text and image) of these videos, and adopt a qualitative approach. We will then present the strengths and limitations of the campaign from a discursive point of view. Finally, based on the analysis of the Ils l’ont vécu section of the Stop-djihadisme website, we will broaden the discussion to consider the necessity to produce a form of narrative that could constitute a relevant alternative discourse: we will thus look into the characteristics and role of testimonies, whether they be testimonies of reformed ex-jihadists, or of direct and indirect victims of terrorism. Stop-djihadisme: the goals of the digital platformDiscrediting the otherThe first campaign video produced by the SIG, Ils te disent, denounces the mendacious nature of jihadist discourse by contrasting the ISO recuiters’ promises with the crimes committed by its members (figure 1): This government campaign relies on non-military weapons (words, images and information) which participate in a war of representations, centered on the other. Indeed, it is based on the discredit-guilt-overturn triad, which promotes a form of “symbolic, psychological and subversive war” (Del Valle, 2014, p. 3) waged in parallel with military operations. In that context, discrediting the other in the eyes of youths who might be receptive to jihadist ideology is achieved through a process of demonization, intended to dehumanize the enemy. The prediction “you will discover hell on earth” thus refers to a deathly place, but also to absolute evil, i.e. the ISO. Yet whereas the designations to name Westerners abound in jihadist propaganda material (449 occurrences of the term “infidels” over the 10 issues of Dar al-Islam, and 194 occurrences of “apostates”), the unspeakable evil denounced by the government campaign is alluded to through pronominalization: the men of the ISO are indeed referred to as “they” 6 times, which is a way to avoid directly naming the opponent. The repetition of that pronoun thus constructs a binary opposition (“they” versus “us”) which aims at rejecting “them” and directly addressing the youth, hence the use of “tu”, the singular, informal, familiar French word for “you”. The unnameable isn’t named, but is nonetheless constantly pictured in the video through its barbaric actions. Indeed that other, which allegorizes death and deceit, is shown as the perpetrator of massacres, via images of war, crucified bodies and beheadings. However, it should be noted that although the SIG mobilizes similar codes to th jihadist propaganda videos by using official ISO images, the Ils te disent video clearly distinguishes two opposite camps: that of the enemy, who chooses barbarity, and that of its detractors, who reject it, and only partially show it. This paradoxical representation of violence, which is shown but understated, thus helps create an unveiling effect: the video asserts that the victims of the ISO’s massacres are not only its claimed target (Bashar al-Assad’s army), but also civilians.The enemy, in addition to being otherized and demonized, is also ridiculed through stereotypes 5 that produce a caricatured vision of jihadist recruiters. The main element used to communicate this form of discredit is language: indeed, the representation of a written interaction on Facebook between a radicalized youth and a recruiter convey the idea of poor language skills in terms of spelling, especially in the Ils te disent video. In that campaign, after viewing djihadist content on Facebook, the protagonist receives a message from a recruiter (figure 2). Figure 1: Example of a sequence from the Ils te disent videoAbove: “They tell you: sacrifice yourself with us, you will defend a just cause” Below: “In reality: you will discover hell on earth and die alone, far from home” (5) The term “stereotype” refers to “shared beliefs regarding the personal attributes of a human group.” (Leyens et al., 1996 : 11).© Cahiers de la sécurité et de la justice - IHEMI90 I DOSSIER “Stop-djihadisme”, or how to thwart radicalization: a critical examination of a plan of prevention through information – Laurène Renaut, Laura AsconeAlthough the graphical choices mimic spelling and style based on orality, in order to achieve a realistic effect and to represent the “youth” language, the effect of these choices could lead some viewers to assign a social category to the depicted characters, and assimilate this lingo to “street slang”. Besides, this possible interpretation also seems to be reinforced phonetically by the speech patterns of the recruiters, which present all the prosodic clues of the so-called “urban” accent, audible in the Toujours le choix video.Furthermore, the recitation-like nature of the indoctrination discourse reproduced in the latter video (“We have to help the Muslims who are suffering. [...] We are gonna help you, you won’t be alone anymore”), and the editing choices that place this sequence right after the scene of the first encounter with young Emma, could contribute to oversimplifying the mechanisms of this recruiting system. Indeed, what is presented in the video as a rapid tipping point for youths on the road to radicalization, is in reality a process, the duration and trajectory of which depend on several social, cultural and psychological factors (Moniquet, 2015). It is also well-known that “such a conversion doesn’t occur like a bolt from the blue, but rather as a slow process” (Haddad, 2015, p. 110). Lastly, in addition to these strategies, the use of remorse, or in other words, the attempt to appeal to a form of guilt, is obvious in the slogan of the Toujours le choix campaign: “Becoming radicalized means destroying one’s family, one’s life and that of others.” Indeed that sentence serves a pathematic function, as it refers to the suffering of one’s relatives. Here, causing guilt hinges upon a logic of victimization, wherein the victim figures are several: the radicalized youths, their loved ones, and more generally, the French population as victim of terrorist attacks. However, within the institutional counter-narrative, guilt is caused by violent actions committed in the name of jihad, whereas in the ISO’s propaganda, it comes on the contrary from inaction in the face of the bombings and humiliations endured by Muslims: “Are you going to let the infidel sleep peacefully in his house while the wives and children of Muslims are shaking in fear, frightened by the of the crusaders’ planes above their heads day and night.” (Dar al-Islam 2, 2015 : 6). The goal, in both cases, is to reveal the immoral acts perpetrated by the enemy and to trigger emotions (fear, shame, anger, etc.) in order to provoke collective behaviors. Orchestrating dissonanceThe device used in the Ils te disent video is based on designing a polyphonic narrative, in which two discordant voices are articulated in order to reveal the mendacious nature of the ISO’s discourse: the voice of the jihadists, kept at a distance by the recourse to direct speech (“they tell you: ‘come start a family with one of our heroes’.”), and the institutional voice introduced by “in reality”. In this way, the government’s communication orchestrates dissonance in order to disrupt the radicalized youths’ belief system and their grasp of reality. To that end, the campaign resorts to semantic inversions which consist in turning the enemies’ own references against them, in order to deconstruct the fantasized vision of jihadist ideology. Thus, the shift from the philosophical sense of the term “truth”, abundantly used by the ISO 6 and associated to its universal value in “the truth is here”, to a plural form of the term in “the (6) There are 111 occurrences of the term “truth”, only in singular form, in the Dar al-Islam magazine. Figure 2: Example of a first contact established by a recruiter with his target on Facebook (screenshot from the Ils te disent video) “Hi The stuff you ‘like’ on here is cool, u interested in what’s going on in the Sham these days? If u have questions feel free, the truth is over there, now is the time to leave!if you want more info, give me ur number I have friends over there fightin, I put u in touch” © Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 91 Cahiers de la sécurité et de la justice – Hors-série only truths you’ll discover are horror and deceit” (figure 3), contributes to the twist in meaning. The figure of the humanitarian worker (“Come and help the Syrian children”) is replaced by that of the war criminal (“You will be complicit in massacres”). Similarly, the sacrificial logic that leads martyrs to paradise, “Sacrifice yourself with us, you will defend a just cause”, becomes a warning: “You will discover hell on earth”. The government plan for the prevention of radicalization through information draws upon a denunciation of jihadist talking points, through a process of reversal of meaning. Furthermore, the representation of the radicalized youth themselves is turned upside down, turning them from mythical heroes to victims manipulated by the jihadist group. From dreams to reality: such is the main shift that operates through the Ils te disent campaign. First and foremost from an iconographic point of view, the jerky editing alternates between images in color for the jihadists’ promises, and black and white images for reality. Regarding the soundtrack, a piece of religious music by Omar Omsen 7, often used in the ISO’s musical propaganda to exalt its recruits’ faith in the fight, is deliberately exploited here by the SIG. However, this melodic chanting is interspersed with static interference, which gives the impression of someone tuning to find the correct frequency, as if to escape the official discourse of the ISO. The cries of weeping women and children can also be heard throughout the video, which is also replete with images of civilians mistreated on caliphate land. This dramatization of the contrast between fiction and reality intends to confront radicalized viewers to information inconsistent with their beliefs, and to re-instill doubt in their minds.Dissuading actionWhile the government’s counter-narrative was presented by then Prime Minister Manuel Valls as a good-intentioned attempt to “bring back these young people [...] to the France that loves them” 8, it is however not free from power struggles. In fact, the Toujours le choix campaign really depicts none other than a showdown between the authorities and individuals susceptible of turning against their own country: it is no coincidence that it shows young Mehdi being arrested and sentenced to jail after fomenting a terrorist attack in France. The threat here is explicit (coming from the State itself), and coercive, as it aims to compel the ISO’s supporters to obey the law, using the spectre of imprisonment. However, the government mostly warns radicalized youths of the risks they incur (suffering a violent death in the Ils te disent video, or Emma being repudiated by her husband in Toujours le choix), in order to build an ethos 9 of authority, where recommendations become commands through the evocation of punishment. The law of the Republic is opposed to the law of Allah in jihadist discourse, and is put to use in a strategy of persuasion through fear, which allows to influence behaviors (Witte and Allen, 2000). It must be noted that, while the institutional counter-narrative places fear at the center of the emotional rhetoric it develops, it is only the better to dissuade the youth from acting, whereas jihadist discourse uses it as a springboard to action. Figure 3: Example of a sequence from the Ils te disent video displaying the iconographic strategy adopted by the SIG On the right: “They tell you: ‘you’re living in a world of impure infidels, the truth is here” On the left: “In reality the only truths you’ll discover are horror and deceit.”(7) The author of the first French-language jihadist videos.(8) Speech given by Manuel Valls on November 18th, 2016.(9) The concept of ethos refers to the discursive image of self, i.e. the presentation of oneself through discourse (Amossy, 2010).© Cahiers de la sécurité et de la justice - IHEMI92 I DOSSIER “Stop-djihadisme”, or how to thwart radicalization: a critical examination of a plan of prevention through information – Laurène Renaut, Laura AsconeStrengths and limits Directed at a diverse audience and consisting of multiple formats (articles and videos), the Stop-djihadisme device is part of a broad plan for the prevention of radicalization, operating through information and communication. As it has spurred, from its outset, controversial reactions within the professional and academic spheres, we offer here a critical reflection, resulting from the previously exposed semiodiscursive analysis.The choice to represent violence The choice to use a website as a tool for the dissemination of its institutional counter-narrative, has allowed the government to articulate its prevention plan around multiple formats: articles, videos, and a toll-free number meant to contribute to the detection of radicalized individuals. In fact, since its opening on April 29th, 2014 to June 30th, 2017 (Benbassa and Troendlé, 2017 : 15), the National Center for Assistance and Prevention of Radicalization (CNAPR) has received 51 429 calls, and a total of 5 723 individuals were reported 10; these reports were passed on to specialized services in charge of assessing and confirming (or not) the level of dangerousness of each person. The heterogeneity of contents is, in our view, one of the strengths of the Stop-djihadisme platform. These campaigns make it possible to share knowledge and provide accessible education on the complex phenomenon of radicalization, by reaching a diverse audience and generating reactions: in that regard, the Ils te disent video registered nearly 2 million views in only two days. And if the video sparked debate, it is especially due to the use of violent images, which have drawn contrasting opinions.Indeed, many of the ISO’s recruits who travelled to Syria or Iraq, admitted in retrospect having been fascinated by the beheading videos circulated by Al-Hayat Media Center and more generally, by violence. A notable example is the case of a young Afghan refugee, aged 25, who was arrested in 2017 by the antiterrorist police force and who, having become addicted to violent images, downloaded no fewer than 2000 videos from the Islamic State’s online networks. 11 This case confirms media sociologist Hasna Hussein’s analysis, who posits that these videos, by alternating images of combat and images of majestic exotic landscapes, trigger “a form of sensory excitation in viewers” (Hussein, 2019). Let us not forget that Salafi jihadist ideology legitimizes the recourse to armed struggle by evoking the theory of the “clash of civilizations” and exploiting current events in the Middle East in order to justify fighting the West, alleged to be waging a war against Islam. Therefore, a video focused on violence, which is considered an end in and of itself within the jihadist community (Crettiez and Ainine, 2017), runs the risk of increasing that fascination in a radicalized audience. Besides, while the strategy of dehumanizing the enemy implemented in these videos can turn out to be applicable to some individuals being initiated to jihadist ideology, and whose cognitive reasoning is still open to other views, it could result inefficient when dealing with ISO supporters who are already deeply engaged in the process, and who approve of the actions conducted in the name of armed jihad. Indeed, as emphasized by Bouzar and Martin (2016), feedback shows that the time dedicated to behavioral change in a radicalized youth is correlated with the individual’s level of radicality: the earlier the process is diagnosed, the fastest the recovery will be. Conversely, the more advanced the youth is in the process, the more completed is the phenomenon of dehumanization of the other: in fact, in jihadist propaganda, the figure of the Westerner is the object of a surge of images of extreme violence, aimed at justifying and encouraging war against him. 12 This dehumanization thus contributes to establishing the psychological distance that is necessary to catalyze violence against a group of people (Semelin, 2005), but also to render inaudible the words of the enemy, in this case, the French government.Lastly, it seems crucial to us to point out a paradox that should be taken into consideration regarding the production of a counter-narrative: “indeed, although victimization is omnipresent in nearly every word in the jihadosphere, it must be noted that its members display utter contempt for the figure of the victim; insults such as ‘he’s playing the victim’ or ‘he’s such a victim’ are frequently (10) As of June 30th, 2017, according to the Benbassa-Troendlé report, op. cit., 2017, p. 15.(11) Lyonmag Website (article published September 29th, 2019) :https://www.lyonmag.com/article/103293/villeurbanne-le-refugie-afghan-etait-devenu-accro-aux-videos-de-decapitations-de-daesh(12) For example, in the nasheed entitled “Avance, avance” (“Keep going”), the enemy is demonized. Qualified as “soldiers of the devil” and “criminals”, the Western others must be annihilated because they are a source of destruction, according to the laws of a binary logic of victimization (it’s either them or me): “You kill them or they kill you”. The song is written in a warlike register, where the other is referred to as an anonymous mass, as emphasized in the anaphoric repetition of “Tue-les” (“Kill them”).© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 93 Cahiers de la sécurité et de la justice – Hors-série used to target the community’s enemies” (Renaut, 2019, p. 28). Therefore, it becomes necessary to question the official language featured in the Ils te disent video and the ambivalent representation it conveys of young jihadists, shown as actors of radicalization but mostly as victims of a discourse of indoctrination. As for the Toujours le choix campaign, on the other hand, it offers the advantage of highlighting the notion of personal responsibility, better suited to an audience reluctant to identify with the figure of the victim. Interrogating the institutional labeling process As we have previously shown, at the heart of the SIG’s strategy of communication lies the dramatization of the gap between the hopes generated by the ISO’s propaganda, and the reality of life in Raqqa (figure 4). The staging of this contrast aims to disrupt the viewer’s belief system. However, to re-instill doubt in a radicalized audience, it is important to anticipate the way the video will be perceived according to the targeted group’s specific understanding of the world. Now in the case of the youths susceptible to be affected by that ideology, they are so stuck in a conspirationist vision of their environment that they risk perceiving this video as just another form of manipulation coming from an impious government.In that respect, the institutional labeling of these videos in itself is an element that should be questioned: indeed, the choice of a government website takes the risk of being deemed fallacious or manipulative, and therefore fueling the conspirationist views of the ISO, thus strengthening the youths’ adhesion to their initial belief system. That danger was in fact highlighted by Mourad Benchellali, an ex-jihadist who was invited by Christian Gravel (former director of the SIG) to collaborate on the creation of videos, and who allegedly refused because “the very format itself [...] was a problem” to him. 13 This problematic element, according to him, would lead to an absence of interest in the campaign from its targeted audience, as he explained in an interview given to newspaper Libération: “I had the opportunity to listen to youths from Vénissieux speak on the matter. First of all, they didn’t know the website, although they are theoretically the targeted audience. Then, I asked if they would want to visit it. Once again, they said no. They were convinced that they would find more propaganda than information there, in their own words.”14 In this respect, the binary discourse system adopted here to oppose jihadist myth to reality seems easy to reverse. For example, the words seen on figure 2 can be turned around as such: “They tell you: ‘the only truths you’ll discover will be horror and deceit’. In reality, you are living in a world of impure infidels, the truth is here.” It should be noted however, that if the government is trying to deconstruct the promises made by the ISO to its supporters, it is precisely because reward is central in jihadist discourse. A beautiful house, a great income, several wives, the forgiveness of Allah are among the most enticing promises. The strategic use of reward aims to motivate radicalized youths to act in the name of jihadist ideology. Leaving for Syria will allow them to lead a life of glory and adventure, and dying as martyrs will guarantee them access to paradise. On the opposite, the institutional counter-narrative seeks to dissuade from action: it is about not joining the ISO, not committing (13) Interview given by Mourad Benchellali to Libération dated February 16th, 2016 : https://www.liberation.fr/france/2016/02/08/rien-ne-bouge-sur-les-contre-discours-a-la-radicalisation_1431915 (14) Ibid.Figure 4: Example of a sequence from the Ils te disent video showing the gap between the ISO’s promises and reality On the right: “They say ‘come start a family with one of our heroes’” On the left: “In reality you will raise your children in war and terror.”© Cahiers de la sécurité et de la justice - IHEMI94 I DOSSIER “Stop-djihadisme”, or how to thwart radicalization: a critical examination of a plan of prevention through information – Laurène Renaut, Laura Asconeattacks on French soil. Similarly, the articles published on the website, directed at the general population rather that the radicalized youths, assert that any action against jihadist terrorism must be undertaken by the State and not by citizens (Ascone, 2018). This is why, although these campaigns don’t exclude the notion of individual freedom, as is emphasized by the title Toujours le choix (“Always a choice”), choice here is mostly limited to renouncement (not leaving for Syria), without a real alternative to fulfil the urge to act. This absence of propositions might be the main weak spot of that discourse, in addition to the very format of a government website, potentially deemed untrustworthy by the targeted audience. All these elements lead us to question who the real targets of that counter-narrative are: while it might not be destined to reach the most indoctrinated individuals, it is more broadly directed at families, educators and members of civil society organizations, in order to raise awareness. Besides, as it was conceived to take up online space and compete against hate speech, it is also a way to strike back through communication, directed at the ISO as well as the global French population in a war context. Global assessment and outlook The government’s campaign thus presents strengths, such as the use of various formats, the accessible dissemination of academic research, and the will to reach out to a diverse audience, but it also has limitations. Indeed, although in January 2015, the anti-jihad video Ils te disent broke records, generating nearly 2 million views, the impact of the campaign remains however difficult to evaluate. The SIG actually admits it: it doesn’t have the technological means to know who is visiting the Stop-djihadisme website, and therefore has no way to know whether it is reaching a radicalized audience, or how such an audience perceives it. In this respect, the study of radicalized profiles on Facebook tends to show that there is little feedback or debate on the Stop-djihadisme campaign within the jihadosphere (Renaut, 2019). Whenever the Ils te disent video is being shared, it is mostly associated with a satirical tone: the content of the videos and the figure of the jihadist recruiter are, in particular, mocked by some of the youths who express on social networks their attachment to the ideology promoted by the ISO. We therefore put forth the hypothesis that the source distributing this counter-narrative is met with a level of mistrust. It must not be forgotten that the individuals who adhere to jihadist ideology reject French democracy and its institutions. This is why we shall not formulate recommendations for the production of an institutional counter-narrative strictly speaking. Instead, we will propose, based on the issues raised by our analysis, to explore other forms of alternative discourse that seem to us relevant in the context of the fight against radicalization.From counter-narrative to alternative discourse: the role of testimonies Although this type of discourse appears peripherally on the Stop-djihadisme website, testimonies are a potential counter-narrative that, according to us, deserves to be developed and given center stage. Through an analysis of the specificities of that discourse, we will show to what extent it could provide a basis for an alternative discursive strategy. In this respect, what is meant by “alternative discourse” is a counter-narrative that isn’t constructed in opposition to another discourse, i.e. a separate discourse that wouldn’t present itself as a counter-narrative. Here, the specific form of the testimony, wherein conflictuality is internal instead of directed against another, fits the definition.Argumentation through lifenarratives: a “weapon of mass persuasion” 15The word testimony refers to the fact of telling, of delivering a narrative, affirming the veracity of what has been seen, heard, perceived or lived. Such a narrative holds the specificity of sharing, rather than asserting domination by hammering one’s point of view: in other words, “it signals instead of grabbing” (Combet-Galland, 1991), and causes the receiver to initiate an effort, which seems relevant and should be taken into account with regards to the specific audience considered within this article. Besides, while in a narrative, argumentation isn’t imposed, it can be latent, expressed subtly in the background: a testimony can thus constitute a strong (15) Phrase uttered by ex-jihadist David Vallat, in a passage of the documentary entitled “Jihad : faut-il croire les repentis ?”shown as part of the TV program Complément d’enquêtes on national channel France 2.© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 95 Cahiers de la sécurité et de la justice – Hors-série argument, which mostly relies on the personality of the speaker 16, as the credibility afforded to the witness always influences that afforded to the testimony (Perelman and Olbrechts-Tyteca, 1988, p. 420). But one of the difficulties that seems to arise in the early stages of producing a counter-narrative, precisely resides in establishing a trust relationship between the source and the receiver, in order to create conditions favorable to self-questioning in a radicalized audience. This point must be considered when choosing which actors to include in the disengagement process, even before determining the rhetorical strategies that will be developed to instill doubt in a radicalized person’s mind. It is in fact the main issue raised by Mourad Benchellali as he evokes his involvement in the prevention of that phenomenon: “You must know what you’re talking about. And who better than a reformed jihadist can speak about the reality of living on a land of jihad, about what it implies?” But while lived experience is necessary to provide a discourse of authenticity, there are obstacles hindering the intervention of reformed jihadist in schools or prisons, because of their criminal records. Nonetheless testimony, as a form of argumentative discourse 17, is set apart by its emotional potency, as it mobilizes three pathematic devices (Lausberg, 1960, p. 257):– First, the speaker shows emotion, expresses emotion, and communicates emotion to the audience through empathy. – Second, the speaker includes other individuals in the testimony, which attests to the referential truth of the delivered narrative (parents, wife, husband, partner, friends susceptible to refute the story in case of inaccuracy). – Finally, the speaker evokes events that are, by definition, emotionally touching.In the present context, testimonies, regardless of the status of those testifying (reformed jihadists or victims of terrorism), could constitute relevant tools in the prevention of radicalization, or even in adressing prison inmates. It is indeed no coincidence that the testimony genre has been exploited in the ISO’s propaganda, for example in the Dar al-Islam magazine, in the form of interviews of women expressing their happiness for having reached the land of the caliphate: “Praise be to Allah who made my journey easy, I have met no difficulty. [...] I feel relieved that I accomplished this obligation al-hamdoulillah.” 18Testimony in Stop-djihadisme : Ils l’ont vécuThe government, it seems, understood the benefits of testimony, and made room for it on the Stop-djihadisme website, more specifically in the Toujours le choix campaign, by creating the Ils l’ont vécu section, where the floor is given to reformed jihadists and relatives of individuals who left for combat zones. However, it is obvious here that the SIG favored the testimonies of families over the testimonies of radicalized individuals themselves. In the campaign, only David Vallat talks about his experience, while Mourad Benchellali declined the invitation to appear in these videos. Nonetheless, in spite of the recurrent problem of coming from an institutional source, these testimonies have the merit of favoring a direct approach to the audience, manifest in the way David Vallat looks straight into the camera, as well as in his use of “tu”, the singular, informal, familiar French word for “you” (for example when he says “on the ment”, meaning “they’re lying to you”). Additionally, a few non-verbal elements, including in the cases where the witnesses’ faces were not filmed (like the ex-girlfriend of a young man, nervously wringing her clenched hands in her lap), contribute to the pathematic power of these video clips. The narrative focuses on the youth who left for Syria, is mostly formulated in third person singular, and is exempt from any kind of opposition or confrontation logic (“us” versus “them”). The goal is indeed to put the radicalized youth at the center of the discourse, via his relatives: “he (16) We are using the terminology of Perelman and Olbrechts-Tyteca (1988). “Thus, for reasons of technical convenience, and in order not to lose sight of the essential role played by the audience, when we use the terms “discourse,” “speaker,” and “audience,” we shall understand by them, respectively, the argumentation, the one who presents the argument, and those to whom it is addressed. We shall not dwell on whether the presentation is spoken or written, or distinguish between formal discourse and the fragmentary expression of thought.” (p. 9) (17) It should be noted that in most testimonies, the persuasion endeavour is indirect or unavowed, which allows us to classify them as discourse with an argumentative dimension, rather than discourse with argumentative intent (Amoss, 2008): here, the persuasive power of narration resides in the reformed jihadists’ will to bring coherence to their story, and to articulate the circumstances in terms of temporality or causality in order to control their meaning. (18) Extract from the article entitled “An interview of the wife of our brother Aboû Basîr Abdoullâh Al-Ifriqî.” It should be noted that the article isn’t signed, and is therefore anonymized. © Cahiers de la sécurité et de la justice - IHEMI96 I DOSSIER “Stop-djihadisme”, or how to thwart radicalization: a critical examination of a plan of prevention through information – Laurène Renaut, Laura Asconefelt rejected, he thought he would not be able to fully live his Islam in France”, “in his eyes, he was on the right path, there was no turning back, no one could hold him back anymore.” It is less about dissuading through warnings or threats, than about delivering authentic life narratives, offering new insight into the phenomenon through lived experience. It must be remembered that the choice of having reformed jihadists speak was actually recommended by Dounia Bouzar, formerly commissioned by the government to establish “deradicalization” strategies, who argued that the testimonies of ex-jihadists would be relatable to radicalized individuals as they were susceptible to resonate with their concerns. Towards practical involvement: leads for research and applicationFollowing the 2015 attacks, several reformed jihadists have expressed their will to get involved in the prevention of radicalization. While these testimonies carry multiple implications, we will evoke here their didactic dimension, their intent to caution against the dangers of jihadism: “Twenty years after reaching the end of the deadly tunnel of radicalized martyrdom, I recover the other meaning of martyrdom, that of testimony. I must now tell what I have lived. [...] To explain from within the mechanism of radicalization” (Vallat, 2016, p. 5). Explaining in order to better understand the phenomenon, and therefore to better counter it: “I seek to understand how all this happened so as to caution those who follow that path against the traps that close in on you when it is too late.” (Ibid., p. 7) These traps, inventoried by the reformed jihadists, are described as the symptoms of a destructive process that leads to depersonalization: – First, they evoke a feeling of inner revolt conjugated with a will to get involved, a point with which many youths engaged in a radicalization process will relate: “Yugoslavia was torn. [...] deep down, unconsciously, I felt like a soldier at heart. [...] Just when I was returning to school, I found myself facing an excruciating dilemma.” (Vallat, 2016, p. 22).– Then, once the involvement begins, all the testimonies describe a sensation of being torn by agonizing uncertainty, which reads as an inner conflict between the remaining doubts and the rules dictated by the jihadist group. This is what Laura Pasoni narrates, regarding her husband whom she is about to join in Syria: “In my head, a harrowing arm-wrestling match began between [...] the obligation to live in the Holy Land, and the desire to turn back. Two little voices were fighting deep in my soul.” (Pasoni and Lorsignol, 2016, p. 74).– Lastly, the inner conflict seems to mutate into a feeling of irreality, of being a stranger to oneself. Laura Pasoni thus feels herself becoming an observer of her own person: “mechanically, like an automaton, I walked towards the exit anyway” (Ibid., p. 12). Mourad Benchellali speaks of himself as another: “I left, but I had the impression that it wasn’t really me leaving.” (Benchellali, 2006, p. 25) Through these descriptions of an inner world in crisis, the testimonies of reformed jihadists intend to point out the objective of dehumanization that drives the ISO’s discourse: “my brain was on autopilot.” (Pasoni and Lorsignol, 2016, p. 75) The role of reformed jihadists narratives’ is thus to bring a new perspective and shed light on radicalization through a discourse of lived experience: as they help decipher the mechanisms of indoctrination, they provide keys to understand, and therefore deconstruct jihadist discourse. As for the good faith of the witness, which is the primary condition for a testimony to be receivable, it is vouched for by the co-authors of the reformed jihadists’ autobiographical works, who act as guarantors of the sincerity of their conversion to the values of the Republic. In order for such work to be efficient, it is essential to operate a shift from remorse, which leaves one stuck in the past, to regret, which is geared towards the future and therefore invites action. And that action, once again, takes precisely the form of testimony: “I hope I will have the opportunity to tell my story in schools. That would be a way for me to redeem myself, make amends, and do what I’ve always wanted to: help others.” (Pasoli and Lorsignol, 2016, p. 194). Helping and getting involved in the fight against radicalization is also what motivates some survivors of terrorism and family members of victims, such as Georges Salines (father of Lola Salines, who died in the Bataclan attack), former president of an association named “13onze15 Fraternité et Vérité” (“November 13th 2015 Fraternity and Truth”). He recently told us, during an interview we conducted with him, that he would like to get further involved in the prevention of that phenomenon by organizing, in addition to the presentations he is already giving to students in schools, activities directed at an audience of individuals incarcerated for their implication in jihadism. And, when asked if this wish is shared by other families of victims, Georges Salines answers: “Yes absolutely, many families, © Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 97 Cahiers de la sécurité et de la justice – Hors-série although not yet by the majority of the victims’ relatives.” If the option of restorative justice 19 hasn’t been made possible yet for acts of terrorism, its spirit is, in the words of Denis Salas, “on the horizon of our expectations” (Walgrave and Salas, 2015), and especially the expectations of the victims, including through the perspective of restorative encounters with jihadist inmates. Such encounters cannot be held without the voluntary participation of the inmates, but other types of encounters could nonetheless be envisioned, for example between different types of victims, primary (the persons directly affected) and secondary (their families or close social circles), or even with more indirect victims such as the parents or relatives of the criminals. If such encounters have a positive impact on the people immediately affected, they might then be extended to voluntary inmates to be organized in prisons. Indeed, in France, restorative justice, which entered the Code of Criminal Procedure after the law of August 15th, 2014, isn’t applicable yet to acts of terrorism, although a task force has been established within the Penitentiary Service of Rehabilitation and Probation (SPIP) of the Val d’Oise department 20 in order to discuss the possibility to develop a program of restorative justice in relation to radicalization issues.ConclusionSince the attacks that hit France in 2015, the SIG launched the Stop-djihadisme government platform in order to counter and prevent jihadist radicalization through information. This contribution, conducted within the framework of discourse analysis applied to issues of security and justice, allowed us to examine the argumentative and rhetorical strategies adopted in the Ils te disent and Toujours le choix video campaigns. By discrediting the jihadist enemy and orchestrating dissonance, the government aims to reveal the atrocities perpetrated by the ISO in order to dissuade the receiver from adhering to the jihadist ideology and, in some cases, from acting in the name of that ideology. However, while the representation of massacres in the Ils te disent video may dissuade some individuals from approving jihadist actions, it also runs the risk of fascinating other youths who might be attracted to violence.It should be noted that our article, which does not claim to be exhaustive, is also limited, both by its subject matter and by the nature of the recommendations considered here. Indeed, we first chose to focus on official discourse, in this case the counter-narrative produced by the French government, and are conscious that we can’t provide in this article a detailed analysis of jihadist discourse itself: although we certainly took into account the diversity of the propaganda material produced by the ISO, the study of its videos and anasheeds does not constitute the heart of our work. Besides, by offering an overview of our respective research projects in the field of jihadist radicalization, we are presenting here a few ideas and reflections to consider rather than actual plans of action. It would be interesting to further this exploration by conducting a comparative study of the counter-narratives produced by all the social actors (including the work of civil society organizations) in the French-speaking world and beyond (UK, Canada, USA and Australia). Nonetheless, the analysis of the strengths and limitations of this government campaign led us to broadening the reflection to include a form of narrative susceptible to constitute a relevant alternative discourse: testimony. Although the Stop-djihadisme does present a section dedicated to the narratives of former jihadists and relatives of radicalized youth, this form of discourse appears only peripherally, overshadowed by the rest of the material that constitute this government campaign. Studying the discursive specificities as well as the argumentative potential of testimonies led us to consider it a relevant form of alternative discourse to counter hatred and division. Therefore it seems to us suitable to give it a real place within the strategies deployed in order to fight the phenomenon of jihadist radicalization n(19) Restorative justice is defined as an “option for doing justice after the occurrence of an offence that is primarily oriented towards repairing the individual, relational and social harm caused by that offence” (Walgrave, 2008, p. 21).(20) The task force’s members are Robert Cario, Géraldine Blin (DFSPIP, SPIP 95), Ouisa Kies (sociologist at CADIS/EHESS and director of the organization “Dialogues Citoyens”), Alexandre Cartier (CPIP, SPIP 95) and Jérémy Bridier (legal expert at the AfVT, French association of terrorism victims). See the website: http://www.justicerestaurative.org/news/radicalisation-terrorisme-et-mesures-de-justice-restaurative/.© Cahiers de la sécurité et de la justice - IHEMI98 I DOSSIER “Stop-djihadisme”, or how to thwart radicalization: a critical examination of a plan of prevention through information – Laurène Renaut, Laura AsconeReferencesAmossy, Ruth, 2008, « Argumentation et Analyse du discours : perspectives théoriques et découpages disciplinaires », Argumentation et Analyse du discours, (1).Amossy, Ruth, 2010, La présentation de soi : Ethos et identité verbale, Paris : PUF, coll. « Interrogation philosophique ».Angenot, Marc, 1989, « Hégémonie, dissidence et contre-discours : réflexions sur les périphéries du discours social en 1889 », Études littéraires, 22(2), p. 11–24.Ascone, Laura, 2018, « La radicalisation à travers l’expression des émotions sur internet », PhD dissertation defended on November 22nd, 2018 at the Université de Cergy-Pontoise.Auboussier, Julien, 2015, « Présentation », Semen, 39. [Online: http://journals.openedition.org/semen/10463].Benbassa, Esther and Catherine Trœndlé, Rapport final de la mission d’information sur le désendoctrinement, le désembrigadement et la réinsertion des djihadistes en France et en Europe, Paris, Sénat, 2017.Benchellali, Mourad, 2006, Voyage vers l’enfer. Éditions Robert Laffont.Bouzar, Dounia and Marie Martin, 2016, « Méthode expérimentale de déradicalisation : quelles stratégies émotionnelles et cognitives ? », Pouvoirs (3), p. 83-96.Combet-Galland, Corina, 1991, « Quand le récit devient communication : Dialogue avec un recueil sur la narration », Revue de philosophie et de théologie, n°123, p. 213-220.Crettiez, Xavier and Bilel Ainine, 2017, Soldats de Dieu, paroles de djihadistes incarcérés, Editions de l’Aube, Fondation Jean-Jaurès.Del Valle, Alexandre, 2009, « Guerre des représentations et virus sémantiques », Géoéconomie, 51 (4), p. 119-145.Haddad, Gérard, 2015, Dans la main droite de Dieu : psychanalyse du fanatisme, Paris : Premier parallèle.Kasiki, Sophie and Pauline Guéna, 2016, Dans la nuit de Daesh : confession d’une repentie, Robert Laffont.Vallat, David, 2016, Terreur de jeunesse, le témoignage d’un ex-djihadiste, Paris, Calmann-Lévy.Lausberg, Heinrich, 1960, Handbuch der literarischen Rhetorik, Munich, Max Hueber.Moniquet, Claude, 2015, Néo-djihadistes : Ils sont parmi nous, qui sont-ils ? Comment les combattre ?, Jourdan.Pasoni, Laura and Catherine Lorsignol, 2016, Au cœur de Daesh avec mon fils, Paris, La Boîte à Pandore.Perelman, Chaïm and Lucie Olbrechts-Tyteca, 1988 (5th edition). Traité de l’argumentation. Bruxelles, Éditions de l’Université libre de Bruxelles.Plantin, Christian, 2011, Les bonnes raisons des émotions. Peter Lang Publishing Group.Reboul, Olivier, 1991, Introduction à la rhétorique : théorie et pratique, Paris, Presses universitaires de France.Renaut, Laurène, 2019, « Radicalisation djihadiste et discours victimaire sur les réseaux sociaux : de la victime au bourreau », Argumentation et Analyse du Discours, 23 [Online: http://journals.openedition.org/aad/3870].Renaut, Laurène and Laura Ascone, 2019, « Contre-discours au discours de haine djihadiste. De l’expression de la conflictualité à la fabrique du doute », Semen, n°47.Salas, Denis, 2018, La Foule innocente, Paris, Desclée de Brouwer.Semelin, Jacques, 2005, Purifier et détruire. Usages politiques des massacres et génocides, Paris, Seuil.Walgrave, Lode, 2008, Restorative Justice, Self-Interest and Responsible Citizenship, Cullompton, Willan Publishing.Walgrave, Lode and Denis Salas, 2015, « Le terrorisme intérieur. Un défi pour la justice restaurative », Les Cahiers de la Justice, 3(3), p. 423-438.Witte, Kim and Mike Allen, 2000, « A meta-analysis of fear appeals: Implications for effective public health campaigns », Health education & behavior, 27(5), p. 591-615.© Cahiers de la sécurité et de la justice - IHEMICahiers de la sécurité et de la justice – Hors-série DOSSIER I 99L’intelligence artificielle : nouvel outil au service de la prévention de la récidive ?Marie NICOLAS-GRÉCIANOe e recours à l’intelligence artificielle 1 comme outil permettant à l’institution judiciaire de gérer la criminalité peut sembler irréaliste, mais la réalité est bien différente, car le mouvement d’automatisation de la justice pénale est en marche. Des algorithmes 2 capables d’anticiper le comportement humain et d’évaluer la dangerosité des individus pourraient, un jour 3, être exploités par les autorités judiciaires, non pas pour prévenir la commission d’une infraction (c’est le rôle revendiqué par la police prédictive 4 avec des outils tels que Predpol), mais pour les aider à prendre des décisions de remise en liberté (durant la garde à vue ou pour une libération conditionnelle) ou à choisir la peine la plus adaptée. Dans un avenir proche, un certain nombre de décisions pourraient donc être prises grâce à des outils d’anticipation du risque ou dits prévisionnels.Précisions terminologiquesDans le langage courant, l’expression d’outils prédictifs est régulièrement employée pour désigner des logiciels d’anticipation du risque pénal. Or, le terme prédictif 5, issu du latin prae- (avant) et dictare (dire), qui signifie littéralement : « dire avant qu’un événement se produise », est impropre à l’utilisation, parce que l’exercice relève davantage de la L’intelligence artificielle pourrait s’imposer comme un outil d’avenir pour la prévention de la récidive en matière pénale. Bien connus du système américain, les algorithmes pourraient être utilisés par l’institution judiciaire pour anticiper la réitération d’infractions, mais ceci à condition de renforcer la fiabilité de ces outils et de respecter les principes fondamentaux du droit pénal.Mots clés : Prévention, Récidive, Dangerosité, Évaluation du risque pénal, Intelligence artificielle, algorithme.LMarie NICOLAS-GRÉCIANOMaître de conférences en droit privé et sciences criminelles à l’Université Clermont-Auvergne, Marie Nicolas-Gréciano est membre du Centre Michel de l’Hospital.(1) « Ensemble de théories et de techniques mises en œuvre en vue de réaliser des machines capables de simuler l’intelligence humaine », Encyclopédie Larousse.(2) « Ensemble de règles opératoires dont l’application permet de résoudre un problème énoncé au moyen d›un nombre fini d’opérations », Encyclopédie Larousse. Ce type d’algorithme existe déjà à l’étranger (COMPAS aux USA, HART au Royaume-Uni). Cf. Infra.(3) Une volonté politique de recourir, de manière maîtrisée, à l’intelligence artificielle dans la justice, s’est manifestée dans le discours de la Garde des Sceaux, le 26 février 2019, mais aucun projet de loi n’a encore été déposé. (4) Voir sur ce thème notamment : Castets-Renard, 2019 : 314-317 ; Institut d’aménagement et d’urbanisme d’Île-de-France, 2019.(5) Il s’agit en réalité d’une locution directement traduite du terme anglais « prediction ».© Cahiers de la sécurité et de la justice - IHEMI100 I DOSSIER L’intelligence artificielle : nouvel outil au service de la prévention de la récidive ? – Marie Nicolas-Grécianopseudoscience (divination) que de la science. Il serait plus pertinent de parler de prévision – notion issue des locutions latines prae- (avant) et visere (voir) – pour désigner le fait de voir un événement avant qu’il se réalise. Au sens strict, l’élaboration d’outils d’anticipation du risque pénal renvoie alors davantage à la prévision qu’à la prédiction (Garapon, 2017 : 6 ; Jean, 2019 : 947).Les atouts de l’intelligence artificielleLes outils prévisionnels présentent un certain nombre d’avantages pour la justice pénale. Ils peuvent soutenir l’institution judiciaire, plus précisément, les décideurs : c’est-à-dire les magistrats de l’ordre répressif. N’y aurait-il pas là une aide incommensurable pour le juge qui trouverait dans cet outil les réponses à des questions cruciales telles que : faut-il condamner l’individu à une peine ferme ? Le condamné va-t-il commettre de nouvelles infractions s’il est libéré de manière anticipée ? Face à ces questionnements quotidiens, nombre de magistrats pourraient être soulagés dans leur prise de décision. De plus, ces technologies pourraient contribuer à renforcer la sécurité juridique au profit du justiciable 6 (victimes et personnes mises en cause), puisque la décision ne serait plus prise par un magistrat, mais par un outil ayant examiné, de manière scientifique, des constantes et des variables objectives et extérieures. Enfin, les outils prévisionnels permettraient de renforcer l’efficacité de la répression pénale en évitant la remise en liberté d’individus dangereux. C’est donc la protection de la société qui serait renforcée. Ce sont là quelques exemples des apports de l’intelligence artificielle pour le système judiciaire 7. Conscients de ces atouts potentiels, les politiques souhaitent mettre les outils prévisionnels au service de la justice pénale 8.L’intelligence artificielle : l’avenir d’une justice 3.0 ?Toutefois, ces techniques représentent aussi un danger possible pour les droits et libertés fondamentales, en raison de résultats parfois erronés ou discriminatoires. De plus, remplacer le juge par une machine limiterait le droit d’accéder à un tribunal indépendant et impartial, tout en limitant l’individualisation de la peine 9. En dépit de ces inconvénients, l’avenir des outils prévisionnels dans les systèmes judiciaires semble tracé par les décideurs politiques, poussés par les concepteurs de ces technologies. Dès lors, la justice pourrait, à l’avenir, devenir algorithmique, automatisée ou simulée. En matière pénale plus précisément, cette technologie pourrait être utilisée pour prévenir le risque pénal entendu largement, qu’il s’agisse d’évaluer le risque de récidive ou de réitération 10 d’un suspect ou d’un condamné. Autrement dit, les outils prévisionnels seraient un nouvel indicateur 11, permettant de limiter le danger de réitération d’infractions pénales. Cette utopie, toujours recherchée dans les sociétés, serait-elle accessible grâce à l’intelligence artificielle ? Pour l’heure, l’utilisation de ces algorithmes par les autorités policières pendant l’enquête ou par le juge dans le cadre d’un procès pénal reste au stade embryonnaire en Europe. Deux interrogations doivent être soulevées : i) peut-on prévenir la réitération ? ii) doit-on la prévenir par le biais de ces outils prévisionnels ?La question de la possibilité de prévenir la réitération au moyen de l’intelligence artificielle est donc posée (I), comme celle de la pertinence du recours à ces outils dans le système pénal (II).La possibilité de prévenir le risque pénal grâce à l’intelligence artificielleLa possibilité de prévenir la réitération d’infractions revêt deux aspects qu’il convient d’envisager tour à tour : la faisabilité technique (A) et le cadre juridique (B).La faisabilité techniqueL’émergence de l’intelligence artificielle en matière pénale s’est réalisée outre-Atlantique avec la création de machines destinées à apporter une certaine dose de prédiction : si l’identification d’un auteur avant la commission d’un crime relève encore de la science-fiction, certains instruments dits d’évaluation du risque (risk assessment tools) sont présentés comme étant capables de mesurer (6) Les attentes (prévisibilité, lisibilité, intelligibilité, compréhensibilité) des citoyens en matière de justice ont été mises en lumière par le rapport annexé au projet de loi de programmation 2018-2022 et de réforme pour la justice, p. 3.(7) Automatiser la justice permet aussi de faire des économies en évitant des procédures longues et coûteuses.(8) Discours de Nicole Belloubet le 26 février 2019 à la conférence d’Helsinki sur l’intelligence artificielle. Selon la Garde des Sceaux : « l’intelligence artificielle favorisera l’analyse jurisprudentielle et constituera un outil d’aide à la décision sans pour autant priver le juge de son rôle ».(9) Cf. infra.(10) Selon l’article 132-16-7 du Code pénal, la réitération intervient lorsque l’individu « commet une nouvelle infraction qui ne répond pas aux conditions de la récidive légale ». La récidive suppose, quant à elle, la commission d’une nouvelle infraction d’un certain type et dans un certain délai, lesquels sont déterminés par le Code pénal aux articles 132-8 à 132-11.(11) Parmi d’autres : casier judiciaire, évaluation psychologique et psychiatrique.© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 101 Cahiers de la sécurité et de la justice – Hors-série les probabilités de réitération d’une infraction pénale par une personne ayant déjà exécuté une mesure privative de liberté. Techniquement, il serait possible d’évaluer le risque de récidive ou de réitération sur la base de certaines données emmagasinées par un algorithme. Pour le prouver, il suffit de tourner son regard vers les États-Unis où les technologies de l’intelligence artificielle sont à la pointe du secteur.États-Unis : une culture juridique propice aux outils prévisionnelsLes outils de prévention de la délinquance ont trouvé un terrain favorable aux États-Unis (Defferrard et Papineau, 2017 : 668) car dans les systèmes de common law, le niveau de risque de récidive (faible, moyen ou élevé) fait, depuis longtemps, partie des critères pris en compte par le juge pour déterminer la peine ou la mesure de réhabilitation à prononcer (Monahan et Skeen, 2016 : 495) 12. Ainsi, les délinquants qui présentent un risque faible de récidive sont, de préférence, condamnés à une peine privative de liberté courte ou bénéficient de mesures alternatives. En revanche, ceux dont le risque de récidive est élevé, sont condamnés à une peine ferme et de longue durée. La détermination du risque de récidive fait donc partie de l’office du juge américain. C’est pour l’aider dans cette tâche délicate que des algorithmes d’évaluation du risque pénal ont été créés et élargis avec le temps. A l’origine, ces outils étaient uniquement employés lors de la phase d’exécution des peines pour évaluer la pertinence d’une mesure de libération conditionnelle (parole) ou de suivi par les services de probation. Aujourd’hui, ils sont aussi employés au moment de la détermination de la peine (sentencing) et au cours d’une mesure de garde à vue (custody) (Kehl et al., 2017 : 9-10). Même si ces outils ne lient pas le juge, le résultat communiqué par l’algorithme exerce, à n’en pas douter, une influence sur sa décision. Les outils prévisionnels, qui ont remplacé les expertises considérées comme désuètes et inefficaces, rencontrent un franc succès aux États-Unis à tel point que le système américain d’évaluation du risque pénal repose exclusivement sur des méthodes de prédiction dites actuarielles (Morvan, 2013 : 326-330).Le perfectionnement des outils américainsLes progrès de l’intelligence artificielle ont conduit à la création d’instruments d’évaluation du risque de nouvelles générations. Le logiciel de référence aux États-Unis, appelé COMPAS (Correctional offender management profiling for alternative sanctions) 13, fonctionne grâce à un algorithme élaboré sur la base : i) des informations tenant à la conduite criminelle rassemblées depuis plusieurs années par les services de police, ii) des décisions de privation de liberté déjà prononcées et iii) des facteurs de risques liés au sexe, à l’âge, à la scolarité, à l’état civil, au statut professionnel, à la situation patrimoniale, aux antécédents judiciaires, au lieu de résidence et à sa stabilité 14. Cet algorithme s’appuie également sur des modèles conçus avec des techniques d’apprentissage automatique (machine learning), c’est-à-dire grâce à des ordinateurs capables d’apprendre tous seuls à partir de données, ceci afin de déterminer la probabilité statistique de récidive de l’individu concerné. Concrètement, l’algorithme procède à une comparaison entre les données statiques acquises et les caractéristiques d’un délinquant. Si celui-ci présente de nombreux facteurs communs avec les individus ayant déjà récidivé, son risque de réitération de l’infraction est indiqué élevé. L’évaluation réalisée par l’algorithme est par la suite communiquée au juge chargé de prendre la décision. Forte de son succès outre-Atlantique 15, la justice actuarielle a fait ses premiers pas sur le Vieux Continent.L’expérimentation britanniqueCréé par l’université de Cambridge, le logiciel HART (Harm assessment risk tool) est utilisé, à titre expérimental, depuis 2017 par la police de Durham au Royaume-Uni 16. Plusieurs étapes ont été nécessaires à la mise en place de cette technologie. D’abord, l’ensemble des archives de la police de Durham entre 2008 et 2012 a été entré dans la machine afin qu’elle prenne connaissance de l’ensemble des décisions prises par les policiers pendant cette période et des statistiques en matière de réitération des suspects. Un algorithme a été conçu sur la base de ces informations préalablement enregistrées pour évaluer le risque de réitération des suspects en les classant selon trois (12) Certains États américains ont intégré l’évaluation des risques dans les lignes directrices (sentencing guidelines), pour déterminer la peine la plus appropriée.(13) Cet algorithme, développé par une entreprise privée, doit obligatoirement être utilisé par le juge dans certains États américains. 137 questions sont posées incluant la présence d’un téléphone à la maison, la difficulté de paiement de factures, les antécédents familiaux ou encore l’histoire criminelle du prévenu. L’algorithme note la personne sur une échelle de 1 (faible risque) à 10 (haut risque). Il s’agit d’une aide à la prise de décision judiciaire, ses conclusions n’étant qu’une des variables à considérer par le juge lors de la définition de la peine.(14) Annexe I de la Charte éthique européenne de l’utilisation de l’intelligence artificielle dans les systèmes judiciaires et leur environnement (ci-dessous « Charte éthique européenne ») intitulée « Étude approfondie sur l’utilisation de l’intelligence artificielle dans les systèmes judiciaires, notamment les applications d’intelligence artificielle assurant le traitement des décisions et des données judiciaires », 3-4 décembre 2018, p. 55, § 128 et s.(15) Cet outil fait l’objet d’importantes contestations. Cf. infra.(16) Charte éthique européenne, § 125 et s.© Cahiers de la sécurité et de la justice - IHEMI102 I DOSSIER L’intelligence artificielle : nouvel outil au service de la prévention de la récidive ? – Marie Nicolas-Grécianocatégories : faible, moyen ou élevé. Il permet d’identifier une trentaine de facteurs, dont certains sont statiques et non liés à l’infraction (adresse, genre, etc.) et de les comparer avec les caractéristiques du suspect (Oswald et al., 2018 : 227-228). Employé dans le cadre de la garde à vue, ce logiciel permet d’évaluer le danger que représente le suspect susceptible de réitérer tout en aidant à la prise de décision policière (prolonger la mesure, remettre en liberté). Il fonctionne également grâce à l’apprentissage automatique, ce qui lui permet de tirer des conclusions des données et de se perfectionner sans nécessairement recevoir de nouvelles instructions. Des tests ont été réalisés, dès 2013, pour déterminer l’efficacité de cet outil. Il a ainsi été possible de constater, en observant le comportement des suspects pendant une période de deux années après la commission de l’infraction, que les prévisions de HART étaient efficaces à hauteur de 98 % en cas de risque faible et de 88 % en cas de risque élevé de récidive 17. Ces résultats élevés seraient dus à l’algorithme exploité dans le logiciel. Celui-ci aurait été élaboré sur la base d’un « compromis entre les faux positifs et les faux négatifs » 18 afin de limiter le nombre de faux négatifs. En d’autres termes, l’algorithme aurait plutôt tendance à étiqueter les suspects comme étant des personnes à risque élevé, même si le danger est en réalité moyen, voire faible, ceci pour éviter les erreurs policières (et la remise en liberté d’individus qui risquent de commettre une nouvelle infraction). Les concepteurs de cette technologie ont donc préféré rester prudents en ayant tendance à sur-qualifier les individus comme dangereux. C’est là un danger considérable, car cet outil privilégie la sécurité nationale par rapport aux droits et libertés fondamentales des individus (Oswald et al. 2018 : 231, 241). Or, aucune concession ne devrait être possible : si des technologies de ce type sont développées et exploitées en France, elles devront respecter un cadre juridique exigeant en matière de droits et libertés fondamentales. Le cadre juridiqueS’il est techniquement possible de créer des outils pour anticiper le risque pénal, il convient encore de définir le cadre juridique dans lequel ils doivent s’inscrire. Trois limites peuvent être posées, puisqu’il faut respecter : les principes du droit pénal fondamental, le droit à un procès équitable et la Charte éthique européenne d’utilisation de l’intelligence artificielle dans les systèmes judiciaires et leur environnement. Les principes du droit pénal fondamentalLa recherche de la prévention du risque pénal et de l’évaluation de la dangerosité fait craindre la résurgence de pensées déterministes 19, basées sur l’absence de libre-arbitre des individus. En effet, il serait facile d’admettre l’idée selon laquelle certains individus sont prédéterminés à commettre des infractions, ce qui justifierait une peine rigoureuse accompagnée de mesures de sûreté 20. Le déterminisme latent se ferait alors au détriment du principe d’individualisation de la peine (Saleilles, 2001 ; Mayaud, 2018 ; Dreyer 2016), largement admis depuis 1945 dans la plupart des systèmes judiciaires européens. En France, c’est la doctrine de la « défense sociale nouvelle » développée par Marc Ancel qui a posé les jalons du droit pénitentiaire moderne (Ancel, 1957 ; Lazerges, 2005). L’approche déterministe et punitive a ainsi été remplacée par un système tourné vers la réadaptation sociale du délinquant. L’idée originelle consiste à dire qu’en (ré-)adaptant l’individu – lequel dispose de son libre-arbitre –, les conditions de la délinquance disparaitront, permettant in fine de prévenir durablement la commission d’infractions. Pourtant, si des algorithmes prévisionnels doivent être développés, les garanties nécessaires doivent être instaurées pour éviter qu’ils ne soient exploités dans une perspective déterministe et que les utilisateurs (magistrats) n’occultent l’objectif de réinsertion (Cassuto, 2017). Par ailleurs, en vertu du principe d’individualisation de la peine 21, le juge doit se fonder sur des éléments objectifs (formation, emploi, prise en charge médico-sociale régulière etc.) pour trouver la peine la plus adaptée au condamné. Ce principe de valeur constitutionnelle 22 doit, lui aussi, être respecté par les outils prévisionnels (Brigant, 2018 : 237). Par conséquent, le recours aux algorithmes de prévision du risque pénal ne doit pas altérer ni écarter les objectifs d’individualisation et de réinsertion de la peine.Le droit à un procès équitableLes outils prévisionnels seront aussi soumis au droit à un procès équitable inscrit à l’article 6§1 de la Convention (17) Annexe I de la Charte éthique européenne, § 126.(18) Les faux positifs qualifient de dangereux des individus alors qu’ils n’ont pas réitéré ou récidivé au moment du constat, tandis que les faux négatifs sont des individus catégorisés comme non dangereux alors qu’ils ont réitéré ou récidivé.(19) Le déterminisme fonde l’école positiviste italienne, représentée par Cesare Lombroso (1835-1909), Enrico Ferri (1856-1929) et Raffaele Garofalo (1851-1934).(20) La mesure de sûreté est une sanction pénale de nature préventive décidée par un juge lorsqu’un individu présente un caractère dangereux. (21) Art. 132-24 du code pénal.(22) Cons. const., déc. N° 2005-520, 22 juillet 2005, Rec. P. 118, cons. 3.© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 103 Cahiers de la sécurité et de la justice – Hors-série européenne de sauvegarde des droits de l’homme et des libertés fondamentales (CESDH) (Meneceur, 2018 ; Ferrié, 2018). Plus précisément, ils devront respecter le droit d’accès à un tribunal, garantie fondamentale énoncée par la Cour européenne des droits de l’homme (CEDH) dans son célèbre arrêt Golder c. Royaume-Uni du 21 février 1975 23 et reconnue par le Conseil constitutionnel24. Selon cette garantie, les justiciables doivent avoir la possibilité de saisir le juge pénal afin qu’il se prononce sur le bien-fondé de l’accusation portée contre eux. Or, il est à craindre que les technologies d’évaluation du risque pénal n’aient pour effet de restreindre, voire de supprimer purement et simplement l’accès au juge. En effet, l’impact de l’évaluation de l’algorithme sur la décision pénale peut être plus ou moins fort. Au stade le moins avancé, le résultat donné par la machine sert seulement d’indicateur dans l’évaluation de la situation et le juge peut le prendre en compte ou l’écarter. Au stade le plus évolué, l’outil informatique se substitue au juge dans certains domaines : c’est ce que l’on appelle la justice automatisée. C’est le logiciel et non le professionnel, le robot et non l’humain, qui rend des décisions ayant des effets de droit. Tel pourrait être le cas pour certains contentieux de masse (par exemple, les infractions au code de la route) qui conduisent déjà à des réponses quasi-systématisées, à l’aide de barèmes (Sayn et al., 2019 ; Gerry-Vernières : 2019). Si cette dernière voie semble encore lointaine25, les outils prévisionnels auront sans doute un rôle de plus en plus prégnant en matière pénale. Il faudra donc veiller scrupuleusement aux garanties du procès équitable, en laissant toujours au juge le soin de prendre la décision et en permettant au justiciable de la contester par les voies de recours.La Charte éthique en matière d’intelligence artificielleLes outils d’évaluation du risque pénal sont également tenus par les exigences de la Charte éthique européenne d’utilisation de l’intelligence artificielle dans les systèmes judiciaires adoptée les 3 et 4 décembre 2018 par la Commission européenne pour l’efficacité de la justice (CEPEJ) (Meneceur, 2019 : 552-558 ; Barbaro, 2019 : 11-14). Cette Charte – relevant du droit souple (soft law)26 – identifie cinq principes : i) le respect des droits fondamentaux, ii) la non-discrimination, iii) la qualité et la sécurité, iv) la neutralité, la transparence et l’intégrité intellectuelle et v) la maîtrise par l’utilisateur. Ainsi, les algorithmes prévisionnels devront être conçus et mis en œuvre de manière à être compatibles avec les cinq principes énoncés. Toute discrimination entre individus ou groupes d’individus devra être bannie, afin d’éviter des résultats biaisés et porteurs d’atteintes aux droits et libertés fondamentales. De plus, l’environnement technique devra être sécurisé pour protéger les données sensibles des personnes concernées par ces outils. Quant aux méthodes d’évaluation et de traitement des données, elles devront être compréhensibles et accessibles pour tous, afin de préserver la confiance des justiciables dans le système judiciaire. Enfin, la Charte recommande que l’utilisateur de l’outil – en l’occurrence l’autorité judiciaire ou policière – soit suffisamment informé et qu’il reste, en tout état de cause, maître de ses choix. Remplacer l’humain par la machine doit donc être exclu au regard des exigences éthiques.Ce n’est qu’en respectant ces principes éthiques et juridiques que les outils prévisionnels pourront être introduits dans notre système, lequel présentera les garanties d’une justice équitable.La pertinence du recours aux outils de prévention Face à l’engouement des systèmes étrangers et des dirigeants pour ces outils prévisionnels, se pose la question de ce qui est de l’ordre du souhaitable : doit-on recourir, voire systématiser l’utilisation des outils de prévention du risque pénal ? A cette question, il est possible de répondre de deux manières : la négative semble s’imposer au regard du manque patent de fiabilité de ces outils (A), mais la positive peut l’emporter à condition de réserver une place spécifique à ces technologies (B).Des instruments défectueuxDeux séries de critiques peuvent être avancées pour refuser, aujourd’hui, le recours aux outils de prévention dans le procès pénal : d’une part, les outils de prévention souffrent de différents biais, d’autre part, leurs résultats sont parfois erronés.(23) CEDH, Golder c. Royaume-Uni, 21 février 1975, § 35, GA, n° 23, Berger, n° 40 (voir aussi Desportes et Lazerges-Cousquer, 2015 : 241).(24) Cons. const., déc. n° 96-373, 9 avril 1996, Polynésie française ; déc. N° 2002-532, 19 janvier 2006, Lutte contre le terrorisme.(25) Notamment parce que l’article 10 de la loi n° 78-17 du 6 janvier 1978 relative à l’informatique, aux fichiers et aux libertés l’interdit expressément. Cela figure aussi à l’article 22 du Règlement (UE) 2016/679 du 27 avril 2016 relatif à la protection des personnes physiques à l’égard du traitement des données à caractère personnel et à la libre circulation de ces données (RGPD).(26) La Charte éthique n’a pas force obligatoire. Elle adresse des recommandations aux acteurs publics et privés chargés de la conception et du développement d’outils et de services d’intelligence artificielle.© Cahiers de la sécurité et de la justice - IHEMI104 I DOSSIER L’intelligence artificielle : nouvel outil au service de la prévention de la récidive ? – Marie Nicolas-GrécianoDes outils biaisésLes outils de prévention sont présentés par leurs concepteurs comme objectifs, contrairement à la subjectivité du juge qu’ils combattent 27. Fondées uniquement sur un traitement statistique de données relatives aux infractions recensées, insensibles aux sentiments et aux préjugés, les prédictions de l’intelligence artificielle se veulent plus fiables que celles de l’être humain. Cependant, ces outils souffrent d’erreurs méthodologiques tenant à la fois à l’approche retenue et aux données introduites (Chouldechova, 2016 ; Angwin et Larson, 2016).Le premier type de biais intervient dans l’approche retenue par les algorithmes : ceux-ci sont élaborés à partir de données statistiques globales relatives à un groupe d’individus (les délinquants) pour des infractions déjà commises. Or, les magistrats doivent juger une individualité caractérisée par son propre degré de culpabilité. Dès lors, la façon dont le crime est appréhendé, puis traité, ne peut se réduire à une approche collective, mais doit suivre une approche individuelle. Cette erreur méthodologique inscrite dans le mode de raisonnement des algorithmes peut rendre ces outils non conformes à certains principes cardinaux du droit pénal, à savoir l’égalité28 qui implique aussi de traiter les personnes différemment lorsqu’elles sont dans des situations différentes, l’individualisation de la peine et sa proportionnalité. Ces outils représentent alors un réel danger pour la justice individualisée et respectueuse de l’égalité. De surcroît, la machine est incapable de reproduire un raisonnement juridique (étude de causalités et non de corrélations dont certaines n’ont pas de sens) ou humain29. De même, la décision humaine peut parfois se fonder sur des valeurs et des considérations sociales qui ne sont pas prises en compte par la machine. À titre d’illustration, un juge pourrait décider d’ordonner la remise en liberté d’une femme délinquante tout en connaissant son risque de récidive élevé car il en appellerait à une valeur supérieure : la nécessité pour elle d’assurer son rôle de mère de famille. C’est l’application de la jurisprudence du « bon juge Magnaud »30. Dans ces hypothèses, l’algorithme déterminerait le risque de réitération de l’infraction sans pouvoir opérer une hiérarchie entre impératifs et valeurs. Le second type d’erreur se manifeste dans les données introduites dans ces logiciels, celles-ci pouvant reproduire des discriminations sociales et raciales préexistantes. Ces données sont constituées, à l’origine, à partir de certaines décisions des pouvoirs publics, tels que les lieux de patrouille des policiers, le profil des personnes qui vont être contrôlées et arrêtées ou encore le type de sanctions qui vont être appliquées. Or, ces données peuvent être entachées de biais sociaux ou raciaux présents dans des choix politiques. En effet, si certaines communautés sont surreprésentées dans les statistiques, parce qu’elles font, plus que d’autres, l’objet de mesures de police, les antécédents criminels des individus appartenant à ces communautés seront en conséquence plus nombreux. L’algorithme aura donc tendance à les considérer comme dangereuses. Par conséquent, ces outils de prévention du risque pénal peuvent avoir pour effet de reproduire des disparités raciales et socio-économiques. Ce défaut a été mis en lumière en 2016 par l’O.N.G. Pro-Publica, qui s’est penchée sur le logiciel américain COMPAS (Angwin et al., 2016). Cet instrument d’évaluation du risque utilisé dans plusieurs États fédérés présente les populations afro-américaines comme ayant un taux de risque de récidive deux fois plus élevé que celui des autres populations, dans les deux années suivant l’application de la peine 31. Les outils mis en place rendent ainsi des résultats discriminatoires et déterministes au détriment de certaines communautés (afro-américaine) ou classes d’âge (les jeunes) qui sont davantage pénalisées que d’autres (Chouldechova, 2016). La prétendue impartialité et l’objectivité d’une justice administrée par les algorithmes se heurte en quelque sorte à une « contamination » des statistiques et des données à l’apparence neutralisée, par l’utilisation de méthodes mathématiques et statistiques (Vigneau, 2018, 2019). Le risque d’une justice dévoyée est d’autant plus grand que ces algorithmes ont pour effet de naturaliser et d’ainsi légitimer des injustices, voire de les amplifier à travers une inégalité de traitement répétée et renforcée par les tribunaux. C’est dire que le fonctionnement même de ces outils de prévention est contestable.Des résultats erronésLes promoteurs de l’intelligence artificielle soutiennent que cet outil est plus exact dans ses déterminations qu’un être humain (en l’occurrence, le juge) 32. Il est vrai que les machines ont une capacité plus importante de traiter et d’établir des liens entre des masses de données, et ceci, (27) Annexe I de la Charte éthique européenne, p. 58, § 135.(28) Art. 1er de la Constitution. Pour une étude en la matière, voir : Dechenaud, 2008.(29) Pour l’opinion d’un magistrat sur ce point, voir Dufour, 2019 : 4-6.(30) Le juge Magnaud a prononcé, le 4 mars 1898, la relaxe de Louise Ménard qui avait volé du pain pour donner à manger à ses enfants.(31) D’autres algorithmes ont été élaborés sur la base d’observations critiques exprimées par la doctrine. Ces outils se fondent sur des variables plus restreintes, plus directement liées au crime commis et moins à la race, le genre ou la condition socioéconomique. Tel est le cas du Public safety assessment tool utilisé dans 30 juridictions américaines.© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 105 Cahiers de la sécurité et de la justice – Hors-série de manière plus performante qu’une personne. Toutefois, l’erreur n’est pas uniquement humaine. Des fautes sont en effet fréquemment commises par les outils prévisionnels, lorsqu’ils classent un individu dans une mauvaise catégorie de risques. Deux types d’erreurs sont possibles : d’une part, les faux négatifs, consistant à libérer un individu en raison de son faible risque pénal, alors qu’en réalité il commettra une nouvelle infraction. D’autre part, les faux positifs, qui conduisent à maintenir en prison une personne considérée comme dangereuse, alors qu’elle n’aurait commis aucun tort si elle avait été libérée. L’étude menée par l’O.N.G. Pro-Publica a mis en lumière certaines de ces erreurs. Ainsi, pour le même type d’infraction – en l’occurrence, un vol – l’algorithme a considéré qu’une femme noire était à haut risque de récidive, tandis qu’un homme blanc était à risque faible (alors que celui-ci avait un casier judiciaire plus important). Dans les deux années suivant l’évaluation du risque, la femme n’a pas récidivé alors que l’homme a commis un vol avec effraction. Dans l’ensemble, il est apparu que l’algorithme avait tendance à surestimer le risque de récidive des personnes de couleur et à le sous-estimer pour les personnes de type caucasien, conformément aux biais humains introduits dans l’algorithme. Les machines peuvent donc, à l’instar des êtres humains, commettre des erreurs. Mais le réel danger réside dans le fait de présenter ces outils comme parfaits. Les vertus attribuées aux outils prédictifs doivent par conséquent être largement relativisées.La place à réserver aux outils de préventionLe tableau dressé n’a pas à être exclusivement noirci, puisque l’intelligence artificielle présente des qualités indéniables, comme l’importante capacité de traitement d’informations et de données par rapport à celles de l’être humain. Dans un contexte de réduction budgétaire, les outils prévisionnels peuvent donc apporter des éléments d’informations supplémentaires au juge. C’est pourquoi le recours à l’évaluation du risque pénal peut être envisagé, à condition de lui réserver une place préalablement déterminée par la loi : celle d’un indicateur pour un juge libre dans son appréciation.Un indicateur capable de donner une estimation du risque pénalPour concilier les avantages de l’intelligence artificielle avec les exigences du procès équitable et de l’individualisation des peines, les logiciels d’analyse du risque pénal doivent être considérés comme des instruments d’aide à la prise de décision policière et judiciaire. Les informations obtenues par le biais de ces outils doivent nécessairement être mises en perspective avec les autres éléments rassemblés par l’institution judiciaire. En somme, le résultat de l’évaluation du risque pénal est une information parmi d’autres (expertise psychologique, passé judiciaire, contexte familial, social et économique) et la décision ne peut pas reposer exclusivement sur celui-ci. De même, il ne peut y avoir de hiérarchie préétablie entre la prédiction et les autres éléments à disposition du juge, car si une valeur supérieure était conférée aux outils prévisionnels, au vu des erreurs possibles, des atteintes aux libertés fondamentales seraient inévitablement commises. Pour être efficace, cette garantie tenant à l’absence de hiérarchisation entre les informations obtenues doit figurer dans la loi et ne souffrir d’aucune exception.Préserver le pouvoir d’appréciation du jugeEn outre, le pouvoir souverain d’appréciation du juge doit être préservé : celui-ci doit pouvoir s’écarter, proprio motu, de la prédiction formulée, notamment lorsque les circonstances de l’infraction et celles tenant à l’individu l’emportent 33. C’est uniquement de cette manière que les garanties d’indépendance du juge 34, du procès équitable et d’individualisation de la peine seront respectées. Autrement dit, les prévisions énoncées ne doivent pas se substituer aux décisions du juge. Cependant, il est légitime de s’interroger sur le risque de dépendance du juge vis-à-vis des résultats avancés par l’algorithme (comme c’est déjà le cas pour les expertises). En effet, il existe une tendance naturelle du juge à s’appuyer sur les conclusions des expertises, surtout lorsqu’elles lui apportent des réponses dans un domaine hors de son champ de compétence, comme la médecine, la balistique, l’évaluation du risque de récidive etc. (Marx, 1964 : 193 ; Vérin, 1980 : 1022 ; Pradel, 1975 : 67). Cette dépendance du juge à l’égard des outils de prévision risque d’être d’autant plus forte que leurs résultats sont présentés comme fiables, par ses concepteurs, généralement des entreprises privées (32) Il s’agit essentiellement des acteurs économiques (entreprises privées) qui mettent en avant les avantages et les promesses des algorithmes et de l’intelligence artificielle.(33) Comme cela a été le cas, par exemple, dans l’affaire du bon juge Magnaud. Cf. supra.(34) Art. 6§1 de la CESDH. Le juge ne doit pas être tenu de suivre le résultat du logiciel prévisionnel. © Cahiers de la sécurité et de la justice - IHEMI106 I DOSSIER L’intelligence artificielle : nouvel outil au service de la prévention de la récidive ? – Marie Nicolas-GrécianoRéférencesAncel, Marc, 1954, la Défense sociale nouvelle, Éditions Cujas, 183 p.Angwin, Julia, Jeff Larson, Surya Mattu et Lauren Kirchner, 2016, « Machine Bias », dans ProPublica, 23 mai 2016.Angwin, Julia et Jeff Larson, 2016, « Bias in criminal risk scores is mathematically inevitable, researchers say », dans Propublica, 30 décembre 2016.Barbaro, Clementina, 2019, « Les travaux de la CEPEJ en matière d’utilisation de l’Intelligence Artificielle dans les systèmes judiciaires. A propos de la Charte éthique européenne d’utilisation de l’IA dans les systèmes judiciaires et dans leur environnement », dans JCP G, p. 11-14.Brigant, Jean-Marie, 2018, « Les risques accentués d’une justice pénale prédictive », dans Archives de philosophie du droit, p. 237-251.Cadiet, Loïc, 2017, « L’open data des décisions de justice », Mission d’étude et de préfiguration sur l’ouverture au public des décisions de justice, novembre. Accessible sur le site de la Cour de cassation : http://www.justice.gouv.fr/publication/open_data_rapport.pdf.Cassuto, Thomas, 2017, « La justice à l’épreuve de sa prédictibilité », dans AJ Pénal, p. 334-337Castets-Renard, Céline, 2019, « L’IA en pratique : la police prédictive aux États-Unis », dans Dalloz IP/IT, p. 314-317dont l’opacité caractérise l’élaboration des algorithmes. Sur ce point, les outils d’évaluation du risque doivent être transparents, avec un code source permettant aux juges, aux avocats de la défense et aux justiciables de comprendre et d’évaluer l’algorithme 35. Il en va du respect de la présomption d’innocence, de l’exercice des droits de la défense et de l’égalité des armes 36. En effet, c’est en connaissant profondément ces outils que l’intéressé sera en mesure de discuter la validité scientifique, les composantes de l’algorithme et les conclusions hâtives.Repenser l’utilisation de l’intelligence artificiellePartant d’un constat en demi-teinte, il semble nécessaire d’envisager une autre utilisation de l’intelligence artificielle en matière pénale. Ainsi, ce n’est pas la place, mais la fonction même de ces outils qui devrait être repensée. Au lieu d’utiliser l’intelligence artificielle comme un outil préventif, déterministe et punitif, les algorithmes pourraient être élaborés pour servir l’individualisation de la peine. En rassemblant les éléments objectifs de personnalité (formation, emploi, prise en charge médico-sociale régulière), ainsi que les informations détenues par les institutions et organismes publics (Pôle emploi, Caisse d’allocations familiales, fisc, etc.), l’intelligence artificielle pourrait centraliser des quantités importantes de données à caractère économique, social et sanitaire et les traiter dans un temps record (Cassuto, 2017 : 334). A l’heure des données ouvertes (open data) des décisions de justice 37 et sur la base du fonctionnement de l’analyse des mégadonnées (big data analytics), les conclusions tirées de ces nombreuses données brutes pourraient être communiquées rapidement au juge qui doit parfois statuer dans des délais très courts, notamment dans le cadre de procédures accélérées 38. Le juge aurait alors les moyens de prononcer la peine la mieux adaptée à la situation de l’individu et à ses ressources. L’intelligence artificielle serait au service de l’humain et contribuerait à la valorisation du travail de l’ensemble de l’institution judiciaire. Toutefois, cet idéal de justice combinant les nouvelles technologies et les droits de l’homme est conditionné à la bonne qualité des données, à la loyauté dans l’utilisation de ces outils, à la transparence du traitement d’algorithmes certifiés et, en tout état de cause, à la préservation de la liberté d’appréciation du juge n(35) Si la transparence doit reposer sur l’engagement et l’éthique de ces acteurs privés, elle doit également être imposée par la loi. Pour l’heure, le code des algorithmes est protégé par le secret des affaires.(36) Ces garanties sont inscrites aux articles 6§1 et 6§3 de la CESDH.(37) Sur ce thème, voir Cadiet, 2017.(38) « Il s’agit de procédures plus courtes que les procédures classiques » (Couvrat, 1994 : 699). Il s’agit, par exemple, de la CRPC, de la comparution immédiate ou encore de l’ordonnance pénale.© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 107 Cahiers de la sécurité et de la justice – Hors-série Couvrat, Pierre, 1994, « Les procédures sommaires en matière pénale », dans RIDC, p. 699Dechenaud, David, 2008, l’Égalité en matière pénale, L.G.D.J., 2008, 620 p.Defferrard, Fabrice et Christelle Papineau, 2017, « Le pouvoir de jurisdictio des algorithmes aux États-Unis : entre fantasme et réalité jurisprudentielle », dans Dalloz IP/IT, p. 668-671Desportes, Frédéric, Lazerges-Cousquer Laurent, 2015, Traité de procédure pénale, Economica, 2480 p.Desportes, Frédéric et Francis Le Gunehec, 2009, Droit pénal général, Economica, 1248 p.Dreyer, Emmanuel, 2016, Droit pénal général, LexisNexis, 2016, 1394 p.Dufour, Olivier, 2019, « Comment les nouvelles technologies ambitionnent de révolutionner la fonction de juger », LPA, 15 avril, p. 4-6.Ferrié, Scarlett-May, 2018, « Les algorithmes à l’épreuve du droit au procès équitable », dans Procédures, avril 2018, p. 4-9.Garapon, Antoine, 2017, « Les enjeux de la justice prédictive », dans JCP G, n° 1, 1er janvier, p. 47-52.Institut d’aménagement et d’urbanisme d’Île-de-France, 2019, « La police prédictive. Enjeux soulevés par l’usage des algorithmes prédictifs en matière de sécurité publique », avril, 39 p.Kehl, Danielle, Priscilla Guo et Samuel Kessler, 2017, « Algorithms in the criminal justice system : assessing the use of risk assessments in sentencing », dans Responsive Communities, juillet 2017, p. 9-10.Lazerges, Christine, 2005, « La défense sociale nouvelle a 50 ans. Actualité de la pensée de Marc Ancel », dans RSC, n° 1, p. 165-170.Marx, Yvonne, 1964, « Observations sur le problème de l’expertise pénale », dans in Mélanges Hugueney, Dalloz, p. 193 et s.Mayaud, Yves, 2018, Droit pénal général, PUF, 784 p.Meceneur, Yannick, 2019, « Les enseignements des éthiques européennes de l’intelligence artificielle », dans JCP G, 25 mars 2019, p. 552-558.Meceneur, Yannick, 2018, « Les systèmes judiciaires européens à l’épreuve du développement de l’intelligence artificielle », dans Revue pratique de la prospective et de l’innovation, dossier n°7, octobre.Monahan, John et Jennifer Skeen, 2016, « Risk Assessment in Criminal Sentencing », dans Annu. Rev. Clin. Psychol., n°153.Morvan, Patrick, 2019, Criminologie, Lexis Nexis, 460 p.Oswald, Marion, Jamie Grace, Sheena Urwin et Geoffrey Barnes, 2018, « Algorithmic risk assessment policing models: lessons from the Durham HART model and « experimental » proportionality », dans Information & Communications Technology Law, p. 223-250.Jean, Jean-Paul, 2019, « À l’ère du numérique, ce que le criminel pourrait apprendre au civil en l’état », dans Recueil Dalloz, p. 947-955.Pin, Xavier, 2019, Droit pénal général, Dalloz, 592 p.Pradel, Jean, 1975, « Les rôles respectifs du juge et du technicien dans l’administration de la preuve pénale », dans in Colloque IEJ, PUF, p. 67-82.Saleilles, Robert, 2001, l’Individualisation de la peine, réédition de la troisième édition, suivi de : L’individualisation de la peine cent ans après Saleilles (plusieurs contributions), Eres, 288 p.Vérin, Jacques, 1980, « L’expertise dans le procès pénal », dans RSC, p. 1022 et s.Vigneau, Vincent, 2018, « Le passé ne manque pas d’avenir. Libres propos d’un juge sur la justice prédictive », dans Recueil Dalloz, p. 1095-1103.© Cahiers de la sécurité et de la justice - IHEMI108 I DOSSIERArtificial Intelligence: a new tool to prevent reoffending? – Marie Nicolas-GrécianoArtificial Intelligence: a new tool to prevent reoffending?Marie NICOLAS-GRÉCIANOesorting to AI 1 as a tool enabling the judiciary institution to manage crime may seem unrealistic, but the reality is quite different as the automatization of criminal justice is underway. Algorithms 2 capable of anticipating human behavior and evaluating the dangerousness of individuals could, some day, 3 be used by judicial authorities, not to prevent crime (it is the role pertaining to the prediction police 4 with tools such as Predpol), but to help them make decisions to release a person who is in custody, or to put somebody on parole), or to choose the most suitable sentence. In the near future, a number of decisions could therefore be made thanks to risk anticipation or so-called prediction tools.Lexicon precisionsIn common speech, the phrase prediction tools is regularly used to refer to criminal risk anticipation softwares. The term prediction 5 which comes from the latin prae- (before) and dictare (tell), which literally means: “tell before an event occurs,” is improper because this practice is more related to a pseudoscience (divination) than to science. It would be more relevant to talk about prevision – a notion originating from prae- (before) and visere (to see) – to mean seeing an event before it happens. Strictly speaking, the creation of criminal risk anticipation tools falls more within the field of prevision than in that of predicting (Garapon, 2017: 6; Jean, 2019: 947).AI could prevail as the tool of the future to prevent recidivism in criminal matters. Well-known in the US system, algorithms could be used by judicial institutions to anticipate reoffending, provided the reliability of these tools is strengthened and the fundamental principles of criminal law are respected. Keywords: prevention, recidivism, dangerousness, criminal risk assessment, AI, algorithm.RMarie NICOLAS-GRÉCIANOA private law and criminal science Lecturer at Université Clermont-Auvergne, Marie Nicolas-Gréciano is a member of Centre Michel de l’Hospital.(1) «Ensemble de théories et de techniques mises en œuvre en vue de réaliser des machines capables de simuler l’intelligence humaine», Larousse Encyclopedia. (2) «Ensemble de règles opératoires dont l’application permet de résoudre un problème énoncé au moyen d’un nombre fini d’opérations», Larousse Encyclopedia. This type of algorithm already exists abroad (COMPAS in the USA, HART in the UK). Cf. Infra.(3) A political will to resort, in a controlled manner, to AI in justice, appeared in the Justice Secretary’s speech on February 26th, 2019, but no bill so far has been registered. (4) On this them, see notably: Castets-Renard, 2019: 314-317; Institut d’aménagement et d’urbanisme d’Île-de-France, 2019.(5) It is actually a term directly translated from the English term “prediction”.© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 109 Cahiers de la sécurité et de la justice – Hors-série The assets of Artificial IntelligencePrediction tools offer a number of advantages to criminal justice. They can support the judicial institution, more precisely, decision-makers: i.e. the magistrates of the repressive order. Wouldn’t they be immeasurably helpful to the judge who would find in these tools the responses to crucial questions such as: is it necessary to sentence the defendant to unconditional imprisonment? Is the convict going to commit new crimes if released early?Faced with this daily questioning, a number of magistrates could be relieved in their decision making. Moreover, these technologies could contribute to reinforce judicial security for the victims and for the persons answerable to the law, 6 since the decision would no longer be made by a magistrate but by a tool having examined, scientifically, objective and external constants and variables. Finally, prediction tools could reinforce the efficacy of penal repression by avoiding the release of dangerous individuals. Therefore, society’s protection would be reinforced. These are a few examples of the benefits of AI for the judicial system. 7 Well aware of these potential benefits, lawmakers wish to put these prediction tools at the service of criminal justice. 8Artificial Intelligence: the future of justice 3.0?However, these techniques also represent a possible danger for the rights and fundamental liberties due to occasionally erroneous or discriminatory results. Furthermore, replacing the judge by a device would limit the right to access to an independent and impartial court, while limiting the individualization of sentences 9 as well. Despite these disadvantages, the future of prediction tools in judiciary systems seems obvious to lawmakers, influenced by the developers of these technologies. As a result, justice might, in the future, become algorithmic, automated or simulated. In criminal matters, more specifically, this technology could be used to prevent the criminal risk in a wider sense, whether it be to assess the risk of reoffending or repeat offense 10 of a suspect or convict. In other words, prediction tools would be a new indicator, 11 enabling to limit the danger of repeated criminal offenses. Would this utopia, always sought by society, be accessible thanks to AI? At the moment, the use of these algorithms by law enforcement authorities during the investigation, or by the judge during a criminal trial, remains embryonic in Europe. Two questions must be addressed: i) can recidivism be prevented? ii) must it be prevented through these prediction tools? Therefore, the question asked is that of whether preventing recidivism through AI is possible (I), as well as whether resorting to these tools is relevant in the criminal system (II).The possibility of preventing criminal risks thanks to AIThe possibility of preventing repeated offenses takes on two aspects which must be examined one by one: technical feasibility (A) and judicial framework (B).Technical feasibilityThe emergence of AI in criminal matters has appeared in the US with the creation of devices meant to offer a certain dose of prediction: although the identification of the perpetrator before the crime has been committed remains only possible in science fiction, certain so-called risk assessment tools are presented as being capable of assessing the probabilities for a person, who has been deprived of liberty, to repeat a criminal offense.Technically, it would be possible to assess the risk of recidivism or repeat on the basis of certain data stored by an algorithm. To find evidence of this, just take a look at what is happening in the US, where AI technologies are leaders in this field.The US: a judicial culture favorable to prediction toolsPrediction tools to prevent crime have found a favorable environment in the US (Defferrard and Papineau, 2017: 668) as, in common law systems, the level of recidivism risk (low, medium or high) has been, for a long time, part of the criteria taken into consideration by the judge, to (6) Citizens expectations in terms of justice (foreseeability, legibility, intelligibility, clarity) have been highlighted by the report annexed to the programming 2018-2022 and justice reform bill, p.3. (7) Automating justice also enables to cut costs by avoiding long and costly procedures.(8) Nicole Belloubet’s speech on February 26th, 2019 at the Helsinki conference on artificial intelligence. According to the Justice Secretary: “Artificial intelligence will enhance jurisprudential analysis and constitute a tool helpful in decision-making without depriving the judge of his role.”(9) Re. infra.(10) Under article 132-16-7 of the Penal Code, reiteration occurs when the individual “commits a new common law offense which does not meet the criteria of legal recidivism.” Recidivism implies committing a new common law offense of a certain type and within a certain timeframe, which are determined by the Penal Code in articles 132-8 to 132-11.(11) Among others: criminal record, psychological and psychiatric assessment.© Cahiers de la sécurité et de la justice - IHEMI110 I DOSSIER Artificial Intelligence: a new tool to prevent reoffending? – Marie Nicolas-Grécianodetermine the sentence or the rehabilitation measure to pronounce (Monahan and Skeen, 2016: 495). 12 Thus, offenders who present a low risk of recidivism are, preferably, sentenced to a short time imprisonment sentence or to an alternative measure. However, those, for whom the risk is high, are sentenced to long term unconditional imprisonment. Determining the risk of reoffending is therefore part of the missions of US judges. To help them with this difficult task, risk assessment algorithms have been created and progressively enlarged. At first, these tools were only used during the execution of sentences to assess the relevance of parole or of follow up by probation services. Today, they are also used at the moment of sentencing and during custody (Kehl et al., 2017: 9-10). Even if these tools are not binding for judges, the result communicated by the algorithm certainly influences their decisions. The prediction tools, which have replaced the opinions of experts, deemed obsolete and inefficient, are extremely popular in the US to the extent that the US system of criminal risk assessment only relies on prediction methods called actuarial (Morvan, 2013: 326-330).The perfecting of US toolsThe progress of AI has led to the creation of new generation risk assessment tools. The software used in the US, called COMPAS (Correctional offender management profiling for alternative sanctions), 13 works thanks to an algorithm elaborated on the basis of: i) information relative to criminal conduct gathered by the police for several years, ii) liberty deprivation decisions previously made, iii) risk factors linked to age, gender, schooling/educational background, civil status, professional status, patrimonial status, criminal record, place of residence and its stability. 14 This algorithm also relies on models developed with machine learning techniques, i.e. thanks to computers capable of learning by themselves from data, in order to determine the statistical probability of the individual’s recidivism. Concretely, the algorithm compares the statistical data acquired and the characteristics of offenders. If an offender presents many common factors with individuals known for repeating offenses, the recidivism risk is deemed high. The assessment made by the algorithm is then communicated to the judge in charge of making the decision. Owing to its success in the US, 15 actuarial justice has begun to be used on the old continent.The British experimentCreated by Cambridge University, the software HART (Harm assessment risk tool) has been experimentally used since 2017 by the Durham police in the UK. 16 Several steps has been necessary to the implementation of this technology. First, all the Durham police records have been uploaded for the machine to access all the decisions made by the police during that period and about the reoffending suspects statistics. An algorithm has been developed on the basis of this information previously recorded to assess the suspects’ recidivism risk by classifying them in three categories: low, medium or high. It enables to identify about thirty factors, some of which are static and not linked to the offense (address, gender, etc.), and compare them with the suspect’s characteristics (Oswald et al., 2018: 227-228). Used during custody, this software enables the police to assess the suspect’s recidivism risk, while helping the police’s decision (prolong the custody, release the suspect). It also works through automatic learning, which helps it draw new conclusions from data and perfect itself without necessarily receiving new instructions. Tests have been made since 2013, to determine the efficiency of this tool. It has thus been possible to notice, by observing the suspects’ behaviour for two years after the offense, that HART’s previsions were 98% efficient for the low risk cases and 88% efficient for the high recidivism risk. 17 These high results might be due to the algorithm used in the software. It might have been developed on the basis of a “compromise between the false positive and the false negative,” 18 in order to limit the number of false negatives. In other words, the algorithm might tend to label suspects as being high risk, even if the danger is in fact medium, or even low, to avoid police errors (and releasing individuals who risk reoffending). The developers of this technology have prefered to be cautious by tending to overqualify (12) Some US states have integrated risk assessment in the sentencing guidelines, to determine the most appropriate sentence.(13) This algorithm, developed by a private company, must by law be used by the judge in certain US states. 137 questions are asked, including the presence of a phone at home, the difficulty in paying bills, family background or the criminal history of the accused. The algorithm rates the person on a scale of 1 (low risk) to 10 (high risk). The idea is to help judicial decision-making, its conclusions being only one of the variables to be taken into account by the judge when defining the sentence. (14) Appendix I of the European Ethical Charter on the use of AI in judicial systems and their environment (below “European Ethical Charter”) entitled “In-depth study on the use of AI in judicial systems, notably AI applications processing judicial decisions and data,” December 3-4th, 2018, p. 55, § 128 et s.(15) This tool has been seriously objected to. Cf. infra.(16) European Ethical Charter, § 125 et s.(17) Appendix I of the European Ethical Charter, § 126.(18) False positives qualify as dangerous individuals although they have not reiterated or repeated offenses at the time of the assessment, whereas false negatives are individuals categorized as not dangerous although they have reiterated or repeated offenses.© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 111 Cahiers de la sécurité et de la justice – Hors-série individuals as dangerous. This is a considerable danger, as this tool favors national security over the rights and fundamental liberties of individuals (Oswald et al. 2018: 231, 241). However, no compromise should be allowed. If these technologies are developed and used in France, they will have to respect a legal framework demanding in terms of rights and fundamental liberties.The legal frameworkIf it is technically possible to create tools to anticipate criminal risk, it is indispensable to define the legal framework in which they can be used. Three limits must be fixed, as it is necessary to respect: the fundamental principles of criminal law, the right to a fair trial and the European Ethical Charter on the use of AI in judicial systems and their environment.Fundamental principles of criminal lawThe search for the prevention of criminal risk and dangerousness assessment alerts us on the resurgence of determinist thinking, 19 based on the absence of free will in individuals. Indeed, it would be easy to admit the idea that certain individuals are predetermined to commit offenses, which would justify a rigorous sentence reinforced by security measures. 20 Underlying determinism would then be favored over the principle of the individualization of sentences (Saleilles, 2001; Mayaud, 2018; Dreyer 2016), largely admitted since 1945 in most European judicial systems. In France, the doctrine of the new social defense by Marc Ancel, has introduced the new penitentiary law (Ancel, 1957; Lazergues, 2005). The determinist and punitive approach was thus replaced by a system aiming to socially readapt the offender. The original idea consisted in saying that by re-adapting the individual – who has got free will – the conditions of crime will disappear, durably and finally leading to the eradication of offenses and crimes. However, if prediction algorithms must be developed, necessary guarantees must be installed to prevent them from being exploited for the sake of determinism and make sure magistrates do not omit the objective of social reintegration (Cassuto, 2017). Besides, under the principle of the individualization of sentences, 21 the judge must base his judgement on objective elements (professional training, employment, regular health and social support etc.) to find the most suitable sentence for the individual. This principle, of constitutional value, 22 must also be respected by the prediction tools (Brigant, 2018: 237). Therefore, resorting to criminal risk algorithms must not alter or eliminate the individualization and social reintegration objectives of the sentence.The right to a fair trialThe prediction tools will also be submitted to the right to a fair trial mentioned in article 6 alinea 1 of the European convention for the safeguard of human rights and fundamental liberties (Meneceur, 2018; Ferrié, 2018). More precisely, they will have to respect the right of access to a court, a fundamental guarantee mentioned by the European Court of human rights (ECHR) in its famous ruling Golder vs. United Kingdom of February 21st, 1975 23 and acknowledged by the Constitutional Council. 24 Under this guarantee, persons who must answer to the law must be able to refer to criminal judges for them to rule over the legitimacy of the accusation held against them. Now, there is some concern about the fact that criminal risk assessment technologies might result in restraining or even merely suppressing access to a judge. Indeed, the impact of the algorithm assessment on the penal decision, may vary in intensity. At the earliest stage, the result provided by the machine only serves as an indicator in the assessment of the situation and the judge can take it into account or not. At the most advanced stage, the IT tool substitutes for the judge in certain areas: that’s what we call automated justice. It is the software and not the professional, the robot and not the human which makes legal decisions. Such could be the case for some mass litigations (for example driving offenses), which already lead to quasi-systematized answers, via sentence charts or scales (Sayn et al., 2019; Gerry-Vernières: 2019). While this alternative seems still faraway, 25 prediction tools will probably play a growing role in criminal matters. It will therefore be necessary to scrupulously ensure the guarantees of fair trial, by leaving to the judge the responsibility to make the decision and (19) Determinism founded the positive Italian school, represented by Cesare Lombroso (1835-1909), Enrico Ferri (1856-1929) et Raffaele Garofalo (1851-1934).(20) A security measure is a preemptive penal sanction decided by a judge when an individual seems dangerous. (21) Art. 132-24 of the Penal Code.(22) Constitutional Council, Decree N° 2005-520, July 22nd, 2005, Rec. P. 118, cons. 3.(23) ECHR, Golder c. Royaume-Uni, February 21st, 1975, § 35, GA, n° 23, Berger, n° 40 (see also Desportes and Lazerges-Cousquer, 2015 241).(24) Constitutional Council, Decree n° 96-373, April 9th, 1996, French Polynesia; Decree N° 2002-532, January 19th, 2006, Counter terrorism.(25) Notably because article 10 of the law of January 6th, 1978 relative to IT, files and freedom forbids it expressly. it also appears in article 22 of the Regulation (EU) 2016/679 of April 2016 relative to the protection of physical persons with regards to personal data processing and the free circulation of this data (RGPD).© Cahiers de la sécurité et de la justice - IHEMI112 I DOSSIER Artificial Intelligence: a new tool to prevent reoffending? – Marie Nicolas-Grécianoby enabling the person being tried the right to protest against it through the right of appeal.The Ethical Charter on the use of AIThe criminal risk assessment tools must also meet the requirements of the European Ethical Charter on the use of AI in judicial systems adopted on December 3rd and 4th, 2018 by the European Commission for the Efficiency of Justice (CEPEJ) (Meneceur, 2019: 552-558; Barbaro, 2019: 11-14). This charter – pertaining to soft law 26 – identifies five principles: i) respect for fundamental rights, ii) non-discrimination, iii) quality and security, iv) transparency, impartiality and fairness and v) “under user control”. Thus, prediction algorithms will have to be designed and implemented so as to be compatible with the five principles stated. Any discrimination between individuals or groups of individuals must be banished in order to avoid biased results which violate rights and fundamental liberties. Furthermore, the technical environment will have to be secured to protect sensitive data of the persons concerned by these tools. As for the assessment and data processing methods, they will have to be understandable by and accessible to all, to preserve trust in the judiciary system. Finally, the Charter recommends that the tool user – in this case the court or the police – be sufficiently informed and that they remain, in any case, in control of their choices. Replacing man by a machine must consequently be excluded with regards to ethical requirements. It’s only by respecting these ethical and legal principles that prediction tools will be able to be introduced in our system, which will offer fair justice guarantees. The relevance of resorting to prevention toolsIn the face of the growing popularity of foreign systems and the attraction of lawmakers to these prediction tools, is it a good idea to resort to or even generalize the use of criminal risk assessment tools? To this question, the answer is twofold: a negative answer seems to prevail due to the undeniable lack of reliability of these tools (A), but a positive answer can prevail provided a specific place is reserved to these technologies (B).Flawed toolsTwo types of criticisms can be put forward to refuse, today, to resort to prevention tools in a criminal lawsuit: first, prevention tools are biased in various ways. Furthermore, their results are sometimes erroneous.Biased toolsPrevention tools are presented by their developers as objective, contrary to the judge’s subjectivity, against which they fight.27 Only based on statistical data processing relative to identified and listed offenses, insensitive to feelings and prejudice, AI predictions intend to be more reliable than those made by man. However, these tools suffer from methodological errors in terms of the chosen approach and the data entered (Chouldechova, 2016; Angwin and Larson, 2016).The first type of bias occurs in the chosen approach by algorithms: they are elaborated from overall statistical data relative to a group of individuals (offenders) for already committed offenses. Now, magistrates must judge an individuality characterized by its own degree of guilt. Therefore, the way the crime is apprehended, then processed, cannot be reduced to a collective approach, but must follow an individual approach. This methodological mistake embedded in the reasoning mode of algorithms, can result in tools non-compliant with certain key principles of criminal law, i.e. equality,28 which also implies treating people differently when they are in different situations, the individualization of sentences and their proportionality. These tools present a real danger for individualized justice and for the respect of equality by justice. Furthermore, a machine is incapable of reproducing legal or human29 reasoning (causality study, not correlation study, some of which make no sense). Likewise, human decision can sometimes be based on social values and considerations which are not taken into account by a machine. For instance, judges might decide to release a convicted woman while being aware of her high recidivism risk as they might refer to a higher value: the need for the woman to take care of her children. It’s the application of the “good judge Magnaud” precedent.30 In these hypotheses, the algorithm would determine the offense repeating risk without being able to operate a hierarchy between imperatives and values.(26) The Ethical Charter is not binding. It addresses recommendations to the public and private players in charge of designing and developing AI tools and services.(27) Appendix I of the European Ethical Charter, p. 58, § 135.(28) Art. 1er of the Constitution. For a study on the matter, see Dechenaud, 2008..(29) For the opinion of a judge on this point, see Dufour, 2019: 4-6.© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 113 Cahiers de la sécurité et de la justice – Hors-série The second type of error is manifest in the data entered in the software, which can reflect pre-existing social and racial discriminations. This data is constituted, originally, by government decisions such as where the police patrol, the profile of controlled or arrested people or even the type of sanctions applied. Now, this data can be biased by social or racial prejudice present in political choices. Indeed, if some communities are overrepresented in statistics, it is because these are more targeted than others by police measures, therefore the criminal records of individuals belonging to these communities will be more abundant. The algorithm will therefore tend to consider these communities to be dangerous. Consequently, these criminal risk assessment tools may result in reproducing racial and socio-economic inequalities. This flaw was highlighted in 2016 by ProPublica, an NGO that studied the US software COMPAS (Angwin et al., 2016). This risk assessment tool used in several US states presents African-American populations as having a recidivism rate twice as high as that of other populations, in the two years following the application of the sentence.31 The implemented tools thus give discriminatory and determinist results at the expense of some communities (African-American) or age groups (youth) who are more criminalized than others (Chouldechova, 2016). The alleged impartiality and objectivity of a justice administered by algorithms is somehow faced with a “contamination” of statistics and data, apparently neutralized by the use of mathematical and statistical methods (Vigneau, 2018, 2019). The risk of misguided justice is all the greater as these algorithms result in naturalizing and thus legitimizing injustices or even amplifying them through recurrent unequal treatment reinforced by courts. In other words, the very functioning of these prevention tools is objectionable.Erroneous resultsThe proponents of AI claim that this tool is more accurate in its determinations than a human being (here, a judge).32 It is true that machines have a greater capacity for dealing with and establishing links between masses of data, and in a more performing way than a person. Nevertheless, erring is not only human. Mistakes are indeed frequently made by prediction tools when they classify an individual in the wrong risk category. Two types of errors are possible: on the one hand false negatives, which consist in releasing an individual due to low criminal risk, when in reality they will commit another offense. On the other hand false positives, which lead to keeping in jail a person considered dangerous, who would not have committed any offense if they had been released. The study led by NGO ProPublica highlighted some of these mistakes. So, for the same type of offense, here a theft, the algorithm considered that a Black woman had a high risk of recidivism, whereas a white man had a low risk (although he had a worse criminal record). In the years following the risk assessment, the woman did not commit a new offense whereas the man committed robbery. Overall, it appeared that the algorithm tended to overestimate the risk for coloured people and underestimate it for caucasien people, in line with the human prejudices introduced into the algorithm. Machines can, therefore, like human beings, make mistakes. But the real danger lies in presenting these tools as perfect. The virtues attributed to these prediction tools must therefore largely be nuanced and put into perspective.The place to give these toolsThe bottom line should not exclusively be negative since AI presents undeniable qualities such as the capacity to process information and data compared to that of a human being. In a low budget context, prediction tools can therefore bring supplementary elements of information to the judge. That’s why resorting to criminal risk assessment can be considered on condition that it is used within a framework beforehand determined by the law: as an indicator for a judge to freely appreciate its value.An indicator capable of estimating criminal riskTo reconcile the advantages of AI with the requirements of fair trial and individualization of sentences, risk assessment software must be considered helpful tools for decision-making by police or justice. Information obtained through these tools must necessarily be put into perspective with the other elements gathered by the judiciary institution. Therefore, the result of the criminal risk assessment is an element of information among others (psychological expertise, criminal record, family background, socio-economic background) and the decision cannot rely only on it. Likewise, there can be no pre-established hierarchy between the prediction and other elements at the disposal of the judge, as, if a superior value was granted to prediction tools, considering the possibility of mistakes, violations of fundamental liberties (30) Judge Magnaud pronounced, on March 4th, 1898, the discharge of Louise Ménard who had stolen bread to feed her children. (31) Other algorithms have been elaborated on the basis of critical observations expressed by the doctrine. These tools are based on more restricted variables, more directly related to the crime and less to the race, gender or socioeconomic conditions. Such is the case with the Public safety assessment tool used in 30 US jurisdictions.(32) It is essentially economic players (private companies) who put forward the upsides and promises of AI.© Cahiers de la sécurité et de la justice - IHEMI114 I DOSSIER Artificial Intelligence: a new tool to prevent reoffending? – Marie Nicolas-Grécianowould be inevitable. To be efficient, the guarantee of the absence of hierarchy between elements of information must appear in the law and suffer zero exception.Preserving the judge’s power of appreciationMoreover, the sovereign power of appreciation of the judges must be preserved: they must be able to move away, on their own initiative (proprio motu), from the formulated prediction, notably when the circumstances of the offense and those of the individual prevail.33 It’s only in this way that the guarantees of independence of the judge,34 of fair trial and individualization of sentences will be respected. In other words, the predictions stated must not replace the judge’s decisions. However, it is legitimate to question the dependence risk of the judge towards the results of the algorithm (it is already the case for expertise results). Indeed, there is a natural tendency for judges to rely on the conclusions of experts, mostly if they bring responses in areas in which they have no expertise, such as medicine, ballistics, recidivism risk assessment, etc. (Marx, 1964: 193; Vérin, 1980: 1022; Pradel, 1975: 67). This dependency of the judge regarding predicting tools may be all the stronger as their results are presented as reliable by their developers, usually private companies whose opacity characterizes the elaboration of these algorithms. In this respect, risk assessment tools must be transparent, with a source code enabling judges, defense lawyers and persons who must answer to the law to understand and evaluate the algorithm.35 The respect of the presumption of innocence, the exercise of defense rights and equality of arms36 depends on it. Indeed, it is by thoroughly knowing these tools that the person concerned will be able to discuss their scientific value, the algorithm components and rushed conclusions.Rethinking the use of Artificial IntelligentTaking into consideration its relatively successful performance, it is necessary to consider another use of AI in criminal matters. Thus, it is not the place but the very function of these tools which must be reevaluated. Instead of using AI as a preventive, determinist and punitive tool, algorithms could be elaborated to serve the individualization of sentences.By gathering objective aspects of a person (professional training, employment, regular health and social support), as well as the information held by public institutions and bodies (job centers, tax authorities, family allowance fund, etc.), AI could centralize a great amount of economic, social, sanitary/health data and could process it in no time (Cassuto, 2017: 334). At a time of open data of justice decisions37 and on the basis of the functioning of big data analytics, the conclusions reached by this abundant raw data could quickly be communicated to the judge who must sometimes make decisions in a very short timeframe, notably in the case of accelerated procedures.38 The judge would then have the means to pronounce the sentence most adapted to the situation of the individuals and their financial resources. AI would be at the service of the human and contribute to enhance the work of the whole judicial institution. However, this ideal of justice combining new technologies and human rights lies in the quality of data, loyalty in the use of these tools, transparency of certified algorithms and, obviously, the preservation of the freedom of appreciation of the judge n(33) As was the case for good judge Magnaud. Cf. supra.(34) Art. 6§1 of the ECHR. The judge does not have to follow the result of the prediction software.(35) If transparency must rely on the commitment and ethics of these private players, it must also be imposed by the law. For the moment, algorithm code is protected by business secrecy.(36) These guarantees appear in articles 6§1 et 6§3 of the ECHR.(37) On this subject, see Cadiet, 2017.(38) “These are shorter procedures than classic ones” (Couvrat, 1994: 699). It is, for instance, immediate trial with admission of guilt, or a criminal order.ReferencesAncel, Marc, 1954, La Défense sociale nouvelle, Éditions Cujas, 183 pp.Angwin, Julia, Jeff Larson, Surya Mattu and Lauren Kirchner, 2016, “Machine Bias”, in ProPublica, May 13th, 2016.Angwin, Julia and Jeff Larson, 2016, “Bias in criminal risk scores is mathematically inevitable, researchers say”, in ProPublica, December 30th, 2016.Barbaro, Clementina, 2019, Les travaux de la CEPEJ en matière d’utilisation de l’Intelligence Artificielle dans les systèmes judiciaires. A propos de la Charte éthique européenne d’utilisation de l’IA dans les systèmes judiciaires et dans leur environnement, in JCP G, pp. 11-14.Brigant, Jean-Marie, 2018, « Les risques accentués d’une justice pénale prédictive », in Archives de philosophie du droit, pp. 237-251.© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 115 Cahiers de la sécurité et de la justice – Hors-série Cadiet, Loïc, 2017, « L’open data des décisions de justice », Mission d’étude et de préfiguration sur l’ouverture au public des décisions de justice, November. Accessible on the Cour de cassation website: http://www.justice.gouv.fr/publication/open_data_rapport.pdf.Cassuto, Thomas, 2017, « La justice à l’épreuve de sa prédictibilité », in AJ Pénal, pp. 334-337Castets-Renard, Céline, 2019, « L’IA en pratique: la police prédictive aux États-Unis », in Dalloz IP/IT, pp. 314-317Couvrat, Pierre, 1994, « Les procédures sommaires en matière pénale », in RIDC, p. 699Dechenaud, David, 2008, l’Égalité en matière pénale, L.G.D.J., 2008, 620 p.Defferrard, Fabrice and Christelle Papineau, 2017, « Le pouvoir de jurisdictio des algorithmes aux États-Unis: entre fantasme et réalité jurisprudentielle », in Dalloz IP/IT, pp. 668-671Desportes, Frédéric, Lazerges-Cousquer Laurent, 2015, Traité de procédure pénale, Economica, 2480 pp.Desportes, Frédéric and Francis Le Gunehec, 2009, Droit pénal général, Economica, 1248 p.Dreyer, Emmanuel, 2016, Droit pénal général, LexisNexis, 2016, 1394 pp.Dufour, Olivier, 2019, « Comment les nouvelles technologies ambitionnent de révolutionner la fonction de juger », LPA, April 15th, pp. 4-6.Ferrié, Scarlett-May, 2018, « Les algorithmes à l’épreuve du droit au procès équitable », in Procédures, April, pp. 4-9.Garapon, Antoine, 2017, « Les enjeux de la justice prédictive », in JCP G, n° 1, January 1st, pp. 47-52.Institut d’aménagement et d’urbanisme d’Île-de-France, 2019, « La police prédictive. Enjeux soulevés par l’usage des algorithmes prédictifs en matière de sécurité publique », April, 39 p.Kehl, Danielle, Priscilla Guo and Samuel Kessler, 2017, “Algorithms in the criminal justice system: assessing the use of risk assessments in sentencing”, in Responsive Communities, July, pp. 9-10.Lazerges, Christine, 2005, « La défense sociale nouvelle a 50 ans. Actualité de la pensée de Marc Ancel », in RSC, No 1, pp. 165-170.Marx, Yvonne, 1964, « Observations sur le problème de l’expertise pénale », in Mélanges Hugueney, Dalloz, p. 193 and following pages.Mayaud, Yves, 2018, Droit pénal général, PUF, 784 pp.Meceneur, Yannick, 2019, « Les enseignements des éthiques européennes de l’intelligence artificielle », in JCP G, March 25th, 2019, pp. 552-558.Meceneur, Yannick, 2018, « Les systèmes judiciaires européens à l’épreuve du développement de l’intelligence artificielle », in Revue pratique de la prospective et de l’innovation, dossier n°7, October.Monahan, John and Jennifer Skeen, 2016, “Risk Assessment in Criminal Sentencing”, in Annu. Rev. Clin. Psychol., n°153.Morvan, Patrick, 2019, Criminologie, LexisNexis, 460 pp.Oswald, Marion, Jamie Grace, Sheena Urwin and Geoffrey Barnes, 2018, “Algorithmic risk assessment policing models: lessons from the Durham HART model and ‘experimental’ proportionality”, in Information & Communications Technology Law, pp. 223-250.Jean, Jean-Paul, 2019, « À l’ère du numérique, ce que le criminel pourrait apprendre au civil en l’état », in Recueil Dalloz, pp. 947-955.Pin, Xavier, 2019, Droit pénal général, Dalloz, 592 pp.Pradel, Jean, 1975, « Les rôles respectifs du juge et du technicien dans l’administration de la preuve pénale », in Colloque IEJ, PUF, pp. 67-82.Saleilles, Robert, 2001, L’Individualisation de la peine, réédition de la troisième édition, suivi de: L’individualisation de la peine cent ans après Saleilles (plusieurs contributions), Eres, 288 pp.Vérin, Jacques, 1980, « L’expertise dans le procès pénal », in RSC, p. 1022 and following pagesVigneau, Vincent, 2018, « Le passé ne manque pas d’avenir. Libres propos d’un juge sur la justice prédictive », in Recueil Dalloz, pp. 1095-1103.© Cahiers de la sécurité et de la justice - IHEMI116 I VARIALe football amateur, « terrain » ou « terreau » de la radicalisation religieuse ? L’exemple de la ligue des Hauts-de-France (2015-2018) – Olivier ChovauxLe football amateur, « terrain » ou « terreau » de la radicalisation religieuse ? L’exemple de la ligue des Hauts-de-France (2015-2018)Olivier CHOVAUXAborder la question de la radicalisation dans le football amateur revient autant à interroger un angle mort de la recherche historiographique que tenter de mettre à distance une question gouvernée par l’empire des émotions. Précédé par un nécessaire travail de contextualisation, cet article livre les résultats d’une enquête conduite sur le territoire de la Ligue des Hauts-de-France de football, entre 2015 et 2018. Elle met en évidence le caractère infinitésimal du phénomène, compte tenu de la méthodologie et du terrain retenus.Mots-clés : football, violences, radicalisation.border la question de la radicalisation dans les clubs de football revient à interroger un véritable angle mort de la recherche historique, et confronte d’emblée l’historien à nombre de difficultés méthodologiques. Si le concept de radicalisation semble aujourd’hui stabilisé pour les pouvoirs publics 1, étudier ce processus complexe dans le champ sportif et dans celui du football amateur suppose d’appliquer une démarche empirique permettant à la fois de le qualifier, le localiser et le quantifier, en l’inscrivant dans un temps immédiat (2015-2018).Déjà privilégié pour nombre de travaux portant sur les violences et incivilités (Nuytens, 2003) ainsi que sur l’ethnicité sportive (Cho-vaux, 2013), le territoire de la ligue des Hauts-de-France a été ici retenu : la connaissance indigène de la cartographie des clubs et un corpus d’archives institutionnelles accessible et a priori exploitable sur ces objets justifient ce choix. Il paraît en revanche difficile de systématiser le « quadrille métho-dologique » (décrire, catégoriser, contextualiser et comprendre) pourtant éprouvé à propos des violences sportives (Chauvaud, 2010), face à un objet particuliè-rement volatil et politiquement sensible, dont il faut d’emblée rappeler qu’il se déploie également dans d’autres structures éduca-tives. Dans le cas du football, les contours de la radicalisation sont avant tout religieux et relèveraient d’une sorte de prédestination liée à l’identité de certains clubs. Qua-lifiés de « communautaires » par les instances, ils constitueraient de facto un terreau fertile au com-munautarisme. Cette analogie mécanique n’étant pas sans rap-peler celle également opérée entre Olivier CHOVAUXProfesseur des universités en histoire contemporaine à l’université d’Artois. Co-fondateur de l’Atelier SHERPAS (composante de l’URePSSS, EA 7369), ses recherches portent sur l’histoire du football, et plus récemment des violences sportives et leurs régulations. Il travaille actuellement sur l’histoire des arbitres et de l’arbitrage en France, de la fin du xixe siècle à nos jours. (1) « La radicalisation se définit comme un processus par lequel une personne (ou un groupe de personnes) conteste l’ordre institué en adoptant une forme extrême d’ac-tion. Celle-ci est liée à une idéologie qui peut être politique, sociale ou religieuse » (Ministère de l’Enseignement supérieur, 2019. Voir aussi Khosrokhavar, 2014).A© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 117 Cahiers de la sécurité et de la justice – Hors-série « clubs de quartiers » et « clubs à risque » à propos des violences et incivilités commises sur les terrains, dans les tribunes, ou à leur périphérie.L’objet de cette étude consiste donc à poser la question de la radicalisation religieuse des clubs de football amateur, en la situant dans un contexte et un territoire particuliers. La présentation des résultats d’enquêtes préalables portant sur d’autres formes de radicalités sportives (les violences et incivilités) permettent de comprendre d’une part, les difficultés méthodologiques liées à une comptabilisation stricte des phénomènes, et d’autre part, la forme d’omerta institutionnelle qui peut être celle des instances face à ces derniers. Considérer que les clubs de football puissent être des foyers potentiels de radicalisation met en effet à mal la doxa généralement véhiculée par les dirigeants du monde sportif : constamment réifiées, les vertus intégratrices et éducatives du sport s’accommoderaient bien mal de pareilles hypothèses et signeraient là un constat d’échec.Pour autant, certains éléments de contexte peuvent expliquer que le monde du football amateur puisse constituer un terreau prédisposé à l’expression de déviances, fussent-elles statistiquement peu significatives, comme le confirme la série d’enquêtes conduites dans les Hauts-de-France. La présence sur ce territoire de clubs dits communautaires constituant a priori un autre facteur aggravant en matière de radicalisation religieuse. Sur ce dernier point, seule une « mise en perspective temporelle » 2 de ces phénomènes permettrait de le ramener à de justes proportions.Des éléments de contexte prédisposant le football aux déviances ? Des violences et incivilités sportives peu significativesC’est sans doute au cours des premières années des Trente Glorieuses que la fonction intégratrice du football, en particulier pour les populations immigrées, devient plus visible, sans qu’elle soit l’apanage des seules associations sportives. Dans la France du Nord, si les Polonais se situent alors dans un entre-deux, oscillant entre maintien de leurs traditions culturelles et volonté d’en effacer les caractéristiques, les immigrés d’Afrique du Nord, compte tenu du contexte de la décolonisation, seront contraints de jouer sur un autre registre : celui de l’invisibilité (Noiriel, 2016). Il n’est donc pas surprenant que ces derniers soient l’objet de phénomènes de stigmatisation dans les années 1970 : les prodromes de la crise économique, la montée du Front national, les comptes non soldés de la « fracture coloniale » (Bancel et Blanchard, 2005) changent le paradigme de la question migratoire et l’inscrivent dans une temporalité qui ne s’est pas encore refermée aujourd’hui. Dans le Nord-Pas-de-Calais, l’arrivée des ressortissants d’Afrique du Nord (Marocains, Algériens et Tunisiens) et d’Europe méditerranéenne (Portugais et Espagnols) pourra s’observer dans la création de clubs dont la géographie se calque sur les concentrations urbaines et industrielles : en 1990, la métropole lilloise (Lille-Roubaix-Tourcoing) compte une quinzaine de clubs de football dont la dénomination contient une référence explicite au pays d’origine (« Algériens de Tourcoing », « Portugais de Roubaix », « Mahorais de Lille »), tandis que d’autres affichent d’autres intentions, telles l’Association musulmane d’animation et de loisirs de Grande-Synthe, ou l’Association sportive et socio-culturelle franco-maghrébine de Douai, dans le district Escaut. À la fin des années 1990, sur les 1000 clubs affiliés à la Ligue du Nord-Pas-de-Calais, moins de 6% affichent dans leur appellation une référence ethnique ou communautaire. On assiste alors à une sorte de retournement de tendance qui voit se modifier l’identité et le regard porté sur ces clubs. Ce souci de l’entre-soi, observé dans l’entre-deux-guerres, persiste et demeure le fondement des sociabilités sportives mais l’affirmation identitaire émerge alors en divers points du territoire : clubs turcs en Alsace, portugais dans le Bordelais, ou encore arméniens dans la Drôme en sont les exemples les plus connus. Leur implantation et leur sociologie les désignent rapidement comme des clubs de quartier ou des clubs de banlieue (Vieille-Marchiset et Coignet, 2015), dont le rôle intégrateur est alors souligné dans le cadre des politiques de la ville alors naissantes. Ce glissement de sens apparaît à un moment où la question des violences et des incivilités devient prégnante dans le football amateur, sans qu’elle soit toutefois quantitativement significative.En 2006, l’installation par la Fédération française de football (F.F.F.) de son Observatoire des comportements autorise une première série d’enquêtes quantitatives permettant de « recenser et analyser les incivilités en tout genre commises sur les terrains de football, pour mieux les combattre, par la prévention ou la répression » 3. L’analyse des feuilles de match des rencontres officielles, gérées par les ligues et les districts, conduit à catégoriser les violences (2) Comme le soulignait Alfred Wahl (cité par Mignon, 1995) à propos des violences sportives. (3) Assemblée Fédérale du 23 juin 2007, Archives de la F.F.F. (extrait) © Cahiers de la sécurité et de la justice - IHEMI118 I DOSSIER Le football amateur, « terrain » ou « terreau » de la radicalisation religieuse ? L’exemple de la ligue des Hauts-de-France (2015-2018) – Olivier Chovauxet incivilités, à partir de la nomenclature fédérale. Ainsi, pour la saison 2008-2009, sur les 709 514 matchs recensés au niveau national, seuls 11 338 avaient fait l’objet d’un incident consécutif à une « violence hostile » et 454 avaient été prématurément arrêtés, soit respectivement, 1,59% et 0,06% des rencontres. L’essentiel des incivilités (69,9%) était le fait des catégories seniors (Chovaux, 2012).Les « incivilités » sont constituées par des gestes obscènes ou crachats.Les violences « instrumentales » sont commises par les joueurs dans le jeu, constituent des infractions légères, généralement sanctionnées d’un simple avertissement.Les violences « hostiles » regroupent des comportements en dehors du jeu et plus répréhensibles : violences verbales (propos injurieux, menaces ou intimidations verbales ou physiques), violences physiques (bousculade volontaire, tentative de coups, brutalités ou coups). 4Pour la période 2006-2009, les études menées dans les trois ligues du Septentrion (Nord-Pas-de-Calais, Picardie, Champagne-Ardenne) aboutissent à des chiffres sensiblement identiques, tant en ce qui concerne le niveau des violences et incivilités que leur distribution.Ce détour préalable par la question des violences indique à quel point le football amateur est finalement un univers pacifié, autant que peut l’être d’ailleurs l’institution scolaire (Nuytens, 2016). La part résiduelle des comportements déviants ici observée fait écho à la théorie du procès de civilisation de Norbert Elias, parfois décriée (Deluermoz, 2010), mais toujours pertinente. Les écarts à la norme demeurent en somme peu significatifs. Dès lors, on voit mal comment le terrain du football amateur, peu sensible à la question des violences hostiles (se situant en dehors du jeu), pourrait constituer un terreau autrement fertile pour celle de la radicalisation religieuse. L’amalgame parfois opéré entre clubs « communautaires » (dont on a vu qu’ils s’inscrivent dans un long processus de construction lié aux différentes vagues d’immigration) et clubs « communautaristes », peut expliquer que le football soit considéré (parfois par ses propres instances) comme particulièrement prédisposé à un processus de radicalisation religieuse, au regard de la sociologie et de l’ancrage territorial de ces mêmes clubs. Pourtant, l’enquête qui suit tend à montrer qu’aucune corrélation sérieuse ne peut précisément être établie entre identité ethnique et récurrence des comportements déviants.Des clubs communautaires peu exposés aux violences mais stigmatisésL’analyse exhaustive des procès-verbaux de la commission régionale d’appel de la ligue du Nord-Pas-de-Calais de football pour la période 2006-2009 (Chovaux, 2013), permet de mesurer une responsabilité plus que relative, au plan statistique, de ces 31 clubs qualifiés de « communautaires » au regard de leur seule dénomination et de leur ancrage territorial. Ultime instance disciplinaire dans la hiérarchie des ressorts, cette commission examine les cas les plus graves. S’agissant des violences hostiles, sur les 142 dossiers disciplinaires examinés (les 108 dossiers de nature juridique ont été écartés) 5, 29 d’entre eux les concernent, soit près du quart des contentieux. (4) Selon les statuts et règlements de la F.F.F., édition 2008-2009, p.193 et suiv. Sur la distinction entre violences « instrumentales » et « hos-tiles » voir Pfister, 1994.(5) On peut toutefois relever que 5% d’entre eux concernent les clubs dits communautaires, ce qui semble indiquer chez leurs dirigeants une faible culture du contentieux juridique, souvent liée à une méconnaissance des voies de recours.Incidents recensés selon les ligues et leurs districts (% rapporté au nombre de matchs enregistrés)Données F.F.F.Champagne-ArdenneNord-Pas-de-CalaisPicardieLigueDistrictLigueDistrictLigueDistrict2006-2007***5,682,47**2007-20081,762,192,695,132,505,141,472008-20091,651,873,781,181,997,851,82Moyennes1,692,033,233,992,326,491,64(*) données non disponibles© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 119 Cahiers de la sécurité et de la justice – Hors-série Reste que leur faible nombre (6% des clubs affiliés) et le phénomène de récurrence parfois observé pour certains d’entre eux (Valenciennes Dutemple, U.S. Antillais, U.S. Portugais de Roubaix-Tourcoing) conduit parfois les commissions de première instance à appliquer de manière différenciée les barèmes disciplinaires, en se montrant plus sévères à leur endroit qu’elles ne peuvent l’être vis-à-vis de clubs considérés comme normaux. Il y aurait fort à dire sur cette « justice au faciès » (Jobard, 2005) donnant parfois l’impression aux contrevenants d’être traités de manière inéquitable, et provoquant chez les clubs une méfiance, voire une défiance à l’endroit des instances et de l’institution sportive elle-même. L’ethnicité sportive est alors considérée comme une ressource, les clubs installant des comportements d’autoprotection, et jouant de la stigmatisation dont ils s’estiment victimes pour euphémiser la nature des infractions commises, notamment en cas de violences hostiles.Victimes de « discrimination négative » (Castel, 2006), ces « clubs à risque » devraient plutôt être qualifiés de clubs « dans le risque » : la fragilité de l’encadrement éducatif, la précarité et l’insuffisance des équipements sportifs mis à leur disposition, l’environnement socio-économique ainsi que le niveau d’engagement des équipes dans les compétitions représentent autant d’invariants à prendre en compte, si l’on veut précisément mettre à distance tout étiquetage et stigmatisation. On voit donc ici combien la question de l’ethnicité sportive peut être à double tranchant : elle entretient d’un côté le mythe d’une intégration par le sport (la victoire de l’équipe de France « Black, Blanc Beur » lors de la finale de la Coupe du monde de 1998 en constitue l’acmé) autant qu’elle ferait, de l’autre, le lit de comportements transgressifs extrémistes qui remettent en cause les valeurs fondatrices, non seulement du sport lui-même (principe de justice et d’équité, idéal humaniste, etc.), mais également du cadre politique et institutionnel dans lequel il se déploie (principe de laïcité, valeurs de la République, concept de démocratie, etc.) L’adoption de postures, attitudes ou comportements liés à la radicalisation religieuse constituerait alors une double forme de transgression : ils s’inscriraient dans un espace social qui s’est historiquement construit autour de valeurs antinomiques aux options extrémistes. Ils utiliseraient le fait sportif à des fins de propagande plus ou moins explicite pour combattre et in fine détruire les fondements démocratiques et valeurs de régimes politiques considérés comme ennemis.Les terrains du football amateur, terreau de la radicalisation religieuse ?Prévenir la radicalisation religieuse ?Comme pour les violences et les incivilités observées dans les Hauts-de-France, l’historien va devoir faire preuve d’ingéniosité afin de trouver un corpus archivistique exploitable qu’il pourrait éventuellement rapprocher des premières statistiques officielles, livrées par le F.S.P.R.T. (Fichier des signalés pour la prévention de la radicalisation à caractère terroriste, créé par décret en mars 2015) : considéré par les pouvoirs publics comme le véritable baromètre de la radicalisation, il agrège les données recueillies par les services de l’État (police, gendarmerie, éducation secondaire, etc.) ainsi que les remontées d’informations issues du numéro gratuit mis en place par le ministère de l’Intérieur en avril 2014. Collectés à l’échelon départemental, les chiffres communiqués pour l’année 2016 font état de 295 personnes signalées dans le Pas-de-Calais, et 502 dans le Nord, pour 11 820 recensements au niveau national, soit respectivement 2,4% et 4,2% des chiffres nationaux. La région se situe toutefois dans le peloton de tête des territoires les plus concernés (Île-de-France, Rhône-Alpes, Provence-Alpes-Côte-d’Azur), qui sont également les zones les plus densément peuplées. D’un point de vue plus qualitatif, les personnes signalées sont des femmes (27%), des convertis (35%) et des mineurs (17%). Si les profils en haut du spectre sont suivis par les services de police selon leur degré de radicalité (Direction générale de la sécurité intérieure, Service central du renseignement territorial, gendarmerie), la majorité des personnes signalées bénéficient d’un accompagnement par d’autres services de l’État (Protection judiciaire de la jeunesse, services sociaux, etc.) Depuis octobre 2017, dans l’académie de Lille, les établissements scolaires ainsi que les universités sont invités à signaler les personnes en situation de radicalisation, dans le cadre du plan national de prévention, destiné à en repérer au plus tôt les signes.Tout comme les mouvements de jeunesse et d’éducation populaire (Augustin, 2017), le monde du sport et ses 16 millions de licenciés se trouve directement concerné par cette montée des périls. Le cadre de la loi de 1901, l’affirmation de valeurs éducatives et le respect du principe de laïcité constituent généralement des invariants qui favorisent un apprentissage de la citoyenneté et du vivre-ensemble. Acteurs de la régulation des comportements, vecteurs d’intégration et de cohésion sociale, les clubs de football peuvent contribuer à lutter contre toute forme de © Cahiers de la sécurité et de la justice - IHEMI120 I DOSSIER Le football amateur, « terrain » ou « terreau » de la radicalisation religieuse ? L’exemple de la ligue des Hauts-de-France (2015-2018) – Olivier Chovauxcommunautarisme et de prosélytisme religieux. La F.F.F. a d’ailleurs récemment réaffirmé ses principes d’universalité et de neutralité, ainsi que leur stricte application, par l’adoption en 2017 d’une Charte d’éthique et de déontologie du football.Reste que la détection et le signalement de ces déviances supposent l’existence de dispositifs de repérage efficients, à l’image de l’Observatoire des comportements déjà évoqué. Publié en mars 2016 par le ministère de la Ville, de la Jeunesse et des Sports, un guide méthodologique définit le rôle des acteurs du sport en matière de repérage et de signalement des situations de radicalisation, qui concerne autant les pratiquants, dirigeants et éducateurs que les associations elles-mêmes : il s’agit pour les pouvoirs publics d’identifier les processus de radicalisation des jeunes et des encadrants, avant qu’ils n’aboutissent à une possible communautarisation des clubs. L’administration peut suspendre de ses fonctions (pour une durée maximale de six mois) tout éducateur dont le comportement pourrait présenter des risques pour la sécurité morale des pratiquants (article L.212-13 du Code du sport). Pour les mêmes raisons, le préfet peut procéder à la fermeture temporaire ou définitive (sauf procédure d’urgence) de toute structure sportive, indépendamment de son statut commercial ou associatif (article L.322-5). Sur les 7 317 contrôles réalisés en 2014, aucune interdiction de la sorte n’a été prononcée. Enfin, les bénéfices de l’agrément accordés à une association ou à une fédération sportive (article L.131-8) peuvent être retirés en cas de manquements constatés (fonctionnement démocratique, transparence financière, égal accès des femmes et des hommes aux instances dirigeantes). Pour autant, l’existence de dispositions juridiques et d’un arsenal administratif conséquent (plan national Citoyens du sport, guides méthodologiques et juridiques, cellules de suivi pour la prévention de la radicalisation et l’accompagnement des familles 6, Centre national d’assistance et de prévention de la radicalisation d’avril 2014) ne préjuge en rien de la capacité des instances sportives à identifier les comportements déviants pouvant se commettre au sein des clubs ou lors des rencontres de football. Les indicateurs de basculement et autres ruptures jugés comme des indices d’un processus de radicalisation religieuse demeurent bien délicats à observer et à objectiver dans le cadre de la pratique sportive. Il faudrait pour cela disposer d’une formation initiale et d’une réelle expertise pour apprécier par exemple des modifications comportementales (immaturité, instabilité, fragilités narcissiques, intolérance à la frustration, agressivité). Ou encore faudrait-il estimer qu’une prière réalisée dans le vestiaire dans les minutes précédant le coup d’envoi d’une rencontre ou le port de la barbe chez certains joueurs constituent en soi un signe avéré de radicalisation religieuse. Considérer également que « l’expression de pulsions agressives » (Delmotte, 2010, 29-36) peut être retenue comme un indicateur néglige les principes du jeu lui-même : sport de contact et d’engagement physique, chaque match de football associe violences physiques et violences instrumentales, sans que leurs auteurs ne puissent être alors suspectés de basculer dans des comportements extrémistes. Sans forcer davantage le trait, on voit bien toutes les difficultés d’utilisation d’une nomenclature peu stabilisée et les précautions nécessaires à son emploi par les acteurs de terrain, sauf à reproduire les phénomènes de stigmatisation déjà évoqués justement à propos des clubs dits communautaires.Une radicalisation statistiquement invisibleSi la détection au plus près des terrains des processus de radicalisation religieuse demeure pour le moins aléatoire, la consultation de documents administratifs pourrait constituer une autre ressource, à l’image de ce qui a été entrepris à propos des violences et des incivilités. Et ce d’autant que parmi les personnes signalées (8 964 en 2016), se trouve un nombre significatif de mineurs (19,3%) et de personnes généralement situées dans la tranche des 1429 ans qui, dans le cas du football, correspond à la majorité de ses licenciés (pour la saison 2015-2016, les catégories U14 à U20 et les seniors représentent à eux seuls 48% des licenciés masculins) 7. L’examen de l’activité des commissions d’éthique, compétentes pour traiter l’ensemble des manquements à l’éthique sportive est une première piste de travail, à supposer que les dispositions de l’article 9 du barème disciplinaire de la F.F.F. (« comportement raciste et/ou discriminatoire ») permettent bien de repérer de telles attitudes, alors notifiées sur un document officiel (rapport de l’arbitre ou du délégué de la rencontre).L’analyse exhaustive des procès-verbaux des commissions d’éthique des districts et de la ligue du Nord-Pas-de-Calais de football, pour les saisons 2014 à 2017 semble toutefois indiquer que la radicalisation ne constitue pas encore un sujet pour lesdites instances. Il faut d’emblée souligner que le champ d’action de ces commissions, tout comme leur pouvoir de sanctions, demeurent mal définis : « étudier les dossiers qui lui sont confiés et prononcer une (6) Circulaire en date du 29 avril 2014, ministère de l’Intérieur.(7) Respectivement 378 166 licenciés (U14 à U20) et 393 899 licenciés senior, pour un total de 1,6 million de licenciés masculins (F.F.F., 2019).© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 121 Cahiers de la sécurité et de la justice – Hors-série sanction, s’il y a lieu » (district maritime Nord, 2015), « et les transmettre à la commission de discipline » (ligue, 2017). Censées veiller à la stricte application de la Charte d’éthique et de déontologie du football sans disposer toutefois d’un véritable appareil coercitif, ces instances instruisent essentiellement des fraudes administratives (joueurs non-qualifiés ou joueurs suspendus inscrits sur la feuille de match, retrait d’un commun accord entre les clubs des sanctions infligées par les arbitres bénévoles au cours du match, etc.), ainsi que les comportements d’acteurs de la rencontre (joueurs, dirigeants, entraîneurs) pouvant constituer un manquement à l’éthique et qui n’auraient pas été sanctionnés au préalable par les commissions de discipline, ou que lesdites commissions leur auraient transmis pour examen (à l’exemple des allégations sur les réseaux sociaux, ou dans le cas de comportements incivils nécessitant un rappel à l’ordre formel, notamment dans les catégories « jeunes »).Sur les trois saisons retenues et compte-tenu des données disponibles, la répartition des dossiers examinés confirme, tous districts confondus, que les fraudes administratives dominent l’univers des contentieux des commissions d’éthique, pour près de la moitié des cas traités, avec une inégale distribution selon les saisons et les districts. Viennent ensuite les violences et incivilités (plus du tiers des dossiers), l’examen des propos diffamatoires sur les réseaux sociaux (12% des cas) et enfin, l’étude des propos ou actes à caractère raciste (moins de 3% des dossiers). Si l’on peut considérer que la majorité de l’activité de ces commissions correspond finalement au périmètre de leurs missions, il est possible de s’interroger sur le nombre Examen des dossiers par les commissions d’éthique. Saison 2016-2017ArtoisCôte d’opaleEscautFlandreMaritime NordLigueViolences et incivilités112630Propos diffamatoires sur les réseaux sociaux100100Fraudes administratives019720Propos ou actes à caractère raciste010011Total23111461Examen des dossiers par les commissions d’éthique. Saison 2014-2015ArtoisCôte d’opaleEscautFlandreMaritime NordLigueViolences et incivilités*355**Propos diffamatoires sur les réseaux sociaux*423**Fraudes administratives*3211**Propos ou actes à caractère raciste*000**Total*10289**(*) données non disponiblesExamen des dossiers par les commissions d’éthique. Saison 2015-2016ArtoisCôte d’opaleEscautFlandreMaritime NordLigueViolences et incivilités*8441*Propos diffamatoires sur les réseaux sociaux*220*Fraudes administratives*7901*Propos ou actes à caractère raciste*0000*Total*171362*(*) données non disponibles© Cahiers de la sécurité et de la justice - IHEMI122 I DOSSIER Le football amateur, « terrain » ou « terreau » de la radicalisation religieuse ? L’exemple de la ligue des Hauts-de-France (2015-2018) – Olivier Chovauxpour le moins confidentiel de cas qu’elles examinent en matière d’actes à caractère raciste. Ce phénomène avait d’ailleurs pu être observé pour les dossiers traités par les commissions de discipline : rares saisines par ces mêmes instances, réticence des officiels à consigner sur les feuilles de match des incidents de cette nature, généralement passés sous silence parce que banalisés par les acteurs du football amateur. Cette sorte de loi du silence peut dès lors expliquer que les comportements susceptibles d’indiquer un glissement vers un processus de radicalisation religieuse ne font l’objet d’aucun signalement. Comme pour les propos et actes à caractère raciste, leur caractère hautement sensible contribue à les rendre complètement invisibles. En l’absence de données statistiques significatives, il devient loisible pour les dirigeants de la fédération, des ligues et des districts de considérer que le monde du football demeure peu concerné par ces phénomènes, et d’exalter les valeurs humanistes et universalistes dont ce sport serait le dépositaire. 8Sur les 121 dossiers consultés, seuls trois d’entre eux concernent des « propos et/ou actes à caractère raciste », dont on pourrait considérer qu’ils constituent un signal faible d’engagement dans un processus de radicalisation religieuse 9. Leur analyse montre bien toutes les limites de l’exercice interprétatif, deux d’entre pouvant d’emblée être écartés : lors de la rencontre seniors (Excellence poule B) du 12 mars 2017 opposant les clubs de l’U.S. Montreuil à Boulogne Aiglons, des joueurs locaux se plaignent de propos racistes tenus par un supporter du club adverse, ce dernier ayant d’ailleurs été exclu du stade par le délégué de la rencontre. En outre, les propos n’ont pas explicitement été reportés sur la feuille de match. Si la commission d’éthique en charge du dossier relève dans ses attendus « qu’une rencontre peut être arrêtée si des propos racistes sont tenus et si l’auteur de ceux-ci est identifié (…) Que le racisme n’est pas une opinion mais un délit passible d’amende et de prison », ce dernier sera néanmoins classé sans suite, au motif de « la bonne réaction du délégué de terrain, des excuses présentées par le club de Montreuil et acceptées par le club des Aiglons, de la volonté des deux clubs de ne pas envenimer cette affaire et donc de rester en bons termes ». Présentant quelques similitudes avec le dossier précédant, le deuxième cas (« menaces racistes de spectateurs / supporters du club recevant », lors d’un match seniors promotion Fleurbaix U.S. / Hem Olympic F.C. du 9 octobre 2016), amènera la commission d’éthique du district Flandre à prendre une sanction plus conséquente, en infligeant un retrait de trois points au classement (avec sursis) à l’équipe de Fleurbaix, au titre de la police des terrains, en vertu de l’article 74 des règlements généraux de la F.F.F. (Procès-verbal de la commission d’éthique du district Flandre de football, 2016). En l’espèce, c’est plus l’attitude menaçante des supporters envers les joueurs et officiels qui aura été retenue que les propos racistes tenus à l’encontre de l’équipe visiteuse, sans que leur nature ne soit d’ailleurs rapportée (idem) :« Après avoir entendu les représentants des deux clubs sur des incidents ayant débuté entre la 65e et la 70e minute provenant de trois ou quatre individus (supporters ou spectateurs du club recevant), cherchant l’affrontement avec des menaces, propos et injures racistes, boissons (canette en verre) à la main. Les membres de la commission s’étonnent que les représentants du club de Fleurbaix ne semblent pas connaître ces protagonistes, l’un d’eux ayant pénétré dans le vestiaire du club recevant à la fin de la rencontre (…) »Synthèse des dossiers examinés par les commissions d’éthique2014-20152015-20162016-2017dossiers%dossiers%dossiers%Violences et incivilités1327%1745%1336%Propos diffamatoires sur les réseaux sociaux919%410%13%Fraudes administratives2534%1745%1953%Propos ou actes à caractère raciste003Total473836(8) « Le football, parce qu’il est le sport le plus pratiqué en France et le plus médiatisé, se doit d’offrir, notamment aux jeunes, une image exemplaire car il doit rester la fête de l’humain et de la fraternité » (Charte d’éthique et de déontologie du football, décembre 2017).(9) Trois conditions cumulatives caractérisent la radicalisation : une évolution du comportement et une transformation progressive de la personne ; une adhésion à une idéologie extrémiste pouvant prendre différentes formes ; l’adoption de la violence comme principal mode d’action. (Ministère de l’Enseignement supérieur, 2019).© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 123 Cahiers de la sécurité et de la justice – Hors-série Le dernier dossier mérite une attention particulière : lors d’une rencontre seniors de division d’honneur régionale opposant le club de Neuville-en-Ferrain à Lille-Sud (en date du 2 octobre 2016), deux joueurs de Lille-Sud auraient refusé de serrer la main de l’arbitre assistante lors du protocole d’avant-match. Examinant une première fois les faits, la commission d’éthique de la ligue des Hauts-de-France ne peut que renvoyer le dossier, en l’absence de tout débat contradictoire, les joueurs incriminés niant les faits, évoquant une « confusion dans la numérotation des maillots » et considèrent l’incident comme « insignifiant » (Procès-verbal de la commission d’éthique de la ligue des Hauts-de-France, 2016). La présence de l’arbitre de la rencontre lors de la seconde réunion permet de confirmer l’identité des joueurs fautifs, ces derniers se voyant infliger six matches de suspension fermes « pour comportements racistes ou discriminatoires » (article 1-10 de l’annexe 5 des règlements généraux de la F.F.F.) Reste que les attendus ne permettent pas d’apprécier les motivations des deux joueurs, ce qui pose d’ailleurs la question de la qualification des faits : misogynie, attitude sexiste, expression de convictions religieuses ou culturelles, absence de civilité, contentieux antérieur entre joueurs du club et officiels ? Sauf à considérer que le club de Lille-Sud est classé dans la catégorie des clubs « dans le risque », rien ne permet d’affirmer que le comportement de ses joueurs serait la traduction d’une forme de prosélytisme religieux.En définitive, à l’image des violences et des incivilités, la question de la radicalisation religieuse dans le football amateur est une terra incognita qu’historiens et sociologues doivent continuer à explorer de conserve. Avec plus de deux millions de licencié(e)s, près de 16 400 clubs, 400 000 bénévoles et 7 000 salariés et plus d’un million de matches disputés chaque saison (F.F.F., 2019), le football tient une place éminente au sein du système des sports justifiant qu’il soit autant exposé que regardé et que l’éventail des formes de radicalité qu’il fabrique puisse être passé au crible, indépendamment des réticences institutionnelles et de la fragilité des outils de recensement disponibles à ce jour.Sur la base des éléments recueillis, pour la période considérée et à partir d’une méthodologie d’enquête strictement comptable, les clubs de football de la Ligue du Nord-Pas-de-Calais ne constituent pas plus un terrain qu’un terreau fertile pour le prosélytisme religieux. Sans doute conviendrait-il de compléter ces quelques données par des enquêtes ethnographiques, au plus près des terrains et de ces clubs considérés comme plus exposés, ce que les chiffres infirment par ailleurs.Terre d’immigration sportive depuis la fin du xixe siècle, le football nordiste contemporain offre un paysage cosmopolite où les clubs communautaires, épousant la géographie urbaine et industrielle de la France des Trente Glorieuses, font déjà l’objet d’une étrange suspicion de la part des instances disciplinaires, en matière de violences et d’incivilités. Pour autant, rien ne permet d’affirmer, par la consultation des archives des commissions d’éthique de 2014 à 2017, que ces clubs « dans le risque » soient un terreau de la radicalisation religieuse. Comme le souligne Laurent Mucchielli à propos du lien quasi-mécanique opéré de nos jours entre immigration et délinquance juvénile (2007), il convient de se garder de pareils amalgames et éviter de considérer que ces clubs représentent a priori un risque potentiel. Ils n’ont pas été identifiés comme tels d’un point de vue statistique. On peut toutefois s’interroger sur le caractère très exceptionnel des dossiers pouvant avoir un lien, fût-il ténu, avec notre objet. L’explication tient à cette double frilosité, observée dans l’identification et le traitement des actes ou propos à caractère raciste : peu consignés lorsqu’ils se produisent lors des rencontres, peu sanctionnés par les commissions d’éthique, ils participent d’une sorte d’omerta partagée, depuis les pâtures du football du dimanche jusqu’aux salons feutrés du siège de la Fédération, et ce en dépit des injonctions du ministère de tutelle. L’historien des violences sportives ne peut qu’être surpris par ce manque de discernement, qui rend invisible une menace bien réelle pour les démocraties (Cusset, 2018), au même titre que d’autres formes d’extrémisme, dont les stades de football peuvent être le creuset (Bodin, 2010) nRéférencesOuvrages et articlesAugustin, Jean-Pierre, 2017, Loisirs des jeunes. 120 ans d’activités éducatives et sportives, La Documentation française/FONJEP.Bancel, Nicolas et Pascal Blanchard, 2005, la Fracture coloniale. La société française au prisme de l’héritage colonial, La Découverte.Bodin, Dominique, 2010, « Le football à l’épreuve du racisme et de l’extrémisme : un état des lieux en Europe », in Crettiez Xavier et Laurent Mucchielli, les Violences politiques en Europe, La Découverte, coll. « Recherches ».Castel, Robert, 2006, « La discrimination négative. Le déficit de citoyenneté des jeunes de banlieue », Cambridge University Press, Volume 61, n°4, août. © Cahiers de la sécurité et de la justice - IHEMI124 I DOSSIER Le football amateur, « terrain » ou « terreau » de la radicalisation religieuse ? L’exemple de la ligue des Hauts-de-France (2015-2018) – Olivier ChovauxChauvaud, Frédéric, 2010, la Dynamique de la violence, Presses Universitaires de Rennes. Chovaux, Olivier, 2012, « Quantifier les violences et les incivilités dans le sport amateur : le cas du football », Revue EPS, vol. 351, mars/avril. Chovaux, Olivier, 2013, « La prise en compte de l’ethnicité sportive dans la régulation des violences et incivilités : le cas du football amateur », International journal on violence and school, vol. 13. Cusset, François, 2018, le Déchaînement du monde. Logique nouvelle de la violence, La Découverte. Delmotte, Florence, 2010, « Termes clés de la sociologie de Norbert Elias », Vingtième Siècle, n°106, p. 29-36.Deluermoz, Quentin, 2010, « Norbert Elias et le XXe siècle : le processus de civilisation à l’épreuve », Vingtième Siècle, vol. 106, avril/juin. Jobard, Fabien, 2005) « Déviances et modalités de contrôle. La France et l’Allemagne en perspective », Déviance et Société, n°3, vol. 29.Khosrokhavar, Farhad, 2014, Radicalisation, Maison des Sciences de l’Homme, coll. « Itinéraires ». Mignon, Patrick, 1995, « La violence dans les stades. Supporters, ultras et hooligans ». Actes des entretiens de l’INSEP, vol. 10. Mucchielli, Laurent, 2002, Violences et Insécurité. Fantasmes et réalités dans le débat français, La Découverte, coll. « Sur le vif ». Noiriel, Gérard, 2016, le Creuset français. Histoire de l’immigration (xixe – xxe), Seuil, coll. « Points ».Nuytens, Williams, 2003, Étiologie des violences dans le football amateur. Enquêtes sociologiques dans dix clubs du département du Nord, DRDJS du Nord-Pas-de-Calais, rapport de recherche financé. Nuytens, Williams, 2016, « Violences à l’école : l’éducation par le sport peut-elle tout régler ? » in Attali Michaël, les Éducations par le sport, Editions Canopée.Pfister, Richard, 1994, « Les interactions agressives dans la pratique sportive des jeunes », Enfance, n°2/3, p. 215-232.Vieille-Marchiset, Gilles et Benjamin Coignet, 2015, Clubs sportifs en Banlieue : des innovations sociales à l’épreuve du terrain, Presses Universitaires de Strasbourg.Comptes-rendus et rapports officielsFédération française de football, 2019, Archives.Fédération française de football, 2007, Assemblée fédérale du 23 juin 2007.Charte d’éthique et de déontologie du football, décembre 2017.Circulaire en date du 29 avril 2014, ministère de l’Intérieur.Ministère de l’Enseignement supérieur, octobre 2019, Prévenir la radicalisation dans l’enseignement supérieur et les organismes de recherche.Procès-verbal de la commission d’éthique de la ligue des Hauts-de-France, réunion du 28 décembre 2016.Procès-verbal de la commission d’éthique du district Flandre de football, réunion du 16 novembre 2016.© Cahiers de la sécurité et de la justice - IHEMIVARIA I 125Cahiers de la sécurité et de la justice – Hors-série Are amateur soccer fields a breeding ground for religious radicalization? Case study of the Hauts-de-France League (2015-2018)Olivier CHOVAUXAddressing the issue of radicalization in amateur soccer means interrogating a “blind spot” in historiographical research as well as trying to distantiate from a debate that has been dominated by emotions. Introduced by a necessary contextualization, this article presents the results of a study that was conducted on the territory of the Hauts-de-France Soccer League between 2015 and 2018, which demonstrated how infinitesimal the phenomenon is with regard to the methodology and the field chosenKeywords: soccer, violence, radicaliza-tion.ddressing the issue of radicalization within soccer clubs means questioning a “blind spot” in historical research, and immediately confronts the histo-rian with several methodological difficulties. While it seems the concept of radicalization now has a stable definition for public au-thorities, 1 studying such a com-plex process in the field of sports and amateur soccer requires an empirical approach, allowing to qualify it, locate it, quantify it, and inscribe it in an “immediate time frame” (2015-2018).The territory of the Hauts de France League, which has previously been the object of abundant academic research on violence and incivility (Nuytens, 2003) as well as ethnicity in sports (Chovaux, 2013), has been selected for this study: this choice is justified by anterior, indigenous knowledge of the cartography of these clubs, and the availability of a corpus of institutional archive documents which could a priori be exploited for the research. However, it seems that it would be difficult to apply systematically the same “methodological quartet” (describe, categorize, contextualize and understand), although efficient in the previous analysis of violence in sports (Chauvaud, 2010), to such a volatile and politically sensitive topic as radicalization.It must be reminded from the outset that radicalization also unfolds in other educational structures. In the case of soccer, the contours of the phenomenon of radicalization can be outlined as mostly religious, and it allegedly results from a sort of predestination, related to the identity of specific clubs. These clubs, qualified by the authorities Olivier CHOVAUXProfessor of Contemporary History at the Artois University. Cofounder of the SHERPAS Workshop (a component of the URePSSS, EA 7369), his research focuses on the history of soccer, and more recently, of violence in sports and its regulations. He is currently working on the history of referees and refereeing in France, from the end of the 19th century to the present day. (1) “Radicalization is defined as the process through which a person (or a group of persons) challenge the status quo by adopting an extreme form of action, related to an ideology which may be political, social or religious.” (Ministry of Higher Education, 2019. See also Khosrokhavar, 2014).A© Cahiers de la sécurité et de la justice - IHEMI126 I DOSSIER Are amateur soccer fields a breeding ground for religious radicalization? Case study of the Hauts-de-France League (2015-2018) – Olivier Chovauxas “separatist”, are considered to be a de facto breeding ground for the identity-based fragmentation of society into separate minority communities. This mechanical analogy often operated between “ethnic minority clubs” and religious radicalization recalls the analogy that conflates “inner city clubs” or “urban neighborhood clubs” with “sensitive clubs”, in terms of violence and incivilities committed on the field or in and around the stands of stadiums. The object of this study thus consists in questioning religious radicalization in amateur soccer clubs, by situating it in its specific context and territory. Reviewing the results of previous research on other forms of radicality in sports (such as violence and incivility) can allow to understand both the methodological difficulties of strictly quantifying these phenomena, and the kind of “institutional omerta” sometimes maintained around them by the institutions. Indeed, considering soccer clubs as potential hotbeds of radicalization contradicts the doxa widely spread by sports leaders: it would mean admitting the failure of the alleged educational and integrative values of sports, which are constantly reified but wouldn’t fare well with such a hypothesis.Nonetheless, some elements of context can explain why the world of amateur soccer could be a “breeding ground”, predisposed to the expression of deviance, however statistically insignificant might they be, as confirmed by the investigation carried out in the region of Hauts-de-France. The presence of “ethnic minority clubs” on that geographical territory constitutes, a priori, a factor that might further aggravate religious radicalization. However, the actual extent of the impact of this factor will only be measured by putting these phenomena in perspective over time. 2The elements of context that might predispose soccer to devianceViolence and incivility in sports are of little significanceIt was probably during the first years of the Trente Glorieuses (the “Glorious Thirty”, years of growth and prosperity from 1945 to 1975) that soccer’s integrative function, especially for immigrant populations, became more visible, not only within sports organizations. In the North of France, while Polish immigrants were then in an ambiguous situation, fluctuating between the preservation of their cultural traditions and the desire to erase their differences, immigrants from North Africa, in the context of decolonization, were forced to adopt a different posture: that of invisibility (Noiriel, 2016). It is therefore not surprising that the latter were subjected to stigmatization in the 1970s: the prodromes of the economic crisis, the rise of the Front National extreme right-wing party, and the unsettled accounts of the “colonial fracture” (Bancel and Blanchard, 2005) shifted the paradigm around migration issues, and inserted them into a timeframe that still hasn’t been closed to this day. In the region of Nord-Pas-de-Calais, the arrival of immigrants from North Africa (Moroccans, Algerians and Tunisians) and Southern Europe (Portuguese and Spanish) was visible in the creation of new soccer clubs, geographically distributed according to urban and industrial concentrations: in 1990, approximately 15 soccer clubs could be found in the Lille metropolitan area (Lille-Roubaix-Tourcoing), whose names often carried a reference to the country of origin (“the Algerians of Tourcoing”, “the Portuguese of Roubaix”, “the Mahorans of Lille”), while others displayed different intentions, such as the Muslim Association of animation and leisure of Grande-Synthe, or the French-Maghrebi socio-cultural sports association of Douai, in the Escaut district. At the end of the 1990s, less than 6% of the 1000 clubs affiliated with the Nord-Pas-de-Calais regional League reflected ethnicity or community references in their name. A sort of trend reversal was then operated, with a shift both in the identity of these clubs and the way they were perceived. The desire to be with one’s own, already present in the interwar period, persisted and remained the basis of sociability in sports, but the affirmation of separate identities then started to surface in various zones of the national territory: Turkish clubs in the Alsace region, Portuguese clubs around Bordeaux, or Armenian clubs in the Drôme are the best-known examples. Their location and their demographics quickly designated them as “urban clubs” or “neighborhood clubs” (Vieille-Marchiset and Coignet, 2015), and their integrative function was emphasized within the then emerging urban policies. This semantic shift appeared at a time when the question of violence and incivility became more discussed in amateur soccer, even though it was insignificant in terms of quantity. (2) As was highlighted by Alfred Wahl (quoted by Mignon, 1995) regarding violence in sports. (3) Federal Assembly of June 23rd, 2007, Archives of the F.F.F. (extract) © Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 127 Cahiers de la sécurité et de la justice – Hors-série In 2006, the French Soccer Federation (F.F.F.) established a Behavior Monitoring Center, which enabled it to conduct a first set of quantitative surveys that allowed to “inventory and analyze the incivilities of all kinds committed on soccer fields, in order to better oppose them, through prevention or repression.”3 Reviewing the game report sheets of the official matches organized by leagues and districts means adopting the federal classification to categorize the types of violence and incivilities: thus, for the 2008-2009 season, incidents of “hostile violence” were reported in only 11 338 games, and only 454 had to be interrupted prematurely, out of the 709 514 that took place on a national level. These figures represent respectively 1,59% and 0,06% of the matches. The majority of these incivilities (69,9% of them) happened in the “senior” category, where players are 20 to 35 years old (Chovaux, 2012). “Incivilities” are characterized by obscene gestures or spitting. Acts of “instrumental violence” are committed by players during the game, and constitute mild violations, often penalized with a simple warning. Acts of “hostile violence” are comprised of behaviors occurring outside of the game, and are more reprehensible: verbal violence (insults, threats or intimidations), physical violence (deliberately shoving, attempting to hit someone, brutalizing or hitting someone).4For the 2006-2009 period, the surveys conducted in the three Northern regional leagues (Nord-Pas-de-Calais, Picardie, Champagne-Ardenne) resulted in almost identical figures, in terms of the amount of violence and incivilities as well as their distribution. This preliminary detour on the question of violence shows how the world of amateur soccer is, ultimately, a pacified world, as much as schools or other educational institutions can be (Nuytens, 2016). The residual deviant behaviors that still occur marginally echo Norbert Elias’ Civilizing Process theory, which has sometimes been criticized (Deluermoz, 2010), but is still relevant. Deviations from the norm remain, after all, of little significance. Subsequently, it is unlikely for the field of amateur football, which is very scarcely exposed to the issue of hostile violence (outside of the game), to be a breeding ground for that of religious radicalization. The fact that “community-based” clubs (which, as we saw, come from a long process of construction related to the waves of immigration) are sometimes mistakenly conflated with “separatist clubs”, might explain why football is considered (at times even by its own institutions) as especially predisposed to religious radicalization, with regard to the demographics and the geographical location of these clubs. Nonetheless, the results of this research tend to show that no serious correlation can precisely be established between ethnic identity and recurrence of deviant behaviors. Community-based clubs are little exposed to violence, but very stigmatizedThe exhaustive analysis of the minutes of the regional commission of appeal of the Nord-Pas-de-Calais soccer league for the 2006-2009 period (Chovaux, 2013), enables to measure the relatively limited impact, in terms of statistics, of these 31 clubs qualified as “community-based” only because of their name and territorial location. Being the highest disciplinary body in the hierarchy of recourse, that commission examines the most serious (4) According to the by-laws and regulations of the F.F.F., 2008-2009 edition, pages 193 and following. Regarding the distinction between “instrumental” and “hostile” violence, see Pfister, 1994.(5) It can however be noted that only 5% of them are related to the “community-based” clubs, which seems to indicate that their leaders have little “litigation culture” and tend not to resort to judicial processes, often for lack of knowledge of legal remedies.Reported incidents in each league and district (indicated as a percentage of the total number of games)Data from the F.F.F.Champagne-ArdenneNord-Pas-de-CalaisPicardieLeagueDistrictLeagueDistrictLeagueDistrict2006-2007***5,682,47**2007-20081,762,192,695,132,505,141,472008-20091,651,873,781,181,997,851,82Average1,692,033,233,992,326,491,64(*) données non disponibles© Cahiers de la sécurité et de la justice - IHEMI128 I DOSSIER Are amateur soccer fields a breeding ground for religious radicalization? Case study of the Hauts-de-France League (2015-2018) – Olivier Chovauxcases. Regarding hostile violence, out of the 142 disciplinary cases that were examined (the 108 cases with judicial implications have been left out), 5 29 are related to these clubs, which amounts to almost ¼ of the disputes. Proportionally, they represent a small number (only 6% of the affiliated clubs) and a phenomenon of recurrence has been observed with some of them, for example the Valenciennes Dutemple FC, the US Antillais (West Indian Sports Union), and the US Portugais (Portuguese Sports Union) of Roubaix-Tourcoing. This has sometimes led the lower commissions to apply the disciplinary scales in a differentiated manner, and be harsher towards these clubs than they are towards other clubs considered to be “normal”. There would be much to say about this discriminatory justice process, this “judicial profiling” (Jobard, 2005), which sometimes gives the defendants the impression that they are treated inequitably, and causes the clubs to be wary of, or even fully distrust the disciplinary bodies and the very sports institutions. Ethnicity then becomes a resource, wherewith the clubs establish self-protective behaviors, playing on the stigmatization they consider themselves victims of in order to minimize the seriousness of the violations committed by their members, especially in cases of hostile violence. As they are subjected to “negative discrimination” (Castel, 2006), these clubs considered to be “at risk” should rather be qualified as “in risk”: the fragility of their educational staff, the precarious and scarce sports equipment and facilities at their disposal, the surrounding socio-economic environment and the level of engagement of their teams in competition are all invariables that must be taken into account, if we are to avoid precisely any labeling or stigmatization. It thus appears how much of a double-edged sword the issue of ethnicity in sports can be: on the one hand, it perpetuates the myth of integration through sports, which reached its peak with the victory of the “Black, Blanc, Beur” French team in the 1998 World Cup final (“Black, Blanc, Beur”, a popular slogan at the time, uses slang terms to describe the ethnic diversity of the 1998 French soccer team, and could be translated “Black, White, Brown” or “Black, White, Arab”). On the other hand, the issue of ethnicity in sports can also be seen as fueling extremist, transgressive behaviors which challenge the founding values not only of sports themselves (principle of fairness and equity, humanistic ideals, etc.), but also of the political and institutional framework in which they are practiced (principle of secularism, Republican values, concept of democracy, etc.). Adopting postures, attitudes or behaviors related to religious radicalization would then constitute a double transgression, as they would be situated in a social space which historically constructed itself on values incompatible with extremist positions. It would mean using sports for purposes of propaganda, whether explicit or not, in order to fight and ultimately destroy the democratic foundations and values of regimes considered as political enemies. Are amateur soccer fields a breeding ground for religious radicalization?How to prevent religious radicalization?In the same way as for the study of violence and incivility in the Hauts-de-France region, in the case of religious radicalization, the historian must be resourceful in order to find an exploitable corpus of archive that can then be cross-checked with the first official statistics that appear in the F.S.P.R.T. (File of Reported individuals for the Prevention of Terrorist Radicalization, created by a decree of March 2015): considered by public authorities to be a real barometer of radicalization, it gathers the data collected by government services (police, gendarmery, public education, etc.) as well as the information provided by citizens through the toll-free phone number launched by the French Ministry of Homeland Security (“Ministry of the Interior”) in April 2014. Collected on a departmental scale, the data published for the year 2016 indicate 295 reported individuals in the Pas-de-Calais department and 502 in the Nord department, out of a total 11 820 at a national level. These two departments thus respectively represent 2,4% and 4,2% of the national figures. And the Hauts-de-France region, that these two departments belong to, is among the four territories most affected, along with Île-de-France (the Paris region), Rhône-Alpes (the Lyon region) and Provence-Alpes-Côte-d’Azur (the region of Marseille and Nice). From a qualitative point of view, 27% of the individuals reported are women, 35% are converts, and 17% of them are under eighteen years old. While the most serious cases are monitored by police services according to their degree of radicalization (General Directorate of Interior Security, Central Service of Territorial Intelligence, gendarmery), most of the reported individuals are provided the support of other public services (Judiciary Protection of Youth, social services, etc.). Since October 2017, in the Lille school district, schools and universities have been required to report individuals who are experiencing radicalization, with the help of the national prevention plan designed to spot its early signs.Just like the youth movements and the movements of popular education (Augustin, 2017), the world of © Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 129 Cahiers de la sécurité et de la justice – Hors-série sports and its 16 millions of licensed members is directly affected by these increasing dangers. The legal framework of freedom of association defined by the 1901 law, the affirmation of educational values and the respect of the principle of secularism (in its French version of laïcité), are generally unvarying elements that promote teachings of citizenship and coexistence. As agents of behavior regulation and vectors of integration and social cohesion, soccer clubs can contribute to the fight against any expression of ethnic separatism and religious proselytism. In fact, the F.F.F. recently reaffirmed its principles of universality and neutrality, as well as their strict enforcement, through the adoption in 2017 of a Charter of ethics and deontology in soccer. Nonetheless, spotting and reporting these deviances requires efficient detection devices, such as the Behavior Monitoring Center mentioned above. A methodological guide published in March 2016 by the Ministry of Cities, Youth and Sports defines the role of sports stakeholders in terms of spotting and reporting situations of radicalization, regarding players, leaders and educators as well as the associations themselves: the goal is, for public authorities, to identify processes of radicalization both in youth and in management, before they result in a possible separatist withdrawal of entire clubs. The administration is allowed to suspend from duty (for up to 6 month) any educator whose behavior might pose risks to the moral security of participants (article L.212-13 of the Code of Sports). For the same reasons, prefects can proceed to close temporarily or definitively (except in emergency cases) any sports organization, regardless of its administrative status (business or association), according to article L.322-5 of the Code of Sports. Out of 7317 inspections performed in 2014, no such closure has been ordered. Lastly, the benefits of the official accreditation granted to a sports association or federation (article L.131-8) may be removed in case of failure to comply with conditions such as a democratic mode of organization, financial transparency, equality of access to leadership positions for women and men… However, the existence of judicial provisions and a substantial legal and administrative arsenal (“Citizens of Sports” national plan, legal and methodological guides, monitoring units for prevention of radicalization and support to affected families 6, National Center of Assistance and Prevention of Radicalization founded in April 2014) does not mean that sports institutions will be able to identify deviant behaviors that might occur within the clubs or during soccer games. Indeed, the indicators of a tipping point or breaking point that might allow the detection of a religious radicalization process remain difficult to observe and appraise objectively in the context of sports activity. That would require initial training and true expertise in order to assess, for example, a change in behavior (immaturity, instability, narcissistic fragility, intolerance to frustration, aggressiveness). Or one might even considered that performing a prayer in the locker room a few minutes before kick-off, or the fact that some players don a beard, inherently constitute a confirmed sign of religious radicalization. But it must also be taken into account that “the expression of aggressive pulsions”, (Delmotte, 2010, 29-36), which can be seen as an indicator, also relate to the very principles of the game: soccer being a contact sport, that involves physical engagement, every soccer match includes occurrences of physical violence and instrumental violence, which doesn’t imply that the offending players may be suspected of having tipped over extremist behavior. There is no need to further hammer the point: it is clear that the use of an unstable nomenclature comes with difficulties, and the endeavor must be undertaken cautiously by the actors on the field, lest they risk perpetuating the phenomena of stigmatization mentioned above precisely about these clubs deemed ethnic.(6) Circular of April 29th, 2014, issued by the Ministry of the Interior.Cases examined by the ethics commissions. Season 2014-2015ArtoisDistrictCôte d’Opale DistrictEscautDistrictFlandreDistrictMaritime Nord DistrictLeagueViolence and incivilities*355**Defamatory comments on social media*423**Administrative fraud*3211**Racist actions or comments*000**Total*10289**(*) données non disponibles© Cahiers de la sécurité et de la justice - IHEMI130 I DOSSIER Are amateur soccer fields a breeding ground for religious radicalization? Case study of the Hauts-de-France League (2015-2018) – Olivier Chovaux(7) Respectively 378 166 U14 to U20 licensed players, and 393 899 senior licensed players, out of a total 1,6 million male licensed players (F.F.F., 2019).(8) “Soccer, because it is the most practiced sport in France, and the one with most media coverage, must provide, especially for the youth, an exemplary model, as it must remain a celebration of humanity and fraternity” (Charter of ethics and deontology in soccer, December 2017).Cases examined by the ethics commissions. Season 2015-2016ArtoisDistrictCôte d’Opale DistrictEscautDistrictFlandreDistrictMaritime NordDistrictLeagueViolence and incivilities*8441*Defamatory comments on social media*220*Administrative fraud*7901*Racist actions or comments*0000*Total*171362*Cases examined by the ethics commissions. Season 2016-2017ArtoisDistrictCôte d’Opale DistrictEscautDistrictFlandreDistrictMaritime Nord DistrictLeagueViolence and incivilities112630Defamatory comments on social media100100Administrative fraud019720Racist actions or comments010011Total23111461A statistically invisible radicalization While the detection of religious radicalization processes at soccer field level remains rather hazardous, consulting administrative documents might turn out to be an interesting source, as it proved to be regarding violence and incivility. All the more so that among the reported individuals (8964 in 2016), an significant amount are under 18 years old (19,3%), or within the 14-29 age group in general. The majority of licensed soccer players fall within that age range (for the 2015-2016 season, the U14 to U20 categories and the seniors, aged 20 through 35, represent 48% of male licensed players). 7 Examining the activity of the ethics commissions, which are competent to deal with any failure to comply with sports ethics, could be a first lead to work on, assuming the provisions of article 9 of the F.F.F.’s disciplinary scale (“racist and/or discriminatory behavior”) allow to spot such attitudes, which would then be notified on an official document (report of the referee or match delegate).An exhaustive analysis of the minutes of the ethics commissions of each district and of the Nord-Pas-de-Calais soccer League for seasons 2014 through 2017 seems to reveal, however, that radicalization still isn’t a relevant issue for said institutions. First of all, it must be noted that what falls within the scope of these commissions’ activity isn’t very well defined, and neither is their power to enforce sanctions: “studying the cases entrusted to its authority, and pronouncing a penalty if appropriate” (Maritime Nord district, 2015) “and transmitting them to the disciplinary commission” (League, 2017). These bodies are supposed to ensure the strict enforcement of the Charter of ethics and deontology in soccer, although they do not dispose of any real coercive machinery. They essentially investigate administrative fraud (non-qualified or suspended players’ names appearing on game sheets, clubs reaching mutual agreements to remove the sanctions inflicted by volunteer referees during matches, etc.), as well as any behavior exhibited by participants in a game (players, leaders, coaches) that might constitute a breach of ethics and not have been previously penalized by disciplinary commissions, or that said commissions might have referred to them for review (for example in the case of allegations on social media, or in the case of incivilities that might need a formal warning, especially in the “youth” categories).(*) données non disponibles© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 131 Cahiers de la sécurité et de la justice – Hors-série In the three seasons selected for research, and given the available data, the distribution of the examined cases confirms, across all districts, that administrative fraud is the main litigation issue faced by ethics commissions, as it represents almost half the cases, although unequally distributed between seasons and between districts. Next are violence and incivilities (over a third of the cases), defamatory comments on social media (12% of cases), and lastly, racist actions or comments (less than 3% of cases). While it can be considered that the majority of these commissions’ activity falls within the perimeter of their mission, one can ponder over the inconsequential number of cases of racism they examine. The same phenomenon can be observed with the cases handled by the disciplinary commissions: such cases are rarely referred to these institutions, as sports officials are reluctant to record incidents of that nature on game sheets, and they are generally hushed up because they have become normalized in amateur soccer. This kind of “code of silence” can explain why behaviors susceptible to indicate a shift towards a process of religious radicalization are never reported. Just like racist actions and comments, their sensitive nature contributes to making them completely invisible. In the absence of significant statistics, it becomes easy for federation, league and district leaders to consider that the world of soccer is little affected by these phenomena, and to exalt the humanistic and universalist values allegedly embodied by that sport. 8Out of 121 reviewed cases, only 3 were about “racist actions and/or comments”, which might be considered as a weak signal of engagement in a process of religious radicalization. 9 But their analysis demonstrates the limits of that interpretation, as two of them can be immediately discarded: during the seniors game (Excellence pool B) of March 12th, 2017 between the US Montreuil club and the Boulogne Aiglons club, local players complained of racist comments made by a supporter of the opposing club, who was actually excluded of the stadium by the match delegate. Besides, the content of his comments was not reported explicitly on the game sheet. The ethics commissions in charge of the case stated, in the arguments prior to its ruling, “that a game can be interrupted in case of racist comments, if the person who made them is identified [...] That racism is not an opinion but an offense, punishable by fine or imprisonment”. However, the case was dropped without further action, because of “the good reaction of the field delegate, the apologies offered by the Montreuil club and accepted by the Aiglons club, and the desire from both clubs to avoid aggravating the situation and remain in good terms”. The second case (“racist threats made by spectators/supporters of the hosting club” during a seniors game between Fleurbaix US Club and the Hem Olympic FC of October 9th 2016) presents some similarities with the previous one. But in that instance, the ethics commission of the Flandre district imposed a more serious sanction: the (suspended) deduction of three points in the ranking from the Fleurbaix team, according to field policing rules, by virtue of article 74 of the F.F.F.’s general bylaws (minutes of the ethics commission of the Flandre soccer district, 2016). In that case, the threatening attitude of the supporters towards the players and the officials was given more importance than the racist comments they made against the visiting team. Indeed, just like in the first case, the exact content of these comments was not reported: “After hearing the representatives of each club regarding the incidents that began between the 65th and 70th minute, caused by three or four individuals (supporters or spectators of the hosting club) who were seeking a clash through threats, racist comments and slurs, holding drinks (glass cans) in their hands. The members of the commission are surprised that the Fleurbaix club representatives claim Overview of cases examined by the ethics commissions.2014-20152015-20162016-2017cases%cases%cases%Violence and incivilities1327%1745%1336%Defamatory comments on social media919%410%13%Administrative fraud2534%1745%1953%Racist actions and/or comments003Total473836(9) Three cumulative criteria are needed to qualify radicalization: a shift in behavior and a progressive transformation of the person; adherence to an extremist ideology, which can manifest in different ways; the adoption of violence as principal mode of action (Ministry of Higher Education, 2019).© Cahiers de la sécurité et de la justice - IHEMI132 I DOSSIER Are amateur soccer fields a breeding ground for religious radicalization? Case study of the Hauts-de-France League (2015-2018) – Olivier ChovauxReferencesBooks and articles Augustin, Jean-Pierre, 2017, Loisirs des jeunes. 120 ans d’activités éducatives et sportives, La Documentation française/FONJEP.Bancel, Nicolas et Blanchard, Pascal, 2005, La Fracture coloniale. La société française au prisme de l’héritage colonial, La Découverte.Bodin, Dominique, 2010, « Le football à l’épreuve du racisme et de l’extrémisme : un état des lieux en Europe », in Crettiez, Xavier and Mucchielli, Laurent, Les violences politiques en Europe, La Découverte, collection « Recherches ».Castel, Robert, 2006, « La discrimination négative. Le déficit de citoyenneté des jeunes de banlieue », Cambridge University Press, Volume 61, n°4, August. not to know these individuals, as one of them entered the hosting club’s locker room after the match (...)”The last of the three cases warrants special attention: during a regional honors division seniors match between the Neuville-en-Ferrain club and the Lille-Sud club on October 2nd, 2016, two players of the Lille-Sud club allegedly refused to shake hands with the female assistant referee during the pregame protocol. During the initial investigation, the ethics commission of the Hauts-de-France league had no choice but to adjourn the case, as in the absence of adversarial debate, the accused players denied the facts, claiming there was a “confusion in the numbering of the jerseys” and the incident should be considered “trivial” (minutes of the ethics commission of the Hauts-de-France league, 2016). During the second meeting of the commission, the presence of the match referee allowed to confirm the identity of the offending players, who were punished with suspension for a duration of six games, without remission, “for racist or discriminatory behaviors” (article 1-10 of annex 5 of the F.F.F.’s general bylaws). However, the commission’s arguments do not clarify the motives of the two players, which raises the issue of the classifications of the facts: misogyny, sexist attitude, expression of religious or cultural beliefs, incivility, previously existing conflict between the players and the officials? Unless one considers that the Lille-Sud club falls under the “clubs in risk” category, there is no evidence to back up the claim that its players’ behavior is a manifestation of religious proselytism. Ultimately, the issue of religious radicalization in amateur soccer, just like that of violence and incivility, is a terra incognita that historians and sociologists must continue to explore together. Having over 2 million licensed players, nearly 16 400 clubs, 400 000 volunteers and 7 000 workers, and over a million matches played every season (F.F.F., 2019), soccer holds a prime position within the sports system, which explains why it is exposed and watched so much, and that the range of expressions of radicality that it produces is scrutinized, in spite of the institutional reluctance and the fragility of the measurement tools available to date. Based on the information gathered so far, over the studied period and within a strictly statistical methodology of investigation, the soccer clubs of the Nord-Pas-de-Calais soccer League do not constitute a breeding ground for religious proselytism. It would certainly be useful to complete this bit of data by conducting an ethnographic research closer to the field, in the clubs considered to be more exposed, although the numbers seem to disprove that.In a land where sports and immigration have been related since the late 19th century, contemporary soccer in the North of France offers a multicultural landscape, where ethnic minority clubs match the urban and industrial geography of the Trente Glorieuses. Such clubs are already the target of a strange suspicion by the disciplinary bodies in terms of violence and incivilities. However, the archives of the ethics commissions from 2014 to 2017 do not contain sufficient evidence to suggest that these “clubs in risk” are a breeding ground for religious radicalization. As Laurent Mucchielli claims regarding the almost mechanical analogy operated between immigration and juvenile delinquency (2007), one must be wary of such conflations and avoid considering these clubs as posing a potential risk a priori. Indeed they have not been identified as such from a statistical point of view. One can nonetheless wonder about why the cases reporting incidents that might be even remotely related to the research topic are so exceptionally rare. The explanation lies in the double reluctance commonly seen in identifying and dealing with racist actions or comments: they are rarely recorded when they happen during matches, rarely punished by ethics commissions, and are protected by a sort of code of silence, a shared “omerta” that ranges from the green grass of weekend amateur soccer games to the quiet lounges of the Federation headquarters, in spite of the Ministry of Sports’ guidelines. As a historian studying violence in sports, one can only wonder about such a lack of regard for the issue, which invisibilizes a real threat to democracy (Cusset, 2018), as well as other forms of extremism, for which soccer stadiums can provide a crucible (Bodin, 2010) n © Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 133 Cahiers de la sécurité et de la justice – Hors-série Chauvaud, Frédéric, 2010, La dynamique de la violence, Presses Universitaires de Rennes. Chovaux, Olivier, 2012, « Quantifier les violences et les incivilités dans le sport amateur : le cas du football », Revue EPS, Volume 351, March/April. Chovaux, Olivier, 2013, « La prise en compte de l’ethnicité sportive dans la régulation des violences et incivilités : le cas du football amateur », International Journal on Violence and School, Volume 13. Cusset, François, 2018, le déchaînement du monde. Logique nouvelle de la violence, La Découverte. Delmotte, Florence, 2010, « Termes clés de la sociologie de Norbert Elias », Vingtième Siècle, n°106, pages 29-36.Deluermoz, Quentin, 2010, « Norbert Elias et le XXe siècle : le processus de civilisation à l’épreuve », Vingtième Siècle, Volume 106, April/June. Jobard, Fabien, 2005) « Déviances et modalités de contrôle. La France et l’Allemagne en perspective », Déviance et Société, n°3, Volume 29.Khosrokhavar, Farhad, 2014, Radicalisation, Maison des Sciences de l’Homme, collection « Itinéraires ». Mignon, Patrick, 1995, « La violence dans les stades. Supporters, ultras et hooligans ». Actes des entretiens de l’INSEP, Volume 10. Mucchielli, Laurent, 2002, Violences et Insécurité. Fantasmes et réalités dans le débat français, La Découverte, collection « Sur le vif ». Noiriel, Gérard, 2016, Le Creuset français. Histoire de l’immigration (xixe – xxe), Seuil, collection. « Points ».Nuytens, Williams, 2003, Étiologie des violences dans le football amateur. Enquêtes sociologiques dans dix clubs du département du Nord, DRDJS of the Nord-Pas-de-Calais, funded research report. Nuytens, Williams, 2016, « Violences à l’école : l’éducation par le sport peut-elle tout régler ? » in Attali Michaël, Les éducations par le sport, Editions Canopée.Pfister, Richard, 1994, « Les interactions agressives dans la pratique sportive des jeunes », Enfance, n°2/3, pages 215-232.Vieille-Marchiset, Gilles and Coignet, Benjamin, 2015, Clubs sportifs en Banlieue : des innovations sociales à l’épreuve du terrain, Presses Universitaires de Strasbourg.Official report and institutional documentsFrench Soccer Federation (F.F.F.), 2019, Archives.French Soccer Federation (F.F.F.), 2007, Federal Assembly of June 23rd, 2007.Charter of ethics and deontology in soccer, December 2017. Circular of April 29th, 2014 issued by the Ministry of Interior. Ministry of Higher Education, October 2019, Preventing radicalization in higher education and research organizations. Minutes of the ethics commission of the Hauts-de-France league, meeting of December 28th, 2016. Minutes of the ethics commission of the Flandre soccer district, meeting of November 16th, 2016.© Cahiers de la sécurité et de la justice - IHEMI134 I VARIARenseignement, délinquance et violences urbaines – Alexis DeprauRenseignement, délinquance et violences urbainesAlexis DEPRAULe terrorisme, le contre-espionnage ou la prolifération des armes de destruction massive sont des enjeux au cœur de l’action des services de renseignement. Si les questions de sécurité internationale ont concentré depuis quelques années la plupart des analyses, les services de renseignement ont aussi une mission très importante en matière de sécurité intérieure et de sécurité nationale. Si la lutte contre la délinquance est l’apanage des services de police, elle est aussi une mission du renseignement, notamment vis-à-vis des bandes de rue, des violences urbaines, et du phénomène de l’hybridation qui caractérise le lien entre délinquance et terrorisme.Mots-clés : Renseignement ; délin-quance ; violences urbaines ; bandes de rue ; terrorismeriorité stratégique du Livre blanc sur la défense et la sécurité nationale de 2013, et clairement exprimée dans l’article L. 811-3 du Code de la sécurité intérieure, « la prévention de la criminalité et de la délinquance organisées » est un axe de travail essentiel des services de renseignement français pour lequel il leur est possible de recourir aux techniques de renseignement définies dans la loi du 24 juillet 2015. Cette mission était principalement dévolue aux Renseignements généraux (R.G.) supprimés en 2008 par la création de la Direction centrale du renseignement intérieur (D.C.R.I. 1) et le remplacement des Renseignements généraux par la Sous-direction de l’information générale (SDIG). Depuis 2015, elle est assumée par le Service central du renseignement territorial (S.C.R.T.) qui succède partiellement à la SDIG 2. En effet, le travail judiciaire est assuré par les services de police. Mais il doit y avoir une prise en compte effective des phénomènes de délinquance organisée avec les bandes de rue, des violences urbaines, mais encore du phénomène d’hybridation par le renseignement territorial. Ces trois phénomènes sont liés (Gayraud, 2017), puisque les bandes qui participent aux violences urbaines sont celles qui participent aussi aux divers trafics, qui eux-mêmes peuvent apparaître comme des sources de financement du terrorisme.Tout l’enjeu de cette démonstration vise à mettre en avant la nécessité du travail du service de renseignement territorial pour mieux comprendre le fonctionnement de ces bandes et pour mieux lutter contre ces menaces criminelles. Il s’agit d’ailleurs de priorités abordées dans la Stratégie nationale du Alexis DEPRAU Docteur en droit public, spécialiste du droit de la sécurité et de la défense.(1) Devenue depuis la Direction générale de la sécurité intérieure (D.G.S.I.)(2) D. n°2016-466 du 9 mai 2014 modifiant le décret n°2008-633 du 27 juin 2008 modifié relatif à l’organisation déconcentrée de la direction centrale de la sécurité publique (D.C.S.P.), J.O.R.F., n°0108 du 10 mai 2014, texte n°24.P© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 135 Cahiers de la sécurité et de la justice – Hors-série renseignement de juillet 2019, un document nécessaire au Plan national d’orientation du renseignement (PNOR) pour fixer les axes et missions des services de renseignement, et notamment concernant la criminalité organisée 3.La surveillance des bandes de rueLes bandes de rue sont un phénomène traditionnel en France, mais elles ont évolué vers ce qu’on peut dorénavant appeler le néo-banditisme.Le phénomène traditionnel des bandes de ruePlus connus aux États-Unis sous le nom de street gangs qui comprend à la fois les gangs de rue et les organisations carcérales, les street gangs regroupent essentiellement en France des bandes de rue. En 2009, la Direction centrale de la sécurité publique (D.C.S.P.) définit la bande de rue comme : « un groupe composé au minimum de trois adolescents ou jeunes adultes. La structure de la bande peut varier mais elle comprend au moins un noyau stable de membres qui se considèrent ou sont considérés par les membres occasionnels comme étant une bande. Ils se regroupent pour des raisons qui peuvent être sociales, culturelles ou autres et commettent de façon désorganisée ou délibérée des actes antisociaux (incivilités), délictueux ou criminels ».Les bandes ont marqué différentes périodes de l’histoire contemporaine française : les « Apaches », les « loubards » dans les années 1950 et 1970, les « blousons noirs » des années 1960, les « bandes de zoulous », les « bandes de skinheads » suprématistes, les antifascistes « red skins », les « sauvageons », etc. Par exemple, les bandes de zoulous ont été observées dans les années 1990 dans les banlieues d’Île-de-France, composées de jeunes Africains. Ces bandes avaient comme principal mode de recrutement la mise à l’épreuve d’un « prétendant » pour entrer dans le groupe : un combat à main nue contre le chef du groupe et le viol d’une femme blanche, épreuves rituelles à franchir (Desrousseaux, 2014 : 89).Le caractère communautaire des bandes a été souligné en 2007 par le rapport de la Direction centrale des Renseignements généraux (actuel S.C.R.T.), qui rapporte le « danger de l’éventuelle fusion entre deux phénomènes a priori distincts de repli communautaire et d’activité délinquante d’une bande »4. C’est aussi la raison pour laquelle le renseignement territorial dispose d’un « département des dérives urbaines, du repli identitaire et du suivi des mouvances radicales »5, mettant d’ailleurs en avant le lien entre les dérives urbaines et la radicalisation.Si les bandes de rue doivent être considérées comme une atteinte potentielle à la sécurité nationale, c’est en raison de leur violence, de la criminalité organisée qui leur est liée, mais encore de leurs liens avec le phénomène du néo-banditisme qui émerge comme une nouvelle menace criminelle contemporaine.L’évolution vers le néo-banditismeL’évolution des bandes de rue tient au fait que « les services spécialisés constatent la dérive vers le crime organisé d’un nombre croissant de bandes délinquantes issues des quartiers dits sensibles » (Aubry, 2009 : 27). Ces bandes liées aux violences urbaines participent à une finalité criminelle et « s’inspirent des mafias (sans pouvoir en reproduire les aspects « familiaux » emblématiques) avec une hiérarchie pyramidale où chaque acteur va avoir un rôle bien défini autour d’un leader reconnu. Les plus jeunes serviront de guetteur. Les 16-22 ans auront la responsabilité de la revente directe ou de l’approvisionnement. Les plus vieux régneront en véritables caïds sur le réseau, en gérant notamment le blanchiment de l’argent issu du trafic à travers son investissement dans la création de commerces ou de petites PME » (Bauer et Soullez, 2010). Le phénomène des bandes de banlieues s’analyse donc aussi bien dans le cadre du trafic organisé que dans le cadre des diverses violences urbaines. Ce lien de cause à effet et d’interdépendance s’explique parce que le trafic devient très organisé et qu’il ne faut pas que les violences urbaines mettent à mal ce trafic. En effet, un « business » florissant demande calme et discrétion. Ainsi, les bandes redoublent d’intensité car les acteurs du trafic vont vouloir repousser hors du territoire celles et ceux qui pourraient remettre en cause ou perturber le trafic comme les forces de l’ordre, ou encore d’autres bandes rivales.Face à ce nouveau phénomène de criminalité, l’analyse est aussi effectuée par un service dépendant de la police judiciaire. Le Service d’information, de renseignement et d’analyse stratégique sur la criminalité organisée (3) Coordination nationale du renseignement et de la lutte contre le terrorisme, Stratégie nationale du renseignement, juillet 2019 : 7.(4) http://www.lemonde.fr/a-la-une/article/2007/09/05/les-bandes-sous-la-loupe-des-rg_951415_3208.html(5) Arr. du 1er février 2011 relatif aux missions et à l’organisation de la direction centrale de la sécurité publique (D.C.S.P.), J.O.R.F.RF, n°30, 5 février 2011, texte n°29, art. 5.© Cahiers de la sécurité et de la justice - IHEMI136 I DOSSIER Renseignement, délinquance et violences urbaines – Alexis Deprau(SIRASCO) a qualifié ces nouvelles entités de « néo-banditisme » qu’il décrit comme « un banditisme nouveau issu des cités sensibles, se différenciant d’un « milieu » traditionnel fort affaibli par le succès des investigations policières, et un changement de générations. Il recouvre des groupes criminels organisés divers mais tous structurés autour du trafic de stupéfiants. Leur particularité est de privilégier les « circuits courts » au sein du réseau, limitant les intermédiaires, et conférant un sentiment de puissance démesuré à de petits malfaiteurs n’en ayant pas l’envergure » (Dufour et Kabssi, 2015 : 143-144).L’activité principale des bandes de rue est le trafic de stupéfiants à l’image de « Shitland » : une cité de Champigny-sur-Marne qui était devenue avant l’opération de police le second point de vente le plus important de région parisienne, où 150 kg de cannabis avaient été saisis 6. Outre le trafic de stupéfiants, les bandes de rue semblent aussi participer au trafic d’armes. À cet égard, « un phénomène inquiétant consiste dans la saisie, désormais non exceptionnelle, d’armes de guerre lors des perquisitions menées dans les cités. Au vu des stocks découverts, il semble facile aux membres du grand banditisme, mais surtout du « banditisme issu des cités » d’acquérir de tels types d’armes peu onéreuses (un pistolet-mitrailleur AK-47 se négocie entre 500 euros et 800 euros) » (Pradel et Dallest, 2012 : 37).La prise en compte des violences urbaines par le renseignement territorialLa problématique des violences urbaines représente une atteinte potentielle à la sécurité nationale. C’est la raison pour laquelle la lutte contre les violences urbaines est effectuée par le S.C.R.T.Un phénomène concourant à l’instabilité intérieureLes événements de Vaulx-en-Velin (Rhône) en 1979 sont considérés comme les premières violences urbaines, impliquant l’incendie de voitures et des affrontements avec la police (Bauer et Soullez, 2010 : 12-13). Ces violences urbaines ne bénéficient pas encore de définition officielle. Pour autant, on peut retenir la proposition des R.G. voyant les violences urbaines « comme des actes juvéniles collectifs commis de manière ouverte et provocatrice et créant dans la population un fort sentiment d’insécurité » (Janet, 2012 : 11). Par ailleurs, les violences urbaines ne sont pas non plus associées à la qualification juridique d’attroupements. À Villiers-le-Bel (Val-d’Oise), en novembre 2007, deux jeunes à moto meurent dans une collision avec une voiture de police, ce qui déclenche une vague de violence. Ici, pour demander le remboursement par l’État des frais liés aux dégâts, au regard de l’article L. 2216-3 du Code général des collectivités territoriales, les violences urbaines devaient être considérées comme des attroupements. Le conseil d’État a censuré l’arrêt de la Cour administrative d’appel, en considérant que les violences urbaines ne sont pas des attroupements, au motif « que cet incendie avait été provoqué par des personnes qui étaient au nombre de celles qui s’étaient spontanément rassemblées, peu de temps auparavant, pour manifester leur émotion après le décès des deux adolescents et que, par ailleurs, l’attaque du restaurant était sans rapport avec cette manifestation » (CE, 30 décembre, Société Covea risks, n°386536, consid. 3).Quoi qu’il en soit, ces bandes sont réelles et les différents affrontements augmentent régulièrement à partir des années 1990 7. Les bandes perdurent encore aujourd’hui, comme le montrent les incendies de voitures lors de la Saint-Sylvestre 8, ou les affrontements entre « bandes rivales » 9. Si les violences urbaines se concentrent sur l’appartenance territoriale des membres des bandes, une évolution s’est faite dans le cadre des différentes manifestations quand un mouvement de revendication sociale sert de prétexte pour les attaques et pillages de magasins. En effet, « on assiste à un cumul de trois types d’opérations surfant sur les manifestations : contre les forces de l’ordre et les bâtiments publics, contre les magasins et contre les manifestants, ce qui fut particulièrement visible lors des manifestations contre le contrat première embauche (CPE) » (Bauer et Soullez, 2010 : 70).Les violences commises par les bandes montrent, de surcroît, une revendication identitaire dans le cadre des manifestations : ces bandes profiteraient de la foule des (6) http://www.lefigaro.fr/actualite-france/2013/04/02/01016-20130402ARTFIG00037-debut-du-proces-de-shitland-a-creteil.php(7) https://www.lefigaro.fr/actualite/2007/09/07/01001-20070907ARTFIG90045-affrontements_entre_bandes_en_hausse_selon_les_rg.php(8) http://www.lefigaro.fr/flash-actu/2016/01/01/97001-20160101FILWWW00125-nouvel-an-en-france-804-vehicules-incendies-en-baisse-de-145.php(9) https://www.20minutes.fr/faits_divers/2361067-20181025-bandes-rivales-place-jeunes-rue-2-heures-matin© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 137 Cahiers de la sécurité et de la justice – Hors-série manifestations pour agresser et voler des victimes ciblées (Pellegrini, 2005 : 74-75) 10 mais également les manifestants eux-mêmes (Bauer et Soullez, 2010 : 70) 11. À la différence des autres mouvements contestataires qui peuvent avoir des revendications politiques ou religieuses, les violences urbaines sont liées à une forme d’irrationalité, puisque ce sont des actes montrant un mépris des institutions sans aucune revendication quelle qu’elle soit.La lutte contre les violences urbaines effectuée par le Service central du renseignement territorialLe renseignement territorial surveille les banlieues depuis le milieu des années 1970. Cette surveillance a été faite à l’origine dans le but d’évaluer la menace des groupes radicaux au sein de la nouvelle population immigrée qui s’y est installée, sans pour autant que des mesures politiques soient prises à la suite des informations recueillies. Pourtant, « dès 1975, de nombreuses informations remontent des travailleurs sociaux, démontrant qu’au cours des années, les risques se sont amplifiés. Typiques du mode de travail spécifique aux R.G., ces données ne sont pas évaluées à leur juste mesure » (Madelin, 2007 : 246). Parallèlement, le travail de surveillance des extrémistes islamistes a permis d’évaluer non seulement les risques possibles sur leur influence à l’égard des jeunes de banlieue, mais aussi ce que pourraient devenir ces quartiers à l’avenir. Enfin, ce renseignement a permis d’établir un lien, à propos des dérives urbaines, entre la petite délinquance et l’extrémisme radical. Plus spécifiquement, «les informations rassemblées retracent avec précision le rôle des prêcheurs salafistes dans les banlieues, ainsi que leur passé, les raisons qui les ont conduits en France. Elles permettent d’imaginer les dérives induites, tant au plan religieux qu’à celui de la petite à la moyenne délinquance astucieuse, le trabendo et le petit trafic de drogue. Avec en perspective le développement rapide et très extensif des zones de non-droit. Ainsi, dès le début des années 1990, on sait tout » (idem : 247).À la fin de l’année 1991, la dixième section des Renseignements généraux, ou section « Villes et banlieues » (appelée encore section des Violences urbaines), eut pour mission première de se focaliser sur les violences urbaines (Rosière, 2011 : 95). C’est un terme générique assez large concernant à la fois les violences commises par les skinheads, dans les banlieues, les violences des bandes, mais encore la surveillance des activités dans les rave parties (Caumer, 2000 : 90). C’est dans le cadre de ce service que la commissaire Lucienne Bui Trong crée une échelle de huit de degrés de violences urbaines allant des simples feux de poubelles à l’émeute urbaine (Bui Trong, 2000 : 63-72) :1. feux de poubelles et de voitures ;2. harcèlement de l’autorité ;3. violence à l’égard de toute personne portant un uniforme (policiers, pompiers) ;4. attroupements contre la police et le « caillassage » des voitures de police ;5. attroupements entravant l’action des policiers ou faits dans le but de récupérer des amis interpellés ;6. intention de blesser volontairement les policiers ou d’attaquer des commissariats ;7. « mini-émeute » sans lendemain appelant néanmoins à une escalade rapide de la violence ;8. émeute à proprement parler qui se produira sur plusieurs lieux et pendant plusieurs nuits.Au regard de cette échelle, les heurts observés dans toutes les manifestations contre la loi Travail à Paris 12 ou en province 13, peuvent être classés au septième degré de l’échelle des violences urbaines. Il n’est pas exagéré de dire que cette situation est similaire pour les heurts qui ont eu lieu en marge des manifestations des Gilets jaunes, avec les différentes exactions commises par des blacks blocs, ou les pillages de l’Arc de triomphe et de différents magasins (10) Par exemple, lors la manifestation du 8 mars 2003 qui a causé plusieurs dizaines de blessés : « des témoins ont entendu à plusieurs reprises des propos tels que : « On va casser des petits Blancs. » Certains lycéens sont les victimes de passage à tabac à dix contre un ; ils reçoivent des coups de poing, des coups de pieds et des coups de bâton, les filles sont traînées par les cheveux. Les insultes racistes fusent : « sale Blanc » est le maître mot de ces lynchages organisés. […] Des dizaines d’adolescents sont conduits dans les hôpitaux. Après enquête, on apprend que la plupart des agresseurs venaient de la Seine-Saint-Denis et des arrondissements du nord de Paris ».(11) Ainsi « à partir de 2005, les mêmes délinquants s’attaquent aussi aux manifestants eux-mêmes. De nombreux manifestants sont agressés, avec comme nouveauté, des agressions visant spécifiquement de jeunes lycéens « blancs » » (Bauer et Soullez, 2010 : 70.)(12) http://www.lexpress.fr/actualite/societe/en-images-loi-travail-paves-molotov-lacrymo-heurts-a-la-manif-parisienne_1791491.html(13) Les violences urbaines dans le cadre des manifestations contre la loi Travail sont très fortes à Rennes, in http://www.ladepeche.fr/article/2016/05/15/2344913-rennes-face-au-dechainement-de-la-violence-urbaine.html© Cahiers de la sécurité et de la justice - IHEMI138 I DOSSIER Renseignement, délinquance et violences urbaines – Alexis Depraupar des bandes de rue, etc. Ces violences urbaines se répercutent aussi des matchs de football de l’Algérie dans les grandes agglomérations françaises 14.Structurellement, la lutte contre les violences urbaines devient (et reste) une des missions principales des R.G. avec la circulaire du 3 janvier 1995. En effet, à côté des trois grandes missions d’intérêt national 15, les violences urbaines ont été regroupées avec d’autres missions de surveillance, n’ayant aucun lien avec les violences urbaines. En 1995, les violences urbaines n’étaient pas encore assimilées à un domaine nécessitant un service totalement dédié. Plus précisément, les violences urbaines ont été accolées aux « questions liées à l’exclusion (problème des squats et des différentes formes de marginalité), le phénomène des sectes ainsi que les répercussions sur l’opinion publique des faits de société à grande résonance médiatique (exemple du sang contaminé, des affaires judiciaires défrayant la chronique, etc.) » (Zamponi, 1997 : 208).Puis en 1999, la D.C.S.P. en collaboration avec les R.G., élabora une nouvelle base de données, plus pour des raisons de rivalités entre services que de cohérence. En conséquence, « l’échelle Bui Trong fut aussi écartée tant pour des rivalités internes entre les R.G. et la sécurité publique que par la volonté des responsables politiques qui ne voyaient pas d’un très bon œil la température monter annuellement » (Bauer et Soullez, 2010 : 24-25). Appelée Système d’analyse informatique des violences urbaines (SAIVU), cette base avait pour but de, répertorier tous les faits en liens avec les violences urbaines. Mais cet outil fut lui aussi remplacé en 2005, par l’Indicateur national des violences urbaines (INVU), dans le cadre d’un travail de concertation entre les différentes directions de la police, la gendarmerie nationale ainsi que la préfecture de police. Mais à partir de 2010, sur les neuf index statistiques de cet outil de données, seuls deux étaient encore utilisés. Sur l’analyse des violences urbaines, le rapport des R.G. sur les émeutes de novembre 2005 apporta un nouvel éclairage, en niant toute implication de quelque groupe que ce soit, d’ordre islamiste, mafieux ou politique 16. Si les bases de données ne sont pas utilisées à bon escient, les rapports du service de renseignement policier territorial montrent d’un autre côté, la nécessité de leur existence et de leur présence, pour effectuer une analyse pertinente des phénomènes de société.Puis le 27 juin 2008, la section Villes et banlieues devint ensuite une « division des dérives urbaines et du repli identitaire » 17, et se maintint avec l’arrêté du 1er février 2011 18. Avec la réforme intervenue par le décret du 9 mai 2014, le nouveau S.C.R.T. a toujours en son sein cette « division des dérives urbaines et du repli identitaire » 19. Par ces missions qui leur sont confiées, les services départementaux du renseignement territorial (SDRT) font remonter l’information au niveau central en assurant un maillage territorial nécessaire à la récolte du renseignement intérieur. Ces services départementaux doivent « contribuer à déterminer les sites où sont constitués les bandes, définir leur structuration, leur comportement délictuel et identifier leurs membres » (idem : 115).C’est en raison du danger qu’elles représentent que les violences urbaines, appelées aussi subversions violentes, ont été prises en compte par le document sur la Stratégie nationale du renseignement de juillet 2019, où « la radicalisation de ces modes d’action appelle à une vigilance accrue des services de renseignement dans leur fonction d’anticipation et de défense de l’État pour prévenir les violences de toute nature et la déstabilisation de nos institutions » 20.Une nécessaire prise en compte de l’hybridation de la criminalité organisée vers le terrorismeLa structure des organisations criminelles vise à réaliser trois objectifs : l’accumulation de pouvoir, la conquête (14) https://www.francetvinfo.fr/sports/foot/coupe-d-afrique-des-nations-apres-la-victoire-de-l-algerie-des-incidents-a-marseille-paris-et-lyon_3536765.html(15) Les grandes missions d’intérêt national sont : le contrôle des activités des groupes islamistes, l’évolution du climat social, et l’analyse et la recherche financière, notamment pour ce qui concerne les flux de capitaux d’origine douteuse, la lutte contre la criminalité organisée et la lutte contre la corruption.(16) http://www.lemonde.fr/societe/article/2005/12/07/selon-les-rg-les-emeutes-en-banlieue-n-etaient-pas-le-fait-de-bandes-organisees_718347_3224.html(17) Arr. du 27 juin 2008 relatif aux missions et à l’organisation de la direction centrale de la sécurité publique (D.C.S.P.), J.O.R.F.RF, n°152, 1er juillet 2008, texte n°10, art. 7.(18) Arr. du 1er février 2011 relatif aux missions et à l’organisation de la direction centrale de la sécurité publique (D.C.S.P.), J.O.R.F.RF, n°30, 5 février 2011, texte n°29.(19) Arr. du 9 mai 2014 modifiant l’arr. du 1er février 2011 modifié relatif aux missions et à l’organisation de la direction centrale de la sécurité publique (D.C.S.P.), J.O.R.F.RF, n°108, 10 mai 2014, texte n°29, art. 4.(20) Coordination nationale du renseignement et de la lutte contre le terrorisme, Stratégie nationale du renseignement, juillet 2019 : 4.© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 139 Cahiers de la sécurité et de la justice – Hors-série d’un « territoire » urbain, l’accumulation de richesses. Cela n’empêche pas une « alliance entre le terrorisme international et le crime organisé pourrait se révéler un problème sérieux à l’avenir » (Bühler, 2007 : 37). Pour des raisons de pouvoir et de richesse, mais aussi par filiation intellectuelle ou religieuse, la collusion entre les organisations criminelles et les groupes terroristes est un phénomène réel. Par exemple, citons le lien entre la mafia albanophone et l’armée de libération du Kosovo (UÇK). La permanence de la liaison entre organisations criminelles et terrorisme (voire avec les guérillas), s’observe avec la livraison d’armes, en échange d’argent et souvent de drogue : « ces procédés se sont généralisés dans les Balkans, idéalement placés sur la route entre les zones de production d’opium et l’Europe : les organisations locales peuvent traiter avec les organisations turques en amont et italiennes en aval » (Chocquet, 2003 : 45).En France, le lien entre terrorisme et banditisme est aussi présent, comme l’a montré le cas du gang de Roubaix appelé aussi groupe des « islamo-braqueurs » 21. Ce « gangsterrorisme » du milieu des années 1990 était caractérisé par un groupe participant au grand banditisme par ses braquages, mais était aussi composé de convertis à l’islam partis se battre en Bosnie du côté des islamistes et acquis à la cause terroriste islamiste en la finançant par des braquages 22. Cette hybridation entre criminalité organisée et terrorisme s’observe enfin en France avec le néo-banditisme issu des banlieues, et son lien progressif et enraciné avec ce qu’il est permis d’appeler « néo-terrorisme ». Ce néo-terrorisme « est l’œuvre de jeunes hommes et de jeunes femmes ayant grandi dans le même environnement, mais ayant choisi à un moment de basculer dans la voie du fanatisme. Elle mêle à ce rapport débridé à la violence l’engagement de jeunes malfaiteurs un peu paumés, voyant dans l’action armée une exaltation, un rôle à jouer que ne leur confère pas la société. La guerre sainte devient alors une raison de vivre, au même titre que peut l’être la quête de l’argent et du pouvoir engendrée notamment par le trafic de drogues » (Dufour et Kabssi, 2015 : 176).Né il y a plus de vingt ans en France, ce phénomène n’a pas fait l’objet réellement d’analyse et de synthèse sérieuse par les services de renseignement et les autorités politiques. Il faut néanmoins reconnaître le fait que depuis novembre 1993, le ministère des Affaires étrangères a créé une Sous-direction de la sécurité, dont la mission est de traiter les « questions relatives à la lutte contre le terrorisme, le trafic de stupéfiants et la criminalité internationale [...] en liaison avec les autres départements ministériels concernés » (Wodka-Gallien, 1995 : 17). En tout état de cause, il est important que la prospective devienne une priorité. Certes, a été lancée la cellule de profilage des auteurs d’actes terroristes, une telle cellule serait la bienvenue pour l’hybridation, ce qui permettrait d’éviter des impairs : « Khaled Kelkal en 1995, le gang de Roubaix en 1996 : la France connaît depuis vingt ans ces hybrides, mi-gangsters, mi-terroristes, naviguant entre divers fichiers de police et échappant ainsi à des services empêtrés dans leurs prés carrés. Mohamed Merah valide ainsi en 2012 le processus décrit en 2006 par Mitch Silber et l’auteur dans leur étude (la Radicalisation en Occident, la menace enracinée), faite par et pour la police de New York. Les frères Kouachi et Amédy Coulibaly confortent l’analyse. Désormais, la menace provient d’entités fugaces et quasi-protoplasmiques. Un continuum criminalo-terroriste émerge, loin des petits casiers doctement préparés pour chacune de ses composantes d’origine » (Bauer, 2015 : 3).Ainsi, la lutte contre le néo-banditisme et le néo-terrorisme doivent être une mission prioritaire de surveillance, dans la mesure où les bandes participent à la criminalité organisée, avec des membres pouvant participer aux violences urbaines, mais être aussi affiliés à des islamistes radicaux, voire des organisations terroristes. De telle sorte que, selon les estimations de la Direction centrale de la police judiciaire, « près de 40% des suspects apparus ces dernières années dans ce type d’affaires ont un passé délinquant » (Dubois et Pelletier, 2017 : 87). Au regard du phénomène d’hybridation, la distinction doit être atténuée, et la délinquance doit même être considérée comme étant liée directement au terrorisme (Gayraud, 2017). En effet, la plus grande majorité des terroristes qui ont commis des attaques ont été condamnés pour des faits liés à la délinquance et la criminalité organisées (Raufer, 2017 :24). Le travail d’analyse fourni par le SIRASCO doit être entendu, et ce service doit voir ses moyens augmentés pour continuer son travail concernant l’hybridation. À plus forte raison, le renseignement intérieur doit impérativement prendre en compte le phénomène de l’hybridation, ce qui est le cas aujourd’hui, puisque la Stratégie nationale du renseignement de 2019, indique que la « criminalité facilite le terrorisme, y compris en Europe, en affaiblissant le niveau de contrôle sécuritaire de certaines zones géographiques et en fournissant des moyens de financement ou d’approvisionnement en armes, faux documents et matériels » 23.(21) https://www.lepoint.fr/societe/il-y-a-20-ans-le-raid-lancait-son-assaut-contre-les-islamo-braqueurs-du-gang-de-roubaix-28-03-2016-2028385_23.php(22) http://www.rfi.fr/france/20160328-france-gang-roubaix-islamo-braqueurs-terrorisme-caze-dumont-bosnie(23) Coordination nationale du renseignement et de la lutte contre le terrorisme, Stratégie national du renseignement, juillet 2019 : 7.© Cahiers de la sécurité et de la justice - IHEMI140 I DOSSIER Renseignement, délinquance et violences urbaines – Alexis DeprauCertes, la réforme de 2008, qui a remplacé les R.G. par la Sous-direction de l’information générale (SDIG), a pu créer une perte de maillage du renseignement territorial liée à un transfert massif de ces policiers au sein de la D.C.R.I. (actuelle D.G.S.I.) Cependant, l’actuel S.C.R.T. n’en reste pas moins une structure importante et nécessaire, malgré les contrecoups de cette réforme de 2008. Ce n’est qu’avec la nouvelle réforme de 2014 remplaçant la Sous-direction de l’information générale (SDIG) par le S.C.R.T. qu’a été reconstitué le maillage territorial, complété par ce qui est essentiel pour la récolte de l’information, à savoir le renseignement de source ouverte (OSINT 23) ou renseignement d’information publique (presse, radio, télévision, sites Internet).Rattaché à la D.C.S.P., le renseignement territorial et opérationnel permet de mettre en avant les enjeux actuels et les menaces liés aux phénomènes de société. Le renseignement territorial a ainsi permis de soulever la question épineuse des bandes de rue, du commerce souterrain, et plus encore des violences urbaines, sur fond de liaison avec les islamistes radicaux. Il est important dorénavant que toutes les alertes lancées par ce service policier et de renseignement soient bien admises par les décideurs pour affronter au mieux les menaces pouvant porter atteinte à la sécurité intérieure, et à plus forte raison à la sécurité nationale n(24) Acronyme anglais signifiant open source intelligenceBibliographieOuvragesBauer, Alain et Christophe Soullez, 2010, Violences et insécurité urbaines, Paris, PUF, coll. « Que sais-je ? ». (12e éd.)Bui Trong, Lucienne, 2000, Violences urbaines. Des vérités qui dérangent, Paris, Bayard.Caumer, Julien, 2000, Leurs dossiers R.G., Paris, Flammarion, 2000.Chocquet, Christian, 2003, Terrorisme et criminalité organisée, Paris, L’Harmattan.Dubois, Christophe et Éric Pelletier, 2017, Où sont passés nos espions ?, Paris, Albin Michel.Dufour, Julien et Abdelfettah Kabssi, 2015, Bandes, dérive criminelle et terrorisme, Paris, MA Éd., 2015.Gayraud, Jean-François, 2017, Théorie des hybrides. Terrorisme et criminalité organisée, Paris, CNRS Éditions.Madelin, Philippe, 2007, Dans le secret des services, Paris, Denoël.Pelligrini, Charles, Banlieues en flammes, Paris, Éd. Anne Carrière.Pradel, Jean et Jacques Dallest, 2012, La criminalité organisée. Droit français, droit international et droit comparé, Paris, Lexis Nexis.Rosière, Stéphane, 2011, Géographie des conflits armés et des violences politiques, Paris, Ellipses.Zamponi, Francis, 1997, Les R.G. à l’écoute de la France. Police et politique de 1981 à 1997, Paris, La Découverte et Syros.RevuesAubry, Gilles, 2009, « Organisations criminelles et structures répressives : panorama français », Cahiers de la sécurité, n°7, pp. 25-40, janvier-mars.Bauer, Alain, 2015, « Qui est l’ennemi ? », Conflits, supplément au n°5, avril-mai-juin.Bühler, B.O., 2007, « L’alliance du crime organisé et du terrorisme. Un nouveau défi pour le futur ? », Défense & Sécurité internationale, n°30, pp.36-37.Raufer, Xavier, 2017, « Les hybrides (terroristes+criminels) – La police et le renseignement en Europe », Atlantico,mars.Wodka-Gallien, Philippe, 1995, « Renseignement et diplomatie : les voies de la coopération », Enjeux atlantiques, n°11, avril.Travaux universitairesDesrousseaux, Thomas, 2014, Les bandes organisées en milieu urbain aujourd’hui, thèse, Université Panthéon-Assas Paris II.Janet, Mathieu, 2012, Sécurité publique et violences urbaines, mémoire, Université Paris II.© Cahiers de la sécurité et de la justice - IHEMIVARIA I 141Cahiers de la sécurité et de la justice – Hors-série Intelligence, delinquency and urban violenceAlexis DEPRAUTerrorism, counterintelligence, the proliferation of weapons of mass destruction: these issues are at the heart of intelligence operations. Although international security has been the main focus of the last few years, intelligence services have also a very important mission in terms of homeland security and national security. Fighting delinquency is the prerogative of police, but it is also an intelligence mission, especially with regards to street gangs, urban violence, and the phenomenon of hybridization observed in the links between delinquency and terrorism.Keywords: Intelligence; delinquency; urban violence; street gangs; terrorismstablished as a strategic priority in the 2013 French White Paper on Defence and National Security, and clearly defined in Article L. 811-3 of the Code of Homeland Security (CSI), “the prevention of organized crime and delinquency” is a crucial part of the French intelligence services’ mission, for which they can resort to the intelligence techniques described in the law of July 24th, 2015. That mission was mostly attributed to the R.G. (Renseignements généraux, or General Intelligence Services), until they were replaced in 2008 by the D.C.R.I. (Direction centrale du renseignement intérieur, or Central Directorate of Interior Intelligence 1) and the SDIG (Sous-Direction de l’information générale, or Sub-Directorate of General Information). Since 2015, it’s been taken over by the S.C.R.T. (Service central du renseignement territorial, or Central Service of Territorial Intelligence), which partially replaced the SDIG. 2 Indeed, the judicial aspect is covered by the police and justice systems. But the phenomenons of organized delinquency, street gangs, urban violence as well as the hybridization of terrorism and delinquency must be taken into account effectively by territorial intelligence services. These phenomena are all connected (Gayraud, 2017), as the gangs that are involved in urban violence are the same ones that are involved in all sorts of trafficking, which in turn can be connected to terrorism as they contribute to its funding.The whole purpose of this demonstration will be to highlight the necessity of this mission of the territorial intelligence services in order to better understand the way these gangs operate, and to better counter the criminal threat. Alexis DEPRAUDoctor in public law, specialized in security and defence law.(1) Which in turn became in 2014 the Direction générale de la sécurité intérieure (D.G.S.I., or General Directorate for Internal Security).(2) Decree number 2016-466 of May 9th, 2014 modifying the decree number 2008-633 of June 27th, 2008, modified, regarding the delegated organization of the Central Directorate for Public Security (D.C.S.P.), JORF, number 0108 of May 10th, 2014, text number 24.E© Cahiers de la sécurité et de la justice - IHEMI142 I DOSSIER Intelligence, delinquency and urban violence – Alexis DeprauThese objectives are defined as priorities in the National Intelligence Strategy policy roadmap of July 2019, an essential document in the National Orientation Plan for Intelligence (PNOR) that aims at determining the goals and missions of intelligence services, including the surveillance of organized crime. 3The surveillance of street gangsStreet gangs are a traditional phenomenon in France, but they have evolved to become what can now be called “neo-gangsterism.”The traditional phenomenon of street gangsBetter known in the United States, where they include both street gangs and jail organizations, in France street gangs are essentially circumscribed to the street. In 2009, the D.C.S.P. (Central Directorate for Public Security) defined street gangs as “groups composed of at least three teenagers or young adults. The structure of the gang may vary, but it consists of at least a stable core of members who consider themselves or are considered by the other occasional members, as forming a gang. They gather for social, cultural or other reasons and commit, in deliberate or disorganized ways, criminal, delinquent or antisocial actions (incivilities).”Gangs have been present in various periods of contemporary French history: the Apaches and the loubards in the 1950s and 1970s, the blousons noirs (“black jackets”) of the 1960s, the Zoulous (“Zulu”) gangs, the white supremacist skinheads and anti-fascist redskins, the sauvageons (“little savages”), etc. For example, the Zoulous gangs, composed of young Africans, were active in the 1990s in the Paris suburbs. These gangs’ main mode of recruitment was for the aspirant member to prove his worth through an unarmed fistfight with the gang leader and the rape of a white woman, as initiation rites and trials by ordeal (Desrousseaux, 2014: 89).The ethnic nature of gangs was noted in 2007 in the report of the D.C.R.G. (Central Directorate of General Intelligence, now replaced by the S.C.R.T.), which highlights the “danger of a possible fusion between two initially distinct phenomena: delinquent gang activity, and the ethnic and religious self-segregation and communitarian withdrawal.” 4 This is why there is a “department of urban excesses, identity-based withdrawal, and monitoring of radical movements” 5 within the territorial intelligence services, which also establishes a link between urban violence and radicalization.If street gangs must be considered a potential risk for national security, it is on account of their violence, the organized crime they are connected to, as well as their links with the phenomenon of neo-gangsterism emerging as a new contemporary criminal threat. The shift towards neo-gangsterismThe evolution of street gangs is reflected in “the drift towards organized crime, observed by special services, of an increasing number of delinquent gangs from so-called ‘sensitive’ neighborhoods.” (Aubry, 2009: 27.) These gangs, related to urban violence, participate in criminal activities and “take inspiration from the mafia (although they are not able to reproduce its typical ‘family’ aspects), with a pyramidal hierarchy in which each player has a definite role, organized around a recognized leader. The youngest ones serve as lookouts. Those aged 16 to 22 are in charge of direct sale or supply. The older ones rule over the network as the real “qaids”, or kingpins, as they manage, among other things, the money laundering operations, investing the money made from the traffic in small businesses or companies” (ibidem). The phenomenon of suburban gangs is to be analyzed both in the context of organized traffics and in the general context of urban violence. This causal link and interconnectedness between the two is explained by the fact that, as the traffic becomes increasingly organized, urban violence becomes undesirable as it could get in the way of the traffic. Indeed, for the “business” to flourish, it requires peace and discretion. Therefore, gang activity increases in intensity, as the actors of the traffic try to push out of their territory anyone who might endanger or disturb their traffic, such as law enforcement or other rival gangs.Facing this new criminal phenomenon, a service of the judiciary police also provides its analysis. The Service (3) Coordination Nationale du Renseignement et de la Lutte contre le Terrorisme (CNRLT, National Coordination of Intelligence and Counter-Terrorism), Stratégie nationale du renseignement (National Intelligence Strategy), July 2019, p. 7.(4) http://www.lemonde.fr/a-la-une/article/2007/09/05/les-bandes-sous-la-loupe-des-rg_951415_3208.html(5) Ruling of February 1st, 2011 regarding the missions and the organization of the Central Directorate for Public Security (D.C.S.P.), JORF, number 30, February 5th, 2011, text number 29, article 5.© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 143 Cahiers de la sécurité et de la justice – Hors-série of Information, Intelligence and Strategic Analysis on Organized Crime (SIRASCO) qualified these new entities as “neo-gangsterism,” defined as “a new mafia, coming from sensitive urban projects, distinct from the traditional ‘mob’, much weakened by the success of police investigations and the new generation. It covers various organized criminal groups, all of them structured around drug traffic. Their specificity is that they favor ‘local distribution channels’ within the network, limiting the number of intermediaries, which gives an exaggerated feeling of omnipotence to small-time hoodlums who don’t have the calibre to justify it” (Dufour et Kabssi, 2015: 143-144)kThe main activity of street gangs is drug traffic, as for example in “Shitland” (meaning Hashish Land in French slang): an urban project in the Parisian suburb of Champigny-sur-Marne, which had become the second biggest drug dealing spot in the whole Paris metropolitan area, until it was raided by police and 150 kg (330 lbs) of cannabis were seized. 6 Besides drug traffic, street gangs also seem to be participating in gun trafficking. On this issue, “a preoccupying phenomenon is the increasingly frequent seizure of war weapons during police searches in urban projects, which is no longer an exceptional occurrence. In view of the considerable amounts that are found, it seems easy for gang members, and especially for the members of ‘urban projects gangsterism’ to purchase such weapons at a low cost (an AK-47 machine gun can be acquired for 500 to 800 euros).” (Pradel et Dallest, 2012: 37).The handling of urban violence by territorial intelligence servicesThe urban violence issue represents a potential risk for national security. That is the reason why it is the Central Service of Territorial Intelligence’s mission to counter them. A phenomenon that contributes to homeland instabilityThe 1979 events of Vaulx-en-Velin (in the suburbs of Lyon) are considered to be the first occurrence of urban violence, identified as such by the car fires and the clashes with the police (Bauer et Soullez, 2010: 12-13). This kind of urban violence still doesn’t have an official definition. However, one may adopt the suggestion of the R.G., to define urban violence “as collective juvenile actions, committed openly and provocatively, that cause a deep feeling of insecurity within the population” (Janet, 2012: 11). Besides, urban violence does not fully overlap with the legal characterization of unlawful assembly. In Villiers-le-Bel (in the suburbs of Paris), in November of 2007, two young boys on a motorcycle died in a collision with a police car, which triggered a wave of violence. In that case, the costs related to the material damage that resulted from the riots could only be covered by the State if the urban violence was considered as deriving from an unlawful assembly, according to article L. 2216-3 of the General Code of Territorial Subdivisions (CGCT). The Council of State (the French Supreme Court for administrative justice) quashed the ruling of the Administrative Court of Appeal, considering that this case of urban violence did not qualify as unlawful assembly, as “the fire had been caused by the same number of people who had spontaneously gathered, shortly before that, to manifest their emotion after the death of the two teenagers, and furthermore, the attack on the restaurant was unrelated to that manifestation” (Council of State, December 30th, Société Covea risks, Nr386536, consid. 3.)Be that as it may, these gangs are real, and the frequency of clashes began to increase progressively from the 1990s 7, until the riots of November 2005. The gangs have persisted until today, as evidenced by the number of car fires that take place each New Year’s Eve, 8 and the clashes between “rival gangs.” 9 Violent clashes are often related to the territorial affiliation of the gang members, however, an evolution was noted during several protest marches, when movements for social change become an excuse for attacking and looting stores. Indeed, “there has been an accumulation of three types of actions, with groups taking advantage of the context of protest marches in order to attack firstly law enforcement and public buildings, secondly stores, and thirdly the protesters themselves, which was made particularly visible during the movement against the CPE (First Employment Contract) labor reform” (Bauer et Soullez, 2010: 70).Furthermore, the acts of violence committed by gangs within protest marches demonstrate an ethnic identity dimension: it seems that these gangs take advantage of (6) http://www.lefigaro.fr/actualite-france/2013/04/02/01016-20130402ARTFIG00037-debut-du-proces-de-shitland-a-creteil.php(7) https://www.lefigaro.fr/actualite/2007/09/07/01001-20070907ARTFIG90045-affrontements_entre_bandes_en_hausse_selon_les_rg.php(8) http://www.lefigaro.fr/flash-actu/2016/01/01/97001-20160101FILWWW00125-nouvel-an-en-france-804-vehicules-incendies-en-baisse-de-145.php(9) https://www.20minutes.fr/faits_divers/2361067-20181025-bandes-rivales-place-jeunes-rue-2-heures-matin© Cahiers de la sécurité et de la justice - IHEMI144 I DOSSIER Intelligence, delinquency and urban violence – Alexis Deprauthe crowd present in protest marches in order to assault and rob targeted victims (Pellegrini, 2005: 74-75) 10 but also the protesters themselves (Bauer et Soullez, 2010: 70) 11 Unlike other protest movements which might have religious or political claims, urban violence seem to stem from irrational motives, as its actions completely disregard and despise the institutions.How the Central Service of Territorial Intelligence works to counter urban violenceTerritorial intelligence services have been surveilling the suburbs since the mid 1970s. That monitoring initially intended to assess the possible threat of radical groups within the new immigrant population that had settled there, but the information gathered didn’t result in political measures. However, “as early as 1975, a great quantity of information had been collected by social workers, which evidenced that through the years, the risks had increased. But this data was not given a proportionate response, which is typical of the way the R.G. worked” (Madelin, 2007: 246). On the other hand, the surveillance of islamist extremists allowed to estimate not only the possible risks that could arise if they were to influence the suburban youths, but also what these sensitive neighborhoods and urban projects might become in the following years. Lastly, this intelligence work allowed to establish a link, in terms of urban excesses, between small-time delinquency and radical extremism. More specifically, “the information gathered tracked precisely the role of salafi preachers in the suburbs, as well as their past and the reasons that led them to immigrate to France. It made it possible to imagine the potential consequences, both in religious terms and in terms of minor delinquency, of confidence schemes and cams, of trabendo and of small-time drug traffic. So that as early as the beginning of the 1990s, everything was known” (idem: 247).At the end of the year 1991, the tenth section of the R.G., also known as the “Cities and Suburbs” section (or section of Urban Violence), was attributed the primary mission of focusing on urban violence (Rosière, 2011: 95). It is quite a broad and generic phrase, that covers violence committed by skinheads, violence in the suburbs, gang violence, as well as activities taking place in rave parties (Caumer, 2000: 90). It was in that service that police commissioner Lucienne Bui Trong devised a scale that establishes 8 degrees of urban violence, from simple garbage can fires to actual urban riots (Bui Trong, 2000: 63-72):1. dumpster fires and car fires;2. harassment of authorities; 3. violence towards any person wearing a uniform (police, firefighters);4. gatherings against the police and the “stoning” of police cars;5. gatherings that disturb police operations, or organized with the intent of helping detained friends to escape;6. deliberate will to injure police and/or to attack police stations;7. “mini-riot” that doesn’t last more than a day, but does call for a swift escalation of violence;8. actual riot, which occurs in several places and lasts several nights.In view of this scale, the clashes that occurred during the protests against the El Khomri labor law in Paris 12 and other regions, 13 may be classified within the seventh degree of urban violence. It would not be an exaggeration to claim the same regarding the clashes that took place on the sidelines of the Yellow Vests movement protests, the various acts of violence committed by the black blocs, the lootings of the Arc de Triomphe and various stores by street gangs, etc. Such acts of urban violence also happen frequently after soccer games of the Algerian national team in several major French metropolitan areas. 14(10) For example, during the protest of March 8th, 2003, which resulted in tens of injured: “witnesses heard, on various occasions, phrases such as: ‘we’re gonna beat up some whities.’ Several high school students were severely beaten up, ten on one; they were punched, kicked and hit with sticks, girls were dragged by the hair. Racist slurs were heard: ‘white bastard’ is the leitmotiv in these organized lynchings. [...] Tens of teenagers were taken to the hospital. Upon investigation, it was discovered that most of the attackers were from the suburbs of Seine-Saint-Denis and the Northern districts of Paris,” in Pellegrini (C.), Banlieues en flammes, Editions Anne Carrière, Paris, 2005, p. 74-75.(11) Thus “starting from 2005, these same delinquents also attack the protesters themselves. Many protesters are assaulted and the the assaults specifically target young “white” high school students, which was previously unheard of” (Bauer et Soullez, 2010: 70.)(12) http://www.lexpress.fr/actualite/societe/en-images-loi-travail-paves-molotov-lacrymo-heurts-a-la-manif-parisienne_1791491.html(13) During the protests against the El Khomri law, the urban violence was especially serious in Rennes, in http://www.ladepeche.fr/article/2016/05/15/2344913-rennes-face-au-dechainement-de-la-violence-urbaine.html(14) https://www.francetvinfo.fr/sports/foot/coupe-d-afrique-des-nations-apres-la-victoire-de-l-algerie-des-incidents-a-marseille-paris-et-lyon_3536765.html© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 145 Cahiers de la sécurité et de la justice – Hors-série In structural terms, countering urban violence began (and still is to this day) one of the main missions of the R.G. after the circular of January 3rd, 1995. Indeed, besides the R.G.’s three main missions of national interest 15, the monitoring of urban violence was aggregated to other surveillance missions that were completely unrelated with urban violence. That is because in 1995, urban violence was not yet considered an issue that needed its own dedicated surveillance agency. Specifically, the issue of urban violence was grouped together with “matters of social exclusion (the issue of squats and various forms of marginality), the phenomenon of cults, as well as the impact on public opinion of events that received intense media coverage (for example, the scandal of the transfusions of HIV-contaminated blood, or exceptional criminal cases that made headlines, etc.)” (Zamponi, 1997: 208.)Then in 1999, the D.C.S.P., in cooperation with the R.G., elaborated a new database, for reasons of rivalry between services rather than for consistency purposes. In consequence, “the Bui Trong scale was also discarded, not only because of internal rivalries between the R.G. and the D.C.S.P., but also because the elected political officials were not too happy about this measuring device that made evident, year after year, the increase in tensions” (Bauer et Soullez, 2010: 24-25). Named System of Computerized Analysis of Urban Violence (in French, Système d’analyse informatique des violences urbaines, or SAIVU), it aimed to establish a record of all urban violence related incidents. But that tool was also replaced in 2005, by the National Indicator of Urban Violence (Indicateur national des violences urbaines, or INVU), in a collaborative effort between all the different Police Directorates, the National Gendarmerie, as well as the Police Prefecture. But by 2010, out of the nine statistical indices included in the database, only two were still in use… With regards to urban violence analysis, the R.G. report on the riots of November 2005 shed a new light on the issue, as it revealed there hadn’t been any involvement from any organization, whether Islamist, mafia-related or political. 16 Although the databases were not used properly, the reports from the service of territorial police intelligence demonstrate that their existence is necessary, in order to provide a relevant analysis of social phenomena.Then on June 27th, 2008, the “Cities and Suburbs” section was renamed “Division of urban excesses and identity-based withdrawal”, 17 which was maintained by the ruling of February 1st, 2011. 18 When the reform introduced by the decree of May 9th, 2014 established the new S.C.R.T., said “Division of urban excesses and identity-based withdrawal” remained in existence within the new agency 19. Through the missions attributed to them, the Departmental Services of Territorial Intelligence (SDRT) report the information they gather to the national level, guaranteeing a dense coverage of the territory, necessary to ensure the collection of information needed for homeland intelligence. These departmental services are meant to “contribute to locating the sites where the gangs operate, determining their structure and their delinquent behaviors, and identifying their members” (idem: 115).Because of the real danger represented by urban violence, also known as “violent subversion”, it has been included in the National Intelligence Strategy document of July 2019, which states that “the radicalization of these modes of action calls for increased vigilance of the intelligence services in their function of anticipation and defense of the State, so as to prevent all kinds of violence and the destabilization of our institutions.” 20The necessity to take into account the hybridization of organized crime and terrorismThe structure of criminal organizations is geared towards the accomplishment three goals: accumulation of power, conquest of urban “territory”, and accumulation of wealth. The scenario of an “alliance between international terrorism and organized crime could become a serious issue in the future” (Bühler, 2007: 37), and is not unlikely to happen. For motives of power and wealth, but also (15) These three missions are the surveillance of islamist groups activities, the social climate evolution and financial analysis and tracking, especially regarding dubious capital flows, the fight against organized criminality and corruption.(16) http://www.lemonde.fr/societe/article/2005/12/07/selon-les-rg-les-emeutes-en-banlieue-n-etaient-pas-le-fait-de-bandes-organisees_718347_3224.html(17) Ruling of June 27th, 2008 regarding the missions and the organization of the Central Directorate for Public Security, JORF, number 152, July 1st, 2008, text number 10, art. 7.(18) Ruling of February 1st, 2011 regarding the missions and the organization of the Central Directorate for Public Security, JORF, number 30, February 5th, 2011, text number 29.(19) Ruling of May 9th, 2014 modifying the ruling of February 1st, 2011, modified, regarding the missions and the organization of the Central Directorate for Public Security, JORF, number 108, May 10th, 2014, text number 29, art. 4.(20) Coordination Nationale du Renseignement et de la Lutte contre le Terrorisme, Stratégie nationale du renseignement, July 2019, p. 4.© Cahiers de la sécurité et de la justice - IHEMI146 I DOSSIER Intelligence, delinquency and urban violence – Alexis Depraureligious or intellectual affiliation, cases of collusion between criminal organizations and terrorist groups have indeed been observed. For example, there are links between the Albanese-speaking mafia and the Kosovo Liberation Army (UÇK). The existence of connections between criminal organizations and terrorist organizations (or even guerilla armies) is evidenced by the delivery and supply of weapons in exchange for money, and often for drugs: “these practices became generalized in the Balkans, because of their ideal geographical location, right on the route between opium production zones and Europe: local organizations deal with Turkish organizations upstream, and with Italian organizations downstream” (Chocquet, 2003: 45).In France, the link between terrorism and gangsterism is also a reality, as demonstrated by the case of the “Roubaix gang”, also known as the “Islamo-robbers” gang. 21 This “gangsterrorist” group, that operated in the mid-90s, was involved in armed robberies, but its members were Muslim converts, who had fought in Bosnia alongside the Islamists, and supported the cause of Islamist terrorism by funding it with their hold-ups. 22 This hybridization between organized crime and terrorism is also seen in France in the neo-gangsterism from the suburbs, and the deeply-rooted links it progressively built with what we can call “neo-terrorism”. Said neo-terrorism “emanates from young men and young women who grew up in that same environment, but who, at one point, chose to slip into fanaticism. It reflects the unbridled relationship to violence of young, confused, criminals who find in armed action an exaltation, a role and a place, while society doesn’t provide them with either. Holy war thus becomes a reason to live, in the same way that for other youths, the quest for money and power generated by drug traffick can be a reason to live” (Dufour et Kabssi, 2015: 176).Although it appeared in France over twenty years ago, that phenomenon has not really been subjected to any serious analysis or overview by intelligence services or political authorities. It must however be acknowledged that, since November 1993, the Ministry of Foreign Affairs created a Sub-Directorate for Security, in charge of dealing with “issues related to counter-terrorism, drug trafficking and international crime [...] in cooperation with the other ministries and departments involved” (Wodka-Gallien, 1995: 17).In any case, gaining more foresight needs to be a crucial priority. In the same way that a special unit was created for the profiling of perpetrators of terrorist acts, a similar unit dedicated to cases of hybridization would be beneficial, which would allow to avoid serious mistakes: “the case of Khaled Kelkal in 1995, the Roubaix gang in 1996: France has been confronted for the past twenty years to these hybrids, half-gangsters, half-terrorists, who navigate between several different police files, thus escaping to services that are too caught up in their own turf to cooperate with each other. The case of Mohamed Merah validated, in 2012, the process described by Mitch Silber in his 2006 book, Radicalization in the West: The Homegrown Threat, a study made by and for the New York Police Department. The cases of the Kouachi brothers and Amédy Coulibaly also confirmed that analysis. From now on, the threat arises from transient, almost protoplasmic entities. A continuum of crime and terrorism has emerged, far from the neat little boxes that conventionally distinguished between each of its initial components” (Bauer, 2015: 3).Thus, countering neo-gangsterism and neo-terrorism must be a priority surveillance mission, as gangs participate in organized crime, and its members may be involved in urban violence, but also affiliated to radical Islamism, or even terrorist organizations. So much so that, according to the Central Directorate of Judiciary Police, “near 40% of the suspects that appeared in such cases have a delinquent past” (Dubois et Pelletier, 2017: 87). In view of this phenomenon of hybridization, the distinction must be nuanced, and delinquency must be considered to be in direct connection with terrorism. Indeed, the vast majority of the persons who committed terrorist attacks were previously convicted in cases related to delinquency and organized crime (Raufer, 2017:24). The analysis provided by the SIRASCO must be heard, and this service’s budget must be increased so as to continue working on the hybridization issue. Furthermore, homeland intelligence must imperatively take into account that phenomenon. It is now the case, as the National Intelligence Strategy published in July of 2019 states that “crime favors terrorism, including in Europe, as it weakens the level of security and control over certain geographical areas and provides means of funding or supplying of weapons, fake documents, and equipment.” 23 (21) https://www.lepoint.fr/societe/il-y-a-20-ans-le-raid-lancait-son-assaut-contre-les-islamo-braqueurs-du-gang-de-roubaix-28-03-2016-2028385_23.php(22) http://www.rfi.fr/france/20160328-france-gang-roubaix-islamo-braqueurs-terrorisme-caze-dumont-bosnie(23) Coordination Nationale du Renseignement et de la Lutte contre le Terrorisme, Stratégie nationale du renseignement, July 2019, p. 7.© Cahiers de la sécurité et de la justice - IHEMIDOSSIER I 147 Cahiers de la sécurité et de la justice – Hors-série Sure, the reform of 2008, which replaced the General Intelligence (R.G.) services by the Sub-Directorate of General Information (SDIG), generated a loss of territorial coverage due to the massive transfer of policemen from that service to the D.C.R.I. (now called DGSI). However, in spite of suffering the consequences of this reform, the Central Service of Territorial Intelligence remains a necessary agency. It was only thanks to the new reform of 2014, which replaced the Sub-Directorate of General Information (SDIG) by the S.C.R.T., that the territorial coverage was rebuilt. It is completed by open source intelligence (OSINT), which is crucial to the collection of information.The operations of territorial intelligence, affiliated to the D.C.S.P. (Central Directorate for Public Security), enable to highlight the current issues and the threats related to social phenomena. Territorial intelligence thus made it possible to raise the thorny questions of street gangs, underground trafficks, and especially of urban violence, in connection with radical Islamism. From now on, it is necessary that any warnings from that intelligence and police service be taken seriously by political authorities, in order to ensure the best reaction to the threats that may affect homeland security, and even more so national security nReferencesBooksBauer, Alain and Christophe Soullez, 2010, Violences et insécurité urbaines, Paris, PUF, 12e ed.Bui Trong, Lucienne, 2000, Violences urbaines. Des vérités qui dérangent, Paris, Bayard.Caumer, Julien, 2000, Leurs dossiers R.G., Paris, Flammarion, 2000.Chocquet, Christian, 2003, Terrorisme et criminalité organisée, Paris, L’Harmattan.Dubois, Christophe and Éric Pelletier, 2017, Où sont passés nos espions ?, Paris, Albin Michel.Dufour, Julien and Abdelfettah Kabssi, 2015, Bandes, dérive criminelle et terrorisme, Paris, MA Éd., 2015.Gayraud, Jean-François, 2017, Théorie des hybrides. Terrorisme et criminalité organisée, Paris, CNRS Éditions.Madelin, Philippe, 2007, Dans le secret des services, Paris, Denoël.Pelligrini, Charles, Banlieues en flammes, Paris, Éd. Anne Carrière.Pradel, Jean and Jacques Dallest, 2012, La criminalité organisée. Droit français, droit international et droit comparé, Paris, Lexis Nexis.Rosière, Stéphane, 2011, Géographie des conflits armés et des violences politiques, Paris, Ellipses.Zamponi, Francis, 1997, Les R.G. à l’écoute de la France. Police et politique de 1981 à 1997, Paris, La Découverte and Syros.JournalsAubry, Gilles, 2009, “Organisations criminelles et structures répressives: panorama français”, Cahiers de la sécurité, n°7, pp. 25-40, January-March.Bauer, Alain, 2015, “Qui est l’ennemi?”, Conflits, supplément au n°5, April-May-June.Bühler, B.O., 2007, “L’alliance du crime organisé et du terrorisme. Un nouveau défi pour le futur?”, Défense & Sécurité internationale, n°30, pp.36-37.Raufer, Xavier, 2017, “Les hybrides (terroristes+criminels) – La police et le renseignement en Europe”, Atlantico, March.Wodka-Gallien, Philippe, 1995, “Renseignement et diplomatie: les voies de la cooperation”, Enjeux atlantiques, n°11, April.Academic thesesDesrousseaux, Thomas, 2014, Les bandes organisées en milieu urbain aujourd’hui, thèse, Université Panthéon-Assas Paris II.Janet, Mathieu, 2012, Sécurité publique et violences urbaines, mémoire, Université Paris II.© Cahiers de la sécurité et de la justice - IHEMI148 I VARIAEnvironnement big data et prise de décision : l’étape de contre-la-montre du Tour de France 2017 – Jordan Vazquez, Cécile Godé, Jean Fabrice LebratyEnvironnement big data et prise de décision : l’étape de contre-la-montre du Tour de France 2017Jordan VAZQUEZ, Cécile GODÉ, Jean Fabrice LEBRATYComme le démontrent Godé et Vazquez (2017), les effectifs de la Police nationale française rencontrent fréquemment des situations inattendues qui imposent des prises de décisions rapides (Godé, 2016). Les environnements big data sont susceptibles d’affecter le processus de prise de décision des policiers. La question que nous posons ici est la suivante : Comment les experts de la sécurité publique prennent-ils des décisions en environnement big data ?. Cette recherche s’intéresse à un évènement en particulier : l’étape de contre-la-montre du Tour de France 2017. La ville de Marseille a accueilli le 21 juillet 2017 les coureurs du Tour de France pour une étape de contre-la-montre : jusqu’à 300 000 personnes étaient attendues pour l’évènement. Afin de coordonner les patrouilles de police et les différentes compagnies de C.R.S. sur le terrain, les équipes du Centre d’Information et de Commandement (C.I.C.) de la Police de Marseille ont pu s’appuyer sur de nombreuses technologies qui constituaient leur environnement big data. Cet environnement big data permet aux décideurs de repérer des situations en contexte changeant, de réévaluer des situations non familières et d’envisager des solutions de retrait pour sécuriser les actions des équipes sur le terrain.Mots-clés : Décision, intuition, environ-nement big data, évènements inattendus, Police nationale, Tour de France 2017.n mai 2018, la police d’Orlando annonçait avoir arrêté un individu suspect grâce à l’intelligence artificielle Rekognition (Dugal, 2018), développée en 2016 par le géant Amazon (Amazon, 2016). Ce système est associé au parc de caméras de surveillance installées sur tout le périmètre de l’agglomération. L’objectif est de prévenir les actions malveillantes d’individus dangereux, grâce à l’analyse et l’exploitation des très nombreuses images captées par l’ensemble des caméras : une tâche difficilement réalisable par le seul recours à des ressources humaines tant les données générées sont massives. Ce partenariat traduit le fort intérêt des entreprises et des institutions publiques pour le big data et les applications qui en découlent. Les capacités de traitement de Rekognition sont considérables : le système peut détecter des objets et des visages, extraire du texte ou lancer des alertes lorsqu’une situation potentiellement à risque est repérée (Pathak et al., 2018). Rekognition permet par exemple d’identifier en temps réel jusqu’à cent personnes par image : ce qui conduit à l’analyse de plusieurs milliards de clichés chaque jour (Amazon, 2018).Intégrer les étapes de collecte, de traitement et d’analyse du big data représente d’importants défis matériels et humains pour les entreprises (Karoui et al., 2014 ; Vassakis et al., 2018). Le big data requiert de puissantes techniques de calcul pour révéler des tendances et des modèles à partir d’ensembles de données (Dallemule et Davenport, 2017). Il devient nécessaire d’investir dans des solutions avancées de traitement des données massives (big data analytics) : ce sont dans ce cas ces systèmes qui sélectionnent les données à traiter et qui transmettent une représentation directement exploitable des éléments agrégés (Vassakis et al., 2018) Rekognition est l’une de ces solutions avancées. Cependant, nombreuses sont les entreprises qui ne possèdent pas de système technologique permettant de stocker, de collecter et donc de traiter le big data. Elles se retrouvent dans ce cas contraintes d’analyser manuellement les données qui présentent un intérêt pour leurs activités (Vassakis et al., 2018). Nous considérons que ces entreprises évoluent en environnement big data. L’environnement big data correspond à « un environnement informationnel E© Cahiers de la sécurité et de la justice - IHEMIVARIA I 149 Cahiers de la sécurité et de la justice – Hors-série dense et hétérogène, constitué d’un ensemble de systèmes d’information (et/ou de technologies) non ou peu intégrés » (Godé et al., 2019). Il inclut de gros volumes de données structurées et non structurées dont la majeure partie est générée en continu.En environnement big data, les technologies d’aide à la décision jouent un rôle clé en fournissant des données significatives pour l’organisation, afin qu’elle puisse prendre des décisions (Assunção et al., 2015 ; Genovese et Prentice, 2014 ; Godé et Lebraty, 2013). Toutefois, c’est au décideur qu’il revient de sélectionner manuellement les données qui présentent un intérêt effectif. Il recoupe les informations et les agrège : il peut prendre en compte des publications Facebook ou Twitter et compléter ces informations en effectuant des recherches en parallèle sur Google pour obtenir une représentation plus complète de la situation.Dans le cas où le décideur est un expert de son domaine, le processus de sélection et de traitement des informations est généralement intuitif (Vassakis et al., 2018). Dane et Pratt définissent l’intuition comme « un jugement chargé affectivement qui émerge au travers d’associations rapides, non-conscientes et holistiques » (Dane et Pratt, 2007). Il existe plusieurs courants abordant la prise de décision intuitive. Nous suivrons le courant initié par l’approche naturaliste (Klein, 2015) qui estime que l’intuition se fonde sur l’expérience (naturalistic decision making). Les décisions intuitives fondées sur l’exploitation de l’environnement big data dépendent donc des capacités d’analyse des décideurs, mais aussi de leur expertise (Janssen et al., 2017).L’un des aspects les plus critiques du big data est relatif à son effet sur les processus décisionnels, a fortiori lorsque le processus de sélection des informations est intuitif. La littérature dans ce domaine s’est considérablement enrichie ces dernières années (Davenport, 2014, 2017 ; Davenport et al., 2012 ; George et al., 2014 ; Van Rijmenam, 2014 ; Vitari et Raguseo, 2017). Cependant, la plupart des travaux actuellement publiés ne s’intéressent à ce phénomène que sous le seul angle d’organisations disposant de solutions permettant le traitement automatisé du big data. Très peu se concentrent sur les problèmes rencontrés par les décideurs contraints de construire des assemblages de données massives en situation, à partir de dispositifs technologiques peu ou non intégrés.De nombreuses évolutions, notamment d’ordre technologique, sont actuellement en cours au sein de la Police nationale (P.N.) Cette institution s’ouvre en effet de plus en plus au big data et aux technologies qui permettent la collecte, le traitement et la visualisation des données. Elle accuse toutefois un retard par rapport aux organisations privées (Casey et al., 2019). Au sein des centres d’information et de commandement (C.I.C.), les équipes, sous l’autorité d’un superviseur, ont pour mission d’assurer la coordination des patrouilles sur le terrain. L’exploitation de l’environnement big data permet par exemple de faciliter l’allocation géographique des ressources (patrouilles pédestres ou motorisées) ou de réduire le taux de criminalité par l’identification de zones à risque (Casey et al., 2019). Pour l’heure, les tâches de collecte et d’analyse des données à disposition sont toutefois à la charge des superviseurs et de leurs équipes.Dans ce cadre, notre question de recherche est la suivante : comment les experts de la sécurité publique prennent-ils des décisions en environnement big data ?Cet article s’intéresse à des décideurs experts évoluant en contexte extrême, c’est-à-dire qui font régulièrement face à des imprévus et qui agissent sous une pression temporelle forte : les superviseurs des C.I.C. de la P.N. Il s’agit de déterminer si l’environnement big data affecte leurs processus décisionnels. En effet, ce nouvel environnement informationnel met à leur disposition des informations supplémentaires et peut, à priori, permettre d’améliorer les décisions mises en œuvre sur le terrain.La première partie de cet article présente le contexte extrême au sein duquel les décideurs de la P.N. évoluent ainsi qu’un modèle décisionnel approprié. La deuxième partie décrit la méthodologie ainsi que le terrain de recherche au sein sur lequel ont été menées les investigations. La troisième partie présente les Jordan VAZQUEZMaître de conférences - Université Polytechnique des Hauts de France, France.Cécile GODÉProfesseur des universités - Aix Marseille Univ, CRET-LOG, Aix-en-Provence, France.Jean Fabrice LEBRATYProfesseur des universités – Université de Lyon, France.© Cahiers de la sécurité et de la justice - IHEMI150 I VARIAEnvironnement big data et prise de décision : l’étape de contre-la-montre du Tour de France 2017 – Jordan Vazquez, Cécile Godé, Jean Fabrice Lebratyprincipales contributions de ce travail de recherche. Enfin, une réponse à la question de recherche est proposée en conclusion.Décision et environnement big data en contexte extrêmeLes policiers sont fréquemment confrontés à des évènements inattendus qui nécessitent d’adapter les modes d’action et les décisions appliquées sur le terrain. Cette spécificité est caractéristique des situations extrêmes de gestion (Aubry et al., 2010 ; Bouty et al., 2012 ; Godé, 2015 ; Godé et Lebraty, 2015). Au sein d’un contexte extrême, le décideur fait parfois face à des problèmes complexes et inattendus (Rittel et Webber, 1973), dont les conséquences sont difficiles à anticiper. Ils imposent au décideur de collecter des informations supplémentaires pour identifier avec précision la situation qu’il rencontre. Ce sont in fine ces informations qui conditionnent la décision qui sera par la suite appliquée.Les différentes situations rencontrées par les équipes policièresDans le cadre de leurs actions sur le terrain, les équipes opérationnelles de la P.N. sont généralement amenées à rencontrer trois principales situations : la situation routinière, la situation inattendue et la situation de crise (Godé, 2015).Comme le précise Godé, les acteurs dans des contextes extrêmes ne sont pas toujours confrontés à des situations imprévues. La plupart de leurs actions sont routinières. Cependant, une situation routinière peut soudainement basculer en situation inattendue. L’équipe doit donc être capable de gérer la transition d’une situation à l’autre et savoir passer de la décision standardisée cadrée par les procédures, en situation de routine, à la flexibilité et à l’adaptation en situation inattendue (Godé, 2015). En octobre 2020, deux agents de Police réalisent une mission de surveillance dans une zone industrielle du Val d’Oise (Le Monde, 2020). Cette intervention est routinière pour la patrouille. Au cours de cette mission, les policiers sont abordés par trois individus qui pensent avoir affaire à des gens du voyage. Les agents présentent leur carte de service mais sont immédiatement pris à parti et extraits de leur véhicule. Les agresseurs subtilisent dans la foulée les armes de service des policiers et ouvrent le feu. L’un des agents est touché à quatre reprises. Il s’en sortira finalement in extremis après plusieurs jours de coma. Ces policiers ont fait face à un dramatique basculement de situation, de la routine vers l’imprévu. Ces basculements imposent de très rapidement changer de mode d’action afin de répondre à la nouvelle menace rencontrée.Les policiers sont soumis à un risque d’imprévu constant, mais sont préparés pour y répondre (Hällgren et al., 2018). La mise en œuvre d’une action (même à priori adaptée) peut toutefois avoir des conséquences négatives, voire conduire à des catastrophes (ibidem), d’autant plus lorsque les patrouilles sont confrontées à des comportements à risque. C’est fréquemment le cas lorsque des individus refusent d’obtempérer et tentent de prendre la fuite. En juin 2019, une patrouille de la Préfecture de police de Paris procède à un banal contrôle routier suite à un accrochage entre deux voitures. L’un des conducteurs refuse d’obtempérer et tente de prendre la fuite. Durant la course-poursuite, le chauffard s’engage à contresens sur le périphérique pour échapper aux forces de l’ordre. Il finit par percuter un véhicule et provoque un accident d’une extrême gravité : cinq personnes sont blessées dont deux très grièvement (Figaro, 2019).Afin d’anticiper ces évènements et d’y répondre et, les équipes doivent être capables de redonner rapidement et intuitivement du sens aux situations rencontrées (Weick, 1993). C’est dans ce cas l’expertise du décideur qui permet de transformer une situation parfois chaotique en représentation cohérente (Klein, 1999).Modèle recognition-primed decision (R.P.D.) et environnement big dataLes tenants de l’approche naturaliste (naturalistic decision making) estiment que l’intuition repose sur l’expérience (Klein, 2015). D’après cette approche, l’intuition se renforce à mesure de l’acquisition de l’expertise (Phillips et al., 2004). Le décideur expert développe ainsi, au fil de ses expériences, sa capacité à répondre de manière réflexe aux évènements qu’il rencontre.Le modèle recognition-primed decision (R.P.D.) a été conçu par Gary Klein et décrit le processus décisionnel intuitif en quatre étapes d’un décideur expert évoluant en contexte extrême (Phillips et al., 2004 ; Ross et al., 2004). Selon ce modèle, le décideur repère tout d’abord une situation problématique (étape 1) puis collecte des informations issues de son environnement proche pour construire une représentation mentale de la situation (étape 2). Il identifie intuitivement des analogies entre la situation rencontrée et d’autres qu’il a vécues par le passé (étape 3). Durant cette phase, le décideur teste mentalement la première © Cahiers de la sécurité et de la justice - IHEMIVARIA I 151 Cahiers de la sécurité et de la justice – Hors-série solution qui lui vient à l’esprit et anticipe ses conséquences potentielles. La décision est finalement appliquée si celle-ci est considérée comme étant satisfaisante pour résoudre le problème rencontré (étape 4). Ce mode décisionnel indique que les experts ne comparent pas les options entre elles mais appliquent intuitivement une décision qu’ils ont déjà mise en œuvre dans un contexte jugé similaire (Klein et al., 2010). L’expert met en œuvre une solution qui lui semble correspondre à la situation à laquelle il est confronté, mais il éprouve des difficultés à expliquer pourquoi celle-ci est si évidente pour lui.Les nouvelles informations dorénavant accessibles par le biais des nouvelles technologies (réseaux sociaux, caméras de surveillance, bases de données, etc.) sont susceptibles d’avoir des effets sur les processus décisionnels des décideurs experts. L’environnement big data ne peut être défini comme le big data stricto sensu. Il comprend toutes les données accessibles (sociales ou non) par le biais des applications du web 2.0 mais aussi toutes les informations captées par des systèmes indépendants et consultables sur des tablettes, ordinateurs ou téléphones. Un décideur qui s’appuie sur des données issues d’un parc de caméras de vidéosurveillance, s’informe sur les réseaux sociaux et consulte en parallèle une multitude de bases de données internes à son organisation évolue au sein d’un contexte informationnel très variés que nous appelons environnement big data (Godé et al., 2019).Les parties qui suivent présentent la méthodologie de ce travail de recherche, détaillent les évènements importants survenus durant le déroulement de l’épreuve de contre-la-montre du Tour de France (T.F.) 2017 et offrent une représentation de l’environnement big data des équipes de Police en place durant le passage des coureurs. Nous tenterons de définir comment cet environnement a permis de fonder les décisions du superviseur qui a été en charge du dispositif de maintien de l’ordre déployé dans le cadre de cette épreuve sportive.Méthodologie et terrain de rechercheMéthodologieDans le cadre de cette recherche, plusieurs allers-retours ont été réalisés entre des périodes dédiées à la conduite d’investigations sur le terrain, et d’autres consacrées à l’étude des travaux théoriques en lien avec la problématique traitée (Charreire-Petit et Durieux, 2003) : la démarche adoptée est donc abductive (Koenig, 1993).La conduite des investigations au sein du C.I.C. de la P.N. des Bouches-du-Rhône a été rendue possible après plusieurs rencontres avec les responsables du département de recherche de l’École nationale supérieure de la police (É.N.S.P.) Ces investigations ont été réalisées dans le cadre d’un travail de thèse et de l’élaboration d’un rapport technique (Godé et Vazquez, 2017) commandé par ledit département. Au total, 28 entretiens ont été conduits, dont 4 réunions préalables, 12 entretiens exploratoires et 12 entretiens confirmatoires. Ces entretiens ont été réalisés individuellement ou collectivement, auprès de commandants et de commissaires de la P.N. La durée des entretiens est comprise entre 45 minutes et 2 heures 26 minutes pour une durée moyenne de 1 heure et 17 minutes. En parallèle, une journée d’observation non participante a été effectuée le 22 juillet 2017. Nous étions installés à l’arrière du superviseur d’une salle de maintien de l’ordre et disposions d’une vision d’ensemble sur les différents postes des opérateurs.Une analyse thématique de contenu a été conduite par le logiciel NVivo 11. Deux grilles de codification ont été élaborées. La première, dédiée à l’analyse des données issues des entretiens exploratoires, comprend un ensemble de thèmes identifiés après une première phase d’étude des travaux théoriques en lien avec la problématique de cet article. De nouvelles catégories ont par la suite émergé au fil des phases d’investigation, conduisant à l’élaboration d’une seconde grille de codification. Cette seconde grille comprend quatre thèmes principaux : (1) les spécificités contextuelles des situations décisionnelles ; (2) la familiarité du décideur par rapport à la situation ; (3) les informations exploitées par l’environnement big data durant le processus décisionnel ; (4) les rétroactions qui ont suivi l’intervention. L’analyse des données des trois premiers thèmes a livré trois enseignements : (1) des changements au niveau de la capacité de l’expert à repérer une situation problématique ou à anticiper un changement ; (2) un croisement systématique des données remontées par les patrouilles sur le terrain avec celles exploitables via les technologies du C.I.C. ; (3) un recours à l’environnement big data pour envisager des solutions de repli à destination des équipes sur le terrain. Ces résultats nous permettent de proposer une version amendée du modèle R.P.D. de Gary Klein.Présentation du C.I.C. de la Police nationale des Bouches-du-Rhône lors du passage du Tour de France 2017 (T.F. 2017)La P.N. est partagée en trois niveaux d’opérations : le tactique, l’opératif et le stratégique. Le niveau tactique englobe les © Cahiers de la sécurité et de la justice - IHEMI152 I VARIAEnvironnement big data et prise de décision : l’étape de contre-la-montre du Tour de France 2017 – Jordan Vazquez, Cécile Godé, Jean Fabrice Lebratyéquipes sur le terrain (patrouilles pédestres et motorisées), le niveau opératif comprend les différents C.I.C. Enfin, le niveau stratégique fait référence aux directives politiques.Le C.I.C. de la P.N. des Bouches-du-Rhône est organisé en neuf zones qui s’étendent sur 300 m² : salle de commandement, cellule d’appui opérationnelle, salle de service administratif, salle de traitement des appels 17, salle de traitement des messages, salle de maintien de l’ordre, etc. « Le rôle actuel des C.I.C. est de recevoir les appels d’urgence, de traiter ces appels, puis de coordonner les interventions. », nous explique un commandant. Le C.I.C. fait office de trait d’union entre les niveaux tactique et stratégique, ce que souligne le commandant C. : « Le C.I.C. est aussi l’échelon qui permet d’informer la hiérarchie afin de prendre de bonnes décisions ».Lorsque les équipes se rendent sur un lieu d’intervention, elles ne disposent bien souvent pas de toutes les informations nécessaires pour prévenir une situation dangereuse. Ce qu’affirme A. : « On est souvent sous-informé lors des interventions, les fonctionnaires ne savent par exemple pas lorsqu’ils se rendent au contact d’un individu déjà connu pour des faits de violence ». Le rôle informatif du C.I.C. est donc capital car les patrouilles n’ont pas le temps nécessaire pour effectuer ces recherches. C’est à lui de collecter et de transmettre l’information, à fortiori lors de la tenue d’un évènement d’ampleur.Les grands évènements programmés, tels que les rencontres sportives d’envergure et autres épisodes festifs, imposent la mise en place de dispositifs de sécurité conséquents. Ce sont la quasi-totalité des patrouilles qui peuvent être mobilisées pour encadrer l’organisation d’un évènement comme le passage du T.F. 2017. Le maintien de l’ordre est piloté depuis les locaux du C.I.C. , au sein d’une salle capable d’accueillir la totalité des acteurs concernés par l’évènement. Durant le déroulement d’un évènement d’ampleur, le dispositif de sécurité est piloté par le superviseur de la salle de commandement. Il prend la main sur ses équipes lorsqu’un évènement inattendu important a lieu. C’est pourquoi les seules décisions analysées dans le cadre de la période d’observation réalisée le 22 juillet 2017 sont celles du superviseur en place durant l’évènement.Cette année-là, une épreuve de contre-la-montre du T.F. doit avoir lieu à Marseille. Pour l’occasion, un périmètre de sécurité a été érigé tout le long du parcours. Des points disposés en des lieux stratégiques permettent le passage des piétons et des équipes organisatrices.Environnement big data du superviseur en salle de maintien de l’ordreLors d’un évènement tel que le passage du T.F. 2017 en épreuve de contre-la-montre, le superviseur bénéficie d’un environnement informationnel très riche et varié. Celui-ci est présenté dans le Tableau 1.Au sein du C.I.C., le superviseur croise en continu les différentes informations émanant de son environnement big data. C’est à lui (avec le soutien de ses équipes) que revient la tâche de collecter, de trier, de nettoyer et d’analyser les données disponibles. Certaines technologies génèrent des flux d’information difficilement traitables manuellement : le superviseur se trouve par exemple face à quatre écrans retranscrivant jusqu’à 16 flux vidéo simultanément. Il doit donc être capable d’identifier très rapidement les informations présentant un intérêt par rapport à la situation qu’il rencontre.Déroulement de l’évènement T.F. 2017Le 22 juillet 2017, la salle de maintien de l’ordre est ouverte dès 7h30 du matin pour préparer la mise en place du dispositif de sécurité pour l’évènement du jour. Pas moins de vingt personnes sont déjà présentes.Les premières heures de la matinée sont dédiées à la seule préparation du dispositif de sécurité. Il faut effectuer la transition de toutes les équipes de nuit avec celles de jour, vérifier les canaux de communication et s’assurer que tous les points de cisaillements soient surveillés par des équipages. Aux alentours de 8h30, les dernières discussions dans la salle se stoppent, les opérateurs sont dorénavant tous concentrés devant leurs écrans. Un adjoint au superviseur fait un point oral dans la salle : « À tous, je réitère une dernière fois, secteur 1 et 2 conférence 37 et secteur 3 et 4, conférence 38 ». À 9h13, tous les points semblent contrôlés.Chaque problème rencontré induit un contrôle par les caméras de surveillance. À 9h38, le superviseur remarque qu’un point de circulation sensible rue de la Bonneterie n’est surveillé par aucune équipe de la police municipale. L’opérateur de la police municipale présent dans la salle de maintien de l’ordre est appelé pour valider cette constatation. À 10h03, plusieurs traversées sauvages sont remontées par les équipes sur place : les gens passent d’une voie à l’autre en escaladant les barrières. Le superviseur décide de ne pas concentrer ses efforts sur ces franchissements : les premiers départs de la course ne doivent en effet pas commencer avant plus d’une demi-heure. À 10h30, le stade Vélodrome est ouvert au public et aucun débordement n’est à déplorer.© Cahiers de la sécurité et de la justice - IHEMIVARIA I 153 Cahiers de la sécurité et de la justice – Hors-série Il est 10h52 lorsque l’ambiance se tend, les franchissements de barrières sont de plus en plus nombreux et le tour de reconnaissance des équipes féminines est démarré. Un point de cisaillement, au niveau du Prado, n’est surveillé par aucune équipe municipale malgré son positionnement stratégique sur le parcours : il est décidé de contacter la mairie pour obtenir des effectifs supplémentaires. Le superviseur s’exclame « On est en train de se faire malmener !». La tension est palpable dans la salle car il y a une mésentente concernant ce point : les opérateurs de la police municipale signalent que ce point n’était pas sur leur plan d’origine. Après concertation, des patrouilles pédestres sont finalement détournées de leurs missions initiales pour se rendre sur place. À 11h14, le départ de la course de mécénat cardiaque est annoncé. Le superviseur voudrait que des patrouilles motorisées soient rapidement envoyées en soutien sur le point problématique, il apprend cependant deux minutes plus tard qu’aucun véhicule n’est disponible : cette option est abandonnée. À 11h31, la caravane se lance sur le parcours, le superviseur donne l’information dans la salle : « À tous, départ de la caravane effectif ! ».Suite à plusieurs remontées de la population, la supervision constate qu’une zone est complètement enclavée par le dispositif de sécurité, personne ne peut en sortir. Plusieurs Environnement big data – Tour de France 2017Sources externesMédia BFM TVChaine de télévision (actualités nationales et internationales en continu)Application Google MapService de cartographie en ligneCaméras du Centre de supervision urbainParc de 240 camérasApplication GoogleMoteur de rechercheRéseaux sociaux (via les renseignements territoriaux)Veille manuelle des échanges sur Twitter et Facebook concernant l’évènementSources internesApplication PEGASEGestion des missions en cours et positionnement des patrouillesFichier des personnes recherchées (FPR II)Base de données contenant les identités des individus recherchés et disparus en FranceFichier des objets et des véhicules signalés (FOVeS)Base de données contenant les immatriculations des véhicules signalés volésFichier de traitement des antécédents Judiciaires (TAJ)Base de données contenant les antécédents judiciaires des délinquants et criminelsSources humainesRenseignement territorialRéseau d’informateurs sur le terrainRéseau des transports marseillais (RTM)Les RTM disposent de caméras installées dans les tramwaysPatrouilles policières (pédestres et motorisées)Comprend les effectifs de la Police nationale, de la Police municipale et les C.R.S.SignaleursRemontées de civils chargés de signaler les incidents sur le parcoursPompiers de MarseilleLes pompiers travaillent en collaboration avec le C.I.C. (croisement des informations)GendarmesLes gendarmes travaillent en collaboration avec le C.I.C. (croisement des informations)Tableau 1 Environnement big data du superviseur© Cahiers de la sécurité et de la justice - IHEMI154 I VARIAEnvironnement big data et prise de décision : l’étape de contre-la-montre du Tour de France 2017 – Jordan Vazquez, Cécile Godé, Jean Fabrice Lebratymouvements de foule sont en parallèle à déplorer au niveau de certains points de cisaillement. Il est 12h05 et l’ambiance s’est de nouveau tendue dans la salle : il y a toujours discorde vis-à-vis du point du Prado. Durant le passage des coureurs, tous les points de cisaillement doivent impérativement rester fermés, c’est pourquoi les franchissements sauvages sont plus fréquents. À 12h21, le départ du tour de reconnaissance des coureurs masculins est annoncé tandis que des groupes de plus de vingt personnes franchissent les barrières au niveau du point de cisaillement de la flamme rouge. La menace terroriste est présente dans l’esprit de tous dans la salle, le superviseur explique : « Non mais le problème c’est que si un criminel commet un attentat, les pouvoirs publics se demanderont pourquoi la zone n’était pas sécurisée ! ». Une patrouille arrive au niveau de la flamme rouge à 12h34 et parvient à mettre un terme aux franchissements.À 12h56, de gros débordements sont annoncés au niveau de la zone dite de la Corderie, les policiers sur place doivent repousser des individus qui tentent de passer les barrières par la force. La situation est cependant rapidement stabilisée et un point de cisaillement est installé au niveau de la zone enclavée. L’épreuve de contre-la-montre démarre sans incident majeur à 13h50. Le principal problème rencontré par les équipes durant l’après-midi concerne la présence d’un sac suspect hors d’une zone piétonne. À 14h54, les caméras sont orientées sur l’endroit où se trouve le bagage. Ce qu’explique le superviseur : « C’est bon, on a le sac à la caméra, on fait monter le service de sécurité ». Aucune équipe ne peut toutefois franchir les voies du fait du passage des coureurs, il est décidé de passer par le métro et d’envoyer un chien capable de déceler les matières explosives. Le Réseau des transports marseillais est contacté pour signaler l’emprunt des voies de métro afin d’accéder au sac. À 15h14, un dispositif de sécurité est mis en place autour de la zone par un opérateur Sentinelle. Un individu apparait à la caméra à 15h25, il fouille dans le sac, personne dans la salle ne sait qui est cette personne. Le comportement de l’individu inquiète le superviseur : « Mais qui est cet homme ?! Il faut faire venir un effectif à pied et aller au contact, je n’aime pas son comportement, il textote, il est accroupi… Il faut envoyer des effectifs traverser tout de suite ! ». L’individu est finalement appréhendé par une patrouille piétonne qui est parvenue à traverser la voie. Un long échange s’ensuit durant lequel l’individu continue de textoter sur son téléphone. Le superviseur indique sur les ondes « Je veux l’identité complète de l’individu, et prenez-lui son téléphone !». Après contrôle du sac, l’équipe sur place signale que la personne en question est un agent de sécurité qui tentait de joindre son responsable avec son téléphone pour expliquer sa présence : tout danger est écarté. Le reste de la course se passe dans le calme. Les Renseignements territoriaux signaleront qu’à peine 40 000 personnes au total se sont finalement déplacées. La Figure 1 présente le dispositif de sécurité mis en place par les services de maintien de l’ordre.Figure 1 Dispositif de sécurité (réalisé) - Tour de France 2017© Cahiers de la sécurité et de la justice - IHEMIVARIA I 155 Cahiers de la sécurité et de la justice – Hors-série Analyse des résultatsNos résultats indiquent que l’environnement big data est exploité par le superviseur durant les quatre étapes mises en avant dans le modèle R.P.D. : repérage d’une situation ; représentation de la situation ; repérage de modèles ; application de la décision. L’environnement big data permet au superviseur de repérer une situation potentiellement problématique ou un contexte changeant (étape 1). Lorsque le décideur fait face à une problématique non familière, il consulte son environnement big data et croise les différentes sources auxquelles il a accès pour améliorer sa représentation de la situation (étape 2). Il collecte ensuite des indices supplémentaires, par l’environnement big data, pour tester la faisabilité de la décision envisagée (étape 3). Enfin, nous constatons qu’en environnement big data, le processus décisionnel ne s’arrête pas en phase 4, le superviseur continue d’exploiter des informations pour améliorer sa décision ou envisager de nouvelles stratégies.Étape 1 : Repérage d’une situation en environnement big dataDurant la mise en place du dispositif de sécurité pour l’épreuve de contre-la-montre du T.F., les caméras de surveillance sont utilisées par le superviseur pour contrôler le jalonnement des C.R.S. le long du parcours et l’état de chaque point de cisaillement. À plusieurs reprises, il est décidé de détourner des patrouilles de leurs missions initiales pour renforcer certains points jugés fragiles par le superviseur. Il anticipe des problèmes futurs susceptibles d’être engendrés par les faiblesses du dispositif de sécurité.Les images fournies en continu par le média BFM TV viennent compléter celles des caméras du centre de supervision urbain. Les prises de vues montrent clairement que la cité a été pour l’occasion désertée dans certaines zones. Ce que confirme le superviseur : « Ce n’est pas très dense au niveau de la population, selon les plans on dirait qu’il y a du monde mais en fait pas trop, ça va ». Très rapidement, le superviseur comprend que, sous réserve d’un évènement imprévu, le dispositif policier mis en place est amplement suffisant pour gérer les incidents. Il décide donc de concentrer ses équipes au niveau des zones les plus chargées en spectateurs.En routine, l’environnement big data offrent la possibilité de surveiller la situation sur le terrain et facilite le repérage des anomalies : le superviseur assure une veille en continu de l’évolution des situations sur le terrain. En ce sens, l’environnement big data renforce la capacité du décideur à anticiper un basculement de la routine à l’inattendu. Il passe de la réaction à la proaction (résultat 1).Étapes 2 et 3 : Représentation de la situation et évaluation de la décision envisagée en environnement big dataDurant le passage des coureurs, les informations sont principalement collectées par le réseau de caméras. Ces images permettent de construire une représentation mentale de la situation et de jauger son niveau de risque. Elles sont susceptibles de conduire le décideur à changer brutalement de mode d’action en fonction de l’évolution de la situation. Lorsque de gros débordements sont par exemple signalés à la Corderie, le décideur ordonne dans un premier temps de fermer le point de cisaillement en question. Après consultation des images remontées par les caméras présentes sur place, il change de stratégie et demande la réouverture du point de passage. En effet, les images permettent de constater que la concentration de spectateurs est bien trop importante sur la zone : le blocage du point pourrait provoquer des tensions. Durant ces étapes du processus décisionnel, le superviseur croise en continu les informations remontées par les équipes sur le terrain avec celles de son environnement technologique.Durant l’épreuve, des policiers municipaux signalent la présence d’un colis suspect. Le superviseur demande que soit réalisé un contrôle du colis. Quelques minutes plus tard, le propriétaire du sac se manifeste. Les images diffusées par la caméra inquiètent très rapidement le superviseur de la salle de maintien de l’ordre. Il demande que soit questionné de toute urgence l’individu. Son objectif n’est alors plus de contrôler le colis suspect, mais d’appréhender rapidement l’individu en question. Comme précédemment, l’environnement big data permet au superviseur de modifier immédiatement la réponse qu’il apporte à la situation rencontrée. Dans le cas présent, la situation de routine devient une urgence et les équipes sur le terrain modifient leur approche (Godé, 2015) en fonction du sens que le superviseur donne aux images qui lui parviennent.L’environnement big data met à disposition du décideur de nouvelles informations qui viennent enrichir la représenta-tion qu’il se fait d’une situation non familière. Ces nouvelles informations sont susceptibles de conduire plus fréquem-ment à un changement de mode d’action (résultat 2).Il peut toutefois arriver que le superviseur ne parvienne pas à obtenir certaines informations nécessaires à la prise de © Cahiers de la sécurité et de la justice - IHEMI156 I VARIAEnvironnement big data et prise de décision : l’étape de contre-la-montre du Tour de France 2017 – Jordan Vazquez, Cécile Godé, Jean Fabrice Lebratydécision en exploitant l’environnement big data. Durant le passage des coureurs, ce problème se présente à plusieurs reprises, notamment lorsque des franchissements sauvages ont lieu en dehors des zones couvertes par les caméras de surveillance. Il s’agit là d’une limite organisationnelle de l’environnement big data. Dans ce cas, le superviseur délègue la décision à un commissaire physiquement présent sur le terrain après une phase de concertation.Étape 4 : Application d’une décision en environnement big dataNos résultats traduisent un changement majeur, induit par l’environnement big data, à l’étape 4 du modèle R.P.D. Selon le modèle initial conçu par Gary Klein, le processus décisionnel s’achève lorsque la décision est appliquée. Or, nous constatons qu’en aval de l’application de la décision, l’environnement big data peut être consulté par le superviseur : le processus décisionnel n’est donc pas stoppé. Les nouvelles informations à disposition sont exploitées pour améliorer la décision en cours d’application (actions des équipes sur le terrain) ou pour envisager des solutions de secours en cas de survenue d’un évènement imprévu.Les informations transmises par Google Maps permettent par exemple d’améliorer une décision en cours d’application lorsque les équipes reçoivent l’ordre de se rendre sur une zone. Durant le T.F. 2017, le superviseur consulte cette application pour définir en temps réel les trajets à suivre afin d’éviter les zones d’embouteillages ou de travaux. Cette technologie permet de réduire le temps nécessaire à une équipe pour se rendre sur un lieu d’intervention (et in fine la pression temporelle). L’environnement big data tend ainsi à maintenir le contact entre le superviseur et les équipes opérationnelles lorsque la décision est en cours d’application.L’environnement big data, via Google Maps et l’interface de gestion des caméras de surveillance, permet par ailleurs d’envisager aisément des solutions de secours en cas de survenue d’une situation dangereuse et imprévue. Lorsque les équipes sont envoyées sur un lieu d’intervention situé dans une zone à risque, la connaissance de la topographie permet de prévoir des scénarios de retrait si elles sont prises à partie. Durant le T.F. 2017, lorsque des tensions éclatent entre les forces de l’ordre et des spectateurs prêts à forcer certains passages, le superviseur envoie une équipe au niveau du secteur concerné. Il identifie en parallèle des zones à proximité, dépourvues de barricades ou d’objets susceptibles d’être utilisés comme des projectiles, afin de permettre à ses équipes de se replier en cas de besoin.Face à l’imprévu, l’environnement big data renforce la capacité du décideur à identifier des solutions de secours et à améliorer une décision en cours d’application (résultat 3).Un modèle R.P.D. adapté à la prise de décision en environnement big dataLes résultats 1, 2 et 3 nous conduisent à proposer une nouvelle version du modèle R.P.D., adaptée aux processus décisionnels en environnement big data. Elle est présentée en figure 2 et inclut une nouvelle étape. L’étape 4 bis met en avant la poursuite du processus décisionnel durant l’application d’une décision. Cette phase est dorénavant employée par le décideur expert pour améliorer sa décision ou pour identifier des solutions de secours à appliquer dans le cas où la situation basculerait.Figure 2 Modèle RPD en environnement big data Conclusion Étape 1 : Repérage d’une situation Étape 2 : Représentation de la situation Étape 3 : Test d’une solution par la simulation mentale Étape 4 : Application de la décision La décision est inadaptée Étape 4 bis : Amélioration de la décision et solutions de secours La décision est adaptée Figure 2 Modèle RPD en environnement big data Le décideur repère et anticipe des situations problématiques via l’environnement big data Le décideur croise les informations de l’environnement big data pour affiner la représentation qu’il se fait de la situation Le décideur consulte l’environnement big data pour améliorer sa décision et sécuriser ses équipes © Cahiers de la sécurité et de la justice - IHEMIVARIA I 157 Cahiers de la sécurité et de la justice – Hors-série ConclusionCet article avait pour but de répondre à la problématique suivante : comment les experts de la sécurité publique prennent-ils des décisions en environnement big data ?L’environnement big data favorise la détection de signaux, conduisant à un basculement d‘une situation d’attente à une posture d’action (résultats 1 et 2). En outre, l’environnement big data induit une poursuite du processus décisionnel lorsque la décision est en cours d’application. Le décideur continue notamment d’exploiter des informations afin de garantir la sécurité de ses équipes en action (résultat 3).Par ailleurs, plusieurs recommandations à destination des équipes du C.I.C. de la police des Bouches-du-Rhône peuvent être formulées afin de faciliter l’exploitation de l’environnement big data. Il serait tout d’abord nécessaire d’allouer une ressource aux tâches de collecte et de traitement des données sociales et de caméras de surveillance : nous proposons de nommer un responsable de cette veille (recommandation 1). Par ailleurs, la mise en place de solutions de traitement automatisé nous paraît essentielle à moyen terme pour renforcer l’exploitation de ces informations. (recommandation 2). Plusieurs expérimentations conduites au sein de l’institution visent le repérage automatique des situations dangereuses (personnes à terre, colis abandonnés, armes à feu, etc.) Des alertes pourraient être automatiquement remontées aux superviseurs afin de les inviter à réaliser un contrôle visuel plus poussé en cas de repérage d’une situation potentiellement problématique.Enfin, nos investigations ouvrent la voie à de nouvelles recherches. En effet, nous avons pu constater qu’en fonction des contextes, l’environnement big data peut parfois affecter négativement la capacité du décideur à se représenter convenablement l’évènement en cours et engendrer la mise en œuvre d’une décision inadaptée à la situation rencontrée. Les nouvelles données à disposition augmentent en effet le risque de se retrouver confronté à des éléments contradictoires ou ambigus (Fisher et Kingma, 2001). C’est notamment le cas lorsque le superviseur constate (par les caméras de surveillance) que ses équipes échangent longuement avec le propriétaire du colis suspect. Il semble très étonné que les agents municipaux ne confisquent pas le portable de l’individu et lance de nombreux messages sur les ondes en enchainant les directives, sans que ces équipes ne réagissent. Il apprend par la suite que l’individu en question contactait simplement son employeur pour justifier sa présence au-delà du périmètre de sécurité. Les policiers municipaux détenaient cette information, contrairement au superviseur. C’est pourquoi ses directives n’avaient pas été suivies nBibliographieAmazon, 2016, « Introducing Amazon Rekognition » (https://aws.amazon.com/fr/about-aws/whats-new/2016/11/introducing-amazon-rekognition/).Amazon, 2018, « Amazon Rekognition » (https://aws.amazon.com/fr/rekognition/faqs/).Assunção, Marcos, Rodrigo Calheiros, Silvia Bianchi, Marco Netto et Rajkumar Buyya, 2015, « Big data computing and clouds : trends and future directions », Journal of parallel and distributed computing, n°79, p. 3-15.Aubry, Monique, Pascal Lièvre et Brian Hobbs, 2010, « Project management in extreme environments », Project management journal, 41, n°3, p. 2-3.Bharadwaj, Anandhi, Omar El Sawy, Paul Pavlou et N. Venkatraman, 2013, « Digital business strategy : toward a next generation of insights », MIS Quarterly, vol. 37, n°2.Bouty, Isabelle, Cécile Godé, Carole Drucker-Godard, Pascal Lièvre, Jean Nizet et François Pichault, 2012, « Coordination practices in extreme situations », European management journal, vol. 30, n°6, p. 475-489.Casey, Don, Phillip Burrell, et Nick Sumner, 2019, « Decision support systems in policing », European law enforcement research bulletin, n°4 special conference edition, p. 97-106.Charreire-Petit, Sandra et Florence Durieux, 2003, « Explorer et Tester : deux voies pour la recherche », in Thietart, Raymond-Alain (dir.), Méthodes de recherche en management, Paris, Dunod.Cukier, Kenneth et Viktor Mayer-Schoenberger, 2013, « The rise of big data : How it’s changing the way we think about the world », Foreign Affairs, n°92, p. 28-40.© Cahiers de la sécurité et de la justice - IHEMI158 I VARIAEnvironnement big data et prise de décision : l’étape de contre-la-montre du Tour de France 2017 – Jordan Vazquez, Cécile Godé, Jean Fabrice LebratyDallemule, Leandro et Thomas Davenport, « What’s your data strategy ? », Harvard business review, vol. 95, n°3, p. 112-121.Dane, Erik et Michael Pratt, 2007, « Exploring intuition and its role in managerial decision making », Academy of management review, vol. 32, n°1, p. 33-54.Davenport, Thomas, 2014, « How strategists use « big data » to support internal business decisions, discovery and production », Strategy and Leadership, vol. 42, n°4, p. 45-50.Davenport, Thomas, 2017, « The 2 Types of data strategies every company needs », Harvard business review (https://hbr.org/2017/05/whats-your-data-strategy).Davenport, Thomas, Paul Barth et Randy Bean, 2012, « How big data is different », MIT Sloan management review, vol. 54, n°1, p. 43.Dugal, Matthieu, 2018, « Arrêté grâce à Amazon : La reconnaissance faciale au service de la police », Radio-Canada (https://ici.radio-canada.ca/nouvelle/1103310/arrete-grace-a-amazon-la-reconnaissance-faciale-au-service-de-la-police).Figaro (le), 2019, « Paris : un chauffard prend le périphérique à contresens, 2 blessés graves », Le Figaro, 8 juin (https://www.lefigaro.fr/flash-actu/paris-un-chauffard-prend-le-peripherique-a-contresens-2-blesses-graves-20190608).Fisher, Craig et Bruce Kingma, 2001, « Criticality of data quality as exemplified in two disasters », Information et Management, vol. 39, n°2, p. 109-116.Genovese, Yvonne et Stephen Prentice, 2011, « Pattern-based strategy : getting value from big data », Gartner Special Report G.George, Gerard, Martine Haas et Alex Pentland, 2014, « Big data and management », Academy of management journal, vol. 57, n°2, p. 321-326.Godé, Cécile, 2015, La coordination des équipes en environnement extrême : Pratiques de travail et usages technologiques en situation d’incertitude, Paris, ISTE Éditions, coll. « Innovation, Entrepreneuriat et Gestion ».Godé, Cécile et Jean Fabrice Lebraty, 2013, « Improving decision making in extreme situations : the case of a military decision support system », International journal of technology and human interaction, vol. 9, n°1, p. 1-17.Godé, Cécile et Jean Fabrice Lebraty, 2015, « Experience feedback as an enabler of coordination : an aerobatic military team case », Scandinavian journal of management, vol. 31, n°3, p. 424-436.Godé, Cécile, Jean Fabrice Lebraty et Jordan Vazquez, 2019, « Le processus de décision naturaliste en environnement big data : le cas des forces de police au sein d’un Centre d’Information et de Commandement (C.I.C.) », Systèmes d’information et management, vol. 24, n°3, p. 67-96.Godé, Cécile et Jordan Vazquez, 2017, Étude É.N.S.P. : la prise de décision en environnement big data, une application aux forces de la Police nationale, Lyon.Hällgren, Markus, Linda Rouleau et Mark De Rond, 2018, « A matter of life or death : how extreme context research matters for management and organization studies », Academy of management annals, vol. 12, n°1, p. 111-153.Janssen, Marijn, Haiko van der Voort et Agung Wahyudi, 2017, « Factors influencing big data decision-making quality », Journal of business research, n°70, p. 338-345.Karoui, Myriam, Grégoire Davauchelle et Aurélie Dudezert, 2014, « Big data, mise en perspective et enjeux pour les entreprises », Ingénierie des systèmes d’information, vol. 19, n°3, p. 73-92.Klein, Gary, 1999, Sources of power : how people make decisions, MIT press.Klein, Gary, 2015, « A naturalistic decision making perspective on studying intuitive decision making », Journal of applied research in memory and cognition, vol. 4, n°3, p. 164-168.Klein, Gary, Roberta Calderwood et Anne Clinton-Cirocco, 2010, « Rapid decision making on the fire ground : The original study plus a postscript », Journal of cognitive engineering and decision making, vol. 4, n°3, p. 186-209.Koenig, Gérard, 1993, Production de la connaissance et constitution des pratiques organisationnelles (http://cat.inist.fr/?aModele=afficheNetcpsidt=97456).Monde (Le), 2020, « Deux policiers blessés par balles dans le Val-d’Oise, leurs armes volées », lemonde.fr, 8 octobre 2020 (https://www.lemonde.fr/societe/article/2020/10/08/deux-policiers-blesses-par-balles-dans-le-val-d-oise-leurs-armes-volees_6055241_3224.html).© Cahiers de la sécurité et de la justice - IHEMIVARIA I 159 Cahiers de la sécurité et de la justice – Hors-série Orlikowski, Wanda et Susan Scott, 2015, The algorithm and the crowd : considering the materiality of service innovation, MIS Quarterly, (39: 1) p. 201-216.Pathak, Ajeet, Manjusha Pandey et Siddharth Rautaray, 2018, « Application of deep learning for object detection », Procedia computer science, n°132, p. 1706-1717.Phillips, Jennifer, Gary Klein et Winston Sieck, 2004, « Expertise in judgment and decision making : a case for training intuitive decision skills », in Kœhler, Derek et Nigel Harvey (dir.), Blackwell handbook of judgment and decision making, Blackwell publishing, p. 297-315.Rittel, Horst et Melvin Webber, 1973, « Planning problems are wicked », Polity, n°4, p. 155-169.Ross, Karol, Gary Klein, Peter Thunholm, John Schmitt et Holly Baxter, 2004, The recognition-primed decision model, DTIC Document.Van Rijmenam, Mark, 2014, Think bigger : developing a successful big data strategy for your business, New York, American management association.Vassakis, Konstantinos, Emmanuel Petrakis, et Ionnis Kopanakis, 2017, « Big data analytics : applications, prospects and challenges », dans Skourletopoulos, Georgios, George Mastorakis, Constandinos Mavromoustakis, Ciprian Dobre et Evangelos Pallis, Mobile big data : a roadmap from models to technologies, Springer, p. 3-20.Vitari, Claudio et Elisabetta Raguseo, 2017, « Digital data, dynamic capability and financial performance : An empirical investigation in the era of Big data », Systèmes d’information et management, vol. 21, n°3, p. 63-92.Weick, Karl, 1993, « The collapse of sensemaking in organizations : the Mann Gulch disaster », Administrative science quarterly, vol. 38, p. 628-652.© Cahiers de la sécurité et de la justice - IHEMI160 I VARIABig data environments and decision-making: the time trial stage of the 2017 Tour de France – Jordan Vazquez, Cécile Godé, Jean Fabrice LebratyBig data environments and decision-making: the time trial stage of the 2017 Tour de France Jordan VAZQUEZ, Cécile GODÉ, Jean Fabrice LEBRATYAs demonstrated by Godé and Vazquez (2017), French National Police teams often encounter unexpected events (Godé, 2016) which compel them to make quick decisions. Big data environments can have an impact on their decision-making processes. The research question of this article is: “How are public safety decisions taken in big data environments?” This research focuses on a specific event: the time trial stage of the 2017 Tour de France, which took place in Marseille in 2017. The city of Marseille thus hosted the famous cyclists on July 21st, 2017 during this special stage of the popular annual French cycling race: up to 300,000 spectators were expected. In order to coordinate the numerous police patrols, the decision-makers of the Center for Information and Command (CIC) were able to rely on the set of technologies that constitute their big data environment. This new informational context is exploited by police decision-makers to identify risky situations, reassess a situation when an unexpected event occurs, and secure the operations of the teams on the ground.Keywords: decision-making, intuition, big data environment, unexpected events, police, 2017 Tour de France.n May of 2018, the Orlando Police Department reported that it had been able to arrest a suspect thanks to Rekognition (Dugal, 2018), an artificial intelligence software developed in 2016 by the industry giant Amazon (Amazon, 2016). This system is linked to the network of surveillance cameras installed across the whole urban area. Its purpose is to prevent the malevolent actions of dangerous individuals by exploiting and analyzing the many images caught by all the cameras: the data generated is so massive that such a task would hardly be achievable by resorting only to human resources. This partnership evidences the growing interest of companies and public institutions in big data and its resulting applications. Rekognition’s processing capacity is considerable: the system can detect objects and faces, extract text, and set off an alert when a potentially dangerous situation is spotted (Pathak et al., 2018). For example, Rekognition makes it possible to identify up to a hundred people per image in real time, which allows to analyze several billions of images per day (Amazon, 2018). Integrating the phases of collection, processing and analysis of big data represents an important human and material challenge for a company (Karoui et al., 2014; Vassakis et al., 2018). Big data requires significant computing power to reveal trends and patterns based on a set of information (Dallemule and Davenport, 2017). It becomes necessary to invest in advanced solutions for massive data processing, also known as “big data analytics”: systems that select the relevant data to be processed and transmit a directly exploitable representation of the aggregated elements (Vassakis et al., 2018). Rekognition is one of these advanced solutions. However, many companies do not have the technology necessary to store, collect and process big data. They are therefore forced to analyze manually the data relevant to their activity (Vassakis et al., 2018). It is considered that such companies I© Cahiers de la sécurité et de la justice - IHEMIVARIA I 161 Cahiers de la sécurité et de la justice – Hors-série operate within a “big data environment”, which is defined as “a dense and heterogeneous informational environment composed of a set of information systems (and/or technologies that are little or not integrated” (Godé et al., 2019). It includes major volumes of structured and unstructured data, most of it generated continuously. In a big data environment, decision support technologies play a key role by providing the organization with significant data so that it can make decisions (Assunção et al., 2015; Genovese and Prentice, 2014; Godé and Lebraty, 2013). It is incumbent upon the decision-makers, however, to manually select the relevant data. They must cross-reference the information and aggregate them: for example, they can take into account Facebook or Twitter posts, then complement the information by conducting parallel Google searches to obtain a more accurate representation of the situation. When the decision-makers are experts in their field, the process of selecting and processing the information is generally intuitive (Vassakis et al., 2018). Dane and Pratt define intuition as “affectively charged judgments that arise through rapid, nonconscious, and holistic associations” (Dane et Pratt, 2007). Several theoretical approaches exist that address intuitive decision-making. We shall adopt the naturalistic decision-making model (Klein, 2015), which claims that intuition is based on experience. Intuitive decisions based on exploiting a big data environment thus depends on the decision-makers’ ability to analyze but also on their expertise (Janssen et al., 2017). One of the most critical aspects of big data relates to its effect on decision-making processes, all the more so when the process of information selection is intuitive. The literature on this topic has been considerably enriched in the past few years (Davenport, 2014, 2017; Davenport et al., 2012; George et al., 2014; Van Rijmenam, 2014; Vitari and Raguseo, 2017). Yet most of the recently published works only address this phenomenon with respect to organizations that work with automated processing solutions for big data analytics. Very few focus on the issues that decision-makers are confronted with when they must build massive data collections on the spot, using little or not integrated technological devices. The French National Police (PN) is currently facing many changes, including a technological evolution. The institution is increasingly open to big data and to the technologies that make it possible to collect, process and visualize the data. However, it is lagging far behind private organizations (Casey et al., 2019). In the Centers for Information and Command (CIC), the staff’s mission is to ensure the coordination of the patrols on the ground, under the authority of a supervisor. Exploiting the big data environment enables, for example, to facilitate the geographical distribution of resources (pedestrian and motorized police patrols) or to reduce criminality by identifying at-risk areas (Casey et al., 2019). As for now, the tasks of collecting and analyzing the available data still remains incumbent on the supervisors and their teams. In such a context, our research question is as follows: how do public safety experts take decisions in a big data environment. This article focuses on expert decision-makers who work in extreme conditions, in the sense that they frequently face unpredictable events and operate under a heavy time pressure: the supervisors of the National Police Centers for Information and Command. The idea is to determine whether the big data environment has an impact on their decision-making processes. Indeed, this new informational environment provides them with more information which, we can assume, could enable them to improve the decisions that will be implemented on the ground. The first part of this article presents the extreme context in which the National Police decision-makers operate, as well as an adequate decision-making model. The second part describes the methodology as well as the research field within which the study was conducted. The third part introduces the main contributions of this research. And lastly, an answer to the research question is offered in the conclusion. Jordan VAZQUEZMaître de conférences - Université Polytechnique des Hauts de France, France.Cécile GODÉProfesseur des universités - Aix Marseille Univ, CRET-LOG, Aix-en-Provence, France.Jean Fabrice LEBRATYProfesseur des universités – Université de Lyon, France.© Cahiers de la sécurité et de la justice - IHEMI162 I VARIABig data environments and decision-making: the time trial stage of the 2017 Tour de France – Jordan Vazquez, Cécile Godé, Jean Fabrice LebratyDecision-making and big data environment in an extreme context Police officers are frequently confronted with unexpected events that require them to adapt their modes of action and the way they implement decisions on the ground. This specificity is characteristic of extreme management situations (Aubry et al., 2010; Bouty et al., 2012; Godé, 2015; Godé and Lebraty, 2015). In an extreme context, decision-makers often face complex and unexpected issues (Rittel and Webber, 1973), whose consequences are difficult to anticipate. Such issues force them to collect additional information in order to identify precisely the situation they are dealing with. Ultimately, this information is what determines the final decision that will be implemented. The different types of situations encountered by police teams According to Godé (2015), French National Police teams are generally faced with three main types of situations in the context of their operations on the ground: routine situations, unexpected situations, and crisis situations. The author explains that operators that work in extreme contexts are not always confronted with unpredictable situations. Most of their actions are routine. However, a routine situation can suddenly take a turn to an unexpected situation. The team must therefore be able to deal with the shift from a situation to another, and to trade the standardized decisions framed by procedure that govern routine situations for the flexibility and adaptation needed to respond to an unexpected situation (Godé, 2015). In October of 2020, two police officers are on patrol to monitor an industrial area in the Val d’Oise department (Le Monde, 2020). This is a routine mission for them. During their patrol, they are accosted by three individuals who mistake them for travellers. The police officers then show their professional ID cards but are immediately are taken on and pulled out of their vehicle. The aggressors steal their service weapons straight after and open fire. One officer is shot four times. He barely survived after several days in a coma. These police officers were confronted with a dramatic shift in a situation, from routine to unpredictable. These shifts force individuals to quickly change their mode of action in order to respond to the new threat faced.Police officers are subjected to the constant risk of unpredictable events, but they are prepared to respond to it (Hällgren et al., 2018). The execution of an action (even one that seems adapted), can however have negative consequences, or even lead to disasters (ibid.), especially when the patrols are confronted with an individual’s dangerous behavior. It is frequently the case when a suspect resists arrest and tries to run away. In June of 2019, a patrol of the Paris Police Department conducted an ordinary roadside check after a bump between two cars. One of the drivers refused to comply and tried to run away. During the chase, the man drove against traffic into the beltway in an attempt to escape the law enforcement officers. He ended up colliding with another vehicle and causing an extremely serious accident: five persons were injured, two of them very severely (Figaro, 2019). In order to anticipate these events and respond to them, police teams need to be able to quickly and intuitively make sense of the situations they encounter (Weick, 1993). In such cases, the expertise of the decision-makers is what enables them to transform a situation that may be chaotic into a coherent representation (Klein, 1999).Recognition-Primed Decision model (RPD) and big data environments The proponents of the naturalistic decision-making approach consider that intuition is based on experience (Klein, 2015). According to this approach, intuition is strengthened as more expertise is acquired (Phillips et al., 2004). Expert decision-makers thus build, as their experience grows, the reflex ability to respond to the situations they encounter.The Recognition-Primed Decision model (RPD), developed by Gary Klein, describes the 4 steps of the intuitive decision-making process of an expert decision-maker operating in an extreme context (Phillips et al., 2004; Ross et al., 2004). The model goes as follows: first, the decision-maker detects a problematic situation (step 1), then collects information from the surrounding environment to build a mental representation of the situation (step 2). Then, the decision-maker intuitively identifies analogies between the present situation and other situations encountered in the past (step 3). During that phrase, the decision-maker mentally tests the first solution that comes to mind, and anticipates its potential consequences. The decision is ultimately put into practice if it is deemed satisfactory to solve the problem encountered (step 4). This decision-making model indicates that experts do not compare different options with one another, but rather intuitively implement a decision that they already applied before in a context deemed similar (Klein et al., 2010). Experts © Cahiers de la sécurité et de la justice - IHEMIVARIA I 163 Cahiers de la sécurité et de la justice – Hors-série execute a solution that seems to them relevant to the situation they are confronted with, but they find it difficult to explain why the solution appears so evident to them. The additional information that is now accessible through new technology (social networks, surveillance cameras, databases, etc.) is likely to have an impact on the decision-making processes of expert decision-makers. A big data environment is not strictly the same as big data per se. A big data environment comprises all the data (social or otherwise) available through web 2.0 applications, but also all the information captured by independent systems and accessible on tablets, computers or phones. Decision-makers who rely on data from surveillance cameras, get information from social networks, and simultaneously search through the many internal databases of their organization operate within a varied informational context that qualifies as a big data environment (Godé et al., 2019).The following sections of this article present the methodology of our research, describe the important events that occurred during the time-trial stage of the 2017 Tour de France (TF), and offer a representation of the big data environment that police teams worked in during the cycling competition. We will try to determine how this environment enabled the supervisor in charge of the policing apparatus deployed for the sports event to make well-founded decisions. Methodology and field of research MethodologyDuring this research, we went back and forth several times between periods dedicated to conducting field investigation, and others devoted to the study of theoretical works related to the issue at hand (Charreire-Petit and Durieux, 2003): the method we adopted is thus abductive reasoning (Koenig, 1993).The investigation conducted within the CIC of the Bouches-du-Rhône regional PN department was made possible after several meetings with the officials in charge of the research division of the Superior National School of Police (ENSP). This investigation was carried out as part of a PhD research and the elaboration of a technical report (Godé and Vazquez, 2017) commissioned by said division. In total, 28 interviews were conducted, including 4 prior meetings, 12 exploratory interviews, and 12 confirmatory interviews. These interviews were carried out individually or collectively, with National Police commanders and captains. The interviews lasted between 45 minutes and 2 hours 26 minutes, with an average duration of 1 hour 17 minutes. Additionally, a day of non-participant observation was conducted on July 22nd, 2017. We were seated behind the supervisor of a public order enforcement room, and therefore could get an overall view of the different operators’ work stations. A thematic content analysis was executed using the NVivo 11 software. Two coding tables were elaborated. The first one, dedicated to the analysis of the information collected in the exploratory interviews, contains a set of themes identified after a first stage of studying theoretical works related to this article’s subject. New categories then emerged as the investigation kept going, which led to the elaboration of a second coding table. That second table comprises 4 main themes: (1) the contextual specificities of decision-making situations; (2) the familiarity of the decision-maker with the situation; (3) the information exploited through the big data environment during the decision-making process; (4) the consequences of the intervention. The analysis of the information pertaining to the first three themes provided three conclusions: (1) In a big data environment, the experts are better able to detect a problematic situation or anticipate a change. (2) The information collected by patrols on the ground is systematically cross-referenced with the data exploitable by the CIC’s technologies. (3) The big data environment is used to envision fallback solutions for the teams on the ground. These results allow us to suggest an amended version of Gary Klein’s RPD model. Presentation of the CIC of the Bouches-du-Rhône regional PN department and during the time-trial stage of the 2017 Tour de France (TF 2017) The National Police is divided in three levels: tactical, operational, and strategic. The tactical level involves the work of the teams on the ground (pedestrian and motorized patrols), the operational level consists of the action of the different CICs. Lastly, the strategic level refers to the political guidelines. The CIC of the Bouches-du-Rhône regional PN department extends over 3200 square meters and is organized into 9 zones: a command room, an operational support unit, an administrative service room, an emergency call dispatch room, a message processing room, © Cahiers de la sécurité et de la justice - IHEMI164 I VARIABig data environments and decision-making: the time trial stage of the 2017 Tour de France – Jordan Vazquez, Cécile Godé, Jean Fabrice Lebratya public order enforcement room, etc. As a commander explained to us : “The current function of CICs is to receive emergency calls, to process them, and to coordinate the interventions.” The CIC acts as the connecting link between the tactical and strategic levels, as commander C. highlights it: “The CIC is also the echelon that makes it possible to inform the hierarchy in order to make good decisions.”When the teams reach the site of an intervention, they are often not in possession of all the information needed to prevent a dangerous situation. This is emphasized by A.: “We are often underinformed during an intervention, for example the officers are not aware of it when they are going to be interacting with an individual who already has a violent record.” The informative function of the CIC is thus crucial, because the patrols do not have enough time to search for the relevant information. It is incumbent upon the CIC to collect and transmit it, all the more so during a major scale event. When a massive event is scheduled, for example a major sports competition or a festive occasion, a substantial security apparatus must be implemented. To secure an event such as the time-trial stage of the TF 2017, nearly all the patrols have to be mobilized. The public order enforcement operations are monitored remotely from the CIC premises, in a room able to accommodate all the staff involved in the event. During a big scale event, the security apparatus is monitored by the supervisor of the command room, who takes over and leads the teams when a significant unexpected situation arises. This is why the only decisions we analyzed in the context of the observation we conducted on July 22nd 2017, were that of the supervisor in charge during the event.That year, the time-trial stage of the TF was to take place in Marseille. A safety perimeter was set up for the occasion all along the cyclists’ route, and strategically placed openings allowed pedestrians and organizers to pass through. The big data environment used by supervisor of the public order enforcement roomDuring an event such as the time-trial stage of the TF 2017, the supervisor benefits from a very rich and varied informational environment, as presented in Table 1.Within the CIC, the supervisor constantly cross-references the information emanating from the surrounding big data environment. The tasks of collecting, classifying, organizing and analyzing the available data is incumbent to the supervisor (supported by the rest of the team). Some technologies generate a stream of information that is difficult to process manually: for example, the supervisor is facing 4 screens that are transmitting up to 16 video streams simultaneously, and must therefore be able to swiftly identify the information that is relevant to the situation at hand. Development of the TF 2017 eventOn July 22nd, 2017, the public order enforcement room opened as early as 7:30 a.m. in order to prepare the set-up of the security apparatus for the event. No less than 20 people were already present. The first hours of the morning were dedicated solely to the preparation of the security apparatus: executing the handover from the night teams to the day teams, checking the communication channels, and making sure all the authorized crossing points were monitored by patrols. Around 8:30, all conversations in the room came to a stop, and from then on the operators were all focused on their screens. An assistant to the supervisor gave an oral brief: “To all, I repeat one last time, sector 1 and 2, conference 37, and sector 3 and 4, conference 28.” By 9:13, all the crossing points seem under control.Every single problem encountered implies a check of the surveillance cameras. At 9:38, the supervisor noticed that a sensitive crossing spot, in the Bonneterie street, wasn’t being monitored by a Municipal Police team. The Municipal Police operator present in the public order enforcement room was called upon to confirm that fact. At 10:03, the teams on the ground reported several unauthorized crossings: people were climbing over the safety barriers to go from one side to the other. The supervisor decided not to focus efforts on these crossings, as the race would not start until over another half hour. At 10:30, the Vélodrome stadium opened to the public, and by then, no incidents had been reported. It was 10:52 when the atmosphere got tense, an increasing number of people were crossing the barriers, and the women’s teams warm-up lap began. An authorized crossing point, on the Prado avenue, was not being monitored by a Municipal team, in spite of its strategic location on the cyclists’ route: the decision was made to get in touch with the city officials to request additional forces. The supervisor exclaimed: “We’re being kicked around!” The tension was palpable in the room, as there had been a miscommunication on this issue: the operators of the Municipal Police reported that this crossing point didn’t appear on the map they had initially been given. After consultation, it was finally decided that a few patrols would be removed from their initial missions © Cahiers de la sécurité et de la justice - IHEMIVARIA I 165 Cahiers de la sécurité et de la justice – Hors-série and reallocated to that site. At 11:14, the heart disease charity race started. The supervisor expressed the wish to have motorized patrols promptly sent in support to the problematic crossing point, but was informed two minutes later that there was no vehicle available: that option was thus dismissed. At 11:31, the publicity caravan set off on the route, and the supervisor informed the room: “To all: caravan departure confirmed!”After several reports from the population, the supervisor observed that a whole area was completely enclosed by the security apparatus and nobody could exit it. Simultaneously, several stampedes were occurring near some of the crossing points. At 12:05 p.m., the atmosphere was getting tense again in the room: the Prado avenue crossing point remained a bone of contention. As the cyclists go through, all the authorized crossing points must imperatively remain closed, which is why more people tend to cross in disorderly, unauthorized ways. At 12:21, as the male cyclists’ warm-up lap started, groups of over 20 people were climbing over the barriers at the crossing point near the flamme rouge (the flag that signals 1km to the finish line). In the room, the terrorist threat was on everybody’s mind, and the supervisor explained: “The problem is, if a criminal commits an attack, public officials will ask why the zone wasn’t secured!” A patrol arrived at the flamme rouge at 12:34 and managed to put an end to the unauthorized crossings. Big data environment – Tour de France 2017External sourcesBFM TVTelevision channel (24-hours national and international news)Google Map ApplicationOnline mapping serviceCameras of the Urban Supervision CenterNetwork of 240 camerasGoogle Application Search engineSocial networks (via the territorial intelligence services)Social media watch, manual monitoring of Twitter and Facebook posts related to the event Internal sourcesPEGASE ApplicationManagement of current missions and patrol positioning File of Wanted People (FPR II)Database containing the identities of wanted and missing individuals in FranceFile of Reported Objects and Vehicles (FOVeS)Database containing the license plates of vehicles reported stolenCriminal Records File (TAJ)Database containing the judicial records of delinquents and criminalsHuman sourcesTerritorial Intelligence servicesNetwork of informants on the groundMarseille Public Transport Network (RTM)The RTM’s tramways are equipped with surveillance camerasPolice Patrols (both pedestrian and motorized)They include National Police, Municipal Police, and riot control units (CRS)InformersReports from civilians in charge of notifying any incident along the route Marseille FirefightersThe firefighters work in cooperation with the CIC to cross-reference the informationGendarmesThe gendarmes work in cooperation with the CIC to cross-reference the informationTable 1 the supervisor’s big data environment© Cahiers de la sécurité et de la justice - IHEMI166 I VARIABig data environments and decision-making: the time trial stage of the 2017 Tour de France – Jordan Vazquez, Cécile Godé, Jean Fabrice LebratyAt 12:56, major disturbances were reported in the Corderie area, where the police had to repel the individuals who were trying to forcibly cross the barriers. The situation was however rapidly stabilized, and a new crossing point was set up to open up the previously mentioned enclosed area. The time-trial race began without any major incident at 1:50. The main issue faced by the teams during the afternoon involved a suspicious package left unattended outside of the area open to pedestrians. At 2:54, the cameras were turned towards the location of the package. The supervisor said: “Good, we have a visual on the bag, let’s get the security unit over there”. However, no team could get across the road, because the cyclists were going through. It was thus decided to get there via the metro station and send a dog trained to detect explosives. The Marseille Public Transportation Network (RTM) was thus contacted to notify them that the metro tracks would be crossed to reach the bag. At 3:14, a security apparatus was set up around the area by an anti-terrorist “Sentinelle” operator. An individual appeared on the camera screen at 3:25, and started searching through the bag. No one in the room recognized him. His behavior seemed worrisome to the supervisor: “Who is that man?! We need to send a walking patrol to make contact, I don’t like his behavior, he is texting, he is squatted down… We must send a team across right now!” The individual was finally apprehended by a pedestrian patrol that managed to get across the road. A lengthy interaction ensued, during which the individual kept texting on his phone. The supervisor called out on the radio: “Give me the individual’s full identity, and take his phone from him!” After controlling the bag, the team on the ground reported that the individual was a security agent who was trying to reach his superior on the phone to explain why he was there: all danger was ruled out. The rest of the race went by without any incidents. The Territorial Intelligence services reported that ultimately, only 40,000 spectators turned out for the event. Figure 1 presents the security apparatus that was set up by law enforcement.Analysis of the results Our results indicate that the big data environment was exploited by the supervisor during all 4 steps described by the RPD model: detecting a situation, building a representation of the situation, identifying patterns, implementing a decision. The big data environment allows the supervisor to detect a potentially problematic situation or a shift in context (step 1). When the decision-maker is facing an unfamiliar issue, he searches through his big data environment and cross-references the different sources of information he has access to in order to improve his representation of the situation (step 2). He then collects additional clues, using the big data environment, to test the feasibility of the decision he is considering (step 3). Lastly, we note that in a big data environment, the decision-making process does not end Figure 1 Security apparatus implemented, 2017 Tour de France© Cahiers de la sécurité et de la justice - IHEMIVARIA I 167 Cahiers de la sécurité et de la justice – Hors-série at step 4, the supervisor continues to exploit the available information to improve his decision or envision new strategies.Step 1: Detection of a situation in a big data environment As the security apparatus was being set up for the TF’s time-trial race, the supervisor used the surveillance cameras to verify that the CRS riot control units were duly positioned along the route and that all the authorized crossing points were being monitored. On several occasions, it was decided to remove patrols from their initial missions in order to strengthen the control of points deemed sensitive by the supervisor, who anticipated potential future issues susceptible to arise from the weaknesses of the security apparatus. The continuous stream of images provided by the BFM TV channel completes those coming from the cameras of the Urban Supervision Center. The images clearly showed that on that occasion, certain areas of the city were deserted. This was confirmed by a statement made by the supervisor: “The crowd is not very dense, in some of the shots it looks like there are a lot of people, but really, it’s not too many, we’re good.” Very quickly, the supervisor understood that, unless something unexpected happened, the police apparatus that had been set up was more than enough to deal with any incident. He thus decided to focus his teams on the zones most packed with spectators. In a routine situation, a big data environment allows for the monitoring of the events on the ground and makes it easier to detect an anomaly: the supervisor is in charge of keeping a continuous watch over the evolution of the situation on the ground. Hence the big data environment strengthens the decision-maker’s ability to anticipate a sudden shift from routine to unexpected. From being reactive, he becomes proactive (result 1). Steps 2 and 3: Representation of the situation and assessment of the envisioned decision in a big data environmentAs the cyclists were passing through the city, the information was mainly being collected through the network of surveillance cameras. These images enabled to build a mental representation of the situation and to assess the risk level. They could lead the decision-maker to abruptly change his mode of action depending on the evolution of the situation. For example, when major incidents were reported in the Corderie area, the decision-maker initially ordered to shut down that crossing point. After checking the feedback from the cameras installed in the zone, he changed his strategy and requested the reopening of the crossing point. Indeed, the images showed him that the area was much too crowded with spectators: blocking the authorized crossing point could stir tensions. Throughout the stages of the decision-making process, the supervisor was continuously cross-referencing the information reported by the teams on the ground with the information gathered from his technological environment.During the race, municipal police officers reported the presence of a suspicious package. The supervisor requested a control of the package to be performed. A few minutes later, the bag’s owner appeared. Immediately, the images displayed on the camera screen were deemed concerning by the supervisor of the public order enforcement room. He thus requested the individual to be interrogated with all urgency. His objective was then no longer to control the suspicious package, but to quickly apprehend the individual in question. Again, the big data environment was what allowed the supervisor to swiftly change his response to the situation at hand. In the case discussed here, the routine situation had become an emergency situation, and the teams on the ground transformed their approach (Godé, 2015) according to the way the supervisor made sense of the images he was receiving. A big data environment provides decision-makers with new information that enriches their representation of an unfamiliar situation. This new information is likely to lead more frequently to a change in the selected mode of action (Result 2). However, it sometimes happens that the supervisor does not manage to obtain some of the information needed to make decisions by exploiting the big data environment. For example, as the cyclists were going through, this issue arose on several occasions, when unauthorized crossings occurred outside of the areas covered by the surveillance cameras. This is an organizational limitation of the big data environment. In such a case, after consultation, the supervisor delegates the decision to a police captain physically present on the ground. © Cahiers de la sécurité et de la justice - IHEMI168 I VARIABig data environments and decision-making: the time trial stage of the 2017 Tour de France – Jordan Vazquez, Cécile Godé, Jean Fabrice LebratyStep 4 : Enforcing a decision in a big data environment Our results evidence a major change induced by the big data environment in step 4 of the RPD model. According to the model initially conceived by Gary Klein, the decision-making process ends when the decision is put into practice. Yet we note that after the decision is implemented, the supervisor can still access his big data environment: the decision-making process is therefore not interrupted. The new information available can be exploited to improve the decision being enforced (the action of the teams on the ground) or to come up with a fallback solution in the event of an unpredictable incident. The information provided by Google Maps, for example, enables to improve a decision being implemented when the teams receive the order to go to a certain area. During the TF 2017, the supervisor used that application to determine in real time the best itinerary to follow in order to avoid heavy traffic or areas under construction. That technology enables to reduce the time it takes for a team to reach the intervention site (and ultimately, to reduce the time pressure). The big data environment thus allows the supervisor to stay in contact with the operational teams as the decision is being implemented.The big data environment, through Google Maps and the surveillance camera management interface, also makes it possible to easily envision a backup plan in case a dangerous, unexpected situation should arise. When the teams are sent to an intervention site in an at-risk area, good knowledge of the topography enables to plan strategic withdrawal scenarios for the police patrols in case they are faced with violent confrontation. During the TF 2017, when tensions burst between law enforcement and the spectators who were trying to force their way across, the supervisor deployed a team in the affected area. He simultaneously spotted nearby zones that were devoid of barriers or objects that could be used as projectiles, so that his teams could withdraw in case it became necessary. When facing the unpredictable, a big data environment strengthens the decision-maker’s ability to come up with backup solutions and to improve a decision already being implemented (Result 3). A new RPD model adapted to decision-making in a big data environment Results 1, 2 and 3 lead us to suggest a new version of the RPD model, adapted to decision-making processes in a big data environment. It is presented below in figure 2 and includes an additional step. Step 4a refers to the continuation of the decision-making process during the implementation of a decision. This new step is the phase when expert decision-makers can now improve their decision or come up with fallback solutions to put into practice in case the situation changes dramatically. Figure 2 RPD model adapted to a big data environment Conclusion The purpose of this article was to answer the following question: how do public security Step 1: Detection of a situation Step 2: Representation of the situation Step 3: Assessment of a solution through mental simulation Step 4: Implementation of the decision The decision is inadequate Step 4a: Improvement of the decision and creation of backup solutions The decision is adequate Decision-makers detect and anticipate problematic situations via the big data environment Decision-makers cross-reference the information provided by the big data environment to sharpen their mental representation of the situation Decision-makers search through their big data environment to improve their decisions and protect their teams. Figure 2 RPD model adapted to a big data environment© Cahiers de la sécurité et de la justice - IHEMIVARIA I 169 Cahiers de la sécurité et de la justice – Hors-série ConclusionThe purpose of this article was to answer the following question: how do public security experts make their decisions in a big data environment? A big data environment favors signal detection, which induces a shift from waiting in expectation to taking action (Results 1 and 2). Furthermore, working in a big data environment leads to a continued decision-making process including when the decision is already being implemented. In particular, decision-makers keep exploiting the available information in order to guarantee the safety of their operating teams (Result 3). Besides, several recommendations directed at the staff of the CIC of the Bouches-du-Rhône regional police department can be formulated in order to facilitate the exploitation of the big data environment. First of all, it is necessary to allocate resources to the tasks of collecting and processing the social data and the data provided by the surveillance cameras: we suggest appointing an agent in charge of keeping such a watch (Recommendation 1). Furthermore, it seems essential to us that automated processing solutions be implemented in a medium-term perspective to better exploit the information (Recommendation 2). Several experiments were conducted within the institution, aiming at automated detection of dangerous situations (person on the ground, packages left unattended, firearms, etc.) Whenever a potentially problematic situation is detected, an automated alert could be sent to the supervisors to invite them to perform an in-depth visual check. Finally, our investigation opens the way for new research avenues. Indeed, we noted that, in some contexts, a big data environment can sometimes negatively affect the decision-maker’s ability to build a correct mental representation of the unfolding event, and lead to enforcing a decision unsuitable for the situation encountered. Indeed, the newly available data increase the likeliness of being confronted with contradictory or ambiguous information (Fisher and Kingma, 2001). It was the case, for example, when the supervisor noted (via the surveillance cameras), that his teams were having a lengthy interaction with the owner of the suspicious package. He seemed very surprised that the municipal police officers had not confiscated the individual’s mobile phone, and sent out several messages on the radio, issuing command after command, without any reaction from his teams. He later learned that the individual in question was simply contacting his employer to explain why he was standing beyond the safety perimeter. The municipal police officers were in possession of that information, while the supervisor wasn’t, which is why his orders were not being executed nReferencesAmazon, 2016, “Introducing Amazon Rekognition” (https://aws.amazon.com/fr/about-aws/whats-new/2016/11/introducing-amazon-rekognition/).Amazon, 2018, “Amazon Rekognition” (https://aws.amazon.com/fr/rekognition/faqs/).Assunção, Marcos, Rodrigo Calheiros, Silvia Bianchi, Marco Netto and Rajkumar Buyya, 2015, “Big Data Computing and Clouds: Trends and Future Directions”, Journal of Parallel and Distributed Computing, n°79, p. 3-15.Aubry, Monique, Pascal Lièvre and Brian Hobbs, 2010, “Project Management in Extreme Environments”, Project Management Journal, 41, n°3, p. 2-3.Bharadwaj, Anandhi, Omar A. El Sawy, Paul A. Pavlou and N. Venkat Venkatraman, 2013, “Digital Business Strategy: Toward a Next Generation of Insights”, MIS Quarterly, vol. 37, n°2.Bouty, Isabelle, Cécile Godé, Carole Drucker-Godard, Pascal Lièvre, Jean Nizet and François Pichault, 2012, “Coordination Practices in Extreme Situations”, European Management Journal, vol. 30, n°6, p. 475-489.Casey, Don, Phillip Burrell and Nick Sumner, 2019, “Decision Support Systems in Policing”, European Law Enforcement Research Bulletin, n°4 special conference edition, p. 97-106.Charreire-Petit, Sandra and Florence Durieux, 2003, « Explorer et Tester : deux voies pour la recherche », in Thietart, Raymond-Alain (Ed.), Méthodes de recherche en management, Paris, Dunod.Cukier, Kenneth and Viktor Mayer-Schoenberger, 2013, “The Rise of Big Data: How it’s Changing the Way we Think about the World”, Foreign Affairs, n°92, p. 28-40.Dallemule, Leandro and Thomas Davenport, “What’s Your Data Strategy?”, Harvard Business Review, vol. 95, n°3, p. 112-121.© Cahiers de la sécurité et de la justice - IHEMI170 I VARIABig data environments and decision-making: the time trial stage of the 2017 Tour de France – Jordan Vazquez, Cécile Godé, Jean Fabrice LebratyDane, Erik and Michael Pratt, 2007, “Exploring Intuition and its Role in Managerial Decision Making”, Academy of Management Review, vol. 32, n°1, p. 33-54.Davenport, Thomas, 2014, “How Strategists Use ‘Big Data’ to Support Internal Business Decisions, Discovery and Production”, Strategy and Leadership, vol. 42, n°4, p. 45-50.Davenport, Thomas, 2017, “The 2 Types of Data Strategies Every Company Needs”, Harvard Business Review (https://hbr.org/2017/05/whats-your-data-strategy).Davenport, Thomas, Paul Barth and Randy Bean, 2012, “How Big Data is Different”, MIT Sloan Management Review, vol. 54, n°1, p. 43.Dugal, Matthieu, 2018, « Arrêté grâce à Amazon : La reconnaissance faciale au service de la police », Radio-Canada (https://ici.radio-canada.ca/nouvelle/1103310/arrete-grace-a-amazon-la-reconnaissance-faciale-au-service-de-la-police).Figaro (le), 2019, « Paris : un chauffard prend le périphérique à contresens, 2 blessés graves », Le Figaro, June 8th (https://www.lefigaro.fr/flash-actu/paris-un-chauffard-prend-le-peripherique-a-contresens-2-blesses-graves-20190608).Fisher, Craig and Bruce Kingma, 2001, “Criticality of Data Quality as Exemplified in Two Disasters”, Information et Management, vol. 39, n°2, p. 109-116.Genovese, Yvonne and Stephen Prentice, 2011, “Pattern-based Strategy: Getting Value from Big Data”, Gartner Special Report G.George, Gerard, Martine Haas and Alex Pentland, 2014, “Big data and Management”, Academy of Management Journal, vol. 57, n°2, p. 321-326.Godé, Cécile, 2015, La coordination des équipes en environnement extrême : Pratiques de travail et usages technologiques en situation d’incertitude, Paris, ISTE Éditions, collection « Innovation, Entrepreneuriat et Gestion ».Godé, Cécile and Jean Fabrice Lebraty, 2013, “Improving Decision Making in Extreme Situations: the Case of a Military Decision Support System”, International Journal of Technology and Human Interaction, vol. 9, n°1, p. 1-17.Godé, Cécile and Jean Fabrice Lebraty, 2015, “Experience Feedback as an Enabler of Coordination: an Aerobatic Military Team Case”, Scandinavian Journal of Management, vol. 31, n°3, p. 424-436.Godé, Cécile, Jean Fabrice Lebraty and Jordan Vazquez, 2019, « Le processus de décision naturaliste en environnement big data : le cas des forces de police au sein d’un Centre d’Information et de Commandement (C.I.C.) », Systèmes d’information et management, vol. 24, n°3, p. 67-96.Godé, Cécile and Jordan Vazquez, 2017, Étude É.N.S.P. : La prise de décision en environnement big data, une application aux forces de la Police nationale, Lyon.Hällgren, Markus, Linda Rouleau and Mark De Rond, 2018, “A Matter of Life or Death: How Extreme Context Research Matters for Management and Organization Studies”, Academy of Management Annals, vol. 12, n°1, p. 111-153.Janssen, Marijn, Haiko van der Voort and Agung Wahyudi, 2017, “Factors Influencing Big Data Decision-Making Quality”, Journal of Business Research, n°70, p. 338-345.Karoui, Myriam, Grégoire Davauchelle and Aurélie Dudezert, 2014, « Big data, mise en perspective et enjeux pour les entreprises », Ingénierie des systèmes d’information, vol. 19, n°3, p. 73-92.Klein, Gary, 1999, Sources of Power: How People Make Decisions, MIT Press.Klein, Gary, 2015, “A Naturalistic Decision Making Perspective on Studying Intuitive Decision Making”, Journal of Applied Research in Memory and Cognition, vol. 4, n°3, p. 164-168.Klein, Gary, Roberta Calderwood and Anne Clinton-Cirocco, 2010, “Rapid Decision Making on the Fire Ground: The Original Study Plus a Postscript”, Journal of Cognitive Engineering and Decision Making, vol. 4, n°3, p. 186-209.Koenig, Gérard, 1993, Production de la connaissance et constitution des pratiques organisationnelles (http://cat.inist.fr/?aModele=afficheNetcpsidt=97456).Monde (Le), 2020, « Deux policiers blessés par balles dans le Val-d’Oise, leurs armes volées », lemonde.fr, October 8th (https://www.lemonde.fr/societe/article/2020/10/08/deux-policiers-blesses-par-balles-dans-le-val-d-oise-leurs-armes-volees_6055241_3224.html).© Cahiers de la sécurité et de la justice - IHEMIVARIA I 171 Cahiers de la sécurité et de la justice – Hors-série Orlikowski, Wanda et Susan Scott, 2015, The algorithm and the crowd : considering the materiality of service innovation, MIS Quarterly, vol. 39, n°1, p. 201-216.Pathak, Ajeet, Manjusha Pandey and Siddharth Rautaray, 2018, “Application of Deep Learning for Object Detection”, Procedia Computer Science, n°132, p. 1706-1717.Phillips, Jennifer, Gary Klein and Winston Sieck, 2004, “Expertise in Judgment and Decision Making: a Case for Training Intuitive Decision Skills”, in Kœhler, Derek and Nigel Harvey (Eds.), Blackwell Handbook of Judgment and Decision Making, Blackwell Publishing, p. 297-315.Rittel, Horst and Melvin Webber, 1973, “Planning Problems are Wicked”, Polity, n°4, p. 155-169.Ross, Karol, Gary Klein, Peter Thunholm, John Schmitt and Holly Baxter, 2004, The Recognition-Primed Decision Model, DTIC Document.Van Rijmenam, Mark, 2014, Think bigger: Developing a Successful Big Data Strategy for Your Business, New York, American Management Association.Vassakis, Konstantinos, Emmanuel Petrakis, et Ionnis Kopanakis, 2017, « Big data analytics : applications, prospects and challenges », dans Skourletopoulos, Georgios, George Mastorakis, Constandinos Mavromoustakis, Ciprian Dobre et Evangelos Pallis, Mobile big data : a roadmap from models to technologies, Springer, p. 3-20.Vitari, Claudio and Elisabetta Raguseo, 2017, “Digital Data, Dynamic Capability and Financial Performance: an Empirical Investigation in the Era of Big Data”, Systèmes d’information et management, vol. 21, n°3, p. 63-92.Weick, Karl, 1993, “The Collapse of Sensemaking in Organizations: the Mann Gulch disaster”, Administrative Science Quarterly, vol. 38, p. 628-652.© Cahiers de la sécurité et de la justice - IHEMIView publication stats | RESEARCH GATE |
Data-driven Model of Temporal Evolution of Solar Mg II h and k Profiles over the SolarCycleJúlius Koza1, Stanislav Gunár2, Pavol Schwartz1, Petr Heinzel2,3, and Wenjuan Liu21 Astronomical Institute, Slovak Academy of Sciences, 05960 Tatranská Lomnica, Slovakia; koza@astro.sk2 Astronomical Institute, The Czech Academy of Sciences, 25165 Ondřejov, Czech Republic3 University of Wrocłav, Center of Scientific Excellence—Solar and Stellar Activity, Kopernika 11, 51-622 Wrocłav, PolandReceived 2021 December 28; revised 2022 March 22; accepted 2022 March 29; published 2022 July 20AbstractThe solar radiation in the cores of the Mg II h and k spectral lines plays a significant role in the illumination ofprominences, coronal mass ejections (CMEs), spicules, flare loops, and surges. Moreover, the radiation in theselines strongly correlates with solar magnetic activity and the ultraviolet solar spectral irradiance affecting thephotochemistry, especially of oxygen and nitrogen, in the middle atmosphere of the Earth. This work provides adata-driven model of temporal evolution of the solar full-disk Mg II h and k profiles over the solar cycle. Thecapability of the model to reproduce the Mg II h and k profiles for an arbitrary date is statistically assessed. Basedon selected 76 IRIS near-UV full-Sun mosaics covering almost the full solar cycle 24, we find the parameters ofdouble-Gaussian fits of the disk-averaged Mg II h and k profiles and a model of their temporal evolutionparameterized by the Bremen composite Mg II index. The model yields intensities within the uncertainties of theobserved data in more than 90% of the reconstructed profiles assuming a statistically representative set of BremenMg II index values in the range of 0.150–0.165. The relevant full-disk Mg II h and k calibrated profiles withuncertainties and spectral irradiances are provided as an online machine-readable table. The model yieldsMg II h and k profiles representing the disk incident radiation for the radiative-transfer modeling of prominences,CMEs, spicules, flare loops, and surges observed at arbitrary time.Unified Astronomy Thesaurus concepts: Solar chromosphere (1479); Solar ultraviolet emission (1533); Solarcycle (1487)Supporting material: machine-readable table1. IntroductionRadiative-transfer models of isolated chromospheric andcoronal structures require an accurate specification of incidentradiation—namely, radiation from the solar disk—as a keyboundary condition in calculations. Incident radiation stronglyaffects the source function and thus must be precisely specifiedfor all considered line and continuum frequencies and for alldirections (Labrosse et al. 2010; Heinzel 2015). This pertains tomodeling prominences (Heinzel et al. 2014, 2015; Schwartzet al. 2015; Vial et al. 2016; Jejčič et al. 2018; Levens &Labrosse 2019; Ruan et al. 2019; Vial et al. 2019; Zhang et al.2019; Barczynski et al. 2021; Peat et al. 2021), spicules(Alissandrakis et al. 2018; Tei et al. 2020; Kuridze et al. 2021),flare loops (Mikuła et al. 2017; Koza et al. 2019), and surges(Kayshap et al. 2021).TheresonancelinesMg II h(2803.53 Å)andMg II k(2796.35 Å) provide important diagnostic tools for these structuresand their disk profiles represent a significant source of externalillumination. The archive4 of the Interface Region ImagingSpectrograph (IRIS; De Pontieu et al. 2014) offers a wealth ofobservations of on-disk and off-limb structures (De Pontieuet al. 2021). Their precise modeling requires specification ofactual background or incident radiation at the moment of dataacquisition. The IRIS full-Sun mosaics, primarily used forquasi-regular monitoring of changes in instrument sensitivity,can be also employed for definition of incident radiation fromthe solar disk. Recently, such a mosaic provided the incidentMg II h and k profiles for prominence modeling by Zhang et al.(2019) and Vial et al. (2019).But the IRIS mosaics have much broader application.Investigating the chromospheric footpoints of the solar wind,Bryans et al. (2020) analyzed three Mg II h mosaics. Theydemonstrated significant spectral variation within the chromo-spheric plasma of coronal holes. The spectral differences outlinethe boundaries of some—not all—coronal holes and point to achromospheric source for the inhomogeneities found in the fastsolar wind. Ayres et al. (2021) used the IRIS full-disk mosaics in astudy of the solar–stellar connection. They showed that the IRISC II (T ∼ 104 K), Si IV (8 × 104 K), and Mg II (8 × 103 K) Sun-as-a-star profiles compare well to tracings of solar-twin α Centauri A(G2 V). As far as the solar–stellar connection is concerned (1) thechromospheric and transition zone lines are progressively moreassociated with surface magnetic features (network lanes andplages) as the formation temperature T rises, and (2) even at themagnetic cycle minimum, all the IRIS reference lines havesignificant emission from the supergranulation network.The Mg II h and k intensities, observed on the solar disk,varyconsiderablybetweendifferentobservedstructures(Schmit et al. 2015), with distance from the disk center (Gunáret al. 2021), and also with the solar cycle. In general, variationin the solar spectral irradiance (SSI) drives short-timescalechanges in the middle atmosphere of the Earth, and on longertimescales it influences the Earthʼs climate (Floyd et al. 2002;The Astrophysical Journal Supplement Series, 261:17 (18pp), 2022 Augusthttps://doi.org/10.3847/1538-4365/ac69cf© 2022. The Author(s). Published by the American Astronomical Society.Original content from this work may be used under the termsof the Creative Commons Attribution 4.0 licence. Any furtherdistribution of this work must maintain attribution to the author(s) and the titleof the work, journal citation and DOI.4https://iris.lmsal.com/data.html1Snow et al. 2019). In particular, changes in the 2000–3000 Åregion of the solar spectrum involving the Mg II h and k linescontrol the creation and destruction of ozone (Heath &Schlesinger 1986; Snow et al. 2019). Heath & Schlesinger(1986) defined the Mg II center-to-wing ratio (also known asthe Mg II index) as an easily measured proxy for solar magneticactivity and relevant ultraviolet SSI variability. The ratio of theirradiances in the centers of the Mg II h and k lines to theirradiances in the nearby continua (i.e., the ratio of thechromospheric contribution to the photospheric contribution)on either side produces a dimensionless quantity that is highlycorrelated to SSI variability throughout the ultraviolet spectralrange. Some of the primary advantages of such a relativemeasurement—rather than an absolute measurement—are thatit is fairly easy to obtain and that most instrument artifacts thataffect both center and wings cancel each other out. Snow et al.(2019) revised this assumption concluding that degradation-corrected data are better but that uncorrected data still producea highly accurate Mg II index. As a test case they used high-resolution spectral measurements obtained by the Solar–StellarIrradiance Comparison Experiment (SOLSTICE; McClintocket al. 2005a, 2005b) on board the Solar Radiation and ClimateExperiment (SORCE; Anderson & Cahalan 2005; Rott-man 2005) covering the period from the end of 2003 to theend of 2017 (i.e., the declining phase of solar cycle 23 and mostof solar cycle 24). They found that uncorrected data wouldproduce an error in the Mg II index that is less than 0.6% of thesolar cycle amplitude over a decade. Using corrected data, thaterror decreases by a factor of 5 to about 0.1% of the solar cycleamplitude. It proves the relative insensitivity of the Mg II indexto instrumental degradation. The Mg II index has a long historyof daily measurements reviewed in Snow et al. (2014, 2019), aswell as several published long-term composites. For example,the University of Bremen maintains the widely used Bremencomposite magnesium II index (Snow et al. 2014), hereafterreferred to as the Bremen Mg II index or BI, of which the dailyvalues from 1978 November 7 to the present are availableonline.To take properly into account the significant temporalvariability of Mg II h and k line radiation from the solar disk inradiative-transfer modeling of chromospheric and coronal struc-tures, disk-averaged profiles should be taken quasi-cotemporallywith observations of the studied structures. However, this isalmost impossible and thus to compensate for the absence ofactual full-disk Mg II h and k profiles a model of their cyclicevolution is needed allowing one to reconstruct them for arbitrarydates. This problem is solved for the Lyα line in Kowalska-Leszczynska et al. (2018, 2020) and Gunár et al. (2020). Theypresented models of temporal evolution of the full-disk Lyα lineprofile driven by the composite Lyα index, i.e., the disk- andwavelength-integrated Lyα spectral irradiance. Although bothmodels use the same driver they differ much in input Lyα data,scopes of applicability, and complexity of reconstruction of thespectral shape of Lyα. Construction of the model in the latterwork was primarily aimed at application in radiative-transfermodeling. Because there does not exist an index for theMg II h and k lines equivalent to the composite Lyα index (i.e.,a disk- and wavelength-integrated Mg II spectral irradiance) theBremen Mg II index is a potential driver representing SSIvariability in a model of cyclic evolution of the full-diskMg II h and k profiles.In the present work, we use a series of 91 IRIS full-Sunmosaics obtained over solar cycle 24 and select 76 mosaics toconstruct a state-of-the-art model of temporal evolution ofcores of the solar full-disk Mg II h and k profiles driven by theBremen Mg II index. We check and discuss the performanceand fidelity of the new model. To facilitate its utilization thecorresponding routines, written in IDL, are made publiclyavailable.5 Their description is given in Section 4.4. Finally, weprovide the 91 disk-averaged intensity profiles with uncertain-ties in units of W m−2 sr−1 nm−1 and mW m−2 sr−1 Hz−1 andthe spectral irradiances at 1 au with uncertainties in units ofmW m−2 nm−1,showingthetemporalvariationsoftheMg II h and k cores in solar cycle 24, in machine-readableformat online. This work is complementary to our previousworks on the quiet-Sun reference Lyα, Mg II h, and Mg II k lineprofiles presented in Gunár et al. (2020) and Gunár et al. (2021,hereafter, Paper I).2. Data2.1. IRIS Full-Sun MosaicsAn important aspect of operating spaceborne instruments isregular monitoring of changes in the instruments’ sensitivityover their lifetimes. In the case of the IRIS instruments it isachieved by a quasi-regular acquisition of spectral maps of theentire solar disk in both the near-UV (NUV) and far-UVspectral ranges and comparison with cotemporal measurementsby SORCE/SOLSTICE (Wülser et al. 2018) or the Solar EUVExperiment (Woods et al. 2005) on board the ThermosphereIonosphere Mesosphere Energetics and Dynamics spacecraft.We refer to the IRIS full-disk NUV spectral maps as the IRISfull-Sun mosaics or just mosaics.At the time of writing, the archive of IRIS full-Sun mosaics6contained 105 observations spanning the period from 2013September 30 to 2022 May 30 with the number of mosaics peryear as follows: 2013 (4 mosaics), 2014 (10), 2015 (13), 2016(14), 2017 (11), 2018 (11), 2019 (14), 2020 (14), 2021 (10),and 2022 (4). The footnote web address of the IRIS mosaicarchive is up to date and those in Paper I are outdated thoughthe data should be the same (B. De Pontieu 2021, privatecommunication). The only difference is that the up-to-dateaddress offers the mosaics in compressed files.In a typical year the first mosaic is usually taken at the end ofFebruary and the last one at the end of October (i.e., outside theIRIS eclipse season of November to February) except in the years2013 and 2014, when the first was taken in September and March,respectively. The source spectra of the mosaics are binned alongthe slit with binning factors of 2 (76 mosaics) and 4 (15). Thespectra are binned along the dispersion (i.e., along the wavelengthrange) with factors of 2 (76 mosaics) and 4 (15), yielding spectralscales of 51.2 mÅ px−1 and 102.4 mÅ px−1, respectively, andspectral resolutions of 106 mÅ and 212 mÅ, respectively (DePontieu et al. 2014, Table 1). By default the mosaic processingcode spectrally rebins Level 2 data to a default value6 of spectralscale of 35 mÅ px−1 common for all mosaics disregarding theoriginal spectral binning. Thus the spectral range of 3.5 Å,covered by each mosaic, is sampled by 101 points.In this paper we utilize the same data in Paper I, i.e., 91mosaics spanning the period from 2013 September 30 to 20205https://github.com/jkidl/IRIS6https://iris.lmsal.com/mosaic_index.html2The Astrophysical Journal Supplement Series, 261:17 (18pp), 2022 AugustKoza et al.October 19. Details on observational characteristics of themosaics, their processing, their radiometric calibration, andtheir uncertainties can be found in Paper I. The Solar InfluencesData Analysis Center (SIDC)7 defined the minimum, themaximum, and the end of solar cycle 24 for 2008 December,2014 April, and 2019 December, respectively. These dates willbe relevant when referring to the aspects of solar cycle 24.Figure 1 shows examples of mosaics in the Mg II k line center(k3) taken on 2014 March 17 and 2019 October 20 shortly beforethe maximum and the end of solar cycle 24. See also Figure 16showing the mosaic taken on 2014 May 27 shortly after themaximum.Variationsofthecorrespondingdisk-averagedMg II h and k profiles in days of high solar activity (2014 March17 and 24) and low activity (2019 October 20) are exemplified inFigure 2. An inspection of disk-averaged profiles from the yearsof low solar activity 2019–2020 shows that they practicallycoincide with that from 2019 October 20. However, Figure 2clearly demonstrates that during high solar activity the variation ofthe full-disk profiles is more pronounced. It illustrates that afull-disk profile measured on a day of high solar activity isnot representative of a time span of weeks or even an entireCarrington rotation.Figure 2 also illustrates an effect of spectral binning 4 on aresulting disk-averaged profile. While the profiles from 2014March 17 and 24 and 2019 October 20 (in orange, red, andblue, respectively) are constructed from mosaics taken withspatial and spectral binning 2 the quiet-Sun profiles from 2019October 15 (in purple) are constructed from mosaics withbinning 4. Although an extremely low level of solar activity iscommon for both days the effect of binning on the line shapesfrom 2019 October 15 is significant as compared to the profilesfrom October 20.Figure 1. The IRIS full-Sun mosaics in the Mg II k line center (k3) taken on 2014 March 17 shortly before the maximum of solar cycle 24 (left) and on 2019 October20 shortly before the cycle end (right). The x-axis and y-axis show the solar X and Y coordinates in arcseconds.Figure 2. Examples of disk-averaged Mg II h (left) and Mg II k (right) profiles from the IRIS full-Sun mosaics. The reference quiet-Sun Mg II h and k profilespublished in Paper I are practically identical to the profiles from 2019 October 20. The quiet-Sun profiles from 2019 October 15 (purple) are constructed from mosaicswith spectral binning 4 while the other profiles correspond to binning 2.7https://wwwbis.sidc.be/silso/cyclesminmax3The Astrophysical Journal Supplement Series, 261:17 (18pp), 2022 AugustKoza et al.2.2. Data AvailabilityA short example of data entering this analysis is given inTable 1. The table is published in its entirety in machine-readableformat. A portion is shown here for guidance regarding its formand content. The wavelength separation Δλ is centered at restwavelengths 280.3530 and 279.6352 nm of the Mg II h and klines, respectively. The example table lists the specific intensities Iand spectral irradiances Ee of the disk-averaged Mg II h and kprofiles with some characteristics of IRIS full-Sun mosaic Nos. 76and 77 taken on 2019 October 15 and 20 (shown in Figure 2): thespatial and spectral binning (Bin), the exposure time (texp), and theflag marking the nonreference (Ref = 0) and the reference mosaic(Ref = 1) used in the definition of the reference Mg II h and kprofiles in Paper I. The latter were taken on 2019 April 21, May27, July 27, September 12, September 22, and October 20 and on2020 March 2, March 23, April 22, June 21, August 23, andSeptember 6. They are marked in Figure 3. The spectralirradiances in Columns (c) are computed byp=⎜⎟⎛⎝⎞⎠( )EIau,1eN2whereN and au represent the nominal values of the solarradius and astronomical unit, respectively, and I is the disk-averaged intensity shown in Columns (a) and (b).2.3. Bremen Mg II IndexThe values of the Bremen Mg II index are listed at the LASPInteractive Solar Irradiance Data Center (LISIRD)8 operated atthe Laboratory for Atmospheric and Space Physics at a dailycadence but they are not daily averages. Actually, themeasurements are taken during a fraction of a day. The up-to-date minimum and maximum values of the index are0.14947 ± 0.00062 and 0.18005 ± 0.00268 for 1985 December2 and 1979 November 10, respectively.We already pointed out in Section 2.1 that a disk-averagedMg II h or k profile measured on a day of high solar activity isTable 1Disk-averaged Mg II h and k Profiles and Characteristics of IRIS Full-Sun MosaicsIRIS Full-Sun MosaicMg II hMg II kNo.YearMonthDayΔλIIEeσIIEeσBintexpRef(nm)(a)(b)(c)(%)(a)(b)(c)(%)(s)76201910150.16109112.3961.95217641.9951.922441076201910150.16459252.4362.90207621.9951.802441076201910150.16809302.4463.24207621.9951.842441076201910150.17159352.4563.58207621.9951.802441076201910150.17509402.4663.92207611.9851.72244107720191020−0.17509752.5666.28198022.0954.51232217720191020−0.17157882.0753.62237682.0052.20242217720191020−0.16805831.5339.63317241.8949.24262217720191020−0.16453921.0326.68446861.7946.62272217720191020−0.16104071.0727.70426951.8147.2727221Notes. (a) In units of W m−2 sr−1 nm−1. (b) In units of 10−7 mW m−2 sr−1 Hz−1. (c) In units of mW m−2 nm−1.(This table is available in its entirety in machine-readable form.)Figure 3. The evolution of the Bremen Mg II index (top) in solar cycle 24(SC24) and the integrated intensities E (bottom) of the IRIS Mg II h (red) andMg II k (blue) line profiles inferred from 91 IRIS full-Sun mosaics from theperiod2013–2020.Thedisk-averagedintensitiesEareinferredoverwavelength intervals of 1.0 Å centered at λ = 2803.530 Å and 2796.352 Åfor Mg II h and Mg II k, respectively. The vertical dashes mark the 12 daysselected for definition of the reference Mg II h and k profiles in Paper I. Theupside-down triangles mark the maximum and the end of solar cycle 24.8https://lasp.colorado.edu/lisird4The Astrophysical Journal Supplement Series, 261:17 (18pp), 2022 AugustKoza et al.not representative of a time span of weeks or even an entireCarrington rotation. The fact is also obvious from the highvariability of the Bremen Mg II index in years of high solaractivity (top panel in Figure 3). It means that an off-limbstructure observed on a day of high activity may experience anincident solar disk radiation different from that when thestructure is close to a visible disk center. Thus, in the modelingof an off-limb structurethe disk-averagedMg II h and kprofiles, dated roughly one week prior to or subsequent toobservation of the structure at the west or east limb,respectively, are relevant (see Sections 4.4 and 8).3. CorrelationsTo be able to reconstruct disk-averaged Mg II h and k profilesfor arbitrary dates our model needs a suitable driver for whichtemporal variations well reflect the SSI variability in time. In thissection we assess the suitability of the Bremen Mg II index as adriver of the sought model. We examine the relations between theindex and some characteristics of the IRIS disk-averagedMg II h and k profiles illustrated in Figures 3, 4, 5, and A1.To facilitate quick recognition of results pertaining to theMg II h and k lines, we use for them the red and blue colors,respectively, in Figure 3 and all following figures in this article,if relevant. Also we will keep a uniform layout of multipanelfigures placing Mg II h- and k-relevant panels on the left andright, respectively.3.1. Bremen Mg II Index versus IRIS Integrated IntensitiesFigure 3 compares the variations of the Bremen Mg II index(top panel) and the wavelength-integrated intensities E (thehistogram-style curves in the bottom panel) of the 91 disk-averaged Mg II h and k profiles inferred over the 1.0 Å wideintervals encompassing the emission cores. The comparisonimplies good agreement between the temporal variations of theparameters. The scatter plots in Figure 4 support the view oftight correlation. Table 2 lists the coefficients of linear fits ofthe intensities E(Mg II h) and E(Mg II k) integrated over theintervals of 1.0 and 3.5 Å. The correlation coefficients largerthan 0.9 in the last column confirm a very good correspondencebetween the integrated intensities of the emission cores and theFigure 4. Correlations of the integrated intensities E of the IRIS Mg II h (left) and Mg II k (right) line profiles with the Bremen Mg II index. The disk-averagedintensities E are integrated over wavelength intervals of 1.0 Å (circles) and 3.5 Å (plus signs) centered at λ = 2803.530 Å and 2796.352 Å for Mg II h and Mg II k,respectively, for the 91 IRIS full-disk mosaics from the period 2013–2020 (Figure 3). The linear fits are indicated by the red and blue lines. The fit parameters and thecorrelation coefficients are shown in Table 2.Figure 5. Correlations between Bremen Mg II index and the ratio betweenaveraged peak intensities and center intensities of the IRIS disk-averagedMg II h (red) and Mg II k (blue) line profiles. The data correspond to the IRISmosaics taken with spatial and spectral binnings 2 (empty circles) and 4 (filledcircles).5The Astrophysical Journal Supplement Series, 261:17 (18pp), 2022 AugustKoza et al.Bremen Mg II index. We note that from here on we use valuesof the Bremen Mg II index interpolated to the start time ofacquisition of the individual IRIS mosaics.3.2. Bremen Mg II Index versus Ratio of Peak Intensity toCenter IntensityA more complex view renders the comparison of the indexwith the ratio of the peak intensity to the center intensity0.5(Ih2v + Ih2r)/Ih3 and 0.5(Ik2v + Ik2r)/Ik3 shown in Figure 5.Here, Ih2v, Ih2r, Ik2v, and Ik2r represent the intensities of theviolet and red peaks of the disk-averaged Mg II h and k profilesand Ih3 and Ik3 stand for their center intensities. The range ofthe x-axis is intentionally set to 0.149−0.180 corresponding tothe full range of measured values of the Bremen Mg II index(Section 2.3). Figure 5 shows a strong bifurcation of the ratiosfor binnings 2 and 4 and suggests an anticorrelation betweenthe ratio and the Bremen Mg II index. This means that dayswith high index values and with bright active regions, coveringlarge portions of the solar disk, have lower ratios than quiet-Sun days with small index values. This is quite understandableas the brightness of active regions increases the Ih3 and Ik3intensities as compared to quiet-Sun days, when they areabsent. But most importantly, the correlation coefficients of theindex and the binning 2 ratios (empty circles in Figure 5) are infact −1 for both Mg II h and k. This suggests that the BremenMg II index perfectly represents both the wavelength-integratedintensities of the disk-averaged Mg II h and k profiles and thevariations of their spectral features. The anticorrelation of theindex and the peak-to-center ratio is documented and analyzedin detail by the Solar Bayesian Analysis Toolkit (SoBAT;Anfinogentov et al. 2021) in Appendix A.Conclusively, the online availability and the high correla-tions with the wavelength-integrated intensities and the peak-to-center ratios of the IRIS disk-averaged Mg II h and k profilesstimulate us to adopt the Bremen Mg II index as the driver ofthe new model. But we will check carefully its ability to specifyspectral shapes of disk-averaged Mg II h and k profiles inSection 5 as the most important feature of the model.4. Model ConstructionWe aim to provide an easily parameterized model that iscapableofreproducingdouble-peakedemissioncoresofMg II h and k profiles with central reversal and with differentpeak intensities (peak asymmetry) in a 3.5 Å wide spectralwindow captured by IRIS mosaics. In constructing a new modelof cyclic evolution of the solar disk–averaged Mg II h and kprofiles we build on the methods presented in Schmit et al. (2015)and Kowalska-Leszczynska et al. (2018). The uncertainties ofobservedintensities(PaperI)areconsistentlyconsideredthroughout the model construction in relevant fitting processes.4.1. Double-Gaussian Fit of the Solar Mg II h and k LinesInitially, to model the 91 disk-averaged Mg II h and k profileswe test a nine-parameter subtractive double-Gaussian modelfunction, represented by Equation (1) in Schmit et al. (2015).For the fitting the SolarSoft function mpfitfun.pro is used.The function calls the core procedure mpfit.pro, whichperforms a Levenberg–Marquardt least-squares minimizationof merit function (Moré 1978; Moré & Wright 1993;Markwardt 2009). We find out that the model functioninvolving Gaussians with positive and negative amplitudesdoes not yield a satisfactory and accurate fit of many observedprofiles. Therefore, we choose an additive double-Gaussianfunction also with nine parameters, which sums two Gaussianssuperimposed on a background represented by a linearfunction. The model function is of the formålllslD=- D- D++D-=⎜⎟⎧⎨⎩⎛⎝⎞⎠⎫⎬⎭()∣∣( )IAabcexp2,2iiiimod122where Ai, Δλi, and σi stand for the peak amplitudes, positions,and widths of the two Gaussian components and the parametersa, b, and c represent the shape of the background. Thecomponentsandparametersofthemodelfunctionareschematically illustrated in Figure 6. Also shown is thecommon marking of the prominent spectral features I1, I2, andI3 of the Mg II h profile (h) in its violet (v) and red (r) portions.The same goes for Mg II k identified with (k).To avoid the Mn I 2801.907 Å line in the violet wing of Mg II h(Figure 7) the fitting is limited within the wavelength range of−1.475 Å to 1.7 Å around the Mg II h line center. The entirewavelength range of ±1.75 Å is considered in fitting Mg II kincluding the weak line Mn I 2795.641 Å in its violet wing(Table 2 in Pereira et al. 2013). We find out that the latter does notinterfere with the fitting process and does not bias the final results.To enhance the accuracy of final fits at the local extremes I2 andI3 we introduce weighting of the input intensities in the meritfunction. Weights of 100 are assigned to the extreme intensitiesgiving them in the fitting process much higher importance thanthat given to other spectral points. In general, this approachrenders very satisfactory results exemplified in Figure 7 togetherTable 2Parameters of the Linear Fit y = q + kx and the Correlation Coefficients (CC)between the Disk-averaged Integrated Intensities E(Mg II h) and E(Mg II k) andthe Bremen Mg II Index in Figure 4Spectral LineIntervalqkCC(Å)(W m−2 sr−1)Mg II h1.0−31629140.943.5−26237300.84Mg II k1.0−40937470.933.5−34042580.85Figure 6. Sketch of the components and parameters of the double-Gaussian fitgiven by Equation (2) used to approximate the observed Mg II h and k profiles.The thick black line is the final profile. Prominent spectral features I1, I2, and I3of the Mg II h profile (h) in its violet (v) and red (r) portions are indicated. Thesame goes for Mg II k identified with (k).6The Astrophysical Journal Supplement Series, 261:17 (18pp), 2022 AugustKoza et al.with the uncertainties of the observed intensities. The fitting byEquation (2) yields 91 × 9 resulting parameters.4.2. Model of Evolution of the Solar Mg II h and k ProfilesAs a result of fitting individual profiles, we obtain 91 sets ofparameters, each describing different Mg II h and k profilesobserved by IRIS in different phases of solar cycle 24characterized by the Bremen Mg II index. The values of thefitting parameters for the Mg II h and k lines are shown inFigures 8 and 9, respectively, as functions of the index. Themagnitudes of their uncertainties, obtained from the fitting inSection 4.1, are comparable to or smaller than the size of thecircles. The figures intentionally discriminate between para-meters of profiles with binnings 2 (black empty circles) and 4(gray filled circles). Apparently, the parameter values are splitinto two subsamples depending on the binning. Becausebinning 4 strongly modifies the intensities of local extremes I2and I3, as already shown in Figure 2, and consequentlybifurcates the sample of parameter values (Figures 8 and 9),only the parameters of 76 profiles with binning 2 are consideredin constructing the model. The model parameter values inFigures 8 and 9 clearly show a correlation with the BremenMg II index. We see that the correlation of all parameters withthe index in the shown range of values 0.150−0.165 can beapproximated by a linear function. We fit each of the nineparameters Pi (i = 1...9) with the linear function Pi = αi + βiBI,where Pi = {A1, Δλ1, σ1, A2, Δλ2, σ2, a, b, c}, αi and βirepresent the fit parameters listed in Table 3, and BI is theBremen Mg II index. The fits are shown in Figures 8 and 9 bythe red and blue lines, respectively. Conclusively, the set of theFigure 7. Disk-averaged intensities (circles) and uncertainties (shaded areas) of the Mg II h (left) and Mg II k (right) line profiles from the IRIS full-Sun mosaics takenon 2014 March 17 (top), 2014 March 24 (middle), and 2019 October 20 (bottom) close to the maximum and the end of solar cycle 24. The observations (circles) areoverplotted with their double-Gaussian fits (red and blue lines) computed by Equation (2).7The Astrophysical Journal Supplement Series, 261:17 (18pp), 2022 AugustKoza et al.parametersa b=(),ii i1 ... 9 represents the resulting model ofcyclic variability of cores of the full-disk Mg II h and k profiles.4.3. Peak Amplitudes A1 and A2 and the Parameter cCloser inspection of the evolution of peak amplitudes A1 and A2(top left and middle left panels of Figures 8 and 9) reveals distinctchanges with increasing solar activity—namely a rise of violetpeak amplitude ΔA1 of approximately 50 W m−2 sr−1 Å−1 isfollowed by a rise of red peak amplitude ΔA2 of about 70–100W m−2 sr−1 Å−1. Thus, growth of activity decreases asymmetryin peak intensities. We can also observe a redshift of the parameterc toward zero with increasing activity in the bottom right panels ofFigures 8 and 9. This effect may be due to the symmetrization ofthe entire profile with the center of gravity being identified withthe parameter c and not with the zero-point of the relativewavelength scale.4.4. Availability of the IDL RoutinesTo facilitate utilization of the presented model in retrievingdisk-averaged Mg II h and k profiles for any date starting from1987 November 79 and for any position on the solar disk theIDL procedures get_mgii_hk.pro and get_mgii_hk_-when_at_meridian.pro with auxiliary IDL routines aremade publicly available at https://github.com/jkidl/IRIS. Theroutines require IDL version 8.2.1 or higher and installation ofthe SolarSoft package (Freeland & Handy 1998, 2012). For agiven date the procedure get_mgii_hk.pro returns in anoutput structure (i) Mg II h and k specific intensities in units ofW m−2 sr−1 Å−1 and mW m−2 sr−1 Hz−1, (ii) Mg II h and kspectral irradiances at 1 au in units of mW m−2 nm−1, and(iii) the Bremen Mg II index and its uncertainty adopted fromthe LISIRD database.Thesecondroutineget_mgii_hk_when_at_meri-dian.pro takes into account the heliocentric Cartesiancoordinates (Thompson 2006) of an off-limb or on-diskstructure on the date and time of its observation and returnsthe disk-averaged Mg II h and k profiles on the date when apotential counterpart of the structure is at the central meridian.Theroutinereliesonthesynodicrotationcoefficientsdetermined from small magnetic features by Howard et al.(1990). After finding the date at the meridian the routine callsget_mgii_hk.pro and returns an output structure with thesame quantities specified in items (i)–(iii) in the previousparagraph. The heliocentric Cartesian coordinates of a structurecan be found by the visualization tool JHelioviewer by Mülleret al. (2017). Alternatively, they can be found by the IDLroutine sdo_featurelocator.pro or gong_feature-locator.pro in the IDL library rridl developed by R. J.Rutten.10,11 For example, for the quiescent prominence withoff-limb heliocentric Cartesian coordinates (x, y) = (818″,Figure 8. Correlations between the Bremen Mg II index and the parameters of the double-Gaussian model (Equation (2)) obtained by fitting Mg II h profiles withspectral binnings 2 (black circles) and 4 (gray filled circles). The size of parameter uncertainties is comparable to or smaller than the size of the circles. The solid linesare linear fits of the parameters of the Mg II h profiles (Table 3) with spectral binning 2. They represent the model of evolution of Mg II h profiles in solar cycle 24.9Start date of the Bremen Mg II index in the LISIRD database.10 https://robrutten.nl/Recipes_IDL.html11 https://robrutten.nl/rridl/00-README/sdo-manual.html8The Astrophysical Journal Supplement Series, 261:17 (18pp), 2022 AugustKoza et al.597″) observed at the western limb on 2013 October 22 at7:30 UT the disk-averaged Mg II h and k profiles from 2013October 15 are relevant. Because the coordinates are ad hocslightly off-limb, the routine uses their radial projection (778″,568″) back to the nearest point on the limb. The date when acounterpart of the structure is at the meridian can be foundalso by the JHelioviewer itself and then entered into get_mgii_hk.pro.5. Model AccuracyIn checking the accuracy of the new model we utilize a greatasset of IRIS data, which is the availability of all inputs forreliably estimating uncertainties of IRIS intensities (Paper I). Theperformance of the new model is demonstrated in Figure 10,where the observed IRIS profiles, the same as those in Figures 2and 7, are overplotted with the model functions (Equation (2))whose parameters are computed by the coefficients α and β inTable 3 for BI values of 0.16174, 0.16426, and 0.15101corresponding to the dates 2014 March 17, 2014 March 24, and2019 October 20, respectively. The figure shows that thereconstruction of the profiles from 2014 March 17 (top panels)and 2019 October 20 (bottom panels) renders intensities mostlywithin the uncertainties of the observations. But the modelapparently underestimates the peak intensities of profiles from2014 March 24 (see middle right panel showing Mg II k). Butwhat is the performance of the new model in general? Is it capableof reproducing accurately the peak and center intensities I2 and I3?To answer these questions we construct three histograms (shownin Figures 11–13) allowing us to assess the new modelstatistically.ThehistograminFigure11showsthenumberofoccurrences of reconstructed Mg II h and k profiles with thetypical deviation of model intensities Δtyp across the entirefitting range given by the formulaållslD=D-DD=( )()()()( )tNItItt1,,,,3iNiiityp1obsmodobswhere Iobs(Δλi, t) andlD()It,imodrepresent the observed andmodel intensity (Equation (2)), respectively, at spectral positionΔλi within the full-disk profile on date t. The value ofσobs(Δλi, t) is the uncertainty of intensity at spectral positionΔλi and N is the number of spectral positions considered in thedouble-Gaussian fitting, namely 91 and 101 for Mg II h and k,respectively (see Section 4.1). Due to normalization withrespect to N and σobs(Δλi, t), the values of Δtyp(t) showwhetherlD()It,imodistypicallywithintheuncertaintyinterval, |Δtyp| 1, or outside it with |Δtyp| > 1. The histogramin Figure 11 implies that all but one model profiles of both linesqualifyasa|Δtyp| 1typewithmostcaseshaving|Δtyp| 0.4, and thus with the model intensities Imod safelywithin the uncertainty intervals.The histogram in Figure 12 demonstrates the accuracy of themodel in reproducing the peak intensities I2. It shows thenumber of occurrences of reconstructed Mg II h and k profileswith the deviation of the model intensities Δ2 given by theFigure 9. The same as Figure 8 but for Mg II k.9The Astrophysical Journal Supplement Series, 261:17 (18pp), 2022 AugustKoza et al.formulasD= áñ- áñ( )( )( )( )( )tItItt,422,obs2,mod2,obswhere 〈I2,obs〉 and áñI2,mod are the average peak intensity of theobserved and model profile, respectively. They are defined as〈I2〉 = 0.5(I2v + I2r), where I2v and I2r are the intensity of theviolet and red peaks, respectively. The value of σ2,obs is takenas the minimum of the uncertainties of I2v,obs and I2r,obs, i.e.,sss=( ){( )( )}tttmin,2,obs2v,obs2r,obs. The histogram impliesthat 95% of the Mg II h model profiles and 91% of the Mg II kmodel profiles are characterized by a deviation |Δ2| 1. Thismeans that the model intensities I2,mod of both lines are mostlywithin the interval of the observation uncertainties σ2,obs.The histogram in Figure 13 illustrates the accuracy of themodel in reproducing the center intensities I3. It shows thenumber of occurrences of reconstructed Mg II h and k profileswith the deviation of the model intensities Δ3 given by theformulasD=-( )( )( )( )( )tItItt533,obs3,mod3,obswhere I3,obs and I3,mod are the center intensity in the reversal ofthe observed and model profiles, respectively, and σ3,obs is theuncertainty. The histogram implies that 96% of the Mg II hmodel profiles and 93% of the Mg II k model profiles arecharacterized by a deviation |Δ3| 1. Thus the modelintensities I3,mod of both lines are mostly within the intervalof the observation uncertainties σ3,obs.Finally, we would like to remind the reader that the modeland the presented analysis of its accuracy pertain to a range ofBremen Mg II index values of 0.150–0.165 while the up-to-datemeasured range is 0.149–0.180 (Section 2.3). Thus the modelshould be taken with caution for index values higherthan 0.165.6. Reconstruction of SSIWedemonstratedinSection5thatthenewmodelaccomplishes a credible reconstruction of the spectral shapesof disk-averaged Mg II h and k profiles relevant for radiative-transfer modeling. In this section we aim to test the ability ofthe model to reconstruct also the SSI over longer time spansand to compare it with the benchmark values measured by theSORCE/SOLSTICE instrument over its mission lifetime from2003 April 4 until 2020 February 26 available through theLISIRD database.The top panels in Figure 14 show the IRIS spectralirradiances reconstructed by the model in Table 3 using dailyvalues of the Bremen Mg II index. The relevant Mg II h and kprofile intensities are converted into spectral irradiances byEquation (1) and integrated over intervals of 1 Å and 1.75 Å,both with midpoints at the line centers. The integration limits ofthe former coincide roughly with the local minima k1(Figures 6, 7, and 10). The integration limits of the seeminglyad hoc interval of 1.75 Å have no correspondence to theproperties of the Mg II h and k profiles observed by IRIS but aredictated by the spectral resolution limitations of the SORCE/SOLSTICE instrument. The middle panels of Figure 14 showthe spectral irradiances computed from measurements bySORCE/SOLSTICE. The observed Mg II h and k profiles areintegrated over the wavelength intervals of 1.75 Å indicated bythe short vertical dashes at the bottom x-axes in the bottompanels of Figure 14. Also here the integration limits coincidewith estimated positions of the local minima k1. The relativelylow spectral resolution of SORCE/SOLSTICE causes anabsence of central reversals in the Mg II h and k profiles. Theabsolute and relative differences of the wavelength-integratedspectral irradiances on the dates of their minima and maximaare summarized in Table 4. The listed values are inferred fromboxcar-averaged irradiances smoothed over 399 days.Comparison of the top and middle panels in Figure 14 shows(1) high correlations of data obtained by the instruments (bothdaily and smoothed) with the relevant correlation coefficientsequal to one, (2) an ability of the model to fill the data gap inSORCE/SOLSTICE measurements from mid-2013 to early2014, (3) that the reconstructed spectral irradiances seem tohave larger daily variations than the SORCE/SOLSTICEmeasurements likely because they capture mainly chromo-spheric variations while SORCE/SOLSTICE with its widerintegration interval may cover also the upper photosphere withsmaller daily variability of spectral irradiance, (4) that thereconstructed spectral irradiances are smaller than the SORCE/SOLSTICE measurements for both integration intervals likelydue to the difference in calibration of instruments documentedin Figure 1 in Paper I, and (5) a substantial reduction ofdifferences between the model-reconstructed and observedirradiances to ∼0.8 mW m−2 when using the integrationinterval of 1.75 Å (middle panels). Of course, the ad hocnumerical identity of the integration intervals does notguarantee an equivalence of relevant irradiances.An inspection of Table 4 shows that (1) the model-reconstructed irradiances feature greater absolute and relativeTable 3Coefficients αi and βi of the Function Pi = αi + βiBI between the Parameter Pi in Equation (2) and the Bremen Mg II IndexiPiUnitsMg II hMg II kαiβiαiβi1A1(W m−2 sr−1 Å−1)−404.03983.0−509.05118.02Δλ1(mÅ)−151.5−37.3−198.2173.23σ1(mÅ)88.485.969.6270.64A2(W m−2 sr−1 Å−1)−501.04388.0−607.05470.05Δλ2(mÅ)109.7284.1130.4251.36σ2(mÅ)−7.3673.733.1476.97a(W m−2 sr−1 Å−1)−39.0436.0−41.0401.08b(W m−2 sr−1 Å−2)49.0−47.051.0−105.09c(mÅ)−247.61411.8−271.11245.710The Astrophysical Journal Supplement Series, 261:17 (18pp), 2022 AugustKoza et al.Figure 10. The same as Figure 7 but the observations (circles) are overplotted with the reconstructed profiles (red and blue lines) from the model parameters inTable 3.Figure 11. Histogram of typical deviations Δtyp (Equation (3)) of the modelMg II h and k profiles (red and blue, respectively), computed by the parametersin Table 3, and the observed disk-averaged profiles. The data correspond toIRIS mosaics taken with spatial and spectral binning 2.Figure 12. Histogram of deviations Δ2 (Equation (4)) of the averaged peakintensities 〈Ih2v,r〉 and 〈Ik2v,r〉 of the model Mg II h and k profiles (red and blue,respectively), computed by the parameters in Table 3, and of the observations.The data correspond to IRIS mosaics taken with spatial and spectral binning 2.11The Astrophysical Journal Supplement Series, 261:17 (18pp), 2022 AugustKoza et al.differences between the minimum and maximum values thanthe observed irradiances and (2) the dates of the minima andmaxima of the irradiances are offset by about several monthsfor both instruments and also in comparison to the minimumand maximum of solar cycle 24 (Section 2.1).7. Treatment of Spectral SmearingAn effect of spectral smearing is inherent also in all IRISspectra. It needs to be taken into account particularly incomparingobservationswithresultsofradiative-transfermodeling via convolving the latter by an instrumental profile.Gaussian shapes of the instrumental profile of the IRIS NUVchannel with slightly different FWHMs of 60, 52, 52, and50.54 mÅ were assumed in spectral modeling by Pereira et al.(2013), Heinzel et al. (2015), Jejčič et al. (2018), and Tei et al.(2020), respectively. These values are very close to the spectralresolution of 53 mÅ of the IRIS NUV channel (De Pontieuet al. 2014, Table 1).To assess the effect of spectral smearing on the disk-averaged Mg II h and k profiles treated throughout this paper(Figures 2, 6, 7, and 10; Table 1) and provided through the IDLprocedure get_mgii_hk.pro (Section 4.4), the double-Gaussian model function (Equation (2)) is modified appro-priately. It is assumed that both Gaussians result from aconvolution with a Gaussian-shaped instrumental profile,having a width s=()FWHM2 2 ln 2IP, with no influenceon the linear term a + b|Δλ − c|. Then an analytic restorationof the model function can be performed (in Appendix B) by theconvolution theorem (e.g., Gray 2008) yielding the formulaållD=´-++D-sssllss=-D- D-⎧⎨⎩⎛⎝⎞⎠⎫⎬⎭()∣∣( )Aabcexp.6iimod1222iiii2IP22IP2An effect of restoration by this formula is exemplified by thereconstructed Mg II h and k profiles in Figure 10 from 2019October 20 and 2014 March 17, i.e., for days close to the endand the maximum of the solar cycle, respectively. The profiles,restored by Equation (6), are shown in Figure 15 by the blackdotted lines together with the relative differences in the bottompanels. Here we assume the width of the IRIS instrumentalprofile σIP = 22.5 mÅ corresponding to a FWHM = 53 mÅ.The differences of a few percent prove that the effect of spectralsmearing can be safely neglected for the results presented hereand also in Paper I. In this context we note that the adoptedvalue of σIP = 22.5 mÅ is much smaller than the widthsσ1,2 ∼ 100 mÅ of the Gaussian components in Figures 8 and 9.Another reason to neglect spectral smearing here is anuncertainty in the shape and width of the instrumental profile ofthe IRIS NUV channel. This is documented by the right panelin Figure 13 of De Pontieu et al. (2014) showing a histogram ofthe full width at half-minimum of a weak absorption line ofMn I 2801.907 Å (Pereira et al. 2013, Table 2) measured in aquiet-Sun region. The histogram shows the occurrence ofprofiles with the widths down to 40 mÅ suggesting that theIRIS NUV instrumental profile may be narrower than theassumed value of 53 mÅ. We conclude this section by recallingthat spectral smearing cannot be neglected in cases when therelevant widths of synthetic Mg II h and k profiles are compar-able with the width of the instrumental profile as in Heinzelet al. (2015), Jejčič et al. (2018), and Tei et al. (2020).8. DiscussionIt has been already pointed out in Kowalska-Leszczynskaet al. (2018) that employment of the composite Lyα indexinvokes an ambiguity issue common also for the Bremen Mg IIindex. Different disk-averaged profiles may correspond to thesame value of the index, depending on the coverage of the solardisk with active regions and filaments. Similarly, because thedefinition of the Bremen Mg II index involves summingspectralirradiances overseveralwavelengths (see, e.g.,Equation (1) in Snow et al. 2019), disk-averaged Mg II h and kprofiles with different spectral shapes may result in the samevalue of the Bremen Mg II index. The testing of the modelaccuracy in Section 5 (Figures 12 and 13) suggests that theissue may become apparent in less than one profile reconstruc-tion out of ten.The possible heliolatitude variation of the disk-integratedMg II h and k line profiles and the Bremen Mg II index is, to ourknowledge, unexplored. Some form of variation should beexpected because, as shown by Schmit et al. (2015), the Mg II hprofile depends on features on the solar disk that are beingobserved and the latitude distribution of these features isinhomogeneous and varies during the solar cycle. Should thedisk-integrated Mg II h and k profiles indeed vary with heliola-titude, this would potentially have consequences for thedefinition of incident radiation for radiative-transfer calcula-tions. However, a recent investigation of the variation of thesolar spectrum with heliolatitude by Kiselman et al. (2011)seems to have brought a negative result, i.e., no variation(Bzowski et al. 2013). A mild heliolatitude dependence of disk-and wavelength-integrated Lyα irradiance is considered inKowalska-Leszczynska et al. (2018, Equation (2)) with refer-ences on previous theoretical and observational studies. Thismight be relevant for Gunar et al. (2020).Exact incident radiation, affecting the mean intensity J, israrely available. The geometry and formalism of the problemare outlined in Heinzel (1983, Figure 1), Heinzel & Rompolt(1987), and Sahal-Brechot et al. (1986, Figures 2 and 3). Atarget is illuminated from all directions and this illuminationvaries spatially. Ideally, one should take into account an actualFigure 13. Histogram of deviations Δ3 (Equation (5)) of center intensities Ik3,h3 of the model Mg II h and k profiles (red and blue, respectively), computed bythe parameters in Table 3, and of the observations. The data correspond to IRISmosaics taken with spatial and spectral binning 2.12The Astrophysical Journal Supplement Series, 261:17 (18pp), 2022 AugustKoza et al.brightness topology of underlying solar features visible from analtitude of the target (e.g., prominences or flare loops)illuminating it at the time of observation (Vial et al. 2019,Section 2.2). For off-limb targets this is fully practicable onlywith the availability of simultaneous stereoscopic observations(Vial et al. 2016). Otherwise, one should use the incidentradiation dated roughly one week prior to or subsequent totarget observation at the west or east limb, respectively(Section 4.4).The problem of incident radiation is consistently solved inZhang et al. (2019, Section 4) in the case of the plasmadiagnostics of an eruptive prominence observed on 2014 May28. Incident radiation is represented by the actual brightnesstopology inferred from the IRIS full-disk mosaic taken fromFigure 14. Top and middle panels: IRIS spectral irradiances reconstructed by the model in Table 3 (gray dots in top panels) and computed by the SORCE/SOLSTICEobservations (gray dots in middle panels) taking integration intervals Λ of 1 Å and 1.75 Å. The red and blue lines represent boxcar-averaged values smoothed over 399days. The different line styles distinguish between the intervals of 1 Å (solid in top panels) and 1.75 Å (dotted in top and middle panels) for the IRIS modelreconstructions and of 1.75 Å for the SORCE/SOLSTICE observations (solid in middle panels). The vertical dashes at the bottom x-axes mark the dates of minimaand maxima of the smooth irradiances by IRIS (top panels) and SORCE/SOLSTICE (middle panels) given in Table 4. The relative variations with respect to theminima are shown at the right y-axes. Bottom panels: The Mg II h and k profiles averaged over the mission lifetime of SORCE/SOLSTICE. The vertical dashes at thebottom x-axes mark the wavelength intervals of 1.75 Å considered in computing the integrated spectral irradiances in the middle panels.13The Astrophysical Journal Supplement Series, 261:17 (18pp), 2022 AugustKoza et al.May 27 to May 28. Although at that time the solar disk featuredseveral active regions (Figure 16) the prominence-relevantincident radiation came solely from a quiet area (Zhang et al.2019, Figure 15). We conclude this section by noting that in theabsence of an actual brightness topology the model presentedin Table 3 yields the best approximation of the incidentMg II h and k radiation currently available for targets observedon an arbitrary day of the solar cycle and at an arbitraryheliolatitude when the Bremen Mg II index is lower than 0.165.For its higher values the accuracy of the reconstructed incidentradiation is uncertain. An updated model, based on new datataken over a solar maximum stronger than that of solar cycle24, would be needed.9. ConclusionsBased on selected IRIS full-Sun mosaics in the NUVcovering almost the full solar cycle 24, we find an analyticrepresentation of the spectral shapes of Mg II h and k line coresparameterized by the Bremen Mg II index. The definition of themodel is given by Equation (2) and the parameters are given inTable 3. For a given value of the Bremen Mg II index, themodel provides the spectral shapes of Mg II h and k line cores.The uncertainties of the present model are difficult to assess.We estimate, however, that the model is accurate in more than90% of profile reconstructions for a set of selected BremenMg II indices from the range 0.150–0.165. By model accuracywe mean that the entire reconstructed profile is within theobservational uncertainties. A follow-up study in this serieswill address the issue of sensitivity of the radiative-transfermodels to Mg II h and k cyclical variability. Another possibleapplication of the series of calibrated IRIS mosaics is toreaddress the question whether the quiet Sun changes over thesolar cycle at the chromospheric layers (Solanki 2007).Measurements by White & Livingston (1981) showed nosignificant systematic variability in the intensity parameters ofthe Ca II K line core for quiet regions near the center of thesolar disk over the years 1975–1980 in the rising phase of solarcycle 21. It bears on the stability of the basic supergranulationTable 4Absolute ΔE and Relative ΔEr Differences of Wavelength-integrated Spectral Irradiances in Figure 14 on the Dates of Their Minima tmin and Maxima tmax for theIntegration Intervals ΛMg II hMg II kInstrument—MethodΛtmintmaxΔEΔErΔEΔEr(Å)(mW m−2)(%)(mW m−2)(%)IRIS—model reconstruction1.002009 Mar2014 Jul2.2262.927IRIS—model reconstruction1.752009 Mar2014 Jul2.4213.124SORCE/SOLSTICE—observations1.752008 Aug2014 Sep1.9162.720Figure 15. The reconstructed Mg II h and k profiles from Figure 10 with (red and blue lines) and without spectral smearing (black dotted lines) and their relativedifferences (bottom panels).Figure 16. The IRIS full-Sun mosaic in the Mg II k line center (k3) taken on2014 May 27 shortly after the maximum of solar cycle 24. The x-axis and y-axis show the solar X and Y coordinates in arcseconds.14The Astrophysical Journal Supplement Series, 261:17 (18pp), 2022 AugustKoza et al.flow process and the quiet network during a solar cycle. Thisconclusion can be reassessed by IRIS mosaics capturing quiet-Sun regions over longer time spans compared to previousmeasurements.Finally, we would like to warn the reader that our data can beused for specific problems, but other situations may requireanother approach. For example if the prominence is located justabove a quiet region (no activity around) at rather low heights,it will be illuminated predominantly by the quiet-Sun radiation,which is closer to the solar cycle minimum data. This can bethe case of, e.g., the polar crown filaments observed duringtimes of enhanced solar activity elsewhere on the disk.J.K. and P.S. acknowledge the project VEGA 2/0048/20. J.K. acknowledges support from grant No. 19-17102S of theCzech Science Foundation (GAČR). S.G., P.H., and P.S.acknowledge support from grant Nos. 19-16890S and 19-17102S of the Czech Science Foundation (GAČR). S.G.acknowledges support from grant No. 19-20632S of the CzechScience Foundation (GAČR). S.G., P.S., P.H., and J.K.acknowledge support from the Joint Mobility Project SAV-18-03 of the Academy of Sciences of the Czech Republic andthe Slovak Academy of Sciences. S.G., P.H., and W.L.acknowledge support from project RVO:67985815 of theAstronomical Institute of the Czech Academy of Sciences. IRISis a NASA small-explorer mission developed and operated byLMSAL with mission operations executed at the NASA AmesResearch Center and major contributions to downlink commu-nications funded by the European Space Agency and theNorwegian Space Center. The IRIS full-Sun mosaics areavailable at iris.lmsal.com/mosaic.html. We acknowledge theuse of data obtained from LISIRD available at lasp.colorado.edu/lisird. This research has made use of NASAʼs Astro-physics Data System Bibliographic Services. The authorsacknowledge important suggestions from the anonymousreferee and the data editor who helped to improve themanuscript.Facilities: Interface Region Imaging Spectrograph (IRIS),SORCE/SOLSTICE.Software: IDL,SolarSoft(SSW; Freeland& Handy1998,2012),LISIRD(https://lasp.colorado.edu/lisird),SoBAT(https://github.com/Sergey-Anfinogentov/SoBAT).Appendix AOn the Anticorrelation between the Mg II h and k Peak-to-center Intensity Ratio and the Bremen Mg II IndexThe almost perfect anticorrelation between the ratio of the peakintensity to the center intensity and the Bremen Mg II index(Table A2) motivated us to look for a best-fit model of the datawith binning 2 shown in Figures 5 and A1. Parameter inference ofthe linear model y = q + kx and the ad hoc noninteger powermodel y = r + a(x − 0.15)p was carried out with the help of theMarkov Chain Monte Carlo (MCMC) sampling algorithmimplemented in the IDL toolkit SoBAT by Anfinogentov et al.(2021). We ran the MCMC fitting of both models with 105samples after the initial burn-in stage with 5 × 104 samples. Thedefinitions of the initial guesses of model parameters and thepriors, common for both the Mg II h and k data, are summarized inTable A1. Figure A2 shows histograms approximating themarginalized posterior distributions of the model parametersobtained by the MCMC sampling and estimates of the maximuma posteriori probability and the uncertainties by 95% credibleintervals. The Bayesian parameter inferences and the Bayesianfactors Kpow/lin are summarized in Table A2. Note the relativelybroad credible intervals of the parameter a. The corresponding fitsby the linear model and the noninteger power model are displayedby the solid and dashed lines, respectively, in Figure A1,intentionally for the full range of measured values of the BremenMg II index 0.149–0.180 (Section 2.3).We carried out several trial runs of the MCMC fitting with thesame parameters in Table A1. They yielded the histograms inFigure A2 with slightly different shapes and Bayesian factorsoscillating around 2. Thus Figure A2 and the factors in Table A2are only illustrative examples. The factor Kpow/lin quantifies thelevel of evidences of the two models. Its values are close to thelimit between inconclusive and positive evidence to the nonintegerpower model in front of the linear model (Arregui 2018, Table 1).Therefore, the limited data currently available, covering only halfof the full range of measured values of the Bremen Mg II index,does not allow us to prefer one model over the other. However,the models predict distinct trends of the ratio for index valueslarger than 0.165 (Figure A1). While the linear model predictsa divergence of the ratios for Mg II h and k toward a value ofFigure A1. Correlations between Bremen Mg II index and the ratio betweenaveraged peak intensities and center intensities of the IRIS disk-averagedMg II h (red) and Mg II k (blue) line profiles. The data correspond to the IRISmosaics taken with spatial and spectral binnings 2 (empty circles) and 4 (filledcircles). The solid and dashed lines represent data fits by the linear and thenoninteger power function, respectively.Table A1Initial Parameters and Priors of the MCMC FittingModelParameterInitialGuessPrior TypePriorParameterslineark3.0uniform1.0, 5.0q−7.5normal−7.5, 3.0nonintegerpowerr1.65uniform0.5, 2.0a−25.0uniform−60.0, 0.0p1.2uniform0.5, 2.015The Astrophysical Journal Supplement Series, 261:17 (18pp), 2022 AugustKoza et al.Figure A2. Histograms approximating marginalized posterior distributions of parameters of the linear model y = q + kx and the noninteger power model y = r + a(x − 0.15)p of the Mg II h (left) and Mg II k (right) data in Figure A1 obtained by 105 MCMC samples. The solid and dashed vertical lines indicate estimates of themaximum a posteriori probability and the uncertainties by 95% credible intervals, respectively (Table A2).16The Astrophysical Journal Supplement Series, 261:17 (18pp), 2022 AugustKoza et al.roughly 1.4, the noninteger power model implies a common trendtoward ∼1.35 at a maximum index value of 0.180. This willhopefully enable us to test the models in the future when theobserved Mg II h and k peak-to-center intensity ratios will beavailable for high solar activity characterized by a range ofBremen Mg II index values of 0.165–0.180.Appendix BRestoration FormulaAssume that a Gaussian component G(x) in Equation (2) is aresult of convolution given by the formula=*( )( )( )()G xxx ,B1Nwhere ( )x is a spectrally unsmeared Gaussian and ( )xNis theunit-area instrumental profile sometimes referred to also as theapparatus function (e.g., Brault & White 1971). For simplicity,here we write x instead of Δλ and without loss of generality wealso assume that the Gaussian peak is at Δλ = 0. The left sideof Equation (B1) can be written asss p=-=⎧⎨⎩⎛⎝⎞⎠⎫⎬⎭( )( )()G xAxAGxexp22,B22Nwhere A and σ are of the same meaning as in Equation (2) andGN(x) is the unit-area Gaussian given bys ps=-⎧⎨⎩⎛⎝⎞⎠⎫⎬⎭( )()Gxx12exp2.B3N2Its Fourier transform yields the formula (Gray 2008)ps=-~ ( ){() }()Gssexp2.B4N2The unit-area instrumental profile in Equation (B1) can berepresented by the formulasps=-⎜⎟⎧⎨⎩⎛⎝⎞⎠⎫⎬⎭( )()Axx12exp2B5NIPIP2with the Fourier transformps=-~ ( ){() }()Assexp2,B6NIP2where σIP is the width of the instrumental profile related to itsFWHM by the conversion formula s=()FWHM2 2 ln 2IP.Thus Equation (B1) can be rewritten with the help of the rightside of Equation (B2) ass p=*( )( )( )()GxAxx12.B7NNWe are seeking the function ( )x . To do so Equation (B7) canbe represented by the convolution theorem (Gray 2008) ass p=~~( )( )( )()GsAss12,B8NNwhere ( )sis the Fourier transform of ( )x . RecastingEquation (B8) and using Equations (B4) and (B6) one canobtains ps ppss==--~~( )( )( ){() }()sAGssAs22exp2.B9NN2IP22Its inverse Fourier transform yields the sought formula for theunsmeared Gaussian used in Equation (6):sssss=---⎧⎨⎩⎛⎝⎜⎞⎠⎟⎫⎬⎭( )()xAxexp2.B102IP22IP22Finally, we note thatòòs p==-¥¥-¥¥( )( )()G x dxx dxA2.B11ORCID iDsJúlius Kozahttps://orcid.org/0000-0002-7444-7046Stanislav Gunárhttps://orcid.org/0000-0003-3889-2609Pavol Schwartzhttps://orcid.org/0000-0001-5986-9948Petr Heinzelhttps://orcid.org/0000-0002-5778-2600Wenjuan Liuhttps://orcid.org/0000-0002-2788-0760ReferencesAlissandrakis, C. E., Vial, J. C., Koukras, A., Buchlin, E., & Chane-Yook, M.2018, SoPh, 293, 20Anderson, D. E., & Cahalan, R. F. 2005, SoPh, 230, 3Anfinogentov, S. A., Nakariakov, V. M., Pascoe, D. J., & Goddard, C. R. 2021,ApJS, 252, 11Arregui, I. 2018, AdSpR, 61, 655Ayres, T., De Pontieu, B., & Testa, P. 2021, ApJ, 916, 36Barczynski, K., Schmieder, B., Peat, A. W., et al. 2021, A&A, 653, A94Brault, J. W., & White, O. R. 1971, A&A, 13, 169Bryans, P., McIntosh, S. W., Brooks, D. H., & De Pontieu, B. 2020, ApJL,905, L33Bzowski, M., Sokół, J. M., Tokumaru, M., et al. 2013, Cross-Calibration of FarUV Spectra of Solar System Objects and the Heliosphere (Berlin:Springer), 67De Pontieu, B., Polito, V., Hansteen, V., et al. 2021, SoPh, 296, 84De Pontieu, B., Title, A. M., Lemen, J. R., et al. 2014, SoPh, 289, 2733Floyd, L., Tobiska, W. K., & Cebula, R. P. 2002, AdSpR, 29, 1427Freeland, S. L., & Handy, B. N. 1998, SoPh, 182, 497Freeland, S. L., & Handy, B. N. 2012, SolarSoft: Programming and DataAnalysis Environment for Solar Physics, ascl:1208.013Gray, D. F. 2008, The Observation and Analysis of Stellar Photospheres(Cambridge: Cambridge Univ. Press)Table A2Correlation Coefficients between the Ratio of Peak Intensity to Center Intensity and the Bremen Mg II Index for the Data with Binning 2 in Figures 5 and A1Spectral LineCCqkrapKpow/linMg II h−0.95-+2.598 0.0810.078--+6.536 0.5050.526-+1.611 0.0030.003--+32.341 26.22816.916-+1.371 0.1730.1392.82Mg II k−0.98-+3.095 0.0710.071--+9.531 0.4600.459-+1.661 0.0050.004--+18.410 16.5587.766-+1.154 0.1260.1493.16Note. Bayesian parameter inferences by the linear fit y = q + kx and the noninteger power fit y = r + a(x − 0.15)p, and the corresponding Bayesian factor Kpow/lin.Posterior summaries are given by the maximum a posteriori probability and uncertainties by 95% credible intervals.17The Astrophysical Journal Supplement Series, 261:17 (18pp), 2022 AugustKoza et al.Gunár, S., Koza, J., Schwartz, P., Heinzel, P., & Liu, W. 2021, ApJS, 255, 16Gunár, S., Schwartz, P., Koza, J., & Heinzel, P. 2020, A&A, 644, A109Heath, D. F., & Schlesinger, B. M. 1986, JGR, 91, 8672Heinzel, P. 1983, BAICz, 34, 1Heinzel, P. 2015, Solar Prominences (Cham: Springer), 103Heinzel, P., & Rompolt, B. 1987, SoPh, 110, 171Heinzel, P., Schmieder, B., Mein, N., & Gunár, S. 2015, ApJL, 800, L13Heinzel, P., Vial, J. C., & Anzer, U. 2014, A&A, 564, A132Howard, R. F., Harvey, J. W., & Forgach, S. 1990, SoPh, 130, 295Jejčič, S., Schwartz, P., Heinzel, P., Zapiór, M., & Gunár, S. 2018, A&A,618, A88Kayshap, P., Singh Payal, R., Tripathi, S. C., & Padhy, H. 2021, MNRAS,505, 5311Kiselman, D., Pereira, T. M. D., Gustafsson, B., et al. 2011, A&A, 535, A14Kowalska-Leszczynska, I., Bzowski, M., Kubiak, M. A., & Sokół, J. M. 2020,ApJS, 247, 62Kowalska-Leszczynska, I., Bzowski, M., Sokół, J. M., & Kubiak, M. A. 2018,ApJ, 852, 115Koza, J., Kuridze, D., Heinzel, P., et al. 2019, ApJ, 885, 154Kuridze, D., Socas-Navarro, H., Koza, J., & Oliver, R. 2021, ApJ, 908, 168Labrosse, N., Heinzel, P., Vial, J. C., et al. 2010, SSRv, 151, 243Levens, P. J., & Labrosse, N. 2019, A&A, 625, A30Markwardt, C. B. 2009, in ASP Conf. Ser. 411, Astronomical Data AnalysisSoftware and Systems XVIII, ed. D. A. Bohlender, D. Durand, & P. Dowler(San Francisco, CA: ASP), 251McClintock, W. E., Rottman, G. J., & Woods, T. N. 2005a, SoPh, 230, 225McClintock, W. E., Snow, M., & Woods, T. N. 2005b, SoPh, 230, 259Mikuła, K., Heinzel, P., Liu, W., & Berlicki, A. 2017, ApJ, 845, 30Moré, J., & Wright, S. 1993, Frontiers in Applied Mathematics: OptimizationSoftware Guide (Philadelphia, PA: SIAM)Moré, J. J. 1978, Numerical Analysis (Berlin: Springer), 105Müller, D., Nicula, B., Felix, S., et al. 2017, A&A, 606, A10Peat, A. W., Labrosse, N., Schmieder, B., & Barczynski, K. 2021, A&A,653, A5Pereira, T. M. D., Leenaarts, J., De Pontieu, B., Carlsson, M., & Uitenbroek, H.2013, ApJ, 778, 143Rottman, G. 2005, SoPh, 230, 7Ruan, G., Jejčič, S., Schmieder, B., et al. 2019, ApJ, 886, 134Sahal-Brechot, S., Malinovsky, M., & Bommier, V. 1986, A&A, 168, 284Schmit, D., Bryans, P., De Pontieu, B., et al. 2015, ApJ, 811, 127Schwartz, P., Gunár, S., & Curdt, W. 2015, A&A, 577, A92Snow, M., Machol, J., Viereck, R., et al. 2019, E&SS, 6, 2106Snow, M., Weber, M., Machol, J., Viereck, R., & Richard, E. 2014, JSWSC,4, A04Solanki, S. K. 2007, in ASP Conf. Ser. 368, The Physics of ChromosphericPlasmas, ed. P. Heinzel, I. Dorotovič, & R. J. Rutten (San Francisco, CA:ASP), 481Tei, A., Gunár, S., Heinzel, P., et al. 2020, ApJ, 888, 42Thompson, W. T. 2006, A&A, 449, 791Vial, J.-C., Pelouze, G., Heinzel, P., Kleint, L., & Anzer, U. 2016, SoPh,291, 67Vial, J. C., Zhang, P., & Buchlin, É. 2019, A&A, 624, A56White, O. R., & Livingston, W. C. 1981, ApJ, 249, 798Woods, T. N., Eparvier, F. G., Bailey, S. M., et al. 2005, JGRA, 110, A01312Wülser, J. P., Jaeggli, S., De Pontieu, B., et al. 2018, SoPh, 293, 149Zhang, P., Buchlin, É., & Vial, J. C. 2019, A&A, 624, A7218The Astrophysical Journal Supplement Series, 261:17 (18pp), 2022 AugustKoza et al. | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/44159004Variations in K/Th on MarsArticle in Journal of Geophysical Research Atmospheres · December 2006DOI: 10.1029/2006JE002676 · Source: OAICITATIONS97READS44927 authors, including:Some of the authors of this publication are also working on these related projects:selenophycis View projectOSIRIS-Rex Asteroid Sample Return Mission View projectWilliam V. BoyntonThe University of Arizona811 PUBLICATIONS 23,713 CITATIONS SEE PROFILEJohn M. KellerUniversity of Colorado68 PUBLICATIONS 1,181 CITATIONS SEE PROFILERobert C ReedyPlanetary Science Institute522 PUBLICATIONS 10,576 CITATIONS SEE PROFILEAll content following this page was uploaded by Kyeong J Kim on 17 December 2013.The user has requested enhancement of the downloaded file.Variations in K/Th on MarsG. Jeffrey Taylor,1 J. D. Stopar,1 W. V. Boynton,2 S. Karunatillake,3 J. M. Keller,2J. Bru¨ckner,4 H. Wa¨nke,4 G. Dreibus,4 K. E. Kerry,2 R. C. Reedy,5 L. G. Evans,6R. D. Starr,7 L. M. V. Martel,1 S. W. Squyres,3 O. Gasnault,8 S. Maurice,8 C. d’Uston,8P. Englert,1 J. M. Dohm,2,9 V. R. Baker,2,9 D. Hamara,2 D. Janes,2 A. L. Sprague,2K. J. Kim,2 D. M. Drake,10 S. M. McLennan,11 and B. C. Hahn11Received 9 January 2006; revised 20 May 2006; accepted 6 June 2006; published 13 December 2006.[1]K/Th determined by the Mars Odyssey Gamma Ray Spectrometer varies by a factorof 3 on Mars (3000 to 9000), but over 95% of the surface area has K/Th between 4000and 7000. K/Th is distinctly lower than average in some areas, including west of OlympusMons in the Amazonis Planitia, the region around Memnonia Fossae, Chryse Planitia,southeastern Arabia Terra, Syrtis Major Planum, and northwest of Apollinaris Patera.On the other hand, K/Th is distinctly higher than average in other areas, including thecentral part of Valles Marineris and the surrounding highlands, and in the northern part ofHellas. The generally modest variation in K/Th may be explained by inherent variations inigneous rocks and by variations in the extent of aqueous alteration.Citation:Taylor, G. J., et al. (2006), Variations in K/Th on Mars, J. Geophys. Res., 111, E03S06, doi:10.1029/2006JE002676[printed 112(E3), 2007].1.Introduction[2] One of the prime goals of the scientific exploration ofMars is to understand the abundance, distribution, andhistory of water on the planet. The Martian crust wasconstructed by igneous processes and subsequently modi-fied by both endogenic and exogenic processes, whichinclude magmatism, tectonism, impact cratering, aqueousactivity, and aeolian processes [e.g., Scott and Tanaka,1986; Tanaka, 1986; Greeley and Guest, 1987; Dohm etal., 2001a, 2001b, 2002; Anderson et al., 2001; Nimmo andTanaka, 2005; Solomon et al., 2005]. The products of allthis geologic activity contain the record of the early differ-entiation of Mars, the evolution of magma composition andproduction rates, the sedimentary history of the planet, andthe nature of interactions among the atmosphere, hydro-sphere, and lithosphere. In this paper, we focus on usingK and Th data from the Mars Odyssey Gamma Ray Spec-trometer (GRS) to assess the effect of aqueous processes onelemental distributions. To do this, we must deconvolve theeffects of igneous and aqueous processes. Understandingthe details of aqueous activity is essential to assessing thepotential for life having arisen on Mars and for decipheringthe igneous evolution of the crust.[3] There is strong evidence that liquid water flowed onthe surface of Mars. Morphological features on Mars suchas gullies, valley networks, and outflow channels suggest atleast the past presence of liquid water [e.g., Scott et al.,1995; Carr, 1996; Masson et al., 2001; Baker, 2001].Gullies [Malin and Edgett, 1999, 2000] might form bymelting snow on crater walls [Christensen, 2003], seepinggroundwater [Malin and Edgett, 2000], eruption of ground-water by cold climate processes [Gaidos and Marion, 2003]and geothermal heating [Mellon and Phillips, 2001], or dryavalanches of sediment [Treiman, 2003]. Valley networkshave been suggested as the end result of groundwatersapping [e.g., Tanaka et al., 1998; Malin and Carr, 1999;Gulick, 2001] or rainfall [e.g., Craddock and Howard,2002]. Rainfall has also been implicated in the creation ofdendritic valley networks [Mangold et al., 2004] and fromestimates of the denudation rate in the Martian highlands[Hynek and Phillips, 2001]. Outflow channels resemble thescars of catastrophic floods [e.g., Baker and Milton, 1974;Carr, 1979; Baker et al., 1992]. Some authors have sug-gested that the northern plains contained a Martian ocean[Baker et al., 1991; Parker et al., 1993; Ivanov and Head,2001; Carr and Head, 2003], including other hypothesized,endogenic-derived water bodies of varying sizes [Faire´n etal., 2003], possibly acidic enough to inhibit carbonateformation [Faire´n et al., 2004]. Other liquid aqueousJOURNAL OF GEOPHYSICAL RESEARCH, VOL. 111, E03S06, doi:10.1029/2006JE002676, 2006 [printed 112(E3), 2007]1Hawaii Institute of Geophysics and Planetology, Honolulu, Hawaii,USA.2Lunar and Planetary Laboratory, University of Arizona, Tucson,Arizona, USA.3Center for Radiophysics and Space Research, Cornell University,Ithaca, New York, USA.4Max-Planck-Institu¨t fu¨r Chemie, Mainz, Germany.5Institute of Meteoritics, University of New Mexico, Albuquerque, NewMexico, USA.6Computer Sciences Corporation, Lanham, Maryland, USA.7Department of Physics, Catholic University of America, Washington,DC, USA.8Centre d’Etude Spatiale des Rayonnements, Centre National de laRecherche Scientifique/Universite´ Paul Sabatier, Toulouse, France.9Department of Hydrology and Water Resources, University of Arizona,Tucson, Arizona, USA.10TechSource, Santa Fe, New Mexico, USA.11Department of Geosciences, State University of New York, StonyBrook, New York, USA.Copyright 2006 by the American Geophysical Union.0148-0227/06/2006JE002676E03S061 of 20processes may have been active at various times during thehistory of Mars: hydrothermal systems formed by impacts[e.g., Hagerty and Newsom, 2003; Abramov and Kring,2005] or magmatic heating [Gulick, 1998; Dohm et al.,2001a, 2001b, 2001c], low temperature lakes [Newsom etal., 1996; Cabrol et al., 1998], and water seeps [e.g., Ferriset al., 2002; Tanaka et al., 2003, 2005; Miyamoto et al.,2004] often controlled by basement structures such as faultsand fractures [e.g., Tanaka et al., 1998; Dohm et al., 2001a].[4] In addition, the Martian landscape may have beenmodified by solid-ice activity [Baker, 2001]. This includesrelatively ancient massive ice sheets and glaciers [Kargeland Strom, 1992; Dohm et al., 2001b; Gaidos and Marion,2003; Kargel, 2004; Head et al., 2004], local to regionalglacial activity during the Amazonian Period, as noted forthe Tharsis Montes region [Scott and Zimbelman, 1995;Scott et al., 1998; Neukum et al., 2004; Shean et al., 2005],and very young, Late Amazonian flows, which resemble iceand rock glaciers [Baker, 2001; Mustard et al., 2001;Whalley and Azizi, 2003; Berman et al., 2005]. Evidencefor ice mobility suggests environmental change, whichmight include snow precipitation and accumulation [Baker,2001; Head et al., 2005]. The presence of large amounts ofH2O in polar regions and less but significant amounts inequatorial regions [Boynton et al., 2002; Feldman et al.,2002] measured by the Odyssey GRS are consistent withthe morphological evidence for significant water action onMars.[5] The compositions of soils and rocks at the Viking[Clark et al., 1982; Clark, 1993], Pathfinder [e.g., Bru¨ckneret al., 2003], and Gusev [Squyres et al., 2004a; Gellert etal., 2004; Christensen et al., 2004a] landing sites indicatethat water played an important role in producing soils andaltering rocks. Chemical compositions and mineralogyindicate the presence of Mg-sulfates and other salts. Evi-dence for extensive action by water is overwhelming at theOpportunity landing site in Meridiani Planum [Squyres etal., 2004b, 2004c; Christensen et al., 2004b; Klingelho¨fer etal., 2004; Rieder et al., 2004], where abundant sulfateminerals and jarosite are present, and rock festoonedcross-stratification in outcrops indicate transport dominatedby water.[6] If aqueous processes have been active, at least inter-mittently, then it seems likely that portions of the Martiansurface have experienced aqueous alteration. The effects ofsuch alteration may be recorded in the elemental composi-tion of the Martian surface. We report here concentrationsand distributions of K and Th determined by the OdysseyGamma Ray Spectrometer (GRS). As explained in section 4,K and Th are potentially useful for assessing the extent ofaqueous alteration as they and their host minerals behavedifferently in aqueous solutions, raising the possibility thatwe can assess the extent to which surface materials inter-acted with water. Uranium and sulfur are useful, but we donot yet have a fully developed data reduction scheme forthese elements (W. V. Boynton et al., Concentration of H, Si,Cl, K, Fe, and Th in the low-latitude and midlatitude regionsof Mars, submitted to Journal of Geophysical Research,2006; hereinafter cited as Boynton et al., submitted manu-script, 2006). While uranium has many peaks, all of thestrongest ones have interferences caused by significant peaksfrom other elements. Sulfur suffers from similar problems inpeak fitting. Uranium data will be available at the end of theextended mission. Fe, Cl, and H are also useful, and wemake use of them in the context of trying to explain K/Thvariations, but interpretation of their abundances is notstraightforward. Hydrogen is affected by the presence ofsubsurface ice and an unknown amount of bound water, Cl isconcentrated in surface materials that might be redistributed[Keller et al., 2006], and Fe might vary considerably in theprimary igneous rocks. The K/Th ratio avoids many of theseproblems. We show in this paper that the GRS data areconsistent with modest amounts of aqueous alteration, butnot with pervasive alteration of the entire surface. Weexplore the reasons why a signal for aqueous alterationcould be muted in the global GRS data set, concentratingon K/Th variation in igneous rocks, processes operating inthe weathering zone, submarine alteration, and how durationof weathering events would affect the distribution of K andTh.2.Methods[7] The Odyssey GRS and data reduction methods aredescribed by Boynton et al. [2004; submitted manuscript,2006]. Gamma ray spectra were collected every 20 secondsas the Mars Odyssey spacecraft orbited Mars from 8 June2002 to 3 April 2005. To improve the signal-to-noise ratio(S/N), these spectra have been summed over 5-�-5� grids.Concentrations have been determined for the elements K,Th, H, Si, Fe, and Cl on the basis of gamma rays from theMartian surface collected during this time period (Boynton etal., submitted manuscript, 2006). In this paper, we presentthree types of data. First, we present maps of the concen-trations of K, Th, and K/Th derived from a 5 � 5-degree basemap binned to 0.5 � 0.5� and smoothed with a 10-degree arc-radius mean filter. For comparison, �50% of the signal at thedetector originates within a 4 � 5� arc-radius distance fromnadir (e.g., Boynton et al., submitted manuscript, 2006).The greater spatial scale of the filter reduces statistical noise,but spatial certainty is decreased as well. The maps are usefulfor showing global variations in concentration and in K/Th.Because the diluting effect of large concentrations of ice nearthe poles leads to poor counting statistics, we restrict K andTh maps to the area from 75� south to 75� north latitudes.[8] Second, for quantitative analysis, we rebin thesmoothed data to 5 � 5�. These data products are particu-larly useful for x-y plots. The 5 � 5 degree data have goodcounting statistics for K and Th, hence a reasonableinstrumental uncertainty in the K/Th ratio, and constitutea sufficient number of points to investigate compositionalvariation across the planet. The typical instrumental uncer-tainty, srms, stemming from counting statistics on a per-binbasis is computed as the root-mean square of the individual5 � 5 bin instrumental uncertainties (sm). Stated as percen-tages of corresponding global means, these are 7% for K,10% for Th, and 12% for K/Th. However, the instrumentaluncertainties (sm) of individual bins vary significantlyacross the planet, depending on concentration and elevation.(Topographically lower points yield lower gamma photoncount rates because of attenuation in the atmosphere. Wecorrect for atmospheric effects; nevertheless, peak heightsare smaller, hence the uncertainty is larger.)E03S06TAYLOR ET AL.: VARIATIONS IN K/Th2 of 20E03S06[9] Third, we used the mean-smoothed 5 � 5 degree binsto produce maps of the distribution of three parametersdesigned to track the effects of aqueous processes: Cl, H2O,and K/Th. H2O is stoichiometrically calculated from theconcentration of H. The Cl and H2O parameters track theamount of brines that could have been involved in alterationof the surface materials. (MER data from the Opportunitysite suggest that S is a more consistent indicator of waterchemistry than is Cl, but we do not yet have fully reduceddata for S.) K/Th is affected by aqueous alteration (seesection 4), so it may estimate the extent of alteration. Datafor Cl and H are restricted to regions where H is not rapidlyincreasing toward polar ice-rich regions. This was done byuse of the H-masking technique described by Boynton et al.(submitted manuscript, 2006).3.Results[10] K, Th, and K/Th are not uniform across the Martiansurface (Figures 1 and 2). The K/Th ratio varies by a factorof 3, but about 95% of the surface area has a ratio between4000 and 7000 (Figure 3). There are regions that aredistinctly higher and lower than the mean, and we highlightsome of these here. Table 1 lists K, Th, and K/Th for severalregions representative of the global average, lower thanaverage, and higher than average K/Th. This is not meant tobe a comprehensive list of spatially clustered ratios in thetails of the general distribution, just interesting examplesthat allow discussion of the processes (igneous, aqueous)that shaped the geochemistry of the Martian crust. Theglobal mean is the ratio of the K and Th compositionsestimated by summing all GRS spectra between 75� southand 75� north latitude. The data for each region identified asbeing significantly different from the global mean weredetermined by summing all the 20-second spectra collectedover it; the listed uncertainty is the instrumental uncertaintyof the mean, sm, and reflects the numerical confidence withwhich the mean is known.[11] The distribution of K/Th values in our 5 � 5 degreedata set (Figure 3) is moderately skewed to the left with along tail of high values (skewness 0.81), and more peakedwith lower flank frequencies (excess kurtosis 2.17) relativeto a random normal distribution. While the Shapiro-FranciaW’ test suggests that the global K/Th population may besignificantly different from a random normal distribution,we feel the scatterplot between normal order statistics andsorted K/Th values, with a correlation of 0.98, is sufficientlylinear to assume normality [e.g., Upton and Fingleton,1985]. This enabled us to meaningfully identify spatiallyclustered bins at the tails of the global K/Th distribution.Qualitatively, comparison of the K/Th map with the map of1sm instrumental uncertainties at each location (Figure 2)indicates that most places with notably higher or lowerK/Th are statistically meaningful. For example, the Amazo-nis and Syrtis areas with low K/Th (4000–4500) also havereasonably low uncertainties (<500) and appear to be signif-icantly different from the global mean, at least at the level ofinstrumental uncertainty.[12] We quantified these observations by computing theprobability that a sample of 2160 data (the number of binsin our 5 � 5 data set between and inclusive of 75� south and75� north latitude) from a random normal population couldcontain a value at least as extreme in magnitude as weobserve for a given bin. In other words, K/Th bins that aresignificantly less or greater than the global mean are lesslikely to be randomly sampled and more likely to be in thetails of the population. To assess this, we evaluated a two-tailed Student’s t-probability for each K/Th bin, using theinstrumental uncertainty, sm,i of the bin from Figure 2b;the global arithmetic mean, mg (Figure 2a values); and thestandard deviation, s. The standard deviation, s, and globalmean, mg, were used as estimators of the random normalpopulation standard deviation (s), and population mean (m),respectively. The test parameter for the ith bin, ti, wascomputed asti ¼ mg � miffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffis2m;i þ s2qwhere sm,i ensures that bins with higher instrumentaluncertainty are given less significance for deviating fromthe mean. While spatial autocorrelation and associatedinformation redundancy affect all spatial data, and particu-larly ours due to smoothing, mg and s are minimally affectedby these issues [e.g., Haining, 2003]. Figure 2c highlightsthe bins with K/Th �32% likely to be from the populationas sampled by the global data. In other words, these are theareas where individual bins are different in the samedirection by at least 1s from the global mean. Areas 1slower than the mean are in blue and those 2s lower are inpurple. Areas 1s above the mean are in yellow. Note thatthere are no bins 2s higher than the mean and that there arefew (only 7 bins) 2s lower than the mean. The spatialcontiguity of blue/purple or yellow bins defining areascomparable to the GRS footprint make them unlikely to bemerely the result of random variability within instrumentaluncertainty. All our areas of interest (labeled in Figure 2c)are dominated by spatially contiguous pixels. In general,however, there is a significant coupling of K and Th in themidlatitudes, as strongly reinforced by multivariate analyses[Karunatillake et al., 2006]. Furthermore, srms and s areapproximately equal, potentially making instrumental un-certainty, not actual variability in the regolith, the dominantcomponent of the observed variability about the globalmean (i.e., standard deviation, s).[13] In spite of such concerns, we identified two regionswith distinctly higher than average K/Th. These are listed inTable 1 with letters corresponding to the areas identified inFigure 2c. (A) Part of the Hellas basin and its northernejecta, one of the largest impact structures on Mars. Thefloor of Hellas is almost certainly the site of extensivesedimentation and fluvial activity [e.g., Moore and Wilhelms,2001], but the K/Th ratio is not uniform on the floor.(B) Areas north and south of Valles Marineris. We identifiedsix regions with distinctly lower than average K/Th: (C) TheMemnonia region in the ancient highlands southwest ofTharsis. (D) The Amazonia Planitia region is located westof the Tharsis Montes along the highland-lowland boundary[Scott and Tanaka, 1986]. Portions of the Medusae Fossaeregion also have low K/Th. Medusae Fossae materials mayconsist of ash-flow tuffs and ignimbrite deposits on thebasis of their morphology [e.g., Scott and Chapman, 1991;Scott and Tanaka, 1982], possibly subjected to alteration byE03S06TAYLOR ET AL.: VARIATIONS IN K/Th3 of 20E03S06Figure 1.(a) Topographic map of Mars (from MOLA data) with significant locations identified andmaps of the distribution of (b) K and (c) Th on Mars as measured by the Mars Odyssey Gamma RaySpectrometer. Data have been smoothed using a 10-degree boxcar filter. The data are displayed over ashaded relief map of Mars, with mission landing sites indicated (V1 and V2, Viking 1 and 2; PF,Pathfinder; M, Opportunity in Meridiani Planum; G, Spirit in Gusev Crater).E03S06TAYLOR ET AL.: VARIATIONS IN K/Th4 of 20E03S06Figure 2E03S06TAYLOR ET AL.: VARIATIONS IN K/Th5 of 20E03S06fluids [Dohm et al., 2005]. (E) There is a distinct area of lowK/Th in Chryse Planitia, a region associated with SurfaceType 2. However, Karunatillake et al. [2006] show that lowK/Th is not characteristic of Surface Type 2. In general,K/Th 1s clusters are spatially unrelated to Surface Types 1and 2 as defined on the basis of thermal emission spectralproperties [Bandfield et al., 2000]; see Karunatillake et al.[2006] for a detailed study of the chemical composition ofSurface Types 1 and 2. (F) Syrtis Major and a region to itsnorthwest and southwest. Syrtis Major is thought to be abroad, low shield volcano [Hiesinger and Head, 2004].(G) The southeastern part of Arabia Terra, which Dohm et al.[2004a] propose is an ancient impact basin (now uplifted) andpossible site of extensive sedimentary deposition. The sedi-ments would have been derived from the older highlands, andmight show the effects of aqueous alteration. (H) The areasurrounding and north-northwest of Apollinaris Patera. Wediscuss these areas with high and low K/Th in more detail insection 5.4.Processes That Cause Variations in K/Th[14] The variation in K, Th, and K/Th may reflect acombination of igneous processes and aqueous alteration.In this section we use compositions of terrestrial, lunar, andMartian rocks to assess the extent to which igneous andaqueous processes fractionate K from Th. We also discusshow time and dissolution kinetics might affect K/Th. Ourgoal is to gauge the extent to which aqueous processesaffected the global variation in the concentrations of K andTh, and the variation in the K/Th ratio. In section 5, we usethese observations to assess the possible causes of thedistinctly higher-than-average and lower-than-averageK/Th in selected regions (Table 1).4.1.Igneous Processes[15] K and Th are highly incompatible lithophile ele-ments. Crystal-melt distribution coefficients for olivine,orthopyroxene, clinopyroxene, and garnet are all substan-tially less than 1 [Beattie, 1993; Borg and Draper, 2003;Hauri et al., 1994]. Thus they should correlate strongly inigneous rocks, and in general they do. However, K/Th is farfrom constant. Furthermore, the distribution coefficients forK and Th are vastly different in phosphate minerals [Jones,1995], Ca-perovskite [Kato et al., 1988], and phlogopiteand amphibole [Halliday et al., 1995], so involvement ofthose phases could lead to fractionation of K from Th. Thereis enough variability in the K/Th ratio in igneous rocks thatthe entire range we observe on Mars could be ascribed toigneous fractionation. However, GRS data and individualrock suites are not equivalent. GRS data represent averagesover very large areas (nominally about 500 km in diameter),whereas chemical analyses of individual rocks represent aspatial resolution of perhaps 10 cm. Thus we must usecaution when comparing data sets acquired on differentscales.Figure 3.Distribution of the K/Th ratio on Mars, between 75 south and 75 north latitude. The labelsrefer to the value at the center of each bin, which are 500 K/Th values wide. Over 95% of thesmoothed 5 � 5 degree resolution elements lie between 4000 and 7000.Figure 2.Maps of the variation of K/Th on Mars (Figure 2a) and of the one-sigma variation in the ratio (Figure 2b); theuncertainty is basically the measurement uncertainty at each point and does not reflect the standard deviation of the samplepopulation. Data have been smoothed using a 10-degree boxcar filter. The data are displayed over a shaded relief map ofMars, with mission landing sites indicated (V1 and V2, Viking 1 and 2; PF, Pathfinder; M, Opportunity in MeridianiPlanum; G, Spirit in Gusev Crater). Figure 2c is based on 5 � 5 degree smoothed data. It shows the locations of areasdiscussed in the text. Grid points shaded in blue are one standard deviation of the sample population below the global meanof all 5 � 5 degree grid points; purple are two standard deviations below the global mean. Yellow indicates points that areone standard deviation above the global mean.E03S06TAYLOR ET AL.: VARIATIONS IN K/Th6 of 20E03S064.1.1.SNC Meteorites[16] SNC meteorites, all of which are igneous rocks withonly modest amounts of aqueous alteration, vary in theirK/Th ratio. In Figure 4b, we plot the K/Th ratio in SNCmeteorites and in our 5 � 5 degree GRS data set, bothnormalized to our inferred K/Th in bulk Mars (5300[Taylor et al., 2006]). Normalization emphasizes the varia-tion rather than the absolute value of the K/Th ratio, whichvaries from planet to planet. The SNC data represent averagesof numerous analyses or use analyses of large sample masses(data sources are listed in the caption to Figure 4). Thisapproach minimizes sampling errors, which can be consid-erable because K and Th are located in different minerals inSNC meteorites. We also eliminated any meteorite with lowTh/U, which would indicate addition of U and perhaps K byterrestrial groundwater. The SNC meteorite data set has K/Thranging from lower than the GRS data set (basaltic shergot-tites) to roughly the mid-range of the GRS variation. This isnot caused by a systematic error in our measurements of K orTh. They are radioactive elements, so emit gamma rays ofwell-determined energy that is not affected by neutronscattering or capture, H concentration, or other factors, andthe peaks do not suffer from significant interferences fromother gamma ray peaks.[17] It appears that SNC meteorites are not representativeof the Martian surface [Hamilton et al., 2003; Taylor et al.,2006]. The basaltic SNC meteorites have K/Th much lowerthan our inferred bulk composition while the rest have closeto the average composition. This suggests that there couldbe other sources in the Martian mantle with greater thanaverage K/Th, which would have produced magmas (not yetsampled among the meteorite collection) with higher thanaverage K/Th (see McLennan [2003] for a discussion ofpossible heterogeneous mantle sources in Mars). The SNCdata are ambiguous. Taken at face value, they suggest thatsome of the K/Th variation on Mars, especially those withhigh K/Th, could be caused by non-igneous processes, but itis possible that we have not sampled the full range ofigneous rocks on Mars.4.1.2.Terrestrial Igneous Rocks[18] Terrestrial rocks also vary in K/Th. To illustrate, weplot three suites of igneous rocks in Figure 5, compared tothe field occupied by our GRS data (Figure 4b). Theterrestrial data have been normalized to K/Th in the bulksilicate Earth, 2900 [Jagoutz et al., 1979; Taylor andMcLennan, 1985; McDonough and Sun, 1995], and theMars GRS data have been normalized to K/Th in bulksilicate Mars (5300) as in Figure 4b. The terrestrial suite ofaverage igneous rocks of different ages compiled by Condie[1993] has a narrow range of normalized K/Th, from 0.5 to1.4, excluding two high points. The exceptional points arefelsic igneous rocks whose origins might have involvedK-rich fluids (as discussed below). This range is smaller thanthe range shown by our GRS data for Mars (0.6 to 1.7). Onthe other hand, lavas from the Hawaiian reference suite[Basaltic Volcanism Study Project, 1981] have a range inK/Th as large as the variation in the GRS data set. Thissuggests that the entire range we observe on Mars could becaused entirely by complex igneous processes, not aqueousalteration.[19] Terrestrial magmas produced at island arc convergentmargins vary widely in K/Th, with normalized values from0.5 to 5.2 (Figure 5). In subduction zones on Earth, magmasform by hydrous partial melting of the mantle wedge abovethe subducting oceanic slab [e.g., Hawkesworth et al., 1997;Grove et al., 2002]. The trace element characteristics appearto come dominantly from the slab [e.g., Grove et al., 2002].Most important, the slab-derived fluid fractionates the traceelements, enriching K in the fluid. The result is that manyarc magmas are enriched in K. Th appears not to be verysoluble in the fluid phase, so K/Th is high [Hawkesworth etal., 1997], as shown in Figure 5. If a similar processoperated on Mars, whether in a plate tectonic environmentor some other setting in which hydrous fluids couldfractionate K from Th in the mantle, K/Th might besignificantly high. As above, this argues that igneousprocesses might have caused much of the variation inK/Th we observe. An important caution, of course, is thatwe are comparing 500-km GRS data with suites of handspecimens.4.1.3.Lunar Samples[20] The anhydrous Moon offers an interesting exampleof a purely igneous system. All variation in K/Th must beby igneous processes. Those processes might be complicatedTable 1. K, Th, and K/Th in Selected Regions on MarsRegionK,a ppmTh,a ppmK/ThbTypical average regionscGlobal3300 ± 100.62 ± 0.025330 ± 220Regions with high K/ThHellas basin (A)3304 ± 1310.37 ± 0.078830 ± 1830Surrounding Valles Marineris (B)3260 ± 600.49 ± 0.046710 ± 580Regions with low K/ThMemnonia Fossae region (C)3480 ± 700.84 ± 0.054230 ± 230Amazonis Planitia (D)2990 ± 1000.75 ± 0.064010 ± 360Chryse Planitia (E)3750 ± 150.95 ± 0.093950 ± 390Syrtis Major+eastern Arabia (F)2890 ± 500.68 ± 0.044240 ± 260Southeastern Arabia Terra (G)3260 ± 1100.87 ± 0.073770 ± 320NW of Apollinaris Patera (H)2530 ± 1300.78 ± 0.083390 ± 370aUncertainty represents 1-sigma counting statistics (sm) for global summed spectrum (total of 2.4 x 107 seconds countingtime for global average) [Taylor et al., 2006], and the summed spectra for each region; this is the instrument uncertainty of themean.bUncertainty calculated from (K/Th)[(sK/K)2 + (sTh/Th)2]1/2, where K and Th are mean concentration of K and Th(columns 2 and 3). All uncertainties represent the confidence with which the mean is known and do not reflect variations in K,Th, or K/Th.cLetters refer to locations in Figure 2.E03S06TAYLOR ET AL.: VARIATIONS IN K/Th7 of 20E03S06because of preservation of the effects of magma oceancrystallization, subsequent mantle overturn, partial meltingof assorted mantle reservoirs, assimilation, and fractionalcrystallization [e.g., Shearer and Papike, 1999], but becausethe Moon lacks endogenous H2O, the variations in K/Thamong igneous rocks cannot be caused by anything exceptigneous processes. As Figure 6 shows, K/Th varies widelyamong lunar igneous rocks. Some of the variation might becaused by non-representative samples being analyzed. This isespecially possible among highland rocks. However, themare basalts plotted are mostly large samples and they showconsiderable range in K/Th. In fact, there is more variation inK/Th among lunar rocks than among arc magmas on Earth.Lunar GRS data [Prettyman et al., 2002] summed to 500-kmspatial resolution (Figure 6) vary widely. However, for K >1000 ppm, where analytical uncertainty is much lower, thevariation in K/Th is small. The Mars GRS data set has asomewhat larger range in K/Th than does the lunar GRS dataset for K > 1000 ppm, suggesting either more complexigneous processes on Mars or aqueous processes affectedthe distribution of K and Th. However, the large range amongindividual lunar mare basalts shows that K/Th can vary bypurely igneous processes.4.1.4.Summary[21] We do not know the extent to which igneous pro-cesses fractionated K from Th on Mars. The examples givenabove suggest that the entire range in K/Th we observecould be caused by magmatic processes, but until we knowmore about how those processes operated on Mars, weassume that some of the variation could have been causedby aqueous alteration of igneous rocks. We explore thatbelow, but emphasize the possibility that none of thevariation we observe was caused by aqueous alteration.Figure 4.(a) K/Th ratio versus K concentration and (b) K/Th ratio normalized to K/Th in bulk Mars(5300) versus K concentration. The range in K/Th is relatively small, about a factor of 3. Note thatalthough the SNC meteorites are shifted downward from the GRS data set, the total range is similar(about a factor of 3 from lowest to highest). SNC data are from Lodders [1998] and Meyer [2006] andreferences therein.E03S06TAYLOR ET AL.: VARIATIONS IN K/Th8 of 20E03S064.2.Submarine Alteration[22] The extensive evidence for water-driven surfaceprocesses on Mars suggests that seawater-rock interactionsmay have happened. On Earth, alteration of Mid-OceanRidge Basalts (MORB) leads to a large increase in K and inK/Th [e.g., Honnorez, 1981; Jochum and Verma, 1996;Staudigel et al., 1996]. This occurs because ocean water isrich in K (about 400mg/kg [Drever, 1997]) and severelyFigure 6.K/Th in lunar samples and lunar GRS data (binned to 10-degree cells, or 500-km spatialresolution) normalized to bulk Moon value (360) compared to K/Th on Mars normalized to bulk Marsvalue. Lunar sample data are from Papike et al. [1998]; lunar GRS data are from Prettyman et al. [2002].Figure 5.K/Th in suites of terrestrial rocks normalized to bulk Earth value (2900), compared to K/Thfor Mars, normalized to bulk Mars value. Terrestrial data sources: Hawaiian reference suite, BasalticVolcanism Study Project [1981]; average igneous suites, Condie [1993]; arc magmas, Grove et al. [2002],Elliott et al. [1997], and Turner et al. [1996].E03S06TAYLOR ET AL.: VARIATIONS IN K/Th9 of 20E03S06depleted in Th (<0.5 ng/kg), leading to very high K/Th(>8 � 108) in seawater. The large increase in K/Th in oceanwater is not caused simply by dissolution of high-Kcontinental rocks because continental rocks are alsoenriched in Th. Instead, it is due to the higher solubilityof K compared to Th in approximately neutral water.Dissolution of typical Martian crust would still lead to highK/Th in Martian surface water. Staudigel et al. [1996] showthat K/Th (normalized to fresh MORB) in a group of alteredbasalts increases as K increases (Figure 7). In this example,K/Th in altered basalt is 3 to 90 times greater than in typicalunaltered MORB. Thus, if conditions of seawater-rockinteraction were the same on Mars as in modern terrestrialoceans, we would expect that Martian seawater alterationwould lead to a large increase in K/Th. We do not observesuch an enrichment in the northern plains, site of theputative northern ocean.[23] Is terrestrial seawater an appropriate analog? Wehave no direct measurements of the compositions of Mar-tian brines, but experiments suggest that Martian seawatercould be enriched in soluble elements compared to terres-trial oceans. Sawyer et al. [2000] performed a series ofexperiments on water interaction with the Nakhla meteorite.They showed that the concentrations of Na, Mg, Cl, andSO4 in solutions co-existing with Nakhla were very similarto terrestrial ocean water. The concentration of K was lowerthan in seawater, but this might merely reflect the low Kcontent of Nakhla. Bullock et al. [2004] did similar experi-ments using a mineral mixture derived from the mineralogyof SNC meteorites. Their results show a clear enrichment inthe same soluble ions as in the experiments done by Sawyeret al. [2000], including K. In fact, K and the other ions (Na+,Mg2+, Ca2+, SO42�, Al3+, Si4+) were enriched substantiallyover those in terrestrial seawater. We have no informationabout the behavior of Th in the experiments, but it doesseem clear that K is likely to be enriched in surface rocksthat react with Martian brines. This would likely have led toenrichment in K/Th in the altered materials. One caveat isthat under low pH conditions, which are likely on Mars[e.g., Burns, 1993; Faire´n et al., 2004; Hurowitz et al.,2006] because of the presence of abundant chloride andsulfate species, the solubility of Th increases (see below).For comparison, terrestrial seawater is buffered by thecarbonate system [Krauskopf and Bird, 1995] and has apH of about 8.4.3.Weathering4.3.1.Weathering Under Neutral Conditions[24] The mobility of elements during chemical weather-ing depends on the compositions of the igneous rockscomposing the crust, dissolution rates of their host phases,temperature, water chemistry, pH, and rock/water ratio(governed by water supply and rock permeability). Ingeneral, K is more mobile than Th [e.g., Nesbitt and Wilson,1992; Daux et al., 1994; Nesbitt and Markovics, 1997;Patino et al., 2003; Amundson, 2004], as shown in Figure 8.Terrestrial studies show that some elements tend to beimmobile (Al, Ti, Fe, Zr, Th), while others tend to be muchmore mobile during weathering (Mg, Ca, K, REE). Althoughthese generalities do not always hold, they are useful enoughto develop indexes of alteration. The choice of the bestweathering index depends on the nature of the parent rock[Price and Velbel, 2003]. In Figure 8, we compare K/Th to theChemical Index of Alteration (CIA = 100[Al2O3/(Al2O3 +CaO + Na2O + K2O)]), where oxides are in wt%. Highervalues of CIA correlate with more extensive weathering.[25] Within each of the three weathering profiles shownin Figure 8, as weathering progresses (CIA increases), Ktends to decrease in the residual host rock and Th tends tobe unchanged or even increase, resulting in generallydecreasing K/Th (Figure 8). The profiles do not show thesame uniform decrease in K/Th with increasing CIA. Datafor the basalt studied by Nesbitt and Wilson [1992] show aclear, monotonic decrease in K/Th with increasing CIA. Incontrast, the basalt studied by Daux et al. [1994] showsthat K/Th remains relatively unchanged until a CIA value ofFigure 7.K/Th in altered seafloor basalts normalized to average MORB. Data are from Staudigel et al.[1996]. Alteration increases as K concentration increases. K/Th ratio is affected strongly by alterationbecause terrestrial ocean water has very high K/Th. A Martian ocean might not have had such high K/Thif Martian waters were highly acidic.E03S06TAYLOR ET AL.: VARIATIONS IN K/Th10 of 20E03S06�80 then decreases with increasing CIA. Data for a weath-ering profile developed from granodiorite [Nesbitt andMarkovics, 1997] are clustered in two groups, relativelyfresh rock (CIA < 65) with high K/Th, and altered rockwith low K/Th (CIA >75).[26] The decrease in K/Th in the host material withincreasing amount of weathering is a consequence of thehigh solubility of K and low solubility of ThO2. However,the details depend on whether secondary phases precipitate,such as kaolinite to trap K or iron oxides that adsorb Th, andwhether the local regolith is open or closed. In some cases,K and Th are redistributed in the column of weathered rock.Furthermore, even immobile elements such as Th can belost during aqueous alteration. Thorium losses up to 40%can occur due to the formation of colloids in which Th canbe trapped and transported elsewhere [Daux et al., 1994].Patino et al. [2003] argue that no element ought to beconsidered immobile. Nevertheless, weathering under typ-ical terrestrial conditions usually fractionates K from Th.Thus transport of weathering fluids from an area mightleave behind a surface enriched in Th because K is removedby dissolution in the fluid, or subsequent mechanicaltransport of those K-depleted materials could producedeposits with low K/Th. This might be why regions suchas Arabia have lower than average K/Th (Table 1). Incontrast, areas that have been affected by flooding withwater that had reacted with rock would tend to have higherK/Th.4.3.2.Weathering Under Acidic Conditions[27] Terrestrial weathering conditions, however, are ap-proximately neutral to slightly alkaline (pH between 5 and 8)because slightly acidic rainwater (buffered by atmosphericCO2) is made more basic by reactions with minerals,whereas on Mars alteration solutions may be more acidic[e.g., Burns, 1993; Faire´n et al., 2004; Hurowitz et al.,2006; Tosca et al., 2005]. Under those conditions thesolubility of Th increases [e.g., Fanghaenel and Neck,2002], potentially leading to less fractionation of K fromTh. The significant increase in Th solubility with decreas-ing pH is shown in Figure 9, where we show Th solubilityversus pH for two cases: Th in pure H2O and Th in watercontaining 0.008 ppm HPO4 (the concentration in accordwith experiments by Bullock et al. [2004]). In both cases,Th solubility increases orders of magnitude as pHdecreases. In reality, the situation is much more compli-cated because Th does not typically precipitate in a purephase such as thorianite. Instead, it usually dissolves as atrace element in other minerals. In addition, its speciationin natural waters depends on the details of the watercomposition [Langmuir and Herman, 1980], factors notknown well for Mars. Nevertheless, it is clear that at lowpH, Th will behave differently than at neutral conditionslike those prevailing during terrestrial weathering.[28] The Berkely Pit lake in Butte, Montana [Gammons etal., 2003], provides an instructive terrestrial analog. Thelake formed in an open pit that began flooding after miningoperations (for copper and other metals) stopped in 1983. Ithas a pH of 2.3–2.6. The mineral deposits were hosted in ahighly altered silicic rock known as the Butte QuartzMonzonite. We can get a rough idea of how pH affectsK/Th by comparing K/Th in the acidic lake to the unalteredquartz monzonite. The fresh host rock has K/Th of 2270,but K/Th in the lake is only 70. It is possible that K wasnot dissolved completely from the altered and unaltered hostquartz monzonite (see section 4.5 for a discussion ofdissolution kinetics) or had been removed preferentiallyduring hydrothermal alteration that created the mineraldeposit, hence does not reflect processes in the acid lake.It is equally likely, however, that it has been removed fromsolution by precipitation into K-bearing phases such asjarosite, which is supersaturated in the lake [Pellicori etal., 2005]. Gammons et al. [2003] did a series of experi-ments to determine the partitioning of K, Th (and otherelements) between solutions and suspended solids. Thoriumconcentrates in the precipitates to a greater extent than doesK. The KD (concentration in suspended solids/concentrationFigure 8.K/Th in sequences of progressive weathering. CIA is the ‘‘chemical index of weathering’’ andis defined by CIA = 100[Al2O3/(Al2O3 + CaO + Na2O + K2O)]), where oxides are in wt%. Althoughspecifics are different for these cases, weathering on Earth clearly leads to an increase in K/Th of theresidual soil as K is carried off in solution.E03S06TAYLOR ET AL.: VARIATIONS IN K/Th11 of 20E03S06in solution) for Th averaged 13 in the experiments versus aKD of <8 for potassium. The bottom line is that highly acidicconditions can fractionate K from Th.[29] Unpublished dissolution experiments by G. Dreibus[see also Dreibus et al., 1996] at relatively low pH are alsoinformative. The experiments used crushed samples of theMartian meteorite Zagami, a shergottite. In Zagami andother shergottites, Th is concentrated in phosphate minerals,while K is contained in mesostasis glass and maskelynite(shocked plagioclase). Zagami contains 18–19 vol% mas-kelynite and 2–3 vol% phosphate minerals [McCoy et al.,1992]. Because phosphate minerals dissolve rapidly (seebelow), Th is released into solution. At moderate pH itprecipitates into other phases, resulting in a low concentra-tion in solution, as found for the typical terrestrial weath-ering profile (Figure 8). At low pH, however, Dreibus’sexperiments show that Th enters solution: after 14 hours ofreaction time, the solutions in contact with crushed samplesof the Zagami shergottite have K/Th of 181 (versus 3160 infresh Zagami). The altered Zagami residue had K/Th of10,000. We do not know if the mineralogical sites for K andTh in the Martian crust are like those in SNC meteorites, butDreibus et al. [1996] point out Mars appears to be enrichedin moderately volatile elements, leading to relatively high(few percent) modal abundances of phosphate minerals andincreases the likelihood that the bulk of the Th resides inphosphate minerals. The high K/Th ratio on Mars comparedto Earth [Taylor et al., 2006] demonstrates the enrichment ofmoderately volatile elements. Thus these experimental dataare consistent with data from the Berkely Pit lake andindicate that K/Th can be affected by aqueous alteration atlow pH more effectively than at more neutral conditions.4.4.Hydrothermal Alteration[30] Newsom [1980], Newsom et al. [1999], and Abramovand Kring [2005] have drawn attention to the potentialimportance of hydrothermal activity associated with eithermeteorite impact or volcanism on Mars. Newsom et al.[1999] showed that the composition of Martian soils couldbe produced from a mixture of the products of two end-member types of hydrothermal fluids, the neutral-chloridetype and the acid-sulfate type [Nicholson, 1993]. The fluidsare very different from typical ocean water or weatheringfluids on Earth. There have been few studies of the behaviorof K and Th during alteration by hydrothermal fluids. Adetailed study [Papike, 1992] of fumeroles and associateddeposits during a large rhyolitic eruption at Katmai, Alaska,showed that many elements are enriched in the deposits,including Cl, K, and Th, but that K/Th remained constant towithin about 30%. It is not clear how hydrothermal alter-ation has affected K/Th on Mars.4.5.Kinetics of Aqueous Alteration[31] K/Th varies substantially during weathering(Figure 8), but the observed total K/Th range is relativelynarrow: 95% of the 5 � 5 degree points on the Martian surfacefall in the range 4000 to 7000 (Figure 3), a range of less than afactor of two. Elemental alteration pathways depend on therates of mineral dissolution. Could this be the cause of therelatively small variation in K/Th? Rates of dissolutiongenerally depend on crystal structure, bond strengths, andchemical activity of solution [Eggleton, 1986]. Dissolutionoccurs at mineral surfaces (including cracks and defects) incontact with aqueous solutions where chemical reactions cantake place. Dissolution rates can be limited either by theavailability of mineral surface area in contact with solution orFigure 9.Th solubility (activity, a, of Th4+) diagrams as a function of pH. Left diagram contains noPO42�; solution in diagram on the right contains 0.008 ppm HPO4 (a reasonable level based onexperiments by Bullock et al. [2004]). Solubilities were calculated using Geochemist’s Workbench1. Thsolubility is enhanced at low pH, especially in the presence of dissolved phosphate. Phosphate may bepresent in Martian aqueous solutions [Bullock et al., 2004]. The calculations are only approximate as This more likely to be dissolved as a trace element in minerals rather than forming thorianite (ThO2), butthey show the tendency for Th to be much more soluble at low pH.E03S06TAYLOR ET AL.: VARIATIONS IN K/Th12 of 20E03S06by rate of fluid transport to/from the mineral-water interface[e.g., Sak et al., 2004]. Channels and pore spaces will limitfluid flow and control dissolution rates. Mineral dissolutionrates are determined from laboratory or field experimentaldata. The derived empirical rate equations try to account for asmany environmental variables as possible (e.g., pH, temper-ature, aqueous species in solution, etc.).[32] Potassium resides in feldspars and residual glass (ifpresent). Using modal data and mineral compositions for theShergotty basaltic meteorite [Stolper and McSween, 1979],we estimate that K is roughly equally divided betweenplagioclase and silica-rich mesostasis. Th would beexpected to reside in late-stage minerals such as phosphatesand in residual glass if phosphate minerals have not crys-tallized. This is consistent with observations of trace ele-ments in Martian meteorites [e.g., Harvey et al. 1993],assuming Th partitions like the rare earth elements do.Thus, during weathering on Mars the concentrations of Kand Th in aqueous fluids will not be constant. K/Th willdepend on the rates of weathering of plagioclase, residual(silica-rich) glass, and phosphates, and the behavior of Kand Th as their host phases dissolve. The glassy surfaces ofbasaltic lava flows will contain most of the K and Th inbasaltic glass. Experimental data (Figure 10) show thatbasaltic glass and apatite dissolve 3–4 orders of magnitudefaster than do feldspars. Silica-rich glasses have fasterdissolution rates than basaltic glass does, but it is muchslower than plagioclase. The dissolution rates of each phaseincreases as pH decreases. Also, in all cases, these mineralsdissolve faster at higher temperatures than at lower ones(Figure 10 shows data for 25�C). White and Brantley [2003]estimate that most mineral dissolution rates based onlaboratory experiments are 102 to 104 times faster than inthe field, but this would not affect the relative rates ofdissolution shown in Figure 10.[33] If weathering took place under roughly neutral pHconditions (where K is mobile but Th is immobile), and arock surface contained little residual glass, significantfractionation of K from Th would not take place untilplagioclase began to dissolve. Th would be released fromrapidly dissolving phosphate minerals, but would immedi-ately precipitate. After enough time has passed to begindissolving feldspars, the host material will lose K whileretaining Th (since Th continues to precipitate from solu-tion). So, in the long-term K/Th will decrease in the host.On the other hand, if a substantial amount of the K was inresidual glass (Shergotty data suggests that half of it couldbe [Stolper and McSween, 1979]), then the loss of K wouldbe more gradual, but would still lead to a decrease in K/Thin the residual materials.[34] If weathering occurred under acidic conditions,where both K and Th are mobile, Th could be lost from arock unit while K remained because phosphate dissolvesmuch faster than plagioclase and somewhat faster thansilica-rich residual glass. In that case, the weathering fluidwould have a low K/Th ratio while the residual rock wouldhave higher K/Th. Some of the observed variation in K/Thon Mars might be caused by fractionation at low pH but fora time short enough that K remained immobile while Thwas dissolved.4.6.Summary[35] The above considerations lead to the followingguidelines to use in interpreting variation in K/Th of theMartian surface.[36] 1. Geological and paleohydrological activities have adirect bearing on the elemental signatures.[37] 2. Weathering on Earth under neutral pH clearlyfractionates K from Th, with residual materials becomingenriched in Th and having low K/Th. The fluids becomeFigure 10.Dissolution rates at 25�C for phases that contain K and Th in Martian meteorites, andexpected to be the main host for K and Th in other rocks on Mars. All rates increase as pH decreases.Except for basaltic glass, Th-bearing phosphate minerals dissolve much faster than do K-bearingplagioclase. Rate data sources: Albite, Chen and Brantley [1997]; anorthite, Blum and Stillings [1995];plagioclase, Stillings and Brantley [1995]; glass, Gislason and Oelkers [2003]; fluorapatite, Guidry andMackenzie [2003].E03S06TAYLOR ET AL.: VARIATIONS IN K/Th13 of 20E03S06enriched in K, hence have high K/Th. This is why terrestrialocean water has very high K/Th.[38] 3. Weathering under acidic conditions might bluntthe fractionation of K from Th because of increased Thsolubility, but studies of acid lakes suggest that K still tendsto be enriched compared to Th in solution unless a K-richphase precipitates.[39] 4. If Martian ocean water had high K/Th, liketerrestrial ocean water, fresh basalts altered in a Martianmarine environment would have had elevated K/Th.[40] 5. Differential solubility of mineral hosts for K andTh lead to changing K/Th with time. Details depend onwater flux, Th solubility, and pH.[41] 6. Igneous processes can produce the observed rangein K/Th, so this ratio alone cannot be used to assess theextent of aqueous alteration on Mars.5.Discussion[42] On a large scale, for example, the Martian highlandsversus the lowland plains, K/Th does not vary dramatically(Figure 2). Nevertheless, there are areas with distinctiveK/Th. The regions with significantly lower K/Th may pro-vide evidence for the igneous and aqueous processes thatshaped their geochemistry and geology. We discuss areaswithin which K/Th is consistently one standard deviation (1s)above and 1s below the global mean and put our K/Thmeasurements into the context of previous geological studies.5.1.Areas With Higher Than Average K/Th[43] To help assess the role of aqueous processes inproducing the observed K/Th ratios, we created a map thathighlights ratios affected by weathering (Figure 11). Thethree parameters used in Figure 11 are Cl (red), H2O(green), and K/Th (blue). Ideally, weathering should tendto produce high Cl and H2O, and low K/Th, although underacidic conditions K/Th might be higher than average.Furthermore, Cl could be affected by highly mobile regolithcomponents or highly localized variations in weathering,and be enriched in volcanic sublimates; see Keller et al.[2006] and H. E. Newsom et al. (Geochemistry of Martiansoil and bedrock in mantled and less mantled terrains withgamma ray data from Mars Odyssey, submitted to Journalof Geophysical Research, 2006) for detailed discussions ofCl behavior and its distribution on Mars. Thus distinctlygreen (high H2O), yellow (high H2O and Cl), or blue (highK/Th) areas are candidates for being sites of significantaqueous alteration. Dark areas (low in all three parameters)may not have been greatly affected by aqueous processes.[44] Two regions appear to have distinctly high K/Th.One is the northern part of the Hellas basin and rimmaterials. The other region is the large expanse south andnorth of Valles Marineris.5.1.1.Northern Hellas[45] The Hellas basin is one of the largest impact struc-tures on Mars and almost certainly the site of extensivefluvial activity and sedimentation. For example, large chan-nels have drained into the basin [e.g., Greeley and Guest,1987; Crown et al., 1992; Scott et al., 1995]. Extensiveancient glaciation [e.g., Kargel and Strom, 1992] or cata-strophic erosion induced by magmatic intrusion into volatile-rich rocks [Tanaka et al., 2002] may have resulted in therelatively flat topography in the Malea Planum region alongthe southern margin of the large impact structure (e.g.,Malea Planum volcanic materials partly infills a largetopographic low over 1200 km wide [see Head and Pratt,2001]). Thick basin infill materials are interpreted to bemade of water-lain sediment [Malin and Edgett, 2000;Moore and Wilhelms, 2001]. Thus the geomorphologyclearly indicates that Hellas is the site of extensive geolog-ical and paleohydrological activity. Chemical compositionsof basin fill are consistent with this. Almost the entire basinis characterized by relatively high K/Th, though only thenorthwest section and associated highlands are significantlyhigher than the global average at 1s levels (Figure 2c). Thearea with high K/Th (blue areas in Figure 11) has lowconcentrations of Cl and H2O.Figure 11.Map of the distribution of Cl (red), equivalent H2O (green), and K/Th (blue) from median-smoothed 5 � 5 degree data using H-mask. Areas discussed in the text are outlined and identified byletters (see Table 1 and Figure 2). Low values are dark in each color. Mixtures of low K/Th but high Cland H2O appear as yellow-orange; light blue reflects high K/Th and H2O but low Cl; high K/Th and Clbut low H2O areas are magenta; and generally dark areas are low in all three parameters.E03S06TAYLOR ET AL.: VARIATIONS IN K/Th14 of 20E03S06[46] The interpretation of K/Th in the Hellas region ismore complicated than simply deposition of altered sedi-ments into a basin. The K/Th anomaly extends to thenorthern rim of the basin. These deposits are rugged,mountainous basin ejecta [Greeley and Guest, 1987].Although likely modified by later aqueous and possiblyigneous events, the high K/Th on the Hellas rim mightindicate that deeper crustal (and mantle?) rock has elevatedK/Th compared to the average surface materials on Mars.5.1.2.Valles Marineris Region[47] The Valles Marineris region with distinctly elevatedK/Th is composed of Noachian to Hesperian volcanicdeposits [Scott and Tanaka, 1986]. The K/Th high alsoincludes Valles Marineris itself, but the canyon is narrowerthan the average GRS pixel (500 km), so the extent to whichValles Marineris has high K/Th is not known with certainty.There is abundant evidence for fluvial processes in this area.Komatsu et al. [2004a] draw an analogy with the Mongolianplateau, a large tectonomagmatic complex on Earth. Similarto that reported for the central part of Valles Marineris[Dohm et al., 2001b], magmatic-driven uplift, rifting, vol-canism, and aqueous processes have played important rolesin creating the Mongolian plateau. Products of ice-magmainteractions have been suggested to occur in the VallesMarineris area [Chapman and Tanaka, 2001; Komatsu etal., 2004b]. Ice-magma interactions might have led to theelevated K/Th signal. On the other hand, it is possible thatthe lavas making up most of the surface are intrinsicallyhigh in K/Th. If so, then this emphasizes the compositionalheterogeneity of the Martian mantle.5.2.Areas With Lower Than Average K/Th5.2.1.Memnonia Fossae Region[48] The region around Memnonia Fossae (C in Table 1and Figures 2 and 11; a dark yellow area in Figure 11) ischaracterized by moderate concentrations of Cl and H2Oand low K/Th. The region with low K/Th records diversestratigraphy [Scott and Tanaka, 1986], a fault system thatfans out to the southwest from the Tharsis Montes region,Memnonia Fossae [e.g., Scott and Tanaka, 1986; Andersonet al., 2001], and a conspicuous outflow channel system thatdebouches into Amazonis Planitia, Mangala Valles [Scottand Tanaka, 1986; Chapman and Tanaka, 1993; Zimbelmanet al., 1994; Craddock and Greeley, 1994]. The channelsystem includes the NSVs region [Dohm et al., 2001b,2001c, 2004b], west-southwest of the huge northeast-trending chain of Tharsis Montes shield volcanoes [e.g.,Scott and Tanaka, 1986]. Mangala Valles, which originatesfrom faults of Memnonia Fossae in the cratered highlands,dissects ancient highland materials. These materials consistof a me´lange of rocks that are interpreted to mainly includelava flows, impact breccias, and eolian, fluvial, and collu-vial deposits [e.g., Scott and Tanaka, 1986; Tanaka, 1986].The Mangala flood outbursts may have been related toexplosive magmatic activity, resulting from structurallycontrolled, magma-water interactions, which includesphreatomagmatic activity caused by dike emplacement,cryospheric cracking, magma groundwater mixing, andexplosive eruption to the surface [Wilson and Head, 2004].[49] The low K/Th in this region shows that ancientMartian crust varied in K/Th and that mapped geologicunits vary in composition. The region with low K/Th alsocontains a ridged unit, presumably affected by faulting [e.g.,Scott and Tanaka, 1986]. Mangala Valles, which formed asa result of at least two episodes of flooding during theHesperian and Early Amzonian [Chapman and Tanaka,1993; Zimbelman et al., 1994; Craddock and Greeley,1994], lies at the eastern end of the area with low K/Th,so the releases of groundwater is not necessarily the causeof the depressed K/Th. However, Hanna and Phillips[2005] have identified a widespread set of outflow channelsin the region around Memnonia Fossae and Sirenum Fossae(which borders the southern end of the low K/Th region).This outflow complex may account for the low K/Th byreaction between discharged groundwater and surface mate-rials. Alternatively, these erosional events might have con-tinuously stripped the surface of altered rock, revealingancient rock with inherently low K/Th.5.2.2.Amazonis Planitia[50] Amazonis Planitia (D in Table 1 and Figures 2 and 11;orange in Figure 11) is located between the Tharsis andElysium volcanic provinces. The region is relativelysmooth, possibly the result of deposition of fluid sedimentsand thin fluid lava flows [Fuller and Head, 2002; Tanaka etal., 2005]. In addition, Amazonis Planitia is the site of ahypothesized ocean [e.g., Parker et al., 1993] or paleolake[e.g., Scott et al., 1995]. Amazonian volcanic depositsdominate the region with low K/Th in Amazonis Planitia[Scott and Tanaka, 1986; Fuller and Head, 2002; Tanaka etal., 2005]. Thus this region might reflect an inherently lowK/Th in the lava flows. An alternative is that a higher K/Thhas been diluted by addition of fluvial sediments from otherregions, such as the Medusae Fossae formation, which lies ata higher elevation to the south. The Medusae Fossaeformation is located west of the Tharsis Montes along thehighland-lowland boundary [Scott and Tanaka, 1986], ismade up of a discontinuous band of wind-etched materialsthat appear to blanket underlying topography [Greeley andGuest, 1987; Scott and Tanaka, 1986; Sakimoto et al.,1999]. Their morphology suggests that these materialsmay consist of ash-flow tuffs and ignimbrite deposits[e.g., Scott and Chapman, 1991; Scott and Tanaka, 1982],possibly subjected to alteration by fluids [Dohm et al.,2005]. It is distinctly enriched in Cl and H2O, but hasaverage K/Th (Figure 2). Keller et al. [2006] outline fourprimary hypotheses for the enrichment in Cl in this region.The explanations include acid-fog reactions involving emis-sions from local volcanism, hydrothermal alteration, low-temperature alteration, and addition of Cl-rich aeolian mate-rials. Sediments derived from the Medusae Fossae formationought to make K/Th in Amazonis similar to the globalaverage, not decrease it, although this depends on local pHduring alteration. Thus it seems most likely that the lowK/Th reflects the inherent K/Th of the Amazonian lavaflows in Amazonis west of Olympus Mons.5.2.3.Chryse Planitia[51] Chryse Planitia records a complex geologic andpaleohydrologic history, which includes the degradation ofthe putative Chryse impact [e.g., Dohm et al., 2001b], floodinundation [Scott and Tanaka, 1986; Rotto and Tanaka,1995; Tanaka et al., 2005], possibly episodic related toTharsis-induced activity [Dohm et al., 2001b, 2001c; Faire´net al., 2003], spring-fed activity along the highland-lowlandboundary [Tanaka et al., 2005], and emplacement of vola-E03S06TAYLOR ET AL.: VARIATIONS IN K/Th15 of 20E03S06tile-enriched materials (including fluvial deposits and debrisflows) and relatively late-stage lava flows [e.g., Tanaka etal., 2005]. The low K/Th area in Chryse Planitia (E in Table 1and Figures 2 and 11; a dark area in Figure 11) is locatednorth of the Viking 1 and Pathfinder landing sites andincludes easternmost Acidalia Planitia. Although buriedchannel materials are present the two most abundant unitsare interpreted to be volcanic in origin, but possiblymodified by eolian deposits [Scott and Tanaka, 1986].Cl and H2O are moderate to low, suggesting that the lowK/Th reflects the composition of late lava flows rather thanmodification by the extensive paleohydrologic processesobserved in the region.5.2.4.Syrtis Major[52] Syrtis Major (F in Table 1 and Figures 2 and 11; adark area in Figure 11) is thought to be a broad, low shieldvolcano [Greeley and Guest, 1987; Hiesinger and Head,2004]. Data from the Mars Global Surveyor ThermalEmission Spectrometer [Bandfield et al., 2000] indicate thatthe surface is dominated by basaltic minerals, clinopyroxeneand plagioclase. Mustard et al. [2005a, 2005b] confirm this,but also show that although augite is most abundant,orthopyroxene is present. The far northwest flanks of Syrtis,in the direction of Nili Fossae, is richer in orthopyroxeneand contains a significant amount of olivine [Hamilton etal., 2003; Hamilton and Christensen, 2005]. Those regionsalso contain hydrated minerals [Mustard et al., 2005b;Bibring et al., 2005; Poulet et al., 2005]. Nevertheless,most of Syrtis consists of dark, relatively dust-free pyrox-ene-rich rock. The region of low K/Th and low Cl and H2Oextends to the northwest and southeast from the Syrtis shieldvolcano (Figures 2 and 11), suggesting that low K/Thvolcanism and low amounts of aqueous alteration arecharacteristic of this region. Harvey and Hamilton [2005]suggested that Syrtis might be the source for the nakhliteMartian meteorites. The nakhlites and Chassigny have K/Thin the range 4500 to 6000, generally higher than the averagevalue we observe in Syrtis Major, 3900 (Table 1). Ourmeasurement is more in line with those of the shergottites,which range from 3500 to 5500, or a mixture of nakhlite-like and shergottite-like magmas. Nevertheless, it is likelythat Syrtis is a basaltic volcano that has not experiencedsignificant aqueous alteration. The lack of clear evidence foraqueous alteration, except for the presence of hydratedminerals near the edge of Syrtis [Mustard et al., 2005b],suggests that since the last eruptions on the volcano, waterhas not been present long enough for significant alterationto have occurred (Figure 10 and section 4.5).5.2.5.Southeastern Arabia Terra[53] Arabia Terra (G in Table 1 and Figures 2 and 11;green in Figure 11) may be a now uplifted ancient impactbasin [Dohm et al., 2004a]. Extensive sedimentary deposi-tion has occurred in it [Malin and Edgett, 2000], and ancientdrainage features debouch into the basin. The sedimentswould have been derived from the older highlands, andmight show the effects of aqueous alteration before, during,and after their deposition. Arabia has substantially elevatedH2O and average Cl, but lower than average K/Th in itssoutheastern half (Figures 2 and 11 and Table 1). The highH2O and numerous features suggesting aqueous processesare consistent with water having played a significant role inthe evolution of the Arabia region. If so, why is K/Th lowerthan average? As discussed in section 4, weathering underacidic conditions causes loss of both K and Th from aweathering horizon. If there was sufficient time for weath-ering to have affected silicic glass or plagioclase, K/Thought to reflect the global average. Since K/Th is low inArabia, perhaps the sediments deposited in it derived from aregion whose original igneous rocks had low K/Th, likethose in the broad region to the east of region G, northwestand southeast of Syrtis Major (region F). It is also plausiblethat much of the K-Th fractionation resulted from prefer-ential dissolution of Th in the source rocks, if weatheringtook place under acidic conditions, leading to depositionfrom water with low K/Th. Whatever the details, theenhancement of H2O and moderate enrichment in Cl pointsto an important role for water in the evolution of surfacematerials in Arabia.5.2.6.Northwest of Apollinaris Patera[54] This region (H in Table 1 and Figures 2 and 11; ayellow area in Figure 11) is dominated by Apollinaris Pateraand deposits from it. A seemingly isolated shield volcano,Apollinaris Patera, has been reported to have involved theinteraction of magma and water [Scott et al., 1993; Robinsonet al., 1993]. The development of the prominent andlong-lived (episodic at least during the Noachian andHesperian) volcano, which is located between Gusev impactcrater to the south (location of the MER rover Spirit) and thehighland-lowland boundary to the north, occurred in awater-enriched region of Mars [Scott et al., 1993, 1995].Evidence for magma and water interactions include partsof the volcano that have broken away to form chaoticterrain and a network of valleys along the southern marginof the Hesperian lava flow that was emplaced on thesouthern flank of the volcano [Scott et al., 1993]. Inaddition, a prominent scarp that forms the base of thevolcano, similar to the basal scarp for Olympus Mons, wasinterpreted as a wave-cut platform among other possiblehypotheses [Mouginis-Mark et al., 1992; Scott et al.,1993]. In addition, valleys dissect stratigraphically young(Middle Amazonian and younger) Medusae Fossae For-mation materials to the east of Apollinaris Patera [Scott etal., 1993, 1995; Dohm et al., 2004b]. Medusae FossaeFormation materials, which occur along the highland-lowland boundary west of the giant shield volcanoes ofTharsis Montes [Scott and Tanaka, 1986], form a discon-tinuous band of wind-etched materials that appear to blanketunderlying topography [Scott and Tanaka, 1986; Greeley andGuest, 1987; Bradley et al., 2002] and embay northern andeastern parts of the basal scarp of Apollinaris Patera [Scott etal., 1993]. Among several potential origins [e.g., Schultz andLutz, 1988], the materials of the Medusae Fossae Formationhave been interpreted to consist of ash-flow tuffs and ignim-brite deposits on the basis of their morphology [e.g., Malin,1979; Scott and Tanaka, 1982; Scott and Chapman, 1991].The hypothesis of such long-lived magma and water inter-actions in the Apollinaris Patera and surrounding regions(episodically from the Noachian to at least the late Hesperian),which point toward a pronounced hydrothermal environmentin the geologic and paleohydrologic records [e.g., Scott etal., 1993, 1995; Dohm et al., 2004b, 2005], is supportedby GRS data that indicates elevated hydrogen and chlorinewhen compared to the rest of the equatorial region ofMars (Figure 11). On the other hand, the elevated chlorineE03S06TAYLOR ET AL.: VARIATIONS IN K/Th16 of 20E03S06signature could also be attributed to the complex lithologicnature of the region [Scott and Tanaka, 1986; Scott et al.,1993], making it difficult to interpret the low K/Th.6.Summary and Conclusions[55] The most striking observation is that K/Th does notvary much on the surface of Mars. Ninety-five percent ofthe surface area has K/Th between 4000 and 7000, in spiteof the evidence for significant aqueous processes havingoccurred on Mars and geochemical studies showing thatK/Th is affected by rock-water interactions. Exactly howrock-water reactions affect K/Th depends on environmentalfactors, especially on solution pH and on the amount of timeavailable for alteration. The modest variation in K/Th mightindicate that weathering solutions were highly acidic, whichmakes both K and Th mobile. Alternatively, the restrictedrange in K/Th might indicate that roughly neutral conditionsprevailed and the duration of weathering events was too shortto significantly affect K/Th. It is also possible that strongeffects from aqueous alteration are only detectable on a scalesmaller than the GRS footprint. There are spatial clusters inthe K/Th distribution, with regions having significantlysmaller than average K/Th and others having significantlyhigher than average K/Th. The K/Th ratio in these areasseems to have been inherited from primary igneous rocks, butcould have been affected by aqueous alteration.[56] Understanding the relative roles of igneous andaqueous processes requires additional data. Uraniumbehaves differently from either K or Th during aqueousprocessing, yet tends to behave similar to K and Th inanhydrous igneous systems. This will provide us with anadditional parameter for testing whether aqueous alterationoccurred in a given region. During weathering, for example,both U and Th are released by phosphate dissolution, butThO2 tends to precipitate while U remains in solution, ifconditions are oxidizing enough for UO2++ to be the dom-inant U species and pH is moderate. Thus, during the earlystages of weathering before K-bearing phases dissolve, K/Uand Th/U will both increase as U is leached from rapidlydissolving phosphates, while K/Th remains roughly con-stant. We will have U data for large regions by the end ofthe extended Mars Odyssey mission. Full understanding ofthe aqueous environment will also be aided by measurementof S. This element produces weak peaks, but we hope tohave good measurements of S over large regions of theplanet. This will be very useful in conjunction with Cl datato assess the abundance of salts on the surface. A betterunderstanding of Martian crustal history will come fromdetailed study of specific regions on Mars, like those wehighlight briefly above, using multiple imaging and spec-troscopic data sets.[57]Acknowledgments.The authors thank the Mars Odyssey mis-sion for support. We thank Pam Clark and an anonymous reviewer for veryuseful comments. This is SOEST publication 6784 and HIGP publication1440.ReferencesAbramov, O., and D. A. Kring (2005), Impact-induced hydrothermal activ-ity on early Mars, J. Geophys. Res., 110, E12S09, doi:10.1029/2005JE002453.Amundson, R. (2004), Soil formation, in Treatise on Geochemistry, vol. 5,chap. 1, 1–35, Elsevier, New York.Anderson, R. C., J. M. Dohm, M. P. Golombek, A. Haldemann, B. J.Franklin, K. L. Tanaka, J. Lias, and B. Peer (2001), Significant centersof tectonic activity through time for the western hemisphere of Mars,J. Geophys. Res., 106, 20,563–20,585.Baker, V. R. (2001), Water and the Martian landscape, Nature, 412, 228–236.Baker, V. R., and D. J. Milton (1974), Erosion by catastrophic floods onMars and Earth, Icarus, 23, 27–41.Baker, V. R., R. G. Strom, V. C. Gulick, J. S. Kargel, G. Komatsu, and V. S.Kale (1991), Ancient oceans, ice sheets and the hydrological cycle onMars, Nature, 352, 589–594.Baker, V. R., M. S. Marley, M. H. Carr, V. C. Gulick, and C. R. Williams(1992), Channels and valley networks, in Mars, edited by H. H. Kieffer etal., pp. 493–522, Univ. of Ariz. Press, Tucson.Bandfield, J. L., V. E. Hamilton, and P. R. Christensen (2000), A globalview of Martian surface compositions from MGS-TES, Science, 287,1626–1630.Basaltic Volcanism Study Project (1981), Basaltic Volcanism on the Ter-restrial Planets, 1286 pp., Elsevier, New York.Beattie, P. (1993), The generation of uranium series disequilibria by partialmelting of spinel peridotite: Constraints from partitioning studies, Earth.Planet. Sci. Lett., 117, 379–391.Berman, D. C., W. K. Hartmann, D. A. Crown, and V. R. Baker (2005), Therole of arcuate ridges and gullies in the degradation of craters in theNewton Basin region of Mars, Icarus, 178, 465–486.Bibring, J. P., et al. (2005), Mars surface diversity as revealed by theOMEGA/Mars Express observations, Science, 307, 1576–1581.Blum, A. E., and L. L. Stillings (1995), Feldspar dissolution kinetics, inChemical Weathering Rates of Silicate Minerals, Rev. Mineral., vol. 31,edited by A. F. White and S. L. Brantley, pp. 291–351, Mineral. Soc. ofAm., Washington, D. C.Borg, L. E., and D. S. Draper (2003), A petrogenetic model for the originand compositional variation of the Martian basaltic meteorites, Meteorol.Planet. Sci., 38, 1713–1731.Boynton, W. V., et al. (2002), Distribution of hydrogen in the near surfaceof Mars: Evidence for subsurface ice deposits, Science, 297, 81–85.Boynton, W. V., et al. (2004), The Mars Odyssey Gamma-Ray Spectro-meter Instrument Suite, Space Sci. Rev., 110, 37–83.Bradley, B. A., S. E. H. Sakimoto, H. Frey, and J. R. Zimbelman (2002),Medusae Fossae Formation: New perspectives from Mars Global Sur-veyor, J. Geophys. Res., 107(E8), 5058, doi:10.1029/2001JE001537.Bru¨ckner, J., G. Dreibus, R. Rieder, and H. Wa¨nke (2003), Refined data ofAlpha Proton X-ray Spectrometer analyses of soils and rocks at the MarsPathfinder site: Implications for surface chemistry, J. Geophys. Res.,108(E12), 8094, doi:10.1029/2003JE002060.Bullock, M. A., J. M. Moore, and M. T. Mellon (2004), Laboratory simula-tions of Mars aqueous geochemistry, Icarus, 170, 404–423.Burns, R. (1993), Rates and mechanisms of chemical weathering of ferro-magnesian silicate minerals on Mars, Geochim. Cosmochim. Acta, 57,4555–4574.Cabrol, N. A., E. A. Grin, R. Landheim, R. O. Kuzmin, and R. Greeley(1998), Duration of the Ma’adim Vallis/Gusev Crater hydrogeologic sys-tem, Mars, Icarus, 133, 98–108.Carr, M. H. (1979), Formation of Martian flood features by release of waterfrom confined aquifers, J. Geophys. Res., 84, 2995–3007.Carr, M. H. (1996), Water on Mars, Oxford Univ. Press, New York.Carr, M. H., and J. W. Head III (2003), Oceans on Mars: An assessment ofthe observational evidence and possible fate, J. Geophys. Res., 108(E5),5042, doi:10.1029/2002JE001963.Chapman, M. G., and K. L. Tanaka (1993), Geologic map of the MTM-05152 and -10152 quadrangles, Mangala Valles region of Mars, U.S.Geol. Surv. Misc. Invest. Ser., Map I-2294, scale 1:500,000.Chapman, M. G., and K. L. Tanaka (2001), Interior trough deposits onMars: Subice volcanoes?, J. Geophys. Res., 106(E5), 10,087–10,100.Chen, Y., and S. L. Brantley (1997), Temperature- and pH-dependence ofalbite dissolution rate at acid pH, Chem. Geol., 135, 275–290.Christensen, P. R. (2003), Formation of recent Martian gullies throughmelting of extensive water-rich snow deposits, Nature, 422, 45–48.Christensen, P. R., et al. (2004a), Initial results from the Mini-TES experi-ment in Gusev Crater from the Spirit rover, Science, 305, 837–842.Christensen, P. R., et al. (2004b), Mineralogy at Meridani Planum from theMini-TES experiment on the Opportunity rover, Science, 306, 1733–1739.Clark, B. E. (1993), Geochemical components in Martian soil, Geochim.Cosmochim. Acta, 57, 4575–4581.Clark, B. E., A. K. Baird, R. J. Weldon, D. M. Tsusaki, L. Schnabel, andM. P. Candelaria (1982), Chemical composition of Martian fines, J. Geo-phys. Res., 87, 10,059–10,067.E03S06TAYLOR ET AL.: VARIATIONS IN K/Th17 of 20E03S06Condie, K. (1993), Chemical composition and evolution of the upper con-tinental crust: Contrasting results from surface samples and shales, Chem.Geol., 104, 1–37.Craddock, R. A., and R. Greeley (1994), Geologic map of the MTM-20147quadrangle, Mangala Valles region of Mars, U.S. Geol. Surv. Misc. Invest.Ser., Map I-2310, scale 1:500,000.Craddock, R. A., and A. D. Howard (2002), The case for rainfall on awarm, wet early Mars, J. Geophys. Res., 107(E11), 5111, doi:10.1029/2001JE001505.Crown, D. A., K. H. Price, and R. Greeley (1992), Geologic evolution ofthe east rim of the Hellas basin, Mars, Icarus, 100, 1–25.Daux, V., J. L. Crovisier, C. Hemond, and J. C. Petit (1994), Geochemicalevolution of basaltic rocks subjected to weathering: Fate of the majorelements, rare earth elements, and thorium, Geochim. Cosmochim. Acta,58, 4941–4954.Dohm, J. M., K. L. Tanaka, and T. M. Hare (2001a), Geologic map of theThaumasia region of Mars. U.S. Geol. Surv. Misc. Invest. Ser., Map I-2650, scale 1:5,000,000.Dohm, J. M., J. C. Ferris, V. R. Baker, R. C. Anderson, T. M. Hare, R. G.Strom, N. G. Barlow, K. L. Tanaka, J. E. Klemaszewski, and D. H. Scott(2001b), Ancient drainage basin of the Tharsis region, Mars: Potentialsource for outflow channel systems and putative oceans or paleolakes,J. Geophys. Res., 106, 32,943–32,958.Dohm, J. M., et al. (2001c), Latent outflow activity for western Tharsis,Mars: Significant flood record exposed, J. Geophys. Res., 106, 12,301–12,314.Dohm, J. M., S. Maruyama, V. R. Baker, R. C. Anderson, J. C. Ferris, andT. M. Hare (2002), Evolution and traits of Tharsis superplume, Marspaper presented at Superplume International Workshop, Tokyo Inst. ofTechnol., Tokyo.Dohm, J. M., N. G. Barlow, J. P. Williams, V. R. Baker, R. C. Anderson, W. V.Boynton, A. G. Faire´n, and T. M. Hare (2004a), Ancient giant basin/aquifersystem in the Arabia region, Mars, Lunar Planet Sci. [CD-ROM], XXXV,Abstract 1209.Dohm, J. M., J. C. Ferris, N. G. Barlow, V. R. Baker, W. C. Mahaney, R. C.Anderson, and T. M. Hare (2004b), The Northwestern Slope Valleys(NSVs) region, Mars: A prime candidate site for the future explorationof Mars, Planet. Space Sci., 52, 189–198.Dohm, J. M., K. Kerry, J. M. Keller, V. R. Baker, S. Maruyama, R. C.Anderson, J. C. Ferris, and T. M. Hare (2005), Mars geological provincedesignations for the interpretation of GRS data, Lunar Planet. Sci. [CD-ROM], XXXVI, Abstract 1567.Dreibus, G., E. Jagoutz, B. Spettel, and H. Wa¨nke (1996), Phosphate-mo-bilization on Mars? Implication from leach experiments on SNCs, LunarPlanet. Sci., XXVII, 323–324.Drever, J. I. (1997), Geochemistry of Natural Waters: Surface and Ground-water Environments, 3rd ed., vol. xii, 436 pp., Prentice-Hall, UpperSaddle River, N. J.Eggleton, R. A. (1986), The relation between crystal structure and silicateweathering rates, in Rates of Chemical Weathering of Rocks and Miner-als, edited by S. M. Colman and D. P. Dethier, pp. 21–42, Elsevier, NewYork.Elliott, T., T. Plank, A. Zindler, W. White, and B. Bourdon (1997), Elementtransport from slab to volcanic front at the Mariana arc, J. Geophys. Res.,102, 14,991–15,019.Faire´n, A. G., J. M. Dohm, V. R. Baker, M. A. de Pablo, J. Ruiz, J. Ferris,and R. C. Anderson (2003), Episodic flood inundations of the northernplains of Mars, Icarus, 165, 53–67.Faire´n, A. G., D. Ferna´ndez-Remolar, J. M. Dohm, V. R. Baker, and R. Amils(2004), Inhibition of carbonate synthesis in acidic oceans on early Mars,Nature, 431, 423–426.Fanghaenel, T., and V. Neck (2002), Aquatic chemistry and solubility phe-nomena of actinide oxides/hydroxides, Pure Appl. Chem, 74, 1895–1907.Feldman, W. C., et al. (2002), Global distribution of neutrons from Mars:Results from Mars Odyssey, Science, 297, 75–78.Ferris, J. C., J. M. Dohm, V. R. Baker, and T. Maddock III (2002), Darkslope streaks on Mars: Are aqueous processes involved?, Geophys. Res.Lett., 29(10), 1490, doi:10.1029/2002GL014936.Fuller, E. R., and J. W. Head III (2002), Amazonis Planitia: The role ofgeologically recent volcanism and sedimentation in the formation of thesmoothest plains on Mars, J. Geophys. Res., 107(E10), 5081,doi:10.1029/2002JE001842.Gaidos, E., and G. Marion (2003), Geological and geochemical legacy of a coldearly Mars, J. Geophys. Res., 108(E6), 5055, doi:10.1029/2002JE002000.Gammons, C. H., S. A. Wood, J. P. Jonas, and J. P. Madison (2003),Geochemistry of the rare-earth elements and uranium in the acidic Ber-keley Pit lake, Butte, Montana, Chem. Geol., 198, 269–288.Gellert, R., et al. (2004), Chemistry of rocks and soils in Gusev Crater fromthe Alpha Particle X-ray Spectrometer, Science, 305, 829–832.Gislason, S. R., and E. H. Oelkers (2003), Mechanism, rates, and conse-quences of basaltic glass dissolution: II. An experimental study of thedissolution rates of basaltic glass as a function of pH and temperature,Geochim. Cosmochim. Acta, 67(20), 3817–3832.Greeley, R., and J. E. Guest (1987), Geologic map of the eastern equatorialregion of Mars, U. S. Geol. Surv. Misc. Invest. Ser., Map I-1802–B, scale1:15,000,000.Grove, T. L., S. W. Parman, S. A. Bowring, R. C. Price, and M. B. Baker(2002), The role of an H2O-rich fluid component in the generation ofprimitive basaltic andesites and andesites from the Mt. Shasta region,N. California, Contrib. Mineral. Petrol., 142, 375–396.Guidry, M. W., and F. T. Mackenzie (2003), Experimental study of igneousand sedimentary apatite dissolution: Control of pH, distance from equili-brium, and temperature on dissolution rates, Geochim. Cosmochim. Acta,67(16), 2949–2963.Gulick, V. C. (1998), Magmatic intrusion and a hydrothermal origin forfluvial valleys on Mars, J. Geophys. Res., 103, 19,365–19,387.Gulick, V. C. (2001), Origin of the valley networks on Mars: A hydrolo-gical perspective, Geomorphology, 37, 241–268.Hagerty, J. J., and H. E. Newsom (2003), Hydrothermal alteration at theLonar Lake impact structure, India: Implications for impact cratering onMars, Meteorol. Planet. Sci., 38, 365–381.Haining, R. (2003), Spatial Data Analysis: Theory and Practice, pp. 273–286, Cambridge Univ. Press, New York.Halliday, A. N., D.-C. Lee, S. Tommasini, G. R. Davies, C. R. Paslick, J. G.Fitton, and D. E. James (1995), Incompatible trace elements in OIB andMORB and source enrichment in the sub-oceanic mantle, Earth Planet.Sci. Lett., 133, 379–395.Hamilton, V. E., and P. R. Christensen (2005), Evidence for extensive,olivine-rich bedrock on Mars, Geology, 33, 433–436.Hamilton, V. E., P. R. Christensen, H. Y. McSween Jr., and J. L. Bandfield(2003), Searching for the source regions of Martian meteorites usingMGS TES: Integrating Martian meteorites into the global distributionof igneous materials on Mars, Meteorit. Planet. Sci., 38, 871–885.Hanna, J. C., and R. J. Phillips (2005), A new system of tectonic outflowchannels in the Memnonia region of Mars, , 86(52), Fall Meet. Suppl.,Abstract P23B-0194.Harvey, R. H., and V. E. Hamilton (2005), Syrtis Major as the source regionof the nakhlite/chassigny group of Martian meteorites: Implications forthe geological history of Mars, Lunar Planet. Sci., XXXVI, Abstract 1019.Harvey, R. H., M. Wadhwa, H. Y. McSween Jr., and G. Crozaz (1993),Petrography, mineral chemistry, and petrogenesis of Antarctic ShergottiteLEW88516, Geochim. Cosmochim. Acta, 57, 4769.Hauri, E. H., T. P. Wagner, and T. L. Grove (1994), Experimental andnatural partitioning of Th, U, Pb, and other trace elements between gar-net, clinopyroxene, and basaltic melts, Chem. Geol., 117, 149–166.Hawkesworth, C., S. Turner, D. Peate, F. McDermott, and P. van Calsteren(1997), Elemental U and Th variations in island arc rocks: Implicationsfor U-series isotopes, Chem. Geol., 139, 207–221.Head, J. W., and S. Pratt (2001), Extensive Hesperian-aged south polar icesheet on Mars: Evidence for massive melting and retreat, and lateral flowand ponding of meltwater, J. Geophys. Res., 106, 12,275–12,299.Head, J. W., III, D. R. Marchant, and G. J. Ghatan (2004), Glacial depositson the rim of a Hesperian-Amazonian outflow channel source trough:Mangala Valles, Mars, Geophys. Res. Lett., 31, L10701, doi:10.1029/2004GL020294.Head, J. W., et al. (2005), Tropical to mid-latitude snow and ice accumula-tion, flow and glaciation on Mars, Nature, 434, 346–351.Hiesinger, H., and J. W. Head III (2004), The Syrtis Major volcanic pro-vince, Mars: Synthesis from Mars Global Surveyor data, J. Geophys.Res., 109, E01004, doi:10.1029/2003JE002143.Honnorez, J. J. (1981), The aging of the oceanic crust at low temperature, inThe Oceanic Lithosphere, edited by C. Emiliani, pp. 525–587, JohnWiley, Hoboken, N. J.Hurowitz, J. A., S. M. McLennan, N. J. Tosca, R. E. Arvidson, J. R.Michalski, D. W. Ming, C. Schro¨der, and S. W. Squyres (2006), In situand experimental evidence for acidic weathering of rocks and soils onMars, J. Geophys. Res., 111, E02S19, doi:10.1029/2005JE002515.Hynek, B. M., and R. J. Phillips (2001), Evidence for extensive denudationof the Martian highlands, Geology, 29, 407–410.Ivanov, M. A., and J. W. Head III (2001), Chryse Planitia, Mars: Topo-graphic configuration, outflow channel continuity and sequence, and testsfor hypothesized ancient bodies of water using Mars Orbiter Laser Alti-meter (MOLA) data, J. Geophys. Res., 106, 3275–3295.Jagoutz, E., H. Palme, H. Baddenhausen, K. Blum, M. Cendales, G. Dreibus,B. Spettel, V. Lorenz, and H. Wa¨nke (1979), The abundances of major,minor, and trace elements in the Earth’s mantle as derived from primitiveultramafic nodules, Proc. Lunar Planet. Sci. Conf. 10th, 2031–2050.Jochum, K. P., and S. P. Verma (1996), Extreme enrichment of Sb, Tl andother trace elements in altered MORB, Chem. Geol., 130, 289–299.E03S06TAYLOR ET AL.: VARIATIONS IN K/Th18 of 20E03S06Jones, J. H. (1995), Experimental trace element partitioning, in A Handbookof Physical Constants: Rock Physics and Phase Relations, AGU Ref. ShelfSer., vol. 3, edited by T. J. Ahrens, pp. 73–104, AGU, Washington, D. C.Kargel, J. S. (2004), Mars: A Warmer Wetter Planet, 557 pp., Springer,New York.Kargel, J. S., and R. G. Strom (1992), Ancient glaciation on Mars, Geology,20, 3–7.Karunatillake, S. K., et al. (2006), Composition of northern low-albedoregions of Mars: Insights from the Mars Odyssey Gamma Ray Spectro-meter, J. Geophys. Res., doi:10.1029/2006JE002675, in press.Kato, T., A. E. Ringwood, and T. Irifune (1988), Experimental determina-tion of element partitioning between silicate perovskites, garnets andliquids: Constraints on early differentiation of the mantle, Earth Planet.Sci. Lett., 89, 123–145.Keller, J., et al. (2006), Equatorial and midlatitude distribution of chlorinemeasured by Mars Odyssey G RS, J. Geophys. Res. , 111, E 03 S0 8,doi:10.1029/2006JE002679. [printed 112(E3), 2007]Klingelho¨fer, G., et al. (2004), Jarosite and hematite at Meridiani Planumfrom Opportunity’s Mo¨ssbauer Spectrometer, Science, 306, 1740–1745.Komatsu, G., J. M. Dohm, and T. M. Hare (2004a), Hydrogeologic pro-cesses of large-scale tectonomagmatic complexes in Mongolia-southernSiberia and on Mars, Geology, 32, 325–328.Komatsu, G., G. G. Ori, P. Ciarcelluti, and Y. Litasov (2004b), Interiorlayered deposits of Valles Marineris, Mars: Analogous subice volcanismrelated to Baikal rifting, Southern Siberia, Planet. Space Sci., 52, 167–187.Krauskopf, K. B., and D. K. Bird (1995), Introduction to Geochemistry, 647pp., McGraw-Hill, New York.Langmuir, D., and J. S. Herman (1980), The mobility of thorium innatural waters at low temperatures, Geochim. Cosmochim. Acta, 44,1753–1766.Lodders, K. (1998), A survey of shergottites, nahklite and chassigny me-teorites whole-rock compositions, Meteorol. Planet. Sci., 33, A183–A190.Malin, M. C. (1979), Mars: Evidence of indurated deposits of fine materials(abstract), NASA Conf. Publ., 2072, 54.Malin, M. C., and M. H. Carr (1999), Groundwater formation of Martianvalleys, Nature, 397, 589–591.Malin, M. C., and K. S. Edgett (1999), Oceans or seas in the Martiannorthern lowlands: High resolution imaging tests of proposed coastlines,Geophys. Res. Lett., 26, 3049–3052.Malin, M. C., and K. S. Edgett (2000), Evidence for recent groundwaterseepage and surface runoff on Mars, Science, 288, 2330–2335.Mangold, N., C. Quantin, V. Ansan, C. Delacourt, and P. Allemand (2004),Evidence for precipitation on Mars from dendritic valleys in the VallesMarineris area, Science, 305, 78–81.Masson, P., M. H. Carr, F. Costard, R. Greeley, E. Hauber, and R. Jaumann(2001), Geomorphic evidence for liquid water, Space Sci. Rev., 96, 333–364.McCoy, T. J., G. J. Taylor, and K. Keil (1992), Zagami: Product of a two-stage magmatic history, Geochim. Cosmochim. Acta, 56, 3571–3582.McDonough, W. F., and S.-s. Sun (1995), The composition of the Earth,Chem. Geol., 120, 223–253.McLennan, S. M. (2003), Large-ion lithophile element fractionation duringthe early differentiation of Mars and the composition of the Martianprimitive mantle, Meteorit. Planet. Sci., 38, 895–904.Mellon, M. T., and R. J. Phillips (2001), Recent gullies on Mars and thesource of liquid water, J. Geophys. Res., 106(E10), 23,165–23,180.Meyer, C., Jr. (2006), Mars Meteorite Compendium, JSC#27672 RevisionC, NASA Johnson Space Cent., Houston, Tex. (Available at http://curator.jsc.nasa.gov/antmet/mmc/index.cfm)Miyamoto, H., J. M. Dohm, R. A. Beyer, and V. R. Baker (2004), Fluiddynamical implications of anastomosing slope streaks on Mars, J. Geo-phys. Res., 109, E06008, doi:10.1029/2003JE002234.Moore, J. M., and D. E. Wilhelms (2001), Hellas as a possible site ofancient ice-covered lakes on Mars, Icarus, 154, 258–276.Mouginis-Mark, P. J., L. Wilson, and M. T. Zuber (1992), Physical volca-nology, in Mars, edited by H. H. Kieffer et al., pp. 424–452, Univ. ofAriz. Press, Tucson.Mustard, J. F., C. D. Cooper, and M. K. Rifkin (2001), Evidence for recentclimate change on Mars from the identification of youthful near-surfaceground ice, Nature, 412, 411–414.Mustard, J. F., F. Poulet, A. Gendrin, J.-P. Bibring, Y. Langevin, B. Gonder,N. Mangold, G. Bellucci, and F. Altieri (2005a), Olivine and pyroxenediversity in the crust of Mars, Science, 307, 1594–1597.Mustard, J. F., et al. (2005b), Crustal formation, volcanism, and alteration inthe Syrtis Major region revealed by OMEGA data, Lunar Planet. Sci.,XXXVI, Abstract 1341.Nesbitt, H. W., and G. Markovics (1997), Weathering of granodioritic crust,long-term storage of elements in weathering profiles, and petrogenesis ofsiliciclastic sediments, Geochim. Cosmochim. Acta, 61, 1653–1670.Nesbitt, H. W., and R. E. Wilson (1992), Recent chemical weathering ofbasalts, Am. J. Sci., 292, 740–777.Neukum, G., et al. (2004), Recent and episodic volcanic and glacial activityon Mars revealed by the high resolution stereo camera, Nature, 432,971–979.Newsom, H. E. (1980), Hydrothermal alteration of impact melts sheets withimplications for Mars, Icarus, 44, 207–216.Newsom, H. E., G. E. Brittelle, C. A. Hibbitts, L. J. Crossey, and A. M.Kudo (1996), Impact crater lakes on Mars, J. Geophys. Res., 101(E6),14,951–14,955.Newsom, H. E., J. J. Hagerty, and F. Goff (1999), Mixed hydrothermalfluids and the origin of the Martian soil, J. Geophys. Res., 104, 8717–8728.Nicholson, K. (1993), Geothermal Fluids, 2nd ed., 263 pp., Springer, NewYork.Nimmo, F., and K. L. Tanaka (2005), Early crustal evolution of Mars, Annu.Rev. Earth Planet. Sci., 33, 133–161.Papike, J. J. (1992), The valley of Ten Thousand Smokes, Katmai, Alaska:A unique geochemistry laboratory, Geochim. Cosmochim. Acta, 56,1429–1449.Papike, J. J., G. Ryder, and C. Shearer (1998), Lunar samples, in PlanetaryMaterials, Rev. Mineral., vol. 36, edited by J. J. Papike, pp. 5-01 to5-234, Mineral. Soc. of Am., Washington, D. C.Parker, T. J., D. S. Gorsline, R. S. Saunders, D. C. Pieri, and D. M.Schneeberger (1993), Coastal geomorphology of the Martian NorthernPlains, J. Geophys. Res., 98(E6), 11,061–11,078.Patino, L. C., M. A. Velbel, J. R. Price, and J. A. Wade (2003), Traceelement mobility during spheroidal weathering of basalts and andesitesin Hawaii and Guatemala, Chem. Geol., 202, 343–364.Pellicori, D. A., C. H. Gammons, and S. R. Poulson (2005), Geochemistryand stable isotopes composition of the Berkeley pit lake and surroundingmine waters, Butte, Montana, Appl. Geochem., 20(11), 2116–2137.Poulet, F., J.-P. Bibring, J. F. Mustard, A. Gendrin, N. Mangold,Y. Langevin, R. E. Arvidson, B. Gondet, C. Gomez, and the OmegaTeam (2005), Phyllosilicates on Mars and implications for early Martianclimate, Nature, 438, 623–627.Prettyman, T., W. C. Feldman, D. J. Lawrence, G. W. McKinney, A. B.Binder, R. C. Elphic, O. M. Gasnault, S. Maurice, and K. R. Moore(2002), Library least squares analysis of Lunar Prospector gamma rayspectra, Lunar Planet. Sci. [CD-ROM], XXXII, Abstract 2012.Price, J. R., and M. A. Velbel (2003), Chemical weathering indices appliedto weathering profiles developed on heterogeneous felsic metamorphicparent rocks, Chem. Geol., 202, 397–416.Rieder, R., et al. (2004), Chemistry of rocks and soils at Meridiani Planumfrom the Alpha Particle X-ray Spectrometer, Science, 306, 1746–1749.Robinson, M. S., P. J. Mouginis-Mark, J. R. Zimbelman, S. S. C. Wu, K. K.Ablin, and A. E. Howington-Kraus (1993), Chronology, eruption dura-tion, and atmospheric contribution of the Martian volcano, Apollinarispatera, Icarus, 104, 301–323.Rotto, S. L., and K. L. Tanaka (1995), Geologic/geomorphologic map ofthe Chryse Planitia region of Mars, U. S. Geol. Surv. Misc. Invest. Ser.,Map I-2441.Sak, P. B., D. M. Fisher, T. W. Gardner, K. Murphy, and S. L. Brantley(2004), Rates of weathering rind formation on Costa Rican basalt, Geo-chim. Cosmochim. Acta, 68, 1453–1472.Sakimoto, S. E. H., H. V. Frey, J. B. Garvin, and J. H. Roark (1999),Topography, roughness, layering, and slope properties of the MedusaeFossae Formation from Mars Orbiter Laser Altimeter (MOLA) andMars Orbiter Camera (MOC) data, J. Geophys. Res., 104, 24,141–24,154.Sawyer, D. J., M. D. McGhee, J. Canepa, and C. B. Moore (2000), Watersoluble ions in the Nakhla Martian meteorite, Meteorit. Planet. Sci., 35,743–747.Schultz, P. H., and A. B. Lutz (1988), Polar wandering on Mars, Icarus, 73,91–141.Scott, D. H., and M. G. Chapman (1991), Geological map of science studyarea 6, Memnonia region, Mars, U. S. Geol. Surv. Misc. Invest. Ser., MapI-2084, scale 1:500,000.Scott, D. H., and K. L. Tanaka (1982), Ignimbrites of Amazonis Planitiaregion of Mars, J. Geophys. Res., 87, 1179–1190.Scott, D. H., and K. L. Tanaka (1986), Geologic map of the western equa-torial region of Mars, U. S. Geol. Surv. Misc. Invest. Ser., Map I-1802–A,scale 1:15,000,000.Scott, D. H., and J. R. Zimbelman (1995), Geologic map of Arsia Monsvolcano, Mars, U. S. Geol. Surv. Misc. Invest. Ser., Map I-2480, scale1:1,000,000.Scott, D. H., J. M. Dohm, and D. J. Applebee (1993), Geologic map ofscience study area 8, Apollinaris Patera region of Mars, U. S. Geol. Surv.Misc. Invest. Ser., Map I-2351, scale 1:500,000.E03S06TAYLOR ET AL.: VARIATIONS IN K/Th19 of 20E03S06Scott, D. H., J. M. Dohm, and J. W. Rice Jr. (1995), Map of Mars showingchannels and possible paleolake basins, U. S. Geol. Surv. Misc. Invest.Ser., Map I-2461, scale 1:30,000,000.Scott, D. H., J. M. Dohm, and J. R. Zimbelman (1998), Geologic map ofPavonis Mons volcano, Mars, U. S. Geol. Surv. Misc. Invest. Ser., Map I-2561, scale 1:1,000,000.Shean, D. E., J. W. Head, and D. R. Marchant (2005), Origin and evolutionof a cold-based tropical mountain glacier on Mars: The Pavonis Monsfan-shaped deposit, J. Geophys. Res., 110, E05001, doi:10.1029/2004JE002360.Shearer, C. K., and J. J. Papike (1999), Magmatic evolution of the Moon,Am. Mineral., 84, 1469–1494.Solomon, S. C., et al. (2005), New perspectives on ancient Mars, Science,307, 1214–1220.Squyres, S. W., et al. (2004a), The Spirit rover’s Athena science investiga-tion at Gusev Crater, Mars, Science, 305, 794–799.Squyres, S. W., et al. (2004b), In situ evidence for an ancient aqueousenvironment at Meridiani Planum, Mars, Science, 306, 1709–1714.Squyres, S. W., et al. (2004c), The Opportunity rover’s Athena scienceinvestigation at Meridiani Planum, Mars, Science, 306, 1698–1703.Staudigel, H., T. Plank, B. White, and H.-U. Schmincke (1996), Geochem-ical fluxes during seafloor alteration of the basaltic upper crust: DSDPSites 417 and 418, in Subduction: Top to Bottom, Geophys. Monogr. Ser.,vol. 96, edited by G. E. Bebout et al., pp. 19–38, AGU, Washington,D. C.Stillings, L. L., and S. L. Brantley (1995), Feldspar dissolution at 25�C andpH 3: Reaction stoichiometry and the effect of cations, Geochim. Cos-mochim. Acta, 59(8), 1483–1496.Stolper, E. M., and H. Y. McSween Jr. (1979), Petrology and origin of theshergottite meteorites, Geochem. Cosmochim. Acta, 43, 1475–1498.Tanaka, K. L. (1986), The stratigraphy of Mars, Proc. Lunar Planet. Sci.Conf. 17th, part 1, J. Geophys. Res., 91, suppl., E139–E158.Tanaka, K. L., J. M. Dohm, J. H. Lias, and T. M. Hare (1998), Erosionalvalleys in the Thaumasia region of Mars: Hydrothermal and seismicorigins, J. Geophys. Res., 103, 31,407–31,419.Tanaka, K. L., J. S. Kargel, D. J. MacKinnon, T. M. Hare, and N. Hoffman(2002), Catastrophic erosion of Hellas basin rim on Mars induced bymagmatic intrusion into volatile-rich rocks, Geophys. Res. Lett., 29(8),1195, doi:10.1029/2001GL013885.Tanaka, K. L., J. A. Skinner Jr., T. M. Hare, T. Joyal, and A. Wenker(2003), Resurfacing history of the northern plains of Mars based ongeologic mapping of Mars Global Surveyor data, J. Geophys. Res.,108(E4), 8043, doi:10.1029/2002JE001908.Tanaka, K. L., J. A. Skinner Jr., and T. M. Hare (2005), Geologic map of thenorthern plains of Mars, U. S. Geol. Surv. Misc. Invest. Ser., Map SIM-2888, scale 1:15000,000.Taylor, G. J., et al. (2006), Bulk composition and early differentiation ofMars, J. Geophys. Res., doi:10.1029/2005JE002645, in press.Taylor, S. R., and S. M. McLennan (1985), The Continental Crust: ItsComposition and Evolution, 312 pp., Blackwell Sci., Malden, Mass.Tosca, N. J., S. M. McLennan, B. C. Clark, J. P. Grotzinger, J. A. Hurowitz,A. H. Knoll, C. Schroder, and S. W. Squyres (2005), Geochemical mod-eling of evaporation processes on Mars: Insight from the sedimentaryrecord at Meridiani Planum, Earth Planet. Sci. Lett., 240, 122–148.Treiman, A. H. (2003), Geologic settings of Martian gullies: Implicationsfor their origins, J. Geophys. Res., 108(E4), 8031, doi:10.1029/2002JE001900.Turner, S., C. Hawkesworth, P. van Calsteren, E. Heath, R. Macdonald, andS. Black (1996), U-series isotopes and destructive plate margin magmagenesis in the Lesser Antilles, Earth Planet. Sci. Lett., 142, 191–207.Upton, G. J. G., and B. Fingleton (1985), Spatial Data Analysis by Exam-ple: Point Pattern and Quantitative Data, J. Wiley, Hoboken, N. J.Whalley, W. B., and F. Azizi (2003), Rock glaciers and protalus landforms:Analogous forms and ice sources on Earth and Mars, J. Geophys. Res.,108(E4), 8032, doi:10.1029/2002JE001864.White, A. F., and S. L. Brantley (2003), The effect of time on the weath-ering of silicate minerals: Why do weathering rates differ in the labora-tory and field?, Chem. Geol., 202, 479–506.Wilson, L., and J. W. Head III (2004), Evidence for a massive phreatomag-matic eruption in the initial stages of formation of the Mangala Vallesoutflow channel, Mars, Geophys. Res. Lett., 31, L15701, doi:10.1029/2004GL020322.Zimbelman, J. R., R. A. Craddock, and R. Greeley (1994), Geologic map ofthe MTM-15147 quadrangle, Mangala Valles region of Mars, U. S. Geol.Surv. Misc. Invest. Ser., Map I-2402, 1:500,000.�����������������������V. R. Baker, W. V. Boynton, J. M. Dohm, D. Hamara, D. Janes, J. M.Keller, K. E. Kerry, K. J. Kim, and A. L. Sprague, Lunar and PlanetaryLaboratory, University of Arizona, Tucson, AZ 85721, USA.J. Bru¨ckner, G. Dreibus, and H. Wa¨nke, Max-Planck-Institu¨t fu¨r Chemie,Postfach 3060, D-55020 Mainz, Germany.D. M. Drake, TechSource, P.O. Box 31057, Santa Fe, NM 87594-1057,USA.C. d’Uston, O. Gasnault, and S. Maurice, Centre d’Etude Spatiale desRayonnements, CNRS/UPS, 9 Avenue du Colonel Roche, Boite Postale4346, F-31028 Toulouse Cedex 4, France.P. Englert, L. M. V. Martel, J. D. Stopar, and G. J. Taylor, Hawaii Instituteof Geophysics and Planetology, 1680 East-West Road, Honolulu, HI 96822,USA. (gjtaylor@higp.hawaii.edu)L. G. Evans, Computer Sciences Corporation, 7700 Hubble Drive,Lanham, MD 20706, USA.B. C. Hahn and S. M. McLennan, Department of Geosciences, StateUniversity of New York, Stony Brook, NY 11794, USA.S. Karunatillake and S. W. Squyres, Center for Radiophysics and SpaceResearch, Cornell University, Ithaca, NY 14853, USA.R. Reedy, Institute of Meteoritics, University of New Mexico,Albuquerque, NM 87131, USA.R. Starr, Department of Physics, Catholic University of America,Washington, DC 20064, USA.E03S06TAYLOR ET AL.: VARIATIONS IN K/Th20 of 20E03S06 | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/341878299�����������������������������������������������������������������������������������������������������DECISION TO PURCHASE IN MARKETING MIX OF M HERB DIETARYSUPPLEMENTS OF CUSTOME...Conference Paper · June 2020CITATIONS0READS541 author:Some of the authors of this publication are also working on these related projects:�������������������������������������������������������������������������������������������������������� View project7th National and International Conference on Administration and Management 24-25 February 2020 Mahanakorn University of Technology, Bangkok, Thailand ViewprojectPashatai CharutawephonnukoonRangsit University44 PUBLICATIONS 2 CITATIONS SEE PROFILEAll content following this page was uploaded by Pashatai Charutawephonnukoon on 03 June 2020.The user has requested enhancement of the downloaded file.การตัดสินใจเลือกซื้อจากส่วนประสมการตลาดของผลิตภัณฑ์เสริมอาหาร เอ็มเฮิร์บของผู้บริโภคในเขตกรุงเทพมหานคร พัชร์หทัย จารุทวีผลนุกูล1 ผู้อ ำนวยกำรหลักสูตรบริหำรธุรกิจมหำบัณฑิต คณะบริหำรธุรกิจ มหำวิทยำลัยรังสิต โชติโรส มูลสท้าน กชกร แหเลิศตระกูล ธัญชนก ยูซบ นักศึกษำปริญญำโทหลักสูตรบริหำรธุรกิจมหำบัณฑิต คณะบริหำรธุรกิจ มหำวิทยำลัยรังสิต มรกต จันทร์กระพ้อ อำจำรย์พิเศษ คณะบริหำรธุรกิจ มหำวิทยำลัยรังสิต บทคัดย่อ กำรวิจัยนี้มีวัตถุประสงค์ (1) เพื่อศึกษำกำรตัดสินใจเลือกซื้อจำกส่วนประสมกำรตลำดของผลิตภัณฑ์เสริมอำหำรเอ็มเฮิร์บของผู้บริโภค และ (2) เพื่อศึกษำควำมแตกต่ำงของกำรตัดสินใจเลือกซื้อจำกส่วนประสมกำรตลำดของผลิตภัณฑ์เสริมอำหำรเอ็มเฮิร์บ จ ำแนกตำมปัจจัยส่วนบุคคลของผู้บริโภค ประชำกรที่ใช้ในกำรวิจัยครั้งนี้ คือ ผู้บริโภคที่ซื้อผลิตภัณฑ์อำหำรเสริมเอ็มเฮิร์บในเขตกรุงเทพมหำนคร โดยก ำหนดจ ำนวนตัวอย่ำงทั้งสิ้น 400 คน ด้วยวิธีกำรสุ่มตัวอย่ำงแบบตำมสะดวก ใช้แบบสอบถำมในกำรเก็บรวบรวมข้อมูล และน ำมำวิเครำะห์ข้อมูลด้วยสถิติเชิงพรรณนำ ประกอบด้วย ค่ำควำมถี่ค่ำร้อยละ ค่ำเฉลี่ย และค่ำส่วนเบี่ยงเบนมำตรฐำน และสถิติเชิงอนุมำน ได้แก่ กำรทดสอบค่ำเฉลี่ย 2 กลุ่มตัวอย่ำงแบบอิสระ และกำรวิเครำะห์ควำมแปรปรวนทำงเดียว ผลกำรวิจัยพบว่ำ ปัจจัยส่วนบุคคลของผู้บริโภคในกรุงเทพมหำนคร ได้แก่ อำยุ สถำนภำพสมรสและรำยได้เฉลี่ยต่อเดือนที่แตกต่ำงกันส่งผลต่อกำรตัดสินใจเลือกซื้อในส่วนประสมกำรตลำดของผลิตภัณฑ์เสริมอำหำรเอ็มเฮิร์บแตกต่ำงกัน อย่ำงมีนัยส ำคัญทำงสถิติที่ระดับ 0.05 โดยด้ำนอำยุและด้ำนสถำนภำพสมรสนั้นจะมีควำมแตกต่ำงกันในด้ำนของรำคำและกำรส่งเสริมกำรตลำด ส่วนด้ำนรำยได้เฉลี่ยต่อเดือนนั้นจะมีควำมแตกต่ำงกันในด้ำนของกำรส่งเสริมกำรตลำด ค าส าคัญ: ส่วนประสมกำรตลำด กำรตัดสินใจเลือกซื้อ ผลิตภัณฑ์เสริมอำหำร 1 ผู้อ ำนวยกำรหลักสูตรบริหำรธุรกิจมหำบัณฑิต คณะบริหำรธุรกิจ มหำวิทยำลัยรังสิต ต.หลักหก อ.เมือง จ.ปทุมธำนี 12000 หมำยเลขติดต่อ: 096-945-2522 อีเมล: pashatai.c@rsu.ac.th การประชุมวิชาการระดับชาติ RTBEC 2020DECISION TO PURCHASE IN MARKETING MIX OF M HERB DIETARY SUPPLEMENTS OF CUSTOMER IN BANGKOK Pashatai Charutawephonnukoon1 Director of Master of Business Administration Program, Faculty of Business Administration Rangsit University Chotiros Moonsatan Kodchakorn haeloedtrakun Thanchanok Yusob Graduate Student, Master of Business Administration, Faculty of Business Administration Rangsit University Morakot Junkrapor Part-time lecturer, Faculty of Business Administration Rangsit University Abstract This research had the objective (1) to study customers’ decision to purchase affected by marketing mix of “M Herb” dietary supplements, and (2) to study difference of customers’ decision to purchase affected by marketing mix of “M Herb” dietary supplements classified by their personal factors. Population of this research was customers buying “M Herb” dietary supplements in Bangkok and samples were of 400 ones. Convenience sampling method was used to select customer. Questionnaires were distributed for data collection. Data analysis employed descriptive statistics including frequency, percentage, mean, and standard deviation and inferential statistics including t-test and F-test. The results found that customers’ different personal factor in terms of age, marital status, and income differently affected customers’ decision to purchase by marketing mix of “M Herb” dietary supplements at the statistical significance at the level of 0.05. In terms of age and marital status, there will be differences in price and promotion. As for income aspect, there are differences in terms of promotion. Keywords: Marketing Mix, Decision to Purchase, Dietary Supplements 1 Director of Master of Business Administration Program, Faculty of Business Administration Rangsit University. Tel: +6696-945-2522 Email: pashatai.c@rsu.ac.th การประชุมวิชาการระดับชาติ RTBEC 2020512บทน า ปัจจุบันกระแสรักสุขภำพเป็นที่นิยมอย่ำงมำกทั้งในประเทศไทยและต่ำงประเทศ ซึ่งประชำกรโลกต่ำงหันมำรับประทำนอำหำรที่มีประโยชน์ต่อร่ำงกำย เช่น ผัก และผลไม้ปลอดสำรพิษ เพรำะปัจจุบันโรคภัยที่มำจำกกำรรับประทำนอำหำรไม่ถูกสุขลักษณะทวีควำมรุนแรงขึ้นทุกปี อำหำรเสริมจึงเป็นทำงเลือกที่ปัจจุบันก ำลังได้รับควำมนิยมส ำหรับกลุ่มผู้ที่เร่งรีบมีเวลำดูแลตัวเองน้อยเช่นกลุ่มคนวัยท ำงำนอำยุ 30 ปีขึ้นไป ที่ต้องกำรดูแลตัวเอง หรือคนที่ไม่ชอบทำนอำหำรชนิดใดชนิดหนึ่ง แล้วท ำให้ขำดสำรอำหำรบำงชนิด ด้วยกำรรับประทำนอำหำรเสริมชดเชยทั้งนี้ ท ำให้กำรแข่งขันในตลำดอำหำรเสริมค่อนข้ำงรุนแรง (สยำมธุรกิจ, 2559) ภำพรวมรำยได้ตลำดอำหำรเสริมทั่วโลกในปี ค.ศ. 2016 ข้ำงต้นนี้มีมูลค่ำอยู่ที่ 121.2 พันล้ำนเหรียญสหรัฐ และมีอัตรำเติบโตเฉลี่ยสะสมต่อปีอยู่ที่ 5.7 เปอร์เซ็นต์ (ปี ค.ศ. 2006-2020) โดยกำรคำดกำรณ์จำกวำรสำร Nutrition Business Journal (NBJ) ระบุว่ำ รำยได้ตลำดอำหำรเสริมทั่วโลกจะมีมูลค่ำสูงถึง155 พันล้ำนเหรียญสหรัฐ ในปี ค.ศ. 2020 ซึ่งรำยได้ดังกล่ำวจะมำจำกกำรเติบโตในภูมิภำคเอเชีย (ไม่รวมประเทศจีน ญี่ปุ่นและอินเดีย) และภูมิภำคยุโรปตะวันออก โดยคำดกำรณ์ว่ำจะมีอัตรำกำรเติบโตเฉลี่ยสะสมต่อปีสูงถึง 9 เปอร์เซ็นต์ (สถิติอุตสำหกรรมอำหำรเสริม, 2562) ซึ่งสอดคล้องกับประเทศไทยในปี พ.ศ. 2562 ที่ทิศทำงของตลำดอำหำรเสริมเริ่มฟื้นตัวดีกว่ำปีที่ผ่ำนมำค่อนข้ำงมำกและรัฐบำลนั้นมีมำตรกำรที่เข้มงวดในกำรตรวจสอบผลิตภัณฑ์เสริมอำหำรที่ท ำผิดกฎหมำย ไม่ถูกต้อง ไม่ได้มำตรฐำน และโฆษณำเกินจริงอย่ำงเข้มงวด ท ำให้ในปีนี้ผู้บริโภคมีควำมมั่นใจว่ำผลิตภัณฑ์เสริมอำหำรที่วำงจ ำหน่ำยในท้องตลำดมีควำมปลอดภัย ซึ่งผลวิจัยจำก TCDC เมื่อปี พ.ศ. 2561 ระบุว่ำ ธุรกิจที่ส่งเสริมให้คนมีสุขภำพดีเติบโตขึ้นทั่วโลกถึงร้อยละ 10.6 หรือคิดเป็น 122 ล้ำนล้ำนบำทต่อปี ทั้งนี้ ธุรกิจที่เกี่ยวข้องกับกำรส่งเสริมสุขภำพเติบโตอย่ำงรวดเร็วในทุกปีและยังมีแนวโน้มเติบโตอย่ำงต่อเนื่อง (TCDC, 2018) ผลิตภัณฑ์เสริมอำหำรที่จ ำหน่ำยในท้องตลำดมีเป็นจ ำนวนมำก ซึ่งล้วนมีส่วนประกอบและประโยชน์ด้ำนกำรส่งเสริมสุขภำพแตกต่ำงกันไป หนึ่งในผลิตภัณฑ์เสริมอำหำรของบริษัทเนเจอร์ เฮิร์บ อินเตอร์เนชั่นแนล โฮลดิ้ง จ ำกัด ซึ่งเป็นที่รู้จักกันอย่ำงกว้ำงขวำงจำกกำรพัฒนำต่อยอดมำจำกสูตรต ำรับของคุณหมอแสง หรือนำยแสงชัย แหเลิศตระกูล ผู้ซึ่งได้รับกำรขึ้นทะเบียนเป็นหมอพื้นบ้ำนที่จังหวัดปรำจีนบุรี โดยใช้ภูมิปัญญำที่ได้รับกำรถ่ำยทอด และพัฒนำสืบต่อกันมำจำกรุ่นสู่รุ่น จนได้เป็นสูตรต ำรับที่ดีที่สุดในปัจจุบัน และได้รับกำรพิสูจน์จำกกำรใช้จริงของผู้คนจ ำนวนมำกมำยำวนำนกว่ำ 10 ปี จนสำมำรถกล่ำวได้ว่ำผลิตภัณฑ์เสริมอำหำรเอ็มเฮิร์บสำมำรถท ำให้สุขภำพร่ำงกำยแข็งแรงขึ้นอีกทั้งยังมีสรรพคุณในกำรต่อต้นโรคมะเร็ง (Nature Herb, 2020) ดังนั้นผู้วิจัยจึงเห็นควำมส ำคัญและมีควำมสนใจที่จะศึกษำ กำรตัดสินใจเลือกซื้อในส่วนประสมกำรตลำดของผลิตภัณฑ์เสริมอำหำรเอ็มเฮิร์บของผู้บริโภคในกรุงเทพมหำนคร เพื่อให้ผู้ที่ด ำเนินธุรกิจผลิตภัณฑ์เสริมอำหำรเพื่อสุขภำพ น ำผลกำรศึกษำที่ได้ ไปใช้ในกำรพัฒนำต่อยอดธุรกิจในด้ำนกำรวำงแผน กลยุทธ์กำรสื่อสำรทำงกำรตลำด เพื่อเข้ำใจกลุ่มเป้ำหมำย และสำมำรถตอบสนองควำมต้องกำร ของกลุ่มเป้ำหมำยได้สูงสุด วัตถุประสงค์ในการวิจัย 1. เพื่อศึกษำกำรตัดสินใจเลือกซื้อในส่วนประสมกำรตลำดของผลิตภัณฑ์เสริมอำหำรเอ็มเฮิร์บของผู้บริโภคในกรุงเทพมหำนคร 2. เพื่อศึกษำควำมแตกต่ำงของกำรตัดสินใจเลือกซื้อในส่วนประสมกำรตลำดของผลิตภัณฑ์เสริมอำหำรเอ็มเฮิร์บ จ ำแนกตำมปัจจัยส่วนบุคคลของผู้บริโภคในกรุงเทพมหำนคร การประชุมวิชาการระดับชาติ RTBEC 2020513สมมติฐานการวิจัย ปัจจัยส่วนบุคคลของผู้บริโภคในกรุงเทพมหำนครที่แตกต่ำงกันส่งผลต่อกำรตัดสินใจเลือกซื้อในส่วนประสมกำรตลำดของผลิตภัณฑ์เสริมอำหำรเอ็มเฮิร์บแตกต่ำงกัน กรอบแนวคิดการวิจัย ตัวแปรอิสระ ตัวแปรตาม ภาพที่ 1 กรอบแนวคิดกำรวิจัย การทบทวนวรรณกรรม แนวคิดและทฤษฎีที่เกี่ยวกับส่วนประสมกำรตลำด หมำยถึง กลยุทธ์ขององค์กำรธุรกิจที่มุ่งเน้นสร้ำงควำมแตกต่ำงทำงกำรตลำดและรักษำควำมได้เปรียบทำงกำรแข่งขันขององค์กำร (Daniel, 2018) ซึ่งมีวัตถุประสงค์เพื่อเสนอขำยสินค้ำให้กับผู้บริโภคด้วยผลิตภัณฑ์ที่มีคุณภำพ ในรำคำที่เหมำะสม สถำนที่ที่ลูกค้ำเข้ำถึงได้ และกลยุทธ์กำรส่งเสริมกำรคลำดที่มีประสิทธิภำพ (Gituma, 2017) ส่วนประสมกำรตลำดเป็นแนวคิดแบบดังเดิมที่ประกอบไปด้วย 4 องค์ประกอบ (4Ps) ได้แก่ ด้ำนผลิตภัณฑ์ (Product) หมำยถึง ผลิตภัณฑ์ที่มีตัวตนหรือไม่มีตัวตนที่ตอบสนองควำมต้องกำรของผู้บริโภค (Ivy, 2008) ด้ำนรำคำ (Price) หมำยถึง รำคำของผลิตภัณฑ์หรือบริกำรที่เสนอขำย (Abdullah Saif, 2015) สถำนที่ (Place) หมำยถึง สถำนที่ตั้งของร้ำนค้ำที่ผู้บริโภคสำมำรถเข้ำถึงได้โดยสะดวก ยังรวมไปถึงควำมสะดวกในกำรขนส่ง คลังสินค้ำ และกำรกระจำยสินค้ำ (Muchiri, 2016) และกำรส่งเสริมกำรตลำด (Promotion) หมำยถึง กระบวนกำรในกำรสื่อสำรกับผู้บริโภค เช่น กำรโฆษณำ ประชำสัมพันธ์ กำรขำยทำงตรง และรวมไปถึงกิจกรรมกำรส่งเสริมกำรขำย เพื่อกระตุ้นให้ผู้บริโภคเกิดกำรตัดสินใจซื้อ (Suherly, Affif, Arief & Guterres, 2016) แนวคิดและทฤษฎีที่เกี่ยวกับกำรตัดสินใจซื้อ หมำยถึง หมำยถึง ขั้นตอนในกำรเลือกซื้อผลิตภัณฑ์จำกสองทำงเลือกขึ้นไป โดยพิจำรณำในส่วนที่เกี่ยวของกับกระบวนกำรตัดสินใจ ทั้งด้ำนจิตใจและพฤติกรรมทำงกำยภำพ (Schiffman & Kanuk, 2007) ซึ่งกิจกรรมเหล่ำนี้ท ำให้เกิดกำรซื้อและเกิดพฤติกรรมกำรซื้อตำมบุคคลอื่นกระบวนกำรตัดสินใจซื้อมีองค์ประกอบส ำคัญจำกสิ่งกระตุ้นทำงกำรตลำด (Stimulus Marketing) ซึ่งนักกำรตลำดเป็นคนเสนอ ประกอบไปด้วย ด้ำนผลิตภัณฑ์ (Product) รำคำ (Price) สถำนที่ (Place) และกำรส่งเสริมกำรตลำด (Promotion) นอกจำกนี้ปัจจัยส่วนบุคคล - เพศ - อำยุ - อำชีพ - ระดับกำรศึกษำ - สถำนภำพสมรส - รำยได้เฉลี่ยต่อเดือน การตัดสินใจเลือกซื้อในส่วนประสมการตลาดของผลิตภัณฑ์เสริมอาหารเอ็มเฮิร์บ - ด้ำนผลิตภัณฑ์ - ด้ำนรำคำ - ด้ำนสถำนที่ - ด้ำนกำรส่งเสริมกำรตลำด การประชุมวิชาการระดับชาติ RTBEC 2020514ขั้นตอนในกำรตัดสินใจซื้อของผู้บริโภคจะมีล ำดับขั้นตอนในกำรตัดสินใจ 5 ขั้นตอน ได้แก่ กำรตระหนักรู้ถึงควำมต้องกำร กำรแสวงหำข้อมูล กำรประเมินทำงเลือก กำรตัดสินใจ และพฤติกรรมหลังกำรซื้อ (Kotler, 2003) งำนวิจัยที่เกี่ยวข้อง เยี่ยม เหอ (2558) ท ำกำรศึกษำเรื่อง ปัจจัยที่มีผลต่อกำรเลือกซื้อผลิตภัณฑ์อำหำรเสริมของบริษัทแอมเวย์ (ประเทศไทย) จ ำกัด ผลกำรศึกษำพบว่ำ ปัจจัยส่วนบุคคล ได้แก่ อำยุ ระดับกำรศึกษำ อำชีพ และรำยได้ที่แตกต่ำงกัน ส่งผลต่อกำรให้ควำมส ำคัญต่อปัจจัยส่วนประสมทำงกำรตลำดที่แตกต่ำงกัน อย่ำงมีนัยส ำคัญทำงสถิติที่ระดับ 0.05 สวงค์ เศวตวัฒนำ (2560) ท ำกำรศึกษำเรื่อง ปัจจัยกำรตลำดอำหำรเสริมผู้สูงอำยุในเขตพื้นที่จังหวัดปทุมธำนี ผลกำรศึกษำพบว่ำ ผู้ตอบแบบสอบถำมที่มี เพศ อำยุ ระดับกำรศึกษำ แตกต่ำงกัน มีระดับควำมส ำคัญส่วนประสมกำรตลำดอำหำรเสริมผู้สูงอำยุในเขตพื้นที่จังหวัดปทุมธำนีไม่แตกต่ำงกัน จินตนำ อ่อนลำ (2557) ท ำกำรศึกษำเรื่อง กลยุทธ์ทำงกำรตลำดที่มีอิทธิพลต่อกำรตัดสินใจซื้อผลิตภัณฑ์เสริมอำหำรของผู้บริโภคในเขตอ ำเภอวำรินช ำรำบ จังหวัดอุบลรำชธำนี ผลกำรศึกษำพบว่ำ ผู้ตอบแบบสอบถำมที่มี อำยุ ระดับกำรศึกษำ อำชีพ และรำยได้ต่อเดือนมีควำมคิดเห็นเกี่ยวกับกลยุทธ์กำรตลำดที่มีอิทธิพลต่อกำรตัดสินใจซื้อผลิตภัณฑ์เสริมอำหำรแตกต่ำงกันอย่ำงมีนัยส ำคัญทำงสถิติที่ระดับ 0.01 ในด้ำนผลิตภัณฑ์ ด้ำนกำรส่งเสริมกำรตลำด วิธีการด าเนินการวิจัย ประชำกรที่ใช้ในกำรวิจัย คือ ผู้บริโภคที่ซื้อผลิตภัณฑ์อำหำรเสริมเอ็มเฮิร์บในกรุงเทพมหำนคร ซึ่งไม่ทรำบจ ำนวนประชำกรที่แน่ชัด จึงก ำหนดขนำดกลุ่มตัวอย่ำงโดยใช้สูตรของ Cochran (1977) ได้จ ำนวนตัวอย่ำงทั้งสิ้น 400 คน โดยใช้วิธีกำรสุ่มตัวอย่ำงแบบตำมสะดวก (Convenience Sampling) เครื่องมือที่ใช้ในกำรวิจัย คือ แบบสอบถำมปลำยปิด ที่แบ่งออกเป็น 3 ส่วน คือ ส่วนที่ 1 ข้อมูลทั่วไปของผู้ตอบแบบสอบถำม ส่วนที่ 2 ข้อมูลพฤติกรรมกำรซื้อผลิตภัณฑ์เสริมอำหำร เอ็มเฮิร์บ สูตรหมอแสง ส่วนที่ 3 ข้อมูลกำรตัดสินใจเลือกซื้อในส่วนประสมกำรตลำดของผลิตภัณฑ์เสริมอำหำรเอ็มเฮิร์บ โดยมีควำมเชื่อมั่นอยู่ระหว่ำง 0.703 – 0.918 ซึ่งมำกกว่ำ 0.7 ขึ้นไปถือว่ำแบบสอบถำมมีควำมน่ำเชื่อถือ (กัลยำ วำนิชย์บัญชำ, 2550) กำรวิเครำะห์ข้อมูล โดยใช้กำรประมวลผลข้อมูลจำกโปรแกรมส ำเร็จรูปทำงสถิติ ได้แก่ สถิติเชิงพรรณนำ (Descriptive Statistics) ประกอบด้วย ค่ำควำมถี่ค่ำร้อยละ ค่ำเฉลี่ย และค่ำส่วนเบี่ยงเบนมำตรฐำน และสถิติเชิงอนุมำน (Inferential Statistics) ได้แก่ t-test และ F-test ผลการวิจัย ผลกำรวิเครำะห์ข้อมูลปัจจัยส่วนบุคคล พบว่ำ ผู้ตอบแบบสอบถำมส่วนใหญ่เป็นเพศหญิง จ ำนวน 228 คน (ร้อยละ 57.0) มีอำยุ 41 – 50 ปี จ ำนวน 152 คน (ร้อยละ 38.00) ประกอบอำชีพเป็นพนักงำนเอกชน จ ำนวน 109 คน (ร้อยละ 27.30) มีกำรศึกษำระดับปริญญำตรี จ ำนวน 228 คน การประชุมวิชาการระดับชาติ RTBEC 2020515(ร้อยละ57.00) มีสถำนภำพสมรส/ อยู่ด้วยกัน จ ำนวน 273 คน (ร้อยละ 68.30) และรำยได้เฉลี่ยต่อ 20,001-30,000 บำท มำกที่สุดจ ำนวน 185 คน (ร้อยละ 46.30) โดยพฤติกรรมกำรซื้อของผู้ตอบแบบสอบถำมส่วนใหญ่ซื้อ 1 ครั้ง/สัปดำห์ มำกที่สุดจ ำนวน 152 คน (ร้อยละ 38.00) ส่วนใหญ่ซื้อครั้งละ 1 กล่อง จ ำนวน 194 คน (ร้อยละ 48.50) ซึ่งซื้อครั้งละ 5,001 - 10,000 บำท จ ำนวน 179 คน (ร้อยละ 44.80) โดยกำรสั่งซื้อออนไลน์มำกที่สุด จ ำนวน 159 คน (ร้อยละ 39.80) ส่วนใหญ่ซื้อเพรำะบุคคลในครอบครัว/ญำติ มำกที่สุด จ ำนวน 146 คน (ร้อยละ 36.50) และ ส่วนใหญ่ซื้อเพื่อช่วยบ ำรุงสุขภำพ มำกที่สุด จ ำนวน 138 คน (ร้อยละ 34.50) ผลกำรวิเครำะห์ข้อมูลกำรตัดสินใจเลือกซื้อในส่วนประสมกำรตลำดของผลิตภัณฑ์เสริมอำหำรเอ็มเฮิร์บ พบว่ำ ผู้ตอบแบบสอบถำมส่วนใหญ่มีควำมคิดเห็นในภำพรวมอยู่ในระดับปำนกลำง (X̅ = 3.35, S.D. = 0.56) โดยเรียงล ำดับจำกมำกไปน้อยได้ดังนี้ ตัดสินใจเลือกซื้อในด้ำนกำรส่งเสริมทำงกำรตลำดมำกที่สุด (X̅ = 3.32, S.D. = 0.57) รองลงมำ ด้ำนสถำนที่ (X̅ = 3.17, S.D. = 0.45) ด้ำนผลิตภัณฑ์ (X̅ = 3.05, S.D. = 0.53) และน้อยที่สุด ด้ำนรำคำ (X̅ = 2.97, S.D. = 0.54) ตำมล ำดับ โดยสำมำรถจ ำแนกควำมคิดเห็นในแต่ละด้ำนดังต่อไปนี้ 1. ด้ำนผลิตภัณฑ์ พบว่ำ ผู้ตอบแบบสอบถำมมีระดับควำมคิดเห็นอยู่ในระดับปำนกลำง (X̅ = 3.05, S.D. = 0.53) โดยพิจำรณำเป็นรำยข้อ พบว่ำ ผลิตภัณฑ์ที่มีประสิทธิภำพใช้แล้วเห็นผล มีค่ำเฉลี่ยมำกที่สุด (X̅ = 3.46, S.D. = 0.89) รองลงมำ เป็นผลิตภัณฑ์ที่มีคุณภำพ (X̅ = 3.38, S.D. = 0.85) และน้อยที่สุด ผลิตภัณฑ์ระบุวันหมดอำยุที่ชัดเจน (X̅ = 2.49, S.D. = 1.08) 2. ด้ำนรำคำ พบว่ำ ผู้ตอบแบบสอบถำมมีระดับควำมคิดเห็นอยู่ในระดับปำนกลำง (X̅ = 2.97, S.D. = 0.54) โดยพิจำรณำเป็นรำยข้อ พบว่ำ กำรมีป้ำยรำคำติดไว้อย่ำงชัดเจน มีค่ำเฉลี่ยมำกที่สุด (X̅ = 3.60, S.D. = 0.98) รองลงมำ ผลิตภัณฑ์มีรำคำที่เหมำะสมกับปริมำณที่บรรจุ (X̅ = 3.45, S.D. = 1.12) และน้อยที่สุด สำมำรถช ำระเงินได้หลำกหลำยช่องทำง เช่น เคำน์เตอร์ธนำคำร เคำน์เตอร์บริกำรของห้ำงสรรพสินค้ำ เคำน์เตอร์เซเว่นอีเลฟเว่น ร้ำนค้ำ Nature Herbs ช ำระผ่ำนบัตรเครดิต/เดบิต เป็นต้น (X̅ = 2.42, S.D. = 0.79) 3. ด้ำนสถำนที่ พบว่ำ ผู้ตอบแบบสอบถำมมีระดับควำมคิดเห็นอยู่ในระดับปำนกลำง (X̅ = 3.17, S.D. = 0.45) โดยพิจำรณำเป็นรำยข้อ พบว่ำ มีช่องทำงสั่งซื้อหลำกหลำย เช่น ทำงเว็บไซต์ ไลน์ หน้ำบ้ำนหมอแสง เป็นต้น มีค่ำเฉลี่ยมำกที่สุด (X̅ = 4.19, S.D. = 0.84) รองลงมำ มีบริกำรจัดส่งผลิตภัณฑ์ถูกต้องและรวดเร็ว (X̅ = 3.13, S.D. = 1.18) และน้อยที่สุด สำมำรถสั่งซื้อได้ง่ำยและสะดวก (X̅ = 2.54, S.D. = 0.86) 4. ด้ำนกำรส่งเสริมกำรตลำด พบว่ำ ผู้ตอบแบบสอบถำมมีระดับควำมคิดเห็นอยู่ในระดับปำนกลำง (X̅ = 3.32, S.D. = 0.57) โดยพิจำรณำเป็นรำยข้อ พบว่ำ กำรจัดโปรโมชั่น เช่น ลดรำคำ มีของแถม เป็นต้น มีค่ำเฉลี่ยมำกที่สุด (X̅ = 4.23, S.D. = 0.92) รองลงมำ พนักงำนขำยมีควำมรู้และสำมำรถแนะน ำผลิตภัณฑ์ได้ดี (X̅ = 3.72, S.D. = 1.12) และน้อยที่สุด กำรประชำสัมพันธ์สินค้ำผ่ำนโซเชียลมีเดีย โดยอยู่ในระดับน้อย (X̅ = 2.47, S.D. = 1.02) การประชุมวิชาการระดับชาติ RTBEC 2020516ตารางที่ 1 แสดงกำรทดสอบสมมติฐำนปัจจัยส่วนบุคคล ปัจจัยส่วนบุคคล การตัดสินใจเลือกซื้อในส่วนประสมการตลาดของผลิตภัณฑ์เสริมอาหารเอ็มเฮิร์บ ผลิตภัณฑ์ ราคา สถานที่ การส่งเสริมการตลาด ภาพรวม - เพศ t = 1.03 Sig = 0.31 t = -0.71 Sig = 0.48 t = -0.47 Sig = 0.64 t = 1.04 Sig = 0.30 t = 0.37 Sig = 0.72 - อำยุ F = 1.95 Sig = 0.09 F = 2.23 Sig = 0.05* F = 1.32 Sig = 0.25 F = 4.38 Sig = 0.00* F = 4.21 Sig = 0.00* - อำชีพ F = 1.91 Sig = 0.09 F = 0.74 Sig = 0.59 F = 0.65 Sig = 0.66 F = 0.78 Sig = 0.56 F = 1.37 Sig = 0.24 - ระดับกำรศึกษำ F = 1.37 Sig = 0.26 F = 0.44 Sig = 0.65 F = 2.35 Sig = 0.10 F = 0.67 Sig = 0.51 F = 1.13 Sig = 0.33 - สถำนภำพสมรส F = 0.49 Sig = 0.61 F = 3.19 Sig = 0.04* F = 2.50 Sig = 0.08 F = 4.83 Sig = 0.01* F = 4.54 Sig = 0.01* - รำยได้เฉลี่ยต่อเดือน F = 0.99 Sig = 0.40 F = 1.21 Sig = 0.31 F = 2.18 Sig = 0.09 F = 9.60 Sig = 0.00* F = 4.88 Sig = 0.00* * มีนัยส ำคัญทำงสถิติที่ระดับ 0.05 จำกตำรำงที่ 1 พบว่ำ ปัจจัยส่วนบุคคลของผู้บริโภคในเขตกรุงเทพมหำนคร ได้แก่ อำยุ สถำนภำพสมรสและรำยได้เฉลี่ยต่อเดือนที่แตกต่ำงกันส่งผลต่อกำรตัดสินใจเลือกซื้อในส่วนประสมกำรตลำดของผลิตภัณฑ์เสริมอำหำรเอ็มเฮิร์บแตกต่ำงกัน อย่ำงมีนัยส ำคัญทำงสถิติที่ระดับ 0.05 1. ด้ำนอำยุนั้นจะส่งผลต่อกำรตัดสินใจเลือกซื้อในส่วนประสมกำรตลำดของผลิตภัณฑ์เสริมอำหำรเอ็มเฮิร์บแตกต่ำงกัน อย่ำงมีนัยส ำคัญทำงสถิติที่ระดับ 0.05 ในด้ำนของรำคำ และกำรส่งเสริมกำรตลำด 2. ด้ำนสถำนภำพสมรสนั้นจะส่งผลต่อกำรตัดสินใจเลือกซื้อในส่วนประสมกำรตลำดของผลิตภัณฑ์เสริมอำหำรเอ็มเฮิร์บแตกต่ำงกัน อย่ำงมีนัยส ำคัญทำงสถิติที่ระดับ 0.05 ในด้ำนของรำคำ และกำรส่งเสริมกำรตลำด 3. ด้ำนรำยได้เฉลี่ยต่อเดือนนั้นจะส่งผลต่อกำรตัดสินใจเลือกซื้อในส่วนประสมกำรตลำดของผลิตภัณฑ์เสริมอำหำรเอ็มเฮิร์บแตกต่ำงกัน อย่ำงมีนัยส ำคัญทำงสถิติที่ระดับ 0.05 ในด้ำนของกำรส่งเสริมกำรตลำด อภิปรายผล สมมติฐำนข้อที่ 1 ปัจจัยส่วนบุคคลของผู้บริโภคในเขตกรุงเทพมหำนครที่แตกต่ำงกันส่งผลต่อกำรตัดสินใจเลือกซื้อในส่วนประสมกำรตลำดของผลิตภัณฑ์เสริมอำหำรเอ็มเฮิร์บแตกต่ำงกัน พบว่ำ อำยุ สถำนภำพสมรสและรำยได้เฉลี่ยต่อเดือนที่แตกต่ำงกันส่งผลต่อกำรตัดสินใจเลือกซื้อในส่วนประสมกำรตลำดของผลิตภัณฑ์เสริมอำหำรเอ็มเฮิร์บแตกต่ำงกัน อย่ำงมีนัยส ำคัญทำงสถิติที่ระดับ 0.05 สอดคล้องกับสมมติฐำนที่ตั้งไว้ ซึ่งสำมำรถจ ำแนกอภิปรำยผลกำรตัดสินใจเลือกซื้อในส่วนประสมกำรตลำดของผลิตภัณฑ์เสริมอำหำรเอ็มเฮิร์บ ในแต่ละด้ำนที่แตกต่ำงกันดังนี้ การประชุมวิชาการระดับชาติ RTBEC 20205171. ด้ำนรำคำ เนื่องจำก โรคภัยไข้เจ็บสำมำรถเกิดขึ้นได้กับทุกบุคคล ทุกเพศทุกวัย ซึ่งผลิตภัณฑ์เสริมอำหำรเอ็มเฮิร์บ สูตรหมอแสงนั้นมีสรรพคุณช่วยท ำให้ร่ำงกำยแข็งแรงขึ้น อีกทั้งยังต่อต้ำนโรคมะเร็ง ซึ่งเป็นที่รู้จักของคนไทยอย่ำงมำก ซึ่งกำรวิจัยนี้พบว่ำ ผู้ที่มีอำยุ 41 – 50 ปี เป็นตัวอย่ำงที่ซื้อผลิตภัณฑ์เสริมอำหำรเอ็มเฮิร์บ สูตรหมอแสงมำกที่สุด ชี้ให้เห็นว่ำเมื่ออำยุสูงขึ้นจะท ำให้เกิดกำรดูแลและรักษำสุขภำพของตนเองมำกขึ้น และจะมีกำรสรรหำผลิตภัณฑ์เสริมอำหำรที่ช่วยท ำให้ร่ำงกำยแข็งแรงสมบูรณ์ นอกจำกนี้ผู้ที่มีสถำนภำพสมรส จะมีพฤติกรรมกำรดูแลตนเองมำยิ่งขึ้น เพื่อที่จะสำมำรถมีร่ำงกำรที่แข็งแรงและพร้อมที่จะท ำงำนหำเลี้ยงครอบครัว อีกทั้งยังช่วยดูแลซึ่งกันและกันระหว่ำงสำมีและภรรยำ ซึ่งมักจะพิจำรณำจำกรำคำที่มีกำรระบุไว้อย่ำงชัดเจน อีกทั้งมีรำคำที่เหมำะสมกำรปริมำณที่เสนอขำย 2. กำรส่งเสริมกำรตลำด พบว่ำ ผู้ที่มีรำยได้เฉลี่ยต่อ 20,001-30,000 บำท จะเป็นกลุ่มเป้ำหมำยหลักในกำรจ ำหน่ำยผลิตภัณฑ์เสริมอำหำรเอ็มเฮิร์บ สูตรหมอแสง เนื่องจำกมีรำคำค่อยข้ำงสูง เพื่อแลกกับกำรสมุนไพรที่ได้รับกำรยอมรับและผลกำรรับรองที่มีกำรกำรันตี ด้วยยอดขำยจ ำนวนมำกจึงให้ควำมใส่ใจในด้ำนกำรรับรู้ข่ำวสำรและข้อมูลจำกโฆษณำและกำรประชำสัมพันธ์ของผลิตภัณฑ์เสริมอำหำรเอ็มเฮิร์บ โดยเฉพำะกำรจัดโปรโมชั่น เช่น ลดรำคำ มีของแถม เป็นต้น ข้อค้นพบดังกล่ำว สอดคล้องกับงำนวิจัยของ เยี่ยม เหอ (2558) ท ำกำรศึกษำเรื่อง ปัจจัยที่มีผลต่อกำรเลือกซื้อผลิตภัณฑ์อำหำรเสริมของบริษัทแอมเวย์ (ประเทศไทย) จ ำกัด ผลกำรศึกษำพบว่ำ ปัจจัยส่วนบุคคล ได้แก่ อำยุ และรำยได้ที่แตกต่ำงกัน ส่งผลต่อกำรให้ควำมส ำคัญต่อปัจจัยส่วนประสมทำงกำรตลำดที่แตกต่ำงกัน อย่ำงมีนัยส ำคัญทำงสถิติที่ระดับ 0.05 และงำนวิจัยของ จินตนำ อ่อนลำ (2557) ท ำกำรศึกษำเรื่อง กลยุทธ์ทำงกำรตลำดที่มีอิทธิพลต่อกำรตัดสินใจซื้อผลิตภัณฑ์เสริมอำหำรของผู้บริโภคในเขตอ ำเภอวำรินช ำรำบ จังหวัดอุบลรำชธำนี ผลกำรศึกษำพบว่ำ ผู้ตอบแบบสอบถำมที่มี อำยุ และรำยได้ต่อเดือนมีควำมคิดเห็นเกี่ยวกับกลยุทธ์กำรตลำดที่มีอิทธิพลต่อกำรตัดสินใจซื้อผลิตภัณฑ์เสริมอำหำรแตกต่ำงกันอย่ำงมีนัยส ำคัญทำงสถิติที่ระดับ 0.01 ในด้ำนผลิตภัณฑ์ ด้ำนกำรส่งเสริมกำรตลำด แต่แตกต่ำงจำกงำนวิจัยของ สวงค์ เศวตวัฒนำ (2560) ท ำกำรศึกษำเรื่อง ปัจจัยกำรตลำดอำหำรเสริมผู้สูงอำยุในเขตพื้นที่จังหวัดปทุมธำนี ผลกำรศึกษำพบว่ำ ผู้ตอบแบบสอบถำมที่มี อำยุ ที่แตกต่ำงกัน มีระดับควำมส ำคัญส่วนประสมกำรตลำดอำหำรเสริมผู้สูงอำยุในเขตพื้นที่จังหวัดปทุมธำนีไม่แตกต่ำงกัน อภิปรำยผลเพิ่มเติมในส่วนของปัจจัยส่วนบุคคลที่ส่งผลกำรตัดสินใจเลือกซื้อในส่วนประสมกำรตลำดของผลิตภัณฑ์เสริมอำหำรเอ็มเฮิร์บไม่แตกต่ำงกัน ในด้ำนผลิตภัณฑ์และสถำนที่ เนื่องจำก ผู้ตอบแบบสอบถำมให้ควำมส ำคัญในสองด้ำนเหมือนกัน เพรำะผลิตภัณฑ์เป็นปัจจัยที่ส ำคัญในกำรค ำนึงกำรเลือกซื้อผลิตภัณฑ์ที่เสริมสร้ำงสุขภำพและควำมสมบูรณ์แข็งแรงให้กับร่ำงกำย ดังนั้นผลิตภัณฑ์ที่น ำเสนอต่อตลำดจะต้องมีประสิทธิภำพใช้แล้วเห็นผลด้วยคุณภำพจำกสำรสกัดสมุนไพรที่เป็นที่ยอมรับในสังคมไทยและกำรรับรองจำกกระทรวงสำธำรณสุขและงำนวิจัยต่ำง ๆ ท ำให้ผู้บริโภคเกิดควำมมั่นใจในเลือกใช้ผลิตภัณฑ์เสริมอำหำรเอ็มเฮิร์บ นอกจำกนี้สถำนที่จัดจ ำหน่ำยของผลิตภัณฑ์ สำมำรถเข้ำถึงได้โดยสะดวกและหลำกหลำยช่องทำง เพื่อท ำให้ผู้บริโภคสำมำรถหำซื้อได้ง่ำย นอกจำกนี้ทำงบริษัทเนเจอร์ เฮิร์บ อินเตอร์เนชั่นแนล โฮลดิ้ง จ ำกัด ยังมีกำรจัดจ ำหน่ำยผ่ำนหน้ำเพจ เว็บไซต์ และมีกระบวนกำรจัดส่งที่รวดเร็ว สอดคล้องกับงำนวิจัยของ สำยพิณ วิศัลยำงกูร และอิทธิกรข ำเดช (2556) ที่ท ำกำรศึกษำเรื่อง พฤติกรรมกำรเลือกซื้อผลิตภัณฑ์เพื่อสุขภำพของคนวัยท ำงำนในกรุงเทพมหำนคร ผลกำรศึกษำพบว่ำ ผู้ตอบแบบสอบถำมให้ควำมส ำคัญในส่วนประสมกำรตลำดในการประชุมวิชาการระดับชาติ RTBEC 2020518ด้ำนช่องทำงกำรจัดจ ำหน่ำยที่สะดวกในกำรหำซื้อ และผลิตภัณฑ์ที่มีคุณภำพ และสอดคล้องกับงำนวิจัยของ ภัทรำ รำดด ำรง และสุจิตรำ รอดสมบุญ (2559) ที่ท ำกำรศึกษำเรื่อง ปัจจัยที่ส่งผลต่อกำรตัดสินใจซื้อผลิตภัณฑ์อำหำรเสริมควำมงำมคอลลำเจนผ่ำนทำงเฟชบุ๊ค ผลกำรศึกษำพบว่ำ ผู้บริโภคให้ควำมส ำคัญที่สุดในด้ำนผลิตภัณฑ์ รองลงมำคือ ด้ำนรำคำ ช่องทำงกำรจัดจ ำหน่ำย และด้ำนกำรส่งเสริมกำรขำย ข้อเสนอแนะที่ได้รับจำกกำรวิจัย 1. ผู้ประกอบกำรและผู้ที่มีส่วนเกี่ยวข้องทำงด้ำนกำรตลำดของผลิตภัณฑ์เสริมอำหำรเอ็มเฮิร์บ ควรที่จะน ำผลกำรวิจัยนี้ไปช่วยในกำรวิเครำะห์และวำงแผนด ำเนินงำนทำงด้ำนกำรตลำด โดยเฉพำะกำรน ำเอำไปใช้ในกำรก ำหนดส่วนกำรตลำดที่สอดคล้องกับควำมต้องกำรของตลำดในกลุ่มผู้ที่เริ่มเข้ำสู่วัยสูงอำยุ เพื่อให้พวกเขำตระหนักถึงกำรดูแลรักษำตนเองให้ปลอดภัยและปรำศจำกโรคภัยไข้เจ็บ อีกทั้งส่งเสริมให้เกิดกำรดูแลซึ่งกันและกันในครอบครัวด้วยกำรเสนอแพ็กเกจทำงกำรตลำดที่สอดรับกับครอบครัวที่ต้องกำรดูแลสุขภำพ นอกจำกนี้ยังควรเอำข้อมูลทำงรำยได้ของผู้บริโภคโดยกำรน ำไปใช้ในกำรสื่อสำรและก ำหนดให้เป็นกลุ่มเป้ำหมำยหลักทำงกำรตลำดของบริษัทต่อไป 2. ผู้ประกอบกำรและผู้ที่มีส่วนเกี่ยวข้องทำงด้ำนกำรตลำดของผลิตภัณฑ์เสริมอำหำรเอ็มเฮิร์บ ควรที่จะน ำผลกำรวิจัยนี้ไปใช้ในกำรเสนอกลยุทธ์ทำงกำรตลำดที่เหมำะสมกับปัจจัยส่วนบุคคลที่แตกต่ำงกัน โดยเฉพำะในด้ำนรำคำ ที่ต้องมีกำรก ำหนดปริมำณสินค้ำให้เหมำะสมกับรำคำ เช่น จ ำหน่ำยสินค้ำในรำคำและปริมำณที่แตกต่ำง เพื่อเพิ่มทำงเลือกในกำรตัดสินใจซื้อให้กับผู้บริโภค อีกทั้งยังสำมำรถน ำข้อมูลควำมแตกต่ำงนี้มำใช้ในกำรวิเครำะห์และคำดกำรณ์กำรขำยและผลิตได้อย่ำงมีประสิทธิภำพ 3. ผู้ประกอบกำรและผู้ที่มีส่วนเกี่ยวข้องทำงด้ำนกำรตลำดของผลิตภัณฑ์เสริมอำหำรเอ็มเฮิร์บ ควรให้ควำมส ำคัญกับกลยุทธ์ทำงด้ำนส่งเสริมกำรขำย ให้ผู้บริโภคเกิดควำมคุ้มค่ำคุ้มรำคำกับต้นทุนที่ต้องใช้ในกำรช ำระต่อกำรซื้อต่อครั้ง นอกจำกนี้ยังควรมีกำรบริหำรและพัฒนำช่องทำงกำรขำยในแต่ละช่องทำงให้ผู้บริโภคสำมำรถเข้ำถึงสินค้ำได้ง่ำย โดยเฉพำะช่องทำงออนไลน์ที่ควรมีกำรประชำสัมพันธ์ทั้งทำงด้ำนข้อมูลข่ำวสำรและออกแบบให้ดึงดูดในกำรใช้งำนมำกยิ่งขึ้น ข้อเสนอแนะในกำรวิจัยครั้งต่อไป 1. กำรกำรศึกษำวิจัยครั้งต่อไป ควรมีกำรศึกษำตัวแปรอื่นๆ เพิ่มเติม เช่น กำรรับรู้ ค่ำนิยม ทัศนคติ ควำมเชื่อ ควำมกังวล ควำมตระหนักในกำรเลือกซื้อผลิตภัณฑ์เพื่อให้ทรำบปัจจัยที่มี ผลต่อกำรตัดสินใจองผู้บริโภคได้อย่ำงครบถ้วน 2. กำรกำรศึกษำวิจัยครั้งต่อไป ควรท ำกำรศึกษำกับกลุ่มประชำกร/กลุ่มตัวอย่ำงในสถำนที่แตกต่ำงกัน หรือกำรก ำหนดช่วงอำยุของกลุ่มประชำกร/กลุ่มตัวอย่ำง เพื่อเปรียบเทียบและตรวจสอบเจตคติต่อกำรตัดสินใจซื้อผลิตภัณฑ์ได้ชัดเจนมำกยิ่งขึ้น 3. กำรกำรศึกษำวิจัยครั้งต่อไป ควรศึกษำด้วยวิธีกำรวิจัยเชิงคุณภำพ โดยใช้กำรสัมภำษณ์เชิงลึก เพื่อให้ได้ข้อมูลที่มีควำมเฉพำะเจำะจงมำกยิ่งขึ้น เอกสารอ้างอิง กัลยำ วำนิชย์บัญชำ. (2550). การวิเคราะห์สถิติ: สถิติส าหรับบริหารและวิจัย พิมพ์ครั้งที่ 10.กรุงเทพฯ: โรงพิมพ์จุฬำลงกรณ์มหำวิทยำลัย. การประชุมวิชาการระดับชาติ RTBEC 2020519จินตนำ อ่อนลำ. (2557). กลยุทธ์ทางการตลาดที่มีอิทธิพลต่อการตัดสินใจซื้อผลิตภัณฑ์เสริมอาหารของผู้บริโภคในเขตอ าเภอวารินช าราบ จังหวัดอุบลราชธานี. (กำรศึกษำค้นคว้ำอิสระปริญญำบริหำรธุรกิจมหำบัณฑิต, มหำวิทยำลัยรำชภัฏอุบลรำชธำนี). ภัทรำ รำดด ำรง และสุจิตรำ รอดสมบุญ. (2559). ปัจจัยที่ส่งผลต่อกำรตัดสินใจซื้อผลิตภัณฑ์อำหำรเสริมควำมงำมคอลลำเจนผ่ำนทำงเฟชบุ๊ค. วารสารวิชาการตลาดและการจัดการ มหาวิทยาลัยเทคโนโลยีราชมงคลธัญบุรี, 3(2), 104-118. เยี่ยม เหอ. (2558). ปัจจัยที่มีผลต่อการเลือกซื้อผลิตภัณฑ์อาหารเสริมของบริษัทแอมเวย์ (ประเทศไทย) จ ากัด. (กำรศึกษำค้นคว้ำอิสระปริญญำบริหำรธุรกิจมหำบัณฑิต, มหำวิทยำลัยเกษมบัณฑิต). สยำมธุรกิจ. (2559). อาหารเสริมคึกคัก!! แบรนด์ไทย-เทศ แย่งส่วนแบ่งตลาด. สืบค้นจำก https://siamturakij.com/news/7868-อำหำรเสริมคึกคัก-แบรนด์ไทย-เทศ-แย่งส่วนแบ่งตลำด สวงค์ เศวตวัฒนำ. (2560). ปัจจัยกำรตลำดอำหำรเสริมผู้สูงอำยุในเขตพื้นที่จังหวัดปทุมธำนี. วารสารวิชาการ สถานบันเทคโนโลยีแห่งสุวรรณภูมิ, 3(1), 315-325. สำยพิณ วิศัลยำงกูร และอิทธิกรข ำเดช. (2556). พฤติกรรมกำรเลือกซื้อผลิตภัณฑ์เพื่อสุขภำพของคนวัยท ำงำนในกรุงเทพมหำนคร. วารสาร การเงิน การลงทุน การตลาด และการบริหารธุรกิจ, 1(2), 113-130 สถิติอุตสำหกรรมอำหำรเสริม. (2562). ภาพรวมอุตสาหกรรมอาหารเสริม. สืบค้นจำก http://www.tcels.or.th/Resources/Market-Analysis/1065?lang=th Abdullah Saif, N. M. (2015). How does marketing strategy influence firm performance? Implementation of marketing strategy for firm success. International Journal of Innovation and Economic Development, 1(3), 7-15. Cochran, W. G. (1977). Sampling Techniques (3rd ed.). New York: John wiley & Sons. Daniel, C. O. (2018). Effect of marketing strategies on organizational performance. International Journal of Business Marketing and Management, 3(9), 1-9. Gituma, M. M. (2017). Effects of marketing mix on sales performance: A case of Unga Feeds Limited. (Doctoral dissertation, United States International University-Africa). Ivy, J. (2008). A new higher education marketing mix: The 7Ps for MBA marketing. International Journal of Educational Management, 22(4), 288-299. Kotler, P. (2003). Marketing Management (11th ed.). Upper Saddle River, N.J.: Prentice Hall. Muchiri, M. N. (2016). The Effectiveness of Marketing Mix Strategies On Performance of Kenol Kobil Limited. (Doctoral dissertation, Thesis. University of Nairobi). Nature Herb. (2020). ข้อมูลเกี่ยวกับบริษัท Nature Herb. สืบค้นจำก https://natureherbinter.com/ การประชุมวิชาการระดับชาติ RTBEC 2020520Schiffman, L. G., & Kanuk, L. L. (2007). Consumer behavior (9th ed.). Englewood Cliffs, N.J.: Prentice Hall. Suherly, H., Affif, F., Arief, H., & Guterres, A. D. (2016). Marketing performance as the impact of marketing mix strategy (7P) with determination of marketing attraction and company’s resources. International Journal of Economics, Commerce and Management, 4(9), 569-587. TCDC. (2018). TREND2018: Health and Well-Being. สืบค้นจำก http://www.tcdc.or.th/articles/others/28293/#TREND2018-Health-and-Well-Being การประชุมวิชาการระดับชาติ RTBEC 2020521 การประชุมวิชาการระดับชาติ RTBEC 2020522View publication stats | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/346741073Investigating fractal dimension, heart rate variability, and memory duringdifferent image sequencing regimes in young adultsArticle in Chaos (Woodbury, N.Y.) · November 2020DOI: 10.1063/5.0002764CITATIONS2READS705 authors, including:Some of the authors of this publication are also working on these related projects:Depression and heart failure View projectVanessa ZarubinNorthwestern University8 PUBLICATIONS 17 CITATIONS SEE PROFILEKatherine R Mickley SteinmetzWofford College28 PUBLICATIONS 611 CITATIONS SEE PROFILECarolyn M MartsbergerLoyola University Chicago19 PUBLICATIONS 255 CITATIONS SEE PROFILEAll content following this page was uploaded by Carolyn M Martsberger on 04 August 2021.The user has requested enhancement of the downloaded file.Chaos 30, 113116 (2020); https://doi.org/10.1063/5.000276430, 113116© 2020 Author(s).Investigating fractal dimension, heart ratevariability, and memory during differentimage sequencing regimes in young adultsCite as: Chaos 30, 113116 (2020); https://doi.org/10.1063/5.0002764Submitted: 30 January 2020 . Accepted: 19 October 2020 . Published Online: 06 November 2020David Aguillard, Vanessa Zarubin, Caroline Wilson, Katherine R. Mickley Steinmetz, and CarolynMartsbergerARTICLES YOU MAY BE INTERESTED INReconstructing regime-dependent causal relationships from observational time seriesChaos: An Interdisciplinary Journal of Nonlinear Science 30, 113115 (2020); https://doi.org/10.1063/5.0020538Using mathematics to diagnose, cure, and predict cardiac arrhythmiaChaos: An Interdisciplinary Journal of Nonlinear Science 30, 113132 (2020); https://doi.org/10.1063/5.0021844Accuracy of neural networks for the simulation of chaotic dynamics: Precision of training datavs precision of the algorithmChaos: An Interdisciplinary Journal of Nonlinear Science 30, 113118 (2020); https://doi.org/10.1063/5.0021264ChaosARTICLEscitation.org/journal/chaInvestigating fractal dimension, heart ratevariability, and memory during different imagesequencing regimes in young adultsCite as: Chaos 30, 113116 (2020); doi: 10.1063/5.0002764Submitted: 30 January 2020 · Accepted: 19 October 2020 ·Published Online: 6 November 2020View OnlineExport CitationCrossMarkDavid Aguillard,1 Vanessa Zarubin,2Caroline Wilson,1 Katherine R. Mickley Steinmetz,2andCarolyn Martsberger1,a)AFFILIATIONS1Department of Physics, Wofford College, 429 N Church Street, Spartanburg, South Carolina 29303, USA2Department of Psychology, Wofford College, 429 N Church Street, Spartanburg, South Carolina 29303, USAa)Author to whom correspondence should be addressed: martsbergercm@wofford.eduABSTRACTThe goal of this study is to investigate patterns that emerge in brain and heart signals in response to external stimulating image regimes.Data were collected from 84 subjects of ages 18–22. Subjects viewed a series of both neutrally and negatively arousing pictures during 2-minand 18-s-long segments repeated nine times. Both brain [electroencephalogram (EEG)] and heart signals [electrocardiogram (EKG)] wererecorded for the duration of the study (ranging from 1.5 to 2.5 h) and analyzed using nonlinear techniques. Specifically, the fractal dimen-sion was computed from the EEG to determine how this voltage trace is related to the image sequencing. Our results showed that subjectsvisually stimulated by a series of mixed images (a randomized set of neutrally or negatively arousing images) had a significantly higher fractaldimension compared to subjects visually triggered by pure images (an organized set of either all neutral or all negatively arousing images).In addition, our results showed that subjects who performed better on memory recall had a higher fractal dimension computed from theEEG. Analysis of EKG also showed greater heart rate variability in subjects who viewed a series of mixed images compared to subjects visuallytriggered by pure images. Overall, our results show that the healthy brain and heart are responsive to environmental stimuli that promoteadaptability, flexibility, and agility.Published under license by AIP Publishing. https://doi.org/10.1063/5.0002764In this study, we capture simultaneous brain and heart data inhealthy subjects to investigate the connection between the under-lying nonlinear nature of voltage patterns to external imageregimes. For example, we found differences in the fractal dimen-sion (FD) of the electroencephalogram (EEG) when subjects werevisually stimulated by a series of categorized pictures. Specifically,subjects visually stimulated by a series of mixed images (a ran-domized set of neutrally or negatively arousing images) had a sig-nificantly higher fractal dimension compared to subjects visuallystimulated by pure images (an organized set of either all neutralor all negatively arousing images). Analysis of electrocardiogram(EKG) also showed greater heart rate variability (HRV) in sub-jects who viewed a series of mixed images compared to subjectsvisually triggered by pure images. In addition to these results,we found that subjects who had better memory recall had signifi-cantly higher fractal dimensions. These results may have medicalimplications in terms of how we might non-invasively investigatememory and physiological fitness. This is also an important resultto highlight the ubiquity of physics techniques in a variety ofrelevant fields.INTRODUCTIONScientists across fields have computed the fractal dimension(FD) of signals in a variety of systems in order to more fullyunderstand how nonlinearity plays a significant role in systembehavior (Goldberger and West, 1987; Katz, 1988; Kesic and Spa-sic, 2016; Stam, 2005; and Wright et al., 1993). FD measures thedegree of nontrivial self-similarity in a system, the persistenceof patterns, or characteristic features of the system’s morphol-ogy over multiple scales (Captur et al., 2017; Nayyeri, 2017; andRiley et al., 2012). In other words, FD quantifies the characteris-tic patterns viewed on a large scale that repeat on smaller scalesChaos 30, 113116 (2020); doi: 10.1063/5.000276430, 113116-1Published under license by AIP Publishing.ChaosARTICLEscitation.org/journal/cha(Arsac and Deschodt-Arsac, 2018). The fractal behavior of a sys-tem, classified by neither exclusively regular nor entirely randomfluctuations, has been shown to be correlated with better healthoutcomes (Iyengar et al., 1996). In many cases, the level of fractalscaling can be used to differentiate between healthy and pathologi-cal subjects (Golinska, 2012). Studies have reported that decreasedcomplexity or altered fractal dimensions reflect impaired adaptiveresponse to physiological stress as a result of aging or disease (Arsacand Deschodt-Arsac, 2018; Golinska 2012; Iyengar et al., 1996; andPeng et al., 1995). Typically, we associate irregular, complex sig-nals with normal physiology, while consistent patterns are indicativeof reduced function (Pincus and Goldberger, 1994). Furthermore,subjects with improved clinical outcomes display a certain range offractal behaviors, which are considered a signature of nonlinearity(Delignieres et al., 2006). These patterns are visible on both longand short time scales and suggest a degree of complexity that maybe indicative of human health; visualization and analysis of thesephysiological signals could play a role in predicting and recogniz-ing a disease (Nayyeri, 2017 and Pincus and Goldberger, 1994). Bylooking at the FDs in EEG data in this study, we aim to obtain infor-mation from standard voltage traces about their complexity underdifferent external image regimes and in relation to a subject’s overallmemory capabilities.In addition to FD, we investigate the influence of these exter-nal image regimes on heart rate variability in EKG data. Short-termheart rate variability is evident and quantifiable by numerous meth-ods (Piskorski and Guzik, 2007). Studies have shown that decreasesin heart rate variability may accompany adverse health states (Pincusand Goldberger, 1994), and healthy heart dynamics often demon-strate high frequency heart rate fluctuations (Iyengar et al., 1996).These beat-to-beat fluctuations shown in the healthy heart may berelated to properties of cardiac and neural systems controlling theiroutput (Peng et al., 1995). Because low heart rate variability has beenassociated with compromised health, while greater fluctuations havebeen linked to normal, healthy physiology, the study of fluctuationsin the heart rate holds promising information for clinical and predic-tive measures (Captur et al., 2017; Pincus and Goldberger, 1994; andShaffer and Ginsberg, 2017). Our work seeks to quantify changes inthe EKG to explore the connections between signal characteristicsand performance as an indicator of health.In this study, we use mathematical techniques such as frac-tal dimension for the brain (EEG) signals and heart rate variabilitymetrics for the heart (EKG) signals to calculate the degree of com-plexity, self-similarity, and variability of voltage signals acquired.We find FD in the brain and short-term heart rate variability met-rics in the heart to investigate how these parameters correlate withvisual stimuli and a subject’s recall capability. We hypothesize thathighly complex signals in the brain and greater variability of signalsin the heart may be correlated with the external image regime andindicative of better overall performance on memory recall.METHODParticipants84undergraduatestudentsfromtheWoffordCollege(58 female) between the ages of 18 and 22 (M = 19.98, SE = 0.125)participated in this study for either course credit or a $20 Amazongift card. All participants provided written informed consent, andprocedures were approved by the Wofford College InstitutionalReview Board. Six participants were excluded from analysis onthe basis of prior exposure to the images (n = 2), left-handedness(n = 1), connection failure (n = 1), and neurological disorder(n = 2). An additional 11 participants were excluded from the analy-sis due to poor signal capture. Three more participants were omittedfrom the heart rate analysis due to a low-quality EKG. This resultedin a sample of 67 participants (47 female) included in the brain anal-yses and 64 participants (44 female) included in the heart analyses.All participants were right-handed. Each participant was assignedto either a pure image regime (n = 32 for the brain; n = 31 for theheart) or a mixed image regime (n = 35 for the brain; n = 33 forthe heart). Pure and mixed image regimes are described below. Par-ticipants were not currently abusing drugs, did not receive generalanesthesia in the two weeks prior to testing, and did not sustain aconcussion in the month prior to testing.Study materials: Image regimes explainedIn this study, two image regimes (pure and mixed) were cre-ated for subjects to view. To create these two image regimes, thefollowing protocol was implemented. 66 negative, 66 related neutral(we will refer to as categorical), and 66 unrelated neutral (we willrefer to as neutral) images resized to 500 by 400 pixels were selectedfrom a pool of 100 unrelated neutral images, 150 related neutralimages, and 150 negative images sourced from the InternationalAffective Picture System (Lang et al., 1995), the Geneva AffectivePicture Database (Dan-Glauser and Scherer, 2011), the EmotionalPicture Set (Wessa et al., 2010), the image pool of Talmi et al. (2007),and Google Images. Participants rated images on arousal (perceivedphysiological state), valence (how pleasant or unpleasant the imagemade the participant feel), and relatedness (how semantically asso-ciated with other pictures a particular image was) in pilot studies.Participants rated both arousal and valence on nine-point scales.For arousal, the scale was graded from “calm/soothing” to “excit-ing/agitating.” For valence, the scale ranged from “very unpleasing”to “very pleasing.” Participants rated images’ relatedness on a seven-point scale from “low association” to “high association.” Participantswere given examples of categorically related images (a handgun anda rifle; walking and running) as well as thematically related images(an umbrella and clouds) prior to rating. To ensure that image cat-egories did not differ in brightness or complexity, the luminosityrating was also calculated for each image using Photoshop, andcomplexity ratings were taken from three independent raters.Subjects were randomly assigned to either a pure image regimeor a mixed image regime. In the pure image regime, subjects viewedimages that were all neutral, all negative, or all categorical for atotal of nine segments. Each segment contained 22 images fromone of the three groups (i.e., negative, neutral, or categorical). Thenine segments were always ordered in the following way: negative(segment 1), categorical (segment 2), neutral (segment 3), categori-cal (segment 4), neutral (segment 5), negative (segment 6), neutral(segment 7), negative (segment 8), and categorical (segment 9).For the mixed image regime, subjects viewed nine segments of amix of neutral, negative, and categorical images for every segment.For example, in the mixed list for segment 1, the 22 images wereChaos 30, 113116 (2020); doi: 10.1063/5.000276430, 113116-2Published under license by AIP Publishing.ChaosARTICLEscitation.org/journal/chaTABLE I. Means and standard errors for pure image selections.Negative meanNegativestandard errorCategorical meanCategoricalstandard errorNeutral meanNeutralstandard errorArousal7.2120.0754.8720.0834.8380.068Valence2.7110.0745.1390.0865.0900.058Relatedness3.6260.0853.5890.0481.9810.017Luminosity106.1053.048106.3531.304107.5163.030Complexity3.0250.0853.0620.1163.0250.134a random sampling of negative, neutral, and categorical images.Overall, image regimes were balanced such that negative imageswere rated significantly lower on valence and significantly higher onarousal than either categorical or neutral, which did not significantlydiffer from each other on either measure (Tables I and II). Both neg-ative and categorical images were rated as significantly more relatedthan neutral images but did not differ from each other in relatedness(Tables I and II). All image types were matched on both complexityand luminosity (Tables I and II). An additional categorical list andan additional mixed list were created to be used as the practice testfrom unselected images rated in the initial pool. The practice testgave participants an opportunity to try the testing procedure beforerecordings started.Participants reported their depression and anxiety levels usingstandardized questionnaires including a 19-item version of the BeckDepression Inventory (BDI-I; Beck et al., 1961) and a 21-item ver-sion of the Beck Anxiety Inventory (BAI; Beck et al., 1988), both ofwhich have been demonstrated as reliable and concise measures ofdepression and anxiety levels.ProcedureAfter participants gave consent to participate in the study,aCortechActiveTwo32ChannelEEGSystem(Manufac-turer: Cortech Solutions; Model Specifications: Model NumberDA-AT_HCL32) from the Behavioral Brain Sciences Center, Birm-ingham, UK, was applied to the participant’s scalp while he or shecompleted the depression and anxiety questionnaires (BDI-I andBAI) and the practice test.The experiment consisted of nine segments. In each segment,participants first viewed 22 images. Images were presented inTABLE II. t-test image-balancing comparisons between selected image pools.Significant differences denoted by <0.001. As intended, negative images were morearousing and negatively valenced than either neutral groups but did not differ in lumi-nosity or complexity. In addition, negative and categorical images were more relatedthan neutral images.Neutral tocategoricalNegative toneutralNegative tocategoricalArousal0.752<0.001<0.001Valence0.634<0.001<0.001Relatedness<0.001<0.0010.706Luminosity0.7240.7430.940Complexity0.8380.9990.801E-Prime 3.0 (Psychology Software Tools, Pittsburgh, PA). Eachimage was presented for 2 s followed by a fixation cross (solid whitescreen with a central black plus sign symbol) that was presentedfor 4 s. The fixation cross between each image was used to reduceany remaining emotional responses carrying over from the previousimage (Talmi and McGarry, 2012). After viewing all 22 images, par-ticipants completed a 1-min arithmetic task, which prevented themfrom having an increased memory for the images that appeared atthe end of the list. Following the arithmetic task, participants wereasked to recall as many images as they could for up to 3 min.Data processing and reductionFor the EEG recording, sites were referenced online to twomastoid electrodes and re-referenced offline using the commonaverage reference. Electrode offsets were between 0 and ±30 mV.Signals were amplified, bandpass filtered (0.03–30 Hz), and dig-itized at 1024 Hz. Two EOG electrodes were used to track eyemovements. Three EKG electrodes also recorded heart rate data dur-ing the experiment. All data were cleaned to remove any artifactsthat impeded the voltage signals. Blinks were removed using theElectromagnetic Source Estimation (EMSE) manual artifact tool.Analyses were performed based on image regimes (pure vsmixed). Fractal dimensions were computed in EEG, and heart ratevariability metrics were computed in EKG to determine the impactof the two image lists (pure vs mixed) on the acquired signals. Com-plexity in EEG and variability in the heart were also analyzed in con-nection with overall memory performance. To investigate EEG andEKG patterns from subjects who viewed the mixed image regimevs the pure image regime, we ran a standard independent samplet-test in SPSS Statistics software to compare the mean from thesetwo groups. All reports used Levene’s Test for Equality of Variances(with equal variances not assumed) to determine p-value statistics.We also computed paired sample t-tests within each subject dur-ing testing to analyze drift and variation throughout the course ofthe study. Furthermore, we looked at the difference in each imageregime to see if patterns emerged within each list type (i.e., analysisentirely within the pure image regime). We also computed fractaldimensions and heart rate variability in correlation with subjects’memory scores (MSs). The level of significance was set at p = 0.05,although in some cases, marginal significance of 0.05 < p < 0.06 isnoted.Analyzing voltage dataData were analyzed using robust methods to determineself-similarity and variability. Voltage signals acquired from theChaos 30, 113116 (2020); doi: 10.1063/5.000276430, 113116-3Published under license by AIP Publishing.ChaosARTICLEscitation.org/journal/chascalp (i.e., EEG signals) were analyzed using the Higuchi methodto determine FD (Higuchi, 1988). Data acquired from the EKG wereanalyzed using short-term heart rate metrics. These were obtainedby analyzing the time between two consecutive R-waves of the EKGknown as the RR interval. We used the RR interval to determinethe percentage of sequential RR intervals that differ by more than50 ms (pNN50), the standard deviation (SD) of the Poincaré plotperpendicular to the line of identity (SD1) and along the line ofidentity (SD2) and the standard deviation of normal to normal inter-vals (SDNN) (Shaffer and Ginsberg, 2017). The results are discussedbelow.Computational methods for the brainTheHiguchimethodwasusedtodetermineamet-ric of complexity for the brain. This method of calculat-ing fractal dimensions determines the length of the curvefor a collection of a time series. The time series is definedas Xmk ; X(m), X(m + k), X(m + 2k), . . . , X�m + N−mk· k�, (m = 1,2, . . . , k), in which m and k are integers representing the initialtime and the interval time, respectively, and N represents the totalnumber of data points. The length of the curve Xmk is then given byLm(k) =���� N−mk�i=1��X(m + ik) − X(m + (i − 1) · k)���N−1N−mk·k�/k.For the time interval k, the curve length ⟨L(k)⟩ represents the aver-age value over k sets of Lm(k). A value for ⟨L(k)⟩ proportionate tok−D indicates a fractal with dimension D (Higuchi, 1988).In order to use the Higuchi algorithm to calculate FD, weneeded to determine a maximum value for the algorithm’s depen-dent parameter k, which we define as Kmax. The parameter k refersto the time interval between time series observations and representsthe number of distinct sets of time series obtained from various ini-tial times. The model was optimized via a computational experimentto determine a Kmax value that would be used for all electrodes. Inorder to determine Kmax, six electrodes were chosen from equallyspaced regions within the electrode grid to give the best visual of theexpanse of the brain. Specifically, electrodes Fp1, T7, Oz, T8, Fp2,and Cz were chosen from the brain (see Fig. 1). Data from thesesix electrodes were used to determine the optimal Kmax to run theHiguchi algorithm so that the fractal dimension plateaued and thecomputational run time was reasonable. A Kmax of 58 was deter-mined to be the optimal value to achieve both goals. To confirm thisresult, we also computed the peak of the sum of differences of thefractal dimension as a function of Kmax (from 1 to 100) in every sub-ject for the six chosen electrodes. Five out of six electrodes had theirpeak at Kmax = 58.After determining a Kmax value of 58 computationally, wedivided the raw EEG from all 32 electrodes of each participant intonine segments, which corresponded to the nine periods when thesubject viewed the images. Each segment lasted 2 min and 18 s andincludes data from 3 s before the subject viewed the first image until2 s after the subject viewed the final image. The nine segments ofEEG data were then run through the Higuchi algorithm for each ofthe 32 electrodes individually so that every participant had 32 FDvalues associated with each segment, one value for each electrode onthe cap. The arithmetic mean of the 32 FD values was also computedFIG. 1. Map of 32 electrodes highlighting the 6 electrodes from which Kmax wascomputed.to find the average fractal dimension (AFD) per participant for eachsegment.To assign a memory score, we analyzed each subject’s per-formance in memory after each segment by asking participants todescribe every image they could remember from that segment. Thememory score was measured by the percentage of images in eachgroup correctly identified as remembered. Subjects were also giventhree separate scores from 0 to 100 for their ability to recall negative,neutral, and categorical images. The subjects’ recall was highest onnegative images, which is consistent with the literature (Talmi et al.,2007; Kensinger, 2009; Talmi and McGarry, 2012; and Schmidt andSaari, 2007).In addition to calculating FDs for each of the nine individ-ual segments, for participants who viewed pure lists that were onlycategorical, negative, or neutral for the entire duration of each seg-ment, we also computed the average FD and the FDs of the 32electrodes across the three categorical, three neutral, and three neg-ative segments to analyze each segment type (i.e., negative, neutral,and categorical). This gave us a single AFD and an electrode-basedfractal dimension associated with each of the image types (negative,neutral, and categorical) for each participant viewing the pure imageregime. We compared the calculated FD values of each image type(negative, neutral, and categorical) to corresponding memory scoresfor each image type. Similarly, for each of the mixed image regimes,we computed the AFD and each electrode’s FD across the nineindividual segments as well as across the triplet of segment type toevaluate FD values in relation to memory scores and in comparisonwith the corresponding pure segment types.Chaos 30, 113116 (2020); doi: 10.1063/5.000276430, 113116-4Published under license by AIP Publishing.ChaosARTICLEscitation.org/journal/chaAnalyzing the heart rate dataThe participants’ EKG data recorded during each of the ninesegments were split into 6-s intervals. Each 6-s interval com-menced when the subject began viewing a given image and endedbefore the following image was shown. These segments werenormalized and filtered through MATLAB’s Savitzky-Golay filter(Sgolay, MathWorks).We used MATLAB’s “findpeaks” algorithm to identify each Rwave peak to extract RR interval data. After initial cleaning, the 6-ssegments were stitched back together, and two independent researchassistants manually checked the data for false and missed RR intervalassignments. These were corrected manually before the voltage datawere analyzed.RR intervals were analyzed for SDNN, pNN50, SD1, andSD2. The standard deviation of normal-to-normal intervals (SDNN)(normal referring to heartbeats originating only in the sinoatrialnode rather than being induced ectopically, or elsewhere) and thepercentage of successive RR intervals that differ by more than 50 ms(pNN50) are measurements that seek to quantitatively address thevariability present in time intervals between heartbeats. SD1 andSD2 are nonlinear measures of heart rate variability that apply stan-dard deviations of points from the Poincaré scatterplot to visualizepatterns and trends of time series.FINDINGS AND ANALYSISBrain results: Fractal dimension in EEG and imageregimesFractal dimensions were computed for each electrode (32 total)from the expanse of the brain. We also report the fractal dimen-sion averages from all 32 electrodes (AFD). Fractal dimensions arepresented for each segment type (negative, neutral, and categorical),over the entire nine segments, and segment-by-segment (i.e., 1–9).Subjects viewing mixed image regimes had overall higher FDswhen compared to the FDs of subjects who viewed pure imageregimes. We created brain maps to identify the electrodes thatexhibited significant behavior between the mixed image regime andthe pure image regime. The brain maps are presented in Fig. 2.Specifically, the fractal dimensions were higher (p < 0.05) in sub-jects viewing the mixed vs the pure lists. For example, for the entireset of all nine segments, 16 of the 32 electrodes displayed signifi-cance as highlighted in Fig. 2(a). This result persisted for the averagecomputed over segment types. To calculate the averages over thesegment type, we implemented the following procedure. First, weaveraged the FD per electrode of all three neutral segments (puresegments 3, 5, and 7) from the pure subjects and compared the resultto the FD averages of the corresponding mixed list segment numbers(mixed segments 3, 5, and 7) of the mixed subjects. We found that 20out of 32 electrodes had a statistically significant higher FD for themixed list compared to the pure neutral list [Fig. 2(b)]. Similarly,comparing the three categorical segments (pure segments 2, 4, and9) from the pure subjects to the corresponding mixed list segmentnumbers (mixed segments 2, 4, and 9) reveals that 14 out of 32 elec-trodes were significant during the categorical segments [Fig. 2(c)].Finally, comparing the three negative segments (pure segments 1, 6,and 8) to the corresponding mixed list segments (mixed segments1, 6, and 8), we found that 10 out of 32 electrodes were significant[Fig. 2(d)]. In each of the cases reported, fractal dimensions werehigher in subjects stimulated by mixed image lists as opposed toimage lists organized according to the image type (categorical, nega-tive, or neutral). The AFD for this result is reported in Fig. 3. Overall,higher FD corresponded to the mixed image regime. Since subjectswho viewed the mixed image regime had an overall higher FD com-pared to subjects who viewed the pure image regime, we hypothesizethat an alternating external image regime may influence EEG. Thesedata may suggest that mixing images that evoke a range of emotionalresponses caused a constant switching in the brain that contributedto a more complex waveform, whereas the methodical nature of thepure lists reduced the complexity of the corresponding brain sig-nals, possibly due to habituation of the type of image the subjectviewed throughout the sequence. A mixed image regime seems toillicit a more malleable, dynamic response in subjects compared tothe consistent sequence resulting from the pure regime.Fractal dimension based on segment number (i.e., 1–9) wasalso computed as a function of the pure or mixed image regime.Specifically, the fractal dimension was averaged over all electrodesper subject during each of the nine segments. The average frac-tal dimension per segment was compared between image regimes(mixed and pure). In comparing each of the segments between thepure and mixed subject assignments, six out of nine segments werestatistically significant (Fig. 4). In particular, FD values from themixed group were significantly higher in the first six segments oftesting (i.e., the first sequential two–thirds of the testing period) cor-responding to the negative, categorical, neutral, categorical, neutral,and negative segments. The last three segments (neutral, negative,and categorical) did not exhibit significance in FD. Comparing thepure vs the mixed groups reveals that participants who viewed mixedlists had a statistically significant higher fractal dimension com-pared to the subjects who viewed pure lists (categorical, negative,or neutral) during the first six segments of testing. This is consistentwith our first reported finding that the mixed image regime showedsignificantly higher FDs than the pure image regime.In summary, subjects stimulated by mixed image lists hadhigher FDs when compared to subjects stimulated by pure imagelists organized according to image type (categorical, negative, orneutral). This suggests that the FD may be influenced by a visu-ally changing external image regime, and variable external patternsof stimulation contribute to activating the brain differently com-pared to an external image regime in which the same thematicgroup of images occurs. This may suggest that an alternating imageregime can evoke a different response in the resulting voltage patterncompared to a constant image regime, which may be applicable todeveloping a form of clinical stress testing to investigate brain fitnessor to crafting a therapeutic exercise, based on a switching stimuluspattern.Brain results: Fractal dimension in EEG and memoryOur finding that the fractal dimensions were higher in themixed list case vs the pure list case suggests that the image regimepattern evoked a response in subjects who scaled with complex-ity. We also investigated the fractal dimension of subjects withtheir respective memory scores in a similar manner. We foundChaos 30, 113116 (2020); doi: 10.1063/5.000276430, 113116-5Published under license by AIP Publishing.ChaosARTICLEscitation.org/journal/chaFIG. 2. Electrode maps illustrating significant fractal dimensions for different segment types and all segments. Highlighted electrodes in (a) illustrates the location of electrodeswith higher FD for the mixed image regime compared to the pure image regime for all segments. (b) shows the electrodes with higher FD for the mixed image regime comparedto the pure image regime for neutral segments. (c) shows the electrodes with higher FD for the mixed image regime compared to the pure image regime for categoricalsegments. (d) shows the electrodes with higher FD for the mixed image regime compared to the pure image regime for the negative segments.Chaos 30, 113116 (2020); doi: 10.1063/5.000276430, 113116-6Published under license by AIP Publishing.ChaosARTICLEscitation.org/journal/chaFIG. 3. Average fractal dimension (AFD) for the mixed image regime (blue) and the pure image regime (orange). Results are presented for the segment types [negative(p-value = 0.079), neutral (p-value = 0.007), and categorical (p-value = 0.027)] as well as for all segments [all (p-value = 0.024)]. Our results show that the mixed imageregime has a significantly higher FD (denoted by *) compared to the pure image regime.that for subjects viewing pure images only, there was a correlationbetween their memory score and their fractal dimensions. Specifi-cally, subjects who had higher memory scores also had higher fractaldimensions. We found this result under the following conditions.For subjects who viewed negative images, a higher memory scorecorrelated with higher fractal dimensions in 21 of the 32 electrodes(with 5 of the 21 electrodes displaying marginal significance). Weillustrate this finding in the brain map [Fig. 5(a)] and in Fig. 6 forthe AFD. Similar results for the categorical data showed a correla-tion between higher fractal dimensions and higher memory score in18 out of the 32 electrodes, displayed in Figs. 5(b) and 6 for the AFD.This supports findings emphasizing the role of semantic related-ness in memory processes, favoring categorical stimuli over random(Talmi and Moscovitch, 2004). For the case of neutral segments, wefound that 4 out of the 32 electrodes were significant. This finding isillustrated in Figs. 5(c) and 6 for the AFD. Overall, these data suggestFIG. 4. Average fractal dimension (AFD) per segment for the mixed image regime (blue) and the pure image regime (orange). Our results show that the mixed image regimehas significantly higher FDs (denoted with *) for the first six segments compared to the pure image regime.Chaos 30, 113116 (2020); doi: 10.1063/5.000276430, 113116-7Published under license by AIP Publishing.ChaosARTICLEscitation.org/journal/chaFIG. 5.Electrode maps illustrating fractaldimensions by memory. Higher fractal dimen-sions were correlated with better memoryscores in highlighted electrodes. (a) illustrateshigher fractal dimension with better mem-ory scores for subjects viewing the negativepure image regime. (b) illustrates higher frac-tal dimension with better memory scores forsubjects viewing the categorical pure imageregime. (c) illustrates higher fractal dimensionwith better memory scores for subjects view-ing the neutral pure image regime. (d) illus-trates that there was no correlation betweenFD and better memory for the subjects whoviewed the mixed image regime. (e) illus-trates that there was a correlation betweenFD and better memory for the subjects whoviewed all nine segments of the pure imageregime. (f) illustrates that there was no cor-relation between FD and better memory forall subjects viewing the mixed or pure imageregimes.Chaos 30, 113116 (2020); doi: 10.1063/5.000276430, 113116-8Published under license by AIP Publishing.ChaosARTICLEscitation.org/journal/chaFIG. 6. Average fractal dimension vs memory. FD as it relates to memory during the pure image regime only is presented in the first three pure results [pure negative(p-value = 0.014), pure categorical (p-value = 0.023), and pure neutral (p-value = 0.172)]. Subjects in the pure image regime with above average memory had higherFDs. FD as it relates to memory for the entire pure regime only [pure (p-value = 0.040)], the entire mixed regime only [mixed (p-value = 0.602)], and the entire data set[mixed + pure (p-value = 0.521)] is also presented. * denotes significance.that greater complexity of a subject’s brain data is correlated witha positive memory outcome, which may provide a foundation forfuture research possibilities.For subjects viewing entirely mixed lists, there was no cor-relation between memory and fractal dimensions. This result isillustrated in Figs. 5(d) and 6 for AFD. Similarly, the data over allpure and mixed segments together demonstrated no clear evidencethat memory was correlated with fractal dimension [Figs. 5(f) and 6].However, for subjects viewing entirely pure lists over all nine seg-ments, we found that 12 out of the 32 electrodes had higher fractaldimensions as the memory score increased. This result is illustratedin Figs. 5(e) and 6 for the AFD. Because all pure segments, whetheraveraged over all nine segments [Fig. 5(e)] or binned either accord-ing to image type [Figs. 5(a)–5(c)], show a correlation between bettermemory scores and higher fractal dimensions, our findings sug-gest that heightened memory recall may be correlated with a morecomplex neurological voltage pattern. On the contrary, all mixedsegments and all pure and mixed segments together do not showsignificance in fractal dimensions with memory, suggesting that theeffect can be washed out by an experience that mitigates recall.The result in this section is supported by our analysis of mem-ory scores. Specifically, there was no significant difference in FDsbetween subjects in the mixed image regime compared to thosein the pure image regime for their corresponding memory scores(MSs) [MS(mixed) = 44.36 vs MS(pure) = 46.21, p = 0.542]. Like-wise, we do not find statistically different MSs when comparingwithin subjects based on list types [negative (MS(mixed) = 52.04vs MS(pure) = 55.36, p = 0.272], neutral [MS(mixed) = 35.28 vsMS(pure) = 37.76, p = 0.474], and categorical [MS(mixed) = 45.47vs MS(pure) = 46.01, p = 0.884]. Because there was no differencebetween memory scores of the mixed and pure lists, we would notexpect to see correlation between fractal dimension and memorywhen we look at the results from subjects in the pure and mixedimage regimes all together. Because the mixed regime-only analysisshows no correlation, it is likely that the mixed image regime resultswash out the effects of the correlation between FD and memorybecause the mixed image regime promotes memory recall differ-ently than the pure image regime. Previous studies support thatrecall is diminished for neutral items in mixed compared to pure lists(Watts et al., 2014 and Barnacle et al., 2018). However, for our study,within subjects exposed to the pure image regime only, there is sta-tistical significance (p < 0.000 001) in MS for all inter-segment typematches (neutral vs negative, negative vs categorical, and categori-cal vs neutral), with the following MS values: MS(negative) = 55.36,MS(category) 46.01, and MS(neutral) = 37.77. This result suggeststhat subjects remember negative images best, followed by categor-ical images and then neutral images. This finding is confirmed byother studies (Talmi and McGarry, 2012 and Talmi, 2013). Viewingthe corresponding FD brain maps in Fig. 5, this result is reaffirmedin our findings of fractal dimensions. For Fig. 5(a), the negativesegments that correspond to the highest memory recall in subjectsalso correspond to the largest number of electrodes (21 electrodes)that are significant in the correlation between memory and fractaldimension. The categorical segments also illustrate this correlationChaos 30, 113116 (2020); doi: 10.1063/5.000276430, 113116-9Published under license by AIP Publishing.ChaosARTICLEscitation.org/journal/chabut not in as many electrodes (18 electrodes) [Fig. 5(b)], andthese segments correlate with the mid-range memory scores fromsubjects. Finally, the neutral segments illustrate the least number ofelectrodes (four electrodes) that show correlation with memory andfractal dimensions [Fig. 5(c)], which is consistent with the fact thatthey also represent the set of images with the lowest memory recallin subjects. Overall, the data suggest that higher memory scoresduring the pure image regime are related to higher fractal dimen-sions, which may have implications for the connection betweenbrain complexity and overall memory capabilities.To protect from family wise errors when making multiplehypotheses, results were presented for the entire cohort (representedabove as the AFD). In addition, results from a Bonferroni correctionare listed below for the 32 electrode cap, segment type, and segmentnumber results. For the 32-electrode cap for the pure vs mixed imageregimes, electrodes P7, PO3, and P8 had a p-value <1.5625 * 10−3(P7 p-value = 5.9023 * 10−4, PO3 p-value = 1.527 39 * 10−3, and P8p-value = 1.340 24 * 10−3). Similarly, for the 32-electrode cap for theabove average memory vs the below average memory pure negativegroup, electrodes CP5, PO3, and Pz had a p-value <1.5625 * 10−3.For the results of pure vs mixed image regimes on segment num-ber, segment 2 had a p-value < 5.555 56 * 10−3. For the pure vs mixedbased on the segment type, the pure neutral vs mixed had a p-valueof 0.007, satisfying the Bonferroni correction as well. A full repre-sentation of all of our data and the corresponding p-values can befound in the supplementary material.HEART RESULTSHeart rate metrics, as recommended by the most recentoverview of heart rate variability metrics and norms (Shaffer andGinsberg, 2017), were also compared to list type, segment num-ber, and memory scores. We computed the standard deviation andmean for the RR intervals of each subject as well as the followingfour short-term heart rate variability metrics: pNN50, SD1, SD2, andSDNN (Shaffer and Ginsberg, 2017). Together, these time-domainmeasurements were also investigated in relation to image regimesand memory. We also looked at how these variables changed overthe course of the study within a subject.Heart results: Heart rate metrics and image regimesWe found significant differences in heart rate metrics whencomparing findings between subjects exposed to the pure vs mixedimage regimes. These results are similar to our findings in the brainwith respect to fractal dimension. Specifically, we found greaterheart rate variability in subjects who viewed the mixed image regimecompared to subjects who viewed the pure image regime, simi-lar to our findings in EEG. For example, in looking at the overallspread in RR intervals, we found that the standard deviation ofthe RR intervals was higher during the mixed image regime com-pared to the pure image regime. The standard deviation (SD) ofthe RR intervals for mixed image regime subjects is SD = 90.60 msand that of the pure image regime subjects is SD = 63.89 ms(p = 0.056). The ratio of the standard deviation of RR intervals tothe mean RR intervals (SDM) was also higher during the mixedlist (SDM = 0.110 ms) compared to the pure list (SDM = 0.078 ms,p = 0.048). This suggests a higher degree of variability in subjects’RR intervals during the mixed list exposure when compared to thepure list. The scaling of variability based on image regime (i.e.,increased variability in response to the more variable mixed imageregime) is consistent with our findings in the brain as well.For the pure vs mixed image regime comparison, wefound that pNN50 was statistically significant for segment 7[pNN50(mixed) = 0.258 vs pNN50(pure) = 0.164, p = 0.048]. Thisresult shows that during this segment, pNN50 was higher for themixed group compared to the pure group. Segment 7 corresponds tothe neutral segment for the pure regime and is the only neutral seg-ment sandwiched between two negative segment types. This findingimplies that the neutral image regime brought out a low heart ratevariability response in subjects, whereas the categorical and nega-tive image regimes inspired variability similar to the mixed list thatincluded all image types. This suggests that neutral image regimescould provide a baseline heart metric of dynamicity in subjects,and this baseline could be used to compare the agility of subjects’responses to other regimes (i.e., a mixed regime) in response toexternally applied stimuli. This is consistent with our findings inEEG, where the neutral image regime also had the lowest, baselineFD in comparison to the mixed image regime (Fig. 3).SDNN also showed significance for the first neutral segment(segment 3) to which the subject was exposed. This result, con-sistent with our other results, showed that for the mixed group,SDNN = 96.079 ms, whereas for the pure list, SDNN = 48.178 ms(p = 0.026). Though not shown in the other neutral segments,this finding in one neutral segment supports our hypothesis thatvariability in EKG is present when stimulating beyond the neutralpalette. It follows that the neutral segments of the pure image regimemay be used as a baseline metric for interpreting deviations awayfrom the baseline as an indicator of resiliency in participants.Findings in SD1 are similar to our findings in the heart ratemetrics above. SD1 conveys the variance of the y = mx line of thePoincaré plot, as discussed in Shaffer and Ginsberg (2017). SD1is a measure of short-term variability; generally, a higher SD1 isindicative of greater heart rate variability. This is related to theidea that the variable nature of signals coming from physiologicalsystems allows organisms to respond with resilience and adaptabil-ity in various quick-response situations. In our study, findings inSD1 mirror previously outlined heart rate metrics to give an insightinto subjects’ heart rate variability. When comparing the pure vsmixed image regimes, SD1 showed significance in the first neu-tral segment (segment 3), with SD1 = 79.59 ms for the mixed listand SD1 = 29.04 ms for the pure list (p = 0.03). The higher heartrate variability is correlated with the mixed regime once again forTABLE III. HRV parameter values for the pure vs mixed image regimes.HRV parameter Mixed (M; SE) (ms) Pure (M; SE) (ms) p-valueSDNN90.57; 11.8163.87; 6.830.056pNN50.255; .035 (%)0.181; 0.027 (%)0.097SD167.83; 12.2439.13; 7.330.05SD21128.63; 27.651164.36; 41.590.477Chaos 30, 113116 (2020); doi: 10.1063/5.000276430, 113116-10Published under license by AIP Publishing.ChaosARTICLEscitation.org/journal/chaFIG. 7. pNN50 and SD2 for the pure image regime only subjects. This shows significance (denoted by *) within subject for different segment types. Specifically, we comparethe heart rate metric of the subject from their negative image segments to their neutral image segments, their negative image segments to their categorical image segments,and their categorical image segments to their neutral image segments. Our study shows significant differences between neutral compared to both negative and categorical.TABLE IV. A summary of the HRV parameters for the pure image regime comparing the negative, neutral, and categorical image types. p-values that are not significant arelabeled as “n.s.”HRV parameter1Negative2Neutral3Categorical(1−2)p-value(2−3)p-valueSDNN (ms)59.6759.2664.92n.s.n.s.pNN50 (%)0.1890.1640.1870.0220.007SD1 (ms)34.3436.2240.32n.s.n.s.SD2 (ms)1170.791155.591167.350.0070.037Chaos 30, 113116 (2020); doi: 10.1063/5.000276430, 113116-11Published under license by AIP Publishing.ChaosARTICLEscitation.org/journal/chathe SD1 parameter. When we look at all the segments averaged,we also find that SD1 is higher for the mixed list (SD1 = 67.83 ms)than for the pure list (SD1 = 39.19 ms, p = 0.050). This reinforcesthe findings above and provides another avenue to see that vari-ability is heightened by the dynamicity of the mixed image regime,which may be important for clinically assessing heart signals basedon environmental or external stimuli.The final heart rate metric we will present in this paper is SD2.SD2 is a measure of the variance on y = mx + b (b = RR intervalaverage) of the Poincaré plot (Shaffer and Ginsberg, 2017). HigherSD2 variance indicates higher heart rate variability (HRV) (Shafferand Ginsberg, 2017). However, SD2 did not illustrate any signifi-cance based on segment number or on overall average of segments.Out of the four heart rate metrics analyzed, this is the only one thatdid not display a result for this type of analysis. As a result, purevs mixed image regimes do not seem to be differentiable with thisparameter. A summary of our findings in HRV parameters for thepure vs mixed regime is shown in Table III.We also looked at patterns in HRV for only subjects assignedthe pure image regime. The goal was to determine differencesbetween image types (i.e., negative, neutral, or categorical) within asubject. SD2 and pNN50 showed significance within subjects basedon the type of images they viewed. These two metrics were inves-tigated in subjects exposed to pure lists only (n = 31). Specifically,we compared each subject’s heart rate metric during one image typesegment (i.e., neutral) to the same subject’s results during anotherimage type segment (i.e., negative). In all cases of intra-subject vari-ability, we found that the corresponding heart rate metric was higherduring either the negative or the categorical segments when com-pared to the neutral segments within subjects (Fig. 7). This largerdegree of variability based on image type suggests that visual arousalis associated with heightened heart rate variability. After all, neg-ative and categorical images promote relatedness, while negativeimages also promote arousal and valence (Table II). Furthermore,the neutral segments seem to function as a baseline when com-pared to more arousing and related images. Categorical comparedto negative images did not have distinguishing differences in heartrate variability, suggesting that both have induced the same level ofarousal (activation) from a cardiac perspective. This suggests thatneutrality might allow the heart to settle into a baseline state thatcan serve as a control similar to the results we found in the brain.Therefore, using these types of image sequencing might be able todistinguish between subjects that easily adapt to a more arousingstimulus beyond the neutral state and those whose heart rate met-rics do not change based on the external stimulus. Reflection on ourcardiac results in the healthy heart reveals that comparison of thepure vs mixed image regimes may provide a safe way to investigateheart rate variability within an unhealthy subject. A summary of ourfindings in HRV parameters for the pure image regime is presentedin Table IV.Heart results: Heart rate metrics and memoryNone of the heart rate metrics showed correlation with memoryscores.Averages for all data were also included in the Heart RateMetric section to protect from any family wise errors. In addition,the Bonferroni correction is needed when comparing across seg-ment type. For these data, the negative vs neutral comparison forSD2 and RR interval has a p-value <0.017 (0.007 and 0.004, respec-tively). For pNN50, the categorical vs neutral has a p-value of 0.007,satisfying this condition as well.CONCLUSIONPhysiological systems depend on the internal function as well asthe ability to react to external, environmental stimuli for adaptationand survival. For example, the human heart beats upon activation byan autonomous pacemaker in the heart while simultaneously beingaffected by numerous neural feedback pathways. These factors worktogether to produce a normal heart rate consisting of complex fluc-tuations that occur both naturally and in response to environmentalfactors (Glass, 2001). Variability has been attributed to overall hearthealth; healthy subjects often display short-term variations in timeintervals between beats and a higher degree of cardiac chaos (Poonand Merrill, 1997). Our research looked at the response of the brainand heart to external image regimes to get a glimpse into how mini-mal interfacing may set up patterns that could be suggestive of healthand fitness (i.e., responsiveness to external stimuli or the propensityfor better memory recall).Our study was conducted in a pool of healthy subjects withno heart or brain abnormalities, and our findings suggest that volt-age patterns in subjects may be related to visual stimulation andrecall capabilities. These findings show that different image regimesmay influence the voltage pattern in the brain and heart. The mixedimage regime, which implements a constant switching between neu-tral, categorical, and negative images, arouses more complex andvariable signals in the brain and heart when compared to the pureimage regime, which has a fixed set of images for the subject toview. This suggests that an external, variable image regime canevoke a physiological response in both the brain and heart, possi-bly as an evolutionary response of the body to be more responsiveto the unpredictable and varying nature of our environment. Fur-thermore, we found a correlation between better memory recall andhigher fractal dimension but no correlation between better memoryrecall and greater heart rate variability. This second result suggeststhat attentiveness may be correlated with more agility in the brainbut does not elicit a distinct response in the heart. This may sug-gest that attentiveness does not demand or need flexible cardiacresponsiveness (possibly due to habituation). Taken together withour previous findings, this reinforces our hypothesis that the phys-iological response of the brain and heart can adjust accordingly totheir external environment.Our results also show that the fractal dimension in EEG scaleswith the complexity of the external image regime. For example, forthe pure vs mixed image regime, the biggest difference in complex-ity by construction of the experiment is between the mixed regime(most arousing) and the neutral regime (least arousing). The imageregimes get closer in their complexity as the pure image regimeis tuned from neutral, through categorical and then negative, withthe negative images being the closest in complexity to the mixedimage regime out of the three image types. The response in EEGas expressed via the fractal dimension mirrors the tuning in com-plexity—we find the biggest differences between the mixed imageChaos 30, 113116 (2020); doi: 10.1063/5.000276430, 113116-12Published under license by AIP Publishing.ChaosARTICLEscitation.org/journal/charegime when compared to the neutral image regime and this dif-ference starts to fade when comparing mixed and categorical imageregimes and then fades even more when comparing mixed and neg-ative image regimes. Likewise, the variability in the heart tracks inthe same fashion, with the most complex external image regime (i.e.,the mixed image regime) illustrating the most variability in the heartwhen compared to the pure neutral image regime. Our results sug-gest that in the heart and brain, a dynamic switching in physiologicalresponse occurs to mirror the external trigger. Other authors havehypothesized that these shift in fractal behavior in EEG, for exam-ple, are a necessity for survival to switch flexibly from one state toanother when triggered. Some authors suggest a theory called the“complexity matching effect” (Allegrini et al., 2006 and West et al.,2008), where the complexity of the external triggers gives rise to thefractal behavior in EEG. Many physiological systems such as theheart also exhibit this type of behavior as a way to be an agile andflexible in response to the stresses of the environment. Some scien-tists have hypothesized that if nature has fractal features, the bodymust also have fractal features as a system that has to continuallyrespond to it (Werner, 2010 and Chiavlo, 2010).Adapting techniques from physics to dynamical systems pro-vides a new lens through which we can understand how neurologicaland cardiac processes behave. Although our work needs to be scaledto a clinical environment, since pure and mixed image regimes arefeasible to mimic in a clinical setting, this method of stimulationmay uncover details of how certain pathologies present in the brainand the heart. Based on the simplicity of the calculations and theaccessibility of computational techniques, our methods of analysismay provide insight to inform the modification of clinical thera-pies and to influence patient treatment and care. Furthermore, thispaper illustrates the ubiquity of techniques from physics in pursuitof understanding the physical world.SUPPLEMENTARY MATERIALSee the supplementary material for the complete set of allresults presented in this paper.ACKNOWLEDGMENTSThe authors would like to thank Dr. Matt Cathey for his sup-port and review of our statistical analysis. The authors would alsolike to thank Timothy Phillips and Catherine (Catie) Cronister forhelp with data cleaning.DATA AVAILABILITYThe data that support the findings of this study are availablefrom the corresponding author upon reasonable request.REFERENCESAllegrini, P., Bologna, M., Grigolini, P., and West, B. J., “Response of com-plex systems to complex perturbations: The complexity matching effect,”arXiv:cond-mat/0612303v1 (2006).Arsac, L. M. and Deschodt-Arsac, V., “Detrended fluctuation analysis in a simplespreadsheet as a tool for teaching fractal physiology,” Adv. Physiol. Educ. 42,493–499 (2018).Barnacle, G. E., Tsivilis, D., Schaefer, A., and Talmi, D., “Local context influ-ences memory for emotional stimuli but not electrophysiological markers ofemotion-dependent attention,” Psychophysiology 55, 1–14 (2018).Beck, A. T., Ward, C. H., Mendelson, M., Mock, J., and Erbaugh, J., “Aninventory for measuring depression,” Arch. Gen. Psychiatry 4, 561–571(1961).Beck, A. T., Epstein, N., Brown, G., and Steer, R. A., “An inventory for measur-ing clinical anxiety: Psychometric properties,” J. Consul. Clin. Psychol. 56,893–897 (1988).Captur, G., Karperien, A. L., Hughes, A. D., Francis, D. P., and Moon, J. C., “Thefractal heart—Embracing mathematics in the cardiology clinic,” Nat. Rev.Cardiol. 14, 56–64 (2017).Chiavlo, D. R., “Emergent complex neural dynamics,” Nat. Phys. 6, 744–750(2010).Dan-Glauser, E. S. and Scherer, K. R., “The Geneva affective picture database(GAPED): A new 730-picture database focusing on valence and normativesignificance,” Behav. Res. Methods 43, 468–477 (2011).Delignieres, D., Ramdani, S., Lemoine, L., Torre, K., Fortes, M., and Ninot, G.,“Fractal analyses for ‘short’ time series: A re-assessment of classical methods,”J. Math. Psychol. 50, 525–544 (2006).Faes, L., Nollo, G., Jurysta, F., and Marinazzo, D. “Information dynamics ofbrain-heart physiological networks during sleep,” New J. Phys. 16, 105005(2014).Glass, L., “Synchronization and rhythmic processes in physiology,” Nature 410,277–284 (2001).Goldberger, A. L. and West, B. J., “Fractals in physiology and medicine,” Yale J.Biol. Med. 60, 421–435 (1987).Golinska, A. K., “Detrended fluctuation analysis (DFA) in biomedical signalprocessing: Selected examples,” Stud. Logic Grammar Rhetoric 29, 107–115(2012).Higuchi, T., “Approach to an irregular time series on the basis of the fractaltheory,” Physica D 31, 277–283 (1988).Ivanov, P. C., Liu, K. K. L., Bartsch, R. P., “Focus on the emerging new fields ofnetwork physiology and network medicine,” New J. Phys. 18, 100201 (2016).Iyengar, N., Peng, C.-K., Morin, R., Goldberger, A. L., and Lipsitz, L. A., “Age-related alterations in the fractal scaling of cardiac interbeat interval dynamics,”Am. J. Physiol. 271, R1078–R1084 (1996).Katz, M., “Fractals and the analysis of waveforms,” Comput. Biol. Med. 18(3), 145(1988).Kensinger, E. A., “Remembering the details: Effects of emotion,” Emotion Rev. 1,99–113 (2009).Kesic, S. and Spasic, S., “Application of Higuichi’s fractal dimensions from basicto clinical neurophysiology: A review,” Comput. Methods Programs Biomed.133, 55–70 (2016).Lang, P. J., Bradley, M. M., and Cuthbert, B. N., International Affective PictureSystem (IAPS): Technical Manual and Affective Ratings (NIMH Center for theStudy of Emotion and Attention, 1999).Mather, M. and Sutherland, M. R., “Arousal-biased competition in perception andmemory,” Perspect. Psychol. Sci. 6, 114–133 (2011).Nayyeri, P., “Analyzing electrocardiography (ECG) signal using fractal method,”Int. J. Curr. Eng. Technol. 7, 498–505 (2017).Peng, C. K., Havlin, S., Stanley, H. E., and Goldberger, A. L., “Quantification ofscaling exponents and crossover phenomena in nonstationary heartbeat timeseries,” Chaos 5, 82–87 (1995).Pincus, S. M. and Goldberger, A. L., “Physiological time-series analysis: What doesregularity quantify?,” Am. J. Physiol. 266, H1643–H1656 (1994).Piper, D., Schiecke, K., Pester, B., Benninger, F., Feucht, M., and Witte, H., “Time-variant coherence between heart rate variability and EEG activity in epilepticpatients: An advanced coupling analysis between physiological networks,”New J. Phys. 16, 11 (2014).Piskorski, J. and Guzik, P., “Geometry of the Poincare plot of RR intervals and itsasymmetry in healthy adults,” Physiol. Meas. 28, 287–300 (2007).Poon, C.-S. and Merrill, C. K., “Decrease of cardiac chaos in congestive heartfailure,” Nature 389, 492–495 (1997).Riley, M. A., Bonnette, S., Kuznetsov, N., Wallot, S., and Gao, J., “A tutorialintroduction to adaptive fractal analysis,” Front. Physiol. 3, 1–10 (2012).Chaos 30, 113116 (2020); doi: 10.1063/5.000276430, 113116-13Published under license by AIP Publishing.ChaosARTICLEscitation.org/journal/chaSchmidt, S. R. and Saari, B., “The emotional memory effect: Differential process-ing or item distinctiveness?,” Mem. Cognit. 35, 1905–1916 (2007).“Sgolay,” MathWorks, Savitzky-Golay filter design—MATLAB, see https://www.mathworks.com/help/signal/ref/sgolay.html.Shaffer, F. and Ginsberg, J. P., “An overview of heart rate variability metrics andnorms,” Front. Public Health 5, 1–17 (2017).Stam, C. J., “Nonlinear dynamical analysis of EEG and MEG: Review of anemerging field,” Clin. Neurophysiol. 116, 226–2301 (2005).Talmi, D. and Moscovitch, M., “Can semantic relatedness explain the enhance-ment of memory for emotional words?,” Mem. Cognit. 32, 742–751 (2004).Talmi, D., Luk, B. T. C., McGarry, L. M., and Moscovitch, M., “The contributionof relatedness and distinctiveness to emotionally-enhanced memory,” J. Mem.Lang. 56, 555–574 (2007).Talmi, D. and McGarry, L. M., “Accounting for immediate emotional memoryenhancement,” J. Mem. Lang. 66, 93–108 (2012).Talmi, D., “Enhanced emotional memory: Cognitive and neural mechanisms,”Curr. Dir. Psychol. Sci. 22, 430–436 (2013).Watts, S., Buratto, L. G., Brotherhood, E. V., Barnacle, G. E., and Schaefer,A., “The neural fate of neutral information in emotion-enhanced memory,”Psychophysiology 51, 673–684 (2014).Werner, G., “Fractals in the nervous system: Conceptual implications for theoret-ical neuroscience,” Front. Physiol. 1, 1–28 (2010).Wessa, M., Kanske, P., Neumeister, P., Bode, K., Heissler, J., and Schonfelder, S.,“Emopics: Subjective and psychophysiological evaluations of new visual mate-rial for clinical-bio-psychological research,” J. Clin. Psychol. Psychother. 1, 77(2010).West, B. J., Geneston, E. L., and Grigolini, P., “Maximizing information exchangebetween complex networks,” Phys. Rep. 468, 1–99 (2008).Wright, J. J., Kydd, R. R., and Liley, D. T. J., “EEG models, chaotic and linear,”Psycholoquy 4, eeg-chaos.l.wright (1993).Chaos 30, 113116 (2020); doi: 10.1063/5.000276430, 113116-14Published under license by AIP Publishing.View publication stats | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/366877840Self-Myofascial Release of the Foot Plantar Surface: The Effects of a SingleExercise Session on the Posterior Muscular Chain Flexibility after One HourArticle in International Journal of Environmental Research and Public Health · January 2023DOI: 10.3390/ijerph20020974CITATIONS0READS9711 authors, including:Some of the authors of this publication are also working on these related projects:Active workplace intervention versus no treatment on return on investment, costs and sickness absence in sedentary affected by musculoskeletal conditions, an:Economic systematic review and meta-analysis of randomized and non-randomized trials View projectMindful Lab Project View projectEleonora MontagnaniChildren's Hospital Los Angeles7 PUBLICATIONS 9 CITATIONS SEE PROFILERiccardo Di GiminianiUniversity of L'Aquila38 PUBLICATIONS 538 CITATIONS SEE PROFILEStefano PalermiUniversity of Naples Federico II84 PUBLICATIONS 446 CITATIONS SEE PROFILEAll content following this page was uploaded by Stefano Palermi on 05 January 2023.The user has requested enhancement of the downloaded file.Citation: Russo, L.; Montagnani, E.;Pietrantuono, D.; D’Angona, F.;Fratini, T.; Di Giminiani, R.; Palermi,S.; Ceccarini, F.; Migliaccio, G.M.;Lupu, E.; et al. Self-MyofascialRelease of the Foot Plantar Surface:The Effects of a Single ExerciseSession on the Posterior MuscularChain Flexibility after One Hour. Int.J. Environ. Res. Public Health 2023, 20,974. https://doi.org/10.3390/ijerph20020974Academic Editors: Paul B.Tchounwou and Britton W. BrewerReceived: 19 November 2022Revised: 1 January 2023Accepted: 3 January 2023Published: 5 January 2023Copyright:© 2023 by the authors.Licensee MDPI, Basel, Switzerland.This article is an open access articledistributedunderthetermsandconditions of the Creative CommonsAttribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).International Journal ofEnvironmental Researchand Public HealthArticleSelf-Myofascial Release of the Foot Plantar Surface: The Effectsof a Single Exercise Session on the Posterior Muscular ChainFlexibility after One HourLuca Russo 1,*,†, Eleonora Montagnani 2,†, Davide Pietrantuono 3, Fabiola D’Angona 3, Tommaso Fratini 1,Riccardo Di Giminiani 4, Stefano Palermi 5, Francesco Ceccarini 6, Gian Mario Migliaccio 7,‡, Elena Lupu 8,‡and Johnny Padulo 9,‡1Department of Human Sciences, Università Telematica degli Studi IUL, 50122 Florence, Italy2Department of Sports and Health Sciences, University of Brighton, Brighton BN2 4AT, UK3Department of Medicine and Health Sciences, University of Molise, 86100 Campobasso, Italy4Department of Biotechnological and Applied Clinical Sciences, University of L’Aquila, 67100 L’Aquila, Italy5Public Health Department, University of Naples Federico II, 80132 Naples, Italy6Department of Psychology, New York University Abu Dhabi, Abu Dhabi 129188, United Arab Emirates7Department of Performance, Sport Science Lab, 09131 Cagliari, Italy8Department of Motor Activities, Petroleum Gas University Ploiesti, 100600 Ploiesti, Romania9Department of Biomedical Sciences for Health, Università degli Studi di Milano, 20133 Milan, Italy*Correspondence: l.russo@iuline.it†These authors contributed equally to this work as first author.‡These authors contributed equally to this work as last author.Abstract: This study evaluated the effects of a single exercise session of Self-Myofascial Release(SMR) on the posterior muscular chain flexibility after one hour from the intervention. Thirty-sixparticipants performed SMR using a rigid ball under the surface of both feet. Participants were testedwith the Sit and Reach (S&R) test at four different times: before (T0), immediately after (T1), 30 (T2),and 60 (T3) minutes after the SMR intervention. The sample (n = 36) was categorized into threegroups: (1) flexible, (2) average, and (3) stiff, based on the flexibility level at T0 (S&R values of >10 cm,>0 but <10 cm and <0 cm, respectively). For the whole sample, we detected significant improvementsin the S&R test between the T1, T2, and T3 compared to T0. The stiff group showed a significant(p < 0.05) improvement between T1–T2 and T1–T3. Results were similar between the average groupand the whole sample. The flexible group did not show any significant difference (p > 0.05) over time.In conclusion, this investigation demonstrated that an SMR session of both feet was able to increaseposterior muscular chain flexibility up to one hour after intervention. Considering that a standardtraining session generally lasts one hour, our study can help professionals take advantage of SMReffects for the entire training period. Furthermore, our results also demonstrate that physical exercisepractitioners should also assess individuals’ flexibility before training, as the SMR procedure used inthis work does not seem necessary in flexible individuals.Keywords: self-myofascial release; training; sit and reach; posterior muscular chain; flexibility1. IntroductionSelf-Myofascial Release (SMR) is a common self-treatment strategy used by a wideplethora of physically active people, such as gym or home fitness users [1,2], high-levelathletes [3–5], and global health and postural trainers [6–8]. According to the literature,SMR can be defined as a subcategory of myofascial release [4,9,10], which is a series ofmanipulative techniques where pressure is applied to soft tissue, with regard to musclesand fascia [11]. In fact, the main declared goal of SMR is to mobilize the targeted softtissues with rolling devices, such as foam rollers or hard balls, which are used by applyinga certain amount of pressure over the surface of the skin, thus creating a self-inducedInt. J. Environ. Res. Public Health 2023, 20, 974. https://doi.org/10.3390/ijerph20020974https://www.mdpi.com/journal/ijerphInt. J. Environ. Res. Public Health 2023, 20, 9742 of 12massage effect. The effectiveness of SMR is well explained in the literature and differentauthors adduce several mechanisms of action. Compression may increase fascial elasticitythanks to a temporary change in water content [12]. After the compression, the local bloodflow is increased, facilitating the removal of metabolites and delivering oxygen [13–15]as well as the warming of the tissue, ripping the restrictions within layers of the fasciaand restoring soft tissue elasticity [16], reducing the inflammation status [9]. The pressureon soft tissue also leads to a mechanoreceptor stimulation, which is then able to reducemuscular and fascia tension [17]. All these mechanisms are well-known and describedin detail in previously published reviews [9,10]. Although some authors are still dubiousabout the proper name to use for these techniques applied on the myofascial tissue [18],a growing amount of evidence suggests positive outcomes of SMR, such as decreasingpain, reduction of delay onset muscle soreness (DOMS), faster physical recovery, and anincrease in muscle flexibility and joint range of motion (ROM) without a reduction in forceproduction [9,19,20].Doubtless, the increase of joint ROM and muscle flexibility are two of the most interest-ing aspects for coaches and sport science practitioners. Flexibility is a key factor for severalsports [21] as well as for daily life tasks and musculoskeletal health [22–24]. Stretchingis classically considered the best way to improve flexibility over time [25], but evidencesuggests both chronic and acute positive effects of SMR on flexibility [9,19,26,27], despitesome doubts remaining [28].The acute effects of SMR on flexibility are the most investigated because of thewidespread use of SMR techniques before and after physical training. In fact, it is verycommon to see athletes and fitness users adopt foam rollers or rubber balls before training,especially on the lower limbs and the back. In addition, the literature gives great attentionto the effects of SMR, especially on the posterior muscular chain [29–31]. The posteriormuscular chain encompasses a series of muscles interlinked by the deep fascia, extendingfrom the foot to the fascial sheath of the eyeball [32]. For this reason, the myofascial chainsare usually related to the whole-body posture [33,34]. The principal acute effects of SMRon the posterior muscular chain are (1) increasing flexibility of the whole chain or specificmuscles (e.g., hamstrings) [28,29] and (2) increasing ankle and spine ROM [29,31]. It mustbe underlined that these effects are evoked by performing SMR on any segment of theposterior muscular chain [31].Although the acute effects of SMR on the posterior muscular chain are supportedby strong evidence, very few data are available on the duration of this “lengthening”effect. Researchers have shown how the SMR effects last up to 10 min [16,35], and otherdata confirmed the duration of such effects for up to 30 min [36]. However, the longer-term effects of SMR on posterior muscular chain flexibility have not been demonstratedyet. Moreover, no data are available regarding the acute duration of SMR in people withdifferent levels of posterior chain flexibility. This information could be very useful in termsof practice and training, to properly organize the exercises along the training sessions anduse SMR stimulation with a specific timing before the training session or competition.To our knowledge, no studies have investigated the duration of the acute effectsof an SMR on the feet plantar muscles in different clusters of people according to theirflexibility level. In fact, it is reasonable to speculate that SMR stimulation offers higherROM gain in individuals with a lack of posterior muscular chain flexibility. Therefore,the aims of this investigation were twofold: (1) measure the flexibility obtained duringa one-hour span after the SMR procedure, and (2) measure the flexibility obtained insubgroups of individuals divided by their flexibility levels, which will be defined beforethe SMR procedure.2. Materials and Methods2.1. ParticipantsThirty-six volunteer individuals (age 23.9 ± 1.9 years; body height 171.7 ± 10.8 cm;body mass 68.4 ± 15.1 kg) were selected for the study. An a priori analysis was conductedInt. J. Environ. Res. Public Health 2023, 20, 9743 of 12for the sample size setting the error probability at 5% and the power of the analysis at 80%.The recommended sample was 24 participants, but we eventually recruited 36 individualsfor the entire sample. Moreover, another a priori analysis was conducted for the sub-groupssetting the effect size at 0.4, the error probability at 5%, and the power of the analysisat 80%, obtaining 10 participants for the group. The entire sample was divided into 17males (age 24.2 ± 1.3 years; body height 180.2 ± 8.3 cm; body mass 80.5 ± 12.9 kg) and19 females (age 23.7 ± 2.3 years; body height 164.1 ± 5.9 cm; body mass 57.5 ± 5.7 kg).Participants were included in the study if they: (1) reported to be in good health andphysically active (declaring to perform at least 150 min/w of physical activity); (2) did notsuffer any accidents or injuries in the 12 months preceding the test; (3) did not report painin the ankles or feet; (4) did not have any musculoskeletal, systemic and/or neurologicaldisease; and (5) did not show any hypermobility or joint laxity sign. The protocol conformedto internationally accepted policy statements regarding the use of human participants inaccordance with the Declaration of Helsinki Declaration and was approved by the OvidiusUniversity of Constanta Nr. 126 din 18 March 2022. All participants gave their writteninformed consent to participate in the study after receiving a thorough explanation of thestudy’s protocol.2.2. InstrumentationBox Sit and Reach: The testing tool consists of a 31 cm high cube with a protrusionat the upper level along which the displacement in centimeters is determined (Figure 1).The external part, which constitutes the protrusion that extends beyond the front edge andtherefore towards the tested participant, has a length of 23 cm, and in the center, there isthe trolley through which its movement defines the elongation and therefore the variousdegrees in negative to positive values of flexibility [37]. individuals divided by their flexibility levels, which will be defined before the SMR pro-cedure. 2. Materials and Methods 2.1. Participants Thirty-six volunteer individuals (age 23.9 ± 1.9 years; body height 171.7 ± 10.8 cm; body mass 68.4 ± 15.1 kg) were selected for the study. An a priori analysis was conducted for the sample size setting the error probability at 5% and the power of the analysis at 80%. The recommended sample was 24 participants, but we eventually recruited 36 indi-viduals for the entire sample. Moreover, another a priori analysis was conducted for the sub-groups setting the effect size at 0.4, the error probability at 5%, and the power of the analysis at 80%, obtaining 10 participants for the group. The entire sample was divided into 17 males (age 24.2 ± 1.3 years; body height 180.2 ± 8.3 cm; body mass 80.5 ± 12.9 kg) and 19 females (age 23.7 ± 2.3 years; body height 164.1 ± 5.9 cm; body mass 57.5 ± 5.7 kg). Participants were included in the study if they: (1) reported to be in good health and phys-ically active (declaring to perform at least 150 min/w of physical activity); (2) did not suffer any accidents or injuries in the 12 months preceding the test; (3) did not report pain in the ankles or feet; (4) did not have any musculoskeletal, systemic and/or neurological disease; and (5) did not show any hypermobility or joint laxity sign. The protocol conformed to internationally accepted policy statements regarding the use of human participants in ac-cordance with the Declaration of Helsinki Declaration and was approved by the Ovidius University of Constanta Nr. 126 din 18 March 2022. All participants gave their written informed consent to participate in the study after receiving a thorough explanation of the study’s protocol. 2.2. Instrumentation Box Sit and Reach: The testing tool consists of a 31 cm high cube with a protrusion at the upper level along which the displacement in centimeters is determined (Figure 1). The external part, which constitutes the protrusion that extends beyond the front edge and therefore towards the tested participant, has a length of 23 cm, and in the center, there is the trolley through which its movement defines the elongation and therefore the various degrees in negative to positive values of flexibility [37]. Figure 1. Sit and Reach test and measuring box. Rubber ball: A 7 cm diameter rubber ball with a weight of 140 g (ATS, Arezzo, Italy), was used for the foot SMR (Figure 2). The rigidity of the ball does not allow any kind of deformation under foot and lower limb pressure. Figure 1. Sit and Reach test and measuring box.Rubber ball: A 7 cm diameter rubber ball with a weight of 140 g (ATS, Arezzo, Italy),was used for the foot SMR (Figure 2). The rigidity of the ball does not allow any kind ofdeformation under foot and lower limb pressure.Int. J. Environ. Res. Public Health 2023, 20, 9744 of 12Int. J. Environ. Res. Public Health 2023, 20, x FOR PEER REVIEW 4 of 13 Figure 2. Rubber ball position under the foot for the Self Myofascial Release (SMR). Image modified by Russo et al. 2016 [38] with permission. 2.3. Procedure and Data Collection Testing was carried out in a Sport Performance Laboratory at a temperature of 20 °C and relative humidity of 51% according to previous studies [39,40]. The testing session was performed in the morning between 9.00 and 11.00 a.m. in order to avoid any kind of circadian influence on muscle flexibility [41]. The day before the testing session, in the same place and at the same time of the day, all the participants were familiarized with the S&R test. The results of this familiarization session were recorded. During the test session, each participant was tested in four different moments across a one-hour testing experimental procedure (Figure 3): • T0: corresponding to the first S&R test carried out on the participants; • T1: corresponding to the second S&R test on the participants immediately after the SMR technique; • T2: corresponding to the third test of the S&R, carried out 30 min after T1; • T3: corresponding the fourth test of the S&R, carried out 60 min after T1. Figure 3. Experimental setting. S&R was chosen because of its unique ability to incorporate the lumbar spine and flexibility of the hamstring simultaneously while tensioning the posterior muscular chain [42]. The test was performed as reported in the original literature [37]: the participants sat on the floor barefoot, the legs were straight, the soles of the feet were placed flat against Figure 2. Rubber ball position under the foot for the Self Myofascial Release (SMR). Image modifiedby Russo et al. 2016 [38] with permission.2.3. Procedure and Data CollectionTesting was carried out in a Sport Performance Laboratory at a temperature of 20 ◦Cand relative humidity of 51% according to previous studies [39,40]. The testing sessionwas performed in the morning between 9.00 and 11.00 a.m. in order to avoid any kindof circadian influence on muscle flexibility [41]. The day before the testing session, in thesame place and at the same time of the day, all the participants were familiarized with theS&R test. The results of this familiarization session were recorded.During the test session, each participant was tested in four different moments across aone-hour testing experimental procedure (Figure 3):•T0: corresponding to the first S&R test carried out on the participants;•T1: corresponding to the second S&R test on the participants immediately after theSMR technique;•T2: corresponding to the third test of the S&R, carried out 30 min after T1;•T3: corresponding the fourth test of the S&R, carried out 60 min after T1.Int. J. Environ. Res. Public Health 2023, 20, x FOR PEER REVIEW 4 of 13 Figure 2. Rubber ball position under the foot for the Self Myofascial Release (SMR). Image modified by Russo et al. 2016 [38] with permission. 2.3. Procedure and Data Collection Testing was carried out in a Sport Performance Laboratory at a temperature of 20 °C and relative humidity of 51% according to previous studies [39,40]. The testing session was performed in the morning between 9.00 and 11.00 a.m. in order to avoid any kind of circadian influence on muscle flexibility [41]. The day before the testing session, in the same place and at the same time of the day, all the participants were familiarized with the S&R test. The results of this familiarization session were recorded. During the test session, each participant was tested in four different moments across a one-hour testing experimental procedure (Figure 3): • T0: corresponding to the first S&R test carried out on the participants; • T1: corresponding to the second S&R test on the participants immediately after the SMR technique; • T2: corresponding to the third test of the S&R, carried out 30 min after T1; • T3: corresponding the fourth test of the S&R, carried out 60 min after T1. Figure 3. Experimental setting. S&R was chosen because of its unique ability to incorporate the lumbar spine and flexibility of the hamstring simultaneously while tensioning the posterior muscular chain [42]. The test was performed as reported in the original literature [37]: the participants sat on the floor barefoot, the legs were straight, the soles of the feet were placed flat against Figure 3. Experimental setting.S&R was chosen because of its unique ability to incorporate the lumbar spine and flex-ibility of the hamstring simultaneously while tensioning the posterior muscular chain [42].The test was performed as reported in the original literature [37]: the participants sat onthe floor barefoot, the legs were straight, the soles of the feet were placed flat against thebox, forming an angle of 90 degrees with the ankles, and both knees were locked andpressed flat to the floor. From this position, the participants slowly stretched their handsInt. J. Environ. Res. Public Health 2023, 20, 9745 of 12forward with outstretched arms and pushed the measuring gage, reaching forward asfar as possible, holding the maximal reach for 2 s [43] in order to allow the operator toread the result. Each participant performed each test (T0–T3) three times, and the averagevalues were considered for the statistical analysis [44,45]. The S&R measures and thefamiliarization results were made by the same expert kinesiologist (having more than 10years of experience).To be consistent with previous works on SMR and S&R [29–31,46,47], no warming-upprocedures were performed before the S&R test.According to the results obtained in T0, three sub-groups were created consideringthe level of posterior muscular chain flexibility. Specifically, a flexible group (FG), a groupwith average flexibility (AG), and a stiff group (SG–11 participants) were created. Thegroups’ categorization was based on flexibility level at T0. In fact, the FG, the AG, and theSG showed S&R values of > 10 cm, > 0 but < 10 cm and < 0 cm respectively [48]. The FGcounted 10 participants (4 males and 6 females), AG had 15 participants (4 males and 11females), and SG had 11 participants (9 males and 2 females). Because of the total numberof participants, the three sub-groups were mixed for gender. Previously published datashow no influence of gender on the SMR acute improvements for S&R performance [47].2.4. InterventionImmediately after the measurement of the posterior muscular chain flexibility, eachindividual performed a session of SMR of the plantar fascia of both feet. Participantsperformed the SMR training in an upright position, holding their hands on the wall to keepthe balance and to allow the application of consistent pressure on the rubber ball, avoidingtoo much or too low-pressure application. The lead researcher (and expert kinesiologist)instructed participants regarding the SMR procedure. The rubber ball was placed under thesole of the foot, precisely on the plantar arch. The kinesiologist instructed each participantto move slowly the foot, seeking a kind of painful mass or nodule perceived within thesame muscle band, called a trigger point. Once the trigger point had been self-identified bythe participant, pressure was applied in order to feel slight pain and temporary discomfortequivalent to a pain level of 7 out of 10 [16,18]. Three bouts of myofascial release massagewere performed for each foot, each bout lasted 30 s. The kinesiologist randomly assignedthe starting foot, and the work sequence was switching back and forth between the leftand right foot. In the first bout, the trigger point was identified and pressed. In the secondbout, slow circling movements were performed around the trigger point. In the third andlast bout, pressure was still applied to the trigger point combining flexion and extensionmovements of the toes. The intervention had a total duration of 3 min.All the participants completed the intervention, and no dropout cases or adverseeffects of the treatment were registered in this study. Immediately after the intervention,each participant performed the T1 S&R test. Then, a rest period of 30 min was observeduntil T2 and again 30 min until the final S&R test at T3. During the rest time, the participantswere free to walk around the laboratory or to take a sit. Any other type of activity wasforbidden (e.g., jumping, squatting).2.5. Statistical AnalysisDescriptive data were reported as mean and standard deviation (SD). The Shapiro–Wilk test was applied to check that data were normally distributed. Data were analyzedusing a repeated-measures analysis of variance (ANOVA) with Fisher values and Bonferronipost hoc corrections to look for differences in the S&R values across the four measurementstaken over 60 min. The effect size was also calculated (eta squared, η2) for a better interpre-tation of the results (values of 0.01, 0.06, and above 0.15 were considered small, medium,and large, respectively). Repeated ANOVA measures were undertaken for the sub-groupsseparately, while a Kruskal–Wallis test was used to look for differences in the flexibility atT0 for the different subgroups. A significance level of α = 0.05 was adopted and all dataInt. J. Environ. Res. Public Health 2023, 20, 9746 of 12were analyzed with Statistical Package for the Social Science (SPSS), version 27.0 (SPSS Inc.,Chicago, IL, USA).3. ResultsData analysis showed an effect of the SMR on the plantar surface of both feet on theS&R results over time (Table 1).Table 1. Results of ANOVA for repeated measures. Participants’ S&R values modification before andafter SMR intervention.S&R (cm)Group (n)T0T1T2T3Fη2p ValueEntire sample (36)4.3 ± 8.75.7 ± 9.1 *6.4 ± 8.4 *7.0 ± 8.7 *#9.3740.4600.000Males (17)0.5 ± 9.31.4 ± 9.72.4 ± 8.82.6 ± 9.41.9750.2970.164Females (19)7.7 ± 6.69.5 ± 6.5 *10.1 ± 6.1 *11.0 ± 5.8 *#10.4570.6620.000FG (10)13.7 ± 2.714.6 ± 3.414.5 ± 4.415.5 ± 4.11.1980.3390.378AG (15)5.9 ± 3.68.1 ± 4.1 *8.6 ± 4.3 *9.0 ± 4.7 *8.8380.6880.002SG (11)−6.5 ± 4.3−5.8 ± 4.5−3.8 ± 3.9 *#−3.4 ± 4.64.9920.6490.032Note: data expressed as mean ± SD; T0 before intervention; T1 immediately after intervention; T2 30 minafter intervention; T3 60 min after intervention; FG flexible group; AG average group; SG stiff group. Post-hoccomparison: * significant differences compared to T0; # significant difference compared to T1.No differences were present for the entire sample between the S&R values measuredin familiarization and test session at T0 (4.0 ± 8.5 and 4.3 ± 8.7 cm, respectively; p = 0.184).For the whole sample (Figure 4), significant differences were found for the S&Rbetween T0 and T1 (4.3 ± 8.7 and 5.7 ± 9.1 cm; p = 0.001), T0 and T2 (4.3 ± 8.7 and6.4 ± 8.4 cm; p = 0.000), and T0 and T3 (4.3 ± 8.7 and 7.0 ± 8.7 cm; p = 0.000). Anothersignificant difference was found for the S&R between T1 and T3 (5.7 ± 9.1 and 7.0 ± 8.7 cm;p = 0.002). small, medium, and large, respectively). Repeated ANOVA measures were undfor the sub-groups separately, while a Kruskal–Wallis test was used to look for diffin the flexibility at T0 for the different subgroups. A significance level of α = 0.adopted and all data were analyzed with Statistical Package for the Social Science version 27.0 (SPSS Inc., Chicago, IL, USA). 3. Results Data analysis showed an effect of the SMR on the plantar surface of both feetS&R results over time (Table 1). Table 1. Results of ANOVA for repeated measures. Participants’ S&R values modificatioand after SMR intervention. S&R (cm) Group (n) T0 T1 T2 T3 F η2 pEntire sample (36) 4.3 ± 8.7 5.7 ± 9.1 * 6.4 ± 8.4 * 7.0 ± 8.7 *# 9.374 0.460 Males (17) 0.5 ± 9.3 1.4 ± 9.7 2.4 ± 8.8 2.6 ± 9.4 1.975 0.297 Females (19) 7.7 ± 6.6 9.5 ± 6.5 * 10.1 ± 6.1 * 11.0 ± 5.8 *# 10.457 0.662 FG (10) 13.7 ± 2.7 14.6 ± 3.4 14.5 ± 4.4 15.5 ± 4.1 1.198 0.339 AG (15) 5.9 ± 3.6 8.1 ± 4.1 * 8.6 ± 4.3 * 9.0 ± 4.7 * 8.838 0.688 SG (11) −6.5 ± 4.3 −5.8 ± 4.5 −3.8 ± 3.9 *# −3.4 ± 4.6 4.992 0.649 Note: data expressed as mean ± SD; T0 before intervention; T1 immediately after interventiomin after intervention; T3 60 min after intervention; FG flexible group; AG average group; group. Post-hoc comparison: * significant differences compared to T0; # significant differenpared to T1. No differences were present for the entire sample between the S&R values mein familiarization and test session at T0 (4.0 ± 8.5 and 4.3 ± 8.7 cm, respectively; p = For the whole sample (Figure 4), significant differences were found for the Stween T0 and T1 (4.3 ± 8.7 and 5.7 ± 9.1 cm; p = 0.001), T0 and T2 (4.3 ± 8.7 and 6.4 ± p = 0.000), and T0 and T3 (4.3 ± 8.7 and 7.0 ± 8.7 cm; p = 0.000). Another significantence was found for the S&R between T1 and T3 (5.7 ± 9.1 and 7.0 ± 8.7 cm; p = 0.00 Figure 4. S&R values modifications over time before and after the SMR intervention. The bline represents the average values and thin grey lines are the real values of each participant.Figure 4. S&R values modifications over time before and after the SMR intervention. The black fatline represents the average values and thin grey lines are the real values of each participant. T0: testbefore intervention. T1: test after intervention. T2: test 30 min after intervention. T3: test 60 min afterintervention. SMR: self-myofascial release. * Significant differences with p < 0.05.Int. J. Environ. Res. Public Health 2023, 20, 9747 of 12A primary stratification of the entire sample was made by dividing by sex. Themale group showed a lower level of flexibility with respect to the female group at theT0 (0.5 ± 9.3 cm and 7.7 ± 6.6 respectively; p = 0.036), T1 (1.4 ± 9.7 cm and 9.5 ± 6.5respectively; p = 0.013), T2 (2.4 ± 8.8 cm and 10.1 ± 6.1 respectively; p = 0.010), and T3(2.6 ± 9.4 cm and 11.0 ± 5.8 respectively; p = 0.009). The male group showed increasingflexibility over time, but no statistical differences were measured: 0.5 ± 9.3 cm at T0;1.4 ± 9.7 cm at T1; 2.4 ± 8.8 cm at T2; 2.6 ± 9.4 cm at T3. For the female group, significantdifferences were found between T0 and T1 (7.7 ± 6.6 and 9.5 ± 6.5 cm; p = 0.000), T0 andT2 (7.7 ± 6.6 and 10.1 ± 6.1 cm; p = 0.004), and T0 and T3 (7.7 ± 6.6 and 11.0 ± 5.8 cm;p = 0.000). Another significant difference was found between T1 and T3 (9.5 ± 6.5 and11.0 ± 5.8 cm; p = 0.016).The entire sample was also divided into three subgroups, namely the FG, AG, andSG, according to the S&R values measured at T0. The three subgroups showed a signif-icant difference for the S&R at T0 (p = 0.000). Obviously, the FG showed higher flexibil-ity (13.7 ± 2.7 cm) than AG (5.9 ± 3.6 cm) and the same was for AG compared to SG(−6.5 ± 4.3 cm), respectively. The behavior of the subgroups over time was different afterthe SMR stimulation (Figure 5). The FG showed no significant increasing trend over time:13.7 ± 2.7 cm at T0; 14.6 ± 3.4 cm at T1; 14.5 ± 4.4 cm at T2; and 15.5 ± 4.1 cm at T3. TheAG showed a behavior very similar to the whole sample, with T0 being significantly lowerthan T1, T2, and T3: 5.9 ± 3.6 cm at T0; 8.1 ± 4.1 cm at T1 (p = 0.001 compared to T0);8.6 ± 4.3 cm at T2 (p = 0.003 compared to T0); and 9.0 ± 4.7 cm at T3 (p = 0.001 comparedto T0). Finally, SG showed a significant difference between T0 and T2 and between T1 andT2: −6.5 ± 4.3 cm at T0; −5.8 ± 4.5 cm at T1 (p = 0.038 compared to T0); −3.8 ± 3.9 cm atT2 (p = 0.010 compared to T0); and −3.4 ± 4.6 cm at T3.Int. J. Environ. Res. Public Health 2023, 20, x FOR PEER REVIEW before intervention. T1: test after intervention. T2: test 30 min after intervention. T3: test 60 mintervention. SMR: self-myofascial release. * Significant differences with p < 0.05. A primary stratification of the entire sample was made by dividing by sex. Thgroup showed a lower level of flexibility with respect to the female group at the T9.3 cm and 7.7 ± 6.6 respectively; p = 0.036), T1 (1.4 ± 9.7 cm and 9.5 ± 6.5 respectiv0.013), T2 (2.4 ± 8.8 cm and 10.1 ± 6.1 respectively; p = 0.010), and T3 (2.6 ± 9.4 cm a± 5.8 respectively; p = 0.009). The male group showed increasing flexibility over tino statistical differences were measured: 0.5 ± 9.3 cm at T0; 1.4 ± 9.7 cm at T1; 2.4 ± at T2; 2.6 ± 9.4 cm at T3. For the female group, significant differences were found bT0 and T1 (7.7 ± 6.6 and 9.5 ± 6.5 cm; p = 0.000), T0 and T2 (7.7 ± 6.6 and 10.1 ± 6.1 0.004), and T0 and T3 (7.7 ± 6.6 and 11.0 ± 5.8 cm; p = 0.000). Another significant diffwas found between T1 and T3 (9.5 ± 6.5 and 11.0 ± 5.8 cm; p = 0.016). The entire sample was also divided into three subgroups, namely the FG, ASG, according to the S&R values measured at T0. The three subgroups showed a cant difference for the S&R at T0 (p = 0.000). Obviously, the FG showed higher fle(13.7 ± 2.7 cm) than AG (5.9 ± 3.6 cm) and the same was for AG compared to SG (−6cm), respectively. The behavior of the subgroups over time was different after thstimulation (Figure 5). The FG showed no significant increasing trend over time: 13cm at T0; 14.6 ± 3.4 cm at T1; 14.5 ± 4.4 cm at T2; and 15.5 ± 4.1 cm at T3. The AG sa behavior very similar to the whole sample, with T0 being significantly lower thT2, and T3: 5.9 ± 3.6 cm at T0; 8.1 ± 4.1 cm at T1 (p = 0.001 compared to T0); 8.6 ± 4.T2 (p = 0.003 compared to T0); and 9.0 ± 4.7 cm at T3 (p = 0.001 compared to T0). FSG showed a significant difference between T0 and T2 and between T1 and T2: −6.cm at T0; −5.8 ± 4.5 cm at T1 (p = 0.038 compared to T0); −3.8 ± 3.9 cm at T2 (p compared to T0); and −3.4 ± 4.6 cm at T3. Figure 5. S&R modifications over the time before and after the SMR intervention in the thrgroups. SG: stiff group; AG: average group; FG: flexible group. T0: test before intervention. after intervention. T2: test 30 min after intervention. T3: test 60 min after intervention. SMmyofascial release. * Significant differences with p < 0.05. Figure 5. S&R modifications over the time before and after the SMR intervention in the threesubgroups. SG: stiff group; AG: average group; FG: flexible group. T0: test before intervention. T1:test after intervention. T2: test 30 min after intervention. T3: test 60 min after intervention. SMR:self-myofascial release. * Significant differences with p < 0.05.4. DiscussionWith this work, we sought to determine the longer terms effects of the posteriormuscular chain flexibility after practicing SMR on the plantar surface of both feet. Thiswas undertaken by performing the S&R test immediately, 30 min, and 60 min after theInt. J. Environ. Res. Public Health 2023, 20, 9748 of 12SMR intervention. The SMR effects were investigated also for individuals categorized bydifferent flexibility levels at baseline.4.1. Posterior Muscular Chain Flexibility, Changes in the Whole SampleThe main result of this research is relative to the duration of the SMR along an hourtime span and the entire sample showed a constant increase of the S&R values, withsignificant differences along the time. To date, this is the first study to measure the effect ofSMR intervention on posterior muscular chain flexibility across one hour. The increasingflexibility immediately after the intervention (T1) is consistent with the literature [20,31,47],confirming the transmission of information along the myofascial chains [49–51]. However,the most important novelty of this research dwells in the prolonged duration of the effect,lasting for a full hour. This result allows sports science practitioners and coaches to useSMR procedures from a new perspective. In fact, the facilitation given by this interventionduring training could accompany the practitioners for almost the entire training session,considering that a typical training session lasts 30–45 min or 60–120 min, respectively, forhealth or performance goals [52,53]. At the same time, it is fair to remember that during theone-hour test session, the participants did not perform physical activity, but they could onlysit or walk around the laboratory, unlike during a training session. Despite the differentconditions between the experiment and field activity, the results are very interesting froma practical point of view. Accordingly, it seems possible to perform SMR exercises underthe foot before the training session, even in the locker room, and then go directly into thefield. In fact, to give a practical use of these findings, it should be considered, for example,that less flexible athletes show more running economy [54,55], but at the same time, therunning economy is related to tendon length [56]. Because SMR works on connective tissueand fascia, it could offer more flexibility without affecting the running performance [4,57];although, some doubts remain about this relationship [58].4.2. Posterior Muscular Chain Flexibility, Changes for SubgroupsOur results show that the female group was more flexible than the male, which isconsistent with previous literature [59,60]. However, it is interesting to underline thebehavior over time of these two groups. In fact, the female group showed a behavioridentical to the entire sample, while no statistical differences were measured for the malegroup. A plausible explanation for the lack of statistical differences could be identifiedconsidering the higher data variation expressed by the male group. Accordingly, thecoefficient of variation (CV) of the male group is 19.8, which is very high compared to theentire sample (CV 2.0) or with the other subgroups (CV 0.9, 0.1, 0.6, 0.2 for female, SG, AG,FG, respectively).Despite the stratification of the sample based on sex, we also divided the entire dataset based on the flexibility levels of the participants. The S&R test values at T0 were used todistinguish three subgroups: SG, AG, and FG. The effect of the SMR intervention on eachgroup was different.FG did not show significant changes over time. Accordingly, individuals with highflexibility levels seem to not be affected by an SMR stimulation of the muscular chain. Thismight happen as such an SMR intervention can be considered an under-threshold stimulus,and probably the short duration of the SMR is not sufficient to evoke any modification inthis subgroup, which is like previous studies [29,30]. The question of the proper durationof the SMR, especially in a flexible cohort, is, however, still open [61], and more researchis needed. Nevertheless, the characteristics of the SMR approach make it reasonable tospeculate that such a procedure would not be indicated for more flexible individuals. Infact, it is well known that individuals that are more flexible answer differently to flexibilitytraining with respect to rigid individuals [62].AG and SG showed significant modifications over time, confirming the starting hy-pothesis, especially for SG. The behavior of AG and SG was totally different because AGshowed a more constant increasing ramp, while SG showed a delayed effect of the SMR.Int. J. Environ. Res. Public Health 2023, 20, 9749 of 12In the AG, it is possible to appreciate a strong acute effect protracting from T0 to T1and then a stable increase of S&R values from T2 and T3. Therefore, in this group, theinitial gain in terms of flexibility seems to be maintained constantly until T3.In the SG, the increasing values of the S&R test can be considered delayed because dif-ferences in flexibility occur after 30 min from the SMR intervention. This is very interestingfrom a practical and sport science point of view because the SG seems to need more time toallow the myofascial chain adaptation to the new stimulus. Such a finding could dependon the higher density and stiffness of the connective tissue, which is a possible reason toexplain the reduced posterior muscular chain flexibility [18,63]. In light of these results,one question is reasonable: could FG participants benefit from longer or more intense SMRsessions? Further research should investigate this aspect.Previous similar studies [29–31,46,47] did not perform any kind of stratification of theentire sample, and therefore the approach of this study could be considered a novelty in theSMR research field. According to these results, trainers and sports coaches should carefullycheck the timing of the application of SMR intervention with their users and athletes, whomust be identified based on their baseline flexibility levels. This could be the reason whyother researchers avoided including participants with hypermobility [29].4.3. Musculoskeletal and Fascial AspectsThe results of this research offer more information about the open discussion on fascialand muscular chains. How could it be possible that a plantar feet stimulation increasesthe S&R test performance by up to one hour? It is well-known that muscles are wrappedby connective tissue with different layers, the so-called fascial system. The fascia is bodywidespread, linking the skeletal muscles [49]. The fascial system has proprioceptive andnociceptive functions [64–67] and is innervated by mechanoreceptors [66]. When pressureor traction is applied, those may create a range of different responses that facilitate move-ment [29]. Because the plantar fascia is the most distal and caudal part of the posteriormuscular chain [32], proper pressure on the plantar surface stimulates the mechanore-ceptors of the fascia, allowing transitory facilitation of the whole chain [31], due to theviscoelastic properties of the fascial connective tissue.4.4. LimitationsThe lack of a control group could be seen as a limitation of this research, but accordingto the previous literature [29,31], it is well known that a difference exists between interven-tion and control groups. Therefore, the authors’ choice was to avoid splitting the wholesample into two smaller groups but to use the entire sample to increase the number ofparticipants. This allowed us to create subgroups based on their flexibility level. Futureresearch can be undertaken to add more information on the behavior of different flexibilitygroups by applying the same methods of this research but using a larger sample, consider-ing at least a control and an intervention group with the same flexibility level (SG, AG, FG).In fact, the sample size should be considered another limitation of the present study.In addition, the research is focused only on the flexibility of the entire posterior mus-cular chain over a one-hour span, but no data are available on the analytic flexibilitymodification (e.g., the flexibility of hamstring and calves) or ankle joint ROM. In fact, itcould be very interesting to repeat the same protocol and to test analytically even theflexibility of plantar muscles, calves, hamstring, and lower back. This would help under-stand which part of the posterior muscular chain is more affected by the SMR interventionunder the foot, and which part is more responsible for the flexibility improvement of thewhole chain.Finally, the last limitation of this research is the lack of information about the timeneeded to return to the flexibility baseline level. According to the results, after an hourfrom the intervention, the flexibility of the posterior muscular chain is still higher withrespect to the baseline. Therefore, it could be interesting to understand how much time isneeded to return to baseline values. Future research should investigate these aspects.Int. J. Environ. Res. Public Health 2023, 20, 97410 of 125. ConclusionsThe acute effects of the feet plantar surface SMR on S&R performance last up to onehour after the intervention. The magnitude of the modification in posterior muscularchain flexibility depends on the individual flexibility level, as the SMR does not seemto produce effects in a flexible cohort, while it works satisfactorily in individuals withaverage flexibility and who were stiff. Trainers and sports professionals should considerand opportunely use the duration of the main effect of up to one hour in their trainingprograms for health, fitness, and sports.Author Contributions: Conceptualization, L.R. and D.P.; methodology, L.R. and J.P.; software, S.P.,G.M.M. and E.L.; validation, L.R., F.C., R.D.G. and E.M.; formal analysis, L.R. and T.F.; investigation,L.R., D.P. and F.D.; resources, L.R. and G.M.M.; data curation, J.P. and T.F.; writing—original draftpreparation, D.P. and F.D.; writing—review and editing, L.R., E.M., S.P. and G.M.M.; visualization,R.D.G. and E.L; supervision, F.C. and J.P.; project administration, J.P. All authors have read andagreed to the published version of the manuscript.Funding: This research received no external funding.Institutional Review Board Statement: The study was conducted in accordance with the Declarationof Helsinki and approved by the Ethics Committee of OVIDIUS UNIVERISTY OF CONSTANTA(protocol code No. 126 din 18 March 2022).Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.Data Availability Statement: The data that support the findings of this study are available from thecorresponding author, upon reasonable request.Conflicts of Interest: The authors declare no conflict of interest.References1.Junker, D.; Stöggl, T. The Training Effects of Foam Rolling on Core Strength Endurance, Balance, Muscle Performance and Rangeof Motion: A Randomized Controlled Trial. J. Sports Sci. Med. 2019, 18, 229–238. [PubMed]2.Shariat, A.; Anastasio, A.T.; Soheili, S.; Rostad, M. Home-based fundamental approach to alleviate low back pain using myofascialrelease, stretching, and spinal musculature strengthening during the COVID-19 pandemic. Work 2020, 67, 11–19. [CrossRef]3.Shalamzari, M.H.; Minoonejad, H.; Seidi, F. The Effects of a Self-Myofascial Release Program on Isokinetic Hamstrings-to-Quadriceps Strength Ratio and Range of Motion of the Knee Joint Among Athletes With Hamstring Shortness. J. Sport Rehabil.2022, 31, 391–397. [CrossRef] [PubMed]4.Sulowska-Daszyk, I.; Skiba, A. The Influence of Self-Myofascial Release on Muscle Flexibility in Long-Distance Runners. Int. J.Environ. Res. Public Health 2022, 19, 457. [CrossRef] [PubMed]5.Russo, L.; Montagnani, E.; Buttari, D.; Ardigò, L.P.; Melenco, I.; Larion, A.; Migliaccio, G.M.; Padulo, J. Track Running Shoes: ACase Report of the Transition from Classical Spikes to “Super Spikes” in Track Running. Appl. Sci. 2022, 12, 10195. [CrossRef]6.Cheatham, S.W.; Stull, K.R. Knowledge of self-myofascial release among allied health students in the United States: A descriptivesurvey. J. Bodyw. Mov. Ther. 2018, 22, 713–717. [CrossRef]7.Ughreja, R.A.; Venkatesan, P.; Balebail Gopalakrishna, D.; Singh, Y.P. Effectiveness of myofascial release on pain, sleep, andquality of life in patients with fibromyalgia syndrome: A systematic review. Complement. Ther. Clin. Pract. 2021, 45, 101477.[CrossRef]8.López-Torres, O.; Mon-López, D.; Gomis-Marzá, C.; Lorenzo, J.; Guadalupe-Grau, A. Effects of myofascial release or self-myofascial release and control position exercises on lower back pain in idiopathic scoliosis: A systematic review. J. Bodyw. Mov.Ther. 2021, 27, 16–25. [CrossRef]9.Beardsley, C.; Škarabot, J. Effects of self-myofascial release: A systematic review. J. Bodyw. Mov. Ther. 2015, 19, 747–758. [CrossRef]10.Kalichman, L.; Ben David, C. Effect of self-myofascial release on myofascial pain, muscle flexibility, and strength: A narrativereview. J. Bodyw. Mov. Ther. 2017, 21, 446–451. [CrossRef]11.McKenney, K.; Elder, A.S.; Elder, C.; Hutchins, A. Myofascial release as a treatment for orthopaedic conditions: A systematicreview. J. Athl. Train. 2013, 48, 522–527. [CrossRef] [PubMed]12.Chaitow, L. Research in water and fascia. Micro-tornadoes, hydrogenated diamonds and nanocrystals. Massage Today 2009, 9, 1–3.13.Aguilera, F.J.; Martin, D.P.; Masanet, R.A.; Botella, A.C.; Soler, L.B.; Morell, F.B. Immediate effect of ultrasound and ischemiccompression techniques for the treatment of trapezius latent myofascial trigger points in healthy subjects: A randomizedcontrolled study. J. Manip. Physiol. Ther. 2009, 32, 515–520. [CrossRef] [PubMed]14.Lavelle, E.D.; Lavelle, W.; Smith, H.S. Myofascial trigger points. Anesthesiol. Clin. 2007, 25, 841–851. [CrossRef]Int. J. Environ. Res. Public Health 2023, 20, 97411 of 1215.Montañez-Aguilera, F.J.; Valtueña-Gimeno, N.; Pecos-Martín, D.; Arnau-Masanet, R.; Barrios-Pitarque, C.; Bosch-Morell, F.Changes in a patient with neck pain after application of ischemic compression as a trigger point therapy. J. Back Musculoskelet.Rehabil. 2010, 23, 101–104. [CrossRef]16.MacDonald, G.Z.; Penney, M.D.; Mullaley, M.E.; Cuconato, A.L.; Drake, C.D.; Behm, D.G.; Button, D.C. An acute bout ofself-myofascial release increases range of motion without a subsequent decrease in muscle activation or force. J. Strength Cond.Res. 2013, 27, 812–821. [CrossRef]17.Schleip, R.; Müller, D.G. Training principles for fascial connective tissues: Scientific foundation and suggested practical applica-tions. J. Bodyw. Mov. Ther. 2013, 17, 103–115. [CrossRef]18.Behm, D.G.; Wilke, J. Do Self-Myofascial Release Devices Release Myofascia? Rolling Mechanisms: A Narrative Review. SportsMed. 2019, 49, 1173–1181. [CrossRef]19.Cornell, D.J.; Ebersole, K.T. Influence of an acute bout of self-myofascial release on knee extension force output and electro-mechanical activation of the quadriceps. Int. J. Sports Phys. Ther. 2020, 15, 732–743. [CrossRef]20.Ferreira, R.M.; Martins, P.N.; Goncalves, R.S. Effects of Self-myofascial Release Instruments on Performance and Recovery: AnUmbrella Review. Int. J. Exerc. Sci. 2022, 15, 861–883.21.Russo, L.; Palermi, S.; Dhahbi, W.; Kalinski, S.D.; Bragazzi, N.L.; Padulo, J. Selected components of physical fitness in rhythmicand artistic youth gymnast. Sport Sci. Health 2021, 17, 415–421. [CrossRef]22.Maniar, N.; Shield, A.J.; Williams, M.D.; Timmins, R.G.; Opar, D.A. Hamstring strength and flexibility after hamstring straininjury: A systematic review and meta-analysis. Br. J. Sports Med. 2016, 50, 909–920. [CrossRef] [PubMed]23.Sadler, S.G.; Spink, M.J.; Ho, A.; de Jonge, X.J.; Chuter, V.H. Restriction in lateral bending range of motion, lumbar lordosis,and hamstring flexibility predicts the development of low back pain: A systematic review of prospective cohort studies. BMCMusculoskelet. Disord. 2017, 18, 179. [CrossRef] [PubMed]24.Hori, M.; Hasegawa, H.; Takasaki, H. Comparisons of hamstring flexibility between individuals with and without low back pain:Systematic review with meta-analysis. Physiother. Theory Pract. 2021, 37, 559–582. [CrossRef] [PubMed]25.Thomas, E.; Bianco, A.; Paoli, A.; Palma, A. The Relation Between Stretching Typology and Stretching Duration: The Effects onRange of Motion. Int. J. Sports Med. 2018, 39, 243–254. [CrossRef] [PubMed]26.Zhang, Q.; Trama, R.; Fouré, A.; Hautier, C.A. The Immediate Effects of Self-Myofacial Release on Flexibility, Jump Performanceand Dynamic Balance Ability. J. Hum. Kinet. 2020, 75, 139–148. [CrossRef]27.Kasahara, K.; Konrad, A.; Yoshida, R.; Murakami, Y.; Sato, S.; Aizawa, K.; Koizumi, R.; Thomas, E.; Nakamura, M. Comparisonbetween 6-week foam rolling intervention program with and without vibration on rolling and non-rolling sides. Eur. J. Appl.Physiol. 2022, 122, 2061–2070. [CrossRef]28.Beier, Z.; Earp, I.; Korak, J.A. Self-Myofascial Release Does Not Improve Back Squat Range of Motion, Alter Muscle Activation, orAid in Perceived Recovery 24-Hours Following Lower Body Resistance Training. Int. J. Exerc. Sci. 2019, 12, 839–846.29.Grieve, R.; Goodwin, F.; Alfaki, M.; Bourton, A.J.; Jeffries, C.; Scott, H. The immediate effect of bilateral self myofascial release onthe plantar surface of the feet on hamstring and lumbar spine flexibility: A pilot randomised controlled trial. J. Bodyw. Mov. Ther.2015, 19, 544–552. [CrossRef]30.Williams, W.; Selkow, N.M. Self-Myofascial Release of the Superficial Back Line Improves Sit-and-Reach Distance. J. Sport Rehabil.2019, 29, 400–404. [CrossRef]31.Fauris, P.; López-de-Celis, C.; Canet-Vintró, M.; Martin, J.C.; Llurda-Almuzara, L.; Rodríguez-Sanz, J.; Labata-Lezaun, N.; Simon,M.; Pérez-Bellmunt, A. Does Self-Myofascial Release Cause a Remote Hamstring Stretching Effect Based on Myofascial Chains? ARandomized Controlled Trial. Int. J. Environ. Res. Public Health 2021, 18, 12356. [CrossRef] [PubMed]32.Prabu Raja, G.; Shyamasunder Bhat, N.; Marie Cruz, A.; Prabhu, A.; Fernandes, S.; Naaz, N. The anatomical myofascial continuumbetween the neck and eyes. Clin. Anat. 2022, 35, 340–346. [CrossRef]33.Hamaoui, A.; le Bozec, S.; Poupard, L.; Bouisset, S. Does postural chain muscular stiffness reduce postural steadiness in a sittingposture? Gait Posture 2007, 25, 199–204. [CrossRef] [PubMed]34.Russo, L.; Giustino, V.; Toscano, R.E.; Secolo, G.; Secolo, I.; Iovane, A.; Messina, G. Can tongue position and cervical ROM affectpostural oscillations? A pilot and preliminary study. J. Hum. Sport Exerc. 2020, 15, S840–S847. [CrossRef]35.Halperin, I.; Aboodarda, S.J.; Button, D.C.; Andersen, L.L.; Behm, D.G. Roller massager improves range of motion of plantarflexor muscles without subsequent decreases in force parameters. Int. J. Sports Phys. Ther. 2014, 9, 92–102. [PubMed]36.Jay, K.; Sundstrup, E.; Søndergaard, S.D.; Behm, D.; Brandt, M.; Særvoll, C.A.; Jakobsen, M.D.; Andersen, L.L. Specific and crossover effects of massage for muscle soreness: Randomized controlled trial. Int. J. Sports Phys. Ther. 2014, 9, 82–91. [PubMed]37.Liemohn, W.; Sharpe, G.L.; Wasserman, J.F. Criterion related validity of the sit-and-reach test. J. Strength Cond Res. 1994, 8, 91–94.[CrossRef]38.Russo, L.; Benis, R.; Livi, S.; Falcone, A.; Ragalmuto, N.; Bartolucci, P.; Barni, L. Esercizio Correttivo®Postura Salute Perfomance;ATS—Giacomo Catalani Editore: Arezzo, Italy, 2016; p. 226.39.Russo, L.; Di Capua, R.; Arnone, B.; Borrelli, M.; Coppola, R.; Esposito, F.; Padulo, J. Shoes and Insoles: The Influence on MotorTasks Related to Walking Gait Variability and Stability. Int. J. Environ. Res. Public Health 2020, 17, 4569. [CrossRef]40.Ardigò, L.P.; Buglione, A.; Russo, L.; Cular, D.; Esposito, F.; Doria, C.; Padulo, J. Marathon shoes vs. track spikes: A crossoverpilot study on metabolic demand at different speeds in experienced runners. Res. Sports Med. 2021, 1–8. [CrossRef]41.Atkinson, G.; Reilly, T. Circadian variation in sports performance. Sports Med. 1996, 21, 292–312. [CrossRef]Int. J. Environ. Res. Public Health 2023, 20, 97412 of 1242.Mayorga-Vega, D.; Merino-Marban, R.; Viciana, J. Criterion-Related Validity of Sit-and-Reach Tests for Estimating Hamstring andLumbar Extensibility: A Meta-Analysis. J. Sports Sci. Med. 2014, 13, 1–14. [PubMed]43.Lemmink, K.A.P.M.; Greef, M.H.G.; Rispens, P.; Kemper, H.C.G.; Stevens, M. The validity of the sit-and-reach test and themodified sit-and-reach test in middle-aged to older men and women. Res. Q. Exerc. Sport 2003, 74, 331–336. [CrossRef] [PubMed]44.Kaminsky, L.A.; Bonzheim, K.A. American College of Sports Medicine (ACSM) Resource Manual for Guidelines for Exercise Testing andPrescription, 5th ed.; Lippincott Williams & Wilkins: Philadelphia, PA, USA, 2006.45.Gonzalez-Suarez, C.B.; Caralipio, N.; Gambito, E.; Reyes, J.J.; Espino, R.V.; Macatangay, R. The association of physical fitness withbody mass index and waist circumference in Filipino preadolescents. Asia Pac. J. Public Health 2013, 25, 74–83. [CrossRef]46.Joshi, D.G.; Balthillaya, G.; Prabhu, A. Effect of remote myofascial release on hamstring flexibility in asymptomatic individuals—Arandomized clinical trial. J. Bodyw. Mov. Ther. 2018, 22, 832–837. [CrossRef]47.Wilke, J.; Kalo, K.; Niederer, D.; Vogt, L.; Banzer, W. Gathering Hints for Myofascial Force Transmission under In Vivo Conditions:Are Remote Exercise Effects Age Dependent? J. Sport Rehabil. 2019, 28, 758–763. [CrossRef]48.Štefan, L.; Kasovi´c, M.; Culej, M. Normative Values for Health-Related Physical Fitness in First-Year Police Officers. J. StrengthCond. Res. 2022, 36, 2530–2535. [CrossRef]49.Wilke, J.; Krause, F.; Vogt, L.; Banzer, W. What Is Evidence-Based About Myofascial Chains: A Systematic Review. Arch. Phys.Med. Rehabil. 2016, 97, 454–461. [CrossRef]50.Krause, F.; Wilke, J.; Vogt, L.; Banzer, W. Intermuscular force transmission along myofascial chains: A systematic review. J. Anat.2016, 228, 910–918. [CrossRef]51.Gabriel, A.; Konrad, A.; Roidl, A.; Queisser, J.; Schleip, R.; Horstmann, T.; Pohl, T. Myofascial Treatment Techniques on the PlantarSurface Influence Functional Performance in the Dorsal Kinetic Chain. J. Sports Sci. Med. 2022, 21, 13–22. [CrossRef]52.Seiler, S. What is best practice for training intensity and duration distribution in endurance athletes? Int. J. Sports Physiol. Perform2010, 5, 276–291. [CrossRef]53.La Greca, S.; Rapali, M.; Ciaprini, G.; Russo, L.; Vinciguerra, M.G.; Di Giminiani, R. Acute and Chronic Effects of SupervisedFlexibility Training in Older Adults: A Comparison of Two Different Conditioning Programs. Int. J. Environ. Res. Public Health2022, 19, 16974. [CrossRef] [PubMed]54.Jones, A.M. Running economy is negatively related to sit-and-reach test performance in international-standard distance runners.Int. J. Sports Med. 2022, 23, 40–43. [CrossRef] [PubMed]55.Trehearn, T.L.; Buresh, R.J. Sit-and-reach flexibility and running economy of men and women collegiate distance runners. J.Strength Cond. Res. 2009, 23, 158–162. [CrossRef]56.Hunter, G.R.; Katsoulis, K.; McCarthy, J.P.; Ogard, W.K.; Bamman, M.M.; Wood, D.S.; Den Hollander, J.A.; Blaudeau, T.E.;Newcomer, B.R. Tendon length and joint flexibility are related to running economy. Med. Sci. Sports Exerc. 2011, 43, 1492–1499.[CrossRef] [PubMed]57.Stroiney, D.A.; Herrick, S.L.; Paolone, V.J.; Matthews, T.D. The Effects of an Acute Bout of Self-Myofascial Release on thePhysiological Parameters of Running. Int. J. Exerc. Sci. 2020, 13, 113–122. [CrossRef] [PubMed]58.Giovanelli, N.; Vaccari, F.; Floreani, M.; Rejc, E.; Copetti, J.; Garra, M.; Biasutti, L.; Lazzer, S. Short-Term Effects of Rolling Massageon Energy Cost of Running and Power of the Lower Limbs. Int. J. Sports Physiol. Perform. 2018, 13, 1337–1343. [CrossRef]59.Höög, S.; Andersson, E.P. Sex and Age-Group Differences in Strength, Jump, Speed, Flexibility, and Endurance Performances ofSwedish Elite Gymnasts Competing in TeamGym. Front. Sports Act. Living 2021, 3, 653503. [CrossRef]60.Lopes, T.J.A.; Simic, M.; Alves, D.S.; Bunn, P.D.S.; Rodrigues, A.I.; Terra, B.S.; Lima, M.D.S.; Ribeiro, F.M.; Vilão, P.; Pappas, E.Physical Performance Measures of Flexibility, Hip Strength, Lower Limb Power, and Trunk Endurance in Healthy Navy Cadets:Normative Data and Differences Between Sex and Limb Dominance. J. Strength Cond. Res. 2021, 35, 458–464. [CrossRef]61.Phillips, J.; Diggin, D.; King, D.L.; Sforzo, G.A. Effect of Varying Self-myofascial Release Duration on Subsequent AthleticPerformance. J. Strength Cond. Res. 2021, 35, 746–753. [CrossRef]62.Agopyan, A.; Tekin, D.; Unal, M.; Kurtel, H.; Turan, G.; Ersoz, A. Acute effects of static stretching on isokinetic thigh strength onmodern dancers. J. Sports Med. Phys. Fitness 2013, 53, 538–550.63.Pavan, P.G.; Stecco, A.; Stern, R.; Stecco, C. Painful connections: Densification versus fibrosis of fascia. Curr. Pain Headache Rep.2014, 18, 441. [CrossRef]64.Stecco, C.; Porzionato, A.; Macchi, V.; Tiengo, C.; Parenti, A.; Aldegheri, R.; Delmas, V.; De Caro, R. Histological characteristics ofthe deep fascia of the upper limb. Ital. J. Anat. Embryol. 2006, 111, 105–110. [PubMed]65.Stecco, C.; Gagey, O.; Belloni, A.; Pozzuoli, A.; Porzionato, A.; Macchi, V.; Aldegheri, R.; De Caro, R.; Delmas, V. Anatomy of thedeep fascia of the upper limb. Second part: Study of innervation. Morphologie 2007, 91, 38–43. [CrossRef]66.Stecco, C.; Porzionato, A.; Lancerotto, L.; Stecco, A.; Macchi, V.; Day, J.A.; De Caro, R. Histological study of the deep fasciae of thelimbs. J. Bodyw. Mov. Ther. 2008, 12, 225–230. [CrossRef] [PubMed]67.Stecco, A.; Gesi, M.; Stecco, C.; Stern, R. Fascial components of the myofascial pain syndrome. Curr. Pain Headache Rep. 2013, 17,352. [CrossRef]Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individualauthor(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury topeople or property resulting from any ideas, methods, instructions or products referred to in the content.View publication stats | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/354435246Deep Convolutional Neural Networks Predict Elasticity Tensors and their Boundsin HomogenizationPreprint · September 2021CITATIONS0READS6101 author:Some of the authors of this publication are also working on these related projects:Deep Learning for Microstructure-Property Linkages View projectReconstruction, Homogenization and Deformation Analysis of Diamond/β-SiC Composite Films View projectBernhard EidelTechnische Universität Bergakademie Freiberg88 PUBLICATIONS 832 CITATIONS SEE PROFILEAll content following this page was uploaded by Bernhard Eidel on 08 September 2021.The user has requested enhancement of the downloaded file.Deep Convolutional Neural Networks PredictElasticity Tensors and their Bounds in HomogenizationBernhard EidelDFG-Heisenberg-Fellow, Institute of Mechanics, Department Mechanical EngineeringUniversity Siegen, 57068 Siegen, Paul-Bonatz-Str. 9-11, Germany∗e-mail: bernhard.eidel@uni-siegen.de, phone: +49 271 740 2224AbstractIn the present work, 3D convolutional neural networks (CNNs) are trained to link randomheterogeneous, two-phase materials of arbitrary phase fractions to their elastic macroscale stiff-ness thus replacing explicit homogenization simulations. In order to reduce the uncertainty ofthe true stiffness of the synthetic composites due to unknown boundary conditions (BCs), theCNNs predict the stiffness for periodic BCs, its upper bound through kinematically uniformBCs, and its lower bound through stress uniform BCs. The workflow of the homogenization-CNN is described, from the microstructure generation over the CNN design, the operations ofconvolution, nonlinear activation and pooling as well as optimization up to performance mea-surements in tests. The CNN predictions are very accurate even for microstructure samples ofthe two-phase diamond-SiC coating material. The CNN that covers all three BCs is virtuallyas accurate as the separate treatment in three different nets. The CNNs of this contributionprovide through stiffness bounds an indicator of the proper RVE size for individual snapshotsamples. Moreover, they enable statistical analyses for the effective elastic stiffness on ensemblesof synthetical microstructures without costly simulations.Keywords: Deep learning; Convolutional neural networks; Homogenization; Structure-property relations; Solid mechanics1IntroductionAs visual perception in the evolution of mammals is intimately related to the developmentof their cognitive abilities, mimicking some structural and process properties of visualperception has boosted artificial intelligence (AI) in its deep learning (DL) advancementsby convolutional neural networks (CNN), the latter compared to the former in the blinkof an eye. CNNs in image classification, still one of the major field of their application,disclose their insights in terms of their feature maps in the freely accessible, so-calledhidden layers. Not only what they have learned, for instance to identify a dog in an image,but also how they have learned it, namely, by recognizing the nose, that largely makes adog a dog in the image classification of [53]. The broad range of relevant applications ofobject recognition (most notably in large images) render CNNs important, the disclosureof learned lessons in an image language renders them fascinating for humans, since thesense of vision is likely their strongest.For a thorough introduction to deep-learning we refer to [18], for CNNs in particular to[18], Chap. 9, and [17]. A comprehensive overview for ANNs with numerous referencesto original work until 2015 is given in [41], recent advances in architectures of CNNs arearXiv:2109.03020v1 [cond-mat.mtrl-sci] 4 Sep 2021Deep CNNs predict elasticity tensors with bounds in homogenization2discussed in [28]. For a recent presentation of the mathematics of deep learning with rootsin learning theory [8] we refer to [6].The aim of the present work is to construct and train convolutional neural networks forpredictions of the macroscopic elastic properties and their bounds for two-phase compos-ites at arbitrary phase fractions and a wide range of morphologies.Composites in their general definition are multiphase materials that consist of at leasttwo different constituents. When the features of their morphology live on much smallerlength scales than the structure of interest at large, computational homogenization isinstrumental for its favorable accuracy and high efficiency of microstructure sampling incomparison to full numerical simulation (FNS). In homogenization, the appropriate sizeof a microdomain in the case of non-periodic, random microstructures refers to the notionof representative volume elements (RVE) and effective properties.According to Hill [22] the microstructural features of the RVE must be statistical repre-sentative for the heterogeneous material and large enough to be insensitive to the appliedboundary conditions which fulfill the requirement of equal, micro-macro energy density.The reference to the statistics of a particular property instead of all (material) propertiesfor the definition of the RVE was introduced by Drugan and Willis [9] resulted in smallervolumes matching the RVE requirement. The consideration of multiple sample and theirstatistics of apparent properties for the definition of effective properties was introducedby Kanit et al. [26] and Jeulin [24]; it was shown that increasing a the samples’ volumedecreases both the bias, the deviation of the mean stiffness for different BCs, and thestatistical variance.Statistical analyses for effective properties are typically based on randomly generated, syn-thetic samples, since real 3D microstructures from tomography and image acquisition areexpensive and therefore limited in numbers. For suchlike snapshot samples the criterionfor the RVE and effective properties is based on simulation results for different boundaryconditions (BCs), periodic (PBC), kinematical uniform (KUBC), and stress/static uni-form (SUBC) boundary conditions. According to Huet [23] stiffness deviations betweendifferent BCs render the properties ”apparent”, and ”effective”, if the size is sufficientlylarge for stiffness insensitivity to the applied BCs. Then, the VE can be considered anRVE. For effective properties the inequalities among stiffnesses for different boundaryconditions C(SUBC) ≤ C(PBC) ≤ C(KUBC) turn into equalities1 where the relationsbetween positive fourth-order tensors are understood in terms of quadratic forms.Building links from microstructures to macroscopic properties by CNNs is a newly emerg-ing but very active field of research in machine learning. The type of microstructuresranges from two-phase binaries created by Gaussian filters [51], over spherical and ellip-soidal inclusions [38] to short fibre composites [7]. A common denominator of all studiesis the application of PBC, typically along with an a posteriori comparison with analyticalbounds, of Voigt and Reuss in [51, 38] and with the Mori-Tanaka method in [7]. Here,sharper bounds by KUBC and SUBC would enrich the CNN predictions with respect tothe required RVE size and effective properties.Novel contribution.One novel aspect of the present work is to include stiffnessbounds for PBC by SUBC and KUBC directly into the CNN instead of an a posteriori1Even for an equality it might turn out that the variance does not vanish, if the statistical distributionwas available.Bernhard Eidel3comparison with analytical bounds mentioned above. The question arises whether onesingle CNN for the augmented output of three elasticity tensors can catch up with theaccuracy of three distinct CNNs. Here, we venture into unchartered terrain in view of exist-ing results, which have considered one single elasticity tensor at maximum, [51, 49, 38, 7].Work on CNNs in the related fields of predicting effective conductivities or permeabilitiescannot provide answers, since the number of output parameters was restricted to one upto three scalar quantities; to permeability from porous image/volume data [48, 43, 25], toeffective diffusivity [47], to ionic conductivity [30], and to the triple of porosity, perme-ability, and tortuosity [20].Including the stiffness bounds into the CNN predictions enables an interesting quanti-tative picture already for the full set of training data which exhibits a broad range ofmorphologies, arbitrary phase fractions at a stiffness contrast of factor 50; how sensitiveto the applied BCs are the samples? Which microstructures show the strongest sensitivity,which ones the least?A second, novel and challenging aspect is that the CNNs shall be assessed not only forthe microstructures stemming from the same generator as for training and validation, butalso for the real, two-phase microstructure of a diamond/β-SiC composite. This materialfabricated by chemical vapor deposition (CVD) is of interest for its outstanding mechanicalproperties such as hardness and wear-resistance. Therefore, diamond/β-SiC is used, e.g.,as protective coatings for vulnerable substrates, but also in biological applications [50].Since the CNNs have seen on their training track only randomly generated syntheticmicrostructures, but not any sample of this real material, the test goes beyond existinganalyses of homogenization CNNs [51, 38, 7, 34] in that it addresses the intricate problemto generalize from synthetic to real microstructures without statistical links.2Generation of Two-Phase MicrostructuresThe set of cubic volume elements (VE) with different, two-phase microstructures servesas the input of the CNN. The generation of the VEs is carried out in three steps; (i) arandom color code is assigned to each voxel in the VE of edge length 100 voxels, (ii) aGaussian filter is applied to the voxel number field, (iii) a binarization is carried out witha randomly selected phase fraction between 0 and 100%. The number of VEs is M = 104.The three steps shall be briefly described. The 3D Gaussian is given by the equationN(x, µ, C) =1�(2π)3|C|e−12 (x − µ)T C−1 (x − µ) ,(1)where x = [x y z]T contains the coordinates of the considered voxel, and µ represents themean of the Gaussian µ = [µx µy µz]T. The symmetric (3 × 3)-covariance matrix CC =sxsxysxzsysyzsymm.sz(2)embodies variances sx, sy, sz on the diagonal, and covariances sxy, sxz and syz off-diagonal.We consider values for the variances sx/y/z ∈ [0.5, 8] along with zero covariances.Deep CNNs predict elasticity tensors with bounds in homogenization4The choice of the interval for variances and the phase fractions generate a large variety ofdifferent microstructure characteristics. Some of them are exemplarily visualized in Fig. 1.(a) 0.55/0.65/0.99(b) 2.31/2.19/2.00(c) 5.85/5.85/5.92(d) 7.73/7.77/7.79(e) 0.53/7.98/7.92(f) 7.71/0.62/7.87(g) 0.51/0.55/5.77(h) 7.09/0.65/0.60(i) 1.23/7.51/4.73(j) 0.5/5.80/5.94(k) 1.17/5.04/7.74(l) 7.89/7.78/5.23Figure 1: Two-phase microstructures. Snapshot samples for different phase fractionsindicate, how the variances (values of sx/sy/sz in the captions) of the Gaussian filterdetermine the type of microstructure independent of the largely varying phase fractions.A homogeneous dispersion of the phases is induced by values of similar magnitude fordifferent variances as displayed in the Subfigs. (a)–(d). The magnitude of the variancesdetermines the granularity of the microstructure. For fine dispersions, the apparent elasticproperties can be expected to be isotropic or close to it. Pronounced lamellar structuresare generated, when two variances are of the same magnitude and larger than the thirdone, (e), (f), which renders elastic properties close to orthotropic or cubic symmetry.Pronounced material orientations into a particular i-direction of the coordinate systemare induced by a large variance value si at smaller values for the other variances (g), (h),which results in an elastic behavior close to transversal isotropy. Triples of variances ofdifferent size generate more or less pronounced anisotropies, some examples are given in(i)–(l).Bernhard Eidel53Computational Homogenization with BoundsThe elasticity tensors from homogenization simulations render the individual labels of theVEs (in total referred to as ground truth) for the latter training of the CNNs.Computational homogenization in terms of two-scale finite element methods solves in eachintegration point of a macro element the associated microproblem; prominent examplesare the FE2 method [35, 36, 14, 31, 37, 42, 40] and the FE-HMM [4, 2, 3]. For the caseof the first-order strain-driven computational homogenization the macroproblem drivesthe boundary value problem (BVP) of the microscale RVE by macroscopic deformationand, vice-versa, the microproblem provides (elastic) stiffness in terms of the macroscopictangent and averaged microstresses, which implies that a constitutive law merely existson the microscale.To put things into perspective, the displacement field u which is the solution to the balanceof linear momentum on the microscale div σ = 0 can be decomposed into displacementsinduced by a homogeneous infinitesimal strain ε and superimposed microscale fluctuations�w according to u = ε x = εx + �w as sketched in Fig. 2. The corresponding additivedecomposition of strain and stress read as ε = ε+ �ε and σ = σ+ �σ, where the macrostressσ is calculated as the volumetric mean of the microstresses σ = 1/V�RVE σ dV .The Hill-Mandel or macrohomogeneity condition [22] as the theoretical cornerstone ofcomputational homogenization postulates the equality of the macroscale stress powerwith the average stress power in the RVEσ : ˙ε = 1V�RVEσ : ˙ε dV←→1V�VEσ : ˙ε dV − σ : ˙ε = 0 .(3)3.1Energetically consistent coupling conditionsFigure 2: Coupling conditions. Periodic coupling conditions on the boundary ∂RVEwith the outward unit normal n of the RVE are displayed as solid lines in the deformedconfiguration, for linear Dirichlet coupling with �w ≡ 0 as dashed lines.Constraint conditions consistent to the Hill-Mandel postulate are ˙ε := ˙ε on the wholeRVE (Voigt condition) and σ = σ on the whole RVE (Reuss condition). Moreover, lin-ear Dirichlet, linear Neumann as well as periodic boundary conditions equally fulfill thecondition. They can be obtained from the following results.Deep CNNs predict elasticity tensors with bounds in homogenization6�RVEσ : ˙ε dV = σ : ˙ε +�RVE(σ + �σ) : ˙�ε dV = σ : ˙ε +�RVE�σ : ˙�ε dV .(4)Inserting (4) into (3), applying the Gauss theorem with the surface tractions t on theboundary ∂RVE results in�RVE�σ : ˙�ε dV =�∂RVE�t : ˙�w dA =�∂RVE(t − σn) · ( ˙u − ˙εx) dA = 0 ,(5)from which the BCs are obtained satisfying the macro-homogeneity condition.1. Kinematically uniform boundary conditions (KUBC)/Dirichlet BCs˙u(x) = ˙ε x∀x ∈ ∂RVE .(6)2. Stress uniform boundary conditions (SUBC)/Neumann BCst(x) = σ n∀x ∈ ∂RVE .(7)3. Periodic boundary conditions (PBC)For the definition of PBC, the boundary ∂RVE is split into pairwise periodic parts∂RVE+ and ∂RVE− with a corresponding sign convention for position vectors x+and x− and outward unit normal vectors n+ = −n− as visualized in Fig. 2. Periodicfluctuation displacements �w+ = �w− = �w along with (5) result in�∂RVE�t· ˙�w dA =�∂RVE+�t+ · ˙�w dA+�∂RVE−�t− · ˙�w dA =�∂RVE−(�t+ + �t−)· ˙�w dA = 0 , (8)which implies the additional condition �t+ = −�t−; in conclusion, it holds�w+ = �w−and�t+ = −�t−∀x ∈ ∂RVE .(9)3.2Stiffness relationsDifferent BCs result in different moduli in Hooke’s law of linear elasticity σ = C ε; theyare related by inequalities that turn into equalities for effective propertiesC(SUBC) ≤ C(PBC) ≤ C(KUBC) .(10)The inequalities in (10) between positive fourth-order tensors are understood in terms ofquadratic forms, i.e. Ca ≥ Cb ⇔ ε : Ca ε ≥ ε : Cb ε ∀ ε .Volumes of heterogeneous materials with arbitrary microstructure are rarely periodic andtherefore do rarely deform according to PBC. For non-periodicity, the true stiffness de-pends on the continuation of the considered microdomain in all three directions of space.If the stiffness response considerably depends on the applied BCs and the spatial contin-uation of the VE is available, the postulate for representativeness suggests to increase theVE size. If the VE can not be increased in size, the true stiffness remains uncertain. TheBernhard Eidel7uncertainty however is bounded, SUBC and KUBC provide lower and upper bounds tothe true apparent stiffness. These bounds do not necessarily bound the effective stiffness.For the VEs computational homogenization is carried out for the three above mentionedenergetically consistent boundary conditions 1.–3.. The corresponding homogenized elas-ticity tensors C(KUBC), C(PBC), and C(SUBC) are obtained by the microsolver of theFinite Element Heterogeneous Multiscale Method (FE-HMM). For the mathematical foun-dation of FE-HMM in linear elasticity we refer to [1] for a finite element formulation andaspects of implementation to [11]. The boundary conditions in the homogenization simu-lations are fulfilled through the Lagrange multiplier method as described in [11] for PBC,for KUBC and by a novel, simple and efficient formulation for SUBC in [15].3.3Range of validity and limitationsFor the material behavior we assume linear elasticity. For general triclinic materials, hencewithout any symmetry, the elasticity matrix exhibits 21 independent moduli. The matrixform of Hooke’s law reads in Voigt notationσ11σ22σ33σ12σ23σ13=C11C12C13C14C15C16C22C23C24C25C26C33C34C35C36C44C45C46C55C56sym.C66����Elastic stiffness matrix Cε11ε22ε332ε122ε232ε13(11)The CNN is trained to predict 9 components which implies the treatment of elastic cou-pling effects (CE) as follows:CE-1 Considered are the direct effect of normal strains on normal stresses throughC11, C22, C33, the direct effect of shear strains on normal stresses throughC44, C55, C66, and the Poisson’s effect through C12, C13, C23.CE-2 Not considered are coupling effects of normal strains on shear stresses throughCij, i = 1, 2, 3, j = 4, 5, 6 (in blue color in (11)).CE-3 Not considered are the coupling effect of shear strain on shear stress in perpendic-ular planes (corresponding to the Poisson’s effect for normal stress and strain inperpendicular directions) through C45, C46, C56 (in red color in (11)).Considering the above 9 elasticity components and discarding the others is based on twoassumptions related to the material symmetries and their representation, [44], [45]:Asm-1 The considered microstructures are restricted to materials of orthotropic, tetragonal,transversely isotropic or cubic symmetry or are even isotropic.Asm-2 The anisotropic materials with the symmetries of Asm-1 are presented and analyzedin their symmetry basis.Deep CNNs predict elasticity tensors with bounds in homogenization8The symmetry classes of Asm-1 share the same skyline of zero and nonzero entries inthe elasticity matrix in the symmetry basis system, but differ in dependencies of theelasticity moduli. In other generic bases the above material classes exhibit in generala fully populated elasticity matrix, where those coupling effects, which vanish in thesymmetry basis, are active. A post-processing of the CNN prediction identifying thesedependencies can inform about the particular symmetry class.The limitation of the present approach in the format of the elasticity matrix is consistentwith the type of generated microstructures, since the variation of the variances implysymmetry planes that are parallel to the planes of the generating coordinate system orobtained by an additional axis rotation.43D Convolutional Neural NetworksFigure 3: 3D-CNN architecture for homogenization with bounds. From the inputlayer of a two-phase microstructure to the output layer of elasticity components for the cou-pling conditions of KUBC, PBC, and SUBC, in between a stack of convolutional layers andfully connected layers (FC). Each convolutional layer contains the operations of activation,nonlinear activation (here by ReLU) and pooling. The transition to the FC block requires aflattening of 3D arrays into a vector.4.1GeneralThe architecture of the present 3D-CNN for homogenization is displayed in Fig. 3 andbriefly described in the caption. The input layer of the CNN consists of a random variableX, the voxel data of the VE, the output layer of a random variable Z, the elastic moduli.The output is the result of an input-output mapping Z = Z(X, Θ) depending on theparameters Θ = (w, b) of the weights w and the biases b. In an approach of supervisedlearning the net is trained to predict the target variable Y , also referred to as label, whichis equally random since the homogenization computation operates on X, Y = Y (X).Since the error function (also called loss or cost function) measures the distance betweenY and Z, a reasonable choice for the loss is given by the expectation E of the squaredBernhard Eidel9difference of Y and ZL(Θ) = E[(Y − Z)2] = E[(Y (X) − Z(X, Θ))2] .(12)The objective of training is finding the parameters Θ for which the cost function becomesminimal, hence Θ∗ = arg minΘ L(Θ). In practice, optimization is carried out for obser-vations or measurements of X and Y given by {(xi, yi)}Mi=1. The fact that each out of Llayers exhibits its individual parameter set Θ = {Θl} = {wl, bl}Ll=1 is consequential forminimizing the cost function giving rise to the so-called backpropagation.The processing of input data and their conversion through the hidden layers are describedin the following.4.2Classifying the taskBefore, a comment on classification is in order. Following the scheme of ML approaches[18], the present CNN for homogenization can be understood as a task for a structuredoutput, since the output is a vector that exhibits relationships between its elements. Therelationships are established by existing material symmetries, compare Asm-1 in Sec. 3.3.For instance, if the elastic properties are close to isotropy, then all components of theelasticity tensors merely depend on the Lam´e-constants λ and µ. A second type of rela-tionships between the three vectors in the output of Fig. 3 is established by homogenizationapplied to the same sample but for different BCs.The task for a structured output is not too far from the task of classification, where theCNN is trained to assign one out of several categories to an input. The output is theneither a number which identifies a category or it is a probability distribution over classes.The most prominent example is image identification and classification as well as objectrecognition in images, the applications for which CNNs have become famous [33, 32] andmany more.An important difference between the tasks of these two cases is that classification re-quires invariance with respect to translation, rotation and scaling2. For the present caseof homogenization, these invariances would falsify the predictions for obvious reasons.4.3ConvolutionA 3D filter (also referred to as kernel) endowed with the weights obtained from training”scans” the phase voxels and applies convolutional operation to produce the feature map.The feature maps reflect the salient features of the CNN training objective.Formally, the input xl−1 of layer l is subject to an affine transformation of the formzl = wl ∗ xl−1 + bl ,(13)where ∗ is the convolution symbol, wl denotes the weights of a convolution filter and blthe biases; wl and bl can be incorporated in the parameter set Θl as mentioned above.2Pooling over spatial regions produces invariance to translations [18], a strategy for invariance torotations is data augmentation; similarly, scaling invariance is supported by adding random crops ofinput images.Deep CNNs predict elasticity tensors with bounds in homogenization10In fact, the application of a convolution can be cast into an affine transformation of theform zl = W lxl−1 +bl; where W l is a matrix with a particular structure. Since the resultzl undergoes activation al = f(zl) by a nonlinear function f, where the outcome servesas input in the consecutive convolutional layer, hence xl−1 → zl → al =: xl, the notationis adjusted to this procedure in the following.The representation (13), though correct and frequently used, does hide many relevantproperties of convolutional operations. Therefore, we prefer a more transparent form;technically, in 3D convolution a kernel as a 3D array of Rl rows (0 ≤ r ≤ Rl − 1), Clcolumns (0 ≤ c ≤ Cl − 1), and a depth of Dl (0 ≤ d ≤ Dl − 1) sums up the product ofits weights wlmqrcd at position (r, c, d) with the entry x(l−1)q(i+r)(j+c)(k+d) of the qth feature mapof the previous layer (l − 1) in terms of (14)3D:xlmijk=fblm +Q(l−1)−1�q=0Rl−1�r=0Cl−1�c=0Dl−1�d=0x(l−1)q(i+r)(j+c)(k+d) wlmqrcd(14)2D:xlmij=fblm +Q(l−1)−1�q=0Rl−1�r=0Cl−1�c=0x(l−1)q(i+r)(j+c) wlmqrc(15)1D:xlmi=f�blm +Q(l−1)−1�q=0Rl−1�r=0����dimensionalityx(l−1)q(i+r) wlmqr�.(16)This operation results in the output xlmijk at position (i, j, k) on the mth out of Ql featuremaps in the lth layer. The first sum in (14) over q indicates that the obtained featuremap is the superposition of the outcome of convolution applied to all feature maps ofthe previous, hence (l − 1)th layer. The weights as well as the biases are obtained in thetraining process minimizing the cost function. To highlight the dimensionality, we presentin (15) the convolution for a 2D image, and in (16) for a 1D sequence of data. Active inconvolution are the subscripts in (14)–(16).In the present case the input and all feature maps are of cubic shape having the dimensionN l × N l × N l, hence 0 ≤ i, j, k ≤ N l − 1. Equally of cubic shape are the convolutionalfilters, which are moreover of fixed size F for all layers, F ×F ×F, hence Rl = Cl = Dl = Fand 0 ≤ r, c, d ≤ F − 1.One convolutional kernel endowed with particular weights and biases is applied acrossthe entire cube and therefore extracts on type of feature. At the beginning, a kernel isplaced on the feature map at the location (i, j, k) of (0, 0, 0) and carries out convolutionas described. Then the kernel is moved by a prescribed increment S referred to as stridefor a consecutive convolution.By the choice of parameters F, S and zero-padding thickness P (adding a zero-layerof thickness P to the boundaries of a feature map for convolution), the feature mapsize is determined. For a 3D CNN with feature maps of cubic shape and the number ofvoxels/neurons per edge of N l−1 in the (l − 1)th layer, with F, P, and S as defined, theoutput has dimension N l in each direction of space withN l = N l−1 − F + 2PS+ 1 .(17)Bernhard Eidel11In the present case of F = 3 along with zero-padding P = 1 and stride S = 1, it holdsN l = N l−1, the feature map size is unaltered by convolution. This property conceptuallyenables nets of arbitrary depth.Nonlinear activation.Equations (14)–(16) contain two out of three operations whichwe understand as part of one convolutional layer [18]; the first one is an affine transfor-mation referred to as linear activation, which is the term in the brackets, the second isa nonlinear activation by a function f, which is frequently taken to be a ReLU (rectifiedlinear unit ReLU(z) := max(z, 0) in order to avoid vanishing gradients. The third stageis pooling which is described in Sec. 4.4.Some comments on common CNN properties are in order:C-1 Weight sharing; locality. For all the convolutional operations of a filter applied toa particular feature map in the previous layer the weights are the same. (The weightsin (14) do not depend on ijk.) This is referred to as weight sharing. The propertythat a neuron in a layer l is only connected to (a finite set of) corresponding neuronsin the lower layer l − 1 is called locality. (In terms of the indices in (14): ijk ← (i +r)(j + c)(k + d)). Weight sharing implies a relatively small number of parameters inconvolutional layers compared to FC layers, locality renders convolution an efficientscheme.C-2 First analysis, then synthesis. The convolution in the lowest layer acts on onesingle ”feature map”, in (14) the sum over q boils down to one single element,which is the VE in the input layer as displayed in Fig. 3. The VE’s microstructureis filtered for its features, which is an analytical process. Consecutive convolutionsin higher layers l ≥ 2 are superpositions of all feature maps of the previous layer(sketched in Fig. 3 for the central feature map) as described by the first sum overq in (14), which can be understood as a synthesis. Notice that for a convolution inhigher layers not only multiple feature maps are involved but also the weights inconvolving each feature map are altered (superscript q in the weights).C-3 A design principle. As a general design rule of CNNs the number of kernelsincreases for higher layers, so does the number of feature maps Ql, while in higherlayers the size of feature maps N l decreases. It is the pooling operation betweenconvolution layers that compresses the feature maps.4.4PoolingIn view of (17), size reduction can be realized by convolution. Frequently (and throughoutin the present work) convolution along with nonlinear activation is followed by poolingwhich carries out downsizing the feature maps.Average pooling reports the average output within a cubic neigborhood. For an input sizeof N × N × N voxels/neurons and a pooling kernel size of G × G × G entries, averagepooling yields the down-sized output at row-column-depth ijk of the feature map m atlayer lxlmijk = 1G3G−1�r=0G−1�c=0G−1�d=0xlm(G i+r)(G j+c)(G k+d) .(18)Deep CNNs predict elasticity tensors with bounds in homogenization12Since pooling applies to the outcome of convolution, the indices for layer l and featuremap m are unaltered.The corresponding case of 3D max(imum) pooling, which reports the maximum outputwithin a cubic neigborhood, leads to the output at position ijk ofxlmijk =maxr,c,d∈{0,1,...,G−1} xlm(G i+r)(G j+c)(G k+d) .(19)As an example, the feature map size is reduced to one eights for G = 2.4.5Fully connected layersAs displayed in Fig. 3 the stack of convolutional layers is followed by a stack of fullyconnected (FC) layers, where each neuron is connected to all neurons in the consecutivelayer. Since the FC layers are 1D, the 3D arrays of the feature maps need to be flattenedat the interface of convolutional to fully connected layers. Notice that the FC layers fortheir full connectivity largely determine the overall size of a CNN.In FC layers, the value xli of each neuron i out of N l in layer l is computed as the sumover the weighted neuron values in the previous layer l − 1 according tozlj = blj +Nl−1−1�i=0xl−1iwlij ,(20)where, as for the convolutional layers, the affine transformation of (20) can be followed bya nonlinear activation f of the output zlj written as alj = f(zlj). Similar to the convolutionallayers, the outcome of an FC layer serves as input in the consecutive layer, hence xl−1 →zl → al =: xl. At the highest FC layer, l = L − 1, the neurons are mapped to theoutput, it holds xL−1j= zj. In the present task for a structured output, the output layerconsists either of 9 elastic moduli for a CNN designed for one case of BCs, or of a vectorof length 27 for one single CNN covering three different BCs. For the final mapping intothe output the linear activation of (20) is used. For object classification this mapping istypically carried out by Softmax; it provides probabilities for detected objects belongingto different classes.4.6OptimizationSupervised learning of a CNN is realized by minimizing the loss function which measuresthe distance of the CNN prediction to the target, which is here the macroscopic elasticstiffness obtained by homogenization simulation.The CNN’s input in terms of VEs as well as its output are understood as known mea-surements of the random variables (X, Y ). The measurements are given by {(xi, yi)}Mi=1.The corresponding loss function (12) as a mean squared error MSE reads asL(Θ) = 1MM�I=1(yI − zI(Θ, xI))2 .(21)For optimization and testing, the total data set {(xI, yI)}MI=1 is decomposed into threesets,Bernhard Eidel13(i) the training set T = {(xI, yI)}MTI=1 with MT = card(T ),(ii) the validation set V = {(xI, yI)}MVI=1 with MV = card(V), and(iii) the testing set for assessment A = {(xI, yI)}MAI=1 with MA = card(A),along with M = MT + MV + MA, frequently in a ratio of 70% to 20% to 10%.Optimization is carried out for the loss function operating on T and V in two steps:1. Minimize the loss L(Θ) of (21) for the training set T , by finding (Θ∗) =arg minΘL(Θ).2. Evaluate the loss L(Θ∗) for the validation set V, at the currently optimal parametervalues (Θ∗).The procedure of 1. and 2. is the standard method in DNNs to detect overfitting inoptimization. It manifests in that training data can be accurately reproduced by theCNN, but new data lead to a large gap to the training accuracy.The training set T is decomposed into mini-batches. The batch size is the number ofsamples that will be propagated through the network in one forward pass inducing anupdate of the parameters Θ∗. The number of forward passes to pipe the full set of trainingdata MT through the network is the number of iterations per epoch. Large batch sizesare not only memory demanding but also lead to a significant degradation of the net’sability to generalize in tests. This observation made in practice was explained by the factthat large-batch methods tend to converge to sharp minimizers in the landscape of theloss function and get trapped therein, whereas smaller batch size methods consistentlyconverge to flat minimizers [27].4.6.1Gradient descent and backpropagationFor gradient-based minimizationmethods, the parameters are corrected through gradient descent according toΘl ← Θl − α ∂L∂Θl(22)with the learning rate α. For gradient descent, recall that the final output is the resultof activation in the highest of all layers aL = fL(Θ, x) and that the first activation takesplace in layer l = 1, where l = 0 is the input layer. This implies a representation of theloss function for a mini-batch Tbatch ⊂ T of size Mbatch ≪ MTL(Θ) =1MbatchMbatch�I=1�y(xI) − aL(Θ, xI)�2 .(23)To obtain the gradients for the descent, the partial derivatives of L with respect to theparameters Θ = {(wl, bl)}Ll=1 are required for each layer. The chain rule enables to deter-mine the partial derivatives ∂L/∂Θl, from the highest to the lowest layer, l = L, . . . , 1,Deep CNNs predict elasticity tensors with bounds in homogenization14which coins the name backpropagation∂L∂ΘL=∂L∂fL∂fL∂ΘL ,(24)∂L∂ΘL−1=∂L∂fL∂fL∂fL−1∂fL−1∂ΘL−1 ,(25). . .∂L∂Θl=∂L∂fL∂fL∂fL−1. . . ∂fl+1∂fl∂fl∂Θl ,(26)where ∂fl+1/∂fl are derivatives of the output with respect to the input of layer l+1 (validfor both convolutional and FC layers), and ∂fl/∂Θl are derivatives of the output of thelayer l with respect to its parameters.Detailed presentations of backpropagation can be found in, e.g., Sec. 12.5. of [17] and withan account of stochastic gradient descent in [21]. A review of the historical developmentof backpropagation is given in Sec. 5.5. of [41] with references to precursor work of themost prominent papers [39, 33].5ResultsFigure 4: Scaling. Speed-up by MPI-parallelization of homogenization compu-tations.After the microstructure generation describedin Sec. 2 the computation of the homogenizedelasticity tensors of each VE for different BCsis carried out by an MPI-parallelized finite el-ement code. The scaling behavior is displayedin Fig. 4 for a reference speed obtained for240 processing cores. Corresponding sets oflinear equations are solved by the direct solverPARDISO of the Intel Math Kernel Library(MKL) 2020 and by the iterative solver GM-RES along with a preconditioner, both fromthe PETSc library. In the simulations, the fi-nite element discretization adopts the origi-nal, uniform voxel grid, although considerablecomputational savings at moderate accuracylosses can be realized by resolution coarsen-ing, by adaptive, octree-based mesh coarsening, or combinations thereof [16].5.1Pre-analysis on stiffness boundsWe wish to know, how strong elastic stiffness depends on the applied boundary condition,which enables conclusions on the suitability of individual VEs to serve as an RVE. TheYoung’s modulus of the stiff phase is 100 GPa, of the compliant phase 2 GPa, hence acontrast of factor 50. Poisson’s contraction is set to ν = 0.3 for both phases.The point clouds in the diagrams of Fig. 5 (a)–(d) show the ratios of stiffness for differentBCs in exemplary terms of components C11 and C13, which are representative for otherBernhard Eidel15elasticity coefficients. They reveal the following characteristics:(a) component C11(b) component C13(c) component C11(d) component C13(e) 7.6/5.6/7.4 (f) 7.5/5.8/6.5 (g) 7.9/6.8/7.1(h) 0.6/0.7/0.9 (i) 0.5/1.4/0.8(j) 1.7/1.1/0.5Figure 5: Scatter of stiffness spread for different BCs. In (a) and (c) the stiff-ness ratio for KUBC/SUBC over PBC stiffness, in (b) and (d) the stiffness ratio forKUBC/PBC (blue) and SUBC/PBC (orange) over PBC for the full dataset of 104 VEs.For intermediate PBC stiffness in terms of C13, the VEs in (e)–(g), highlighted in (b),have the largest KUBC/SUBC stiffness ratio, the VEs in (h)–(j), equally marked in (b),have the smallest KUBC/SUBC ratio. The values of the variances sx/sy/sz are given inthe captions of Subfigs. (e)–(j).(i) The stiffness of the majority of VEs significantly depends on the applied BCs. TheKUBC stiffness exceeds the SUBC stiffness by a factor of 4 at maximum. The ratio ofKUBC to PBC stiffness exhibits factor 2 at maximum for compliant composites andmonotonously decreases for increasing overall stiffness, the minimal ratio of PBC toSUBC stiffness exhibits factor 0.3 at intermediate overall stiffness of the composites.(ii) Only in the limit of mono-phase systems, hence of highest or lowest stiffness, theVEs exhibit BC-invariance of stiffness.Deep CNNs predict elasticity tensors with bounds in homogenization16(iii) For intermediate phase fractions there is a white space between the point cloudsto the case of BC-invariance (stiffness ratios equal to one) in Fig. 5 (a)–(d). Thiswhite space follows from minimal values of variances close to 0.5 in the virtualmicrostructure generation. The VEs coming closest to BC-invariance all exhibitsmall variance values, the VEs showing strongest sensitivity to BCs exhibit largevariance values as underpinned in the captions of Fig. 5 (e)–(j). This suggests thatfor s → 0 the gap of the white space is decreased or even closed, given that the voxelresolution is consistent to that limit. The error due to the finite voxel resolution ofimage-based microstructures is introduced into the error framework of two-scalefinite element methods for homogenization in [12].In conclusion, the deviation of PBC stiffness from the lower or the upper bound indicatesthat the majority of samples can not serve as an RVE for their small size. As a consequence,the application of PBC implies an error of the apparent stiffness of uncertain magnitude,which is at least bounded by the SUBC and KUBC results. These bounds, however, donot necessarily apply to a true RVE of sufficient size.Two comments are in order:C-4 The full dataset of VEs by its construction shall empower the net with predic-tions for various classes of different microstructures. The statistics of the ensembleis available, but of course not specific enough to predict for a single snapshot-VEthe effective elastic properties by a statistical analysis as proposed by Kanit et al.[26]. Quite in contrast, each VE can be understood as a realization of an unknownensemble with its own statistics. For that reason, the conclusion for the size re-quirement of the RVE in the present context must be based on the criterion ofboundary-invariance of stiffness. Similarly, costly 3D image acquisition by tomogra-phy typically provides only a small number of snapshot specimens in voxel resolutioninstead of a statistical ensemble [5, 19].C-5 The criterion of boundary-insensitivity is of course blind for periodicity. If, for theupper bound of stiffness, KUBC are applied to a VE of edge length δ, which is largerthan the periodic unit cell length ϵ, the convergence to boundary insensitivity, i.e.of KUBC and PBC towards each other, scales with 1/δ according to [10] (Thm.1.2). The corresponding error is the modeling error in the FE-HMM frameworkof errors. It has its origin in the boundary layer which feels the rigidity of theDirichlet constraint. With increasing distance to the boundary the periodic solutionis continuously recovered. In this scenario of convergence, the SUBC results are notconsidered.5.2The present CNN architecture, hyperparameters and implementationThe present CNN exhibits the architecture displayed in Fig. 3, which adopts the archi-tecture of the pioneering work of [51] to a large extent. The exception is the size of theFC layers. We choose for the FC block two layers of widths 256 and 128 instead of 2048and 1024 of reference [51] and thereby reduce the CNN size by more than 82%. The con-volutional block consists of 5 convolutional layers, each one endowed with an increasingnumber of kernels, from 16 to 256. The convolutional kernel size is constantly 3 × 3 × 3,Bernhard Eidel17strides are 1×1×1. Nonlinear activation is carried out by ReLU. Average pooling and, asan alternative, max pooling with kernel size 2×2×2 are carried out after each convolution.For regularization, the L2 value is set to 0.001.For the input layer, 1003 neurons are used according to the number of voxels in eachVE. The voxel data are transformed from binary values 0 and 1 to -0.5 and 0.5 followingthe proposal of [51]. It turned out to improve activation and overall performance of theconvolution layers.The implementation is realized with Keras 2.4.0 as part of Tensorflow 2.4.1 using Python3.7. The training simulations were carried out on NVIDIA TESLA V100-GPUs with 32GB memory for each GPU.5.3Training(a) KUBC(b) PBC(c) SUBC(d) KUBC, PBC, SUBCFigure 6: Training and validation losses. Losses over epochs for different BCs, (a)for KUBC, (b) for PBC, (c) for SUBC, and (d) for the three BCs in one CNN. The lossesrefer to the cost function (GPa) for the training data and the validation data.For optimization, the Adam algorithm [29] is used with the default parameters of theKeras library except of the learning rate α (α = 0.0001 instead of the default α = 0.001).Deep CNNs predict elasticity tensors with bounds in homogenization18CNN-casePoolingKUBCPBCSUBCall-BCsAverageLoss0.58401.97911.83834.9503Epoch4738219627613687MaxLoss1.36882.67813.16597.8501Epoch232312619501286Table 1: Validation losses. Minimal validation losses (GPa) with corresponding epochsfor average and max pooling.The set of 104 VEs is randomly decomposed into sets for training, validation and testingin the ratio of 70:20:10%. For the reasons mentioned in Sec. 4.6 the training set is splitinto mini-batches of 32 VEs, which results in 7000/32→219 iterations per training epoch.Figure 6 shows the losses over the number of training epochs for the four different CNNs.The validation losses decrease monotonically, and turn into an increase thus indicatingthe start of overfitting. The hyperparameters are taken at the sweet spot of the minimumof the validation losses, the corresponding epochs are listed in the diagrams (a)–(d) ofFig. 6. It turns out that average pooling is more accurate than max popling as shown inTab. 1. Dropout in the FC layers with a dropout rate of 0.2 does not improve the results.5.4Accuracy5.4.1The mean absolute stiffness error (MASE)The CNN performance is eval-uated by calculating the mean absolute stiffness error (MASE) [51] for each tensor com-ponent, and an average MASE combining the results for all tensor componentsMASE(Cij) :=1MTMT�k=1���Ck(target)ij− Ck(prediction)ij���average{C(target)ij},MASE := 199�I=1MASEI ,(27)where MT denotes the total number of VEs in the test set, Ck(target)ijand Ck(prediction)ijarethe computed apparent stiffness components and CNN-predicted stiffness components forthe kth VE, respectively. A non-dimensionalization is carried out by the average of Cij ofall the VEs in MT.Table 2 provides the values for testing-MASE for each component of the homogenized elas-ticity tensor separately as well as the averages MASE. The results indicate the followingcharacteristics:• The predictions are very accurate, testing-MASE is below two percent throughout.• Testing-MASE increases for decreasing stiffness of the applied BCs; they are smallestfor KUBC (MASE < 1%), largest for SUBC, for PBC in between.• Only very minor accuracy losses for one CNN for all BCs compared to the case ofthree distinct CNNs. This is true for all BCs, but most pronounced for PBC andSUBC, which is best seen for the average of testing-MASE.Bernhard Eidel19Testing-MASE (%)MASE (%)CaseBCC11C22C33C12C13C23C44C55C66CijKUBC0.560.520.530.670.720.710.540.580.550.603 CNNsPBC1.111.091.141.291.321.221.091.121.151.17SUBC1.511.511.531.711.641.751.731.571.721.63KUBC0.830.750.790.961.000.950.770.830.780.851 CNNPBC1.231.131.131.341.401.361.151.191.181.23SUBC1.621.561.501.791.701.761.691.681.721.67Table 2: CNN Accuracy. Testing-MASE for each tensor component for various CNNs.5.4.2The relative error in boxplotsRelative errors according toerel :=�C(target)ij− C(prediction)ij�/ C(target)ij(28)are shown in the boxplots of Fig. 7; in (a) for the case of three distinct CNNs each forone BC, in (b) for one CNN covering all BCs. The accuracy is best for KUBC, lowest forSUBC, for PBC in between. The observed property of ”the stiffer, the more accurate”manifests in the median, the upper and lower quartile defining the interquartile rangeIQR, and is equally visible by the whiskers marking 1.5 of the IQR. Outliers above 10%are rare in either case. The accuracy of one CNN for all BCs is virtually indistinguishablefrom the separate treatment of each BC case.(a) Three distinct CNNs for KUBC, PBC, and SUBC.(b) One single CNN for all BCs.Figure 7: Boxplots. Errors erel for the test set with MA = 103.Deep CNNs predict elasticity tensors with bounds in homogenization20Three distinct CNNsOne CNNKUBCPBCSUBCTable 3: Target versus CNN prediction for Cij. The results refer to three distinctCNNs (left column) and one CNN for all BCs (right column). The diamond/β-SiC resultsare marked by yellow diamond symbols for 200 voxels per edge, for 100 voxels in red.Bernhard Eidel215.4.3Target stiffness versus CNN predictionThe results in Tab. 3 underpin thehigh accuracy of the CNN predictions for all BCs and both cases of one and three CNNs;it requires a zoom-in to recognize differences between target predicted values of Cij.6Assessing the CNNs by Diamond/β-SiC Composite SamplesThe CNN predictions for the virtual microstructures achieved high accuracy. Next, thereal microstructure of a diamond/β-SiC thin film shall assess the potential to generalize.(a) Dark: β-SiC, bright: diamond.(b) Diamond phase.(c) Cubic test samples of length 2 µm and resolution of 2003 voxels.Figure 8: Diamond/β-SiC. The specimen exhibits the length, height, and width of10 × 3.3 × 2µm and is resolved by 1000 × 330 × 200 voxels. Hence, a voxel exhibits thelength of 10 nm. Notice in (a) and (b), that the bottom layer as the interface to thesubstrate is purely β-SiC. The coordinate system in (c) refers to (a) and (b) likewise. Forthe indices of the elastic moduli it holds 1 ← x, 2 ← y, 3 ← z.Diamond thin films are of interest for their outstanding mechanical properties such as highhardness, low friction coefficient and high wear resistance. However, their application asprotective coatings is inhibited by the poor adhesion on many substrates. The reasonare high stresses at the interface of diamond to substrate, which are induced by differentthermal expansion coefficients. An additional problem is a catalytic effect, which resultsDeep CNNs predict elasticity tensors with bounds in homogenization22in soot and graphite formation in the context of iron-, cobalt- and nickel-based materials.A solution provide nanocrystalline diamond/β-SiC composite films as transition layers;they are effective as adhesion layers and, serving as a barrier, they prevent the catalyticeffect of the substrate elements. Diamond/β-SiC composites are made by chemical vapordeposition (CVD). For synthesis, characterization and applications we refer to [54, 46, 50].The test samples are obtained by cutting out cubes of edge length 100 and 200 voxels,respectively, which renders 60 and 5 non-intersecting VEs. By flipping the phase proper-ties the number of samples is doubled. The 200-voxel samples are reduced to 100-voxelresolution in order to fit the CNN specifications. Figure 8 (c) indicate that the 2003 voxelcubes are cut out in some distance to the bottom of the CVP-grown specimen, since thebottom layer is made of pure β-SiC for the reasons mentioned.For the test, the diamond-SiC samples adopt the elastic properties of the phases as usedin the training.Figure 9: Boxplots. CNN errors erel for the 200-voxel diamond/β-SiC samples in com-parison to the standard test set. The results refer to three distinct CNNs.The results displayed in the left column of Tab. 3 indicate, that the CNN predictions forthe diamond/β-SiC samples of the 200 voxel size (marked by yellow diamonds) are veryaccurate and thereby come close to the accuracy for the synthetic microstructures of thestandard test set. More descriptive for the assessment are the boxcharts in Fig. 9. Thepredictions for the VEs of edge length 100 voxels (marked by red diamonds), which aredisplayed in the right column of Tab. 3, are less accurate than for the 200 voxel case. Here,the predictions for the stiff KUBC case are much better than for the cases of PBC andSUBC. For the latter BCs the general trend to underestimate the true elastic stiffness forall components can be observed. An explanation for these findings is that the 100 voxelcase exhibits very large phase domains with respect to the VE size, a case, which is hardlycovered by the training set with variances having an upper limit of 8.If a particular materials system is envisaged, the approach with training data into theparticular direction of that target material system is a reasonable choice as, e.g., in [34]aiming at elastic predictions for shale. Here in contrast, the generation of VEs described inSec. 2 was not restricted to the type of CVP-grown, two-phase microstructures. Notwith-standing, the sample in Fig. 1 (g) resembles the real microstructure samples in Fig. 8(c) in its parallel alignment of phases by the choice of corresponding variance parameterssy > sx, sz. This type of microstructure induces elastic properties very close to transversalisotropy as shown for diamond/β-SiC in [13].Bernhard Eidel237Discussion and ConclusionsIn the present work, CNNs are constructed and trained to link 3D two-phase microstruc-tures with arbitrary phase fractions to their elasticity tensors from homogenization withdifferent BCs. The CNN predicts the apparent elastic stiffness for VE subject to PBCalong with its lower bound by SUBC and its upper bound by KUBC in one thrust.Thereby it indicates, whether the VE is large enough to serve as an RVE.The pre-analysis of the randomly generated microstructures has revealed a considerablestiffness deviation for KUBC from SUBC for the majority of the generated VEs. Beyondthe CNN-related focus of this paper the results underpin, that PBC for random, non-periodic heterogeneous matter imply an uncertainty of the true elastic stiffness and canlead to gross errors, if the VE is not sufficiently large. This is of general interest, but inparticular for the design of digital twins of heterogeneous materials.The CNN predictions are in excellent agreement with the target of numerical homoge-nization for the standard test track. The CNN covering all three boundary conditionsshows only very minor accuracy losses compared to three distinct CNNs. Remarkably, theCNN predictions turned out to be similarly accurate for real, CVP-grown diamond/β-SiCmicrostructures, a material of industrial relevance for its use as protective and biologicalcoating. Since the CNNs were not trained into the particular direction of these microstruc-tures, and learning was based merely on synthetic samples, the CNNs of this work can beseen as a step towards universal CNNs covering several classes of microstructures with areal-world impact. Similarly, more generality is achieved by arbitrary phase fractions ofthe composite.The proposed CNNs, which embody the learned lessons about elastic homogenization intheir parameters of weights and biases, can be used in two different types of applications.First, for real microstructure samples obtained by tomography and image processing;since in that case of one or a few snapshots the statistics are missing, the prediction ofthe apparent stiffness and the necessary RVE size must be based on the PBC-results alongwith the KUBC- and SUBC-bounds. Second, the CNNs can be used for statistical analysisof an ensemble of (artificially generated) microstructures in order to determine effectiveproperties following the approach of Kanit et al. [26].In either case of application, the predictions are fast and cheap, since they are independentof finite element simulations similar to the off-line stage of Reduced Basis Methods (RBM).Furthermore, the overall CNN-size of [52] could be reduced by approximately 80% by asize reduction of the FC layers.Compared to the wealth of experience with the design and optimization of CNNs in imageidentification and classification, homogenization-CNNs are still in their infancy.Acknowledgements. The author acknowledges support by the Deutsche Forschungsge-meinschaft (DFG) within the Heisenberg program (grant no. EI 453/5-1). Simulationswere performed with computing resources granted by RWTH Aachen University underproject ID BUND0005 and by the University of Siegen. The author thanks Prof. XinJiang for providing the diamond-SiC sample and Dr. Lorenz Holzer for images thereof byFIB-tomography from another joint project. Assistance of Ajinkya Gote and MohamedImran Peer Mohamed in simulations and postprocessing is gratefully acknowledged.Deep CNNs predict elasticity tensors with bounds in homogenization24References[1] A. Abdulle. Analysis of the Heterogeneous Multiscale FEM for problems in elasticity.Mathematical Models and Methods in Applied Sciences, 16(04):615–635, 2006.[2] A. Abdulle. The Finite Element Heterogeneous Multiscale Method: A computationalstrategy for multiscale PDEs. GAKUTO Int. Ser. Math. Sci. Appl., 31:133–181, 2009.[3] A. Abdulle, W. E, B. Engquist, and E. Vanden-Eijnden. The Heterogeneous Multi-scale Method. Acta Numerica, 21:1–87, 2012.[4] A. Abdulle and C. Schwab. Heterogeneous Multiscale FEM for Diffusion Problemson Rough Surfaces. Multiscale Modeling & Simulation, 3(1):195–220, 2005.[5] H. Andr¨a, N. Combaret, J. Dvorkin, E. Glatt, J. Han, M. Kabel, Y. Keehm,F. Krzikalla, M. Lee, C. Madonna, M. Marsh, T. Mukerji, E. H. Saenger, R. Sain,N. Saxena, S. Ricker, A. Wiegmann, and X. Zhan. Digital rock physics benchmarks—Part II: Computing effective properties. Computers & Geosciences, 50:33–43, 2013.[6] J. Berner, P. Grohs, G. Kutyniok, and P. Petersen. The Modern Mathematics ofDeep Learning: arXiv preprint arXiv:2105.04026, 2021.[7] K. Breuer and M. Stommel. Prediction of Short Fiber Composite Properties by anArtificial Neural Network Trained on an RVE Database. Fibers, 9(2):8, 2021.[8] F. Cucker and D.-X. Zhou. Learning theory: An approximation theory viewpoint,volume v.24 of Cambridge Monographs on Applied and Computational Mathematics.Cambridge University Press, Cambridge, 2007.[9] W. J. Drugan and J. R. Willis. A micromechanics-based nonlocal constitutive equa-tion and estimates of representative volume element size for elastic composites. Jour-nal of the Mechanics and Physics of Solids, 44(4):497–524, 1996.[10] W. E, P. Ming, and P. Zhang. Analysis of the Heterogeneous Multiscale Method forelliptic homogenization problems. Journal of the American Mathematical Society,18(01):121–157, 2005.[11] B. Eidel and A. Fischer. The heterogeneous multiscale finite element method forthe homogenization of linear elastic solids and a comparison with the FE2 method.Computer Methods in Applied Mechanics and Engineering, 329:332–368, 2018.[12] B. Eidel, A. Fischer, and A. Gote. From image data towards microstructure infor-mation – Accuracy analysis at the digital core of materials. ZAMM, 101(6), 2021.[13] B. Eidel, A. Gote, M. Ruby, L. Holzer, L. Keller, and X. Jiang.Estimating theeffective elasticity properties of a diamond/ β-SiC composite thin film by 3D recon-struction and numerical homogenization. Diamond and Related Materials, 97:107406,2019.[14] F. Feyel and J.-L. Chaboche. FE2 multiscale approach for modelling the elastovis-coplastic behaviour of long fibre SiC/Ti composite materials. Computer Methods inApplied Mechanics and Engineering, 183(3-4):309–330, 2000.Bernhard Eidel25[15] A. Fischer and B. Eidel. Convergence and error analysis of FE-HMM/FE2 for en-ergetically consistent micro-coupling conditions in linear elastic solids.EuropeanJournal of Mechanics - A/Solids, 77:103735, 2019.[16] A. Fischer and B. Eidel.Error analysis for quadtree-type mesh coarsening algo-rithms adapted to pixelized heterogeneous microstructures. Computational Mechan-ics, 16(04):615, 2020.[17] R. C. Gonzalez and R. E. Woods. Digital image processing. Pearson, New York NY,2018.[18] I. Goodfellow, Y. Bengio, and A. Courville. Deep Learning. MIT Press, 2016.[19] A. Gote, A. Fischer, C. Zhang, and B. Eidel. Computational Homogenization ofConcrete in the Cyber Size-Resolution-Discretization (SRD) Parameter Space: arXivpreprint arXiv:2103.08957, 2021.[20] K. M. Graczyk and M. Matyka. Predicting porosity, permeability, and tortuosity ofporous media from images by deep learning. Scientific reports, 10(1):21488, 2020.[21] C. F. Higham and D. J. Higham. Deep Learning: An Introduction for Applied Math-ematicians. SIAM Review, 61(3):860–891, 2019.[22] R. Hill. Elastic properties of reinforced solids: Some theoretical principles. Journalof the Mechanics and Physics of Solids, 11(5):357–372, 1963.[23] C. Huet. Application of variational concepts to size effects in elastic heterogeneousbodies. Journal of the Mechanics and Physics of Solids, 38(6):813–841, 1990.[24] D. Jeulin. Random texture models for material structures. Statistics and Computing,10(2):121–132, 2000.[25] S. Kamrava, P. Tahmasebi, and M. Sahimi. Linking Morphology of Porous Mediato Their Macroscopic Permeability by Deep Learning. Transport in Porous Media,131(2):427–448, 2020.[26] T. Kanit, S. Forest, I. Galliet, V. Mounoury, and D. Jeulin. Determination of the sizeof the representative volume element for random composites: statistical and numericalapproach. International Journal of Solids and Structures, 40(13-14):3647–3679, 2003.[27] N. S. Keskar, D. Mudigere, J. Nocedal, M. Smelyanskiy, and P. T. P. Tang. OnLarge-Batch Training for Deep Learning: Generalization Gap and Sharp Minima:arXiv-preprint arXiv:1609.04836, 2017.[28] A. Khan, A. Sohail, U. Zahoora, and A. S. Qureshi.A survey of the recent ar-chitectures of deep convolutional neural networks.Artificial Intelligence Review,53(8):5455–5516, 2020.[29] D. P. Kingma and J. Ba.Adam: A Method for Stochastic Optimization: arXivpreprint arXiv:1412.6980, 2014.Deep CNNs predict elasticity tensors with bounds in homogenization26[30] R. Kondo, S. Yamakawa, Y. Masuoka, S. Tajima, and R. Asahi.Microstructurerecognition using convolutional neural networks for prediction of ionic conductivityin ceramics. Acta Materialia, 141:29–38, 2017.[31] V. Kouznetsova, W. A. M. Brekelmans, and F. P. T. Baaijens. An approach to micro-macro modeling of heterogeneous materials. Computational Mechanics, 27(1):37–48,2001.[32] A. Krizhevsky, I. Sutskever, and G. E. Hinton. ImageNet Classification with DeepConvolutional Neural Networks. In F. Pereira, C. J. C. Burges, L. Bottou, and K. Q.Weinberger, editors, Advances in Neural Information Processing Systems 25, pages1097–1105. Curran Associates, Inc, 2012.[33] Y. LeCun, B. Boser, J. S. Denker, D. Henderson, R. E. Howard, W. Hubbard, andL. D. Jackel. Backpropagation Applied to Handwritten Zip Code Recognition. NeuralComputation, 1:541–551, 1989.[34] X. Li, Z. Liu, S. Cui, C. Luo, C. Li, and Z. Zhuang. Predicting the effective mechanicalproperty of heterogeneous materials by image based modeling and deep learning.Computer Methods in Applied Mechanics and Engineering, 347:735–753, 2019.[35] J. C. Michel, H. Moulinec, and P. Suquet. Effective properties of composite materi-als with periodic microstructure: a computational approach. Computer Methods inApplied Mechanics and Engineering, 172(1-4):109–143, 1999.[36] C. Miehe, J. Schr¨oder, and J. Schotte. Computational homogenization analysis infinite plasticity simulation of texture development in polycrystalline materials. Com-puter Methods in Applied Mechanics and Engineering, 171(3-4):387–418, 1999.[37] D. Peri´c, E. A. de Souza Neto, R. A. Feij´oo, M. Partovi, and A. J. C. Molina.On micro-to-macro transitions for multi-scale analysis of non-linear heterogeneousmaterials: unified variational basis and finite element implementation. InternationalJournal for Numerical Methods in Engineering, 87(1-5):149–170, 2011.[38] C. Rao and Y. Liu.Three-dimensional convolutional neural network (3D-CNN)for heterogeneous material homogenization.Computational Materials Science,184:109850, 2020.[39] D. E. Rumelhart, G. E. Hinton, and R. J. Williams. Learning Internal Representa-tions by Error Propagation. In D. E. Rumelhart and J. L. Mcclelland, editors, ParallelDistributed Processing: Explorations in the Microstructure of Cognition, Volume 1:Foundations, pages 318–362. MIT Press, Cambridge, MA, 1986.[40] S. Saeb, P. Steinmann, and A. Javili. Aspects of Computational Homogenizationat Finite Deformations: A Unifying Review From Reuss’ to Voigt’s Bound. AppliedMechanics Reviews, 68(5):050801, 2016.[41] J. Schmidhuber. Deep Learning in Neural Networks: An Overview. Neural Networks,61(3):85–117, 2015.Bernhard Eidel27[42] J. Schr¨oder. A numerical two-scale homogenization scheme: the FE2-method. InF. Pfeiffer, F. G. Rammerstorfer, E. Guazzelli, B. Schrefler, P. Serafini, J. Schr¨oder,and K. Hackl, editors, Plasticity and Beyond, volume 550 of CISM InternationalCentre for Mechanical Sciences, pages 1–64. Springer Vienna, Vienna, 2014.[43] J. Tian, C. Qi, Y. Sun, and Z. M. Yaseen. Surrogate permeability modelling of low-permeable rocks using convolutional neural networks. Computer Methods in AppliedMechanics and Engineering, 366:113103, 2020.[44] T. C. T. Ting. Anisotropic Elasticity: Theory and Applications. Oxford UniversityPress, USA, 1996.[45] P. Vannucci. Anisotropic Elasticity, volume 85. Springer Singapore, Singapore, 2018.[46] T. Wang, S. Handschuh-Wang, Y. Yang, H. Zhuang, C. Schlemper, D. Wesner,H. Sch¨onherr, W. Zhang, and X. Jiang. Controlled surface chemistry of diamond/β-SiC composite films for preferential protein adsorption. Langmuir : the ACS journalof surfaces and colloids, 30(4):1089–1099, 2014.[47] H. Wu, W.-Z. Fang, Q. Kang, W.-Q. Tao, and R. Qiao. Predicting Effective Diffu-sivity of Porous Media from Images by Deep Learning. Scientific reports, 9(1):20387,2019.[48] J. Wu, X. Yin, and H. Xiao. Seeing permeability from images: fast prediction withconvolutional neural networks. Science Bulletin, 63(18):1215–1222, 2018.[49] C. Yang, Y. Kim, S. Ryu, and G. X. Gu.Using convolutional neural networksto predict composite properties beyond the elastic limit.MRS Communications,9(2):609–617, 2019.[50] N. Yang, editor. Novel Aspects of Diamond. Topics in Applied Physics. SpringerInternational Publishing, Cham, 2015.[51] Z. Yang, X. Li, L. Catherine Brinson, A. N. Choudhary, W. Chen, and A. Agrawal.Microstructural Materials Design Via Deep Adversarial Learning Methodology. Jour-nal of Mechanical Design, 140(11), 2018.[52] Z. Yang, Y. C. Yabansu, R. Al-Bahrani, W.-k. Liao, A. N. Choudhary, S. R. Kalidindi,and A. Agrawal. Deep learning approaches for mining structure-property linkages inhigh contrast composites from simulation datasets. Computational Materials Science,151:278–287, 2018.[53] M. D. Zeiler and R. Fergus. Visualizing and Understanding Convolutional Networks.In D. Fleet, T. Pajdla, B. Schiele, and T. Tuytelaars, editors, Computer Vision –ECCV 2014, volume 8689 of Lecture Notes in Computer Science, pages 818–833.Springer International Publishing, Cham, 2014.[54] H. Zhuang, B. Song, V. V. S. S. Srikanth, X. Jiang, and H. Sch¨onherr. ControlledWettability of Diamond/β-SiC Composite Thin Films for Biosensoric Applications.The Journal of Physical Chemistry C, 114(47):20207–20212, 2010.View publication stats | RESEARCH GATE |
1Preprint:Pleasenotethatthisarticlehasnotcompletedpeerreview.Endotrachealtubecuffpressure–howfastparamedicscanlearntheoptimalinflation?-comparisonofdifferentmethodsCURRENTSTATUS:UNDERREVIEWTomaszPiotrIlczakAkademiaTechnicznoHumanistycznawBielsku-Bialejtilczak@ath.bielsko.plCorrespondingAuthorORCiD:https://orcid.org/0000-0003-2478-9045MichalTomaszCwiertniaUniversityofBielsko-BialaPiotrBialonUniversityofBielsko-BialaMichalSzlagorUniversityofBielsko-BialaBeataKudlacikUniversityofBielsko-BialaMałgorzataRakUniversityofBielsko-BialaSzymonBialkaSlaskiUniwersytetMedycznywKatowicachAdamUbychSlaskiUniwersytetMedycznywKatowicachArkadiuszStasickiUniversityofBielsko-BialaWiolettaWaksmanskaUniversityofBielsko-BialaJanBujokUniversityofBielsko-Biala2MonikaMikulskaUniversityofBielsko-BialaRafalBobinskiUniversityofBielsko-BialaMarekStanisławKaweckiUniversityofBielsko-BialaDOI:10.21203/rs.3.rs-25600/v1SUBJECTAREASCriticalCare&EmergencyMedicineKEYWORDSparamedic,intubation,cuffpressure3AbstractBackgroundTrachealintubationistheoptimalmethodforopeninguptheairways.Performedcorrectly,itpreventsstomachcontentsfromenteringtherespiratorytract,andallowsasynchronousCPRtobeconductedduringsuddencardiacarrest.Animportantelementofcorrectintubationisproperinflationoftheendotrachealtubecuff.Researchhasshownthatwhenmedicalpersonnelusethepalpationtechnique,thecuffisusuallyinflatedincorrectly.Thiscanresultinnumeroushealthcomplicationsforthepatient.MethodsThisresearchwasconductedin2020onagroupofparamedicsparticipatinginthe15thInternationalWinterChampionshipofMedicalRescuersinBielsko-Biała.Theaimoftheresearchwastoassesstwomethodsofinflatingtheendotrachealtubecuff.MethodAinvolvedinflatingthecuffusingasyringeandassessingthepressureinthecontrolcuffusingthepalpationtechnique.MethodBinvolvedinflatingthecuffusingamanometer.Duringtheinflation,boththecuffinflationpressureandthetimerequiredtocompletetheprocedurewererecorded.AnalysiswasalsoconductedonwhethercompletionofcertifiedALSandACLStraininghadanyinfluenceontheeffectivenessoftheinflationprocedure.ResultsTheresearchshowedthatparamedicsusingmethodBsignificantlymoreofteninflatedtheendotrachealtubecufftothecorrectpressurethanthoseusingmethodA.However,whenmethodBwasused,theproceduretooklongertoconduct.ThestudyalsoshowedthatcompletionofcertifiedALSorACLStrainingdidnothaveasignificantinfluenceonproperinflationofthecuff.ThosewhohadcompletedcertifiedtrainingcoursestooksignificantlylongertoinflatetheendotrachealtubecuffwhenusingmethodA.ConclusionsInflationoftheendotrachealtubecuffbyuseofasyringe,followedbythepalpationtechniqueforassessingtheinflationofthecuffballoonisineffective.Paramedicteamsshouldbeequippedwithmanometerstobeusedforinflatingtheendotrachealtubecuff.Background4Theendotrachealintubationprocedureremainsthemosteffectivemethodofopeningtheairwaysduringpre-hospitaltreatment.Properinsertionoftheendotrachealtubepreventsstomachcontentsfromenteringtherespiratorytractandallowsasynchronouscardiacresuscitationtobeconducted,whichminimizesthebreaksbetweenchestcompressions,thusimprovingthequalityoftheresuscitationprocedure[1].Animportantfactorinproperuseofendotrachealintubationiscorrectinflationoftheendotrachealtubecuff.Thepressurerequiredforcorrectinflationofthecuffshouldbeintherangeof20to30cmH2O.Assessmentofthecorrectpressureshouldbepreciseandshouldbecarriedoutusingamanometertoindicatethecuffsealinflationpressure.Researchinothercountrieshasshownthatassessmentofcuffinflationusingthepalpationtechniqueisrelativelyineffectiveandrarelyresultsincorrectcuffinflation.Experiencedparamedicspecialistsusingthepalpationtechniquetoassesstheamountofairinthecuffduringintensivecareconducttheprocedurecorrectlyinlessthan30%ofcases[2–5].Ithasbeenproventhatduringpalpationassessment,thecuffsealisoverinflated,eventoashighas100cmH2O[4].Applicationofincorrectpressureintheendotrachealtubecuffsealcanbeextremelyworryingasitmaycausenumerouscomplications[6–11].Inpre-hospitalparamedictreatment,intubationisoftenusedasthechosenmethodforclearingtheairways.Unfortunately,attentionisrarelypaidtothecorrectcuffseal,andinstrumentsforensuringthesealduringtheprocedureareseldomused.Duringparamedictreatment,inmostcasesasyringeisusedandtheamountofairisassessedusingthepalpationtechnique[12].Inthispaper,wetrytoanswerthequestionofwhatthedifferencesareinthequalityofendotrachealtubecuffinflationandthetimerequiredfortheprocedure,dependingonwhetherasyringeoradeviceforinflatingthecuffareused.Theaimofthestudyistodeterminetheeffectofusingadeviceforinflatingtheendotrachealtubecuffincomparisontotheuseofasyringe,intermsofthecuffinflationvolumeandthetimerequiredfortheprocedure.WewillalsoassesstheinfluenceofcompletedcertifiedAdvancedLifeSupport(ALS)andAdvancedCardiovascularLifeSupport(ACLS)trainingcoursesontheachievedinflationvolumeandthetimeneededtoinflatethecuff.Methods5Theresearchwasconductedonthe30thJanuary2020duringasimulatedrescueatthe15thInternationalWinterChampionshipofMedicalRescuersinBielsko-Biała,Poland.Theresearchinvolved108paramedicsworkinginambulanceresponseteams.Thepersonstakingpartintheresearchwerefrom32ambulancestationsfromacrossPoland.DemographicdataonthestudyparticipantsispresentedinTable1below.Allstudyparticipantsgaveinformedconsentforthestudytobecarriedout.TheresearchwasapprovedbytheethicalcommitteeofUniversityinBielsko-BiałaunderthereferencenumberRNN/305/2019–approval2019/12/7/2.Table1.StudygroupdemographicdataGenderFrequency-nPercentageMale10092.60%Female87.40%EducationFrequencyPercentageTertiary8982.40%Secondary1917.60%WorkexperienceFrequencyPercentageUpto5years5046.30%6-8years2321.30%Over8years3532.40%TrainingFrequencyPercentageCompletedALSorACLS7468.50%Notraining3431.50%6Thestudyusedasimulationtechniqueandwasconductedduringasimulatedrescueinordertomaximizethesimilarityoftheproceduretothatusedinreal-lifesituations.Theresearchwasprecededbyparticipantscompletingashortquestionnairewithdemographicdataandthemethodsusedbytheparticipantsforsealingtheendotrachealtubecuffinclinicaltreatment.TheprocedurewascarriedoutusingaSIMManALS(Laredal)mannikin,serialnumberSN.23548170915,anendotrachealID–7.5tube,anda20mlsyringeandmanometerforsealingtheVBM(Germany)endotrachealtubecuff.Measurementswerealsotakenusingavalidatedstopwatchbyaccreditedresearchstaff(no.AB1701.PolishAccreditationCenter)fromtheparamediclaboratoryattheUniversityinBielsko-Biała.ThemannikinwascorrectlyintubatedbyalaboratorytechnicianoverseeingthestudyusinganID7.5endotrachealtubeandaMackintoshlaryngoscopetoadepthof21cm,measuredfromtheleftcornerofthemouth.Everyprocedurewasconductedusingabrand-newendotrachealtube,whichwasinsertedintothemannikin’stracheaimmediatelybeforethecuffwasinflated.Thestudycomprisedtwoproceduremethodsconductedoneaftertheother.MethodA–involvedsealingtheendotrachealtubecuffusinga20mlsyringe.Thetimewasmeasuredfromthemomentthesyringewaspickeduptothemomentthecuffinflationprocedurewascompletedandthesyringeplacedbesidethemannikin.Immediatelyaftertheprocedurewascompleted,thepressureinthecuffsealwasmeasuredusingaclockmanometer.MethodB–involvedsealingtheendotrachealtubecuffusingaVBM(Germany)deviceforinflatingtheendotrachealtubecuff.Thetimewasmeasuredfromthemomentthedevicewaspickeduptothemomentthecuffinflationprocedurewascompletedandthedeviceplacedbesidethemannikin.Immediatelyaftertheprocedurewascompleted,thepressureinthecuffsealwasmeasured.AnalysiswasalsoconductedoftheeffectofcompletedcertifiedAdvancedLifeSupport(ALS)orACLStrainingcoursesonthequalityofendotrachealtubecuffsealinflation.AlldatawasenteredintoaspeciallypreparedExcel2019spreadsheetandsubjectedtostatisticalanalysis.StatisticalAnalysis7Theassumedlevelofsignificanceadoptedwasp = 0.05.Variablesexpressedattheordinalornominallevelwereanalyzedusingtestsbasedonchi-squaredistribution.For2 × 2tables,acontinuitycorrectionwasapplied,whileifconditionswerenotmetfortheapplicationofachi-squaredtest,apreciseFishertestwasusedandextendedfortableslargerthan2 × 2.Parametrictestswereusedtoanalyzequantitativevariablesdividedintogroups(thestudent’sTtestorANOVAanalysisofvariance),ortheirnon-parametricequivalents(theMann-WhitneyUtestortheKruskal-Wallistest).Theselectionoftestswascarriedoutonthebasisofthedistributionofthevariables,whichwasverifiedusingtheShapiro-Wilktest.ThecalculationswereconductedusinganRstatisticsenvironmentversion3.6.0,PSPPsoftwareandMSOffice2019.ResultsAnalysisoftheresultsobtainedduringthestudyshouldbecomplementedbykeyinformationcontainedinthequestionnairescompletedbeforetheprocedurewascarriedout.Allthestudyparticipantsindicatedthatthemethodtheyalwaysusedinconductingadvancedparamedictreatmentwasthatofinflatingtheendotrachealtubecuffbysyringe.Themaindeterminanttheyindicatedforthischoiceofmethodwasthetimerequiredforcarryingouttheprocedure.Noneofthestudyparticipantsgaveinflatedcuffvolumeasaconditionthatdeterminedcorrectcompletionoftheprocedure.Table2.StatisticaldescriptionofanalyzedparametersNMSDMinMaxQ25MepressureA(cmH2O)10883.2535.458.00120.0055.5093.00pressureB(cmH2O)10826.703.3720.0030.0024.0028.00cuffinflationtime-MethodA1087.543.693.3024.865.216.40cuffinflationtime-MethodB1089.713.694.4022.167.108.98N–amount;M–mean;SD–standarddeviation;Min–minimum;Max–maximum;Q25–firstquartile;Me–median;Q75–thirdquartile8Table3.ComparisonofMethodAandMethodBintermsofendotrachealcuffinflationpressureUpMinMaxQ25MeCuffinflationpressureMethodA(cmH2O)1062.00<0.0018.00120.0055.5093.00MethodB(cmH2O)20.0030.0024.0028.00U – test statistic; p –statistical significance; Me – median; Min – minimum result; Max – maximumresult;Q25–firstquartile;Q75–thirdquartileThepressureincuffsinflatedusingMethodAwassignificantlyhigherstatisticallythanincuffsinflatedusingMethodBp < 0.05.Table4.ComparisonofcuffinflationpressuresaccordingtopressurerangeforbothmethodsMethodAMethodBΧ2dfPressure<20cmH205(4.63%)0(0%)193.2632Pressure20–30cmH206(5.56%)108(100%)Pressure>30cmH2097(89.81%)0(0%)InflationofendotrachealtubecuffsusingMethodAsignificantlylessoftenstatisticallyresultedinachievementoftherequiredpressure(20cmH2O–30cmH2O)thanduringuseofMethodB.Table5.ComparisonofMethodAandMethodBintermsofendotrachealtubecuffinflationtime9UpMinMaxQ25MeCuffinflationtimeMethodA–(sec)3324.00<0.0013.3024.865.216.40MethodB–(sec)4.4022.167.108.98U – test statistic; p – statistical significance; Me – median; Min – minimum result; Max – maximumresult;Q25–firstquartile;Q75–thirdquartileThetimerequiredtoinflatedtheendotrachealtubecuffusingMethodBwassignificantlylongerstatisticallythanthetimerequiredusingMethodAp < 0.05.Table6.ComparisonofMethodAandMethodBcuffinflationpressuredependingoncompletedALSandACLStrainingcoursesUpMinMaxQ25MeMethodApressure(cmH2O)completedALS,ACLS1060.000.1858.00120.0046.0082.00notraining34.00120.0076.50101.00MethodBpressure(cmH2O)completedALS,ACLS1063.500.18620.0030.0024.0028.00notraining20.0030.0022.5028.00U–teststatistic;p–statisticalsignificance;Me–median;Min–minimumresult;Max–maximumresult;Q25–firstquartile;Q75–thirdquartileCompletionofcertifiedALSandACLStrainingcoursesdidnothaveastatisticallysignificantinfluenceoncorrectinflationoftheendotrachealtubecuffrelativetothemethodappliedp>0.05.10Table7.Comparisonoftimerequiredforinflationofendotrachealtubecuffrelativetothemethodappliedandcompletedcertifiedtrainingcourses11UpMinMaxQ25TimeA(sec)completedtraining926.500.0293.7224.865.44notraining3.3013.244.75TimeB(sec)completedtraining1195.500.6824.4021.007.16notraining5.2022.167.10PressureA(cmH2O)completedtraining1060.000.1858.00120.0046.00notraining34.00120.0076.50PressureB(cmH2O)completedtraining1063.500.18620.0030.0024.00notraining20.0030.0022.50U–teststatistic;p–statisticalsignificance;Me–median;Min–minimumresult;Max–maximum12result;Q25–firstquartile;Q75–thirdquartileParticipantswhohadcompletedcertifiedALSandACLStrainingcoursestooksignificantlylongertoinflatetheendotrachealtubecuffusingmethodAthanparticipantswhohadnotcompletedatrainingcourse.Table8.ComparisonofcuffinflationvolumebysyringeinrelationtocompletedcertifiedALSandACLStrainingcoursesCompletedALS/ACLStraining(n=74)NoALS/ACLStraining(n=34)Χ2dfPressure<20cmH205(6.76%)0(0%)5.62720.05Pressure20–30cmH206(8.11%)0(0%)Pressure>30cmH2063(85.13%)34(100%)Amongtheparticipantsinthestudygroup(n = 74)whohadcompletedcertifiedALSorACLStraining,6.76%achievedaresultoflowerthan20cmH2O,8.11%achievedaresultaroundthenorm,while85.13%achievedaresultthatexceededavolumeof30cmH2O.Amongthestudygroupparticipants(n = 34)whohadnotcompletedcertifiedtraining,allachievedaresultthatexceededavolumeof30cmH2O.Completionofcertifiedtrainingdoesnotthereforehaveastatisticallysignificantinfluenceonachievingthecorrectinflationvolumeoftheendotrachealtubecuffp = 0.059.13DiscussionInflationoftheendotrachealtubecuffcanbedoneusingasyringeorbyemployingamanometer[12–14].Theauthorsofthisstudyanalyzednumerouspublicationsthatassessedtheeffectivenessofusingamanometerduringinflationoftheendotrachealtubecuff.Innoneofthesepaperswastheredetailedinformationontheavailabilityofthistypeofdeviceamongparamedicteams.Thequestionnaireusedinourstudyshowedthatduringtheirworkinemergencymedicalteams,noneofthestudyparticipantsusesamanometertoinflatetheendotrachealtubecuff.Thisisexplainedbythenecessitytosavetimewhenworkingintwo-personteams.Inflationoftheendotrachealtubecufftothecorrectpressureisoneoftheelementsinproperlyconductedintubation.Carryingouttheprocedurecorrectlyisdifficultwhenitisdoneusingasyringe[2–5].Numerousresearchstudies,bothinclinicalandinsimulatedconditions,haveshownuseofasyringealmostalwaysresultsinexcessiveinflationofthecuff[4,5,13,14].Over-inflationoftheendotrachealtubecuffleadstopatientcomplicationsintheareaofthetrachea,resultinginischemia,infectionornarrowingofthewindpipe[15].Intheirresearchconductedonhumancorpses,Sudhoffetal.demonstratedthatexcessiveinflationoftheendotrachealtubecuffcancauseruptureofthetrachea[16].Ourresearchhasshownthatwhenparamedicsusedamanometer,theycorrectlyinflatedtheendotrachealtubecuffoneveryoccasion,whichdemonstratesastatisticallysignificantdifferencebetweenthismethodandthatinvolvinguseofasyringe.Whenasyringewasused,accompaniedbyassessmentofinflationusingthepalpationmethod,overinflationoccurredin89%ofcases.Inotherresearchesintoproperendotrachealtubecuffinflation,authorsindicatethattoolowpressureinthecuffsealcancausecomplicationssuchastubedisplacementorinhalationofstomachcontentsintotheairways[9–11].Ourresearchshowedthatinsufficientcuffinflationoccurredin4.63%ofcases,andtookplaceonlywhenasyringewasused.TrainingcertifiedbytheAHAandERCinadvancedresuscitationtechniquesfocusonteachingpracticalskillsthatcanbelaterusedduringclinicaltreatment.Thistrainingincludesworkshopsoncorrectlyensuringandmaintainingunobstructedairways[17–20].Manystudiesthatincludeanassessmentofthecorrectprocedureformaintainingunblockedairwayspointtotheneedforimprovementintheskillsrequiredforproper14inflationoftheendotrachealtubecuff[4,5,12,13].OurresearchhasprovedthatparticipationincertifiedALSandACLStrainingcoursesdoesnothaveasignificantinfluenceoncorrectinflationoftheendotrachealtubecuff,irrespectiveofwhichmethodisused.TheresearchdemonstratedthatparamedicswithanALSorACLScoursecompletioncertificatetookastatisticallysignificantlongertimetoinflatethecuffbysyringethanthosewhohadnotcompletedatrainingcourse.Thismayindicatethatincontrasttomedicalstaffwhohavenotcompletedatrainingcourse,paramedicswhohavecompletedsuchacoursearemoreawareoftheimportanceandnecessityofcorrectlyinflatingthecuff,althoughunfortunatelythisisnotreflectedinachievementofthepropercuffpressure.ConclusionsEmergencymedicalteamsshouldbeequippedwithdevicesforinflatingtheendotrachealtubecuff.Thepalpationmethodisineffectiveinassessingtheproperinflationoftheendotrachealtubecuff.MoreemphasisshouldbeplacedonproperinflationoftheendotrachealtubecuffduringworkshopsonmaintainingunobstructedairwaysconductedduringcertifiedALSandACLStrainingcourses.AbbreviationsCPR-cardiopulmonaryresuscitationALS–AdvancedLifeSupportACLS-AdvancedCardiovascularLifeSupportDeclarationsEthicsapprovalandconsenttoparticipate:TheresearchwasapprovedbytheethicalcommitteeofUniversityinBielsko-BiałaunderthereferencenumberRNN/305/2019–approval2019/12/7/2.Consentforpublication:NotApllicableAvailabilityofdataandmaterials:Thedatasetsusedand/oranalysedduringthecurrentstudyareavailablefromthecorrespondingauthoronreasonablerequest.Competinginterests:Theauthorsdeclarethattheyhavenocompetinginterests.15Funding:ThestudywasfinancedfromthefundsoftheFacultyofHealthSciencesUniversityofBielsko-Biala.Thisresearchdidnotreceiveanyspecificgrantfromfundingagenciesinthepublic,commercial,ornot-for-profitsectors.Authors'contributions:TI,MK,MM,WW,conceivedthestudy,designedthetrial,RB,MĆ,MR,PB,MS,BK,AS,JBsupervisedthedatacollectionTI,MM,MR,SB,AU,providedstatisticaladviceonstudydesignandanalysedthedata,MR,TIdraftedthemanuscript,andallauthorscontributedsubstantiallytoitsrevisionLimitations:Thestudyauthorsareawarethattheresearchhasacertainlimitationinthatitwasconductedinsimulatedconditionsonarelativelysmallgroupofparticipants.Theresearchwasalsoconductedusingamanometerwithapressurerangeuptoamaximumof120cmH2O,anditwasthereforenotpossibletorecordhighervalues.Broaderresearchisplannedonalargergroupofparticipantsinasrealisticconditionsaspossible.References1.SoarJ,MaconochieI,WyckoffMH,OlasveengenTM,SingletaryEM,GreifR,etal.2019InternationalConsensusonCardiopulmonaryResuscitationandEmergencyCardiovascularCareScienceWithTreatmentRecommendations.Resuscitation.2019Dec;145:95-150.doi:10.1016/j.resuscitation.2019.10.016.Epub2019Nov14.–dostęp09.02.20202.LimH,KimJH,KimD,LeeJ,SonJS,KimDC,etal.Trachealruptureafterendotrachealintubation:areportofthreecases.KoreanJAnesthesiol.2012Mar;62(3):277-803.SenguptaP,SesslerDI,MaglingerP,WellsS,VogtA,DurraniJ,etal.Endotrachealtubecuffpressureinthreehospitals,andthevolumerequiredtoproduceanappropriatecuffpressure.BMCAnesthesiol.2004Nov;4(1):8.164.HoffmanRJ,ParwaniV,HahnIH.Experiencedemergencymedicinephysicianscannotsafelyinflateorestimateendotrachealtubecuffpressureusingstandardtechniques.AmJEmergMed.2006Mar;24(2):139-43.5.MichligSA.Anaestheticstaffcannotidentifyextremelyhightrachealtubecuffpressuresbypalpationofthepilotballoon.BrJAnaesth.2013Aug;111(2):300-1.6.BouattourK,Prost-LapeyreA,Hauw-BerlemontC,DiehlJ-L,GuerotE.Apost-intubationtrachealruptureinintensivecareunit.AnnFrAnesthReanim2014;33:590-5927.Alvarez-MaldonadoP,VidalE,Ceron-DiazU.Trachealulcersduetoendotrachealtubecuffpressure.JBronchologyIntervPulmonol2011;18:288-2898.HameedAA,MohamedH,Al-MansooriM.Acquiredtracheoesophagealfistuladuetohighintracuffpressure.AnnThoracMed2008;3:23-259.EstesRJ,MeduriGU.Thepathogenesisofventilator-associatedpneumonia:I.Mechanismsofbacterialtranscolonizationandairwayinoculation.IntensiveCareMed1995;21:365-383.10.HamiltonVA,GrapMJ.Theroleoftheendotrachealtubecuffinmicroaspiration.HeartLung2012;41:167-172.11.LauACW,SoHM,TangSL,YeungA,LamSM,YanWW.Preventionofventilator-associatedpneumonia.HongKongMedJ2015;21:61-68.12.PetersJH,HoogerwerfN.Prehospitalendotrachealintubation;needforroutinecuffpressuremeasurement?EmergMedJ.2013Oct;30(10):851-3.doi:10.1136/emermed-2012-201388.13.HarmF,ZuercherM,BassiM,UmmenhoferW.Prospectiveobservationalstudyontrachealtubecuffpressuresinemergencypatients--isneglectingtheproblemtheproblem?ScandJTraumaResuscEmergMed.2013Dec4;21:83.doi:10.1186/1757-177241-21-83.14.HedbergP,EklundC,HogqvistS.IdentificationofaVeryHighCuffPressurebyManualPalpationoftheExternalCuffBalloononanEndotrachealTubeAANAJournal2015;83(3):179-182.15.ChanSM,WongCS,CherngCH.Determininganoptimaltrachealtubecuffpressurebythefeelofthepilotballoon:atrainingcoursefortraineesprovidingairwaycareActaAnaesthesiolTaiwan2009;47(2):79-83.16.SudhoffTH,SeidlRO,EstelB,CoordesA.ClinExpOtorhinolaryngol2015;8(4):409–415.17.SoarJ,NolanJP,BöttigerBW,PerkinsGD,LottC,CarliP,etal.EuropeanResuscitationCouncilGuidelinesforResuscitation2015Section3.Adultadvancedlifesupport.Resuscitation2015;95:100-147.18.LinkMS,BerkowLC,KudenchukPJ,HalperinHR,HessEP,MoitraVK,etal.Part7:AdultAdvancedCardiovascularLifeSupport:2015AmericanHeartAssociationGuidelinesUpdateforCardiopulmonaryResuscitationandEmergencyCardiovascularCare.Circulation2015;132:444-464.19.MonsieursKG,NolanJP,BossaertLL,GreifR,MaconochieIK,NikolaouNI,etal.EuropeanResuscitationCouncilGuidelinesforResuscitation2015:Section1.Executivesummary.Resuscitation2015;95:1-80.20.NeumarRW,ShusterM,CallawayCW,GentLM,AtkinsDL,BhanjiF,etal.Part1:Executivesummary:2015AmericanHeartAssociationGuidelinesUpdateforCardiopulmonaryResuscitationandEmergencyCardiovascularCare.Circulation2015;132:315-367. | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/360248013Consequences of Chinese tariff and US MFP payments on world soybean‐complex marketsArticle in Journal of the Agricultural and Applied Economics Association · April 2022DOI: 10.1002/jaa2.12CITATION1READS223 authors:Some of the authors of this publication are also working on these related projects:Farm Policies View projectTrade Policy View projectAndrii BaryshpoletsTexas Tech University2 PUBLICATIONS 2 CITATIONS SEE PROFILEStephen DevadossTexas Tech University205 PUBLICATIONS 2,132 CITATIONS SEE PROFILEEthan SabalaUnited States Department of Agriculture6 PUBLICATIONS 33 CITATIONS SEE PROFILEAll content following this page was uploaded by Stephen Devadoss on 01 May 2022.The user has requested enhancement of the downloaded file.Received: 7 March 2022 |Accepted: 12 April 2022DOI: 10.1002/jaa2.12A R T I C L EConsequences of Chinese tariff and US MFPpayments on world soybean‐complex marketsAndrii Baryshpolets1|Stephen Devadoss2|Ethan Sabala31PricewaterhouseCoopers, Los Angeles,California, USA2Department of Agricultural and AppliedEconomics, Texas Tech University, Lubbock,Texas, USA3USDA, ERS, MTED, ITD, Washington,District of Columbia, USACorrespondenceStephen Devadoss, Department of Agriculturaland Applied Economics, Texas Tech University,Lubbock, TX, USA.Email: stephen.devadoss@ttu.eduFunding informationNational Institute of Food and Agriculture,Grant/Award Number: 2020‐67023‐30962AbstractThis study develops a theoretical and empirical spatialequilibrium model to determine the effects of removingthe Chinese soybean tariff and US MFP payments onworld soybean‐complex markets. The model includesvertical and horizontal linkages of the soybean, soy oil,and soy meal markets, thus capturing the downstreameffects of tariff and MFP payment removal. The studyfinds that removing these policies generates overall gainsfor the United States and the world, but China incurswelfare losses due primarily to reduced tariff revenues.These impacts of tariff and MFP payment eliminationreverberate to all countries participating in soybean, soyoil, or soy meal trade.K E Y W O R D Ssoybeans, soy meal, soy oil, spatial equilibrium model, subsidies, tariffJ E L C L A S S I F I C A T I O NF13, F141|INTRODUCTIONChina, in its tit‐for‐tat tariff war, imposed a 25% tariff on soybean imports from the United States.Imposition of this tariff was particularly tactful as US soybean exports to China were valued at12.22 billion dollars and accounted for 57% of US soybean exports in 2017 (FAS, 2021a). Before thedispute, the United States was the world's largest soybean producer with 116.93 million metric tons(MMT) in 2017 (FAS, 2021b). The tariff curtailed Chinese demand for US soybeans, and USproduction fell to 96.67 MMT in 2020 (a 17.33% reduction), which resulted in Brazil surpassing theJnl of Agr & App Econ Assoc. 2022;1–12.wileyonlinelibrary.com/journal/jaa2|1This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproductionin any medium, provided the original work is properly cited.© 2022 The Authors. Journal of the Agricultural and Applied Economics Association published by Wiley Periodicals LLC on behalf ofAgricultural and Applied Economics Association.The findings and conclusions in this manuscript are authors' and should not be construed to represent any official USDA or US Governmentdetermination or policy.United States as the world's largest soybean producer with 137 MMT in 2020 (FAS, 2021b). Becauseof the export loss in the Chinese market, the US government implemented the Market FacilitationProgram (MFP) to support US soybean producers. The purpose of this study is to examine theeffects of removing China's bilateral soybean tariff and MFP payments on United States, Chinese,and world soybean, soy oil, and soy meal markets.Soybeans are an intermediate input used in the production of soy oil and meal. Consequently,trade reallocations and production and consumption changes caused by the tariff had ripple effectsthat affected downstream soy oil and meal markets globally. For instance, after the implementationof the Chinese soybean tariff, China's soy meal production decreased from 71.28 MMT in 2017/18 to67.32 MMT in 2018/19 (FAS, 2021b). This is a significant decrease considering China's soy mealproduction had been steadily increasing every year before the tariff implementation. Additionally,China's domestic consumption of soy oil decreased from 16.50 MMT in 2017/18 to 15.89 MMT in2018/19. Though this drop may seem small, it is noteworthy as it interrupted a previously positivetrend in consumption. By contrast, US domestic consumption of soy oil increased from 9.70 MMTin 2017/18 to 10.38 MMT in 2018/19 (FAS, 2021b). Another interesting aspect of the world soybeancomplex is that China imports soybeans and exports meal. Consequently, changes in the Chinesesoybean market also affect the international soy meal market. Thus, the Chinese soybean tariff hassignificant impacts on not only the world soybean market, but also soy oil and meal markets, whichhas not been explored in the literature to the best of our knowledge.Chinese tariffs created significant market disruptions as the large Chinese import market was shutout for US soybeans, and US exporters were not able to find other markets to fully reallocate exportlosses. Because of the lost export revenues, the US government responded by implementing the MFPwhich provided about $25 billion in payments for many agricultural commodity producers in 2018and 2019 (Glauber, 2021). Because soybeans are one of the largest agricultural products, US soybeanfarmers were leading recipients of MFP payments, with rates of $1.65/bu in 2018 and $2.05/bu in 2019(Paulson et al., 2020). Though these payments helped US soybean producers, it brought additionalmarket interventions, caused shifts in world trade, and led to inefficient resource allocations.In early 2020, the United States and China officially agreed to a Phase One trade deal in whichChina committed to, throughout 2020, purchasing $12.50 billion worth of US agricultural productsabove what it purchased in 2017. In reality, China only purchased $6.50 billion more in 2020 than itdid in 2017 (Nigh & Nepveux, 2021). However, the value of US soybean exports to China increasedfrom $8.0 billion in 2019 to $14.08 billion in 2020, which is greater than the $12.22 billion in USexports in 2017, just before the dispute (FAS, 2021a). This drastic change generated by the PhaseOne trade agreement demonstrates the large potential gains of reaching a permanent resolve to theUS–China trade dispute.Our study contributes to the literature by undertaking in‐depth policy analyses of the impacts ofremoval of the Chinese soybean tariff and US MFP payments on United States, Chinese, and worldsoybean, soy meal, and soy oil markets by incorporating vertical and horizontal linkages. Theseimpacts include changes in prices, production, consumption, bilateral trade volumes, and welfarefor all major trading countries. Out results show that the effects of Chinese soybean tariff removalreverberates throughout the international soybean, oil, and meal markets. These ripple effectsimpact production, consumption, bilateral trade, and welfare for every country that trades in theworld soybean complex. Notably, the United States sees a large welfare gain as US producers expandproduction and exports to China. China incurs an overall welfare loss as a consequence of tariffremoval in large part due to the reduction in tariff revenue. Most exporting countries endure lossesin welfare as the United States becomes more competitive in the Chinese import market, but otherimporters experience welfare gains as these exporters sell more to these importing countries.Overall, the total global welfare increases by $2.08 billion across the three commodities. The USMFP removal results in a net gain in the United States as the government cost saving exceeds theloss to producers. Furthermore, producers in other countries also benefit from US MFP paymentremoval.2|CONSEQUENCES OF CHINESE TARIFF AND US MFP PAYMENTS2|PAST STUDIESSeveral studies have investigated the impacts of the United States–China trade dispute on specificcommodity markets,1 and in the interest of brevity, we review only studies related to soybeans.Taheripour and Tyner (2018) examined the effects of 25% Chinese tariffs on US soybeans and othercommodities using the Global Trade Analysis Project (GTAP) model. They estimated that USsoybean production fell by about 11%–15%, US soybean producer price dropped roughly 4%–5%,and US soybean exports to China decreased between 47.7% and 90.6%. Zheng et al. (2018) foundthat because of Chinese retaliatory tariffs, US soybean exports to China decreased by 34.2%, USprices fell 3.9%, and the United States suffered a $1.85 billion loss in producer surplus. Regmi (2019)observed Chinese soybean imports from the United States dropped by 72% from $11.2 billion in July2017 to $3.1 billion in June 2018.Adjemian et al. (2019) noted that US soybean exports to China drastically declined by 20.1MMT in 2018/2019 marketing year from an average of 30.9 MMT over the three year periodbetween 2015/2016 and 2017/2018. Hitchner et al. (2019) observed that the 25% Chinese importtariff on US soybeans led Brazil and Argentina to supply 92% of Chinese imports between July andDecember 2018. In response, the United States diverted its soybean exports to the European Union,Egypt, Pakistan, Taiwan, Vietnam, and Mexico. Moreover, these exports are considerably lower thanthe exports to China, resulting in a significant decline in US soybean total exports. Based on an expost analysis using gravity equations, Grant et al. (2021) found that US soybean export loss due tothe Chinese tariff was the most drastic, resulting in an average of $10.7 billion in losses in 2018 and2019. Adjemian et al. (2021) employed the relative price of a substitute (RPS) method to determinethe effects of the Chinese 25% soybean tariff on US and Brazilian soybean prices. They found thatthe tariff lowered the US soybean price by 65 cents per bushel and increased the Brazilian price by95 cents per bushel.Giri et al. (2018) estimated that soybean producers who met all the requirements for MFPpayments would be more than compensated for any short‐term losses caused by the Chinese tariffon soybeans. The findings of Janzen and Hendricks (2020) also corroborate that MFP paymentsmore than offset the adverse impacts of trade disruptions caused by the trade war (also see Paulsonet al., 2020). Glauber (2021) observed that MFP payments could run counter to WTO commitmentson allowable domestic supports, which can bring formal complaints from other countries tothe WTO.Sabala and Devadoss (2018) is relevant to our current work as it employed the spatialequilibrium model (SEM) to analyze the effects of the Chinese soybean tariff on world soybeanmarkets, but not soy meal and soy oil. Their analysis indicated that the soybean tariff led to manytrade diversions, generating welfare gains for many countries, but welfare losses for others. Forinstance, China significantly increased its imports of Brazilian soybeans, leading to a large welfaregain for Brazil. However, as China reduced its imports of US soybeans, the United States diverted itsexports to other importers such as the European Union, which undercut Canada's exports to theEuropean Union and led to an overall decrease in Canadian exports and welfare.Devadoss et al. (2019) utilized a dynamic optimization model to analyze Argentina's export taxpolicies on soybeans, soy oil, soy meal, cattle, and beef. They observed that a removal of these exporttaxes would increase production, exports, and prices for Argentina's producers. Baryshpolets andDevadoss (2021) developed a SEM for sunflower complex with sunflower seed as an intermediateinput and sunflower oil and meal as final outputs, and used this model to study the effects of theUkraine‐EU free trade agreement on the world sunflower complex markets. The modelingframework of Baryshpolets and Devadoss (2021) is relevant for our study in modeling the vertical1For the impacts of United States–China trade dispute on other commodities see Nti et al. (2019) for pork, Sabala and Devadoss (2021) for cotton, Sabala and Devadoss(2022) for sorghum, and Morgan et al. (2022) for several commodities.BARYSHPOLETS ET AL.|3and horizontal linkages in the soybean complex. Building on the work of Sabala and Devadoss(2018) and using the methodology of Devadoss et al. (2019) and Baryshpolets and Devadoss (2021),we developed a SEM for the soybean complex to determine the impacts of removing the Chinesesoybean tariff and US MFP soybean payments on world soybean, soy oil, and soy meal markets.23|THEORETICAL MODELThe soybean‐complex market consists of the following components of supply and demand for nregions (i jn,= 1, …, ). First, we model the soybean supply function aspcd Qi=+,,iSsiiis∀(1)where piSs is the supply price of soybeans in country i Q,is is quantity supplied in country i, and ciand di are supply parameters. Demand for soybeans is split between food consumption andcrushing. The demand function for food consumption ispab Qi=−,,iSdiiiFD∀(2)where piSd is the demand price of soybeans in country i Q,iFD is quantity demanded for foodconsumption in country i, and ai and bi are demand parameters.Soy oil q( )iand meal z( )iare two byproducts produced from soybeans used for crushing Q()iCDand a vector of other processing inputs X( ) such as labor, machinery, and energy. The production ofoil and meal byproducts from soybeans capture the vertical link in the soy complex market. Consider aproduction technologyXSQg= min [,( )]iCD, where Sqz=+ii.3 This technology implies byproductsare produced in a fixed proportion using soybeans and other inputs that are substitutable amongthemselves, but not for soybeans. The cost function emanating from this production technology isWWC S ppSc S( ,,) =+( ,)iSdiSd ∗, where W is the vector of input prices. Using this cost function, theprofit function can be written asWπpqpzpSc S=+−− ( ,)iOsiiMsiiSd∗∗∗, where pOs is the price ofsoy oil and pMs is the price of soy meal. Let the fixed‐proportion production of byproducts fromsoybeans be represented as qAQ=CD and zBQ=CD, where A =qQCD and B =zQCD are, respectively,fixed‐proportion conversion coefficients AB(+= 1) indicating the amount of oil and meal producedper‐unit of soybeans. This fixed‐proportion technology implies the demand for soybeans used forcrushing translates into supply functions for oil and meal. Furthermore, the fixed proportion of oil andmeal implies that if we determine production of one product, the production of the other productis also determined. For this production technology, we can rewrite the profit functionasWπpAQpBQpQc Q=+−− (,)OsCDMsCDSdCDCD∗∗∗ ∗∗. The first‐order condition with respecttoQCDisWpApBpc Q+−−(,) = 0OsMsSdCD∗∗,wherethecrushingmarginisCMpApBp=+−OsMsSd∗∗, that is, revenue from oil and meal minus the price of soybeans, on aper‐unit basis. Solving this first‐order condition for QCD and expressing it in inverted linear functionresults in2The SEM model has been extensively used for trade policy analyses: the impacts of Mexican tariff on US apple exports (Devadoss & Ridley, 2014), the influence offormer Soviet Union on world wheat market (Jones et al., 1996), a free trade analysis in the wheat market (Gómez‐Plana & Devadoss, 2004), the consequences ofUnited States–Canadian lumber dispute (Devadoss, 2006), and the effects of United States–Brazil cotton dispute on the world market (Ridley & Devadoss, 2014). Theadvantage of SEM is that it analyzes in great detail the effects of policy changes on trade flow reallocations, prices, supply, demand and welfare of each country in thepartial equilibrium or cross commodity framework. Furthermore, since the SEM is an optimization model, it allows for corner solutions, that is, zero trade flows wheneconomic or policy shocks lead to choke prices.3See Devadoss et al. (2019) for production of joint products of soy oil and meal from soybeans.4|CONSEQUENCES OF CHINESE TARIFF AND US MFP PAYMENTSCMkl Qi=+,,iiiiCD∀(3)where ki and li are parameters.The demand function for soy oil ispef qi=−,,iOdiiid∀(4)where piOd is demand price of soy oil, qid is quantity demanded for soy oil, and ei and fi are demandparameters. The demand function for soy meal ispgh zi=−,,iMdii id∀(5)where piMd is demand price of soy meal, zid is quantity demanded for soy meal, and gi and hi aredemand parameters. Given the fixed proportion of oil and meal supply, the interaction between thetwo products captures the horizontal linkage in the soy‐complex market.The SEM incorporates trade costs, consisting of transport costs and tariffs. The per‐unittransport costs from region i to j are denoted as tijk, where kS=(soybeans), O (soy oil), M (soymeal). The ad valorem tariff imposed by country j on imports from country i is denoted as τijk. Theshipments between countries are denoted as xijk, where country i exports to country j. The SEMdifferentiates between regional price p( ) and market price ρ( ). The difference between regional pricep( ) and market price ρ( ) should be understood as follows (Devadoss, 2013): if regional demand priceis less than the market demand price, that is, buyers are willing to pay less than the market price,then quantity demanded is zero, resulting in a corner solution. If regional and market prices areequal, then quantity demanded is positive, leading to an interior solution. On the supply side, ifregional supply price is higher than the market supply price, that is, marginal cost of production isgreater than the market supply price, quantity supplied is zero, yielding a corner solution. If marketand regional supply prices are equal, then quantity supplied is positive, resulting in an interiorsolution. Three approaches are used to solve the SEM: primal, dual, and mixed complementaryproblem (MCP). As Devadoss (2013) has shown, all three approaches are equivalent. We follow theMCP approach as it is the most concise and intuitively the simplest. The supply and demandfunctions presented above are used in the MCP approach, which consists of Kuhn–Tuckerconditions for (1) demand condition, (2) supply condition, (3) optimal consumption, (4) optimalsupply, and (5) spatial arbitrage. These five Kuhn–Tucker conditions are developed for eachmarket–soybeans, soy oil, and soy meal–by incorporating the horizontal and vertical linkages in thesoy‐complex markets, which are presented in the Supporting Information Appendix. Furthermore,the data and empirical approach are also discussed in the Supporting Information Appendix.4|RESULTSThis section presents simulation results of the impacts of two scenarios: (1) elimination of Chinesetariff and (2) elimination of Chinese tariff and US MFP payment.4.1|Scenario 1: Effects of Chinese tariff removalSupporting Information Appendix Table A1 presents the impacts of Chinese tariff elimination onsupply, demand, and prices for all three commodities. Supporting Information AppendixTables A2–A4 present the trade flow impacts on soybeans, soy oil, and soy meal, respectively.BARYSHPOLETS ET AL.|5Supporting Information Appendix Table A5 presents the welfare impacts. We discuss the results inthese tables concurrently for soybeans first, then for soy oil and meal.4.1.1|SoybeansElimination of the 25% Chinese tariff causes US soybean exports to China to increase by 40.52MMT (174.77%), and the United States diverts its sales from domestic and other importingcountries' markets to China.4 As a result, US food consumption of soybeans decreases by 0.50MMT (4.40%) and use of soybean for crushing declines by 4.29 MMT (7.29%). Because theUnited States increases its sales to China, US exports to the European Union and Mexicodecrease by 5.69 MMT and 5.17 MMT, respectively. The large increase in US exports to Chinaoffsets its reductions in domestic sales and exports to other countries, leading to an increase intotal sales, which causes US price to increase by $66.35 per MT (19.98%) and US production toexpand by 7.76 MMT (6.39%).Removal of the Chinese tariff and the ensuing higher exports to China causes a significantreduction (by $15.45 per MT or 3.60%) in the soybean price in China. In response to the increasedcompetition and lower price, Chinese producers decrease production by 0.61 MMT (1.62%).Chinese soybean use increases by 0.26 MMT (0.72%) for food consumption and 3.69 MMT (4.03%)for crushing. Elimination of the Chinese tariff causes significant trade reallocations which haverepercussions on other countries that participate in world soybean trade. For instance, because ofthe lower price in China, other exporters such as Brazil and Argentina decrease their exports toChina by 31.45 and 4.46 MMT, respectively.5 Brazilian and Argentinian production declines by 2.30MMT (1.45%) and 1.10 MMT (1.42%), respectively. A reduction in exports to China increases theavailability of soybeans in Brazil and Argentina which lowers their domestic prices by 4.26% and4.43%, respectively. These price reductions lead to increased soybean use for food consumption andcrushing in Brazil and Argentina. The remaining exporting countries, Canada, and Paraguay, alsoundergo similar directional changes.An interesting result emerges with regard to importing countries. Intuitively, soybean importingcountries, with the exception of China, will incur losses if China removes its import tariff on USsoybeans. This is because trade liberalization causes the United States to divert its exports fromother importers to China, which decreases the availability of soybeans in other importing countries,increases the price in these countries, and reduces consumption. However, with other exporterslosing a large share of sales in China, these exporters divert their soybeans from China to otherimporting countries, which is enough to offset their losses in imports from the United States. This isseen in Supporting Information Appendix Table A2 where each importing country's loss in importsfrom the United States is exceeded by its gains in imports from other countries. This result can alsobe explained by the fact that removal of the Chinese tariff increases total world production by 3.14MMT (0.70%). As a result, importing countries experience decreases in price and production andincreases in food consumption and crush use. For example, prices in the European Union andMexico decrease by 4.09% and 3.61%, respectively. These price reductions cause domesticproduction in the European Union and Mexico to decline by 43.18 and 3.49 thousand metric tons,respectively, but soybean used for crushing to increase by 176.81 and 619.64 thousand metric tons,respectively.4Earlier studies examine the impacts of Chinese tariff, and therefore are not directly comparable to our study of eliminating this tariff. However, it is worth noting thatSabala and Devadoss (2018) report US soybean exports to China decreased by 51.2 MMT and Adjemian et al. (2019) find US exports fell from 30.9 MMT to 20.1 MMTas a result of the 25% Chinese tariff.5These results corroborate the findings of Sabala and Devadoss (2018) and Hitchner et al. (2019) that United States and Brazil compete intensively in the Chinesemarket.6|CONSEQUENCES OF CHINESE TARIFF AND US MFP PAYMENTS4.1.2|Soy oil and mealIncreases in US exports to China due to the elimination of the Chinese soybean tariff reduce theavailability of soybeans for domestic use in the United States. For instance, soybean use for crushingin the United States declines by 4.29 MMT (7.29%), which leads to a decline in the domestic supplyof soy oil and meal. In the oil market, lower supply results in a price increase (5.49%), and a decreasein domestic demand for oil (0.40%). However, in the meal market, the price decreases (2.49%) andas a result demand for meal increases (0.77%). This is because China is an exporter of soy meal andnot of oil. Removal of the soybean tariff reduces the price of soybeans in China and increasessoybean use for crushing by 3.69 MMT (4.03%). This causes a reduction in the price of Chinese soymeal, an increase in exports, and a decrease in the world price of meal. China is able to capture someof the US meal export markets, reduce US exports, and increase the availability of soy meal in theUnited States. This, in turn, reduces the domestic price of meal in the United States and increasesdemand. In fact, the large increase in Chinese meal exports reduces the world price such that everycountry included in the model experiences a decrease in domestic price and increase in demand forsoy meal. China expands soy meal exports to Canada, EU, Japan, Korea, Vietnam, and ROW (seeSupporting Information Appendix Table A4).Because of increased soybean crushing in China, the oil market experiences similar changes inthat almost every country sees a decrease in domestic price and an increase in consumption, exceptfor the United States and Mexico which experience an increase in domestic price and a decrease inconsumption. In the United States, the domestic oil price increases because China is not a majorexporter of oil (rather a large user of soy oil) and, therefore it does not capture some of the US oilexport markets. Consequently, with a reduction in US soy oil supply, caused by diminished soybeancrushing, availability of soy oil for US domestic market decreases in spite of lower exports. InMexico, domestic supply of soy oil and meal expands. In the meal market, Mexico's increase indomestic supply exceeds its loss in imports from the United States, resulting in lower meal price andhigher meal consumption. However, in the oil market, domestic supply does not increase enough tooffset the lost imports, and thus the availability of oil in Mexico declines, the domestic priceincreases, and consumption decreases.4.1.3|WelfareChanges in prices, consumption, and production impact consumer surplus6, producer surplus, tariffrevenue, MFP payments, and thus net welfare. The United States faces a loss in soybean consumersurplus of $736 million in response to its increase in domestic price. However, this price increaseleads to an increase in surplus of $8.31 billion for US soybean producers. US producers of soy oiland meal suffer a $3.78 billion loss in surplus as their crushing margin decreases due to soybeanprice increase. US soy oil consumers lose $373 million in consumer surplus due to increased soy oilprices. Interestingly, US soy meal consumers see a $334 million increase in surplus despite decreasedUS soy meal production. This is because, as explained above, China captures some of the US soymeal export market share and, in turn, increases the availability of soy meal in the United States,which reduces the domestic price of meal, increases consumption, and expands consumer surplus.The United States incurs cost of $1.94 billion as a result of the MFP payments. US total welfare in allthree markets, after accounting for the MFP payments, increases by $3.27 billion in response to theremoval of the Chinese soybean tariff, which underscores the importance of resolving this trade war.6Note that throughout this section and in Supporting Information Appendix Table A5, by soybean consumer surplus we mean surplus that is acquired by soybean foodconsumption only (please refer to Equations 2 and 3 in Section 3). Consumer surplus in the soybean‐crushing market is equivalent to producer surplus in outputmarkets (Just et al., 2005). Hence, producer surplus for soybean crushers is equivalent to either producer surplus in the soy oil and soy meal market, or consumersurplus in the soybean‐crushing market. To avoid any confusion, we refer to it as producer surplus for crushers.BARYSHPOLETS ET AL.|7Chinese tariff removal causes an increase in the availability of soybeans in China, which reducesthe domestic price and causes a $562 million increase in consumer surplus for soybean foodconsumers and a $576 million decrease in producer surplus. Chinese oil and meal producers benefitfrom the reduction in domestic soybean price and secure $485 million in additional producersurplus. The reduction in China's soybean price increases domestic crush, which in turn reduces theprice of oil and meal in China, causing increases in surplus of $306 million and $639 million for oiland meal consumers, respectively. However, elimination of the Chinese tariff on imports of USsoybeans decreases China's soybean tariff revenue by $1.87 billion. Furthermore, increased domesticproduction of soy oil decreases China's imports of soy oil, leading to a decrease in tariff revenuesfrom oil imports by $26 million. Overall, China's welfare decreases by $482 million dollars as a resultof removing the tariff on US soybeans. China, as a large country, loses its terms of trade advantage.Other soybean exporters, such as Brazil and Argentina, suffer losses because they lose marketshare to the United States in the Chinese soybean market. Specifically, soybean producer surplus inBrazil and Argentina decreases by $2.36 billion and $1.16 billion, respectively. Increased availabilityof soybeans domestically in Brazil and Argentina leads to a price decrease and an increase inconsumption. Consequently, consumer surplus for Brazil and Argentina rises by $617 million and$496 million, respectively. Similarly, soy oil and meal producers in Brazil and Argentina experiencean increase in surplus of $220 million and $194 million, respectively. However, overall, Brazil andArgentina have net losses in welfare with Brazil losing $1.24 billion and Argentina losing $395million. Other exporters each see reductions in overall welfare, similar to Brazil and Argentina, withthe exception of India. India exports meal but does not participate in world soybean trade. Increasesin China's soy meal exports lower the world meal price and subsequently reduce India's mealexports. Therefore, India's soybean producers and crushers suffer losses in surplus as a result of theChinese tariff removal. However, these losses are outweighed by consumer surplus gains in India'ssoybean, oil, and meal markets, and India obtains a net increase in welfare of $63 million.As discussed previously, Chinese tariff removal benefits most other importing countries, whichis due to other exporters' reallocation of trade flows from China to these importing countries. Forexample, the European Union loses 5.69 MMT of imports from the United States as the UnitedStates diverts its exports to China; however, this is offset by an increase of 7.36 MMT in importsfrom Brazil. Consequently, the European Union increases soybean food consumer surplus by $10.19million, oil consumer surplus by $53.72 million, and meal consumer surplus by $279.52 million.Although the European Union incurs losses in soybean producer surplus and oil tariff revenue, thenet welfare change in the European Union is a gain of $357.99 million. Other importers each havesimilar directional changes in surplus and welfare with different magnitudes, with the exception ofMexico. In Mexico, soybean food and soy meal consumer surplus increase, but soy oil consumersurplus decreases. This is because Mexico does not increase its imports of soy oil from otherexporting countries and its increase in domestic production does not offset its loss in imports fromthe United States. Consequently, there is less supply of soy oil in Mexico relative to the baselinescenario and the price increases, leading to a decrease in consumption. Additionally, Mexico'ssoybean and oil tariff revenues decrease, and Mexico suffers an overall loss in welfare of $180.95million. Combining the effects of Chinese tariff removal across all three markets results in anincrease in total world welfare of $2.08 billion. This finding highlights the importance ofundertaking free trade in the soybean‐complex market.4.2|Scenario 2: Effects of elimination of Chinese tariffs and US MFPpaymentsIn this scenario, we examine the effects of simultaneous removal of tariffs and MFP payments, thatis, moving to a pre‐trade war situation, on the world soybean complex market. Since the calibrationof parameters for the baseline scenario is based on the data that assimilates the soybean payments,8|CONSEQUENCES OF CHINESE TARIFF AND US MFP PAYMENTSwe set the soybean payments to zero (as well as the Chinese tariff to zero) to run the simulation forthis scenario.The results ofthis scenario arereportedinthe SupportingInformationAppendixTables A6–A10. Comparisons of the results in Supporting Information Appendix Tables A6–A10to the respective Supporting Information Appendix Tables A1–A5 reveal the impacts of MFPremoval alone, which is the focus of the discussion, and we present important qualitative impacts ofMFP elimination by concurrently discussing the results in Supporting Information AppendixTables A6–A10.The benefits accrued by US soybean farmers from the Chinese tariff removal are diminishedfrom MFP elimination. Because US farmers do not receive MFP payments, production increases lessand US price increases more compared to those under the tariff elimination scenario. With lesssupply, US soybean exports to China increase less (149.88% compared to 174.77% under Scenario1), yet the United States diverts all its exports from other countries to China. Furthermore, with lesssupply, availability of soybeans for US domestic food consumption and crushing falls more thanthose under tariff elimination alone. As a result, soy oil and soy meal production decline isexacerbated. Consequently, soy oil price increases more, leading to a larger decline of soy oilconsumption and the United States does not export soy oil to any countries. However, soy mealprice decreases less and soy meal use increases slightly less than that under Scenario 1. This increasein US soy meal use is attributed to expanded soy meal exports by China (resulting from the highersoybean imports from the United States as a result of the tariff removal), which displace US mealexports to importing countries and more US soy meal is available for the domestic market.With US soybean exports to China, though Chinese production does not decrease much, totalsoybean supply in China is less than under Scenario 1. As a result, Chinese soybean price does notfall as much as in Scenario 1. Consequently the availability of soybeans for crushing is less, and soyoil and soy meal supply increase less, leading to a smaller increase in oil and meal use in China.Since China uses all its oil domestically, it does not export any soy oil. However, China does exportmeal to Canada, the European Union, Japan, Korea, Vietnam, and ROW, but less than it does underScenario 1.US MFP removal mitigates the adverse effects of Chinese tariff elimination on US competitors,such as Brazilian, Argentinian, and Canadian producers. Production in these countries does notdecline as much as in Scenario 1. Brazilian and Argentinian exports to China also do not decline asmuch as those under Scenario 1. Because Brazil diverts more of its exports to China, compared toScenario 1, it exports less to other importing countries. Canada, on the other hand, could notrecover the lost market share in China and also in Japan, Korea, Taiwan, and Vietnam because, inthis scenario too, US exports displace Canadian exports in these markets. However, because of itsclose proximity to the European Union, Canada is able to recover some of the lost sales in thismarket under Scenario 1. With all three (Brazil, Argentina, and Canada) countries exporting moresoybeans in this scenario, compared to those under Scenario 1, their domestic consumption andcrushing demand do not increase much, and consequently they produce and export less oil and mealcompared to those under Scenario 1. Soybean prices in importing countries (such as India, theEuropean Union, Japan, Korea, Taiwan, Vietnam, and ROW) do not fall as much as in Scenario 1,resulting in smaller increases in consumption and smaller decreases in production.Because of the MFP payment removal, the US market price for soybeans increases, which lowersconsumer surplus for soybean consumption more than in Scenario 1. This higher soybean pricetranslates into a higher price for oil, resulting in a larger consumer surplus loss from oilconsumption than in Scenario 1. Soybean producer surplus gain is smaller in this scenario becauseof the elemination of MFP payments. A higher input (soybean) price does not benefit crushers,resulting in a larger producer surplus loss compared to that in Scenario 1. Meal users experience asmaller gain in consumer surplus than in Scenario 1 because the soybean price increase lessens theeffects of the soy meal price decrease resulting from export competition from China in the worldmarket. Chinese gain in consumer surplus in bean, oil, and meal markets are smaller, leading to aBARYSHPOLETS ET AL.|9greater welfare loss: $1.4 billion compared to $482 million in Scenario 1. US MFP removal alsobenefits other exporters as net welfare losses to Brazil, Argentina, Canada, and Paraguay aremitigated. In contrast, benefits accrued to importing countries (the European Union, Japan, Korea,Taiwan, Vietnam, and ROW) from the Chinese tariff removal are reduced by the US MFPelimination. However, the overall world welfare increase ($2.25 billion) is higher in this scenariothan in Scenario 1 ($2.08 billion) because of more efficient reallocation of resources andconsumption due to tariff and MFP removal.5|CONCLUSIONSThe Chinese soybean tariff on imports from the United States is of particular interest due to thelarge volume of soybeans traded between the two countries. The tariff caused a significant decreasein US soybean exports to China, which dramatically affected world soybean trade. In response to theChinese retaliatory tariff, the US government implemented the MFP to compensate US soybeanfarmers for the lost export market. The purpose of this study is to determine the impacts ofeliminating the Chinese soybean tariff and US MFP payments on world soybean, oil, and mealmarkets.We develop a theoretical spatial equilibrium model, which incorporates the vertical andhorizontal linkages of an intermediate good along with two downstream final goods. This SEMis used empirically to analyze the effects of removing (a) the Chinese soybean tariff and (b) theChinese tariff and US MFP payments. The results show that elimination of the tariff hassignificant effects on production, consumption, prices, and bilateral trade flows for the UnitedStates, China, and all other exporting and importing countries. Consequently, producer andconsumer surplus and tariff revenues are affected by the elimination of the tariff and severalcountries have net gains in welfare, while others suffer losses. Chinese tariff removal results in awelfare gain for the United States, but causes an overall loss for China, in large part due to itssignificant reduction in tariff revenues. Most of the importing countries gain as a result of theelimination of the Chinese tariff and other exporters lose. However, due to the complexity oftrade reallocations and the inclusion of vertical and horizontal linkages, there are exceptions:India, an exporting country, gains and Mexico, an importing country, loses. The results showthat elimination of the tariff is welfare improving for the world as a whole, and that the overallglobal welfare gain among the three markets is $2.08 billion.Bilateral trade barriers are generally welfare reducing for both countries involved. However, it ispossible for a large‐importing country to benefit from a tariff. This seems to be the case for theChinese soybean tariff, as this study has shown that removing this tariff reduces overall welfare inChina. However, global welfare is enhanced with the removal of the Chinese soybean tariff.Though US MFP removal takes away the benefits to US producers and soybean users lose fromhigher soybean prices, the savings in the cost of payments outweigh the losses, leading to a net gainin the United States. Furthermore, producers in other countries (both exporting and importingcountries) benefit from US MFP elimination. Consequently, this policy reform also enhances globalwelfare. The policy implications of this study are synergistic with those of nearly every trade study inmodern economics, that is, freer trade allocates resources efficiently and improves the world welfare.These results underscore the importance of moving to freer trade in agricultural markets, which isbeneficial globally.ACKNOWLEDGMENTSThe authors thank the Area Editor Shon Ferguson for his effective review coordination and twoanonymous reviewers for providing useful suggestions. This study is supported by the USDANational Institute of Food and Agriculture, Agricultural and Food Research Initiative CompetitiveProgram, Agriculture Economics and Rural Communities, grant #2020‐67023‐30962.10|CONSEQUENCES OF CHINESE TARIFF AND US MFP PAYMENTSDATA AVAILABILITY STATEMENTThe authors will make the data available upon request.REFERENCESAdjemian, Michael, Shawn Arita, Vince Breneman, Rob Johansson, and Ryan Williams. 2019. “Tariff Retaliation Weakenedthe U.S. Soybean Basis.” Choices 34(4): 1–9.Adjemian, Michael K., Aaron Smith, and Wendi He. 2021. “Estimating the Market Effect of a Trade War: The Case ofSoybean Tariffs.” Food Policy 105: 1–13.Baryshpolets, Andrii, and Stephen Devadoss. 2021. “The Effects of EU‐Ukraine Free Trade Agreement on the World'sSunflower Complex.” European Review of Agricultural Economics 49(1): 1–15.Devadoss, Stephen. 2006. “Is There an End to US‐Canadian Softwood Lumber Disputes?” Journal of Agricultural and AppliedEconomics 38(1): 137–53.Devadoss, Stephen. 2013. “Ad Valorem Tariff and Spatial Equilibrium Models.” Applied Economics 45(23): 3378–86.Devadoss, S., J. Luckstead, and W. Ridley. 2019. “A Dynamic Analysis of the Impacts of Export Taxes: The Case ofArgentinean Soy and Beef Markets.” The World Economy 42(8): 2427–51.Devadoss, Stephen, and William Ridley. 2014. “Effects of the Mexican Apple Tariff on the World Apple Market.” Review ofDevelopment Economics 18(4): 763–77.FAS. 2021a. “Global Agricultural Trade System.” Technical report, United States Department of Agriculture, ForeignAgricultural Sevice. https://apps.fas.usda.gov/gats/ExpressQuery1.aspxFAS. 2021b. “Production, Supply, and Distribution.” Technical report, United States Department of Agriculture, ForeignAgricultural Service. https://apps.fas.usda.gov/psdonline/app/index.html#/app/homeGiri, Anil, E. Wesley F. Peterson, and S. Sharma. 2018. “The Impact of the Market Facilitation Program on Us Soybean,Sorghum and Corn Producers.” Choices 33(4): 1–7.Glauber, Joseph. 2021. “US Trade Aid Payments and the WTO.” Applied Economic Perspectives and Policy 43(2): 586–603.Gómez‐Plana, Antonio G., and Stephen Devadoss. 2004. “A Spatial Equilibrium Analysis of Trade Policy Reforms on theWorld Wheat Market.” Applied Economics 36(15): 1643–48.Grant, Jason, Shawn Arita, R. Johansson, Chaoping Xie, and Charlotte Emlinger. 2021. “Agricultural Exports and RetaliatoryTrade Actions: An Empirical Assessment of the 2018/2019 Trade Conflict.” Applied Economic Perspectives and Policy43: 619–40.Hitchner, J., K. Menzie, and S. Meyer. 2019, 4th Quarter. “Tariff Impacts on Global Soybean Trade Patterns and U.S. PlantingDecisions.” Choices 34(4): 1–9.Janzen, Joseph. P., and N. P. Hendricks. 2020. “Are Farmers Made Whole by Trade Aid?” Applied Economic Perspectives andPolicy 42(2): 205–26.Jones, James, S. Li, S. Devadoss, and C. Fedane. 1996. “The Former Soviet Union and World Wheat Trade.” American Journalof Agricultural Economics 78: 869–78.Just, Richard E., D. L. Hueth, and Andrew Schmitz. 2005. The Welfare Economics of Public Policy: A Practical Approach toProject and Policy Evaluation. Northampton MA: Edward Elgar Publishing.Morgan, Stephen, Shawn Arita, Jayson Beckman, Saquib Ahsan, Dylan Russell, Philip Jarrell, and Bart Kenner. 2022. “TheEconomic Impacts of Retaliatory Tariffs on US Agriculture.” Technical report, ERS/USDA. Economic Research Report304.Nigh, V., and M. Nepveux. 2021. “U.S., China Phase One Agreement–Year One.” https://www.fb.org/market‐intel/u.s.‐china‐phase‐one‐agreement‐year‐oneNti, Frank., Lindsay Kuberka, and Keith Jones. 2019. “Impact of Retaliatory Tariffs on the U.S. Pork Sector.” Choices 34(4):1–8.Paulson, N. D., Allen M. Featherstone, and J. C. Hadrich. 2020. “Distribution of Market Facilitation Program Payments andTheir Financial Impact for Illinois, Kansas, and Minnesota Farms.” Applied Economic Perspectives and Policy 42(2):227–44.Regmi, A. 2019, September. “Retaliatory Tariffs on US Agriculture.” Report to Congress R45903, Congressional ResearchService, Washington, DC.Ridley, William, and S. Devadoss. 2014. “US‐Brazil Cotton Dispute and the World Cotton Market.” The World Economy37(8): 1081–100.Sabala, Ethan, and Stephen Devadoss. 2018. “Impacts of Chinese Tariff on World Soybean Markets.” Journal of Agriculturaland Resource Economics 44(2): 291–310.Sabala, Ethan, and Stephen Devadoss. 2021. “Spatial Equilibrium Analysis of Chinese Tariff on World Cotton Markets.” TheWorld Economy 44(7): 2188–202.Sabala, Ethan, and Stephen Devadoss. 2022. “Analysis of Chinese Tariff on Sorghum Market Under Varying MarketStructure.” Journal of Agricultural and Resource Economics 46(3): 1–15.Taheripour, F., and W. E. Tyner. 2018. “Impacts of Possible Chinese 25% Tariff on U.S. Soybeans and Other AgriculturalCommodities.” Choices 33(2).BARYSHPOLETS ET AL.|11Zheng, Yuqing, Dallas Wood, Holly H. Wang, and J. P. Jones. 2018. “Predicting Potential Impacts of China's RetaliatoryTariffs on the US Farm Sector.” Choices 33(2): 1–6.SUPPORTING INFORMATIONAdditional supporting information can be found online in the Supporting Information section atthe end of this article.How to cite this article: Baryshpolets, Andrii, Stephen Devadoss, and Ethan Sabala. 2022.“Consequences of Chinese Tariff and US MFP Payments on World Soybean‐ComplexMarkets.” Journal of the Agricultural and Applied Economics Association 1–12.https://doi.org/10.1002/jaa2.1212|CONSEQUENCES OF CHINESE TARIFF AND US MFP PAYMENTSView publication stats | RESEARCH GATE |
Simulationsstudien zurortsspezifischen Biokonjugationmaßgeschneiderter PolymereDissertation zur Erlangung des naturwissenschaftlichenDoktorgrades der Julius-Maximilians-Universität Würzburgvorgelegt vonJosef Kehreinaus Bad KreuznachWürzburg 2022This document is licensed under the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License (CC BY-NC-SA 4.0): http://creativecommons.org/licenses/by-nc-sa/4.0 This CC license does not apply to third party material (attributed to another source) in this publication. Eingereicht bei der Fakultät für Chemie und Pharmazie amGutachter der schriftlichen Arbeit1. Gutachter:2. Gutachter:Prüfer des öffentlichen Promotionskolloquiums1. Prüfer:2. Prüfer:3. Prüfer:Datum des öffentlichen PromotionskolloquiumsDoktorurkunde ausgehändigt am nil admirari Sämtliche Arbeiten dieser Dissertation wurden im Zeitraum von Februar 2018 biseinschließlich Dezember 2021 unter der Aufsicht von Prof. Dr. Christoph Sotriffer amInstitut für Pharmazie und Lebensmittelchemie der Julius-Maximilians-UniversitätWürzburg durchgeführt. Die hier vorgestellten Ergebnisse wurden teilweise in wis-senschaftlichen Publikationen veröffentlicht:1. N. Hauptstein, P. Pouyan, J. Kehrein, M. Dirauf, M. D. Driessen, M. Raschig,K. Licha, M. Gottschaldt, U. S. Schubert, R. Haag, L. Meinel, C. Sotriffer und T.Lühmann, „Molecular Insights into Site-Specific Interferon-α2a BioconjugatesOriginated from PEG, LPG, and PEtOx“, Biomacromolecules, Jg. 22, Nr. 11,S. 4521-4534, 2021Bei folgenden Posterbeiträgen wurde mitgewirkt:1. J. Kehrein und C. Sotriffer, „Simulation Studies on the Site-Specific Biocon-jugation of Polymers“, 14th German Conference on Cheminformatics (Mainz),20182. J. Kehrein und C. Sotriffer, „Simulation Studies on the Site-Specific Bioconju-gation of Polymers“, 12th European Workshop in Drug Design (Siena), 20193. J. Kehrein und C. Sotriffer, „Simulationsstudien zur ortsspezifischen Biokonju-gation maßgeschneiderter Polymere“, ProMatLeben - Polymere: 1. Doktoranden-und PostDoc-Konferenz (Berlin), 20194. J. Kehrein, C. B. Nkemeto, T. Lühmann, L. Meinel und C. Sotriffer, „Der EffektPEG-basierter Biokonjugationen auf die konformative Proteinstabilität: Ein si-mulationsbasiertes Vorhersagemodell“, ProMatLeben - Polymere: 2. Doktoranden-und PostDoc-Konferenz (Berlin, virtuell), 20215. N. Hauptstein, P. Pouyan, M. Dirauf, J. Kehrein, M. Raschig, K. Licha, M.Gottschaldt, U. S. Schubert, L. Meinel, C. Sotriffer, R. Haag und T. Lühmann,„Site-selective decoration of Interferon-α2a with PEG alternatives and theirphysical, chemical, and biological characterisation“, ProMatLeben - Polymere2. Doktoranden- und PostDoc-Konferenz (Berlin, virtuell), 20216. J. Kehrein, C. B. Nkemeto, T. Lühmann, L. Meinel und C. Sotriffer, „Effects ofPEG-based bioconjugation on protein conformational stability: A simulation-based model“, ProMatLeben - Polymere: Midtermkonferenz (Berlin, virtuell),20217. N. Hauptstein, P. Pouyan, M. Dirauf, J. Kehrein, M. Raschig, K. Licha, M.Gottschaldt, U. S. Schubert, L. Meinel, C. Sotriffer, R. Haag und T. Lühmann,„Correlation of IFN-α2a bioconjugate stability and polymer hydrophilicity“,ProMatLeben - Polymere: Midtermkonferenz (Berlin, virtuell), 20218. N. Hauptstein, P. Pouyan, K. Wittwer, J. Kehrein, M. Dirauf, C. Gizem, M.Raschig, M. Tully, I. Nischang, U. S. Schubert, C. Sotriffer, C. Pfaller, R. Haag,T. Lühmann und L. Meinel, „Site-selective decoration of Interferon-α2a withPEG alternatives and their physical, chemical, and biological characterisation“,ProMatLeben - Polymere: 3. Doktoranden- und PostDoc-Konferenz (Berlin,virtuell), 2021Bei folgenden Präsentationen wurde mitgewirkt:1. M. Tully, N. Engel, M. Dirauf, N. Hauptstein, J. Kehrein, C. B. Nkemeto, P.Pouyan, K. Wittwer, M. Gottschaldt, K. Licha, T. Lühmann, L. Meinel, C.Pfaller, U. S. Schubert, C. Sotriffer und R. Haag, „PEG-alternative Polymerefür die Biokonjugation an Proteine - Next-PEG“, ProMatLeben - Polymere: 1.Doktoranden- und PostDoc-Konferenz (Berlin), 20192. C. Sotriffer und J. Kehrein, „Simulation-driven model building: Towards pre-diction of site-specific bioconjugation“, 7th Strasbourg Summer School in Che-moinformatics (Straßburg, virtuell), 20203. N. Hauptstein, M. Dirauf, N. Engel, J. Kehrein, C. B. Nkemeto, P. Pouyan,M. Tully, K. Wittwer, R. Haag, K. Licha, M. Gottschaldt, L. Meinel, C. Pfal-ler, U. S. Schubert, C. Sotriffer und T. Lühmann, „PEG-alternative Polymerefür die Biokonjugation an Proteine - Next-PEG“, ProMatLeben - Polymere: 2.Doktoranden- und PostDoc-Konferenz (Berlin, virtuell), 20214. N. Hauptstein, P. Pouyan, M. Dirauf, J. Kehrein, M. Raschig, K. Licha, M.Gottschaldt, U. S. Schubert, C. Sotriffer, L. Meinel, R. Haag und T. Lühmann,„Correlation of IFN-α2a bioconjugate stability and polymer hydrophilicity“,ProMatLeben - Polymere: Midtermkonferenz (Berlin, virtuell), 20215. K. Wittwer, N. Hauptstein, M. Dirauf, N. Engel, J. Kehrein, C. B. Nkemeto,P. Pouyan, M. Tully, R. Haag, K. Licha, M. Gottschaldt, L. Meinel, C. Pfal-ler, U. S. Schubert, C. Sotriffer und T. Lühmann, „PEG-alternative Polymerefür die Biokonjugation an Proteine - Next-PEG“, ProMatLeben - Polymere: 3.Doktoranden- und PostDoc-Konferenz (Berlin, virtuell), 2021Folgende Arbeiten, denen andere Forschungsthemen zugrunde liegen, wurden im ge-nannten Zeitraum ebenfalls publiziert:1. C. Kersten, E. Fleischer, J. Kehrein, C. Borek, E. Jaenicke, C. Sotriffer und R.Brenk, „How To Design Selective Ligands for Highly Conserved Binding Sites:A Case Study Using N-Myristoyltransferases as a Model System“, Journal ofMedicinal Chemistry, Jg. 63, Nr. 5, S. 2095–2113, 20202. J. Wiest, J. Kehrein, M. Saedtler, K. Schilling, E. Cataldi, C. A. Sotriffer, U.Holzgrabe, T. Rasmussen, B. Böttcher, M. Cronin-Golomb, M. Lehmann, N.Jung, M. Windbergs und L. Meinel, „Controlling Supramolecular Structures ofDrugs by Light“, Molecular Pharmaceutics, Jg. 17, Nr. 12, S. 4704–4708, 20203. S. Eltschkner, J. Kehrein, T. A. Le, S. Davoodi, B. Merget, S. Basak, J. D.Weinrich, J. Schiebel, P. J. Tonge, B. Engels, C. Sotriffer und C. Kisker, „ALong Residence Time Enoyl-Reductase Inhibitor Explores an Extended Bin-ding Region with Isoenzyme-Dependent Tautomer Adaptation and DifferentialSubstrate-Binding Loop Closure“, ACS Infectious Diseases, Jg. 7, Nr. 4, S.746–758, 20214. L. Hahn, T. Zorn, J. Kehrein, T. Kielholz, A.-L. Ziegler, B. Sochor, E. S. Li-sitsyna, N. A. Durandin, T. Laaksonen, V. Aseyev, C. Sotriffer, M. Windbergs,A.-C. Pöppler und R. Luxenhofer, „Unravelling a novel mechanism in polymerself-assemblies: An order-order transition based on molecular interactions bet-ween hydrophilic and hydrophobic polymer blocks“, ChemRxiv Preprint, 2021Hinweise zu Tabellen und Abbildungen, die aus Publikationen entnommen, adaptiertoder mithilfe von Rohdaten reproduziert wurden:• Tab. 1.4: Reproduziert und mit neuen Referenzen erweitertNachdruck mit Genehmigung aus: S. L. Baker, A. Munasinghe, H. Murata, P. Lin, K.Matyjaszewski, C. M. Colina und A. J. Russell, „Intramolecular Interactions of Conju-gated Polymers Mimic Molecular Chaperones to Stabilize Protein–Polymer Conjuga-tes“, Biomacromolecules, Jg. 19, Nr. 9, S. 3798–3813, 2018. Copyright 2018 AmericanChemical Society (Link: https://pubs.acs.org/doi/10.1021/acs.biomac.8b00927).• Abb. 3.6, 3.7, 3.8, 3.9: Graphen auf der linken Seite entnommenNachdruck mit Genehmigung aus: P. B. Lawrence, Y. Gavrilov, S. S. Matthews, M.I. Langlois, D. Shental-Bechor, H. M. Greenblatt, B. K. Pandey, M. S. Smith, R.Paxman, C. D. Torgerson, J. P. Merrell, C. C. Ritz, M. B. Prigozhin, Y. Levy und J.L. Price, „Criteria for Selecting PEGylation Sites on Proteins for Higher Thermody-namic and Proteolytic Stability“, Journal of the American Chemical Society, Jg. 136,Nr. 50, S. 17 547–17 560, 2014. Copyright 2014 American Chemical Society (Link:https://pubs.acs.org/doi/10.1021/ja5095183).• Abb. 3.17: komplette Abbildung entnommen• Abb. 5.1: Grafiken zu den 10-kDa-Varianten entnommen• Abb. 5.12: komplette Abbildung entnommen• Abb. 5.20: Grafik in (B) entnommen• Abb. 5.23: Balkendiagramm reproduziert• Abb. 5.24: Graphen reproduziert, Grafik in (D) entnommen• Abb. 5.27: Graphen reproduziert, Grafiken in (B) und (C) entnommen• Abb. 5.29: komplette Abbildung entnommenNachdruck mit Genehmigung aus: N. Hauptstein, P. Pouyan, J. Kehrein, M. Dirauf,M. D. Driessen, M. Raschig, K. Licha, M. Gottschaldt, U. S. Schubert, R. Haag, L.Meinel, C. Sotriffer und T. Lühmann, „Molecular Insights into Site-Specific Interferon-α2a Bioconjugates Originated from PEG, LPG, and PEtOx“, Biomacromolecules, Jg.22, Nr. 11, S. 4521-4534, 2021. Copyright 2021 American Chemical Society (Link:https://pubs.acs.org/doi/10.1021/acs.biomac.1c00775).• Abb. 5.30, 5.32: Balkendiagramme reproduziert• Abb. 5.33: Korrelation in (A) adaptiertDaten entnommen aus: N. Hauptstein, P. Pouyan, K. Wittwer, G. Cinar, O. Scherf-Clavel, K. Licha, T. Lühmann, I. Nischang, U. S. Schubert, C. Pfaller, R. Haag und L.Meinel, „Polymer selection impacts the pharmaceutical profile of site specifically con-jugated Interferon-α2a“, ChemRxiv Preprint. Veröffentlicht unter der Creative Com-mons Attribution License (Link: https://doi.org/10.26434/chemrxiv-2022-j85br).Ein Teil der in dieser Arbeit vorgestellten Molekulardynamik-Simulationen wurde aufHPC-Clustern verschiedener Rechenzentren durchgeführt. Im Folgenden sind die aufBasis der NAMD-Logdateien aufsummierten Simulations- und Rechenzeiten relevan-ter Produktionsläufe aufgelistet:• Curta-Cluster der ZEDAT der FU Berlin11.85 µs bzw. 339.82 Tage Rechenzeit.Für die zur Verfügung gestellte Rechenzeit möchte ich mich beim HPC-Dienst derZEDAT, Freie Universität Berlin bedanken.Link: http://dx.doi.org/10.17169/refubium-26754• Julia-Cluster des Rechenzentrums der JMU Würzburg34.43 µs bzw. 593.21 Tage Rechenzeit.Für die zur Verfügung gestellte Rechenzeit möchte ich mich beim Rechenzentrum derJMU Würzburg bedanken.Link: https://www.rz.uni-wuerzburg.de/dienste/rzserver/high-performance-computing/• Linux-Cluster des Leibniz-Rechenzentrums der Bayerischen Akademie der Wis-senschaften (Garching bei München)18.40 µs bzw. 145.74 Tage Rechenzeit.Für die zur Verfügung gestellte Rechenzeit möchte ich mich beim Leibniz-Rechenzentrumbedanken.Link: www.lrz.de DanksagungIch danke Prof. Dr. Christoph Sotriffer recht herzlich für seine Betreuung und denregelmäßigen Austausch in persönlichen Gesprächen. Ebenso danke ich Prof. Dr. Dr.Lorenz Meinel und Prof. Dr. Tessa Lühmann für meine Einbindung in das Verbund-projekt PEG-alternative Polymere für die Biokonjugation an Proteine (Next-PEG)und weitere Forschungsprojekte der pharmazeutischen Technologie.Weiterhin danke ich dem BMBF und dem VDI für die Finanzierung und Organi-sation dieses Projekts, der ZEDAT der FU Berlin, dem Leibniz-Rechenzentrum inGarching bei München und dem Rechenzentrum der JMU Würzburg für ihre Cluster-Ressourcen, sowie allen Kooperationspartnern (JMU Würzburg, FU Berlin, FSU Je-na, PEI) für die Zusammenarbeit: Prof. Dr. Michael Gottschaldt, Prof. Dr. RainerHaag, Prof. Dr. Kai Licha, Prof. Dr. Tessa Lühmann, Prof. Dr. Dr. Lorenz Meinel,Prof. Dr. Michael Schirner, Prof. Dr. Ulrich S. Schubert, Prof. Dr. Christoph Sotriffer,Dr. Christian Pfaller, Dr. Michael Dirauf, Nora Engel, Dr. Paria Pouyan, Dr. MichaelTully, Kevin Wittwer, sowie ganz besonders Niklas Hauptstein und Cynthia BessemNkemeto.Ich bedanke mich bei allen Kollegen des Arbeitskreises für ihre Unterstützung und dieschöne Zeit in Würzburg: Dr. Sebastian Bothe, Stephan Böhler, Mathias Diebold, ErikEndres, Carina Herbst, Gerald Keller, Dr. Maximilian Kuhn, Dr. Yogesh Narkhede,Dr. Yesid Ramirez, Natalia Yuan Chen, Markus Zehe und Thomas Zimmermann. EinDank gilt ebenso allen Assistenten des ersten Semesters für ihr tatkräftiges Mitwirken:Dr. Sebastian Bothe, Stephan Böhler, Sara Endres, Dr. Paul Güntzel, Gerald Keller,Mohamed Marzouk, Dr. Ruben Pawellek, Dr. Joseph Skaf, Natalia Yuan Chen undMarkus Zehe.Zusätzlich danke ich meiner ehemaligen Praktikantin Katarzyna Miśkiewicz und al-len weiteren Personen, die mir geholfen haben, insbesondere Dorothee Haas, EduardoMerino und Magdalena Nowak. Ebenfalls danke ich Prof. Dr. Ruth Brenk, Dr. Chris-tian Kersten und Dr. Raphael Klein, die mein Interesse für diesen Forschungsbereichbereits während meines Pharmaziestudiums gefördert haben.Allen Personen, mit denen ich in zusätzlichen Projekten zusammenarbeiten durfte,danke ich für die Kooperation: Prof. Dr. Robert Luxenhofer, Jun.-Prof. Dr. Ann-Christin Pöppler, Dr. Sandra Eltschkner, Lukas Hahn, Dr. Johannes Wiest, PD Dr.Frank Waller, Simon Hanio, Jonas Schlauersbach und Markus Zehe.Zuletzt gilt ein ganz besonderer Dank meinen Freunden, vor allem Markus, Mohamedund Paul, sowie meiner Familie für den nötigen Beistand in guten sowie schlechtenZeiten. Inhaltsverzeichnis1Einleitung11.1Polymer-Biokonjugate . . . . . . . . . . . . . . . . . . . . . . . . . . .11.1.1Einführung . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11.1.2Herstellungsverfahren. . . . . . . . . . . . . . . . . . . . . . .51.1.2.1Grafting-Methoden. . . . . . . . . . . . . . . . . . .51.1.2.2Polymersynthese . . . . . . . . . . . . . . . . . . . . .61.1.2.3Konjugationsstrategien. . . . . . . . . . . . . . . . .71.1.3Polymerarten . . . . . . . . . . . . . . . . . . . . . . . . . . . .91.1.3.1Polyethylenglykol. . . . . . . . . . . . . . . . . . . .91.1.3.2Polyvinylpyrrolidon . . . . . . . . . . . . . . . . . . .101.1.3.3Polyglycerole . . . . . . . . . . . . . . . . . . . . . . .101.1.3.4Polyoxazoline . . . . . . . . . . . . . . . . . . . . . . .111.1.3.5Weitere Polymere. . . . . . . . . . . . . . . . . . . .111.1.4Therapeutische Beispiele . . . . . . . . . . . . . . . . . . . . . .121.1.4.1Adagen® . . . . . . . . . . . . . . . . . . . . . . . . .121.1.4.2Neulasta®. . . . . . . . . . . . . . . . . . . . . . . .131.1.4.3Pegasys® . . . . . . . . . . . . . . . . . . . . . . . . .131.1.4.4Macugen®. . . . . . . . . . . . . . . . . . . . . . . .131.1.5Thermodynamische Charakterisierung . . . . . . . . . . . . . .141.1.5.1Temperaturabhängige CD-Spektroskopie. . . . . . .141.1.5.2Thermal Shift Assay . . . . . . . . . . . . . . . . . . .161.1.6Stabilisierungsmechanismen . . . . . . . . . . . . . . . . . . . .171.1.6.1Bisherige Hypothesen . . . . . . . . . . . . . . . . . .171.1.6.2Arbeiten von Joshua L. Price . . . . . . . . . . . . . .181.1.6.3Zusammenfassende Übersicht . . . . . . . . . . . . . .191.2Molekulardynamik-Simulationen. . . . . . . . . . . . . . . . . . . . .211.2.1Grundlagen . . . . . . . . . . . . . . . . . . . . . . . . . . . . .211.2.1.1Abbildung atomarer Wechselwirkungen . . . . . . . .211.2.1.2Bewegungsgleichungen . . . . . . . . . . . . . . . . . .211.2.1.3Wahl des Zeitschritts. . . . . . . . . . . . . . . . . .231.2.1.4Strukturaufbereitung und Equilibrierung. . . . . . .231.2.1.5Periodische Randbedingungen. . . . . . . . . . . . .241.2.2Biokonjugat-Simulationen . . . . . . . . . . . . . . . . . . . . .251.2.3Konvergenzkriterien. . . . . . . . . . . . . . . . . . . . . . . .261.2.3.1Self-consistency checks. . . . . . . . . . . . . . . . .271.2.3.2Weitere Konvergenztests. . . . . . . . . . . . . . . .281.2.4Gaußbeschleunigte Molekulardynamik-Simulationen. . . . . .291.2.5Wohltemperierte Metadynamik-Simulationen . . . . . . . . . .301.3QSPR-Modelle zur Stabilitätsvorhersage . . . . . . . . . . . . . . . . .331.3.1Grundlagen . . . . . . . . . . . . . . . . . . . . . . . . . . . . .331.3.1.1Hintergrund. . . . . . . . . . . . . . . . . . . . . . .331.3.1.2Deskriptoren . . . . . . . . . . . . . . . . . . . . . . .331.3.2Statistische Validierungsparameter . . . . . . . . . . . . . . . .341.3.2.1Regressionsparameter . . . . . . . . . . . . . . . . . .341.3.2.2Klassifikationsparameter. . . . . . . . . . . . . . . .351.3.3Modellarten dieser Studie . . . . . . . . . . . . . . . . . . . . .371.3.3.1Multiple lineare Regression . . . . . . . . . . . . . . .371.3.3.2Logistische, binomiale Regression. . . . . . . . . . .381.3.3.3Künstliches neuronales Netz. . . . . . . . . . . . . .391.3.3.4XGBoost-Modell . . . . . . . . . . . . . . . . . . . . .391.3.4Limitierungen . . . . . . . . . . . . . . . . . . . . . . . . . . . .401.4Constrained-Network-Analyse . . . . . . . . . . . . . . . . . . . . . . .421.4.1Theoretischer Hintergrund . . . . . . . . . . . . . . . . . . . . .421.4.2Awendung des CNA-Programms. . . . . . . . . . . . . . . . .431.5Untersuchte Strukturen. . . . . . . . . . . . . . . . . . . . . . . . . .441.5.1Modellproteine . . . . . . . . . . . . . . . . . . . . . . . . . . .441.5.1.1Pin 1 WW-Domäne . . . . . . . . . . . . . . . . . . .441.5.1.2Her2-Affibody. . . . . . . . . . . . . . . . . . . . . .451.5.1.3Src SH3-Domäne . . . . . . . . . . . . . . . . . . . . .461.5.1.4GCN4 und coil-VaLd. . . . . . . . . . . . . . . . . .471.5.2Biologika. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .481.5.2.1Interferon-α2a . . . . . . . . . . . . . . . . . . . . . .482Zielsetzung493Materialien und Methoden513.1Modellproteine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .513.1.1Wildtyp-Simulationen und unkonjugierte Mutanten . . . . . . .513.1.2Simulierte Konjugate . . . . . . . . . . . . . . . . . . . . . . . .513.1.2.1Konjugate mit publizierten Stabilitätsdaten . . . . . .513.1.2.2Zusätzliche Konjugate . . . . . . . . . . . . . . . . . .523.1.3Ermittlung geeigneter Simulationsparameter . . . . . . . . . . .543.1.3.1PyRED-Methode . . . . . . . . . . . . . . . . . . . . .543.1.3.2antechamber-Methode . . . . . . . . . . . . . . . . . .553.1.3.3Monomerbasierte Parametrisierung. . . . . . . . . .573.1.3.4Problemstellungen . . . . . . . . . . . . . . . . . . . .573.1.3.5Vergleich der Simulationen mit Referenz . . . . . . . .583.1.4Modellierung der Strukturen. . . . . . . . . . . . . . . . . . .633.1.4.1Aufbereitung der Peptidstrukturen . . . . . . . . . . .633.1.4.2Einbau modifizierter Aminosäuren . . . . . . . . . . .633.1.4.3Solvatation der Strukturen . . . . . . . . . . . . . . .643.1.5Simulationsprotokoll . . . . . . . . . . . . . . . . . . . . . . . .643.1.5.1Equilibrierung und erster Produktionslauf . . . . . . .643.1.5.2Fortführung des Produktionslaufs. . . . . . . . . . .653.1.6QSPR-Vorhersagemodelle . . . . . . . . . . . . . . . . . . . . .663.1.6.1Deskriptoren . . . . . . . . . . . . . . . . . . . . . . .663.1.6.2Modellgenerierung: Ansatz 1 . . . . . . . . . . . . . .663.1.6.2.1Aufbereitung der Daten . . . . . . . . . . . .673.1.6.2.2Modellgenerierung . . . . . . . . . . . . . . .693.1.6.2.3Evaluation . . . . . . . . . . . . . . . . . . .703.1.6.3Modellgenerierung: Ansatz 2 . . . . . . . . . . . . . .703.1.6.3.1Ausgeschlossene Konjugate . . . . . . . . . .713.1.6.3.2Definition unterschiedlicher Datensätze . . .713.1.6.3.3Clustering der Daten. . . . . . . . . . . . .723.1.6.3.4Modellgenerierung . . . . . . . . . . . . . . .733.1.6.3.5Evaluation . . . . . . . . . . . . . . . . . . .753.1.6.3.6Finales Modell . . . . . . . . . . . . . . . . .753.1.7Metadynamik-Simulationen . . . . . . . . . . . . . . . . . . . .763.1.7.1Untersuchte Systeme. . . . . . . . . . . . . . . . . .763.1.7.2Kollektivvariablen . . . . . . . . . . . . . . . . . . . .763.1.7.3Simulationsparameter . . . . . . . . . . . . . . . . . .763.1.8Vergleich der Polymere . . . . . . . . . . . . . . . . . . . . . . .773.2Biologika. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .783.2.1Wildtyp-Simulationen . . . . . . . . . . . . . . . . . . . . . . .783.2.2Simulierte Konjugate . . . . . . . . . . . . . . . . . . . . . . . .783.2.3Parameterfindung. . . . . . . . . . . . . . . . . . . . . . . . .803.2.4Modellierung der Strukturen. . . . . . . . . . . . . . . . . . .803.2.5Gaußbeschleunigte Molekulardynamik-Simulationen. . . . . .823.2.6Coarse-grained-Simulation . . . . . . . . . . . . . . . . . . . . .823.2.7Experimentelle Arbeiten . . . . . . . . . . . . . . . . . . . . . .834Ergebnisse: Modellproteine854.1QSPR-Vorhersagemodelle . . . . . . . . . . . . . . . . . . . . . . . . .854.1.1Ansatz 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .864.1.1.1Referenzmodelle . . . . . . . . . . . . . . . . . . . . .864.1.1.2Modelle C1 und D1 . . . . . . . . . . . . . . . . . . .874.1.1.3Modelle E und F . . . . . . . . . . . . . . . . . . . . .894.1.1.4Modelle G und H. . . . . . . . . . . . . . . . . . . .914.1.1.5Leistung der Modelle. . . . . . . . . . . . . . . . . .914.1.1.6Verwendete Deskriptoren . . . . . . . . . . . . . . . .934.1.1.7MD-Analysen. . . . . . . . . . . . . . . . . . . . . .944.1.1.7.1Interaktionen und Solvatation. . . . . . . .944.1.1.7.2Linker verschiedener Länge . . . . . . . . . .954.1.1.7.3D-/L-Asparagin. . . . . . . . . . . . . . . .974.1.1.7.4Linker verschiedener Rigidität. . . . . . . .984.1.1.7.5Desolvatation apolarer Flächen . . . . . . . .994.1.1.8Fazit. . . . . . . . . . . . . . . . . . . . . . . . . . .994.1.2Ansatz 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1004.1.2.1Referenzmodelle . . . . . . . . . . . . . . . . . . . . . 1014.1.2.2Regressionsmodell A2-R6 . . . . . . . . . . . . . . . . 1014.1.2.2.1Allgemeine Leistung . . . . . . . . . . . . . . 1014.1.2.2.2Einzelne Konjugate. . . . . . . . . . . . . . 1024.1.2.2.3Applikationsdomäne . . . . . . . . . . . . . . 1034.1.2.2.4Vergleich zur strukturbasierten Richtlinie . . 1044.1.2.2.5Modelldeskriptoren. . . . . . . . . . . . . . 1054.1.2.2.6Alternative Polymerarten . . . . . . . . . . . 1064.1.2.2.7Modell A2-R6v2 . . . . . . . . . . . . . . . . 1074.1.2.2.8Fazit. . . . . . . . . . . . . . . . . . . . . . 1084.1.2.3Klassifikationsmodell A2-C1. . . . . . . . . . . . . . 1094.1.2.3.1Allgemeine Leistung . . . . . . . . . . . . . . 1094.1.2.3.2Einzelne Konjugate. . . . . . . . . . . . . . 1104.1.2.3.3Modelldeskriptoren. . . . . . . . . . . . . . 1104.1.2.3.4Alternative Polymerarten . . . . . . . . . . . 1114.1.2.3.5Fazit. . . . . . . . . . . . . . . . . . . . . . 1114.1.2.4Zusätzliche Modelle . . . . . . . . . . . . . . . . . . . 1114.1.2.4.1Parametervergleich. . . . . . . . . . . . . . 1114.1.2.4.2Modell A2v2-R1 . . . . . . . . . . . . . . . . 1134.1.2.4.3Modell A2v2-C1 . . . . . . . . . . . . . . . . 1144.1.2.5Fazit. . . . . . . . . . . . . . . . . . . . . . . . . . . 1154.2Metadynamik-Simulationen . . . . . . . . . . . . . . . . . . . . . . . . 1174.2.1Energiediagramme . . . . . . . . . . . . . . . . . . . . . . . . . 1174.2.2Konvergenz . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1174.2.3Energiedifferenzen. . . . . . . . . . . . . . . . . . . . . . . . . 1194.2.4Polymerdynamik . . . . . . . . . . . . . . . . . . . . . . . . . . 1204.2.5Fazit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1224.3Vergleich der Polymere . . . . . . . . . . . . . . . . . . . . . . . . . . . 1234.3.1Pin 1-Konjugate. . . . . . . . . . . . . . . . . . . . . . . . . . 1244.3.1.1Konjugationsstelle 14. . . . . . . . . . . . . . . . . . 1244.3.1.2Konjugationsstelle 16. . . . . . . . . . . . . . . . . . 1264.3.1.3Konjugationsstelle 17. . . . . . . . . . . . . . . . . . 1284.3.1.4Konjugationsstelle 18. . . . . . . . . . . . . . . . . . 1304.3.1.5Konjugationsstelle 19. . . . . . . . . . . . . . . . . . 1324.3.1.6Konjugationsstelle 21. . . . . . . . . . . . . . . . . . 1344.3.1.7Konjugationsstelle 23. . . . . . . . . . . . . . . . . . 1364.3.1.8Konjugationsstelle 26. . . . . . . . . . . . . . . . . . 1384.3.1.9Konjugationsstelle 27. . . . . . . . . . . . . . . . . . 1404.3.1.10 Konjugationsstelle 28. . . . . . . . . . . . . . . . . . 1424.3.1.11 Konjugationsstelle 29. . . . . . . . . . . . . . . . . . 1444.3.1.12 Konjugationsstelle 30. . . . . . . . . . . . . . . . . . 1464.3.1.13 Konjugationsstelle 32. . . . . . . . . . . . . . . . . . 1484.3.2Her2-Konjugate . . . . . . . . . . . . . . . . . . . . . . . . . . . 1504.3.2.1Konjugationsstelle 7 . . . . . . . . . . . . . . . . . . . 1504.3.2.2Konjugationsstelle 8 . . . . . . . . . . . . . . . . . . . 1524.3.2.3Konjugationsstelle 15. . . . . . . . . . . . . . . . . . 1544.3.2.4Konjugationsstelle 21. . . . . . . . . . . . . . . . . . 1564.3.2.5Konjugationsstelle 23. . . . . . . . . . . . . . . . . . 1584.3.2.6Konjugationsstelle 25. . . . . . . . . . . . . . . . . . 1604.3.2.7Konjugationsstelle 29. . . . . . . . . . . . . . . . . . 1624.3.2.8Konjugationsstelle 33. . . . . . . . . . . . . . . . . . 1644.3.2.9Konjugationsstelle 39. . . . . . . . . . . . . . . . . . 1664.3.2.10 Konjugationsstelle 43. . . . . . . . . . . . . . . . . . 1684.3.2.11 Konjugationsstelle 46. . . . . . . . . . . . . . . . . . 1704.3.2.12 Konjugationsstelle 47. . . . . . . . . . . . . . . . . . 1724.3.2.13 Konjugationsstelle 49. . . . . . . . . . . . . . . . . . 1744.3.2.14 Konjugationsstelle 50. . . . . . . . . . . . . . . . . . 1764.3.2.15 Konjugationsstelle 53. . . . . . . . . . . . . . . . . . 1784.3.3Zusammenfassender Vergleich . . . . . . . . . . . . . . . . . . . 1804.3.3.1Pin 1 WW-Domäne . . . . . . . . . . . . . . . . . . . 1804.3.3.1.1Vergleich der berechneten Eigenschaften . . . 1804.3.3.1.2PEG und LPG . . . . . . . . . . . . . . . . . 1844.3.3.1.3Effekt des Linkers . . . . . . . . . . . . . . . 1854.3.3.1.4PMeOx und PEtOx . . . . . . . . . . . . . . 1864.3.3.2Her2-Affibody. . . . . . . . . . . . . . . . . . . . . . 1874.3.3.2.1Vergleich der berechneten Eigenschaften . . . 1874.3.3.2.2PEG. . . . . . . . . . . . . . . . . . . . . . 1924.3.3.2.3LPG . . . . . . . . . . . . . . . . . . . . . . . 1934.3.3.2.4PMeOx und PEtOx . . . . . . . . . . . . . . 1944.4Zusammenfassung. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1955Ergebnisse: Biologika1995.1Gaußbeschleunigte Molekulardynamik-Simulationen. . . . . . . . . . 1995.1.1Polymerdichten . . . . . . . . . . . . . . . . . . . . . . . . . . . 1995.1.2Kontaktprofile. . . . . . . . . . . . . . . . . . . . . . . . . . . 2015.1.2.1IFN-α2a-K31-BCN-10-kDa. . . . . . . . . . . . . . . 2025.1.2.1.1BCN-PEG-10-kDa . . . . . . . . . . . . . . . 2025.1.2.1.2BCN-LPG-10-kDa . . . . . . . . . . . . . . . 2035.1.2.1.3BCN-PEtOx-10-kDa . . . . . . . . . . . . . . 2065.1.2.2IFN-α2a-K31-DBCO-10-kDa. . . . . . . . . . . . . . 2065.1.2.2.1DBCO-PEG-10-kDa . . . . . . . . . . . . . . 2065.1.2.2.2DBCO-LPG-10-kDa . . . . . . . . . . . . . . 2085.1.2.2.3DBCO-PEtOx-10-kDa . . . . . . . . . . . . . 2095.1.2.3Acyl-PEG-10-kDa . . . . . . . . . . . . . . . . . . . . 2105.1.2.4Größere BCN-Konjugate (20 - 50 kDa). . . . . . . . 2115.1.2.4.120 kDa . . . . . . . . . . . . . . . . . . . . . 2115.1.2.4.230 - 50 kDa . . . . . . . . . . . . . . . . . . . 2125.1.3Fazit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2165.2Coarse-grained-Simulation . . . . . . . . . . . . . . . . . . . . . . . . . 2175.2.1Vergleich mit GaMD-Simulationen . . . . . . . . . . . . . . . . 2175.2.2Verwendetes Kraftfeld . . . . . . . . . . . . . . . . . . . . . . . 2195.3Untersuchungen zur thermischen Stabilität. . . . . . . . . . . . . . . 2205.3.110-kDa-Varianten . . . . . . . . . . . . . . . . . . . . . . . . . . 2205.3.1.1CN-Analysen . . . . . . . . . . . . . . . . . . . . . . . 2215.3.1.1.1Generierte Netzwerke . . . . . . . . . . . . . 2235.3.1.1.2Perkolationsindizes . . . . . . . . . . . . . . . 2235.3.1.2Native Kontakte . . . . . . . . . . . . . . . . . . . . . 2235.3.1.3RMSF-Verhältnis. . . . . . . . . . . . . . . . . . . . 2275.3.1.4Vergleich mit LiP-Assays . . . . . . . . . . . . . . . . 2275.3.2Größere Konjugate . . . . . . . . . . . . . . . . . . . . . . . . . 2285.4Untersuchungen zur Bioaktivität . . . . . . . . . . . . . . . . . . . . . 2315.5Zusammenfassung. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2346Fazit und Ausblick2376.1Fazit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2376.2Ausblick . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2406.2.1Studien zu Modellproteinen . . . . . . . . . . . . . . . . . . . . 2406.2.1.1Vergrößerung des Datensatzes. . . . . . . . . . . . . 2406.2.1.2Weitere QSPR-Modelle . . . . . . . . . . . . . . . . . 2406.2.1.3Sampling der ungefalteten Varianten . . . . . . . . . . 2416.2.1.4Analyse anderer Eigenschaften . . . . . . . . . . . . . 2416.2.1.5Experimentelle Ansätze . . . . . . . . . . . . . . . . . 2426.2.1.6Alternative Simulationsansätze . . . . . . . . . . . . . 2426.2.1.6.1Metadynamik-Simulationen . . . . . . . . . . 2426.2.1.6.2Erhöhte Temperaturen. . . . . . . . . . . . 2426.2.1.6.3Gerichtete MD-Simulationen . . . . . . . . . 2436.2.1.6.4Entfaltung über Gō-Modelle. . . . . . . . . 2436.2.1.7MD-Alternativen . . . . . . . . . . . . . . . . . . . . . 2436.2.2Größere Konjugate . . . . . . . . . . . . . . . . . . . . . . . . . 2436.2.2.1Verbesserung des Samplings. . . . . . . . . . . . . . 2436.2.2.1.1Abbildung des Solvens . . . . . . . . . . . . . 2436.2.2.1.2GaMD-Alternativen . . . . . . . . . . . . . . 2446.2.2.1.3Vergröberte Kraftfelder . . . . . . . . . . . . 2446.2.2.2MD-Alternativen . . . . . . . . . . . . . . . . . . . . . 244Zusammenfassung245Summary247Anhang249A Ausgleichungsrechnungen251A.1 Zugrundeliegende Gleichungen. . . . . . . . . . . . . . . . . . . . . . 251A.2 Fehlerfortpflanzung . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253B Deskriptoren255B.1Mittlere Abweichung (RMSD) . . . . . . . . . . . . . . . . . . . . . . . 255B.2Mittlere Fluktuation (RMSF) . . . . . . . . . . . . . . . . . . . . . . . 255B.3Solvenszugängliche Oberfläche (SASA) . . . . . . . . . . . . . . . . . . 256B.4Radiale Verteilungsfunktion (RDF) . . . . . . . . . . . . . . . . . . . . 256B.5Gyrationsradius . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256B.6HB/Lebensdauer-Analyse . . . . . . . . . . . . . . . . . . . . . . . . . 257B.7Interaktionsenergie . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257B.8Rückgrat-Diederwinkel . . . . . . . . . . . . . . . . . . . . . . . . . . . 257B.9Ordnungsparameter. . . . . . . . . . . . . . . . . . . . . . . . . . . . 258B.10 Native Kontakte. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259B.11 Diffusionskoeffizient. . . . . . . . . . . . . . . . . . . . . . . . . . . . 259B.12 HB-Autokorrelation. . . . . . . . . . . . . . . . . . . . . . . . . . . . 259B.13 CNA-basierte Deskriptoren. . . . . . . . . . . . . . . . . . . . . . . . 260C Konjugate263D Vorhersagemodelle273D.1 Berechnete Deskriptoren . . . . . . . . . . . . . . . . . . . . . . . . . . 273D.2 Ansatz 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 278D.2.1Modelltypen. . . . . . . . . . . . . . . . . . . . . . . . . . . . 278D.2.2Modellleistungen . . . . . . . . . . . . . . . . . . . . . . . . . . 280D.2.3ADAN-Analysen . . . . . . . . . . . . . . . . . . . . . . . . . . 281D.2.4Pin 1-unspezifische Deskriptoren . . . . . . . . . . . . . . . . . 281D.3 Ansatz 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283D.3.1Pin 1-unspezifische Deskriptoren . . . . . . . . . . . . . . . . . 283D.3.2Modelltypen. . . . . . . . . . . . . . . . . . . . . . . . . . . . 284D.3.3Modellleistungen . . . . . . . . . . . . . . . . . . . . . . . . . . 284D.3.4Analysen. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286Literatur291Abbildungsverzeichnis323Tabellenverzeichnis329Abkürzungsverzeichnis3311Kapitel 1Einleitung1.1Polymer-Biokonjugate1.1.1EinführungDurch die Optimierung und Weiterentwicklung neuer biotechnologischer Verfahrenseit der Markteinführung rekombinant hergestellten Humaninsulins im Jahre 1982 [1]haben makromolekulare Arzneistoffe in der Pharmazie stetig an Bedeutung hinzu-gewonnen. Biopharmazeutika bzw. Biologika und die davon abgeleiteten Biosimilarsstellen Peptide, Proteine oder nukleinsäurebasierte Strukturen wie Aptamere dar undsind vielseitig einsetzbar. Besonders monoklonale Antikörper dienen oftmals zur The-rapie von Autoimmunerkrankungen wie Morbus Crohn oder rheumatoider Arthtritis,können aber auch gegen Krebs eingesetzt werden [2]. Zur Klasse der Biologika zäh-len ebenso Impfstoffe [3]. Acht der zehn weltweit umsatzstärksten Medikamente desJahres 2019 stellten Biopharmazeutika dar [4] und es ist zu erwarten, dass dieserMarkt noch weiter wachsen wird. Aufgrund ihrer komplexen Struktur binden diesemakromolekularen Substanzen selektiv an ihr Angriffsziel [3, 5].Die Nutzung biotechnologisch hergestellter Makromoleküle, die parenteral appliziertwerden müssen [6], ist jedoch keineswegs unproblematisch. Durch enzymatische De-gradation und renale Filtration wird ihre Wirkung in vivo deutlich beeinträchtigt.Auch hinsichtlich der Wirkstoffformulierung ergeben sich Fragen zur Stabilität, dadie Funktion dieser Arzneistoffe wesentlich von ihrer dreidimensionalen Struktur ab-hängt, welche in der Regel sensibel auf Änderungen des pH-Werts, der Temperaturund andere äußere Einflüsse reagiert [3, 5]. Der aufwendige und kostspielige zell-basierte Herstellungsprozess definiert maßgeblich das Produkt, weshalb Biosimilarsauch nicht mit Generika gleichzusetzen sind [2, 3, 7]. Ergänzend zu pharmakokineti-schen und pharmazeutisch-technologischen Problematiken besitzen diese Wirkstoffeein gewisses immunogenes Potenzial, wodurch es zu unerwünschten Abwehrreaktionenkommen kann [3, 8].2Kapitel 1. EinleitungDurch kovalente Verknüpfung biologischer Makromoleküle mit Polymeren könnenneuartige Hybride erhalten werden, die im Idealfall die Vorteile beider Material-welten miteinander vereinen. Dieser Prozess wird Biokonjugation genannt und kanndie physikochemischen und pharmakokinetischen Eigenschaften von Arzneistoffen we-sentlich verbessern [9]. Während besonders die Medizin von solchen Konjugaten pro-fitiert hat, sollte nicht unerwähnt bleiben, dass diese Verbindungen bspw. auch in derLebensmittel- und Kosmetikindustrie zunehmend Anwendung finden [10]. Verbesser-te Biopharmazeutika werden indes auch als Biobetters bezeichnet [11]. Unter smartenBiokonjugaten versteht man gar Systeme, bei denen die Konformation des Polymersdurch äußere Stimuli wie Licht, Temperatur oder pH-Wertänderungen kontrolliertwerden kann [12].Die ersten Arbeiten zu Polymer-Biokonjugaten (PBK) wurden bereits in den Jahren1952 und 1953 publiziert, bei denen Chymotrypsin und Albumin mit Polyglycinenverknüpft wurden [13–15]. Das eigentliche medizinische Potential wurde jedoch erstdurch die 1977 von Abuchowski et al. hergestellten Konjugate deutlich [16, 17]. ZurBiokonjugation ist in der Regel eine spezielle Linkerstruktur nötig, über die das Po-lymer mit dem Protein verknüpft werden kann. In den Arbeiten von Abuchowskiet al. wurde Polyethylenglykol (PEG) unterschiedlichen Molekulargewichts, hier undin den meisten anderen Fällen in Form von Monomethoxy-PEG (mPEG), mittelstoxischem Cyanurchlorid als Linker kovalent an Aminogruppen des bovinen Serum-albumins und der bovinen Leberkatalase geknüpft. Erst 1990 wurde dann das ersteKonjugat von der amerikanischen Behörde für Lebens- und Arzneimittel (Food andDrug Administration, FDA) zugelassen. Es handelt sich um Adagen®, eine mit 5 kDaPEG konjugierte Adenosin-Desaminase zur Behandlung des schweren kombiniertenImmundefekts (SCID) [18]. Aktuell werden 18 von der FDA zugelassene PBK in denUSA vermarktet (Tab. 1.1, Stand: 2020), alle auf Basis von PEG [9, 19, 20]. Danebenkönnen aber auch kleinere Moleküle mit Polymeren verknüpft werden. Ein Beispielhierfür ist Movantik®, ein als Naloxegol bezeichnetes, PEGyliertes Naloxon. Die Kon-jugation soll hier den Übertritt über die Blut-Hirn-Schranke verhindern [21]. Ebenfallswurden bereits mit PEG beschichtete, niedermolekulare Wirkstoffe enthaltende Li-posomen von der FDA zugelassen: Onivyde® mit Irinotecan und Doxil®/Caelyx®mit Doxorubicin als aktive Komponenten [20, 22]. PEG kann hierbei wesentlich dieProtein-Corona des Partikels beeinflussen, d. h. das Ausmaß der Adsorption von Pro-teinen in vivo [23, 24]. 1993 wurde in Japan mit SMANCS/Lipiodol auch ein PBKgegen Leberzellkarzinome zugelassen, welches nicht auf PEG basiert. Es handelt sichum ein an jeweils zwei Aminosäuren mit einem Styrol-Maleimid-Copolymer konjugier-tes Neokarzinostatin, welches in einer lipophilen Formulierung intraarteriell appliziertwird, um sich im Tumorgewebe anzureichern [25, 26]. Eine weiterführende Liste zu ak-tuell in der Entwicklung befindlichen PBK für unterschiedlichste Anwendungsgebieteist in [27] zu finden.1.1. Polymer-Biokonjugate3Tabelle 1.1: Aktuell in den USA vermarktete und von der FDA zugelassene Biokonjugate(Stand: 2020). Handelsnamen, unPEGylierte Wirkstoffe, Indikationsbeispiele und das Jahrder Zulassung, nach dem die Substanzen geordnet sind, sind aufgeführt. In der dritten Spalteist die Art der PEGylierung näher beschrieben: einfach (E) oder mehrfach (M), sowie linear(L) oder verzweigt (V), inklusive Angabe des Molekulargewichts der PEG-Ketten in kDa [9,19, 20].HandelsnameWirkstoffPEGIndikationZulassungJivi®Faktor VIIIE/V/60Hämophilie A2018Asparlas™L-AsparaginaseM/L/5Akute lymphoblastischeLeukämie2018Revcovi™Adenosin-DesaminaseM/L/5.6SCID2018Palynziq®Phenylalanin-Ammoniak-LyaseM/V/20Phenylketonurie2018Rebinyn®Faktor IXE/V/40Hämophilie B2017Adynovate®Faktor VIIIM/V/20Hämophilie A2015Plegridy®Interferon-β1aE/L/20Multiple Sklerose2014Sylatron™Interferon-α2bE/L/12Melanome2011Krystexxa®UricaseM/L/10Urikopathie2010Cimzia®CertolizumabE/V/40Morbus Crohn2008Mircera®Epoetin betaE/L/30Renale Anämie2007Macugen®Aptamer gegenvaskulären endothelialenWachstumsfaktor (VEGF)E/V/40Makuladegeneration2004Somavert®Somatropin-AntagonistM/L/5Akromegalie2003Neulasta®Granulozyten-Kolonie-stimulierenderFaktor (G-CSF)E/L/20Neutropenie2002Pegasys®Interferon-α2aE/V/40Hepatitis B und C2001Pegintron®Interferon-α2bE/L/20Hepatitis C2000Oncaspar®L-AsparaginaseM/L/5Akute lymphoblastischeLeukämie1994Adagen®Adenosin-DesaminaseM/L/5SCID1990Im Prinzip stellen PBK mit ihrer direkten Modifikation des betreffenden Arznei-stoffs eine mögliche Alternative zur Nutzung von Nanovesikeln dar, deren Bedeutungin den letzten Jahren ebenso stieg. Beide Ansätze verfolgen im Grunde ein ähnli-ches Ziel: das Erreichen des Zielorts schlecht bioverfügbarer oder besonders instabilerSubstanzen zu erleichtern. Somit kann der für die Pharmazie relevante chemischeRaum, welcher lange Zeit lediglich durch einige wenige zentrale Dogmen geprägt war(Induced-Fit-Theorie nach Koshland [28], Lipinski’s Rule of Five [29, 30]), mit Fo-kus auf eine verbesserte Pharmakokinetik erweitert werden. Vor allem aber könnenbereits bestehende Therapien bzgl. dieser Problematiken optimiert werden. Während4Kapitel 1. EinleitungNanovesikel durch den EPR (Enhanced permeability and retention)-Effekt die Appli-kation besonders zytotoxischer Substanzen im Zuge einer Chemotherapie erlauben,wird durch eine Biokonjugation der Abbau sowie die Elimination der Moleküle in vivodeutlich verlangsamt [23, 31].In den vergangenen Jahren haben sich unterschiedliche Polymerisations- und Kon-jugationsmethoden zur Synthese von PBK etabliert. Während frühere Konjugateunspezifisch an verschiedenen Stellen PEGyliert waren und große Polydispersitätenaufwiesen, existieren heute Verfahren zur ortsspezifischen Konjugation mit maßge-schneiderten Polymeren definierter Länge. Diese Techniken sorgen chargenübergrei-fend für gleichmäßigere Eigenschaften der Konjugate, was besonders hinsichtlich ei-ner zu vermeidenden Variabilität der Bioverfügbarkeit von Vorteil sein kann [9, 18].Gleichzeitig stärken diese neuen Methoden zunehmend auch den Wunsch nach sys-tematischeren Vorgehensweisen zur Synthese von Konjugaten mit bestmöglichen Ei-genschaften. Letztere sind grundsätzlich abhängig vom Konjugationsort, der Anzahlan verknüpften Polymerketten, der Linkerstruktur und dem Aufbau sowie der Län-ge der Polymere. Rationale Ansätze zur Herstellung von PBK profitieren vor allemvon computertechnischen Methoden, mit denen das Verhalten auf molekularer Ebenenäher analysiert werden kann [32]. Die bisherigen Befunde dieses Forschungsgebietswerden in Kap. 1.1.6.1 näher erläutert.PBK weisen in aller Regel eine bessere Löslichkeit, ein vermindertes Aggregations-verhalten und eine erhöhte thermische und proteolytische Stabilität auf und werdendurch ihren größeren hydrodynamischen Radius schlechter renal filtriert, sodass sielänger im Blutkreislauf verbleiben. Ebenso wird das immunogene Potenzial des Wirk-stoffs gesenkt [33]. Es wird vermutet, dass PEG wie ein Schutzschild fungiert und dasProtein vor proteolytischen Enzymen und Antikörpern, die an Epitope des Arznei-stoffs binden könnten, schützt. Dahingehend wird in der Literatur auch oft von einemsogenannten Stealth-Effekt geschrieben, der im Hinblick auf Nanovesikel auch einermöglichen Bindung von Albumin durch PEG zugeschrieben wird [23]. Zusätzlich solldas Polymer wie eine Art Sieb wirken können: Kleinere Moleküle diffundieren bis zurProteinoberfläche, während größere Moleküle, wie bspw. Antikörper, bereits an derPolymeroberfläche abgeblockt werden [9]. Dieser Effekt wurde an α-Chymotrypsin,das mit Dendrimeren verschiedener Generationen konjugiert worden war, studiert [9,34].Neben den genannten Vorteilen zeigen PBK jedoch zumeist eine verminderte Aktivi-tät. Vermutlich beruht dies auf einer sterischen Abschirmung der Binderegionen desMakromoleküls durch das Polymer [35]. Es wurde jedoch auch bereits eine durch dieKonjugation modulierte Proteindynamik für die Aktivitätsminderung verantwortlichgemacht, weshalb die Nähe der Konjugationsstelle zum aktiven Zentrum nicht zwangs-läufig der entscheidende Faktor für die Aktivitätsmodulation ist [36]. Ebenfalls berich-ten einige Forscher von anti-PEG-Antikörpern. Dies kann zur beschleunigten Filtrati-on der Wirkstoffe aus dem Blutkreislauf und zu Hypersensibilitätsreaktionen führen.1.1. Polymer-Biokonjugate5Omontys®, ein PEGyliertes Erythropoeitin-Mimetikum, wurde aufgrund schwererImmunreaktionen vom Markt zurückgezogen [20, 37]. Dennoch werden Biokonjuga-te, vornehmlich aufgrund der durch die verringerte Clearance deutlich verlängertenHalbwertszeit im Blutkreislauf, zunehmend eingesetzt. Hierdurch kann die Frequenzder parenteralen Applikationen verringert werden, wodurch die Therapie für den Pa-tienten deutlich erleichtert wird [18]. Die Vor- und Nachteile einer Konjugation sindin Tab. 1.2 zusammengefasst.Tabelle 1.2: Mögliche Vor- und Nachteile einer Polymer-Biokonjugation an therapeutischverwendete Biologika [18, 20, 33, 37].VorteileNachteileErhöhte LöslichkeitVerringerte BioaktivitätVerminderte ImmunogenitätImmunogenes PotenzialVerminderte AggregationHohe ProduktionskostenVerminderte ClearanceVariable BioverfügbarkeitErhöhte proteolytische StabilitätPotenzielle ToxizitätThermische StabilisierungThermische Instabilisierung1.1.2Herstellungsverfahren1.1.2.1Grafting-MethodenGrundsätzlich wird zwischen drei verschiedenen Ansätzen der Biokonjugation un-terschieden (Abb. 1.1): Die Grafting to-Methode ist die am häufigsten angewandte.Hierbei wird das Polymer separat synthetisiert und anschließend über eine geeigneteLinkerstruktur an das Protein geknüpft. Dadurch sind Polymersynthesen auch in apo-laren Lösungsmitteln möglich und die Makromoleküle können vor der Konjugationcharakterisiert werden, jedoch ist die Ausbeute der Konjugation in der Regel gering,da zwei Makromoleküle miteinander reagieren müssen. Bei der Grafting from-Methodewird ein geeigneter Initiator in das zu modifizierende Protein integriert, sodass die Po-lymerisation direkt am Makromolekül stattfinden kann. Sie muss daher in einem fürdas Protein geeigneten Milieu stattfinden, die Ausbeute ist jedoch größer als bei derGrafting to-Methode und die Produkte können leicht von den niedermolekularen Mo-nomeren abgetrennt werden. Die letzte Variante stellt die Grafting through-Methodedar und wird aufgrund des geringen Polymerisationsgrades nur selten genutzt. Hier-bei werden die Proteine vor der Polymerisation mit Monomeren verknüpft. Es bildensich dadurch Polymerketten mit mehreren kovalent gebundenen Proteinen [9, 12, 32,38].6Kapitel 1. EinleitungAbbildung 1.1: Schematische Darstellung der (A) Grafting to-, (B) Grafting from- und(C) Grafting through-Methoden [12].1.1.2.2PolymersyntheseZur Synthese der Polymere werden vor allem bei der Grafting from-Methode oftkontrolliert radikalische Polymerisationen (CRP) durchgeführt, welche die Bildungäußerst komplexer makromolekularer Architekturen erlauben. Bei der Atomtrans-ferradikalpolymerisation (ATRP) wird ein organisch gebundenes Halogenid an einenCuI-Liganden, z. B. Tris(2-pyridylmethyl)amin (TPMA), mittels Redoxreaktion über-tragen. Dadurch entsteht ein zur Polymerisation geeignetes Radikal, welches durchBindung des Halogenids des CuII-Liganden wieder deaktiviert wird. Es herrscht eingut kontrollierbares Gleichgewicht zwischen der aktiven und der inaktiven Polymerisa-tion. Als eine Alternative zur ATRP kann die reversible Additions-Fragmentierungs-Kettenübertragungs-Polymerisation (RAFT) angesehen werden. Hierbei stellt sichein Gleichgewicht unter Verwendung eines Kettentransferreagenzes (CTA) ein, wo-für in der Regel eine Thiocarbonylthioverbindung genutzt wird. CTAs oder ATRP-Initiatoren können z. B. mittels N-Hydroxysuccinimid (NHS)-aktivierter Carbonsäu-ren an das Makromolekül geknüpft werden [9, 12, 32, 38].1.1. Polymer-Biokonjugate71.1.2.3KonjugationsstrategienMithilfe verschiedener Linkerstrukturen und variabler Reaktionsbedingungen könnendie Polymerketten bei der Grafting to-Methode an unterschiedliche Aminosäuren ei-nes Proteins konjugiert werden. Eine Auswahl ist als Übersicht in Tab. 1.3 dargestellt.Besonders Aminogruppen von Lysinen dienen z. B. über Bildungen von SchiffschenBasen mit Aldehyden oder NHS-vermittelte Amidierungen als Konjugationsstellen.Die Aminogruppe des N-Terminus kann spezifisch bei pH-Werten adressiert werden,bei denen Lysine noch protoniert vorliegen. Cysteine stellen gute Nukleophile und,aufgrund ihres relativ seltenen Vorkommens in freier Form, zumeist auch spezifischeKonjugationsorte in Proteinen dar, die über die Bildung eines Thioethers bspw. mitMaleimiden oder Vinylsulfonen reagieren können. Alternativ können auch Disulfid-brücken spezifisch angegriffen werden und zu überbrückten PEGylierungen führen.Möglich sind ebenfalls Konjugationen über Histidine, die in manchen Fällen auch zueiner Art Prodrug führen können, da instabile Acyl-Histidin-Bindungen in vivo ineiner allmählichen Freisetzung des unkonjugierten Wirkstoffs resultieren (wie bspw.bei Pegintron®). Zu erwähnen sind schließlich auch komplexere Linkerstrukturen, diees ermöglichen, verzweigte PEG-Ketten mit größerem Stealth-Effekt zu konjugieren[12, 18, 19, 39].Auch wenn die 20 natürlich vorkommenden Aminosäuren bereits eine breite Palettean möglichen Konjugationen erlauben, kann eine ortsspezifische Verknüpfung leich-ter über nicht-natürliche Aminosäuren erreicht werden. Diese können mittels gene-tischer Modifizierung über ein Amber-Codon während der Translation an der ge-wünschten Position in die Aminosäuresequenz eingebaut werden [40]. Alternativ exis-tieren für kleinere Peptide auch Einbaumöglichkeiten über Festphasenpeptidsynthe-sen. Nicht-natürliche Aminosäuren ermöglichen den Einsatz eleganter und selektiverClick-Chemie-Verfahren [12, 41]. Ein Beispiel stellt Nε-[(2-Azidoethoxy)carbonyl]-L-lysin (AZK, auch NAEK genannt) dar, welches über seine Azidfunktion mittels 1,3-dipolarer Huisgen-Cycloaddition an Alkingruppen geknüpft werden kann. In umge-kehrter Weise kann auch Nε-[(Allyloxy)carbonyl]-L-lysin verwendet werden [40, 42,43]. Alternativ zur zumeist kupferkatalysiert durchgeführten Addition kommen Cy-clooctinderivate ohne toxikologisch bedenkliche Metallkatalysatoren aus, indem bio-orthogonale, ringspannungsgeförderte Azid-Alkin-Cycloadditionen (Strain-promotedazide-alkyne click chemistry reaction, SPAAC) durchgeführt werden [41]. Beispie-le für solche Derivate sind der Dibenzocyclooctin (DBCO)- und der Bicyclononin(BCN)-Linker (Abb. 1.2) [12]. Erst kürzlich wurden als Alternative zu bereits etablier-ten ringspannungsgeförderten Cycloadditionen auch Konjugationen mittels trans-Cyclohepten über inverse Elektronenbedarfs-Diels-Alder-Reaktionen beschrieben [44].8Kapitel 1. EinleitungTabelle 1.3: Beispiele möglicher Kupplungsreagenzien und daraus entstehende Produkte fürverschiedene Aminosäuren. R stellt dabei den Polymerrest dar [12, 19].AminosäureKupplungsreagenzKupplungsproduktReferenzen[45–48][49, 50][51, 52][53][54][55][56][57, 58][19]1.1. Polymer-Biokonjugate9Abbildung 1.2: Kupplungsreaktionen für PEGylierte DBCO- und BCN-Linker mit AZK.Neben kovalenten Verknüpfungen von Polymeren an Makromoleküle kann eine orts-spezifische Bindung auch nichtkovalent über Biotin-Streptavidin- oder HisTag-Ni2+-Nitrilotriessigsäure-Wechselwirkungen erfolgen [12]. Schließlich können auch enzyma-tische Kupplungsreaktionen verwendet werden, die aufgrund ihrer Sequenzabhängig-keit zur ortsspezifischen Konjugation geeignet sind. Hierfür werden bspw. Transglut-aminasen oder auch Tyrosinasen genutzt [18, 41, 59]. Berücksichtigt man auch kova-lente Konjugationen peptidischer Elemente an Proteine, so sollte schließlich auch dieSpyTag/SpyCatcher-Methode erwähnt werden, bei der eine Isopeptidbindung zwi-schen einer Asparaginsäure des kleinen SpyTag-Peptids und einer Lysinseitenkettedes SpyCatcher-Proteins gebildet wird. Diese Technologie wird z. B. zur Herstellungvon Vakzinen und Protein-Hydrogelen genutzt [60].1.1.3PolymerartenDie Welt der Polymere bietet schier unendliche Kombinationen möglicher biokom-patibler Strukturen, die zur Konjugation verwendet werden können. Im Folgendensollen nur einige kurz erwähnt werden.1.1.3.1PolyethylenglykolPEG ist das für Biokonjugationen am häufigsten genutzte Polymer. Es handelt sichum ein, je nach Molekulargewicht, in der Regel gut wasserlösliches Makromolekül mitsimpler, linearer Struktur, welches in der Pharmazie schon seit langer Zeit eine breiteAnwendung findet. Es kann unter anderem als Cosolvens, Bindemittel, Gelbildner10Kapitel 1. Einleitungoder auch Wirkstoff gegen Obstipation verwendet werden [61]. Aufgrund der bereitslangen Nutzung wird es von der FDA allgemein als biokompatibel, nicht toxischund Goldstandard zur Biokonjugation angesehen. Die Etherstrukturen resultieren ineiner starken Hydratation und damit in einem größeren hydrodynamischen Radius,wodurch die in Kap. 1.1.1 schon erwähnten Vorteile entsprechender PBK realisiertwerden. Der amphiphile Charakter des Polymers ermöglicht es ebenfalls, Konjugatefür apolare Lösungsmittel zur Durchführung organischer Synthesen herzustellen [26,62, 63].Durch die intensive Anwendung dieses Polymers sind jedoch auch die Nachteile diesesStrukturtyps hinlänglich bekannt. In Patienten nachgewiesene anti-PEG-Antikörpervom Typ IgG und IgM können bei wiederholten Gaben zur schnelleren Clearancedes Wirkstoffs (Accelerated blood clearance, ABC-Phänomen) und schweren Immun-reaktionen führen [20, 26, 37]. Eine kürzlich publizierte Kristallstruktur illustriertdie Bindung des flexiblen Polymers an einen Antikörper [64]. Anaphylaktische Re-aktionen aufgrund einer PEG-Allergie wurden auch bei Arzneimitteln beobachtet,in denen das Polymer als Hilfsstoff, z. B. zur Stabilisation von Lipidnanopartikeln,eingesetzt wird, so auch im Falle des gegen das Coronavirus SARS-CoV-2 verwende-ten mRNA-Vakzins Tozinameran. Auch handelsübliche Pflegeprodukte wie Shampoosenthalten oft PEG [65]. Eine Prophylaxe durch Gabe von freien PEG-Ketten könnteden Titer an Antikörpern jedoch für eine gewisse Zeit verringern [66]. Ebenso sinddie zur Biokonjugation genutzten PEG-Ketten nicht biologisch abbaubar, wodurchsich Polymerrückstände in Leber-, Nieren- und Milzzellen ansammeln und Vakuolenbilden können [26]. Aufgrund dieser Nachteile wurden bereits verschiedene biokom-patible PEG-Alternativen für die Konjugation in Betracht gezogen. Sie lassen sich inbiologisch nicht abbaubare und natürlich abbaubare Polymere einteilen. Ausgewähl-te Strukturen sind in Abb. 1.3 gezeigt. Bisher existieren zu diesen jedoch deutlichweniger in vivo-Studien.1.1.3.2PolyvinylpyrrolidonPolyvinylpyrrolidon (PVP) stellt eine hydrophile, nicht toxische Alternative zu PEGdar. Es kann durch eine RAFT-Polymerisation synthetisiert werden. Ein Vergleichder Biodistributionen konjugierter Tumornekrosefaktoren mit unterschiedlichen Po-lymerarten ergab für das PVP-Konjugat eine im Vergleich zur PEG-Variante erhöhteHalbwertszeit im Blutkreislauf [63, 67]. Allerdings zeigte PVP bei der Untersuchungkonjugierter Uricasen auch eine im Vergleich zu PEG erhöhte Immunogenität [68, 69].1.1.3.3PolyglycerolePolyglycerol (PG) ist ein hydrophiles PEG-Derivat mit zusätzlichen Hydroxymethyl-seitenketten, das in Studien eine sehr gute Zell- und Hämokompatibilität und eineerhöhte Halbwertszeit im Vergleich zu PEG zeigte [63, 70, 71]. Bisher wurde auch invivo keine nennenswerte Toxizität beobachtet [72]. Ein weiterer Vorteil zu PEG liegt1.1. Polymer-Biokonjugate11in der geringeren Viskosität [73]. PG kann mittels Glycidol über eine Ringöffnungspo-lymerisation erhalten werden. Durch die freie Hydroxymethylseitenkette des linearenPolyglycerols (LPG) können im Gegensatz zu PEG auch weitere Derivatisierungenerfolgen [70]. Hyperverzweigte Polyglycerole stellen kompakte, äußerst komplexe Ar-chitekturen dar und zeigen eine längere Halbwertszeit im Blutkreislauf als LPG [71].Allerdings sind auch PG-Ketten nicht bioabbaubar und können sich in der Leber undder Milz anreichern [63, 72].1.1.3.4PolyoxazolineEbenfalls nicht bioabbaubar sind Poly(2-alkyl-2-oxazoline) (POx), welche durch ei-ne kationische Ringöffnungspolymerisation (CROP) erhalten werden können [74]. Siesind schon seit über 50 Jahren bekannt, ihre Bedeutung im Bereich der Biowissen-schaften nahm jedoch erst im letzten Jahrzehnt deutlich zu [75]. Wie bei PG ergibtsich hier durch die Derivatisierung der Oxazolinseitenkette eine große strukturelleVielfalt. Es können bspw. Poly(2-methyl-2-oxazolin) (PMeOx) und Poly(2-ethyl-2-oxazolin) (PEtOx) synthetisiert werden und mittels dieser Variation der Seitenket-tenlänge die Polarität des Polymers beeinflusst werden [76]. POx wird ein ähnli-cher Stealth-Effekt wie PEG zugerechnet, und stark verlängerte Halbwertszeiten imBlutkreislauf wurden bereits für POxylierte G-CSF-Konjugate beschrieben [26, 63].Auch wenn das Polymer nicht bioabbaubar ist, so wurde bereits von einer komplettenAusscheidung eines radionuklidmarkierten Polymers berichtet [77], und es wurden inHasen eine geringere Immunogenität als PEG und keine histopathologischen Verände-rungen von Geweben festgestellt [78]. PEtOx-Ketten können eine geringere Viskositätals PEG aufweisen und sind in verschiedensten Medien löslich. Aufgrund der Vortei-le wurden bereits eine Reihe von Biokonjugaten auf Basis von POx hergestellt [76].PMeOx gilt als etwas hydrophiler, PEtOx hingegen als etwas lipophiler als PEG.Entsprechende Strukturen wurden vor einigen Jahren erstmals auf atomarer Ebenemittels Molekulardynamik-Simulationen untersucht [79].1.1.3.5Weitere PolymereEin Beispiel für ein synthetisches, bioabbaubares PEG-Derivat stellt das Pfropfcopo-lymer Poly(oligo(ethylenglykol)methacrylat) (POEGMA) dar, dessen PEG-Ketten invivo aufgrund der Esterbindung an die Hauptkette abgebaut werden können. Kon-jugate dieses Polymers zeigen eine stark verlängerte Halbwertszeit im Blutkreislaufund eine geringfügig verminderte Aktivität [26, 63]. Auch die Immunogenität derPEG-Strukturen kann bei diesem Strukturtyp vermindert sein [80].Schließlich ist auch zu erwähnen, dass, so wie auch Glykosylierungen von Proteinen inZellen stattfinden, viele zuckerbasierte Konjugationsmöglichkeiten existieren. Bspw.wurden bereits Konjugate von Erythropoeitin, G-CSF und Interferon-α mit Hydro-xyethylstärke (HES) hergestellt. Diese zeigten verbesserte pharmakokinetische Ei-genschaften. HES kann als Plasmaexpander genutzt und durch α-Amylase abgebaut12Kapitel 1. Einleitungwerden. Allerdings mehren sich Berichte zu toxischen Akkumulationen von HES inLeber und Nieren [81]. Eine Alternative zu dieser Struktur stellt bspw. Dextrin dar,welches bereits mit endothelialen Wachstumsfaktoren konjugiert wurde. Aber auchPolysialinsäuren sind als mögliche Alternativen zu nennen [26].Abbildung 1.3: Strukturformeln ausgewählter Polymere (Bezeichnungen gemäß Text).1.1.4Therapeutische BeispieleIm Folgenden sollen einige der in Tab. 1.1 aufgeführten PBK kurz beschrieben werden.1.1.4.1Adagen®Wie in Kap. 1.1.1 erwähnt stellt Adagen® das erste vermarktete PBK dar. Mutatio-nen des für den Purinmetabolismus wichtigen Proteins Adenosin-Desaminase könnenzu einem SCID führen. Eine mit 5 kDa PEG konjugierte bovine Variante des Prote-ins dient zur Therapie. In Adagen® sind zwischen 11 und 17 Polymere, die zuvor miteiner Succinimidylsuccinatgruppe für die Konjugation aktiviert wurden, unspezifisch1.1. Polymer-Biokonjugate13an Lysinseitenketten geknüpft. Durch die deutlich erhöhte Halbwertszeit sind mehr-malige Bluttransfusionen nicht mehr notwendig und das Biokonjugat muss lediglicheinmal wöchentlich appliziert werden [18, 19].1.1.4.2Neulasta®Neulasta® stellt das erste vermarktete Konjugat dar, welches ortsspezifisch konjugiertist. Es handelt sich um PEGyliertes, rekombinant hergestelltes, humanes G-CSF, dasbei Neutropenie eingesetzt wird. Zur Synthese wird 20 kDa PEG mit einer terminalenAldehydfunktion selektiv an die N-terminale Aminogruppe von Methionin über einereduktive Aminierung unter sauren Bedingungen geknüpft. Auch hier ergeben sichdurch die PEGylierung verbesserte pharmakokinetische Eigenschaften und damit eineerleichterte Therapie: Das Arzneimittel muss im Gegensatz zur unPEGylierten Formnur einmalig subkutan appliziert werden [18, 19].1.1.4.3Pegasys®Pegasys® ist ein Beispiel für ein Konjugat mit verzweigter PEG-Kette. Es handeltsich um Interferon-α2a, das gegen Hepatitis B und C eingesetzt wird und an sei-nen solvenszugänglichen Lysinseitenketten (K31, K121, K131 und K134, selten auchan K70 und K83) mit der in Tab. 1.3 am Ende aufgeführten PEG2-NHS-Struktur,die aus einem an beiden Aminogruppen verknüpften Lysin besteht, konjugiert ist.Die Konjugate an den Positionen K31 und K134 zeigen die größte Aktivität, die fürPegasys® insgesamt dennoch nur 7 % beträgt (bezogen auf die unPEGylierte Vari-ante). Trotzdem ist der antivirale Effekt durch eine einmal wöchentliche Applikationaufgrund der längeren Halbwertszeit stärker ausgeprägt als bei dem unkonjugiertenProtein, welches dreimal wöchentlich appliziert werden muss. Die Clearance der un-PEGylierten Variante beträgt 6000 ml/h, jene von Pegasys® 60 bis 100 ml/h [18,19].1.1.4.4Macugen®Zuletzt soll noch das intravitreal applizierte Macugen® erwähnt werden. Es han-delt sich hierbei um ein Anti-VEGF-Aptamer aus 28 Nukleotiden, das gegen eineMakuladegeneration eingesetzt werden kann. Es stellt somit das bisher einzige nukle-insäurebasierte, in den USA zugelassene Biokonjugat dar. Aptamere können in vivoschnell von Nukleasen abgebaut werden. Auch hier wird die Halbwertszeit durch diePEGylierung mit einer verzweigten 40 kDa PEG-Struktur deutlich erhöht. Die beidenlinearen PEG-Ketten sind, wie im Falle von Pegasys®, zunächst an die Aminogrup-pen eines Lysins geknüpft. Dieses bildet hier jedoch kein Amid mit der Seitenketteeines weiteren Lysins, sondern ist über eine kurze Alkylkette mit der Phosphatgruppedes 5’-Ende des Aptamers verknüpft [18, 19].14Kapitel 1. Einleitung1.1.5Thermodynamische CharakterisierungWie in Kap. 1.1.1 angemerkt ist die Funktion eines biologischen Makromoleküls maß-geblich von dessen dreidimensionaler Struktur abhängig. Insofern ist die thermody-namische Stabilität von besonderer Bedeutung für therapeutisch genutzte Biologikaund sollte einen wichtigen Aspekt in der Entwicklung neuer Biokonjugate darstellen.Zumeist wird eine simple Zwei-Phasen-Kinetik angenommen, bei der das ungefaltetebzw. denaturierte Protein beim Faltungsprozess einen Übergangszustand hin zur en-ergetisch stabileren, gefalteten bzw. nativen Konformation durchläuft (Abb. 1.4). Inder Realität kann das Energiediagramm durch viele verschiedene Übergangszuständeoder auch metastabile Konformationen eine deutlich größere Komplexität aufweisen.Die thermodynamische Stabilität kann durch die Differenz der Gibbs-Energie zwi-schen beiden Endzuständen beschrieben werden. Sie steht in unmittelbarem Zusam-menhang mit der Equilibriumskonstanten K, sowie den dazugehörigen kinetischen(Ent-)faltungskonstanten kf und ku (Glg. 1.1). Schon geringe Änderungen in der Pro-teinsequenz können die Stabilität des Wildtyps verändern und dadurch entscheidendebiologische und physikochemische Eigenschaften beeinflussen. So basiert das Vorkom-men verschiedener Krankheiten bspw. auf Punktmutationen in Proteinen. Gleichzeitigerlauben Modifikationen, die die thermodynamische Stabilität von Proteinen optimie-ren, das Vorkommen thermophiler Bakterien bei erhöhten Temperaturen. Zur Cha-rakterisierung dieser Stabilität existieren verschiedenste experimentelle Methoden,von denen zwei im Folgenden kurz vorgestellt werden sollen [82–85].∆G0 = −RT lnK = −RT ln(kfku)(1.1)1.1.5.1Temperaturabhängige CD-SpektroskopieEine temperaturabhängige Circulardichroismus (CD)-Spektropolarimetrie kann zumStudium des Effekts der Konjugation auf die thermodynamische Stabilität des Prote-ins herangezogen werden. Zur Untersuchung der Sekundärstrukturen von Proteinenwird hierbei zirkular polarisiertes Licht unterschiedlicher Wellenlängen, welches auseiner links- und einer rechtsdrehenden Komponente besteht, durch eine Lösung desMakromoleküls gestrahlt. Optisch aktive Strukturen absorbieren beide Komponen-ten in unterschiedlichem Maße. Dies kann durch chirale Moleküle, aber auch durchasymmetrische Sekundärstrukturen wie α-Helices und β-Faltblätter verursacht wer-den [86]. Die Differenz ∆A zwischen den Absorptionen der links- und rechtsdrehen-den Lichtanteile AL/R ist gemäß dem Lambert-Beer’schen Gesetz proportional zu denExtinktionskoeffizienten ϵL/R (Einheit:dm3mol cm), multipliziert mit der Konzentrationm in moldm3 und der Küvettenlänge l in cm:1.1. Polymer-Biokonjugate15Abbildung 1.4: Schematische Darstellung des Faltungsprozesses von Proteinen auf Basiseiner Zwei-Phasen-Kinetik.(AL − AR) = (ϵL − ϵR) m l(1.2)Die Elliptizität [θ]λ der Wellenlänge λ (gleiche Einheit wie l), angegeben in Radianten,kann wiederum unter Nutzung der entsprechenden Absorptionsindizes n′L/R aus demVerhältnis der Achsenlängen der aus der unterschiedlichen Absorption resultierendenEllipse des polarisierten Lichts berechnet werden. Angegeben in Grad mithilfe derAbsorptionen AL/R ergibt sich schließlich ein Umrechnungsfaktor von 32.98 [87].[θ]λ = tan−1( ba)(1.3)[θ]λ (Radianten) = π lλ (n′L − n′R)(1.4)= 2.303 l m4(ϵL − ϵR)(1.5)[θ]λ (Grad) = 32.98 m l (ϵL − ϵR)(1.6)= 32.98 (AL − AR)(1.7)= 32.98 ∆A(1.8)Dabei stellt b die kürzere und a die längere Achse der Ellipse dar. Die durchschnittli-che Elliptizität pro Aminosäure [θ]mrw,λ (Mean residue ellipticity, Angabe in deg cm2dmol res)und die molare Elliptizität [θ]molar,λ (Angabe in deg cm2dmol ) können durch folgende Um-rechnungen erhalten werden:16Kapitel 1. Einleitung[θ]mrw,λ = [θ]λMRW10 l C=[θ]λMN−110 l C(1.9)[θ]molar,λ = 100 [θ]λm l(1.10)Dabei sind M das Molekulargewicht in g/mol, N die Anzahl an Aminosäuren, N − 1die Anzahl an Peptidbindungen,MN−1 das davon abgeleitete mittlere Aminosäurege-wicht (Mean residue weight, MRW), l die Küvettenlänge in cm, C die Konzentrationin g/ml und m die molare Konzentration in mol/l.1 Typische Sekundärstrukturenergeben im Bereich von etwa 190 bis 240 nm charakteristische Banden in einem CD-Spektrum, welches zur schnellen Überprüfung der Proteinstabilität oder einer erstenInformation im Rahmen einer Strukturaufklärung dienen kann. Für die Absorptionunter 240 nm sind vor allem n → π∗ (220 nm)- und π → π∗ (190 nm)-Übergänge derPeptidbindungen verantwortlich [86, 88].Da die Elliptizität maßgeblich durch die Sekundärstrukturen bestimmt wird, kanndie Proteinentfaltung mit einer temperaturabhängigen CD-Spektroskopie untersuchtwerden. Hierzu werden [θ]mrw,λ oder [θ]molar,λ für eine festgelegte Wellenlänge (für α-helikale Strukturen normalerweise bei 222 nm, für β-faltblattstrukturreiche Proteinebei 218 nm) gegen die Temperatur aufgetragen. Die Messungen werden jeweils nacheiner Equilibrierungszeit von bspw. 2 Minuten für jeden Temperaturwert in einem fest-gelegten Intervall (z. B. von 1 bis 95 °C, mit 2 °C-Sprüngen) durchgeführt. Erhaltenwird eine Entfaltungskurve, mit deren Wendepunkt die Schmelz- bzw. Entfaltungs-temperatur Tm des Makromoleküls bestimmt werden kann. Über eine Van-’t-Hoff-Analyse können daraus wichtige thermodynamische Parameter abgeleitet werden, soz. B. die freie Faltungsenergie ∆G0f. Weitere Details zu den hierfür notwendigen Aus-gleichungsrechnungen sind in Anhang A aufgeführt.1.1.5.2Thermal Shift AssayEine Alternative zur temperaturabhängigen CD-Spektroskopie stellt ein fluoreszenz-basierter Thermal Shift Assay (TSA) bzw. die Differential Scanning Fluorimetry-Methode (DSF) dar. Hierbei wird die Temperatur einer Lösung des Proteins ebenfallssukzessive erhöht und währenddessen die allmählich ansteigende Fluoreszenzintensi-tät, die aus einem geeigneten proteinbindenden Farbstoff wie bspw. SYPRO™ Orangeresultiert, detektiert. Über den Wendepunkt, der über die erste Ableitung ermitteltwerden kann, wird die Schmelztemperatur erhalten. Alternativ kann hierzu die Ei-genfluoreszenz von Tryptophanen oder Tyrosinen genutzt werden [89–91].1In diesem Abschnitt wurden die Bezeichnungen aus [86] und [87] verwendet. Üblicherweise wirddie Stoffmengenkonzentration mit c gekennzeichnet, die Massenkonzentration hingegen mit β. Gemäßder angegebenen Literatur ist die Angabe der SI-Einheit mol in Form von Dezimol (dmol) und dieVerwendung von Klammern für die Größen [θ]λ, [θ]mrw,λ und [θ]molar,λ in der CD-Spektroskopieüblich.1.1. Polymer-Biokonjugate171.1.6Stabilisierungsmechanismen1.1.6.1Bisherige HypothesenEine Vielzahl an publizierten Arbeiten beschäftigt sich mit der Frage, wie es zur Er-höhung der thermischen bzw. konformativen Stabilität eines Proteins durch die Kon-jugation mit PEG kommen kann. 2008 bestätigten Rodríguez-Martínez et al. mithil-fe von Wasserstoff-Deuterium (H/D)-Austauschexperimenten unter Verwendung vonFourier-Transform-Infrarotspektroskopie an konjugiertem α-Chymotrypsin die bereitsbestehende Hypothese, dass eine PEGylierung mit einer Reduktion der Proteindyna-mik einhergehen kann. Ebenfalls wurde beobachtet, dass der Stabilisierungseffekt einPlateau ab vier Monomeren erreicht und bei längeren Polymerketten stagniert [92].Molekulardynamik (MD)-Simulationen aus dem Jahre 2011 von Yang et al. mit kon-jugiertem Insulin zeigten eine stabilere Sekundärstuktur und eine bei erhöhten Tem-peraturen niedrigere mittlere Abweichung (RMSD) von der nativen Proteinstruktur,sowie einen Rückgang an solvenszugänglicher Oberfläche (SASA) des Proteins. Indiesen Simulationen interagierte das Polymer hauptsächlich über hydrophobe Wech-selwirkungen mit dem Protein, während es eine Vielzahl an Wasserstoffbrückenbin-dungen (HB) zum Solvens ausbildete [93]. Eine Reduktion der SASA ist zu erwartenund wurde auch in anderen Publikationen beschrieben. Ein solcher Effekt könnte dieProteinrückgrat-HB durch Abschirmung des Solvens verstärken [94]. Weng et al. stu-dierten 2012 die Stabilität PEGylierter SH3-Domänen mithilfe von Guanidiumchloridals Denaturierungsreagenz und leiteten hiervon ein physikalisches Modell zur Erklä-rung der Stabilisierung ab. Dieses Modell beschäftigt sich mit dem Verhältnis zwischenproteinoberflächennahen Polymer- und Wasseratomen und basiert auf der Annahme,dass die PEGylierung die SASA der entfalteten und der nativen Proteinstruktur mo-duliert. Da das für das Solvens zugängliche Volumen in unmittelbarer Umgebung desProteins bei der Entfaltung steigt (Excluded volume-Effekt bei Faltungsvorgängen),das eingenommene Volumen der Polymerkette jedoch gleich bleibt, wird die SASAdes Proteins im gefalteten Zustand durch die PEG-Kette stärker beeinflusst als imdenaturierten. Die Autoren schlussfolgerten, nach Analyse des sogenannten m-Werts,der die Faltungskooperativität des Proteins beschreibt und mit der SASA-Modulationkorreliert, dass die Stabilisierung maßgeblich aus einer Senkung der Entfaltungsge-schwindigkeit (ausgedrückt durch die kinetische Konstante ku) resultiert. Auch dieseArbeit zeigte auf, dass der Stabilisierungseffekt vom Molekulargewicht der größerenPEG-Ketten unbeeinflusst blieb, jedoch Konjugationen an mehreren Stellen gleich-zeitig diesen erhöhten [95]. Eine weitere Studie zum Verhalten von Biokonjugatenmit PEG-Alternativen bei verschiedenen pH-Werten stellte vor kurzer Zeit die Hy-pothese auf, dass Polymere in ähnlicher Weise wie Chaperone in der Lage seien, dieRückfaltung partiell ungefalteter Proteine zu begünstigen [96].18Kapitel 1. Einleitung1.1.6.2Arbeiten von Joshua L. PriceWeitere Befunde auf diesem Forschungsgebiet sind vor allem den Arbeiten von Jo-shua L. Price (Bringham Young University, Utah) zu verdanken. Auf Basis zahlreichersystematischer Untersuchungen zu kleinen Oligomer-Konjugaten der WW-Domäneder 34 Aminosäuren langen, β-faltblattstrukturreichen, humanen Peptidyl-Prolyl-cis-trans-Isomerase Pin 1 wurden verschiedene Hypothesen aufgestellt, wobei sich auchhier zeigte, dass PEG-Tetramere die ideale Länge zur Stabilisierung besitzen [97].PEG führte bei verschiedenen Mutationen in Konjugationsnähe im Gegensatz zurN-Glykosylierung zur Stabilisierung [98]. Während zunächst Sekundärstrukturände-rungen und PEG-Wechselwirkungen mit spezifischen Aminosäuren für die Stabilisie-rung verantwortlich gemacht worden waren [99], zeigten weiterführende Simulationenund CD-basierte Stabilitätsmessungen, auch unter Nutzung von Deuteriumoxid alsSolvens, dass die Desolvatation von in der Nähe der Konjugationsstelle befindlichenHydroxylseitenketten (von Serin, Threonin oder Tyrosin) zu einem entscheidendenentropischen Gewinn führen könnte. Faltungskinetiken, die durch DSF-Messungenmithilfe der Fluoreszenz von Tryptophan ermittelt wurden, deuteten bei stabilisiertenVarianten, wie zuvor auch schon in [97], auf eine Erniedrigung der Entfaltungskon-stante ku ohne Beeinträchtigung der Faltungskonstante kf hin. Auch wurde die kon-formative Stabilität in dieser Arbeit erfolgreich mit der Proteolysestabilität und da-mit mit verbesserten pharmakokinetischen Eigenschaften korreliert. Aufgrund dieserBeobachtungen definierten Price et al. einen kristallstrukturbasierten Winkel θPrice.2Dieser wird zwischen zwei Vektoren a und b gemessen. Ersterer verläuft vom Cα-Atomder zu konjugierenden Aminosäure durch deren Seitenkettenmassenschwerpunkt. Derzweite wird von diesem Massenschwerpunkt aus zum nächstgelegenen Sauerstoffatomeiner Hydroxylseitenkette gezogen (Abb. 1.5). Kleinere Werte für θPrice verdeutli-chen die Orientierung der Konjugationsstelle zur Hydroxylgruppe und korreliertenmit einer Stabilisierung bzw. negativeren ∆∆G0f-Werten (Differenz freier Faltungs-energien zwischen PEGylierter und unPEGylierter Variante, siehe Kap. 1.1.5.1 undAnhang A). Getestet wurde dieses kristallstrukturbasierte Vorhersagemodell an einerPEGylierten SH3-Domäne, welche korrekt als stabilisiert klassifiziert wurde [100].Abbildung 1.5: Illustration des von Price et al. definierten Winkels θP rice am Beispiel derzu konjugierenden Aminosäure S19 der WW-Domäne von Pin 1 [100].2Zur Vermeidung von Verwechslungen mit der zuvor beschriebenen Elliptizität wurde hier dieBezeichnung Price ergänzt.1.1. Polymer-Biokonjugate19Zusätzliche Studien von Price et al. deuten daraufhin, dass rigide, planare Linker-strukturen mit einer ganz bestimmten Länge am besten geeignet sind, um den Desol-vatationseffekt zu maximieren [101]. Die Desolvatation kann jedoch nicht nur in derNähe von Hydroxylseitenketten zum Stabilitätsgewinn führen. Ebenso scheint PEGdurch eine Abschirmung des Solvens intramolekulare Salzbrücken zwischen verschie-den geladenen Aminosäuren verstärken zu können (getestet an der WW-Domäne vonPin 1 und den Homo-(Di-/Tri-)meren 1CW und GCN4, siehe auch Kap. 1.5) [102].Auch ist es in der Lage, vornehmlich exponierte, hydrophobe Areale zu bedecken [103]und NH-π-Wechselwirkungen zwischen Aminosäuren zu verstärken [104]. Arbeiten zueinem α-helikalen Trimer deuten zudem darauf hin, dass PEG-Konjugationen mittelseines Cys-Maleimid-Linkers an der Position i zu benachbarten Lysinen an den Stelleni+3 und i+4 am besten geeignet sind, um eine Stabilisierung α-helikaler Strukturenzu erreichen (Abb. 1.6). Während MD-Simulationen generell auf Interaktionen derPEG-Ketten mit Lysinen hindeuten [105–107], wird eine Stabilisierung nach Price etal. in einer solchen Position vermutlich durch Interaktionen des Maleimids mit denLysinen erreicht. Konjugierte PEG-Ketten führen dagegen zu einer Destabilisierungdieser Wechselwirkungen, wobei eine Stabilisierung in Relation zur reinen Cystein-Mutante dennoch erreicht wird [108].Abbildung 1.6: Potentielle Konjugationsstelle A46 in der Struktur des Her2-Affibodys. K49und K50 sind an den Positionen i+3 und i+4 in der Helix lokalisiert.1.1.6.3Zusammenfassende ÜbersichtInsgesamt scheint die PEG-basierte Erhöhung der konformativen Stabilität also ausmehreren Effekten zu resultieren. Da in der Vergangenheit von sich im Gleichge-wicht befindenden proteinnahen und solvensexponierten PEG-Konformationen be-richtet wurde, besteht vermutlich eine feine Balance zwischen enthalpisch günstigenWirkungen (spezifische Protein-PEG-Interaktionen und Verstärkung intramolekula-rer Salzbrücken) und einem entropischen Gewinn, der anscheinend eine noch bedeu-tendere Rolle spielt (Verdrängung schwach koordinierten Wassers um Hydroxylgrup-pen und um solvensexponierte, apolare Oberflächen). Diese Effekte sind umgebungs-abhängig und können durch einzelne Mutationen beeinflusst werden [99]. Eine nochweiterführende, zusammenfassende Übersicht möglicher Stabilisierungsmechanismenbzw. Beobachtungen aus der Literatur zu verschiedenen PBK ist in Tab. 1.4 gezeigt.20Kapitel 1. EinleitungTabelle 1.4: Mechanismen zur thermischen / chemischen PBK-Stabilisierung (Reproduzie-rung mit Genehmigung aus [96], Copyright 2018 American Chemical Society).3ProteinPolymerStabilisierungsmechanismus / BeobachtungRef.1CW / GCN4PEGAbschirmung von Rückgrat-HB,Abschirmung von intramolekularen Salzbrücken,Interaktionen mit Lysin an Positionen i+3, i+4[94, 102]CellulaseAm, DMAm,DMAm-AAm,DMAm-DMAEMAKeine Verbesserung thermischer / chemischer Stabilität[109]ChymotrypsinPEGBindung hydrophober Regionen, reduzierteProteinflexibilität[92]ChymotrypsinSpermidinHydrophobe Wechselwirkungen und HB[110]ChymotrypsinPCBAm, POEGMA,PDMAEMA, PAM, PSMABegünstigte Rückfaltung durchStabilisierung partiell entfalteter Proteine[96]ChymotrypsinPCBAm, PQA,PSMA, POEGMASäurestabil durch Reduktion elektrostatischerInteraktionen, Strecken des Polymers ins Solvens[111]Cytochrom CPEGThermisch instabil, aber Erhalt einermetastabilen, bioaktiven Konformation[112]Faktor VIIaGlycoPEGPEG verlangsamt Aggregation[113]InsulinPEGHydrophobe Wechselwirkungen, Wasserverdrängung[93]LysozymPEGWechselwirkung über HBoder keine Verbesserung thermischer Stabilität[114, 115]LysozymAm, DMAm, OEOA,Am/PCMA, Am/AA,Am/DMAEMA, AGAThermisch instabil durch Entropieänderungbeim Faltungsprozess, chemisch stabileionische Polymere interagieren mit Protein[116]Methionyl-G-CSFPEGReduktion der Aggregation[117]Pin 1 WW-DomänePEGEntropisch günstige Desolvatation, Stabilisierungdes Übergangs- und nativen Zustandswährend (Ent)faltung durch Wechselwirkungen,verringerte Entfaltungskonstante und Proteinflexibilität,Abschirmung hydrophober Flächen,intramolekularer Salzbrücken und NH-π-Interaktionen[33, 97, 99–101][102–104, 118]PyrophosphatasePOEGMA,PNIPAAmPolymer sollte länger als Distanzzwischen Konjugations- und Bindestelle sein,um durch eine Hydrathülle nahe letzterer dieAggregation des Proteins zu senken[119, 120]StaphylokinasePEGFlexibles Polymer bildet Hydrathüllezur sterischen Abschirmung[121]Src SH3-DomänePEGErniedrigte Entfaltungskonstante kf,SASA-Modulation stärker in nativer Struktur[95]TrypsinPEGWechselwirkung über HB[122]TrypsinDextrin, ST-HPMAThermische Stabilität und Schutz gegen Autolysedurch sterische Abschirmung und HB[123]3AA: Acrylsäure, AGA: N-Acryloyl-D-glucosamin, Am: Acrylamid, DMAEMA: Dimethylami-noethoxymethacrylat, DMAm: Dimethylacrylamid, OEOA: Oligo(ethylenoxid)methyletheracrylat,PAM: Poly(ammoniummethacrylat), PCBAm: Poly(carboxybetainacrylamid), PCMA: Phosphoryl-cholinmethacrylat, PEG: Poly(ethylenglykol), PNIPAAm: Poly(N-isopropylacrylamid), POEGMA:Poly(oligo(ethylenglykol)methacrylat), PSMA: Poly(sulfonatmethacrylat), ST-HPMA: semiteleche-lisches Poly[N-(2-hydroxypropyl)-methacrylamid.1.2. Molekulardynamik-Simulationen211.2Molekulardynamik-Simulationen1.2.1GrundlagenZum grundlegenden Verständnis des Verhaltens inhärent flexibler und dynamischerBiomoleküle auf atomarer Ebene stellt die in den 1950er- entwickelte und in den1970er-Jahren erstmals auf Proteine angewandte Technik der MD-Simulationen einwichtiges Werkzeug der pharmazeutischen Forschung dar [124–127], dessen Hinter-grund im Folgenden beschrieben werden soll.1.2.1.1Abbildung atomarer WechselwirkungenDa es sich bei Makromolekülen in kondensierten Phasen zumeist um Systeme mittausenden Atomen handelt und somit quantenmechanische (QM) Methoden mit ex-pliziter Berücksichtigung elektronischer Wellenfunktionen durch unterschiedliche Ba-sissätze äußerst zeit- und ressourcenintensive Berechnungen darstellen, werden fürsolche Simulationen klassische molekularmechanische Ansätze verfolgt. Ganze Atomewerden hierbei durch einzelne, in der Regel nicht polarisierbare Partikel mit definierterMasse und Ladung in einer Simulationsbox, die das Makromolekül und das unmit-telbar umgebene Solvens beinhaltet, repräsentiert. Sowohl intra- als auch intermole-kulare Wechselwirkungen der Partikel werden über einen Satz verschiedener Termebeschrieben, der in seiner Summe als Kraftfeld bezeichnet wird und die Gesamtener-gie Etotal des Systems definiert. Dieses umfasst Konstanten zu jedem darin definiertenAtomtyp und basiert in aller Regel auf einer Vielzahl festgelegter empirischer Para-meter, die zuvor durch experimentelle oder quantenmechanische Verfahren ermitteltwurden. Üblicherweise werden zur Beschreibung von intramolekularen Bindungen undWinkeln harmonische Potentiale und für Diederwinkel periodische Kosinusfunktionengenutzt. Elektrostatische Wechselwirkungen werden durch das Coulombsche Gesetzund Van-der-Waals (VdW)-Interaktionen mithilfe des Lennard-Jones-Potentials be-rechnet (Abb. 1.7). Zusätzliche Terme, z. B. für Out-of-Plane-Schwingungen, könnenebenfalls im Kraftfeld integriert sein [128–130]. Kraftfelder für biologische Systemesind bspw. AMBER, CHARMM, GROMOS und OPLS [131]. Sie können mit Pro-grammen wie AMBER, GROMACS und NAMD genutzt werden [132].1.2.1.2BewegungsgleichungenMithilfe des verwendeten Kraftfelds kann die Kraft Fi(t), die auf das Atom i mit derMasse mi und der Position xi(t) im euklidischen Raum in einer bestimmten Konfigu-ration des Systems zur Zeit t wirkt, bestimmt werden. Fi(t) stellt die Vektorsummeder Interaktionen mit anderen Partikeln im System dar. U(x(t)) ist die gesamte po-tentielle Energie, die wiederum eine Funktion der Positionen x(t) der Atome ist:Fi(t) = −∂U(x(t))∂xi(t)= mi∂2xi(t)∂t2= mi∂vi(t)∂t= mi ai(t)(1.11)22Kapitel 1. EinleitungAbbildung 1.7: Illustration exemplarischer Kraftfeld-Terme (am Beispiel von AMBER [133])zur Beschreibung inter- und intramolekularer Wechselwirkungen und damit der Gesamtener-gie Etotal des Systems. Bindungslängen werden mit einem harmonischen Potential unter Ver-wendung einer Konstanten Kr beschrieben. Hierfür ist die Abweichung der aktuellen Bin-dungslänge r von der Equilibriumsbindungslänge req entscheidend. In analoger Weise werdendie Winkel θ zwischen drei benachbarten Atomen beschrieben. Eine Torsion wird über eineperiodische Kosinusfunktion definiert, mit der energetischen Barriere Vn, der Periodizität nund der Phase γ. Nichtkovalente Wechselwirkungen werden über das Lennard-Jones-Potentialmit den Konstanten Aij und Bij, sowie elektrostatische Interaktionen mittels des coulomb-schen Gesetzes (Ladungen qi und qj und Dielektrizitätskonstante ϵ) beschrieben [129].Durch die Kraft Fi(t) können über das zweite Newtonsche Gesetz die Beschleunigungai(t) und, in Kombination dieser mit der Position xi(t) und der Geschwindigkeit vi(t)des Partikels zur Zeit t, auch die Position und Geschwindigkeit zur Zeit t+δt ermitteltwerden. Zur Integration der Bewegungsgleichungen dienen verschiedene Methodenauf Basis von Taylorreihen, wie z. B. der Verlet-Algorithmus, der die Position imvorangegangenen Zeitschritt t − δt zur Berechnung nutzt:xi(t + δt) = 2 xi(t) − xi(t − δt) + δt2 ai(t)(1.12)vi(t) = xi(t + δt) − xi(t − δt)2 δt(1.13)Gemäß Glg. 1.13 werden Geschwindigkeiten bei diesem Ansatz nach Erhalt der Posi-tionen des nächsten Zeitschritts berechnet. Alternativ kann das Leapfrog-Verfahrengenutzt werden, welches die Geschwindigkeiten und Positionen asynchron ermittelt.Das Programm NAMD nutzt als Integrator eine Form der Velocity-Verlet-Methode,bei der zunächst die neuen Positionen zum Zeitpunkt t + δt, im Anschluss darandie Geschwindigkeiten zum Zeitpunkt t + 12δt, danach über die Kraft Fi(t + δt) dieBeschleunigung ai(t + δt) und schließlich die Geschwindigkeit zum Zeitpunkt t + δtkalkuliert werden:1.2. Molekulardynamik-Simulationen23xi(t + δt) = xi(t) + δt vi(t) + 12 δt2 ai(t)(1.14)vi(t + 12δt) = vi(t) + 12 δt ai(t)(1.15)ai(t + δt) = Fi(t + δt)mi(1.16)vi(t + δt) = v(t + 12δt) + 12 δt a(t + δt)(1.17)Initiale Momente pi(0) und damit Geschwindigkeiten vi(0) aller Partikel werden in derRegel zu Beginn auf Basis der Maxwell-Boltzmann-Verteilung zufällig entsprechendder gewünschten Temperatur T(t) generiert:T(t) =1(3N − n) kBN�i=1|pi(t)|2mi(1.18)wobei N die Anzahl an Atomen im System, kB die Boltzmann-Konstante und n dieMenge an eingeschränkten Freiheitsgraden (z. B. durch den unten erwähnten SHAKE-Algorithmus) darstellen [134–136].1.2.1.3Wahl des ZeitschrittsDurch die Berechnungen mit dem vorher definierten Schritt δt kann somit eine Trajek-torie erhalten werden, die die Momente und Positionen aller Partikel der Simulations-box über die Zeit hinweg beschreibt. Um jedoch ein stabiles System zeiteffizient simu-lieren zu können, ist die Wahl des Zeitschritts entscheidend. Dieser sollte nicht größerals die höchste Frequenz sein, welche den Schwingungen der Schweratom-Wasserstoff-Bindungen zuzuordnen ist (etwa 0.1 fs). Daher werden die Schweratom-Wasserstoff-Bindungslängen in der Regel fixiert, zumeist mit dem sogenannten SHAKE- [137]oder auch, im Falle von NAMD, mit dem davon abgeleiteten RATTLE-Algorithmus[138]. Auf Basis dieser Eliminierung hochfrequenter Schwingungen wird der üblicheZeitschritt für Simulationen biomolekularer Systeme auf 2 fs gesetzt [134, 136].1.2.1.4Strukturaufbereitung und EquilibrierungStartkoordinaten des zu untersuchenden Proteins werden für eine MD-Simulation zu-meist auf Basis von Kristall- oder Nuklearmagnetresonanz (NMR)-Strukturen aus derProteindatenbank (PDB) [139] aufbereitet. Das System wird zunächst in einem ge-eigneten Protonierungszustand solvatisiert und für einige Schritte energieminimiert.Dies bedeutet, dass die Struktur auf Basis des verwendeten Kraftfelds in das, jenach eingestelltem Gradienten, nächstgelegene Energieminimum des hochdimensio-nalen Konformationsraums gelangt. In einer dem eigentlichen Produktionslauf voran-gestellten Equilibrierung wird sichergestellt, dass zentrale physikalische Kenngrößen(v. a. die kinetische und potentielle Energie, sowie Temperatur, Volumen und Druck)und Eigenschaften des Systems, welche relevant für die zu untersuchende Fragestel-lung sind, einen stabilen Fluktuationsbereich erreichen, bevor es zur Probennahme24Kapitel 1. Einleitung(dem sogenannten Sampling) für die Analyse kommt. Hierfür wird das System immikrokanonischen Ensemble (NVT), d. h. bei konstanter Anzahl an Partikeln (N),sowie bei konstantem Volumen (V) und einer an ein externes Wärmebad gekoppelten,konstanten Temperatur (T) simuliert. Da es sich meistens um Kristallstrukturen han-delt, die unter Kryobedingungen aufgenommen wurden, wird die Temperatur anfangssukzessive von 100 K auf den gewünschten Wert erhöht (meistens 300 K). Nachdemsich ein stabiler Energiewert Etotal und eine konstante Temperatur im NVT-Ensembleeingestellt haben, wird für kurze Zeit eine Equilibrierung im isothermal-isobarischenNPT-Ensemble zum Erreichen der gewünschten Dichte durchgeführt. Hierbei werdendie Partikelanzahl, der Druck (P) und die Temperatur konstant gehalten. Der Druckkann über ein Barostat kontrolliert werden. Anschließend wird der Produktionslaufim NPT-Ensemble gestartet, wobei Momentaufnahmen des Systems in regelmäßigenSchritten herausgeschrieben werden (üblicherweise jede Pikosekunde). Je nach zu un-tersuchendem System bieten sich verschiedene Wärmebäder zum Erhalt einer stabilenTemperatur an, so z. B. das Andersen-, das Langevin-, das Nosé-Hoover- oder dasBerendsen-Thermostat. Analog hierzu können das Andersen-, das Parrinello-Rahman-oder auch das Martyna-Tuckerman-Tobias-Klein-Barostat zur Einstellung eines kon-stanten Drucks dienen [129, 130].1.2.1.5Periodische RandbedingungenZur Verhinderung von Artefakten an der Oberfläche der Simulationsbox und zurgenerellen Minimierung von Effekten, die aus der Analyse eines endlichen Systemsresultieren, werden Simulationen biomolekularer Strukturen üblicherweise unter pe-riodischen Randbedingungen (Periodic boundary conditions, PBC) durchgeführt, dieeine Annäherung an Bulk-Bedingungen darstellen. Dies bedeutet, dass das Systemvon Abbildern seinesgleichen umgeben ist. Um intermolekulare Wechselwirkungenmit dem selben Partikel aus einer Nachbarbox zu unterbinden, werden bestimmte Li-mitierungen (Cut-off-Werte) gesetzt, die nicht größer sein sollten als die halbe Längeder Box. Im Falle von Van-der-Waals-Interaktionen und Coulomb-Wechselwirkungenauf kurzer Distanz wird üblicherweise eine Switching-Funktion ab einer Reichwei-te von 10 Å eingesetzt, mit der die Wechselwirkungen bis 12 Å graduell auf Nullabfallen. Da es im Falle von elektrostatischen Interaktionen aber zu großen Arte-fakten durch eine vollständige Annullierung ab einer bestimmten Reichweite käme(diese fallen nur mit einem Faktor von r−1 ab, das Lennard-Jones-Potential mit ei-nem Faktor von r−6), werden solche Wechselwirkungen über weite Distanzen hin(Long-Range-Interaktionen) mithilfe recheneffizienter, gitterbasierter Methoden ange-nähert. Hierfür werden vor allem die Particle-Mesh-Ewald-Methode [140], bei der eineschnelle Fourier-Transformation zum Einsatz kommt, oder entsprechende Abwand-lungen hiervon genutzt (Particle-Particle-Mesh- oder Smooth-Particle-Mesh-Ewald-Methode) [130].1.2. Molekulardynamik-Simulationen251.2.2Biokonjugat-SimulationenSeit der ersten MD-Simulation eines PEG-Protein-Konjugats auf atomarer Ebene[141] steigt die Menge pro Jahr publizierter Studien, die Simulationen PEGylierterProteine beschreiben. Dagegen verbleibt die Anzahl an Publikationen, die sich mitder Entwicklung bzw. Optimierung geeigneter Kraftfelder beschäftigen, auf einemrelativ niedrigen Niveau [142]. Die Simulation eines Biokonjugats stellt insofern ei-ne besondere Herausforderung dar. Kraftfelder werden weiterhin nur mit Fokus aufbestimmte Molekülarten entwickelt. AMBER14ffSB [133] basiert bspw. auf quanten-mechanischen Berechnungen mit dem Hartree-Fock (HF)-6-31G*-Basissatz und ei-ner Ladungsverteilung mittels der RESP (Restrained electrostatic potential)-Methode[143] und ist speziell für Proteine konzipiert. Daneben existieren auch Kraftfelder fürDNA (OL15 [144]), Kohlenhydrate (GLYCAM_06j [145]), Lipide (LIPID14 [146])oder kleinere organische Moleküle (GAFF2 [147]). Erste Parameter für PEG-Kettenwurden in den frühen 1990er Jahren entwickelt [148]. Für Hybrid-Moleküle mit kova-lent gebundenen Polymeren gibt es jedoch noch keine weitreichend etablierten Para-meter. Für PEG-Protein-Konjugate wird oft das CHARMM-Kraftfeld C35r [149] fürEtherstrukturen genutzt, welches mit Experimenten übereinstimmende Werte bzgl.hydrodynamischer Radien und Persistenzlängen zeigte [32]. Alternativen stellen auchdas TraPPE-UA- und eine modifizierte Version des GAFF-Kraftfelds dar [150–152].Doch auch andere Kraftfelder wurden in der Vergangenheit für Simulationen vonPEG-Konjugaten genutzt (z. B. AMBER99SB-ILDN [153] für die WW-Domäne vonPin 1 [100] oder das United-Atom (UA)-Kraftfeld GROMOS96 43a1 [154] für Insulin[93]).Weiterhin stellt die Größe der Biokonjugate ein Problem dar. Da in der Regel Polyme-re mit einem Molekulargewicht von 5 bis 40 kDa konjugiert werden, stellen Simulatio-nen dieser Strukturen auf atomarer Ebene auch heute noch einen immensen Zeitauf-wand dar. Diese Konjugate repräsentieren Systeme mit zahlreichen Freiheitsgradenund ein ausreichendes Sampling ist dadurch nur bedingt zu erreichen. Um diesesProblem zu lösen, werden entweder erweiterte Sampling-Verfahren (z. B. gesteuerte[155] oder beschleunigte [156] MD-Simulationen) oder aber, sehr viel öfter, vergröber-te (Coarse-grained, CG) Kraftfelder eingesetzt [32]. Bei letzterer Methode wird derDetailgrad der atomaren Struktur durch einen Zusammenschluss mehrerer Atome zueinzelnen Partikeln bzw. Kugeln (Beads) drastisch gesenkt. Bei UA-Kraftfeldern, wiedem oben genannten GROMOS96 43a1, werden Wasserstoff-Atome mit den jewei-ligen Schweratomen zusammengeführt. Deutlich gröber ist das MARTINI-Kraftfeld[157]. Das zunächst für Lipide entwickelte CG-Kraftfeld wurde auf Proteine [158],Kohlenhydrate [159] und Nukleinsäuren [160, 161] erweitert und vereint in der Re-gel vier Atome zu einem von 20 möglichen Partikeltypen (je nach Polarität, Ladungund Größe). Durch CG-Ansätze wird die Anzahl an Dimensionen des Konformati-onsraums verringert und Simulationen großer Konjugate werden so in annehmbarer26Kapitel 1. EinleitungZeit durchführbar (Abb. 1.8). Dies führt jedoch auch zu einem deutlichen Detailver-lust. Zusätzlich müssen CG-Methoden für neue Moleküle (bspw. Polymere) je nachzu analysierender Eigenschaft auf Basis experimenteller Messungen (z. B. über Gyra-tionsradien) neu definiert und iterativ über einen Vergleich mit ausreichend langen,klassischen MDs mit atomarem Detail optimiert werden [162]. Für PBK existierennoch keine allgemein etablierten CG-Ansätze [32], wobei erst kürzlich ein Simulati-onsprotokoll für das MARTINI 3-Kraftfeld veröffentlicht wurde [163].pm nm Größe des zu untersuchenden Systems fs μs Zeit QM-Methoden Simulationen mit atomarem Detail Vergröberte Kraftfelder Mizellare Systeme μm ms Lokale Bewegungen in Proteinen Einzelne Moleküle Gößere Konformationsänderungen in Proteinen Abbildung 1.8: Geeignete Simulationsansätze und üblicher Zeitraum zur Beobachtung be-stimmter Dynamiken in Systemen unterschiedlicher Größe. Rechenintensive QM-Methodeneignen sich für Atome bis hin zu einzelnen Molekülen. Klassische MD-Simulationen auf ato-marer Ebene können lokale Bewegungen von Proteinen registrieren. Vergröberte Kraftfelderreduzieren die Anzahl möglicher Freiheitsgrade und ermöglichen größere Konformationsände-rungen. Angelehnt an Abb. 1 aus [162].1.2.3KonvergenzkriterienEin ausreichendes Sampling der MD-Simulationen ist nötig, um zuverlässige Analysender Trajektorien durchführen zu können. Klare Richtlinien für minimale Simulations-zeiten molekularer Systeme existieren jedoch nicht. Eine Simulation gilt üblicher-weise als konvergiert, wenn aus dem möglichen Konformations- bzw. Phasenraum(Phase space) ausreichende Probennahmen gezogen wurden, die einer Boltzmann-Gewichtung entsprechen. Nur dann ist gemäß der Ergodenhypothese garantiert, dassdie Eigenschaften über die Zeit hinweg mit jenen eines makroskopischen Systemsgleichzusetzen sind [136, 164]. Es gilt also, fernab von Nichtgleichgewichtssystemen,die Annahme:1.2. Molekulardynamik-Simulationen27⟨G⟩eq = limt→∞⟨G⟩t(1.19)wobei ⟨G⟩eq den Wert der betreffenden Eigenschaft im Equilibrium beschreibt und⟨G⟩t das entsprechende Ensemblemittel. Diese Annahme ist generell kritisch zu sehen,da sie streng genommen nur für unendlich lange Simulationen gegeben ist. Alternativkönnen ausreichend viele Replika zur Berechnung der gesuchten makroskopischenEigenschaften durchgeführt werden (Ensemble averages) [164]. Eine Konvergenz istultimativ nicht beweisbar, da, egal nach welcher Simulationszeit, nie eruiert werdenkann, ob es noch weitere, unbesuchte Konformationen im Phasenraum gibt.1.2.3.1Self-consistency checksAuch wenn ein ausreichendes Sampling nicht klar bestimmbar ist, so existieren den-noch Methoden, um zu überprüfen, ob zumindest der bisher besuchte Konformati-onsraum ein in sich konsistentes System abbildet (sogenannte Self-consistency checks,SCC). Zur Evaluation einzelner Variablen aus Simulationen ist eine Berücksichtigungder Korrelation zeitlich aufeinanderfolgender Momentaufnahmen von Bedeutung. DerStandardfehler (Standard error, SE) des Mittelwerts kann folgendermaßen berechnetwerden:SE =σ√Nind∼� τtsim(1.20)wobei σ die Standardabweichung, Nind die Anzahl an unabhängigen Messungen, τ dieaus der Autokorrelationsfunktion berechenbare Korrelationszeit und tsim die gesam-te Simulationszeit darstellen [136, 165, 166]. Im Folgenden sollen drei SCC-Kriterienbeschrieben werden, die eine allgemeine, strukturelle Konvergenz signalisieren [167].Beim SDVG-Kriterium (benannt nach den Autoren Smith, Daura und van Gunste-ren [168]) wird ein Clustering durchgeführt, z. B. auf Basis des RMSD-Werts derCα-Atome. Die Anzahl an gefundenen Clustern sollte sich hierbei gegen Ende derSimulation hin nicht mehr verändern. Dies deutet darauf hin, dass keine neuen Berei-che des möglichen Konformationsraums mehr erreicht werden. Dieses Kriterium kannjedoch nicht detektieren, ob die Struktur in einem metastabilen Zustand gefangenist. Als ergänzende Analyse kann daher die Clusterentropiekonstanz (Constancy ofcluster entropy, CCE) dienen, definiert nach:CCE =�pi log(pi)(1.21)Sie beschreibt die Wahrscheinlichkeit pi, die Struktur zu gegebener Zeit in einemder bereits gefundenen Cluster i wiederzufinden. Im Falle einer in einem Zustandgefangenen Struktur sinkt die CCE stetig. Für eine ausreichende Konvergenz solltesich dieser Wert gegen Simulationsende hin daher nicht mehr stark ändern [167].Ein drittes Kriterium stellt die BBCOM/BCOM-Methode von Romo und Grossfielddar (Bootstrapped block covariance overlap/block covariance overlap) [169]. Hierbei28Kapitel 1. Einleitungwerden Hauptkomponentenanalysen (Principal component analysis, PCA), bspw. aufBasis der Cα-Atome, durchgeführt, um deren Kovarianz zu bestimmen. Für eine PCAwird für N Atome im System eine 3N x 3N-Korrelationsmatrix erstellt:Cij = ⟨xi − xi⟩⟨xj − xj⟩(1.22)wobei xi die x, y oder z-Koordinate des Atoms i darstellt und xi den Wert derdurchschnittlichen Struktur repräsentiert. Über eine Diagonalisierung der erhaltenenMatrix werden Eigenvektoren erhalten, die die zentralen Bewegungsrichtungen derStruktur im Raum beschreiben, sowie Eigenwerte, die die durchschnittliche Fluk-tuation entlang dieser Vektoren beschreiben [168]. Anschließend wird das Verhältniszwischen der Kovarianzübereinstimmung von Simulationsblöcken ansteigender Größemit der gesamten Trajektorie und der Kovarianzübereinstimmung von Simulations-blöcken, die zunächst einem Bootstrapping-Verfahren unterzogen wurden, mit dergesamten Trajektorie berechnet. Im Idealfall sollte das Verhältnis einen Wert von 1erreichen. Beim Bootstrapping-Verfahren werden zufällige Momentaufnahmen aus derTrajektorie ausgewählt, wodurch vermieden wird, dass ausschließlich Strukturen mithoher Korrelation zueinander selektiert werden. Die Kovarianzübereinstimmung zwi-schen zwei Hauptkomponenten mit den Eigenwerten λi und λj und den Eigenvektoren−→v i und −→v j der Ensembles A und B ist definiert nach [169]:ΩA,B = 1 −�����NModii(λAi + λBi ) − 2 �NModii�NModij�λAi λBi (−→v Ai · −→v Bi )2�NModii(λAi + λBi )(1.23)1.2.3.2Weitere KonvergenztestsFür Markow-Chain-Monte-Carlo (MCMC)-Verfahren existieren weiterhin verschiede-ne Tests, um eine stationäre Verteilung einzelner Parameter zu evaluieren. Hierbeiwird geprüft, ob die Probennahmen aus einer sich nicht mehr ändernden und damitzeitinvarianten Verteilung stammen. Dies bedeutet also, dass die Übergangsmatrixder Markow-Kette über die Zeit hinweg unverändert bleibt [170, 171]. Während essich bei diesen Monte-Carlo-Verfahren um Prozesse handelt, bei denen im Gegen-satz zu molekulardynamischen Ansätzen zufällige Stichproben ohne zeitliche Abfolgegeneriert werden, so sollte mit diesen Tests analog auch eine Stationarität einzel-ner Variablen, die aus einer Zeitreihe wie einer MD-Simulation gewonnen wurden,überprüfbar sein.Zwei dieser Konvergenztests wurden in dieser Arbeit getestet: Beim Heidelberger-Welch-Test wird die Stationarität mit einem Cramér-von-Mises-Test über die gesamteTrajektorie geprüft. Wird dieser Test nicht erfüllt, werden die ersten 10 % der Zeitrei-he verworfen und die Konvergenz erneut geprüft. Dies wird so lange durchgeführt, bisdie Stationarität entweder erfüllt wird oder aber 50 % der Zeitreihe verworfen wurden,1.2. Molekulardynamik-Simulationen29wobei die Nullhypothese, dass es sich um eine stationäre Verteilung handelt, in letz-terem Fall nicht erfüllt wurde. Wurde dieser erste Teil des Heidelberger-Welch-Testserfüllt, so wird in einem zweiten Verfahren geprüft, ob das 95%-Konfidenzintervalldes Mittelwerts der nicht verworfenen Zeitreihe kleiner ist als 10 % dieses Mittel-werts [172–174]. Beim Geweke-Test werden hingegen die Mittelwerte der ersten undder zweiten Hälfte der Zeitreihe miteinander verglichen, um einen Z-Wert für derenÜbereinstimmung zu erhalten. Eine Stationarität ist wahrscheinlicher, je näher dieserWert bei 0 liegt (idealerweise unter 2) [174, 175].1.2.4Gaußbeschleunigte Molekulardynamik-SimulationenEine Methode des erweiterten Samplings stellen gaußbeschleunigte MD-Simulationendar (Gaussian accelerated MDs, GaMDs), eine Weiterentwicklung der beschleunig-ten MD-Simulationen (Accelerated MDs, aMDs) [176, 177]. Bei dieser Methode wirdein harmonisches Zusatzpotential ∆V (−→r ) zum zugrundeliegenden Potential V (−→r )addiert, welches die energetischen Barrieren aller Konformationsänderungen verrin-gert. Somit wird eine modifizierte Energielandschaft mit dem Potential V ∗(−→r ) =V (−→r )+∆V (−→r ) erhalten. Nur selten beobachtbare Übergänge in herkömmlichen MD-Simulationen (Conventional MDs, cMDs) werden so schneller realisiert. Im Gegensatzzum aMD-Ansatz folgt dieses Zusatzpotential jedoch einer Gauß-Verteilung, wodurcheine verbesserte energetische Neugewichtung mithilfe einer Kumulanten zweiter Ord-nung ermöglicht wird. Neugewichtungen mit geringen Fehlern können so auch für grö-ßere Systeme durchgeführt werden, im Falle von aMD-Simulationen liegt die Grenzefür eine mögliche Neugewichtung typischerweise bei Proteinen mit 40 Aminosäuren.Das Potential ∆V (−→r ) wird nur addiert, sofern das Grundpotential geringer als dieReferenzenergie E ist:∆V (−→r ) =�12k0Vmax−Vmin (E − V (−→r ))2,V (−→r ) < E0,V (−→r ) ≥ E(1.24)Vmax und Vmin stellen die Maxima und Minima des Potentials dar und k0 ist eineKraftkonstante zwischen 0 und 1, wobei größere Werte eine stärkere Modifizierung derEnergielandschaft bedeuten. Die Referenzenergie E kann Werte zwischen Vmax undVmin + 1k annehmen, wobei die eigentliche Kraftkonstante k über folgende Gleichungdefiniert ist:k =k0Vmax − Vmin(1.25)Standardmäßig ist E jedoch als Vmax festgesetzt. Die Kraftkonstanten und die Re-ferenzenergie werden somit automatisch bestimmt. Lediglich das obere Limit derStandardabweichung des Potentials σ0 wird vor der Simulation durch den Nutzereingestellt. Es steht mit den Parametern k und E in folgendem Zusammenhang:σ∆V = k(E − Vav)σV ≤ σ0(1.26)30Kapitel 1. EinleitungDabei stellen σ∆V die Standardabweichung des Zusatzpotentials, σV jene des Poten-tials und Vav das durchschnittliche Potential dar. Mit E = Vmax gilt, dass k0 einenMaximalwert von 1 erreichen kann [176, 177]:k0 = min(1.0, σ0σVVmax − VminVmax − Vav)(1.27)Ein Zusatzpotential kann über den GaMD-Ansatz zum Diederpotential im System,zum gesamten Potential, sowie zu beiden Elementen gleichzeitig addiert werden (Du-al Boost-Schema). Die Übereinstimmung von ∆V mit einer tatsächlichen GaußschenVerteilung kann an der Anharmonizität γ der Verteilung des Zusatzpotentials bemes-sen werden, welche bei größeren Werten keine genaue Neugewichtung erlaubt. Hierzuwird die Entropie S∆V von ∆V von der jeweiligen maximalen Entropie Smax eineridealen Gauß-Verteilung subtrahiert [178, 179]:γ = Smax − S∆V = 12 ln(2πeσ2∆V ) +� ∞0p(∆V ) ln(p(∆V )) d∆V(1.28)Der Parameter p(∆V ) stellt die normierte Wahrscheinlichkeitsverteilung von ∆V fürdie Momentaufnahmen in einer der Klassen (Bins) der zur Neugewichtung gewähl-ten Reaktionskoordinate(n) dar, e ist die Eulersche Zahl. Zur Simulation wird übli-cherweise eine kurze Phase einer cMD-Simulation vorangeschaltet. Nachdem sich dasSystem in dieser über einige Zeitschritte hinweg stabilisiert hat, werden Werte fürdie Parameter Vmax, Vmin, Vavg und σV gesammelt. Am Ende dieser Phase werdenebenfalls E und k0 ermittelt. In einer nachgeschalteten, kurzen Phase der GaMD-Equilibrierung wird das Zusatzpotential dann entsprechend dieser Parameter zumGrundpotential addiert. Nach dieser Equilibrierung werden E und k0 dann im eigent-lichen Produktionslauf gemäß eines vordefinierten Zeitintervalls regelmäßig aktua-lisiert. Nach abgeschlossener Simulation können mithilfe des PyReweighting-SkriptsNeugewichtungen auf Basis von Kumulanten, einer Maclaurin-Serie oder des expo-nentiellen Durchschnitts erfolgen [178]. Durch die, im Gegensatz zu Metadynamik-Simulationen (vide infra), ungerichtete Beschleunigung bei einer GaMD-Simulation,können ohne Vorwissen geeigneter Kollektivvariablen das Sampling generell verbessertund anspruchsvolle Konformationsänderungen wie bspw. Protein(ent-)faltungen oderLigand-Protein-Bindungsprozesse beobachtet werden [176, 177]. Auch eine Modifika-tion des Algorithmus zur Berechnung von Ligand-Bindungskinetiken wurde kürzlichpubliziert [180].1.2.5Wohltemperierte Metadynamik-SimulationenErweitertes Sampling kann auch gerichtet über vorher definierte Kollektivvariablen(Collective variables, CV), bspw. bestimmte Distanzen oder Winkel im System, er-folgen. Dies eignet sich besonders im Falle von im Vorfeld bekannten Konformati-onsänderungen, die mit der Simulation beobachtet werden sollen und bei denen dieVariablen, die zu einem Übergang führen, bekannt sind. Eine Methode, die diesen1.2. Molekulardynamik-Simulationen31Ansatz verfolgt, ist jene der wohltemperierten Metadynamik-Simulationen mit mul-tiplen „Walkern“ (Multiple walkers well-tempered metadynamics simulations, MW-wtMetaMDs). Bei MetaMDs wird ein kontextabhängiges Zusatzpotential entlang derCV zum Grundpotential des Systems addiert. Dies bedeutet, dass zunächst v. a.die Minima der Energielandschaft mit einem Zusatzpotential „aufgefüllt“ werden, inwelcher sich die Struktur anfangs befindet. Durch das Zusatzpotential werden ener-getische Barrieren überwunden und alle Minima im Hinblick auf die definierten CVsukzessive erreicht [181–183]. Das Potential des Systems ist hier definiert nach:VG(s, t) = ω�t′=τG,2τG,...(t′<t)e−(s(t)−s(t′))22σ2(1.29)Dabei ist s(t) der Wert der Kollektivvariablen s zur Zeit t. Die Gaußsche Höhe ω re-präsentiert die Größe des hinzugefügten Potentials, welches alle τG Zeitschritte zumPotential addiert wird. Zusätzlich wird die Breite der Gauß-Kurve entlang der be-trachteten Variable mit dem Parameter σ definiert. Im Falle mehrerer CV (Anzahld) ergibt sich folgende modifzierte Gleichung:VG(s, t) = ω�t′=τG,2τG,...(t′<t)e− �dα=1(sα(t)−sα(t′))22σ2α(1.30)Durch ein ausreichendes Sampling kann vom modifizierten Potential auf die zugrunde-liegende freie Energie F(s) der unmodifizierten Energielandschaft geschlossen werden:limt→∞ VG(s, t) ∼ −F(s)(1.31)Das Sampling entlang der CV wird maßgeblich durch höhere Werte von ω, σ undniedrigere τG-Werte begünstigt. Allerdings sollten diese Werte mit Vorsicht justiertwerden, da geringere Energiebarrieren durch große repulsive Gauß-Kurven mit hoherFrequenz gleichzeitig auch die Qualität des rekonstruierten, zugrundeliegenden En-ergieprofils beeinträchtigen [181–183]. Im Falle von wtMDs wird die Konvergenz desEnergieprofils deutlich begünstigt [184]. Bei dieser Abwandlung wird ein Bias-Faktorγ = T+∆TTdefiniert (oft mit Werten zwischen 6 und 15), der den Wert ∆T bestimmt,durch welchen wiederum die Gaußschen Kurven mit der Anfangshöhe ω0 über dieSimulationszeit hinweg ihre Höhe reduzieren:ω = ω0e−VG(s,t)kb∆T(1.32)Dabei würde ∆T = 0 einer reinen cMD und ∆T → ∞ einer normalen MetaMDentsprechen. Durch diese Modifikation ergibt sich folgende Beziehung:limt→∞ VG(s, t)T + ∆T∆T∼ −F(s)(1.33)32Kapitel 1. EinleitungDurch multiple „Walker“, d. h. mehrere parallel geschaltete, miteinander „kommuni-zierende“ Replika, wird das Sampling weiter verbessert [185]. Diese „Walker“ (AnzahlNw) füllen die Minima gleichzeitig auf und fließen lediglich in einem gemeinsamen Po-tential VG in folgender Beziehung zusammen:VG(s, t) = ωτGNw�i=1� ti=0dt′ e−|s(t)−s(t′)|22|σ|2(1.34)Durch eine höhere Anzahl an multiplen „Walkern“ wird die Qualität der freien Ener-gielandschaft nicht verbessert, jedoch wird die nötige Simulationszeit zum Erreicheneiner Konvergenz erniedrigt.In der Praxis werden üblicherweise zusätzliche, hohe energetische Barrieren als Mi-nima bzw. Maxima entlang der Reaktionskoordinaten, den CV, definiert. Dies dientdazu, dass die CV nur den für die Fragestellung relevanten Konformationsraum er-reichen und es in diesem zu einem ausreichenden Sampling kommt. Für geeigneteWerte von ω, σ, τG und γ existieren prinzipiell keine Richtwerte und es sollten fürjedes spezifische System verschiedene Kombinationen getestet werden. Je nach ge-wählter Kollektivvariable finden sich in der Literatur jedoch immer wiederkehrende,typische Bereiche zur Untersuchung von Proteinen, die als erster Anhaltspunkt dienenkönnen. Als Richtwert für σ wird bspw. ein Drittel oder die Hälfte der Standardabwei-chung der Kollektivvariablen in einer klassischen MD verwendet [186]. Als beliebtesMetaMD-Modul hat sich bisweilen PLUMED (Plugin for molecular dynamics) eta-bliert, welches in Kombination mit verschiedenen Simulationsprogrammen genutztwerden kann [187, 188].1.3. QSPR-Modelle zur Stabilitätsvorhersage331.3QSPR-Modelle zur Stabilitätsvorhersage1.3.1Grundlagen1.3.1.1HintergrundModellezurBeschreibungvonquantitativenStruktur-Wirkungs/Eigenschafts-Beziehungen (Quantitative structure-activity/property relationships, QSAR/QSPR)werden in der medizinischen Chemie schon seit mehreren Jahrzehnten eingesetzt. Siestellen eine Schnittstelle zwischen der Informatik und der Chemie dar und habensich v. a. durch die immer noch rasant steigenden Mengen an verfügbaren Messdatenals hilfreiches Werkzeug zur Aufklärung möglicher Korrelationen zwischen Molekülei-genschaften und gewünschten biologischen oder physikochemischen Eigenschaften vonMolekülen und darauf aufbauende Wirkstoffentwicklungen etabliert [189]. Die Studienvon Hansch und Fujita zu Effekten des Octanol-Wasser-Verteilungskoeffizienten pflan-zenwachstumsbeeinflussender Moleküle werden üblicherweise als zentraler Grundsteindieses Forschungsbereichs angesehen [189–193]. Durch eine steigende Rechenleistungund die Etablierung von Hochdurchsatz-Screenings in der Chemie haben sich diemethodischen Ansätze stark weiterentwickelt, sodass nicht nur simple lineare Korre-lationen Anwendung finden, sondern, unter Zuhilfenahme von Werkzeugen aus dervon Wold begründeten Chemometrik [194] und dem maschinellen Lernen (Machinelearning), auch andere Methoden wie bspw. Support Vector Machines oder künstlicheneuronale Netze [195].1.3.1.2DeskriptorenAls Variablen zur Modellgenerierung werden sogenannte Deskriptoren verwendet, diebestimmte Eigenschaften eines Moleküls beschreiben. Diese werden mit einer odermehreren Antwortvariablen, zumeist der biologischen Aktivität, korreliert. Deskrip-toren lassen sich in 1D (z. B. Summenformeln), 2D (Strukturformeln), 3D (Formelnmit Informationen zur räumlichen Orientierung bzw. Konformation), 4D (z. B. übereine Berücksichtigung eines zeitlichen Verlaufs oder mehrerer Konformationen oderProtonierungszustände gleichzeitig), 5D (Berücksichtigung der induzierten Passfor-men von Protein-Ligand-Komplexen) oder gar 6D (Berücksichtigung verschiedenerSolvensmodelle) einteilen [189, 196]. Während 4/5/6D-Deskriptoren noch sehr seltengenutzt werden, haben sich 3D-QSAR-Methoden besonders durch die Verwendungder PLS-Methode (Partial least squares) [197] nach Erscheinen des CoMFA (Compa-rative molecular field analysis)-Verfahrens [198] weithin in diesem Forschungsbereichetabliert [189]. Theoretische Hintergründe zu in dieser Arbeit relevanten Deskriptorenwerden in Anhang B aufgezeigt.34Kapitel 1. Einleitung1.3.2Statistische ValidierungsparameterZur Evaluation der prädiktiven Leistung von QSPR-Modellen haben sich verschiedeneParameter etabliert, von denen im folgenden Kapitel einige erläutert werden sollen.1.3.2.1RegressionsparameterIn [199] werden die bekanntesten Validierungsparameter für regressionsbasierte Fra-gestellungen zusammengefasst. Der am häufigsten genutzte Parameter ist der Deter-minationskoeffizient R2, auch Bestimmtheitsmaß genannt:R2 =1 −�(yi − ˆyi)2�(yi − ¯y)2(1.35)=1 − RSSTSS(1.36)Hierbei stellen RSS (Residual sum of squares) und TSS (Total sum of squares)die Residuenquadratsumme und die Quadratsumme der Abweichungen zum Mittel-wert dar. yi sind die einzelnen Messwerte, ¯y der Mittelwert der Messwerte, sowie ˆyidie Vorhersagen. Zusätzlich kann ein adjustiertes Bestimmtheitsmaß R2adj ermitteltwerden, welches die Anzahl verwendeter Deskriptoren p und die Datensatzgröße nberücksichtigt [200]:R2adj =1 − (1 − R2)n − 1n − p − 1(1.37)Mit R2 wird das Bestimmtheitsmaß zumeist für die Messwerte, auf denen die Mo-dellgenerierung basiert, bezeichnet. Wird der Determinationskoeffizient für Werte be-rechnet, die nicht Teil des Trainingsdatensatzes sind (n Datenpunkte eines externenSets EXT), wird dieser zumeist mit Q2 gekennzeichnet:Q2 =1 −�nEXTi=1(yi − ˆyi)2�nEXTi=1(yi − ¯y)2(1.38)=1 − PRESSTSS(1.39)Die entsprechende RSS wird hier als PRESS (Predicted residual sum of squares)bezeichnet. Hier kann entweder die TSS des Trainingsdatensatzes (Q2F1 nach Shi [201])oder, damit der Parameter tatsächlich nur auf den externen Daten basiert, die desTestdatensatzes genutzt werden (Q2F2 nach Schüürmann [202]). Beide Q2-Variantenzeigen unter Umständen unrealistische Abschätzungen der prädiktiven Leistung desModells, wenn die Verteilung des externen Datensatzes nicht vergleichbar mit der desTrainingsdatensatzes ist [203]. Ein sehr viel robusterer Validierungsparameter stellthingegen der Konkordanzkorrelationskoeffizient (Concordance correlation coefficient,1.3. QSPR-Modelle zur Stabilitätsvorhersage35CCC) nach Lin dar [204, 205], welcher folgendermaßen definiert ist:CCC =2 �nEXTi=1(yi − ¯y)(ˆyi − ¯ˆy)�nEXTi=1(yi − ¯y)2 + �nEXTi=1(ˆyi − ¯ˆy)2 + nEXT (¯y − ¯ˆy)2(1.40)Die Variable ¯ˆy stellt den Mittelwert der Vorhersagewerte dar. Der Validierungspa-rameter CCC berücksichtigt sowohl die Residuen der Werte zur Regressionslinie alsauch die Übereinstimmung letzterer mit den experimentellen Werten (gekennzeich-net durch eine annähernde Steigung von 1 beim Auftragen von experimentellen gegenprädiktive Werte) [199]. Daneben existieren ebenso verschiedene Rangkorrelationsko-effizienten, die besonders dann bedeutend sind, wenn es vorwiegend darum geht, einModell zu generieren, welches, ungeachtet der Abweichungen von den experimentellenMesswerten, die korrekte Reihenfolge der Datenpunkte vorhersagt. Zu erwähnen isthierbei das Kendall’sche Tau τb [206, 207]:τb =C − D�(C + D + X0)(C + D + Y0)(1.41)C ist die Anzahl an übereinstimmenden Paaren, D jene an nicht übereinstimmendenund X0 sowie Y0 die Anzahl an punktgleichen Paaren entlang der x- bzw. y-Achse[208]. Zur Beschreibung des Modellfehlers kann entweder der mittlere absolute Fehler(Mean absolute error, MAE) oder die Wurzel der Summe der quadrierten Fehler(Root-mean-square error, RMSE) verwendet werden, wobei letztere größere Fehlerstärker gewichtet:MAE =�nEXTi=1|yi − ˆyi|nEXT(1.42)RMSE =��nEXTi=1(yi − ˆyi)2nEXT(1.43)1.3.2.2KlassifikationsparameterZur Beurteilung der Leistung eines Klassifikationsmodells mit zwei verschiedenenKlassen kann die korrekte Klassifizierungsrate (Correct classification rate, CCR), dieauch der Richtigkeit (Accuracy, Acc) entspricht, herangezogen werden:CCR = NcorrNtot =TP + TNTP + TN + FP + FN(1.44)Hier wird die Anzahl an korrekt klassifizierten Fällen Ncorr mit der Gesamtzahl Ntotin Relation gesetzt. Dies entspricht also der Summe an richtig positiven (True po-sitives, TP) und negativen Fällen (True negatives, TN), geteilt durch die gesam-te Anzahl an Fällen, inklusive falsch positiver (False positives, FP) und falsch ne-gativer (False negatives, FN). Handelt es sich um relativ ungleiche Größen zwei-er Klassen, so sollte eine entsprechende Gewichtung berücksichtigt werden, z. B.über die ausgeglichene Richtigkeit (Balanced accuracy, BAcc), welche definiert ist als36Kapitel 1. EinleitungBAcc = Spezifit¨at+Sensitivit¨at2[209]. Die Spezifität beschreibt den Anteil an tatsächlichnegativen Fällen, welche korrekt als negativ eingestuft wurden (True negative rate,TNR). Die Falsch-positiv-Rate (FPR) ist entsprechend definiert als 1 − Spezifit¨at.Die Sensitivität stellt dagegen die Richtig-positiv-Rate (True positive rate (TPR) oderrecall) dar und beschreibt den Anteil an tatsächlich positiven Fällen, der korrekt alspositiv eingestuft wurde. Zusätzlich existiert die Präzision (Positive predictive value,PPV). Sie setzt die Anzahl an korrekt positiv eingestuften Fällen in Relation zur Ge-samtzahl an positiv eingestuften Fällen. Das Gegenteil der Präzision ist die Segreganz(Negative predictive value, NPV) [210, 211]. Die beschriebenen Kriterien können ausder entsprechenden Konfusionsmatrix abgeleitet werden (Abb. 1.9).TPR =TPTP + FN(1.45)TNR =TNTN + FP(1.46)FPR =1 − TNR =FPTN + FP(1.47)PPV =TPTP + FP(1.48)NPV =TNTN + FN(1.49)Abbildung 1.9: Darstellung einer Konfusionsmatrix, bei der die tatsächlichen Klassifizie-rungen den Vorhersagen gegenübergestellt werden. Es ergeben sich richtig positive, falschpositive, richtig negative und falsch negative Fälle. Die im Text beschriebenen Kriterien nut-zen jeweils einen bestimmten Anteil an Fällen, der in der Abbildung farblich markiert ist.Aufbauend auf diesen Definitionen stellt die Berechnung der Fläche unter der Grenz-wertoptimierungskurve (ROC) eine etwas weiterführende Beschreibung der Modell-leistung dar (Area under the receiver operating characteristic curve, AUROC bzw.AUC). Zur Erstellung einer ROC-Kurve wird die ermittelte TPR gegen die FPRaufgetragen. Eine früh und stark ansteigende Kurve weit über der Diagonalen desGraphen deutet auf ein Modell mit guter Klassifikationsleistung hin. Dagegen re-präsentiert eine Kurve nahe der Diagonalen eine Modellleistung, die einer Zufalls-klassifizierung entspricht. Die Kurve verdeutlicht somit die Leistung des Modells beiverschiedenen Grenzwerten.Handelt es sich um einen Datensatz mit vielen negativen und wenigen positiven Fällen,kann es von besonderem Interesse sein, die Leistung des Modells im Hinblick auf dieDetektion letzterer beurteilen zu wollen. Bei einer solch ungleichen Verteilung kanndaher auch die AUPRC (Area under the precision-recall curve) berechnet werden,1.3. QSPR-Modelle zur Stabilitätsvorhersage37die die Fläche unter der Kurve darstellt, wenn die TPR gegen die PPV aufgetragenwird. Hierbei fließt die Anzahl an tatsächlich negativen Fällen nicht mit ein. ImGegensatz zu einer ROC-Kurve, bei der die Diagonale einer Zufallswahrscheinlichkeitmit dem AUC-Wert von 0.5 entspricht, stimmt der Basiswert der AUPRC mit demAnteil an positiven Fällen an der Gesamtzahl überein. Je nach Situtation kann daherauch ein niedriger AUPRC-Wert eine gute Detektion der tatsächlich positiven Fällesignalisieren. Die Segreganz spielt dagegen eine bedeutende Rolle bei der Beurteilungder Detektion negativer Fälle [212–214].1.3.3Modellarten dieser StudieDa in dieser Arbeit QSPR-Modelle auf Basis sehr kleiner Datensätze generiert wur-den, wurde der Fokus v. a. auf multilineare Regressions- und logistische Klassifi-kationsmodelle gelegt. Durch Nutzung von Datenbanken wie ChEMBL [215] liegendie Datensatzgrößen für QSAR-/QSPR-Ansätze heutzutage oftmals in einem Bereichzwischen 104 und 106 Molekülen [216], dagegen kommen Datensätze im Bereich umlediglich 100 Verbindungen, wie in dieser Arbeit, deutlich seltener vor [217, 218].Im zuerst durchgeführten Ansatz 1 in Kap. 3.1.6.2 wurden jedoch auch eine Viel-zahl anderer, komplexerer Modelle getestet. Hierbei lieferten neuronale Netze undXGBoost-Modelle erwähnenswerte Resultate. Der theoretische Hintergrund der viererwähnten Modellarten soll im Folgenden kurz erläutert werden.1.3.3.1Multiple lineare RegressionBei einer multiplen linearen bzw. multilinearen Regression werden die N Messergeb-nisse y1, . . . , yi, . . . , yN mit den entsprechenden Werten x1, . . . , xi, . . . , xp der p Predik-toren bzw. Deskriptoren in einer linearen Beziehung korreliert. Es wird also davonausgegangen, dass die Antwortvariable über eine bestimmte Funktion f(x) mit denDeskriptoren in Zusammenhang steht. Letztere werden hierzu mit den Koeffizien-ten β1, . . . , βp gewichtet (β0 repräsentiert den Ordinatenabschnitt). Eine auf einerGauß-Verteilung basierende, grundsätzlich vorhandene Abweichung (Mittelwert = 0)der tatsächlichen Werte von den Messwerten der Antwortvariablen wird zusätzlichmit dem Fehler ϵ berücksichtigt. Dieser schließt also unbekannte, durch das Modellprinzipiell nicht abdeckbare, zufällige Messabweichungen ein.f(x) = β0 +p�j=1xjβj + ϵ(1.50)Zur Ermittlung optimaler Koeffizienten während der Modellgenerierung wird übli-cherweise die Methode der kleinsten Quadrate (least squares) angewandt. Es wird dieRSS der Messpunkte yi und deren Vorhersagen f(xi) minimiert:38Kapitel 1. EinleitungRSS(β) =N�i=1(yi − f(xi))2(1.51)=N�i=1(yi − β0 −p�j=1xijβj)2(1.52)Dies geschieht durch Nullsetzen der ersten Ableitung δRSSδβ . Es wird eine Hyperebeneim IRp+1-dimensionalen Raum konstruiert, bei der die Distanzen der Messpunkte zudieser minimiert werden und deren orthogonale Projektionen die jeweiligen Vorher-sagen f(xi) darstellen [219, 220].1.3.3.2Logistische, binomiale RegressionÜber eine logistische, binomiale Regression kann die Wahrscheinlichkeit p(X) =Pr(Y = 1|X) modelliert werden, welche Auskunft darüber gibt, ob eine MessungY auf Basis der Eigenschaft X der Klasse 1 anstatt der Klasse 0 zugeordnet werdenkann. Die Eigenschaft X resultiert hierbei aus einem oder mehreren Deskriptoren.Um Werte von p(X) < 0 und p(X) > 1 zu vermeiden, wird im Gegensatz zur linearenRegression eine logistische Funktion genutzt, die ein sigmoidales Verhalten aufweist:p(X) =eβ0+β1X1 + eβ0+β1X(1.53)Für alle Werte von X wird eine Wahrscheinlichkeit der Klassenzugehörigkeit p(X)zwischen 0 und 1 erhalten. Es besteht hier ein linearer Zusammenhang von X mitder logarithmierten Chancep(X)1−p(X), das sogenannte Logit:log(p(X)1 − p(X)) = β0 + β1X(1.54)Daraus folgt, dass eine Änderung von X zu (X + 1) einer Multiplikation der Chancemit eβ1 entspricht. Die Änderung der eigentlichen Wahrscheinlichkeit p(X) bei einersolchen Zunahme von X hängt jedoch vom vorherigem Wert von X ab, da zwischenWahrscheinlichkeit und Deskriptor keine lineare Beziehung herrscht. Im Gegensatzzur Methode der kleinsten Quadrate bei einer multilinearen Regression wird zur Be-stimmung von β0 und β1 die Methode der maximalen Mutmaßlichkeit (Maximumlikelihood) angewandt. Es wird folgende Likelihood-Funktion optimiert:ℓ(β0, β1) =n�i=1p(xi)yi(1 − p(xi))1−yi(1.55)Auf Basis einzelner Bernoulli-Verteilungen resultiert diese Funktion für n Fälle mitden tatsächlichen Klassenzugehörigkeiten y = 1 oder y = 0 also aus dem Produktaller Wahrscheinlichkeiten p(x). Die Koeffizienten werden hierbei so gewählt, dass diemaximale Anzahl an Messpunkten des Datensatzes durch die sigmoidale Funktion indie korrekte Klasse eingeteilt werden [219, 221].1.3. QSPR-Modelle zur Stabilitätsvorhersage391.3.3.3Künstliches neuronales NetzBei einem künstlichen neuronalen Netz handelt es sich um eine nichtlineare Modell-art, die für Klassifikationen und Regressionen genutzt werden kann. Die Deskriptorenwerden, ähnlich der Vorstellung eines Netzes aus Nervenbahnen, über Neuronen (Li-nearkombinationen der Variablen) mit der Antwortvariablen f(x) korreliert. Letzteresteht in der Regel in einer linearen Beziehung zu den Werten der Neuronen h1, . . . , hH(mit den Gewichtungen γk und dem Ordinatenabschnitt γ0).f(x) = γ0 +H�k=1γk hk(1.56)Jedoch stehen diese H Neuronen, die zur Generierung der Vorhersagefunktion f(x)dienen, zumeist über nichtlineare Funktionen mit den ursprünglichen P Deskriptorenin Beziehung. Steht der Wert des k-ten Neurons hk bspw. in einer logistischen Be-ziehung zu den P Deskriptoren, so gilt für diesen (mit der Gewichtung βj des j-tenDeskriptors):hk(x) =11 + e−(β0+�Pi=1 xjβj)(1.57)Für jedes Neuron werden die Deskriptoren unterschiedlich stark gewichtet. Durch dienichtlineare Beziehung der Deskriptoren mit den Neuronen, einem drastischen Anstiegan veränderbaren Parametern schon bei einem geringen Anstieg der Neuronenanzahlund der Möglichkeit mehrerer hintereinandergeschalteter Neuronen ist diese Modell-art sehr anfällig für eine Überanpassung. Diese Modelle eignen sich daher in der Regelnur für sehr große Datensätze [211].1.3.3.4XGBoost-ModellDer Begriff XGBoost steht für Extreme gradient boosting und stellt eine noch re-lativ neue Modellart dar, die 2016 von Chen und Guestrin entwickelt wurde [222]und in den darauffolgenden Jahren an Popularität gewann, da hiermit viele MachineLearning-Wettbewerbe gewonnen wurden [223]. Zunehmend findet diese Modellartauch in der Chemie Anwendung [224, 225]. Es handelt sich um ein parallelisiertesVerfahren, welches zu einem Ensemble an nacheinander generierten Enscheidungs-bäumen führt. Dieses Ensemble wird über ein Gradientenverfahren zur Minimierungeiner Zielfunktion obj(t) genutzt. Beim überwachten Lernen (Supervised learning) be-steht letztere aus einer Verlustfunktion ℓ, die den Fehler zwischen Messwert yi undVorhersage ˆyi beschreibt, und einem Regularisierungsterm Ω, der die Modellkomple-xität berücksichtigt:obj(t) =n�i=1ℓ(yi, ˆy(t)i ) +t�i=1Ω(fi)(1.58)40Kapitel 1. EinleitungDie Art der Verlustfunktion (z. B. der RMSE-Wert) kann bei diesem Modelltyp freigewählt werden. Die Vorhersage ˆy(t)istellt den prädiktiven Wert nach t Trainingschrit-ten mit zusätzlichen Entscheidungsbaumgenerierungen dar. Der RegularisierungstermΩ bestraft komplexere Modelle und soll damit automatisch einer Überanpassung ent-gegenwirken (ähnlich der Regularisierung bei einer Ridge- oder Lasso-Regression):Ω(fi) = γT + 12λT�j=1w2j(1.59)T repräsentiert die Anzahl an endständigen Ästen im jeweiligen Baum, wj derenBepunktung, λ und γ sind vom Nutzer einstellbare Hyperparameter. Die Bepunktungder Äste hängt von der Anzahl an Messwerten ab, die diese jeweiligen Enden desBaums erreichen. Die Komplexität der Bäume wird ebenso durch ein Beschneiden(Pruning) beschränkt [226].1.3.4LimitierungenDie breite Anwendung von QSPR-Modellen in verschiedensten Forschungsbereichendurch Nicht-Statistiker birgt einige Gefahren. Eine falsche Interpretation der prä-diktiven Leistung solcher Modelle kann vielfältige Ursachen haben: eine zu hoheHeterogenität des Datensatzes, ein unvollständiger Datensatz, Nutzung nicht inter-pretierbarer oder kollinearer Deskriptoren, zu hohe Fehlerbereiche der verwendetenDeskriptoren, eine inadäquate Applikationsdomäne (Applicability domain, AD) undnicht zuletzt eine Überanpassung (Overfitting) an den Datensatz durch Verwendungvon Modelltypen mit vielen Freiheitsgraden oder zu vielen Deskriptoren in Relationzur Datensatzgröße [189].Das Problem, dass durch ein komplexes Modell im Prinzip aus jedem arbiträren Da-tensatz eine scheinbar vielversprechend wirkende Korrelation generiert werden kann,die jedoch keinen tatsächlich gegebenen Zusammenhang der Deskriptoren mit derGröße der Antwortvariablen begründen muss, wird durch folgendes Zitat von Ungerund Hansch treffend beschrieben:„One must rely heavily on statistics in formulating a quantitative model but, ateach critical step in constructing the model, one must set aside statistics andask questions. […] Without such a qualitative perspective, one is apt to generatestatistical unicorns, beasts that exist on paper but not in reality […] it has recentlybecome all too clear that one can correlate a set of dependent variables using ran-dom numbers as independent variables. Such correlations meet the usual criteriaof high significance“ [227].1.3. QSPR-Modelle zur Stabilitätsvorhersage41Aufgrund der geschilderten Problematiken wird empfohlen, sich bei der Modellge-nerierung an gewisse Richtlinien, z. B. jene von Tropsha [228], zu halten. Nicht zuletzthat auch die Organisation für wirtschaftliche Zusammenarbeit und Entwicklung (Or-ganisation for Economic Co-operation and Development, OECD) solche Richtlinienpubliziert:„To facilitate the consideration of a QSAR model for regulatory purposes, itshould be associated with (1) a defined end point, (2) an unambiguous algorithm,(3) a defined domain of applicability, (4) appropriate measures of goodness-of-fit,robustness, and predictivity, (5) a mechanistic interpretation, if possible“ [229].Auch aufgrund einer oftmals nicht gegebenen Publikation aller wesentlichen Modell-parameter streben wissenschaftliche Journale immer öfter gewisse Grundsätze fürQSPR-Modelle an, um eine Reproduzierbarkeit sicherzustellen [230]. Dennoch istdies ein schweres Unterfangen, da solche Modelle, wie bereits oben erwähnt, in ver-schiedensten Forschungsgebieten für unterschiedliche Problemstellungen Anwendungfinden.42Kapitel 1. Einleitung1.4Constrained-Network-Analyse1.4.1Theoretischer HintergrundEine sogenannte Constrained Network-Analyse (CNA) kann dazu dienen, die Stabili-tät von Strukturen auf Basis ihrer Rigidität zu beurteilen [231–239]. Diese basiert aufdem FIRST-Programm (Floppy inclusions and rigid substructure topography) und des-sen Pebble Game-Algorithmus zur Identifizierung rigider Cluster und flexibler Regio-nen in Biomakromolekülen. Die Atome des Moleküls werden hierbei zur Generierungeines Netzwerks als Körper mit zunächst jeweils sechs Freiheitsgraden gehandhabt.Durch jede kovalente und nichtkovalente Wechselwirkung (letzteres umfasst HB, Salz-brücken und hydrophobe Interaktionen) werden diese Körper bzw. Eckpunkte überStäbe (Bars) mit anderen Atomen des Moleküls im Netzwerk verbunden und somitin ihrer Flexibilität und ihren Freiheitsgraden eingeschränkt. Nach Generierung ei-nes solchen Body-and-bar-Netzwerks aus Körpern und Stäben werden in diesem rigideAreale determiniert, welche sich dadurch kennzeichnen, dass aufgrund der Vernetzungkeine relativen Bewegungen zwischen den Körpern ermöglicht wird [234].Das Prinzip basiert auf der Graph-Theorie [236] und findet seinen Ursprung in denArbeiten von Maxwell im 19. Jahrhundert [240]. Interne Freiheitsgrade F werdenauch als Floppy modes bezeichnet und deren Anzahl kann in einem d-dimensionalenRaum über folgende Beziehung beschrieben werden:F = dN − (Nc − Nr) − d(d + 1)2(1.60)Dabei stellen N die Anzahl an Eckpunkten im Netzwerk, Nc die Anzahl an Stä-ben bzw. Resktriktionen (Constraints) und Nr die redundanten Constraints dar. Alsredundant werden Resktriktionen aufgefasst, wenn sie nicht zur Stabilisierung desNetzwerks beitragen [241]. Zur Anwendung der Theorie auf Proteine werden Ein-fachbindungen mit jeweils fünf Stäben, nichtrotierbare Bindungen mit sechs Stäbenrepräsentiert. HB werden ebenso mit fünf Constraints, hydrophobe Wechselwirkungenmit zwei Constraints abgebildet. Letztere werden berücksichtigt, sofern die Distanzeines Kohlenstoff- oder Schwefelatoms zu einem anderen geringer ist als die Sum-me ihrer VdW-Radii, addiert mit einem spezifischen Cut-off-Wert (normalerweise beiD = 0.25 Å). HB werden in ihrer Stärke EHB über folgende Gleichung definiert:EHB = D0{5(R0R )12 − 6(R0R )10}f(θ, Φ, φ)(1.61)wobei D0 = 8 kcal/mol das Minimum des Interaktionspotentials darstellt, R0 = 2.8 Ådie Equilibriumsdistanz zwischen Donor und Akzeptor, sowie R die tatsächliche Di-stanz. Die Funktion f(θ, Φ, φ) hängt von der Hybridisierung der beteiligten Atome ab.Bspw. nimmt die Funktion die Form f = cos2(θ) cos2(Φ − 109.5) im Falle von sp3-hybridisierten Donoren und Akzeptoren an. Die Variablen θ, Φ und φ beschreiben1.4. Constrained-Network-Analyse43dabei die Winkel zwischen den beteiligten Atomen (Donor, Akzeptor und Wasser-stoffatom). Die Einschränkung der internen Freiheitsgrade wird schließlich über den3D-(6,6)-Pebble Game-Algorithmus definiert. Hierbei werden die Freiheitsgrade als6N −6 Steine (Pebbles) visualisiert und zunächst auf alle Eckpunkte gleichmäßig ver-teilt. Gemäß gewissen Regeln werden diese dann auf die bestehenden Stäbe zwischenden Körpern verteilt. Die Anzahl an verbleibenden Pebbles auf den Eckpunkten gibtschließlich Auskunft über die Rigidität des Systems und beschreibt die Freiheitsgerade[242–244].1.4.2Awendung des CNA-ProgrammsDas CNA-Programm erweitert die Funktionen des FIRST-Programms, indem es ther-male Entfaltungssimulationen auch auf Basis von Strukturensembles ermöglicht undwährend diesen auch die Verstärkung hydrophober Wechselwirkungen mit zunehmen-der Temperatur realisieren kann. Durch die Nutzung von Ensembles, die bspw. ausMD-Simulationen gewonnen werden, können die Rigiditätsanalysen der Entfaltungssi-mulationen deutlich verbessert werden. Bei diesen Simulationen wird die Stärke EHBvon HB im Netzwerk über die oben beschriebene Glg. 1.61 [236] bestimmt. Es han-delt sich um die Cut-off-Werte, bei denen diese aufgelöst werden (in vergleichbarerWeise können Salzbrücken zerstört werden). Über eine zuvor empirisch bestimmtelineare Beziehung (Tm = 300K − 20 EHB) werden diese Cut-off-Werte im CNA-Programm mit spezifischen Temperaturwerten gleichgesetzt. Die genannte Beziehungbasiert auf einem systematischen Vergleich der mit dem CNA-Programm ermitteltenEntfaltungstemperaturen mit experimentellen Daten zu 20 verschiedenen Proteinenvon mesophilen Bakterien und homologen thermophilen Varianten [231]. Das anfangsgenerierte Netzwerk wird somit steigenden Temperaturen ausgesetzt, indem der Cut-off-Wert der Energie sukzessive geändert wird und schwache HB nach und nach ausdem System entfernt werden. Nach jedem Temperatursprung werden das Netzwerkund rigide Cluster neu evaluiert. Das CNA-Programm ermöglicht im Anschluss eineBeurteilung der thermischen Stabilität auf Basis verschiedener globaler und lokalerRigiditätsparameter. Eine nähere Beschreibung ist in Anhang B.13 gegeben.Auch Wassermoleküle und Liganden können bei der CN-Analyse berücksichtigt wer-den (wobei das Ergebnis in der Regel durch eine Einbeziehung von Wassermolekülenauf der Proteinoberfläche nicht signifikant beeinflusst wird). Findet das Programmkeine Informationen zu Bindungsordnungen der Heteroatome in der PDB-Datenbank(basierend auf dem dreistelligen PDB-Code des Moleküls), so werden diese mithilfevon OpenBabel [245, 246] über eine distanzbasierte Konnektivität errechnet und dieHybridisierungen der Atome somit bestimmt [234]. Eine Visualisierung der gefun-denen Cluster sowie der nichtkovalenten Wechselwirkungen des Netzwerks und derbeschriebenen globalen und lokalen Parameter kann über das kostenlos zur Verfü-gung stehende PyMOL-Plugin visualCNA [246] erfolgen.44Kapitel 1. Einleitung1.5Untersuchte StrukturenIm Folgenden werden die für diese Arbeit relevanten Peptide und Proteine beschrie-ben. Dabei sollen zunächst die kleineren Modellproteine vorgestellt werden, derenKonjugate in dieser Arbeit zum Studium des Effekts von PEG auf die konformativeProteinstabilität, sowie zum Vergleich der alternativen Polymere LPG, MeOx undPEtOx dienten (Kap. 4). Danach wird das größere, therapeutisch bedeutsame Pro-tein Interferon-α2a (IFN-α2a) beschrieben, dessen PBK in Kap. 5 hinsichtlich favo-risierter Protein-Polymer-Wechselwirkungen, thermischer Stabilität und Bioaktivitätnäher untersucht wurden.1.5.1Modellproteine1.5.1.1Pin 1 WW-DomäneDie humane Peptidyl-prolyl-cis-trans-Isomerase bzw. Rotamase Pin 1 spielt eine wich-tige Rolle in der G2/M-Phase des Zellzyklus. In Eukaryoten ist dieses Protein hoch-konserviert. Es besteht aus einer N-terminalen, 39 Aminosäuren langen WW-Domäneund einer größeren C-terminalen Domäne. Während letztere für die Funktion des Pro-teins als Rotamase verantwortlich ist, dient die kleinere WW-Domäne zur Etablierungvon Protein-Protein-Interaktionen über prolinreiche Areale von Substraten (Erken-nung von PPXY-Sequenzen). Der Name resultiert aus den zwei Tryptophanen derDomäne (W11 und W34 im Falle der humanen Variante). Das Protein wurde 1997kristallisiert, die Struktur besitzt eine Auflösung von 1.35 Å (PDB: 1PIN) und istin Abb. 1.10-A dargestellt. Ein PEG-Pentamer findet sich hier zwischen den Ami-nosäuren Y23, S32 und W34 der WW-Domäne und N90 und K97 der C-terminalenDomäne. Im Bereich zwischen L22, C113 und M130 letzterer ist weiterhin ein Dipep-tid kristallisiert [247, 248].Die WW-Domäne besteht aus drei antiparallelen β-Faltblättern, die über zwei Schlei-fen (Reverse turns) miteinander verbunden sind (Abb. 1.10-B). Die Stabilität des Pro-teins wird maßgeblich durch zwei gegenüberliegende, hydrophobe Zentren bestimmt.Weiterhin sind eine Salzbrücke zwischen E12 und R14, sowie ein ausgedehntes HB-Netzwerk zwischen den Aminosäuren P9, E12, N26, H27, I28 und T29 als stabi-lisierende Elemente beschrieben [249]. Die Aminosäuren M1 bis E6 wurden in derKristallstruktur nicht aufgelöst und auch nicht von Price et al. in deren Studien be-rücksichtigt (siehe Kap. 1.1.6.2), sodass ein Konstrukt aus 34 Aminosäuren verwendetwurde.Aufgrund der geringen Größe und einer anzunehmenden, simplen Zwei-Phasen-Kinetikwährend der Entfaltung (was allerdings in einer Infrarotspektroskopiestudie infragegestellt wurde [250]) wurde die β-faltblattreiche WW-Domäne bereits in verschie-denen Mutations- und Faltungsstudien als Modellprotein verwendet. Die Entfaltungfindet vermutlich zunächst durch eine Loslösung des kleineren, dritten β-Faltblatts1.5. Untersuchte Strukturen45von der Tertiärstruktur statt, wohingegen eine Denaturierung des ersten β-Faltblattseine höhere Energiebarriere darstellt [247, 251–260].Abbildung 1.10: (A) Darstellung der Pin 1-Kristallstruktur 1PIN [247] mit unterschiedli-cher Färbung der beiden Domänen (blaue WW-Domäne: Aminosäuren 1 bis 39, orangene C-terminale Domäne: Aminosäuren 45 bis 163). (B) Illustration der WW-Domäne. HydrophobeZentren sind violett und orange gefärbt, am besagten HB-Netzwerk teilnehmende Aminosäu-ren grün.1.5.1.2Her2-AffibodyIm Gegensatz zur β-faltblattreichen Struktur der Pin 1 WW-Domäne stellt der Her2-Affibody ein α-helikales Peptid dar. Als antigenbindende Alternativen zu Antikör-pern leiten sich die wesentlich kleineren, 58 Aminosäuren langen Affibodies von derIgG-bindenden Domäne des Proteins A aus Staphylococcus aureus ab. Durch einigeModifikationen dieser B-Domäne wurde 1987 die stabilere Z-Domäne erhalten. Ran-domisierte Mutationen von 13 Aminosäuren der ersten zwei Helices brachten mittelseines Phagen-Displays anschließend Peptide hervor, die mit einer hohen Affinität z.B. an Insulin, TNF (Tumor-Nekrose-Faktor)-α oder auch den humanen, epiderma-len Wachstumsfaktor 2 (Human epidermal growth factor receptor 2, Her2) binden.Bei einer Überexpression dieses Faktors, v. a. im Falle von Mammakarzinomen, kannein solcher Affibody als ein wichtiges, diagnostisches Werkzeug eingesetzt werden.Es bindet an ein anderes Epitop als therapeutisch verwendete Antikörper wie bspw.Trastuzumab und bietet aufgrund seiner geringen Größe im Vergleich zu Antikör-pern einige Vorteile, bspw. eine verbesserte Löslichkeit. Da Her2-Affibodies von derZ-Domäne abstammen, werden sie auch ZHer2-Affibodies genannt [261].Für diese Studie diente eine NMR-Struktur der ZHer2:342-Variante (PDB: 2KZJ) alsAusgangspunkt (Abb. 1.11). Die bei der Strukturaufklärung erhaltenen Messdaten er-lauben ebenso eine alternative Struktur dieses Peptids mit einer nicht durchgängigen46Kapitel 1. Einleitungersten Helix (PDB: 2KZI). In Lösung kommt es möglicherweise zu einer Interkon-version beider Zustände, wobei in der Kristallstruktur nach Bindung an Her2 (PDB:3MZW) die durchweg helikale Konformation aufgelöst wurde [262].Abbildung 1.11: (A) Darstellung des Her2-Affibodys in der Kristallstruktur 3MZW, gebun-den an die extrazelluläre Domäne von Her2. Die Bindestelle ist in einem Ausschnitt unterhalbder Struktur vergrößert dargestellt. In (B) ist der Affibody ohne Her2 vergrößert dargestellt(NMR-Struktur 2KZJ). In dieser Arbeit untersuchte Konjugationsstellen sind in beiden Bil-dern türkis gefärbt.1.5.1.3Src SH3-DomäneWährend in dieser Studie primär Konjugate der Pin 1 WW-Domäne und des Her2-Affibodys untersucht werden sollten, wurde ein PBK der Src SH3-Domäne des Banki-vahuhns (Gallus gallus) von Price et al. als eine Validierungsmöglichkeit für seinekristallstrukturbasierte Vorhersagerichtlinie zur PEG-basierten Stabilisierung genutzt[100]. Es handelt sich um ein der Pin1 WW-Domäne strukturell sehr ähnliches Pro-tein, welches ebenso über viele β-Faltblattstrukturbereiche verfügt (Abb. 1.12). EineNMR-Struktur wurde 1993 veröffentlicht (PDB: 1SRL) [263]. Price et al. charakte-risierten eine T20N-Mutante und die entsprechend konjugierte Asn(PEG)4-Variante.Eine starke Stabilisierung mit einem ∆∆G0f-Wert von -1.20 ±0.10 kcal/mol wurdebestimmt [100]. Wie auch die WW-Domäne dient diese Struktur in der Zelle zurErkennung prolinreicher Elemente [247].1.5. Untersuchte Strukturen47Abbildung 1.12: Darstellung der Src SH3-NMR-Struktur mit der von Price et al. unter-suchten Mutationsstelle T20.1.5.1.4GCN4 und coil-VaLdIn ihren neuesten Publikationen [102, 264] charakterisierten Price et al. ebenfalls Kon-jugate der α-helikalen, Doppelwendeln (Coiled coils) bildenden Peptide GCN4 undcoil-VaLd (Abb. 1.13). Während es sich bei GCN4 um einen homodimeren Leucin-Zipper der Hefe (Saccharomyces cerevisiae) handelt, welcher als Transkriptionsfaktorfungiert (PDB: 2ZTA) [265], stellt das Trimer coil-VaLd ein artifizielles bzw. entwor-fenes Konstrukt dar, das von Price et al. auch als 1CW bezeichnet wurde (PDB: 1COI[266] und das strukturell ähnliche Trimer mit dem Code 5UXT [267]).Abbildung 1.13: Strukturen (A) GCN4, sowie von (B) coil-VaLd und (C) dem davonabgeleiteten Trimer mit dem PDB-Code 5UXT. Die von Price et al. untersuchten Konjugati-onsstellen sind grün gefärbt und jeweils für ein Monomer gekennzeichnet.48Kapitel 1. Einleitung1.5.2Biologika1.5.2.1Interferon-α2aDas Protein IFN-α2a bindet an die Interferon-α-Rezeptoren IFNAR1 und IFNAR2.Über eine dadurch induzierte Signalkaskade mit anschließender Gentranskription be-sitzt das therapeutisch verwendete Protein eine immunmodulierende und antiviraleWirkung. Aus diesem Grund wird es gegen unterschiedliche Tumoren, sowie Hepa-titis B und C eingesetzt (in letzterem Falle zumeist in Kombination mit Ribavirin).Die konventionelle interferonbasierte Therapie gegen Hepatitis C spielt in aktuellenLeitlinien im Gegensatz zu interferonfreien DAA-Therapien (Direct acting agents) mitverschiedenen Protease/NS5A/NS5B-Inhibitoren nur noch eine untergeordnete Rolle.Die unPEGylierte Variante (Handelsname: Roferon-A®) besitzt eine äußerst gerin-ge Halbwertszeit von lediglich etwa sechs bis neun Stunden, weshalb die PEGylierteVariante Peginterferon-α2a mit dem Handelsnamen Pegasys® (siehe Kap. 1.1.4.3)entwickelt wurde, welche deutliche pharmakokinetische Vorteile bietet [18, 268–270].Untersuchungen in Bezug auf Hepatozyten deuten allerdings darauf hin, dass einelängere Aktivierung des Jak/STAT-Signalwegs nicht den alleinigen Grund für eineverbesserte Therapie darstellen könnte. Durch ein verändertes pharmakodynamischesProfil könnten weitere immunmodulierende Gene aktiviert werden [271]. In Abb. 1.14-A ist die Kristallstruktur mit dem PDB-Code 3SE3 abgebildet, in der das Protein anseine Rezeptoren gebunden ist [272]. Die NMR-Struktur 1ITF des isolierten Proteinswurde für die in dieser Arbeit durchgeführten Simulationen genutzt (Abb. 1.14-B).Es besteht aus fünf α-helikalen Abschnitten mit insgesamt 165 Aminosäuren [273].Abbildung 1.14: (A) Kristallstruktur 3SE3 mit IFN-α2a und den beiden Rezeptoren IF-NAR1 und IFNAR2. Die beiden potentiellen, in Kap. 3.2.2 genannten KonjugationsstellenK31 und K134 sind gekennzeichnet. In (B) ist die isolierte IFN-α2a-Struktur aus der PDB-Datei 1ITF gezeigt.49Kapitel 2ZielsetzungDie in Kap. 1.1.6 geschilderten Hypothesen zu Stabilisierungsmechanismen von PBKverdeutlichen, dass das Verhalten von Konjugaten auf molekularer Ebene noch nichtgänzlich erforscht ist, erst recht nicht im Falle der selten genutzten PEG-Alternativen.Ein besseres Verständnis der Polymerdynamiken nach kovalenter Knüpfung an Pro-teine kann dazu beitragen, Veränderungen der pharmakokinetischen, pharmakody-namischen und thermodynamischen Eigenschaften von PBK in Relation zu ihrenunkonjugierten Varianten zu erklären. Übergeordnetes Ziel dieser Studie ist es, dieseDynamiken mithilfe von MD-Simulationen näher zu analysieren und einen rationa-len Ansatz zur Auswahl der geeignetsten Konjugationsstellen und Polymerarten zuentwickeln, ohne experimentell kostspielige und zeitaufwendige Screenings durchfüh-ren zu müssen. Speziell sollen die Polymere LPG, PMeOx und PEtOx als potenzielleAlternativen mit dem Goldstandard PEG verglichen werden.Hierfür sollen in dieser Arbeit in adäquater Zeit simulierbare Oligomer-Konjugatezweier kleinerer Modellproteine, der humanen Pin 1 WW-Domäne und des Her2-Affibodys, analysiert werden. Die bereits von Price et al. experimentell charakteri-sierten Varianten der WW-Domäne (siehe Kap. 1.1.6.2) eignen sich dabei als Aus-gangspunkt zum Studium der PEG-basierten Proteinstabilisierung. Das Verhaltender Konjugate in den Simulationen soll mit deren (De-)stabilisierungen in Zusam-menhang gebracht werden, bspw. mittels QSPR-Modellen. Experimentelle Daten zuHer2-Affibody-, LPG-, PMeOx- und PEtOx-Konjugaten stehen im Vorfeld nicht zurVerfügung, sodass hier lediglich Schlussfolgerungen auf Basis der in silico-Daten ge-zogen werden können. Zusätzliche Synthesen weiterer, von Price et al. noch nichtcharakterisierter Konjugate sind im Rahmen der Kooperation mit Prof. Dr. Dr. Lo-renz Meinel und Prof. Dr. Tessa Lühmann jedoch denkbar.Die Untersuchungen sollen schließlich für das therapeutisch relevante Protein IFN-α2afortgesetzt werden. Die Arbeiten der Verbundpartner in diesem vom Bundesministe-rium für Bildung und Forschung finanzierten Projekt mit dem Titel PEG-alternativePolymere für die Biokonjugation an Proteine (Next-PEG) fokussieren sich auf dieSynthese und Charakterisierung von Biokonjugaten dieses Therapeutikums (mit Po-lymerketten im Größenbereich zwischen 5 und 50 kDa). In der Arbeitsgruppe von50Kapitel 2. ZielsetzungProf. Dr. Rainer Haag sollen an der Freien Universität Berlin die Polymere PEG undLPG synthetisiert werden, im Arbeitskreis von Prof. Dr. Ulrich S. Schubert an derFriedrich Schiller-Universität in Jena die Polymere PMeOx und PEtOx.Im Arbeitskreis von Prof. Dr. Dr. Lorenz Meinel und Prof. Dr. Tessa Lühmann ander Julius-Maximilians-Universität Würzburg werden diese Polymere mittels Click-Chemie-Verfahren an Proteine konjugiert und hinsichtlich ihrer Größe, thermischenStabilität und Bioaktivität charakterisiert. Unter der Aufsicht von Dr. Christian Pfal-ler am Paul-Ehrlich-Institut sollen die Arbeiten durch in vivo-Studien der bereitge-stellten Konjugate weitergeführt werden.Simulationsstudien zu IFN-α2a-Konjugaten sollen in der hier vorliegenden Arbeitdurchgeführt werden, um die Ergebnisse der experimentellen Charakterisierungen derVerbundpartner zur thermischen Stabilität und Bioaktivität auf molekularer Ebenezu interpretieren.Zusammengefasst soll in dieser Arbeit mithilfe von Simulationsverfahren und chemo-informatischen Methoden die Dynamik unterschiedlicher PBK studiert werden. DerEffekt der Konjugation auf pharmazeutisch relevante Eigenschaften soll anhand derErgebnisse dieser Studien gedeutet und darauf aufbauend die Entwicklung optimierterKonjugate mit rationalen Ansätzen vorangetrieben werden. Übergeordnetes Ziel desVerbundprojekts ist die Charakterisierung und Etablierung von PEG-Alternativen fürdie Biokonjugation, wozu die vorliegende Dissertation mit Hilfe computergestützterModelle und Simulationen auf atomarer Ebene einen Beitrag leisten will.51Kapitel 3Materialien und Methoden3.1ModellproteineIm Folgenden sollen die simulierten Konjugate der Modellproteine aus Kap. 1.5.1.1vorgestellt (Kap. 3.1.2), sowie die Methoden zur Parametrisierung (Kap. 3.1.3), Mo-dellierung (Kap. 3.1.4) und Simulation (Kap. 3.1.5) dieser Strukturen beschriebenwerden. Im Anschluss werden die Vorgehensweisen zur Generierung von QSPR-Modellen(Kap. 3.1.6), zur Durchführung von Metadynamik-Simulationen (Kap. 3.1.7) und zumsystematischen Vergleich der Dynamik der Polymere (Kap. 3.1.8) geschildert. Die Er-gebnisse werden in Kap. 4 vorgestellt und diskutiert.3.1.1Wildtyp-Simulationen und unkonjugierte MutantenDie Wildtypen (WT) der Modellproteine aus Kap. 1.5.1.1 wurden mit den untenbeschriebenen Einstellungen für Konjugate simuliert (Kap. 3.1.4 und 3.1.5).Unkonjugierte Mutanten der in Kap. 3.1.2 beschriebenen Konjugate mit entspre-chender Linkerstruktur (zumeist also Asparagin) und ohne Polymer wurden analogsimuliert. Zur Messung des von Price et al. postulierten Winkels diente die Strukturohne Mutation der Konjugationsstelle (d. h. ohne Linker), was in den meisten Fäl-len dem Wildtyp entsprach. Für einige Varianten, die neben der Konjugationsstellezusätzliche Punktmutationen enthielten (z. B. 19(S16A)-1, vgl. Namen in Tab. C.2)wurden die entsprechenden Mutanten des Wildtyps zusätzlich simuliert.3.1.2Simulierte Konjugate3.1.2.1Konjugate mit publizierten StabilitätsdatenVon den in Kap. 1.5.1 vorgestellten Konstrukten wurden, mit Ausnahme der Konju-gate von 1CW mit Asn(PEG)45-Struktur aus [108], alle von Price et al. experimentellcharakterisierten Konjugate der Pin 1 WW-Domäne, der Src SH3-Domäne und derDoppelwendeln GCN4 und coil-VaLd simuliert (insgesamt 131 Varianten). Es handel-te sich hierbei zum größten Teil um PEG-Tetramere mit einer endständigen Meth-oxygruppe, welche kovalent an Seitenkettenamide von Asparaginen geknüpft wurden.52Kapitel 3. Materialien und MethodenEine komplette Übersicht aller in dieser Arbeit simulierten Konjugate der Modell-proteine geben die Tab. C.1 (chemische Formeln der konjugierten Aminosäuren) undC.2 (Auflistung aller PBK mit Sequenzen) in Anhang C. In Tab. C.2 wurden die Bio-konjugate nach einem einheitlichen Schema benannt. Diese Namen werden in dieserDissertation durchweg verwendet, wobei die Pin 1-Konjugate ohne PDB-Code undentsprechende unkonjugierte Varianten lediglich mit der Position der Konjugations-stelle (also ohne Kennzeichen für den Linker) bezeichnet werden. Die untersuchtenKonjugationsstellen für die Src SH3-Domäne und die Doppelwendeln GCN4 und coil-VaLd wurden bereits in den Abb. 1.12 und 1.13 gezeigt. Im Falle der Pin 1 WW-Domäne sind die analysierten Varianten in Abb. 3.1 illustriert.3.1.2.2Zusätzliche KonjugateZusätzlich zu den von Price et al. charakterisierten Varianten der Pin 1 WW-Domänewurden weitere PBK mit Asn(PEG)4-Struktur simuliert (Abb. 3.2). Diese könntenin zukünftigen Arbeiten analog synthetisiert und mittels temperaturabhängiger CD-Spektroskopie charakterisiert werden. Ebenso wurden zum Vergleich der alternati-ven Polymere LPG, PMeOx und PEtOx jeweils 13 Asn(LPG)5-, Asn(PMeOx)5- undAsn(PEtOx)5-Strukturen modelliert (siehe Kennzeichen ϵ, λ und ψ in Tab. C.1). EineAnzahl von fünf Monomeren wurde hierbei ausgewählt, da zu diesem Zeitpunkt derArbeit zur Aussicht stand, Konjugate mit einer solchen Struktur im Rahmen der Ko-operation mit den Verbundpartnern analog zu vermessen. Hierzu sollten experimen-telle Daten von niedermolekularen Polyoxazolingemischen mit einem Polymerisations-grad von vier bis sechs Monomeren gewonnen werden. Die Länge der LPG-Variantewurde dementsprechend analog auf fünf Monomere gesetzt.Da bisher keine PBK des Her2-Affibodys experimentell charakterisiert wurden, wur-den in diesem Falle 15 verschiedene Konjugationsstellen ausgewählt und ebenso mit(PEG)4-, (LPG)5-, (PMeOx)5- und (PEtOx)5-Oligomeren modelliert. Hierbei wur-de jedoch eine Cys-Maleimid-Struktur als Linker verwendet (siehe Kennzeichen 4,φ, γ und ω in Tab. C.1), da von den Verbundpartnern geplant war, diese Variantenebenfalls zu charakterisieren. Diese sollten nicht, wie im Falle der WW-Konjugatevon Price et al., mittels Einbau von PEGylierten Aminosäuren über eine Festphasen-peptidsynthese synthetisiert werden, sondern mithilfe eines Click-Chemie-Verfahrensnach erfolgter Expression des Affibodys mit entsprechender Cystein-Mutation. DieStellen sind über das gesamte Peptid verteilt, jedoch nicht im unmittelbaren Bereichder Binderegion zu Her2 (vergleiche Vergrößerung in Abb. 1.11-A). Ein weiteres Aus-wahlkriterium hierbei waren die zwei von Price et al. bereits publizierten Richtlinien,beschrieben in Kap. 1.1.6.2. Lysine an den Positionen i + 3 und i + 4 sollten zu einerStabilisierung gegenüber der Cystein-Mutante führen [108], weshalb die AminosäurenA46 und D53 ausgewählt wurden. Zusätzlich wurden Konjugationsstellen mit relativkleinen oder großen Werten für den Winkel θPrice bei der Auswahl priorisiert. PBK3.1. Modellproteine53Abbildung 3.1: Untersuchte Konjugationsstellen der Pin 1 WW-Domäne mit bereits publi-zierten Stabilitätsdaten. Diese sind in der Struktur rechts benannt und türkis gefärbt. Diedazu aufgelisteten, farblichen Punkte symbolisieren die an der jeweiligen Position simuliertenVarianten, deren Strukturen links in den selben Farben dargestellt sind. Die jeweils unkon-jugierte Aminosäure ist ebenfalls illustriert. Zusätzlich zu den hier gezeigten Varianten desWildtyps wurden eine Vielzahl an Konjugaten mit zusätzlichen Punktmutationen untersucht(siehe hierzu Tab. C.2).54Kapitel 3. Materialien und Methodenohne publizierte Stabilitätsdaten wurden in Tab. C.2 an entsprechender Stelle mitder Abkürzung NA (Not available) gekennzeichnet.Abbildung 3.2: Zusätzlich simulierte, nicht von Price et al. charakterisierte Konjugations-stellen der WW-Domäne.3.1.3Ermittlung geeigneter Simulationsparameter3.1.3.1PyRED-MethodeZur Durchführung klassischer MD-Simulationen wurden für proteinogene Amino-säuren die Parameter des Amberff14SB-Kraftfeldes [133] genutzt. Für die PBK je-doch mussten zusätzliche Elemente definiert werden, wobei zur Modellierung MOE2018.01, 2019.01 und 2020.01 (Molecular Operating Environment) [274] verwendetwurde. Hierzu wurde, mit Ausnahme der Asn(PEG)1-, Asn(PEG)2-, Asn(PEG)3-,Asn(PEG)8- und Asn(PEG)45-Strukturen (siehe hierfür Kap. 3.1.3.3), die konjugier-te Aminosäure mitsamt der (PEG)4-Kette als eine einzelne Aminosäure beschrieben.Da es sich um ein an das Protein kovalent gebundenes Strukturelement handelte, wur-de im Falle der Asn(PEG)4-Struktur (entspricht Kennzeichen 1 in Tab. C.1) zunächstder PyRED-Server verwendet [275–278]. Dieser eignet sich aufgrund der möglichenFestsetzung von Ladungen und anschließender Entfernung bestimmter Atome beson-ders, um Fragmente mit offenen Valenzen für größere Strukturen zu definieren (z.B. Zucker in glykosylierten Proteinen, Monomere in Polymeren, nichtproteinogeneAminosäuren in Peptiden). Die konjugierte Aminosäure wurde hierzu mit amino-säurerückgratähnlichen Endgruppen modelliert: eine Acetylgruppe am N-Terminus,sowie eine Aminomethylgruppe am C-Terminus (Abb. 3.3). Die entsprechende PDB-Struktur wurde dann auf Basis des Amber14:EHT-Kraftfelds [133, 279] energiemini-miert (Gradient: 0.001 kcal/mol/Å) und mitsamt den nötigen Einstellungen auf denServer hochgeladen: Festsetzung der Gesamtladung der Termini auf Null (mithilfe derINTRA-MCC1-Einstellung), Zuweisung von Amber14ffSB-Atomtypen und Nutzungvon Gaussian 16 Rev. B.01 [280] zur Bestimmung von RESP-Ladungen [143]. Die3.1. Modellproteine55vom Server erhaltenen Kraftfelddateien geben an, in welchen Fällen von den im Am-berff14SB-Kraftfeld festgesetzten Parametern abweichende Werte festgelegt wurden,wobei sich diese zumeist minimal von den Standardwerten unterschieden (Tab. 3.1).Abbildung 3.3: Beschreibung des PyRED-Verfahrens am Beispiel von Asn(PEG)4. Die Ge-samtladung der terminalen Acetyl- und Aminomethylgruppen wurde auf Null festgesetzt.Nach Entfernung wurden geeignete RESP-Ladungen für das mittlere Fragment, mit einerNettoladung von Null, erhalten.Tabelle 3.1: Darstellung eines Auszugs aus der von PyRED generierten Parameterdatei fürdie Asn(PEG)4-Struktur. Abweichungen von den Amber14ffSB-Standardparametern sind inder rechten Spalte angegeben (rot markiert).Bindungk [kcal/mol/Å2]Distanz [Å]UrsprungC-CT3151.522adapted from parm10.dat 317.0C-N4901.335taken from parm10.datC-O5701.229taken from parm10.dat3.1.3.2antechamber-MethodeDa der PyRED-Server nach erfolgter Definition der Asn(PEG)4-Struktur für fast zweiJahre aufgrund von Wartungsarbeiten nicht erreichbar war, wurde für alle weiterenVarianten ein ähnlicher Ansatz lokal durchgeführt. Hierzu wurden RESP-Ladungenfür die vorbereitete PDB-Struktur der konjugierten Aminosäure mittels Gaussian 09Rev. C.01 [281] ermittelt und geeignete Amberff14SB-Atomtypen über antechambervon AmberTools18 [282] zugewiesen. Im Anschluss wurde die Struktur in xleap gela-den und die Endgruppen wurden entfernt. Die Ladungen des Proteinrückgrats wurdendann den üblichen Werten im Amber14ffSB-Kraftfeld angeglichen. Diese unterschie-den sich zumeist nur minimal von den lokal mit Gaussian 09 Rev. C.01 erhaltenen56Kapitel 3. Materialien und MethodenWerten. Die Gesamtladung wurde anschließend berechnet und ein Überschuss an po-sitiver oder negativer Ladung wurde schließlich gleichmäßig auf die restlichen Atomeder konjugierten Aminosäure verteilt. Da dieser Überschuss insgesamt in der Regelin einem Bereich von -0.2 bis +0.2 lag, wurden die zugewiesenen Ladungen einzelnerAtome durch diesen Vorgang nur in einem äußerst geringen Maße modifiziert undnicht in ihrem Vorzeichen verändert (Abb. 3.4). Während hierbei lokal durch an-techamber wie beim PyRED-Ansatz zunächst Amberff14SB-Atomtypen zugewiesenwurden, wurden bei dieser Methode jedoch anschließend mit diesem Kraftfeld kompa-tible GAFF2-Parameter [147] hinzugezogen, was in den erhaltenen Dateien einsehbarwar (Tab. 3.2). Sogenannte Strafpunkte (Penalty scores) gaben dabei an, wie stark dieermittelten Parameter auf Extrapolationen bekannter Kraftfeldwerte beruhten [283].Abbildung 3.4: Illustration der alternativ durchgeführten, lokalen antechamber-Methodezur Parametrisierung der konjugierten Aminosäuren, am Beispiel einer Cys-Maleimid-(PEG)4-Struktur (Kennzeichen 4 in Tab. C.1). Nach Entfernung der Endgruppen und Zuwei-sung der üblichen Ladungen des Proteinrückgrats wurde die Nettoladung berechnet und derÜberschuss möglichst gleichmäßig auf die übrigen Atome verteilt.Tabelle 3.2: Darstellung eines Auszugs aus der von antechamber generierten Parameterdateifür eine Cys-Maleimid-(PEG)4-Struktur (Kennzeichen 4 aus Tab. C.1). Die zugrundeliegendenGAFF2-Parameter werden in der rechten Spalte angezeigt, zusammen mit entsprechendenStrafpunkten.Bindungk [kcal/mol/Å2]Distanz [Å]UrsprungC-CT313.001.524same as c-c3, penalty score= 0.0C-N427.601.379same as c- n, penalty score= 0.0C-O637.701.218same as c- o, penalty score= 0.0Auch wenn für manche Parameter je nach verwendeter Methode leicht unterschiedli-che Werte erhalten wurden, so ergab ein Vergleich von Simulationen der Asn(PEG)4-Struktur mit den von PyRED zugewiesenen Ladungen und Atomtypen und jenen mit3.1. Modellproteine57diesem lokalen antechamber-Ansatz keine Unterschiede im Interaktionsprofil des Poly-mers, weshalb diese alternative Methode für alle weiteren konjugierten und nichtprote-inogenen Aminosäuren verwendet wurde, solange der PyRED-Server nicht zur Verfü-gung stand. Später konnten die Varianten Asn(LPG)5, Asn(PEtOx)5, Asn(PMeOx)5,Cys-Maleimid-(LPG)5, Cys-Maleimid-(PEtOx)5 und Cys-Maleimid-(PMeOx)5 wie-der analog zu Asn(PEG)4 über den PyRED-Server parametrisiert werden.3.1.3.3Monomerbasierte ParametrisierungDer Fokus dieser Arbeit lag zunächst auf den von Price et al. studierten PEG-Tetrameren, weshalb die Oligomere, wie bei Price et al. auch [100], als einzelneFragmente parametrisiert wurden. Da jedoch im Zuge der Erweiterung des Daten-satzes an Modellprotein-Konjugaten auch Asn(PEG)X-Strukturen mit variablen Ket-tenlängen simuliert werden sollten, wurde für die Varianten Asn(PEG)1, Asn(PEG)2,Asn(PEG)3, Asn(PEG)8, Asn(PEG)45 und alle größeren PEG-Konjugate aus Kap.5 in analoger Weise zum lokalen antechamber-Ansatz ein PEG-Monomer (FragmentCH2-CH2-O, Abb. 3.5) als vielseitig einsetzbares Strukturelement parametrisiert. Ent-sprechend wurden hierbei zusätzliche PEG-Monomere als Endgruppen auf beidenSeiten während der Ladungsbestimmung genutzt. Zur Parametrisierung einer ter-minalen PEG-Methoxygruppe (Fragment CH2-CH2-O-CH3) wurde analog mit einerEndgruppe auf einer einzelnen Seite verfahren.Abbildung 3.5: Darstellung der zur Parametrisierung eines einzelnen PEG-Monomers ver-wendeten Struktur. Die endständigen Monomere außerhalb des markierten Bereichs wurdenentfernt und überschüssige Ladungen analog zum beschriebenen antechamber-Ansatz verteilt.3.1.3.4ProblemstellungenFür die von Price et al. charakterisierten Cys-Maleimid-Strukturen sind in den Re-ferenzpublikationen keine stereochemischen Informationen angegeben, sodass davonausgegangen werden kann, dass hierbei experimentell ein Racemat untersucht wordenist. Für die Simulationen wurde hierfür jeweils die (S)-Konfiguration modelliert.Ein ähnliches Problem stellen die LPG-Strukturelemente in dieser Arbeit dar. Durchdie Hydroxymethylseitenketten ergeben sich Stereozentren. Da die Taktizität diesesPolymers während der Synthese zumeist nicht kontrolliert wird [70], wurden in die-sen Fällen alternierende (S)- und (R)-Konfigurationen modelliert (als Baustein dergrößeren Konjugate wurde ein Dimer mit einem (S)- und einem (R)-Monomer para-metrisiert).58Kapitel 3. Materialien und MethodenSchließlich stellte auch die unkonjugierte Aminosäure Azidohomoalanin ein Problembei der Atomtypenzuweisung dar (siehe Kennzeichen § aus Tab. C.1). Der PyRED-Server konnte den endständigen Stickstoffatomen keinen Typ zuweisen (diese wurdenvom Server mit N? gekennzeichnet), den endständigen Atomen der Azidgruppe wur-de durch den lokalen Ansatz mittels antechamber ebenso der Atomtyp DU (Dummyatom) zugewiesen. Alternative Parameter aus älteren Publikationen auf Basis der Ar-beiten zum Arzneistoff Azidothymidin, in denen neue Atomtypen definiert worden wa-ren (um eine möglichst realisitische Mischung aus sp- und sp2-Hybridisierungsanteilender Stickstoffatome wiederzugeben), wurden im Anschluss getestet, führten jedochdurchweg zu instabilen Simulationen und wurden daher wieder verworfen [284, 285].Schließlich wurde den terminalen Stickstoffatomen der GAFF2-Atomtyp n1 zugewie-sen. Dies entspricht einer sp-Hybridisierung, wodurch die reale Verteilung der Elek-tronen möglicherweise nicht ideal abgebildet wird. Eine zu erwartende annäherndlineare Struktur der Azidgruppe in den Simulationen wurde dadurch allerdings bei-behalten. Da die Parameter des Azidohomoalanins jedoch lediglich für einige wenigeunkonjugierte Varianten des gesamten Datensatzes genutzt wurden, kann mit hoherWahrscheinlichkeit ein entscheidender Einfluss der eventuell unzulänglichen Hybridi-sierung dieser endständigen Gruppe auf die berechneten Deskriptoren ausgeschlossenwerden.3.1.3.5Vergleich der Simulationen mit ReferenzPrice et al. untersuchten in einer ihrer Studien [100] die Interaktionsprofile zehnverschiedener simulierter WW-Konjugate, sowie die RDF-Werte (Radial distributionfunction, radiale Verteilungsfunktion) des Solvens um bestimmte Seitenkettenatomefür vier WW-Konjugate. Daher wurden zum Vergleich analoge Analysen mit denin dieser Arbeit durchgeführten Simulationen der relevanten Asn(PEG)4-Konjugatevorgenommen (Abb. 3.6, 3.7, 3.8 und 3.9).Die Interaktionspartner der Polymere entsprechen in den meisten Fällen qualitativjenen aus der Referenzpublikation; Unterschiede sind im Falle der Konjugate 27-1 und29-1 zu erkennen. Diese Differenzen können einerseits darauf zurückzuführen sein, dassnicht bekannt ist, mit welchem Analysenprogramm die Wechselwirkungsenergien inder Referenzpublikation berechnet wurden (in der hier vorliegenden Arbeit wurde derlie-Befehl aus cpptraj [282] genutzt, welcher das Solvens lediglich implizit berücksich-tigt). Vor allem aber nutzten Price et al. Mulliken-Ladungen [286] in Zusammenhangmit dem älteren Amberff99SB-ILDN-Kraftfeld [153, 287], was im Vergleich zur Nut-zung von RESP-Ladungen [275] mit dem neueren Amberff14SB-Kraftfeld [133] mithoher Wahrscheinlichkeit zu einer abweichenden Polymerdynamik führt.Die RDF-Werte stimmen mit den Ergebnissen von Price et al. überein. Dies verdeut-licht, dass die generierten Parameter geeignet sind, um die von Price et al. beschrie-benen Desolvatationsvorgänge um spezifische Aminosäuren herum wiederzugeben.3.1. Modellproteine59Abbildung 3.6: Vergleich der Interaktionsprofile von Price et al. Die Namen entsprechenjenen aus der Originalpublikation (vgl. Tab. C.2). Auf der linken Seite (Nachdruck mit Ge-nehmigung aus [100], Copyright 2014 American Chemical Society) sind die Ergebnisse vonPrice et al. dargestellt, auf der rechten Seite analog jene aus dieser Arbeit.60Kapitel 3. Materialien und MethodenAbbildung 3.7: Vergleich der Interaktionsprofile von Price et al. (Fortführung zu Abb.3.6).Die Namen entsprechen jenen aus der Originalpublikation (vgl. Tab. C.2). Auf der linkenSeite (Nachdruck mit Genehmigung aus [100], Copyright 2014 American Chemical Society)sind die Ergebnisse von Price et al. dargestellt, auf der rechten Seite analog jene aus dieserArbeit.3.1. Modellproteine61Abbildung 3.8: Vergleich der Interaktionsprofile von Price et al. (Fortführung zu Abb.3.6).Die Namen entsprechen jenen aus der Originalpublikation (vgl. Tab. C.2). Auf der linkenSeite (Nachdruck mit Genehmigung aus [100], Copyright 2014 American Chemical Society)sind die Ergebnisse von Price et al. dargestellt, auf der rechten Seite analog jene aus dieserArbeit.62Kapitel 3. Materialien und MethodenAbbildung 3.9: Vergleich der RDF-Profile um bestimmte Atome von Aminosäureseiten-ketten (OH für Tyr, Cz für Phe, Nϵ1 für Trp, Oγ1 für Thr). Auf der linken Seite sind dieErgebnisse von Price et al. (Nachdruck mit Genehmigung aus [100], Copyright 2014 Ame-rican Chemical Society) dargestellt. Dabei zeigen die schwarzen Linien die RDF-Werte derunkonjugierten, die roten jene der konjugierten Varianten. Die Namen entsprechen jenen ausder Originalpublikation, WW steht hierbei für den Wildtyp. Auf der rechten Seite sind dieErgebnisse aus den in dieser Arbeit durchgeführten Simulationen analog dargestellt.3.1. Modellproteine633.1.4Modellierung der Strukturen3.1.4.1Aufbereitung der PeptidstrukturenDie verwendeten Peptide wurden im Falle von Kristallstrukturen in MOE bei einempH-Wert von 7 protoniert (Pin 1 WW-Domäne 1PIN, sowie die Doppelwendeln mitden PDB-Strukturen 2ZTA, 1COI und 5UXT), nachdem das enthaltene Kristallwas-ser und sonstige Liganden entfernt worden waren. In die Pin 1 WW-Domäne (Ami-nosäuren K6 bis G39) wurde vor der Protonierung eine W34F-Mutation eingefügtund das entsprechende Phenylalanin lokal energiminimiert. Diese Mutation wurde inallen von Price et al. charakterisierten Varianten dieses Peptids verwendet [99], da dieTryptophan-Variante, so die Erklärung von Price auf Anfrage, „eigenartige“ Signalewährend der DSF-Mesungen verursachte.1 Es wurde durch MOE keine Protonierungder Seitenkette von K6 durchgeführt, diese fungierte hierbei über ihr Stickstoffatomals HB-Akzeptor für die Hydroxylgruppe von S38. Da es sich jedoch um die Terminihandelt, welche in ersten Simulationen ein sehr flexibles Verhalten zeigten und keinedauerhafte Koordination von K6 zu S38 zu beobachten war, wurde diese Seitenkettenachträglich manuell protoniert.Der Protonierungszustand der verwendeten NMR-Strukturen (Her2 Affibody 2KZJund Src SH3-Domäne 1SRL) wurde beibehalten, da eine Protonierung in MOE beipH 7 (2KZJ wurde bei pH 6 aufgenommen) zu keinen Veränderungen führte. DieTermini aller Strukturen wurden bei der WW-Domäne, dem Her2-Affibody und derSrc SH3-Domäne, analog zu Price et al., nicht mit zusätzlichen Endgruppen verse-hen, sondern mit geladenem Zustand modelliert. Im Falle der Doppelwendel aus derKristallstruktur 1COI wurden die Termini beibehalten: Es handelt sich um eine C-terminale Acetylgruppe und ein N-terminales Amid. Um Sequenzen zu modellieren,die den von Price et al. charakterisierten Varianten entsprechen, mussten im Falleder Kristallstrukturen 2ZTA und 5UXT, die lediglich eine Acetylgruppe enthielten,die N-terminalen Amide nachträglich eingefügt werden. Ebenso mussten dazu in der1COI-Struktur die C-terminalen Aminosäuren H28 und G29 entfernt und durch einlokal energieminimiertes Tyrosin ersetzt werden. Möglicherweise wurde dieses Tyrosinvon Price et al. zur Durchführung von DSF-Messungen eingebaut.3.1.4.2Einbau modifizierter AminosäurenDie konjugierten Aminosäuren wurden mittels MOE an den jeweiligen Mutationsstel-len eingefügt. Diese wurde lokal mit dem Amber14:EHT-Kraftfeld energieminimiert.1Erklärung von Prof. Joshua L. Price in einer E-Mail-Korrespondenz: „Years ago, Prof. Jeff Kelly’slab at Scripps observed some unusual behavior in fluorescence-monitored temperature jump kineticexperiments on the WW variant with Trp at position 34, possibly due to the different environmentsof the two Trp residues during the unfolding process. In any case, they found that replacing Trp34with Phe eliminated this unusual behavior and resulted in a variant that could be readily analyzed viakinetic and thermodynamic experiments based on a two-state folding model.“64Kapitel 3. Materialien und MethodenAnschließend wurde zur Ermittlung einer geeigneten Startstruktur des Polymers je-weils eine stochastische Konformationssuche durchgeführt, wobei die restlichen Ami-nosäuren fixiert wurden. Es wurden folgende Einstellungen verwendet und die sta-bilste Konformation nach erfolgter Suche gespeichert: Ablehnungslimit = 100, Itera-tionslimit = 2000, Gradient = 0.001 kcal/mol/Å und Konformationslimit = 1000. ImFalle der unPEGylierten Varianten wurde die betreffende Aminosäure mutiert undlokal energieminimiert.3.1.4.3Solvatation der StrukturenZur Simulation wurden die Peptide mit dem Programm sander aus dem AMBER18-Programmpaket [282] initial energieminimert und anschließend mit TIP3P-Wasser[288] durch tleap von AmberTools18 [282] solvatisiert. Um die Anzahl an Wasser-molekülen bei diesem Schritt möglichst gering zu halten, wurden die Hauptachsender Strukturen zuvor mit den Achsen des Koordinatensystems aligniert, sodass dieProteine in der Simulationsbox optimal ausgerichtet waren. Es wurde jeweils einerechteckige Solvensbox mit einer minimalen Distanz von 10 Å der Peptidatome zuden Grenzflächen der Box generiert. Im Falle der länglichen Doppelwendeln wurdenjedoch kubische Boxen zur Solvatation genutzt, da diese Peptide nicht als annäherndglobulär angesehen werden können und sie dadurch während der Simulationen undnach Alignierung der Hauptachsen über Rotationen evtl. mit den periodischen Nach-barn wechselwirken konnten. Natrium- oder Chloridionen wurden nach erfolgter Sol-vatation zum Ladungsausgleich des Systems hinzugefügt.3.1.5Simulationsprotokoll3.1.5.1Equilibrierung und erster ProduktionslaufDie solvatisierten Peptide wurden im Anschluss mit NAMD 2.12 oder 2.13 simuliert[289]. Ein Zeitschritt von 2 fs wurde gewählt. Es wurden periodische Randbedingun-gen unter Verwendung des RATTLE-Algorithmus [138] genutzt. Zur Abbildung vonInteraktionen auf längere Distanz wurde die Particle-Mesh-Ewald-Methode [140] miteinem Cut-off-Wert von 1.2 nm angewandt [140]. Zunächst wurden 10000 Minimie-rungsschritte durchgeführt, bevor das System in einer ersten Equilibrierung von 100auf 300 K über ein Zeitintervall von 500 ps langsam aufgeheizt wurde. Eine Kraftkon-stante von 0.5 kcal/mol/Å2 wurde verwendet, um die Peptidatome zunächst in ihrerBeweglichkeit zu begrenzen. Auf diese Weise konnten sich die durch tleap artifiziellgesetzten Wassermoleküle zuerst reorganisieren. Die Kraftkonstante wurde über wei-tere 1.6 ns sukzessive bis zum Wert Null verkleinert. Eine konstante Temperatur von300 K im NVT-Ensemble wurde durch Nutzung eines Langevin-Thermostats erreicht.Nach vollständiger Annullierung der Kraftkonstante wurde für eine weitere Nanose-kunde in diesem Ensemble simuliert, bevor eine zweite, 1 ns lange Equilibrierungim NPT-Ensemble durchgeführt wurde, um die Dichte des Systems zu equilibrieren.3.1. Modellproteine65Dieses Ensemble wurde durch Festlegung des Drucks mit dem Nosé-Hoover-Langevin-Piston-Barostat auf 1.01325 bar erreicht. Nach dieser Equilibrierung wurde ein ersterProduktionslauf im NPT-Ensemble für 100 ns vorgenommen, wobei Momentaufnah-men für jede Pikosekunde herausgeschrieben wurden.3.1.5.2Fortführung des ProduktionslaufsDer oben beschriebene Produktionslauf wurde für weitere 100 ns fortgeführt, sofernbestimmte Konvergenzkriterien nicht erfüllt wurden. Hierzu wurden die drei in Kap.1.2.3.1 beschriebenen SCC angewandt. Folgende Limitierungen wurden gesetzt: Ge-mäß dem SDVG-Kriterium durfte kein neuer Cluster in den letzten 25 ns der Simulati-on gefunden werden. Hierzu diente eine hierarchische RMSD-Clusteranalyse (Average-Linkage) mittels cpptraj [282]. Mit ϵ = 1 wurde festgelegt, dass das Clustering bis zueiner Minimaldistanz von 1 Å zwischen den gefundenen Clustern durchgeführt wur-de. Die Clusterkonfigurationsentropie durfte sich in den letzten 25 ns höchstens um5 % ändern. Schließlich musste das BBCOM/BCOM-Verhältnis, welches idealerweiseeinen Wert von 1 erreichen sollte, bei ansteigenden Blocklängen bis zur Hälfte derSimulationszeit einen Wert von 1.20 erreichen. Dieses Verhältnis wurde mithilfe desLOOS-Pakets berechnet [290, 291]. Die drei beschriebenen SCC wurden einerseits fürdie Cα-Atome der Peptide und in einem zweiten Durchlauf für die Sauerstoffatomeder Polymerketten evaluiert. Exemplarische Plots zur Veranschaulichung aller dreiKriterien sind in Abb. 3.10 gezeigt. Zwei der drei gewählten Kriterien mussten erfülltwerden (sowohl für die Cα- als auch für die Polymer-Sauerstoffatome), ansonsten wur-de der Produktionslauf für 100 ns fortgeführt und alle Kriterien wurden erneut unter-sucht. Waren die Bedingungen dann immer noch nicht erfüllt, wurden wiederum 100ns simuliert. Im Falle der unPEGylierten Varianten dienten lediglich die Cα-Atome alsTest. Letztendlich wurden die meisten WW-Konjugate für 200 oder 300 ns simuliert.Die größeren Her2-Konjugate benötigten deutlich mehr Zeit (zumeist 600 bis 800 ns),um die geforderten Kriterien zu erfüllen. Asn(PEG)45-Varianten wurden jeweils für800 ns simuliert, zeigten aber auch danach keine ausreichende Konvergenz bzgl. derPolymer-Sauerstoffatome. Diese wurden daher für die QSPR-Modellgenerierungen alsnicht konvergiert eingestuft (siehe Kap. 3.1.6.2 und 3.1.6.3).Abbildung 3.10: Exemplarische Darstellung des Verlaufs der drei SCC-Kriterien einer Si-mulation. Die violette Zahl im SDVG-Plot (links) gibt die Zeit an, ab der kein weiterer Clustergefunden wurde. Im CCE-Plot gibt die Zahl die Veränderung des Werts in den letzten 25 nsan. Im rechten BBCOM/BCOM-Plot stellt die Zahl das Verhältnis bei der größten Blocklängedar. Letztere erreicht als Maximum die Hälfte der Simulationszeit, da ansonsten nicht mehrals ein einziger Block generiert werden kann.66Kapitel 3. Materialien und Methoden3.1.6QSPR-VorhersagemodelleIm Verlauf der Arbeit wurde der Datensatz an Simulationen und daraus berech-neter Deskriptoren sukzessive erweitert. Ein erster erweiterter Modellierungsansatzwurde auf Basis des R-Pakets caret (Classification And REgression Training [211,292]) entwickelt. Dieses R-Paket von Max Kuhn (Entwickler von RStudio, ehemali-ger Biostatistker bei Pfizer [293]) vereint verschiedenste Modellierungspakete aus derData Science-Community von R mit einer einheitlichen Syntax und zusätzlichen Mög-lichkeiten zur Datenaufbereitung und Evaluation. Es gehört neben scikit-learn undTensorFlow zu einem der beliebtesten Programme für Prozesse des maschinellen Ler-nens. In diesem Ansatz, der in Kap. 3.1.6.2 im Detail beschrieben werden soll, wurdeder Fokus v. a. darauf gelegt, eine Vielzahl der in diesem Paket verfügbaren Mo-dellarten des maschinellen Lernens zu testen. Da die simulierten Konjugate jedocheinen für Machine Learning-Verfahren relativ kleinen Datensatz darstellten, wurdeim Anschluss ein hiervon abgewandelter zweiter Ansatz entwickelt (Kap. 3.1.6.3), dersich auf die Generierung von multilinearen Regressions- und binomialen, logistischenKlassifikationsmodellen beschränkte.3.1.6.1DeskriptorenDie Trajektorien der untersuchten Modellproteine wurden genutzt, um daraus eineVielzahl an Deskriptoren zu berechnen. Dies wurde für die PBK und deren unkonju-gierten Varianten durchgeführt (im Falle des Winkels von Price et al. ebenso für diemutierten Wildtyp-Varianten ohne Konjugationsstelle). Hierzu wurde hauptsächlichcpptraj von AmberTools18 [282] genutzt, aber auch das mdanalysis-Modul (Versi-on 1.0.0) und das CNA-Programm (Version 2.0). Der theoretische Hintergrund zudiesen Deskriptoren wird in Anhang B beleuchtet. Eine Gesamtübersicht aller kal-kulierten Variablen ist in Tab. D.1 gezeigt. Im Falle der CNA-Parameter wurdenMomentaufnahmen ohne Wassermoleküle und Natrium- bzw. Chloridionen aus denTrajektorien für jede Nanosekunde als PDB-Datei herausgeschrieben. Die Deskripto-ren wurden anschließend in zwei unterschiedlichen Ansätzen (siehe Kap. 3.1.6.2 und3.1.6.3) verwendet, um QSPR-Vorhersagemodelle zur PEG-basierten Stabilisierungzu generieren. Für den ersten Ansatz (vide infra) wurden, wo möglich, Medianwerteberechnet, für den zweiten Mittelwerte.3.1.6.2Modellgenerierung: Ansatz 1Für diesen Ansatz wurden die bis zu diesem Zeitpunkt simulierten 85 WW-Konjugateverwendet (siehe Tab. C.2 zur Auswahl). Eine zusammenfassende Übersicht über denAblauf ist in Abb. 3.11 gegeben, die Resultate werden in Kap. 4.1.1 aufgeführt unddiskutiert.3.1. Modellproteine673.1.6.2.1Aufbereitung der DatenFilterung der KonjugateZunächst wurden sieben Konjugate, die einen experi-mentellen Messfehler von ≥0.10 kcal/mol aufwiesen, aus der Modellgenerierung ausge-schlossen: 14-1, 23-1, 32(Y23F)-1, 19(S16F:R21T)-1, 23-2, 19(R21A)-1 und 19(R21L)-1. Im Falle von 14-1 und 23-1 handelte es sich um Konjugate, die von Price et al.ebenso aus weiteren Analysen ausgeschlossen wurden, da ihre CD-Spektren deutli-che Umstrukturierungen des Proteins nicht ausschließen konnten [100]. Die siebenAsn(PEG)45-Konjugate wiesen keine strukturelle Konvergenz der Trajektorien auf(siehe Kap. 3.1.5.2) und wurden daher auch entfernt.Abbildung 3.11: Übersicht über den ersten Modellgenerierungsansatz. Die Schritte wurdeneingeteilt in eine Datenaufbereitungs-, eine Modellgenerierungs- und eine Evaluationsphase.Filterung der Deskriptoren4368 Deskriptoren wurden berechnet (Tab. D.1),von denen jene mit fehlenden Werten für einzelne Konjugate und solche, die identischoder sehr ähnlich für den gesamten Datensatz waren, entfernt wurden. Ebenso wur-den Variablen mit hoher Interkorrelation exkludiert (Limit: 0.85). Dabei wurden die-jenigen Deskriptoren herausgefiltert, die eine durchschnittliche Korrelation von über0.85 zu jedem anderen Deskriptor aufwiesen (mean absolute correlation), wobei dieseKorrelationen nach jeder Filterung neu evaluiert wurden. Die Samplingqualität der68Kapitel 3. Materialien und Methodeneinzelnen Deskriptoren wurde als unzureichend eingestuft und führte damit zu einerFilterung, sofern der Standardfehler des Mittelwerts ≥15 % des entsprechenden De-skriptorwerts für mehr als 10 % der Konjugate im Datensatz zeigte. Die Berechnungdes Standardfehlers wurde mithilfe des R-Paktes coda [174] durchgeführt, welchesdie effektive Stichprobengröße unter Berücksichtigung der Autokorrelation aufeinan-derfolgender Momentaufnahmen berechnete (siehe Glg. 1.20). Weitere Prüfungen aufStationarität (Heidelberger-Welch- und Geweke-Tests gemäß Kap. 1.2.3.2) wurdenanfänglich ebenfalls als Filter verwendet, entfernten jedoch einen relativ großen An-teil der berechneten Variablen und wurden daher als zu stringent angesehen und ausdem Ansatz wieder herausgenommen.Finale DatensätzeDie genannten Filterschritte führten zu zwei Deskriptorsets für71 WW-Konjugate. Ein Deskriptorset enthielt die 526 selektierten Deskriptoren, eindavon abgeleitetes, reduziertes Set lediglich die 33 Pin 1 WW-unspezifischen. Letz-tere dienten zur Generierung von Modellen, welche prinzipiell auf andere Proteineangewandt werden können. Hierfür wurden lediglich jene Deskriptoren ausgewählt,die die Eigenschaften der Polymerketten oder das Verhältnis zwischen der PEGy-lierten und der unPEGylierten Struktur beschreiben (und keine Eigenschaften ein-zelner, spezifischer Aminosäuren oder Strukturbereiche). Diese 33 Deskriptoren sindin Anhang D.2.4 aufgelistet. Der aufbereitete Datensatz wurde anschließend in einenTrainingsdatensatz (TS) von 56 Konjugaten (79 %) und einen zufällig ausgewählten,externen Validierungsdatensatz (Holdout-Set, HS) mit den restlichen 15 Variantenaufgeteilt. Die Verteilung der Faltungsenergien beider Datensätze mit Auflistung derHS-Konjugate ist in Abb. 3.12 gezeigt.Abbildung 3.12: Verteilung der freien Faltungsenergien im TS und im HS des ersten Mo-dellgenerierungsansatzes (Median-Wert jeweils -0.37 kcal/mol). Rechts daneben sind alle 15Konjugate aufgelistet, die dem HS angehören.3.1. Modellproteine693.1.6.2.2ModellgenerierungAusgewählte ModellartenMithilfe des Trainingsdatensatzes wurden anschlie-ßend 19 verschiedene Regressions- und 17 unterschiedliche Klassifikationsmodelle er-stellt (siehe Tab. D.2). Für letztere wurden die Konjugate in zwei Klassen eingeteilt,je nachdem, ob die Differenz der freien Faltungsenergie unter -0.40 kcal/mol lag odernicht. Diese Grenze wurde gewählt, da der Median im gesamten Datensatz einen Wertvon -0.37 kcal/mol betrug. Im Validierungsset fanden sich sieben Konjugate aus derKlasse 1 (< -0.40 kcal/mol) und acht aus der Klasse 2 (> -0.40 kcal/mol).Deskriptor-SelektionsmethodenDrei Selektierungsmethoden zur Auswahl be-deutender Deskriptoren aus dem Trainingsdatensatz wurden angewandt: (1) Rekur-sive Feature-Eliminierung (RFE) [294], (2) simuliertes Annealing [295], sowie (3) ge-netische Algorithmen [296].Bei der RFE handelt es sich um eine Backward-Eliminationsmethode, bei der zu-nächst ein Modell (eingestellt wurde hierbei die Anwendung von kreuzvalidiertenRandom Forest-Modellen und Entscheidungsbäumen mit Bagging-Methode) mit al-len Deskriptoren erstellt wird und von diesem jener Deskriptor mit der geringstenBedeutung für das Modell entfernt wird. Anschließend wird mit dem verbliebenenDatensatz ein neues Modell generiert. Für jede mögliche Anzahl an Deskriptoren (inAnsatz 1 eingestelltes Maximum: alle, Minimum: 2) wird die Auswahl zur darauffol-genden Modellgenerierung herausgeschrieben.Beim simulierten Annealing werden sukzessive zufällig generierte Deskriptorsubsetsgeneriert. Verbessert sich die Leistung des Modells gegenüber der vorherigen ran-domisierten Auswahl, so wird das neue Deskriptorset beibehalten. Mit einer gewis-sen Wahrscheinlichkeit, die mit Zunahme der Iterationen allmählich absinkt, könnendabei ebenso Modelle mit etwas schlechterer Leistung als das aktuell beste Modellausgewählt werden.Genetische Algorithmen stellen ebenso wie das simulierte Annealing zufallsbasierteMethoden dar. Hierbei wird die Deskriptorauswahl durch das Generieren von Toch-tergenerationen und Mutationen, ausgehend von zufällig erstellten Chromosomen,durchgeführt. Chromosomen stellen dabei Binärvektoren dar, die für jedes Gen (dieDeskriptoren) die Information enthalten, ob dieses für die Modellgenerierung genutztwird. Jene Chromosomen mit der besten Leistung werden an zufälligen Positionengeschnitten und miteinander vermischt, um die Tochtergenerationen zu bilden [211].Weitere DatentransformationenDie Deskriptoren des Datensatzes wurden zurModellgenerierung ebenfalls zentriert (Subtraktion des jeweiligen Mittelwerts) undanschließend skaliert (Division durch die Standardabweichung).70Kapitel 3. Materialien und MethodenGenerierung der ModelleDanach wurden Modelle mit wiederholt zehnfacherKreuzvalidierung und Testung unterschiedlicher Hyperparameter-Werte erstellt (ent-sprechend den in caret standardmäßig eingestellten Intervallen). Der gesamte Modell-bildungsprozess wurde mit verschiedenen Einstellungen der oben genannten Selekti-onsmethoden wiederholt, um verschiedene Deskriptorsubsets zu testen, und in einemder insgesamt acht Durchläufe auch ohne jegliche Selektion durchgeführt. Dies führtezu insgesamt 576 verschiedenen Modellen, deren Leistung anschließend mithilfe desValidierungsdatensatzes überprüft wurde.3.1.6.2.3EvaluationModellleistungDie drei besten Modelle aus vier verschiedenen Kategorien (Abis D) wurden anschließend für eine nähere Evaluation ausgewählt, basierend aufdem Konkordanzkorrelationskoeffizienten für Regressions- und dem AUROC-Wertfür Klassifikationsmodelle in Bezug auf das HS. Kategorie A entsprach hierbei WW-spezifischen Regressions-, Kategorie B WW-spezifischen Klassifikations-, KategorieC WW-unspezifischen Regressions- und Kategorie D WW-unspezifischen Klassifika-tionsmodellen. Die ausgewählten Modelle werden in Kap. 4.1.1 evaluiert und inter-pretiert.ApplikationsdomäneMit dem ADAN-R-Paket (Applicability domain analysis)[297] wurden die Applikationsdomänen (AD) der Regressionsmodelle über sechs ver-schiedene Kriterien evaluiert. Über eine Dimensionsreduktion mittels der PLS-Methodewurden die Modelldeskriptoren des Trainingsdatensatzes hierbei in einen Raum derentsprechenden latenten Variablen gesetzt. In diesem Raum wurden dann ebenfalls dieKonjugate des Validierungsdatensatzes lokalisiert und deren (1) euklidische Distanzzum Zentroid, sowie (2) zum nächstgelegenen Punkt des Trainingssatzes evaluiert.Ebenso wurden Ausreißer des PLS-Modells über einen auf den X-Residuen basieren-den (3) Distanz-Score detektiert. Die Vorhersage des untersuchten Testkonjugats fürdie Differenz der freien Faltungsenergie ∆∆G0f wurde weiterhin (4) mit dem Durch-schnittswert und (5) dem nächstgelegenen Wert im Trainingsdatensatz verglichen.Schließlich wurde ebenso (6) die Leistung der Prädiktion für die fünf nächstgelege-nen Konjugate im PLS-Raum beurteilt. Basierend auf diesen sechs Kriterien wurdeevaluiert, ob sich das Konjugat innerhalb der Applikationsdomäne befand (max. 1Kriterium missachtet), mit einem erhöhten Vorhersagefehler zu rechnen war (2 - 3Kriterien nicht erfüllt) oder sich außerhalb der AD des Modells befand (4 - 6 Kriteri-en nicht erfüllt). Ein Kriterium gilt als nicht erfüllt wenn der hierfür ermittelte Wertüber dem 95%-Perzentil des Trainingsdatensatzes liegt.3.1.6.3Modellgenerierung: Ansatz 2Ein modifizierter Ansatz wurde entwickelt, der sich auf Regressions- und Klassifi-kationsmodelle beschränkt, welche auch für kleinere QSPR-Datensätze geeignet seinsollten: multivariate lineare Regressionen und binomiale, logistische Klassifikationen3.1. Modellproteine71(Tab. D.5). Dieser Ansatz wurde an das Prinzip der wiederholt doppelten Kreuz-validierung (Repeated double cross validation, rdCV) angelehnt. Dieses wurde vonFilzmoser et al. [298] für PLS-Modelle konzipiert und beruht darauf, dass die opti-male Anzahl an latenten Variablen in einer inneren For-Schleife determiniert wird,die prädiktive Leistung in einer äußeren. Dieser Prozess wird für verschiedene Ka-librierungsdatensätze wiederholt, um ultimativ eine Verteilung der zur erwartendenModellleistung zu erhalten [298, 299].3.1.6.3.1Ausgeschlossene Konjugate5UXT-Strukturen (siehe Kap. 1.5.1.4) wurden generell von der Modellgenerierungausgeschlossen, da diese über keine Aminosäuren verfügen, mit denen der von Priceet al. postulierte Winkel berechenbar gewesen wäre. Ebenso wurden Daten zu vierweiteren Konjugaten (26(W11F)-1, 26(W11ν)-1, 26(W11Y)-1 und 26-9) erst nachErstellung der Modelle veröffentlicht [104], sodass diese auch nicht berücksichtigtwurden.3.1.6.3.2Definition unterschiedlicher DatensätzeIm Gegensatz zu Ansatz 1 wurden neun verschiedene Datensätze unterschiedlicherGröße untersucht (Tab. 3.3). Dies sollte dazu dienen, die Leistung von Modellenzu beurteilen, die jeweils nur bestimmte Arten von Proteinen und / oder Linker-typen berücksichtigten. Ausgehend vom in Ansatz 1 gewählten Datensatz (dieserentspricht Datensatz 1 aus diesem zweiten Ansatz, nur ohne 1SRL.20-1) wurde einerweiterter Datensatz 2 mit zu diesem Zeitpunkt neu simulierten WW-Konjugatendefiniert. Datensatz 3 schließt zusätzlich die Konjugate der Doppelwendeln (PDB-Strukturen 2ZTA und 1COI) ein. Datensatz 4 beschränkt sich dagegen allein auf jene1PIN/1SRL/2ZTA/1COI-Konjugate mit Asparagin-Linker. Der noch weiter reduzier-te Datensatz 5 beinhaltet lediglich die WW-Konjugate aus Datensatz 4. Da in diesemAnsatz 2 weiterhin ein clusterbasiertes Verfahren (vide infra) zur Definition der HSangewandt wurde, wurde dieses für den ursprünglichen Datensatz 1 auch genutzt,was zu Datensatz 6 führte (es handelt sich also um Datensatz 1 mit einem verän-derten HS). Die weiteren Datensätze 7 und 8 entsprechen Datensätzen 2 und 1 ohne1SRL.20-1, mit der Ausnahme, dass hierbei auch alle Pin 1-spezifischen Deskriptorenberücksichtigt wurden. Datensatz 9 entspricht Datensatz 6 mit einem leicht modifi-zierten Clustering-Ansatz. Eine Liste Pin 1-unspezifischer Deskriptoren in Ansatz 2ist in Anhang D.3.1 gezeigt.72Kapitel 3. Materialien und MethodenTabelle 3.3: Auflistung der getesteten Datensätze im zweiten Modellgenerierungsansatz mitder jeweiligen Anzahl an verwendeten Konjugaten nach Filterung (TS + HS).Nr.BeschreibungKonjugateHS1Datensatz ohne Doppelwendeln,Konjugate mit Kennzeichen §und Cyclohexylalanin-Mutationen.Das HS wurde hier noch ohneClustering-Ansatz zufällig ausgewähltund eine exp. Fehlergrenze von0.10 kcal/mol wurde genutzt.61 + 1117-1, 27-1, 16(S32A)-1, 27-9,16-Z, 19-2, 23(E12L:R14L)-Z,23(E12L:R14A)-Z, 19(S16F)-1,16+26-1, 1SRL.20-12Alle Konjugate mit β-Faltblattstrukturen.78 + 1126-1, 29(S32A)-1, 19-4, 23-9,29-Z, 16-2, 23(E12A:R14F)-Z,19(S16F)-1, 23(E12A:R14CHA)-Z,23-§, 1SRL.20-13Alle Konjugate.100 + 1427-1, 32-9, 16-Z, 27-2, 23(E12L:R14A)-Z,19-U, 16+26-1, 23-1, 19(R21A)-1,23(E12CHA:R14CHA)-Z, 23-§,1SRL.20-1, 2ZTA.10-1,2ZTA.18(R25A)-14Nur Asn-Konjugate.52 + 718(S16D:R17A)-1, 18(R17A)-1, 23-1,19(R21A)-1, 1SRL.20-1,2ZTA.7-1, 2ZTA.10-15Nur Asn-Konjugate der WW-Domäne.25 + 416-1, 27-1, 23-1, 19(R21A)-16Analog zu Datensatz 1, aber mit beschriebenemClustering-Ansatz für das HS.61 + 1118-1, 28-1, 30-1, 23-9, 19-3, 16-Z,32-Z, 19-2, 23(E12L:R14A)-Z,19-5, 1SRL.20-17Pin 1-spezifische Deskriptoren mit allenKonjugaten der WW-Domäne.78 + 1027-1, 19(Y23F)-1, 27-9, 19-3, 19-2,23(E12F:R14A)-Z, 23(E12A)-Z,16+26-1, 32(Y23F)-1, 23-§8Analog zu Datensatz 7 mit der Auswahl vonDatensatz 1.61 + 1017-1, 27-1, 16(S32A)-1, 27-9, 16-Z,19-2, 23(E12L:R14L)-Z,23(E12L:R14A)-Z, 19(S16F)-1,16+26-19Analog zu Datensatz 6, wobei die Anzahl anKonjugaten der „extremen“ Cluster lediglichmit 2 multipliziert wurde.61 + 1116-1, 27-1, 19-8, 19-9, 19-3, 16-Z,29-Z, 18-2, 18(R17A:R21A)-1,23(R14A)-Z, 1SRL.20-13.1.6.3.3Clustering der DatenFür die Datensätze 2 bis 9 wurde ein klar definierter Ablauf zur randomisierten Aus-wahl eines geeigneten HS erstellt. Der zu untersuchende Datensatz wurde hierzu übereinen k-Means-Algorithmus mit k = 5 auf Basis der ∆∆G0f-Werte geclustert. FünfCluster wurden ausgewählt, da bei Nutzung des gesamten Datensatzes 3 (114 Kon-jugate) bei dieser Anzahl visuell ein Kurvenknick beobachtet werden konnte (Abb.3.13), gemäß der oft genutzten „Ellenbogen“-Methode [300]. Anschließend wurde ei-ne Zufallsauswahl getroffen, die folgende Bedingungen erfüllen musste: Die relativenAnteile (1) der definierten Cluster, (2) der Linkertypen (Kennzeichen nach Tab. C.1)und (3) Strukturen, die aus anderen Proteinen als aus der WW-Domäne entstammen3.1. Modellproteine73(1SRL, 1COI, 2ZTA), mussten etwa jenen im gesamten Datensatz entsprechen. Fürdie Datensätze 1 bis 4, sowie 6 und 9, in denen die Src SH3-Domäne enthalten ist,musste diese zur Validierung in jedem Falle in das HS einsortiert werden.Abbildung 3.13: SSE (Sum of squared errors)-Werte in Abhängigkeit des Werts von k beimClustering des gesamten Datensatzes an 114 Konjugaten. Diese SSE-Werte repräsentieren dieSumme der quadrierten Abweichungen aller ∆∆G0f-Werte von den dazugehörigen Cluster-Zentroiden.Die Anzahl an ausgewählten Strukturen wurde jedoch für die zwei „extremen“ Clus-ter mit einem Faktor von 3 multipliziert, um besonders viele stark (de-)stabilisierteVarianten im HS zu erhalten. Darunter sind die beiden Cluster zu verstehen, diedie am stärksten (de-)stabilisierten Konjugate beinhalteten. Entsprechend wurde dieAnzahl in den übrigen Clustern gleichmäßig reduziert. Dies wurde durchgeführt, daim Datensatz besonders viele Konjugate in einem engen Bereich zwischen etwa -0.70und -0.20 kcal/mol präsent waren, die eine Modellbildung deutlich erschwerten. ImFokus der Modellgenerierung stand also die Detektion besonders (de-)stabilisierterKonjugate. Der Clustering-Ansatz ist am Beispiel des Datensatzes 3 in Abb. 3.14exemplarisch geschildert.3.1.6.3.4ModellgenerierungAufbereitung der DatenDer durchgeführte Ansatz ist in Abb. 3.15 für den Falldes Datensatzes 6 (relevant für Kap. 4.1.2.2) zusammengefasst. Im Gegensatz zuAnsatz 1 wurde der Datensatz hierbei noch vor Filterung der Variablen in TS und HSaufgetrennt, um einen Einfluss des letzteren auf den Modellbildungsprozess komplettzu verhindern. Es wurden dann wie im vorherigen Ansatz Konjugate gefiltert, die denKonvergenztest nicht bestanden haben. Die Filterung auf Basis der experimentellenFehler wurde mit einem neuen Cut-off-Wert von 0.16 kcal/mol durchgeführt. So wurdelediglich das Konjugat 23-1 mit einem Fehler von 0.40 kcal/mol entfernt. Deskriptorenwurden, wie in Ansatz 1 bereits beschrieben, gefiltert. Eine Ausnahme bildete dasgewählte Interkorrelationslimit, was von 0.85 auf 0.99 erhöht wurde.74Kapitel 3. Materialien und MethodenAbbildung 3.14: Durchführung des Clustering-Ansatzes am Beispiel von Datensatz 3. DieLinkerstrukturen sind gemäß Tab. C.1 benannt.SegmentierungsprozessDer Trainingsdatensatz wurde dann zufällig in gleichgro-ße Segmente aufgeteilt, sodass in jedem Segment etwa zehn Konjugate vorzufindenwaren. Ein Segment wurde nun als Testset ausgewählt. Die restlichen Segmente fun-gierten als Kalibrierungsset und wurden zur Modellgenerierung verwendet.Deskriptor-SelektionsmethodenHierzu wurden verschiedene Subsets der De-skriptoren ausgewählt: Im Falle der Regressionen durch RFE, sowie per Backward-,Forward- und Sequential-Feature-Elimination [220], im Falle der Klassifikationsmo-delle nur über RFE. Hierbei wurde die Anzahl an Deskriptoren auf max. 10 beschränkt(in Ansatz 1 gab es keine Limitierung). Backward-, Forward- und Sequential-Feature-Eliminationen wurden auf Basis von linearen Regressionen mithilfe des R-Paketsleaps durchgeführt. Im Gegensatz zum bereits in Ansatz 1 beschriebenen Konzeptder Backward-Feature-Elimination, werden bei einer Forward-Elimination sukzessiveDeskriptoren zum Modell hinzugefügt. Bei jedem Schritt wird hierbei die Inklusionjedes verbliebenen Deskriptors getestet und im Anschluss jener mit dem gerings-ten p-Wert bzw. dem stärksten Einfluss auf das Modell hinzugefügt. Der Prozesswird fortgeführt, bis keine Variable mit einem signifikanten Einfluss mehr zum Mo-dell hinzugefügt werden kann oder das Limit von 10 Deskriptoren erreicht wird. DieSequential-Feature-Elimination stellt eine Kombination der beiden anderen Ansätzedar. Zunächst werden Deskriptoren gemäß der Forward-Methode sukzessive hinzu-gefügt, nach jedem Schritt wird jedoch ein Entfernen bereits inkludierter Variablenevaluiert [211].Generierung der ModelleFür jedes generierte Subset wurden anschließend einewiederholt zehnfach kreuzvalidierte Modellgenerierung durchgeführt und statistischeParameter bzgl. des zum Training genutzten Kalibrierungssets als auch des Test-sets ermittelt. Der Prozess der Deskriptorenselektierung, der Modellgenerierung und3.1. Modellproteine7583 Konjugate 128 Deskriptoren Trainingsset (TS) 72 Konjugate Filterung: Konjugate: Konvergenztests Exp. Fehler ≥ 0.16 kcal/mol Deskriptoren: Fehlende Werte Geringe Varianz Standardfehler ≥ 15 % Interkorrelationslimit: 0.99 Simulationen Deskriptoren Trainingsset (TS) 61 Konjugate 81 Deskriptoren Testset 1 Segment Kalibrierungsset 5 Segmente Subset-Selektion ≤ 10 Deskriptoren backward, forward, sequential, RFE Für i Subsets Zehnfach wiederholte CV-Regression Statistiken Kalibrierungsset Testset Bestes Subset Höchster CCCTest CCCCalibration > 0.50 Testsets > 9 Innere Schleife 6 Iterationen Jedes Segment als Testset Äußere Schleife 500 Iterationen Neue zufällige Segmente Generierung mit ganzem TS und bestem Deskriptorensubset Finales Modell HS-Evaluation Holdout-Set (HS) 11 Konjugate Aufteilung in 6 Segmente ≈ 10 Konjugate pro Segment AUFBEREITUNG GENERIERUNG EVALUATION Abbildung3.15:Übersicht über den zweiten Modellgenerierungsansatz am BeispielvonDatensatz6.DieSchrittewurdeneingeteiltineineDatenaufbereitungs-,eineModellgenerierungs- und eine Evaluationsphase.der statistischen Evaluation (innere For-Schleife) wurde dann für mehrere Durchläufewiederholt, da jedes Segment einmal als Testset fungierte. Anschließend wurde dasTrainingsset in neue Segmente aufgeteilt und der Prozess wurde wiederholt. Schließ-lich wurde diese äußere For-Schleife für 500 Iterationen durchgeführt.3.1.6.3.5EvaluationDas beste Deskriptorensubset wurde anschließend ausgewählt. Im Falle der Regressio-nen wurde jenes selektiert, welches den höchsten Median-Wert des Konkordanzkorre-lationskoeffizienten in Bezug auf die Testsets aufwies, unter der Voraussetzung, dassdas Subset auch einen Median-Wert dieses Koeffizienten von über 0.50 für die Kali-brierungssets zeigte und, für eine ausreichende Statistik, mindestens mit zehn Testsetsverwendet wurde. Für die Klassifikationsmodelle wurde die gewichtete Genauigkeitals Auswahlkriterium genutzt, wobei das Subset mindestens einen Median-Wert von0.80 für die Kalibrierungssets zeigen musste.3.1.6.3.6Finales ModellDas auf diese Weise jeweils ausgewählte, beste Deskriptorensubset wurde anschließendgenutzt, um auf Basis des gesamten Trainingsdatensatzes mit wiederholt zehnfacherKreuzvalidierung ein finales Modell zu generieren. Dieses wurde dann anschließendmit dem externen Holdout-Set (HS) geprüft. Abwandlungen des Deskriptorensubsetsdurch verschiedene Grenzwerte eines anschließenden Interkorrelationslimits (0.75 bis0.99) wurden ebenfalls getestet. Die Ergebnisse werden in Kap. 4.1.2 geschildert. Wie76Kapitel 3. Materialien und Methodenin Ansatz 1 wurden die Applikationsdomänen der Modelle mit dem ADAN-R-Paketuntersucht.3.1.7Metadynamik-Simulationen3.1.7.1Untersuchte SystemeUm den Unterschied in der konformativen Stabilität zwischen einem PEGylierten undeinem unPEGylierten Peptid in silico erfassen zu können, wurden wohltemperierteMetadynamik-Simulationen mit multiplen „Walkern“ durchgeführt. Als Fallbeispielezur Untersuchung dienten hier die WW-Konjugate 16-1 und 27-1 (sowie deren un-PEGylierte Mutanten). Mit einer starken Stabilisierung von -0.90 ± 0.03 kcal/molbei einer PEGylierung an Position 16 und einer relativ starken Destabilisierung von0.38 ± 0.04 kcal/mol an Position 27 eigneten sich diese Varianten hierfür sehr gut.3.1.7.2KollektivvariablenAuf Basis der Ergebnisse aus [255] wurden zwei Kollektivvariablen (CV 1 und CV2) unter Nutzung des PLUMED-Plugins (Version 2.5.0) definiert: die Summe derProteinrückgrat-HB-Distanzen zwischen (1) dem ersten und zweiten, sowie (2) demzweiten und dritten β-Faltblatt. Dies entsprach folgenden Werten: für CV 1 die Sum-me aus den Distanzen G10O-H27H, F25O-E12H, E12O-F25H, Y23O-R14H, R14O-Y23Hund R21O-S16H, für CV 2 die Summe aus den Distanzen A31O-N26H, Y24O-Q33Hund Q33O-Y24H (Abb. 3.16).3.1.7.3SimulationsparameterDie vorbereiteten Modellproteine wurden wie in den Kap. 3.1.4.1 und 3.1.5 beschrie-ben solvatisiert und equilibriert. Für den Produktionslauf wurden acht „Walker“ mitder Endkonformation der NPT-Equilibrierung als Startposition gestartet. Diese wur-den für jeweils 300 ns simuliert, was einer Gesamtzeit von 2.4 µs pro Protein ent-sprach. Folgende Einstellungen wurden zur Kontrolle der CVs genutzt: Kurvenbreiteσs = 0.05, Höhe ω = 0.05 und Frequenz τG = 10 ps. Eine Temperierung der MetaMDswurde mithilfe eines Bias-Faktors von 8 erreicht. Der erlaubte Sampling-Bereich derCV 1 lag bei 0 bis 5.2 nm, jener von CV 2 bei 0 bis 2.8 nm. Um keinen Konformati-onsraum außerhalb des definierten Gitters zu erreichen, wurden die oberen Energie-barrieren 4.8 nm für CV 1 und 2.4 nm für CV 2 mit der Konstanten κ mit einemWert von 1000 kJ/mol eingestellt. Diese Konstante wird genutzt, um die Stärke deszurückhaltenden Potentials VCV zu definieren, das auf das System ausgeübt wird,sobald dieses die genannten Gittergrenzen überschreitet:VCV = κ(x − a + os)e(3.1)Dabei ist s ein Skalierungsfaktor mit Standardwert 1, e definiert das Potenzgesetzmit Standardwert 2, a ist das vom Nutzer eingestellte Limit der CV und o (offset)3.1. Modellproteine77eine mögliche Verschiebung dieses Limits, die jedoch standardmäßig auf Null gesetztist [301]. Die Dateien mit den Kurvenhöhen (Hills) wurden durch die Walker alle100 Zeitschritte gelesen. Die CVs wurden alle 500 Zeitschritte herausgeschrieben. ImAnschluss wurden die Simulationen mithilfe des R-Pakets metadynminer analysiert[302]. Die Resultate der Simulationen werden in Kap. 4.2 beschrieben.Abbildung 3.16: Darstellung der für die Kollektivvariablen genutzten Distanzen (CV 1 inmagenta, CV 2 in türkis). Die relevanten Atome sind in den selben Farben gekennzeichnet.3.1.8Vergleich der PolymereFür einen systematischen Vergleich des in silico-Verhaltens der Polymere PEG, LPG,PMeOx und PEtOx wurden verschiedene Eigenschaften für eine Reihe von WW- undHer2-Affibody-Konjugaten näher untersucht: (1) die Interaktionsenergien der Poly-meratome mit jeder Aminosäure und (2) dem gesamten Protein (jeweils mit -1 mul-tipliziert), (3) die vom Polymer bedeckte Proteinoberfläche und (4) der Anteil derPolymeroberfläche, die bedeckt wird, (5) Protein-Polymer HB pro Polymeroberflä-che in Ų, (6) die Anzahl an Wassermolekülen innerhalb von 5 Å der Polymeratomepro Oberfläche in Ų, (7) das massengewichtete RMSF (Root-mean-square fluctuati-on)-Verhältnis des Proteinrückgrats pro Aminosäure im Vergleich zur Simulation derunPEGylierten Variante, sowie schließlich (8) die massengewichteten RMSF-Werteder Polymerschweratome.Im Falle der Pin 1 WW-Domäne wurden die Asn-Konjugate XX-1, XX-ϵ, XX-λ undXX-ψ für die 13 Konjugationsstellen XX = 14, 16, 17, 18, 19, 21, 23, 26, 27, 28, 29, 30und 32 miteinander verglichen. Für den Her2-Affibody wurden die 15 Stellen XX =7, 8, 15, 21, 23, 25, 29, 33, 39, 43, 46, 47, 49, 50 und 53 hinsichtlich der Cys-Maleimid-Konjugate XX-4, XX-φ, XX-γ und XX-ω untersucht. Die Ergebnisse sind in Kap. 4.3aufgeführt.78Kapitel 3. Materialien und Methoden3.2BiologikaIm Folgenden werden die simulierten Konjugate von IFN-α2a beschrieben (Kap.3.2.2). Analog zur Vorgehensweise bei den Modellproteinen sollen die Parametrisie-rung (Kap. 3.2.3), die Modellierung (Kap. 3.2.4) und das Simulationsprotokoll (Kap.3.2.5) beschrieben werden. Schließlich werden die Durchführung einer Coarse-grained-Simulation eines PEG-Konjugats (Kap. 3.2.6) und die experimentellen Arbeiten derVerbundpartner zur Charakterisierung der thermischen Stabilität und Bioaktivität(Kap. 3.2.7) geschildert. Ergebnisse und Diskussion sind in Kap. 5 zu finden.3.2.1Wildtyp-SimulationenIm Falle von IFN-α2a wurden die in der PDB-Datei 1ITF aufgelisteten, ersten zehnder 24 Konformationen für jeweils 150 ns simuliert (insgesamt 1.5 µs), mit den inden Kap. 3.1.4.3 und 3.1.5.1 bereits geschilderten Einstellungen für konventionelleMD-Simulationen der Modellproteine.3.2.2Simulierte KonjugateIFN-α2a wurde experimentell an der Position K31 mit den Polymeren PEG, LPGund PEtOx konjugiert, indem an dieser Stelle die nicht-natürliche Aminosäure AZKeingebaut wurde und diese über eine SPAAC mit einem DBCO- oder BCN-Linkerkonjugiert wurde (siehe Kap. 1.1.2.3 und 3.2.7). Diese Konjugationsstelle stellt eineder ebenfalls in Pegasys® adressierten Positionen dar und eignete sich, unter Berück-sichtigung der Binderegionen des Proteins zu den Rezeptoren IFNAR1 und IFNAR2(vgl. Abb. 1.14-A), neben K134 als Angriffsort des Linkers. K31 und K134 stellen dieKonjugationsstellen in Pegasys für PEG dar, die vermutlich zur geringsten Verminde-rung der Bioaktivität führen [303]. Ebenso zeigen die in dieser Arbeit durchgeführtenWT-Simulationen, dass die Seitenketten beider Aminosäuren weit entfernt von denBinderegionen des Proteins vorzufinden sind und sich nicht zu diesen hin ausrich-ten (Abb. 3.17). K134 lieferte experimentell bei der Expression allerdings geringereAusbeuten, sodass der Fokus auf K31 gelegt wurde [304].In silico wurden BCN-Varianten mit unterschiedlich langen Polymeren (10 - 50 kDa)simuliert. DBCO-Konjugate wurden lediglich mit 10-kDa-Polymerketten simuliert,da auch nur für Konjugate dieser Größe experimentelle Daten erhoben wurden (Tab.3.4).3.2. Biologika79Abbildung 3.17: (A) Darstellung der Seitenketten von K31 und K134 für jede Nanosekundeeines Replikons der WT-Simulationen, nach Alignierung des Proteins auf die Kristallstruktur3SE3. (B) Die mittlere Distanz von K31 zum nächstgelegenen Atom P49-Cβ von IFNAR2beträgt 11.48 Å, (C) jene von K134 zu S64-Oγ von IFNAR1 16.69 Å (Nachdruck mit Geneh-migung aus [304], Copyright 2021 American Chemical Society).Tabelle 3.4: Auflistung aller Konjugate von IFN-α2a, die für die Arbeiten in Kap. 5 simuliertwurden. Dabei sind die zum Zeitpunkt der Modellierung experimentell angestrebten Moleku-largewichte, die Anzahlen modellierter Monomere und die entsprechenden Massen letztereraufgelistet.LinkerPolymerZielgewicht [kDa]Modellierte MonomereMasse [kDa]BCNPEG1022710.02BCNLPG1013610.07BCNPEtOx1010110.03DBCOPEG1022710.02DBCOLPG1013610.07DBCOPEtOx1010110.03BCNPEG2045219.92BCNLPG2029021.45BCNPEtOx2020019.84BCNPEG4090339.80BCNLPG5065048.15BCNPEtOx3030029.7680Kapitel 3. Materialien und Methoden3.2.3ParameterfindungDie Parametrisierung von PEG-Monomeren wurde bereits in Kap. 3.1.3.3 beschrie-ben. Mit den LPG-, PMeOx- und PEtOx-Monomeren wurde analog verfahren. Dadie LPG-Strukturen Stereozentren enthielten, wurde hierbei ein Dimer als einzelnerBaustein definiert, welches über ein (S)- und ein (R)-konfiguriertes Monomer ver-fügte. Die konjugierte Lysinseitenkette wurde mitsamt des konjugierten DBCO- oderBCN-Linkers als eine neue Aminosäure definiert. Während für das Proteinrückgratdieselben Endgruppen wie in Kap. 3.1.3.1 und 3.1.3.2 verwendet wurden, wurde diejeweils endständige Amid-Funktion, die mit der Polymerkette verbunden war, mitPEG-Monomeren während der Ladungsverteilung über die in Kap. 3.1.3.2 beschrie-bene antechamber-Methode versehen (Abb. 3.18).Abbildung 3.18: Parametrisierte BCN- (links) und DBCO-Linker (rechts). Die Endgruppen,die dabei gemäß der beschriebenen antechamber-Methode in xleap entfernt wurden, sind blaugekennzeichnet.3.2.4Modellierung der StrukturenDie NMR-Struktur von IFN-α2a, die bei pH 6 aufgenommen worden war, wurdeohne Änderung der Protonierung verwendet. MOE deprotonierte den N-Terminusund änderte die Tautomerie einiger solvensexponierter Histidine bei pH 7, Maestro2019.1 von Schrödinger [305] hingegen nicht. Die Polymerketten mit den in Tab. 3.4aufgeführten Größen wurden zunächst mittels tleap von AmberTools18 in einer ausge-streckten Konformation generiert. Diese wurden dann an den BCN-Linker konjugiertund in die Position K31 der NMR-Struktur des IFN-α2a eingefügt. DBCO-Variantenwurden analog modelliert.3.2. Biologika81Zunächst wurden die Aminosäuren mit Ausnahme der Position K31 in MOE fixiert.Die potentielle Energie des Systems wurde minimiert (Amber14:EHT-Kraftfeld, Gra-dient: 0.001 kcal/mol/Å), bevor eine kurze Simulation mit implizitem Solvens (1 ns)durchgeführt wurde. Hierfür wurde der Noisé-Poincare-Andersen-Ansatz [306, 307]unter Anwendung des generalisierten Born-Solvensmodells [308] in MOE genutzt.Dies führte zu komprimierten Startstrukturen, bei denen sich das Polymer auf derProteinoberfläche befand (Abb. 3.19). Erste klassische MD-Simulationen mit langge-streckten Polymerketten als Startpositionen wurden getestet, führten jedoch zu keinerAnnäherung des Polymers auf die Proteinoberfläche in annehmbarer Zeit. Auf dieseWeise konnte ebenso die Anzahl nötiger Solvensmoleküle deutlich verringert werden.Zehn verschiedene Startstrukturen wurden für jedes Konjugat generiert, wobei sichdas Polymer dem Protein in den Simulationen mit implizitem Solvens jeweils aus ver-schiedenen Richtungen annäherte und somit in den darauffolgenden gaußbeschleu-nigten Molekulardynamik-Simulationen (Kap. 3.2.5) Wechselwirkungen mit unter-schiedlichen Arealen abgebildet werden konnten. Erste Simulationen mit rechtecki-gen Solvensboxen wiesen Interaktionen des Polymers mit Proteinen der periodischenNachbarn auf. Daher wurden die generierten Konjugate in isometrischen, trunkiertenOktaedern mit einem Mindestabstand von 15 Å des Konjugats bis zum Rande derBox mit TIP3P-Wasser solvatisiert und mit Natrium- oder Chloridionen das Systemneutralisiert.IFN-α2a Abbildung 3.19: Langgestreckte Strukturen des Polymers PEG 10 kDa nach Generierungdurch tleap und Einfügung an Position K31. Rechts sind die Startkonformationen der Simu-lationen dargestellt, die nach initialer Simulation in implizitem Solvens erhalten wurden.82Kapitel 3. Materialien und Methoden3.2.5Gaußbeschleunigte Molekulardynamik-SimulationenFür die zehn generierten Startkonformationen jedes Konjugats wurde eine NVT-Equilibrierung analog zu jener der Modellproteine (siehe Kap. 3.1.5.1) durchgeführt.Anschließend wurde eine auf Basis des GaMD-Ansatzes [176, 177] 20 ns lange NPT-Equilibrierung getätigt. Hierbei wurden zunächst 2 ns einer konventionellen MD-Simulation durchgeführt. In der zweiten Hälfte dieser Phase (1 ns) wurden Wertefür Vmax, Vmin, Vavg und σV gesammelt und entsprechend E und k0 berechnet. Dieletzten Werte für E und k0 wurden dann genutzt, um ein konstantes, gaußbeschleu-nigtes Zusatzpotenzial für eine weitere Nanosekunde anzuschalten. Dieses Zusatzpo-tenzial wurde dann in den darauffolgenden, restlichen 17 ns der Equilibrierung durchÄnderung von E und k0 nach jedem Zeitschritt verändert. Für den GaMD-Ansatzwurde die duale Boost-Methode gewählt, bei der das Gesamt- als auch das Dieder-potential im System durch das harmonische Zusatzpotenzial angehoben werden. Dasobere Limit der Standardabweichung σ0 der beiden Potenziale wurde auf 6 kcal/molgesetzt, dem Standardwert in NAMD. Momentaufnahmen wurden alle zehn Pikose-kunden herausgeschrieben. Im Anschluss an diese Equilibrierungsphase wurde dannder Produktionslauf mit randomisierten Startgeschwindigkeiten für insgesamt 150 nsdurchgeführt, in welchem die Werte E und k0 weiter verändert werden konnten.Jede Polymervariante wurde mit zehn verschiedenen Startkonformationen und somitfür 1.5 µs simuliert. Dies ähnelt den Simulationsansätzen aus [107] und [309], wobeidie zehn unterschiedlichen Startkonformationen in diesen Referenzstudien mittels derpysimm-Software [310, 311] über das Prinzip eines selbstmeidenden Pfades generiertwurden. Dabei wurden sukzessive neue Monomere an die Konjugationsstelle gebautund die entsprechende Kette nach jedem Schritt minimiert. Wenn jedoch bedachtwird, dass die Kette experimentell über eine Grafting to-Methode mit ihrer ganzenLänge an das Protein konjugiert wurde, sollte der in dieser Arbeit angewandte Ansatzüber eine Energieminimierung und eine Simulation der ausgestreckten Polymerketteeine mögliche Alternative darstellen. Die Dynamik der PBK wird in Kap. 5.1 geschil-dert.3.2.6Coarse-grained-SimulationEine an Position K31 mit 10 kDa PEG konjugierte Variante von IFN-α2a wurdemithilfe des Kraftfelds MARTINI 2 für 10 µs simuliert [158, 312]. Geeignete Parame-ter und Startstrukturen für PEG wurden mithilfe des PolyPly-Programms (Version1.1.0) vorbereitet [313, 314]. Zum Coarse-graining des Proteins wurde zunächst dasmartinize.py-Skript genutzt. Die erhaltene Struktur wurde in eine kubische Simu-lationsbox mit 8 nm Mindestabstand des Proteins zum Rand gesetzt. Eine initia-le Energieminimierung (100 Schritte) der vergröberten Struktur im Vakuum wur-de dann durchgeführt. Anschließend wurde die Polymerkette mithilfe von PolyPlysukzessive in einer randomisierten Konformation ohne Linker an die terminale Sei-tenkettenkugel von K31 konjugiert und verlängert (227 Monomere). Vor und nach3.2. Biologika83einer anschließenden Solvatation mit nicht polarisierbarem MARTINI-Wasser undNeutralisation durch Natrium- oder Chloridionen wurde das System jeweils noch-mals energieminiert, bevor eine Equilibrierung für 10000 Schritte mit 2 fs-Intervallendurchgeführt wurde. Schließlich folgte der eigentliche Produktionslauf für 10 µs mit20-fs-Zeitschritten. Coulomb-Interaktionen wurden über die Reaktionsfeld-Methodeund VdW-Wechselwirkungen gemäß eines Verlet-Schemas jeweils mit einer Cut-off-Distanz von 1.1 nm berücksichtigt. Zur Temperaturkontrolle wurde die Velocity-Rescale-Methode, zur Druckkontrolle das Parrinello-Rahman-Barostat angewandt.Eine Struktur des Konjugats vor Start des Produktionslaufes ist in Abb. 3.20 ge-zeigt. Die Beobachtungen werden in Kap. 5.2 beschrieben.Abbildung 3.20: Startstruktur der MARTINI-Simulation vor dem Produktionslauf. PEG-Kugeln sind in cyan dargestellt, Proteinrückgratkugeln in rot und Seitenkettenkugeln in gelb.Die Konjugationsstelle der K31-Seitenkettenkugel ist markiert. Rechts daneben ist eine sche-matische Darstellung der Vergröberung der PEG-Monomere illustriert.3.2.7Experimentelle ArbeitenLPG wurde an der FU Berlin über eine anionische und PEtOx an der FSU Jena übereine kationische ringöffnende Polymerisation gewonnen, entsprechende mPEG-Kettenkonnten kommerziell erworben werden (Abb. 3.21-A, B und C). Alle weiteren Arbeitenin Kap. 3.2.7 wurden durch Niklas Hauptstein an der JMU Würzburg durchgeführt.IFN-α2a und funktionalisierte Varianten wurden durch Nutzung einer Amber-Codon-Mutation in BL21 Star™(DE3)-Zellen exprimiert, erhaltene Einschlusskörperchenisoliert und aufbereitet, das Protein rückgefaltet und über Ionenaustausch- und Grö-ßenausschlusschromatographien aufgereinigt. Anschließend wurde die SPAAC durch-geführt (Abb. 3.21-D), die Proteine wurden durch verschiedene Methoden weiter cha-rakterisiert (CD-Spektroskopie, Polyacrylamid-Gelelektrophorese, dynamische Licht-streuung und Massenspektrometrie).84Kapitel 3. Materialien und MethodenUm die Bioaktivität der Konjugate zu überprüfen, wurden SEAP (Secreting alkalinephosphatase)-Assays mit HEK (Human embryonic kidney)-IFNα/β-Zellen durchge-führt. Durch Bindung aktiven Interferons setzen diese Zellen über den JAK-STAT-Signalweg SEAP in das umgebende Medium frei. Die Aktivität der Phosphatase kannwiederum durch eine QUANTI-Blue™-Lösung detektiert werden [315].Zur Untersuchung der thermischen Stabilitäten wurden Thermal Shift Assays durch-geführt, bei denen die Entfaltung über die Bindung des Farbstoffs SYPRO™ Orangebei erhöhten Temperaturen beobachtet wird. Durch die Bindung des hydrophobenFarbstoffs kommt es zur Änderung der Fluoreszenzintensität, wodurch die Schmelz-bzw. Entfaltungstemperaturen im Anschluss über den Wendepunkt der erhaltenenFluoreszenzkurve ermittelt werden können [89, 90].Zusätzlich wurde die Proteolysestabilität der Konjugate untersucht, indem sogenannteLimited Proteolysis (LiP)-Assays durchgeführt wurden [316]. Hierbei wurden die PBKmit einer Proteinase K versetzt. Anschließend wurden zu verschiedenen Zeitpunkten(nach 1, 3, 5, 10 und 30 min) Proben entnommen und aufbereitet. Schließlich wurdendie durch die Proteolyse erhaltenen Peptidfragmente mit massenspektrometrischenMethoden charakterisiert. Die Ergebnisse werden in den Kap. 5.3 und 5.4 mit denBeobachtungen der Simulationen in Verbindung gebracht.Abbildung 3.21: (A) Synthese von LPG über eine anionische ringöffnende Polymerisation.(B) Synthese von PEtOx über eine kationische ringöffnende Polymerisation. (C) Strukturder verwendeten mPEG-Kette. (D) Funktionalisierung der gewonnenen Polymere durch dieLinker BCN und DBCO, sowie anschließende SPAAC-Reaktion über eine Azidgruppe desProteins zur Konjugation der synthetisierten Polymere.85Kapitel 4Ergebnisse: Modellproteine4.1QSPR-VorhersagemodelleDie systematischen Vorarbeiten von Price et al. zum Modellprotein Pin 1-WW, wel-ches mit PEG-Oligomeren verknüpft wurde, deuten darauf hin, dass die PEG-basierteStabilisierung durch deutlich simplere Systeme als die üblicherweise therapeutisch ge-nutzten Biologika mit Polymerketten von 5 bis 40 kDa beurteilt werden kann. SolcheModellproteine sind auch durch in silico-Ansätze in größerem Maßstab analysierbar.Während erweiterte Simulationsmethoden prinzipiell zur Ermittlung von Energiedif-ferenzen genutzt werden können (vgl. Kap. 4.2), stellen diese zum einen sehr rechen-intensive Ansätze zur Untersuchung mehrerer Systeme dar, zum anderen handelt essich im vorliegenden Fall um äußerst geringe energetische Differenzen. Im Hinblickauf die Genauigkeit computerchemischer Methoden zur Ermittlung von Energiewer-ten kleinerer Moleküle wird üblicherweise von einem Mindestfehler von 1 kcal/molgesprochen, im Falle makromolekularer Systeme wie PBK ist dieser sicherlich nochgrößer [317, 318].Daher kann davon ausgegangen werden, dass simulationsbasierte Energieberechnun-gen für die hier vorliegenden Konjugate nur schwer zu realisieren sind. QSPR-Ansätzestellen in diesem Falle einen recheneffizienteren Ansatz dar und können zusätzlichEinblick in die wichtigsten Stabilisierungsfaktoren zur Vorhersage weiterer Konjuga-tionseffekte geben. Da sich die Polymerketten von PBK jedoch sehr flexibel verhaltenund verschiedene Areale des Proteins adressieren können, ist ein wie in dieser Ar-beit gewählter simulationsbasierter Ansatz unter Berücksichtigung der Dynamik zurGenerierung geeigneter Deskriptoren sinnvoll.Im Folgenden sollen die Ergebnisse der in Kap. 3.1.6 beschriebenen Modellgenerierun-gen geschildert werden. Zunächst werden die QSPR-Modelle aus Ansatz 1 beschrie-ben (Kap. 4.1.1), in welchem unterschiedliche Methoden des maschinellen Lernens aufBasis der Vorhersagen für ein einzelnes Holdout-Set getestet wurden. Im Anschlusswerden die Resultate zu Ansatz 2 beschrieben (Kap. 4.1.2), in welcher ein erweiter-tes Kreuzvalidierungsverfahren angewandt wurde, um multilineare Regressions- undbinomiale logistische Klassifikationsmodelle zu erstellen.86Kapitel 4. Ergebnisse: Modellproteine4.1.1Ansatz 1Im ersten Modellierungsansatz zur Vorhersage der PEG-basierten Stabilisierung vonPBK wurden verschiedene Modellarten getestet. Dabei wurden Modelle mithilfe einesTrainingssets (TS) erstellt und auf Basis der Leistung bzgl. eines einzelnen Holdout-Sets (HS) selektiert. Dies wurde gemäß Kap. 3.1.6.2 für vier verschiedene Katego-rien durchgeführt: (A) Pin 1-spezifische und (C) unspezifische (auf andere Systemetheoretisch transferierbare) Regressionsmodelle, sowie (B) Pin 1-spezifische und (D)unspezifische Klassifikationsmodelle. In den Kategorien A und B zeigten die Pin 1-spezifischen Modelle jedoch eine vollständige Überanpassung (Overfitting) an das TSund nutzten dabei eine in Relation zur Datensatzgröße viel zu große Anzahl an De-skriptoren. Die Resultate dieser Kategorien sollen daher im Folgenden nicht weiterdiskutiert werden. Eine Übersicht der drei ausgewählten Modelle jeder Kategorie istdennoch in Tab. D.3 zu finden.Auch im Falle der hier nun näher diskutierten Pin 1-unspezifischen Modelle mit deut-lich weniger Deskriptoren sollte angemerkt werden, dass es sich um sehr komplexeModelltypen mit vielen Hyperparametern handelt, die bei solch kleinen Datensät-zen auch bei einer geringen Anzahl an Deskriptoren zum Overfitting neigen können.Weiterhin wurden die besten Modelle bzgl. der Vorhersageleistung für ein einzelnesHoldout-Set selektiert, weshalb es sich um kein ausreichend externes Validierungs-set handelt. Ansatz 1 soll wegen dieser Limitierungen lediglich aufgrund folgenderAspekte präsentiert werden: (1) Die Grenzen der Nutzbarkeit erweiterter Methodendes maschinellen Lernens in Bezug zum zur Verfügung stehenden Datensatz sollenaufgezeigt werden, (2) die Eignung der berechneten Deskriptoren zur Wiedergabeder in der Literatur postulierten Stabilisierungsmechanismen soll an exemplarischenSimulationen evaluiert werden, (3) der Ansatz dient als Vorarbeit zu den besser vali-dierten Modellen im hierzu abgewandelten Ansatz 2 (Kap. 4.1.2).4.1.1.1ReferenzmodelleZunächst ist zur Beurteilung der Leistung der generierten Regressionsmodelle ent-scheidend, wie gut ein Modell abschneiden würde, welches jedem Konjugat des HSden Medianwert des TS (-0.37 kcal/mol) zuordnen würde. Dieses Modell soll als Re-ferenz dienen und liefert einen RMSEHS-Wert von 0.39 kcal/mol. Somit sind nur Vor-hersagemodelle mit einem deutlich kleineren Wert als 0.39 kcal/mol für eine weitereAnalyse von Relevanz.Analog hierzu kann ein Referenzmodell für die Klassifikationsmodelle generiert wer-den. Hierfür wurde jedes Konjugat des HS einmalig zufällig einer der beiden Klassenzugeordnet (also mit einer Wahrscheinlichkeit von jeweils 50 %, da die Verteilung imTS auch annähernd 1:1 ist). Für eine solche Einteilung wird eine CCRHS von 0.67erhalten. Geeignete Vorhersagemodelle sollten also zu einer deutlich besseren CCRHSführen.4.1. QSPR-Vorhersagemodelle874.1.1.2Modelle C1 und D1Im Folgenden sollen die beiden besten Modelle der Kategorien C und D beschriebenwerden (genannt C1 und D1). C1 stellt ein entscheidungsbaumbasiertes XGBoost-Modell dar, mit RMSEHS- und CCCHS-Werten von 0.24 kcal/mol bzw. 0.75 (Tab.4.1). Beim Blick auf die Vorhersagen wird deutlich, dass eine korrekte Rangordnungder Konjugate v. a. in der Nähe des Medians des Datensatzes nur bedingt möglich ist(Abb. 4.1). Außerhalb der ADAN-Applikationsdomäne (siehe Kap. 3.1.6.2) befindensich 4 der 15 HS-Konjugate: 19-4, 16-Z, 19-J und 16+19-1 (siehe Tab. D.4 für dieEinteilungen pro Konjugat).Das Konjugat der Src SH3-Domäne, welches von Price et al. zur Validierung ge-nutzt wurde (genannt 1SRL.20-1), wird durch das Modell korrekterweise als stabili-siert eingestuft. Mit dem kristallstrukturbasierten Vorhersagemodell von Price et al.(Funktion der Regressionsgeraden: ∆∆G0f = −1.05 + 0.0073 θPrice [100]) wird eineFaltungsenergie von -0.49 kcal/mol erhalten. Das Modell C1 enthält zusätzliche In-formationen von Konjugaten mit anderen Linkerstrukturen und erzielt eine ähnlicheLeistung, es sagt eine Differenz der freien Faltungsenergie von -0.78 kcal/mol voraus.Dies ist zwar immer noch relativ weit vom experimentellen Wert von -1.20 kcal/molentfernt, 1SRL.20-1 stellt mit einem solch negativen Wert und einem experimentellenFehler von 0.10 kcal/mol jedoch auch ein Konjugat dar, welches im Vergleich zu denanderen PBK im Datensatz eine sehr große Stabilisierung zeigt und theoretisch durchden Filterschritt in Ansatz 1 herausgefiltert werden würde. ADAN klassifiziert die-ses Konjugat weiterhin in Kategorie 3, was bedeutet, dass diese Vorhersage als eherunsicher einzustufen ist.Das Modell nutzt 11 verschiedene Deskriptoren. Die bedeutendsten sind protpegh-bonds (Anzahl an Protein-PEG-HB), protprothbondsratio (Verhältnis der Anzahl anProtein-Protein-HB zwischen PEGylierter und unPEGylierter Variante), shelloratio(Verhältnis der Anzahl an Wassern in der zweiten Hydrathülle zwischen PEGylier-ter und unPEGylierter Variante), rmsdp (RMSD-Wert der Polymerschweratome) undrdfpeg_200 (RDF-Wert um PEG innerhalb von 3 - 4 Å).Das beste Modell der Kategorie D, ein neuronales Netz, erzielt einen CCRHS-Wertvon 0.80 und einen AUROCHS-Wert von 0.91. Es nutzt eine verborgene Schicht (Hid-den layer) mit 5 Neuronen und 2 zusätzlichen Bias-Neuronen (Abb. 4.2). 1SRL.20-1wird mit einer Wahrscheinlichkeit von 58 % korrekterweise der Klasse 1 zugeordnet.Das Modell nutzt jedoch andere Deskriptoren als das Regressionsmodell C1, u. a.pgbrg (PEG-PEG-HB über überbrückende Wassermoleküle), rdfoh_200ratio (RDF-Verhältnisse um S/T/Y), radgyrpeg (Gyrationsradius der Polymerkette) und rmsdra-tio (Proteinrückgrat-RMSD-Verhältnisse zwischen PEGylierter und unPEGylierterVariante).88Kapitel 4. Ergebnisse: ModellproteineTabelle 4.1: Statistische Validierungsparameter für die Modelle C1 und D1, sowie derenverwendete Deskriptoren (geordnet nach der Bedeutung für das Modell). Für das XGBoost-Modell wird der Gewinn (Gain) dargestellt. Dieser repräsentiert den Bruchteil des Beitragsdieses Deskriptors für alle im Prozess generierten Entscheidungsbäume (Summe = 1) [226].Für das neuronale Netz wird eine Methode genutzt, die die Absolutwerte aller Gewichtungenim Netzwerk bzgl. der einzelnen Deskriptoren kombiniert [319].Modell C1#ArtRMSETSMAETSCCCTSR2TSRMSEHSMAEHSCCCHSR2HSC1xgbTree0.140.100.950.920.240.200.750.63DeskriptorGainprotpeghbonds0.23protprothbondsratio0.20shelloratio0.12rmsdp0.12rdfpeg_2000.07rdf_oh400ratio0.06rmsdratio0.06totalint0.06pgbrg0.05radgyrpeg0.02entropysideratio0.01Modell D1#ArtAUROCTSCCRTSAUROCHSCCRHSD1nnet0.950.930.910.80DeskriptorGewichtungenpgbrg17.70rdfoh_200ratio16.56radgyrpeg16.04rmsdratio15.74rmsdp11.10shellohratio11.04pg4dist7.22entropysideratio4.62Abbildung 4.1: Auftragung der durch das Modell C1 vorhergesagten Differenzen der freienFaltungsenergien gegen die experimentellen Werte der (A) TS- und (B) HS-Konjugate.4.1. QSPR-Vorhersagemodelle89Abbildung 4.2: Visualisierung des Netzwerks von Modell D1. Die Neuronen der Eingabe-schicht (Input layer, links) sind grau gefärbt, in der Mitte sind die Neuronen der verborgenenSchicht (H) und der Bias-Neuronen (B) dargestellt. Die Ausgabeschicht (Output layer, O) istrechts zu sehen. Die Neuronen sind mit Linien vernetzt, deren Breite die Größe und Farbedie Vorzeichen der Gewichtungen symbolisieren (schwarz: +, grau: -).4.1.1.3Modelle E und FIm Anschluss wurden weitere Modellgenerierungen mithilfe verschiedener Kombina-tionen der bedeutendsten Pin 1-unspezifischen Modelldeskriptoren aus den Katego-rien A bis D getestet. Dabei wurden die in Ansatz 1 erfolgreichsten Modelltypengenutzt: Regressionen mithilfe des XGBoost-Ansatzes und neuronale Netze zur Klas-sifikation. Daraus ergaben sich ultimativ die Modelle E und F (Tab. 4.2, sowie Abb.4.3 und 4.4).Modell E zeigt im Vergleich zu C1 einen leicht verbesserten RMSEHS-Wert (0.21kcal/mol), ein RMSETS-Wert von nur 0.02 signalisiert allerdings auch, vergleichbarmit Modellen aus der Kategorie A, eine deutliche Überanpassung. Die Vorhersage von-0.63 kcal/mol für 1SRL.20-1 liegt im Bereich jener des Modells C1, doch auch hier-bei wird das Konjugat bei der ADAN-Analyse einer erhöhten Kategorie 2 zugeordnet,was für eine unsichere Vorhersage spricht. 5 der 15 Konjugate des HS werden ebensohöheren ADAN-Kategorien zugeordnet (siehe Tab. D.4 für die Einteilungen pro Kon-jugat), was vergleichbar mit den Resultaten zu Modell C1 ist (4/15). Im Gegensatzzu letzterem stellt shelloratio für das Modell E den bedeutendsten Deskriptor dar.Verglichen mit D1 zeigt Modell F eine etwas schlechtere Leistung bzgl. des TS, jedocheinen verbesserten CCRHS-Wert von 0.87. Ebenso klassifiziert es 1SRL.20-1 richtig.Der wichtigste Deskriptor ist hierbei radgyrpeg. Eine Visualisierung des Modells zeigt,dass das Netzwerk aus einer verborgenen Schicht mit 5 Neuronen und 2 zusätzlichenBias-Neuronen besteht (Abb. 4.4). In Relation zur Anzahl genutzter Deskriptorenstellt dies eine relativ große Anzahl dar. Zur Vermeidung einer möglichen Überanpas-sung sollte die Größe der verborgenen Schicht in der Regel kleiner als die Größe derEingabeschicht (Input layer) sein [320].90Kapitel 4. Ergebnisse: ModellproteineTabelle 4.2: Statistische Validierungsparameter für die Modelle E und F, sowie deren verwen-dete Deskriptoren (geordnet nach der Bedeutung für das Modell). Für das XGBoost-Modellwird der Gewinn (Gain) dargestellt. Dieser repräsentiert den Bruchteil des Beitrags diesesDeskriptors für alle im Prozess generierten Entscheidungsbäume (Summe = 1) [226]. Für dasneuronale Netz wird eine Methode genutzt, die die Absolutwerte aller Gewichtungen im Netz-werk bzgl. der einzelnen Deskriptoren kombiniert [319].Modell E#ArtRMSETSMAETSCCCTSR2TSRMSEHSMAEHSCCCHSR2HSExgbTree0.020.021.001.000.210.180.780.72DeskriptorGainshelloratio0.21rmsdratio0.16molbur0.13pgbrg0.11radgyrpeg0.11rdfoh_200ratio0.11protpeghbonds0.09protprothbonds0.09Modell F#ArtAUROCTSCCRTSAUROCHSCCRHSFnnet0.850.750.910.87DeskriptorGewichtungenradgyrpeg29.47rmsdratio27.58rdfoh_200ratio24.64pgbrg18.31Abbildung 4.3: Auftragung der durch das Modell E vorhergesagten Differenzen der freienFaltungsenergien gegen die experimentellen Werte der (A) TS- und (B) HS-Konjugate.4.1. QSPR-Vorhersagemodelle91Abbildung 4.4: Visualisierung des Netzwerks von Modell F. Die Neuronen der Eingabe-schicht (links) sind grau gefärbt, in der Mitte sind die Neuronen der verborgenen Schicht(H) und der Bias-Neuronen (B) dargestellt. Die Ausgabeschicht (O) ist rechts zu sehen. DieNeuronen sind mit Linien vernetzt, deren Breite die Größe und Farbe die Vorzeichen derGewichtungen symbolisieren (schwarz: +, grau: -).4.1.1.4Modelle G und HAufgrund einer möglichen Überanpassung durch die Nutzung sehr komplexer Modell-arten mit einem relativ kleinen Datensatz wurden die Deskriptoren der Modelle Eund F weiterhin verwendet, um eine multivariate lineare Regression (Modell G) undeine binomiale logistische Regression (Modell H) zu generieren. Die Ergebnisse sindin Tab. 4.3 und in Abb. 4.5 gezeigt. Die deutlich schlechtere Leistung bzgl. des TS imVergleich zu den Modellen E und F legt die Vermutung nahe, dass auch die komple-xen Modelltypen dieses ersten Ansatzes außerhalb der Kategorien A und B zu einerÜberanpassung neigen. Gleichzeitig zeigen die Modelle G und H eine bessere Leistungbzgl. des HS im Vergleich zum TS. Dies kann daraus resultieren, dass die Modelle beidiesem ersten Ansatz auf Basis der Leistung bzgl. des HS selektiert wurden.Beide erwähnten Probleme, die mit einer hohen Modellkomplexität einhergehendeÜberanpassung und der Bias bzgl. des HS, sollen daher in Ansatz 2 mit einem erwei-terten Kreuzvalidierungsverfahren gelöst werden (siehe Kap. 4.1.2).4.1.1.5Leistung der ModelleDie in diesem ersten Ansatz generierten Pin 1-spezifischen Modelle aus den Katego-rien A und B nutzen eine Vielzahl an Deskriptoren und zeigen damit ein komplettesOverfitting. Zu erwähnen sind lediglich die Pin 1-unspezifischen Modelle der Katego-rien C, D, E und F, die alle überzeugende Statistiken aufweisen. Dennoch signalisierendie Leistungen der Modelle G und H, dass es auch in diesen Kategorien, v. a. bei denModellen E und F, zu einer gewissen Überanpassung kommt. Somit sind die statisti-schen Werte kritisch zu sehen.92Kapitel 4. Ergebnisse: ModellproteineTabelle 4.3: Statistische Validierungsparameter für die Modelle G (Art lm = lineare Re-gression) und H (Art glm = logistische Regression), sowie deren verwendete Deskriptoren(geordnet nach der Bedeutung für das Modell, mit Angabe der Regressionskoeffizienten fürden skalierten Datensatz). Die Bedeutung basiert auf den t- (Modell G) bzw. z-Werten (Mo-dell H) der einzelnen Koeffizienten, die aus dem Verhältnis des jeweiligen Koeffizienten unddes dazugehörigen Standardfehlers resultieren.Modell G#ArtRMSETSMAETSCCCTSR2TSRMSEHSMAEHSCCCHSR2HSGlm0.340.270.640.470.270.230.760.58DeskriptorKoeffizientpgbrg0.24rdfoh_200ratio-0.11molbur-0.21protprothbonds-0.08rmsdratio-0.06protpeghbonds-0.09radgyrpeg-0.05shellohratio-0.04Modell H#ArtAUROCTSCCRTSAUROCHSCCRHSHglm0.730.710.930.93DeskriptorKoeffizientradgyrpeg1.54pgbrg-13.50rdfoh_200ratio1.81rmsdratio14.24Abbildung 4.5: Auftragung der durch das Modell G vorhergesagten Differenzen der freienFaltungsenergien gegen die experimentellen Werte der (A) TS- und (B) HS-Konjugate.Ungeachtet dessen enthielten die Modelle aus den Kategorien C, D, E und F In-formationen von PBK mit diversen Linkerstrukturen und konnten dennoch teilweisebessere Stabilitätsvorhersagen erbringen als die von Price et al. postulierte, kristall-strukturbasierte Methode. Diese basierte lediglich auf einem einzigen Deskriptor undeiner kleinen Auswahl an 10 Asn(PEG)4-Konjugaten. ADAN-Analysen deuten je-doch auf große Unsicherheiten bzgl. der Vorhersage zu 1SRL.20-1 und damit eine4.1. QSPR-Vorhersagemodelle93limitierte Anwendbarkeit auf andere Proteine hin. Im Vergleich zum Referenzmodell,welches jedem HS-Konjugat den Medianwert des TS von -0.37 kcal/mol zuweisenwürde (RMSEHS = 0.39 kcal/mol), zeigen die erstellten Regressionsmodelle bessereLeistungen. Auch das Referenzmodell zur Klassifikation zeigt mit einem CCRHS-Wertvon 0.67 eine geringere Leistung als die Modelle D1, F und H.4.1.1.6Verwendete DeskriptorenDie Auswahl der Deskriptoren shellohratio, rdfoh_200ratio und rdfoh_400ratio durchdie Modelle aus den Kategorien C und D unterstützen die Hypothese von Price etal., dass PEG vornehmlich die Solvatation von Hydroxylgruppen moduliert [100].Während solche Solvatationseffekte zu entropisch günstigen Wasserverdrängungen aufder Proteinoberfläche führen können, beschreibt der wichtigste Deskriptor von Mo-dell C1, protpeghbonds, die Anzahl an Protein-PEG-HB und damit eher spezifische,enthalpisch günstige Wechselwirkungen. Dieser Parameter zeigt auch eine schwacheKorrelation (Abb. 4.6-A) mit der Antwortvariablen. Ein weiterer HB-Deskriptor, derdie Protein-Protein-Wechselwirkungen beschreibt (protprothbondsratio), zeigt, dassPEG durch eine Abschirmung möglichweise intramolekulare HB stärkt (vgl. [94]).Das Modell E gewichtet hingegen den Solvatationsparameter shelloratio deutlich stär-ker. Weiterhin nutzt es den Deskriptor molbur, mit welchem eine schwache Korrelationder bedeckten Proteinoberfläche mit der Differenz der freien Faltungsenergie festge-stellt werden kann (Abb. 4.6-B). Modell F nutzt den Deskriptor radgyrpeg, was vor-nehmlich auf doppelt PEGylierte Konjugationsstellen zurückzuführen ist, da hierbeialle PEG-Atome zur Berechnung berücksichtigt wurden. Dies spiegelt jedoch auchdie Hypothese von Price et al. wieder, dass eine weitere PEGylierung an einer weitentfernten Position einen synergistischen Effekt haben kann (16+26-1, 19+26-1), wäh-rend bei einer zweiten PEGylierung in unmittelbarer Nähe die Stabilisierung durchbeide Konjugationen eher beeinträchtigt wird (26+29-1, 16+19-1). In letzteren Fällenadressieren beide Ketten ähnliche Areale [100].Abbildung 4.6: Lineare Regressionen ausgewählter Deskriptoren. Dargestellt sind protpegh-bonds (A, R2 = 0.31) und molbur (B, R2 = 0.18).94Kapitel 4. Ergebnisse: Modellproteine4.1.1.7MD-AnalysenZur weiteren Interpretation wurden relevante Deskriptoren aller Modellkategorienin exemplarischen Simulationen näher untersucht. Dadurch soll auch die Eignungdieser berechneten Deskriptoren zur Wiedergabe der von Price et al. postuliertenEffekte einer PEG-basierten Konjugation demonstriert werden. Diese Analysen sollenim Folgenden vorgestellt werden.4.1.1.7.1Interaktionen und Solvatation16-1 (-0.90 ± 0.03 kcal/mol) und 27-1 (0.38 ± 0.04 kcal/mol) stellen stark stabilisierte(16-1) bzw. stark destabilisierte (27-1) Varianten dar. Im Falle von 16-1 wird mehrProteinoberfläche durch PEG bedeckt (molbur = 71.42 ± 3.08 Å2 für 16-1, 42.13± 3.09 Å2 für 27-1) und die Anzahl an Protein-PEG-HB ist entsprechend höher(protpeghbonds = 1.14 ± 0.01 für 16-1, 0.70 ± 0.01 für 27-1).Beides kann in den Simulationen über die Zeit hinweg beobachtet werden (Abb. 4.7und 4.8): Im Falle von 27-1 interagiert der Asparagin-Linker mit E12, während diePEG-Kette ins Solvens hinausragt. Teilweise adressiert diese den N-Terminus oderdas Rückgrat von G10. Am Ende der Simulation ist 27-1 zwischen G10 und I28 posi-tioniert, was zu einer Erhöhung der bedeckten Proteinoberfläche führt. Im Gegensatzhierzu wendet sich das Polymer von 16-1 nur sehr selten von der Oberfläche ab. Eswechselwirkt entweder mit S32 und F34 des dritten β-Faltblattes oder mit Amino-säuren der ersten Schleife (R17, S18, S19, R21).Um die Solvatation der Proteine zu vergleichen, wurden RDF-Differenz-Plots zwi-schen der PEGylierten und der unPEGylierten Mutante erzeugt (Abb. 4.8). EineDesolvatation ist deutlicher bei der stabilisierten Variante 16-1 zu erkennen, v. a. umdie Aminosäuren R17, S19, Y23 und F34. Im Falle des destabilisierten Konjugats27-1 erhöht sich der RDF-Wert um G20. Analoge Plots weiterer stark (de-)stabili-sierter PBK zeigen einen ähnlichen Trend (Abb. D.1). Auch wenn die Solvatation vonG20 maßgeblich durch die Konformation der ersten Schleife bestimmt wird, spielteder Deskriptor glycinflip (φ-Diederwinkel von G20, dessen Änderung eine Umorien-tierung signalisieren kann) keine bedeutende Rolle für die Pin 1-spezifischen Modelle.Eine Umorientierung bzw. ein Flip von G20 wurde in einigen Simulationen visuellbeobachtet (Abb. D.2).Durch diesen beispielhaften Vergleich zwischen 16-1 und 27-1 werden die Unterschiedein der Desolvatation und den Deskriptoren protpeghbonds und molbur aufgezeigt undauf molekularer Ebene veranschaulicht.4.1. QSPR-Vorhersagemodelle95Abbildung 4.7: Momentaufnahmen aus den Simulationen von (A) 16-1 und (B) 27-1. In derersten Momentaufnahme, der Startposition, sind Dichten der PEG-Atome (Isowert: 0.02) fürdie gesamte Simulation mit roten Gittern illustriert. Drei kleinere Darstellungen zeigen jeweilshäufig beobachtete bzw. repräsentative Konformationen des Polymers in den Simulationen.Aminosäuren in der Nähe sind grün dargestellt und gekennzeichnet, die Konjugationsstelleist in violett gezeigt.4.1.1.7.2Linker verschiedener LängeEine weitere Hypothese von Price et al. besagt, dass rigide, planare Linker mit einergewissen Distanz für eine optimale Desolvatation benötigt werden [101]. Asparaginpositioniert PEG durch die geringere Länge näher an die Hydrathüllen des Proteinsals Glutamin. Zum Vergleich wurden die Konjugate 29-1 (-0.36 ± 0.04 kcal/mol)und 29-9 (0.16 ± 0.06 kcal/mol) untersucht. Beide PEG-Ketten sind auf der unterenSeite der β-Faltblätter lokalisiert, aber im Falle von 29-9 weiter weg vom Proteinpositioniert. Dies wird durch den Deskriptor pg4dist deutlich (2.99 ± 0.21 Å für 29-1,3.71 ± 0.24 Å für 29-9). Die Aminosäuren von 29-1 sind stärker desolvatisiert (Abb.4.9), v. a. W11, A31, Q33, P37 und S38.96Kapitel 4. Ergebnisse: ModellproteineAbbildung 4.8: Vergleich zwischen den Simulationen von 16-1 (links) und 27-1 (rechts). Dieerste Reihe zeigt die durch das Polymer bedeckte Proteinoberfläche über die Zeit, währenddie zweite die Anzahl an beobachteten Protein-PEG-HB darstellt. Zusätzlich dargestellt sindjeweils Violin-Plots. Die dazugehörigen Deskriptorwerte und deren Standardfehler sind jeweilsoben rechts dargestellt. Die letzte Reihe zeigt RDF-Differenz-Plots (PEGyliert - unPEGyliert)für jede Aminosäure zwischen 0.2 und 0.8 nm. Blaue Areale (Werte zwischen -0.1 und -2) re-präsentieren eine stärkere Solvatation der PEGylierten Variante, rote eine der unPEGylierten(Werte von 0.1 bis 2).4.1. QSPR-Vorhersagemodelle97Abbildung 4.9: Darstellung von Dichten der PEG-Atome für 29-1 (blau) und 29-9 (rot),analog zu Abb. 4.7. Darunter sind die RDF-Differenz-Plots dargestellt (links: 29-1, rechts:29-9).4.1.1.7.3D-/L-AsparaginEin weiterer Vergleich kann zwischen den Varianten 19-1 (-0.70 ± 0.04 kcal/mol) und19-8 (0.01 ± 0.04 kcal/mol) hergestellt werden. 19-8 verfügt über ein D-Asparagin an-statt eines L-Asparagins als Linker. Price et al. schlussfolgerten, dass die Orientierungder PEG-Kette durch die Änderung der Linkerkonfiguration entscheidend verändertwird: die PEG-Kette sollte bei 19-8 vom Protein abgewandt sein. In den entsprechen-den Simulationen orientiert sich das Polymer von 19-1 in Richtung der Aminosäurenvon Schleife 1, was zu einer Desolvatation um M15, S16 und G20 führt (Abb. 4.10).Im Gegensatz hierzu wechselwirkt PEG bei 19-8 mit S32 und F34, das Polymer istalso nicht von der Proteinoberfläche abgewandt.Vorherige Simulationen von Price et al. deuteten auf eine Stabilisierung der Schlei-fenkonformation im Falle von 19-1 hin [99]. In den hier durchgeführten Simulatio-nen kommt es auch nur im Falle von 19-8 zu einer deutlichen Konformationsän-derung der ersten Schleife. Vielleicht ist diese Änderung auf die Integration der D-Aminosäure zurückzuführen. Allerdings sollte angemerkt werden, dass in Kraftfeldernfür MD-Simulationen (in diesem Falle AMBERff14SB) prinzipiell keine von den L-Konfigurationen abweichenden Parameter für die D-Varianten genutzt werden, sodasshier lediglich die Startkonformation der Konjugationsstelle zu Unterschieden in denSimulationen geführt haben dürfte. Da die Polymerkette im Falle von 19-8 zur Prote-inoberfläche geneigt ist, zeigt diese Variante eine größere bedeckte Proteinoberfläche98Kapitel 4. Ergebnisse: Modellproteine(molbur = 55.04 ± 1.49 Å2 für 19-1, 85.19 ± 6.42 Å2 für 19-8). Diese Beobachtungzeigt, dass eine von Price et al. postulierte Abkehr der PEG-Kette vom Protein durcheine D-Konfiguration des Asparagins nicht bestätigt werden kann. Allerdings ist dieAnzahl an Protein-PEG-HB im Falle von 19-1 tatsächlich höher (protpeghbonds =0.81 ± 0.01 für 19-1, 0.43 ± 0.01 für 19-8), was eine Destabilisierung von 19-8 imVergleich zu 19-1 begründen könnte.Abbildung 4.10: Darstellung der PEG-Dichten für 19-1 (rot) und 19-8 (blau), analog zuAbb. 4.7. Daneben sind die RDF-Differenz-Plots dargestellt (links: 19-1, rechts: 19-8).4.1.1.7.4Linker verschiedener RigiditätZum Vergleich eines rigiden mit einem flexibleren Linker wurden die Trajektorien von19-1 (-0.70 ± 0.04 kcal/mol) und 19-4 (0.24 ± 0.02 kcal/mol) verglichen (Abb. 4.11).19-4 enthält einen sterisch anspruchsvolleren Cystein-Maleimid-Linker. Bei diesemstark destabilisierten Konjugat ist eine deutliche Konformationsänderung der erstenSchleife zu erkennen. M15 orientiert sich hierbei weg von V22, letztere Aminosäure istdadurch stark solvensexponiert. Ein deutlich höherer Wert für rmsfp wurde bestimmt(25.52 Å für 19-1, 93.84 Å für 19-4).Abbildung 4.11: Darstellung der PEG-Dichte für 19-4, analog zu Abb. 4.7. Daneben zeigteine weitere Momentaufnahme die veränderte Ausrichtung von M15. Rechts ist der entspre-chende RDF-Differenz-Plot gezeigt.4.1. QSPR-Vorhersagemodelle994.1.1.7.5Desolvatation apolarer FlächenZuletzt wurde zur Untersuchung der Wasserverdrängung um solvensexponierte apola-re Aminosäuren [103] die Simulation von 23-Z (-0.29 ± 0.03 kcal/mol) mit jenen derbeiden Mutanten 23(E12L:R14L)-Z (-0.89 ± 0.04 kcal/mol) und 23(E12F:R14F)-Z(-0.33 ± 0.02 kcal/mol) verglichen (Abb. 4.12). Letztere besitzen jeweils zwei zu-sätzliche apolare Aminosäuren im Bereich des ersten β-Faltblatts. Gemäß der Hypo-these einer Desolvatation solvensexponierter hydrophober Areale ist die PEG-Ketteim Falle der beiden Mutanten eher zum ersten β-Faltblatt orientiert. Der Parameterrdfapol_200ratio, der die Solvatation in der ersten Hydrathülle um apolare Amino-säuren beschreibt, signalisiert eine Desolvatation dieser Mutanten (0.774 ± 0.002 für23(E12L:R14L)-Z, 0.884 ± 0.003 für 23(E12L:R14F)-Z).Abbildung 4.12: Darstellung der PEG-Dichten und der dazugehörigen RDF-Differenz-Plotsfür 23-Z (links), 23(E12L:R14L)-Z (Mitte) und 23(E12F:R14F)-Z (rechts).4.1.1.8FazitDie Pin 1-unspezifischen Regressionsmodelle C1 und E lieferten zu Price et al. ver-gleichbare Vorhersagen für 1SRL.20-1 und die Klassifikationsmodelle D1 und F zufrie-denstellende statistische Werte, ADAN-Analysen und die Modelle G und H deutenjedoch auf Vorhersagen mit erhöhten Fehlern und mögliche Überanpassungen hin.Die Modelldeskriptoren unterstreichen die Bedeutung von Solvenseffekten durch einePEGylierung, berücksichtigen aber auch enthalpisch günstige Wirkungen, bspw. überdie Anzahl an Protein-PEG-HB. Verschiedene Hypothesen von Price et al. und da-mit zusammenhängende, bedeutende Modelldeskriptoren wurden in exemplarischenTrajektorien verschiedener Konjugate untersucht, was eine Eignung klassischer MD-Simulationen zur Wiedergabe der postulierten Effekte unterstreicht.100Kapitel 4. Ergebnisse: ModellproteineZur weiteren Optimierung und einer Vermeidung von Überanpassungen ist jedochein deutlich größerer Datensatz an Konjugaten zur Nutzung der hier angewandtenMethoden des maschinellen Lernens nötig. Ein solcher Datensatz sollte eine vergleich-bare Anzahl α-helikaler und β-faltblattreicher Strukturen enthalten (zur Etablierunggeneralisierbarer Vorhersagemodelle). Teilweise enthielt der hier genutzte DatensatzLinkerstrukturen, die nur in einem einzelnen Konjugat vorkamen und im HS vertretenwaren (z. B. 19-J, 19-4 oder 19-5). Allerdings würde der Datensatz durch eine Restrik-tion auf Asparagin-Linker-basierte Konjugate deutlich verringert werden. Zusätzlichsollte erwähnt werden, dass der Effekt auf die Stabilität des Konjugats natürlichauch aus einer Veränderung der Linkerstruktur während der PEGylierung resultierenkann. So existieren gewisse Konjugationsstrategien (z. B. bei den Konjugaten mit denKennzeichen Z oder X), die zu einer deutlichen strukturellen Modifizierung führen.So wird nicht nur der Effekt der PEGylierung berücksichtigt.Wie am Anfang dieses Kapitels bereits angemerkt, wurden diese Modelle auf Basiseines einzigen HS selektiert, und es kamen sehr komplexe Modellarten zum Einsatz,die in der Regel nur bei größeren Datensätzen angewandt werden sollten, um eineÜberanpassung aufgrund der Hyperparameter zu vermeiden. Die erhaltenen statisti-schen Werte der beschriebenen Modelle sind daher kritisch zu sehen. Der beschriebeneAnsatz dient jedoch als Grundlage für den angepassten zweiten Ansatz (vide infra).Weiterhin kann er als Grundlage für mögliche zukünftige Arbeiten dienen, auf die diein diesem Kapitel erläuterten Modelltypen (neuronale Netze und XGBoost-Modelle)angewandt werden könnten.4.1.2Ansatz 2Im Folgenden sollen das beste Regressions- und das beste Klassifikationsmodell ausAnsatz 2 vorgestellt werden. In diesem Ansatz wurde ein erweitertes Kreuzvalidie-rungsverfahren genutzt, sodass das zugehörige HS im Gegensatz zu Ansatz 1 keinenEinfluss auf die Modellauswahl hatte und es sich somit um eine externe Validierunghandelt. Weiterhin wurden lediglich multivariate lineare Regressions- und binomialelogistische Klassifikationsmodelle erstellt. Gemäß Kap. 3.1.6.3 wurden hierbei 9 ver-schiedene Datensätze untersucht (siehe Tab. 3.3), die sich in der Art und der Anzahlder Konjugate, sowie der Nutzung Pin 1-spezifischer und -unspezifischer Deskriptorenunterschieden. Die statistischen Parameter der finalen Modelle zu jedem Datensatzsind in den Tab. D.6 und D.7 zu finden. Hier sollen das Regressionsmodell zum Da-tensatz 6 (genannt A2-R6) und das Klassifikationsmodell zu Datensatz 1 (genanntA2-C1) mit den Korrelationslimits von 0.99 vorgestellt werden, die die jeweils bestenErgebnisse erzielten.Die Modelle der anderen Konjugate zeigen eher mäßige Leistungen und sollen hiernicht näher diskutiert werden. Dies ist vermutlich auf die Zusammensetzungen derjeweiligen Datensätze zurückzuführen. Der von Price et al. experimentell ermittelteEffekt von Cyclohexylalanin- und Azidohomoalanin-Mutationen (siehe Kap. 3.1.3.4)4.1. QSPR-Vorhersagemodelle101in den Datensätzen 2 und 3 konnte durch die Simulationen evtl. nicht korrekt ab-gebildet werden. Weiterhin stellen die α-helikalen Strukturen der Doppelwendeln inDatensatz 2 von den WW-Domänen strukturell deutlich abweichende Peptide dar.Hierbei wird der Effekt der PEGylierung auf Di- und Trimere berücksichtigt. DieDatensätze 4 und 5 bestehen wiederum grundsätzlich aus einer sehr geringen Anzahlan Konjugaten. Dies führt unter anderem dazu, dass Regressionsmodelle zu Daten-satz 5 bzgl. des HS mit nur vier Konjugaten zwar recht hohe R2-Werte zeigen (Tab.D.6), diese jedoch lediglich auf jeweils zwei im Hinblick auf ihre Faltungsenergie nahbeieinander liegenden stark stabilisierten und destabilisierten Varianten beruhen. DieLeistungen für das TS sind deutlich schlechter. Für Datensätze 8 und 9 wurden prote-inspezifische Deskriptoren genutzt, welche nicht für generalisierte Modelle verwendetwerden können und auch in Ansatz 1 bereits zu unbrauchbaren Modellen führten.4.1.2.1ReferenzmodelleWie auch in Ansatz 1 sollten zunächst die Leistungen von Referenzmodellen für diejeweiligen Datensätze untersucht werden. Ein Modell, welches für jedes Konjugat desTS von Datensatz 6 den Median- oder den Mittelwert vorhersagen würde, zeigt einenRMSETS-Wert von 0.44 kcal/mol und einen RMSEHS-Wert von 0.53 kcal/mol.Für das Klassifikationsmodell A2-C1 wurden zunächst die Anteile an Konjugaten derstabilisierten Klasse 1 (61 von 72, bzw. 84.72 %) und der un- bzw. destabilisiertenKlasse 2 (11 von 72, bzw. 15.28 %) ermittelt. Bei einer zufälligen Einteilung derTS- und der HS-Konjugate des Datensatzes 1 in eine der beiden Klassen mit dengenannten Wahrscheinlichkeiten würden ein CCRTS-Wert von 0.74, ein BAccTS-Wertvon 0.44, ein CCRHS-Wert von 0.64 und ein BAccHS-Wert von 0.44 erhalten werden.4.1.2.2Regressionsmodell A2-R64.1.2.2.1Allgemeine LeistungDie Leistung des Regressionsmodells A2-R6 ist in Tab. 4.4 für das TS und das HS zu-sammengefasst. Eine Illustration für das TS und das HS ist in Abb. 4.13 dargestellt.Das Modell erzielt mit einem RMSETS-Wert von 0.25 kcal/mol, einem R2TS-Wertvon 0.68 und einem CCCTS-Wert von 0.81 eine gute Leistung für das TS. Der Wertfür τB(TS) ist jedoch mit 0.45 sehr gering. Dies resultiert aus der hohen Anzahl anKonjugaten im Bereich zwischen etwa -0.70 und -0.20 kcal/mol. Die Fähigkeit des Mo-dells, diese Varianten in einer sehr dichten „Punktwolke“ korrekt zu ordnen, ist, wieim ersten Ansatz, sehr schlecht. Dieser enge Bereich repräsentiert einen für QSPR-Modelle nur limitiert zugänglichen Raum. Die Leistung ist für das HS etwas geringer,aber immer noch akzeptabel (RMSEHS-Wert von 0.37 kcal/mol und CCCHS-Wertvon 0.74). Von entscheidender Bedeutung ist, dass das Modell extrem stabilisierteund destabilisierte Varianten, welche aufgrund des bei der Erstellung angewandtenCluster-Ansatzes (siehe Kap. 3.1.6.3) im HS überrepräsentiert sind, korrekt einord-net. Durch deren Überrepräsentation zeigen R2HS und τB(HS) sogar bessere Werte102Kapitel 4. Ergebnisse: Modellproteineals im TS. Sowohl im Hinblick auf das TS (0.25 kcal/mol) als auch das externe Va-lidierungsset (0.37 kcal/mol) zeigt das Modell A2-R6 bessere RMSE-Werte als daszu Beginn beschriebene Referenzmodell (RMSETS = 0.44 kcal/mol, RMSEHS = 0.53kcal/mol).Tabelle 4.4: Statistische Validierungsparameter für das Modell A2-R6, sowie dessen verwen-dete Deskriptoren (geordnet nach der Bedeutung für das Modell, mit Angabe der Regressi-onskoeffizienten für den skalierten Datensatz). Die Bedeutung basiert auf den t-Werten dereinzelnen Koeffizienten, die aus dem Verhältnis des jeweiligen Koeffizienten und des dazuge-hörigen Standardfehlers resultieren.Modell A2-R6RMSETSτB(TS)CCCTSR2TSRMSEHSτB(HS)CCCHSR2HS0.250.450.810.680.370.820.740.76DeskriptorKoeffizientmolbur-0.96rdf_500ratio1.76rdfpeg_500-0.67rdf_700ratio-0.83pgbrg0.20shellpego0.50shellpolratio0.13rdf_400ratio-0.81rdf_200ratio-0.29pegangle-0.08Abbildung 4.13: Auftragung der durch das Modell A2-R6 vorhergesagten Differenzen derfreien Faltungsenergien gegen die experimentellen Werte der (A) TS- und (B) HS-Konjugate.4.1.2.2.2Einzelne KonjugateDie Vorhersagen zu den einzelnen Konjugaten des HS sind in Tab. 4.5 aufgelistet. Dieentsprechenden ADAN-Kategorien sollen weiter unten diskutiert werden. Die drei be-sonders stabilisierten Konjugate 19-2, 19-3 und 1SRL.20-1 (∆∆G0f < -1.00 kcal/mol)werden vom Modell korrekterweise als die stabilisiertesten Varianten im HS einge-stuft. Ebenso wird ein positiver Wert für die destabilisierte Variante 23-9 berechnet.4.1. QSPR-Vorhersagemodelle103Diese Beobachtungen deuten darauf hin, dass das Modell prinzipiell gut geeignetist, deutlich (de-)stabilisierte Konjugate zu detektieren. Schaut man sich die Diffe-renzen zu den experimentellen Werten an (Tab. 4.5), so zeigt sich, dass das Modellden Stabilisierungseffekt der Konjugate generell überbewertet. Dies führt dazu, dassdie Varianten 18-1, 28-1 und 30-1 als schwach stabilisiert eingestuft werden, obwohlderen PEGylierungen auf Basis der experimentellen Befunde keinen Effekt auf diethermische Stabilität ausüben. Unter Berücksichtigung der von der ADAN-Analyseausgegebenen 95%-Konfidenzintervalle und den experimentellen Fehlern stimmen dieVorhersagen des Modells für die Konjugate des HS mit den Messungen überein, dieADAN-Konfidenzintervalle für 28-1, 16-Z und 32-Z decken jedoch auch einen Groß-teil des Energiebereichs, in welchem die experimentellen Befunde zu verorten sind,ab. Im Falle eines kleineren Konfidenzintervalls von ±0.25 kcal/mol gäbe es keineÜbereinstimmung bei 32-Z.Tabelle 4.5: Vorhersagen von Modell A2-R6 zu den einzelnen Konjugaten des HS. Dazuist die Differenz zum experimentell ermittelten Wert aufgelistet. Die Vorhersagen sind mitden 95%-Konfidenzintervallen versehen, die aus der ADAN-Kategorisierung resultieren. Dabeiführen niedrige Kategorien (0 - 1, grün) zu einem geringen Wert (±0.25 kcal/mol) und höhere(2 - 3, rot) zu einem größeren (±0.50 kcal/mol). Die Differenzen sind grün gefärbt, wenn dasADAN-Konfidenzintervall den Bereich des exp. Werts einschließt, andernfalls sind sie rotgefärbt.KonjugatExperimentellerWert [kcal/mol]Vorhersage [kcal/mol]Differenz[kcal/mol]ADAN18-10.00 ± 0.07-0.66 ± 0.25-0.660/628-10.00 ± 0.07-0.26 ± 0.50-0.262/630-10.00 ± 0.07-0.28 ± 0.25-0.280/623-90.23 ± 0.040.03 ± 0.25-0.200/619-3-1.13 ± 0.04-1.09 ± 0.250.040/616-Z0.07 ± 0.05-0.13 ± 0.50-0.203/632-Z-0.33 ± 0.03-0.72 ± 0.50-0.392/619-2-1.19 ± 0.02-1.46 ± 0.25-0.271/623(E12L:R14A)-Z-0.31 ± 0.02-0.69 ± 0.25-0.380/619-50.01 ± 0.02-0.57 ± 0.25-0.581/61SRL.20-1-1.20 ± 0.10-0.85 ± 0.250.350/64.1.2.2.3ApplikationsdomäneGemäß der ADAN-Analyse liegen fast alle Konjugate des HS innerhalb der Appli-kationsdomäne des Modells A2-R6. 3 von 11 HS-Konjugaten (28-1, 16-Z und 32-Z)wurden höheren Kategorien zugeordnet, diese weisen damit größere Fehlerbereicheauf (±0.50 kcal/mol statt ±0.25 kcal/mol). Dies deutet darauf hin, dass das generier-te Modell die Wertebereiche der relevanten Deskriptoren dieser Konjugate prinzipiellabdeckt und somit für eine Vorhersage geeignet sein sollte, trotz des Vorkommensrelativ vieler Konjugate mit extremen ∆∆G0f-Werten im HS. Auch 1SRL.20-1 wirdim Gegensatz zu den Modellen aus Ansatz 1 (Kap. 4.1.2) nicht in eine höhere ADAN-Kategorie eingestuft, was auf eine Anwendbarkeit des Modells auf andere Proteinehinweist.104Kapitel 4. Ergebnisse: Modellproteine4.1.2.2.4Vergleich zur strukturbasierten RichtlinieEin direkter Vergleich mit der strukturbasierten Richtlinie von Price et al. [100]ist schwierig, da diese lediglich auf einer kleinen Anzahl (10) an Konjugaten mitAsparagin-Linkern beruht (16-1, 17-1, 18-1, 19-1, 26-1, 27-1, 28-1, 29-1, 30-1 und32-1). Sie wurde außerdem nur mit einem einzigen zusätzlichen Konjugat getestet(1SRL.20-1). In Tab. 4.6 sollen zunächst die bereits beschriebenen statistischen Wer-te des Modells A2-R6 zum TS und HS (Spalte 1) den korrespondierenden Werten zurRichtlinie von Price et al. mit den dort untersuchten 10 Konjugaten (Spalte 2) gegen-übergestellt werden. Dabei zeigt sich, dass die RMSETS-, CCCTS und R²TS-Werte desModells A2-R6 vergleichbar mit jenen der strukturbasierten Richtlinie sind. Letzterezeigt jedoch eine deutlich bessere Rangordnung (τB). Der von Price et al. postulier-te Winkel wird auf Basis der Kristallstruktur ermittelt. Wird der Mittelwert diesesWinkels in der Simulation des WT der WW-Domäne berechnet, so werden die Wertein Spalte 3 erhalten. Mit Ausnahme eines leicht erniedrigten τB-Werts werden diestatistischen Werte leicht verbessert, was auf verringerte Residuen für 16-1 und 19-1zurückzuführen ist (Regression nicht gezeigt). Dies verdeutlicht, dass die Dynamikder Strukturen hierbei eine wichtige Rolle spielen kann, was prinzipiell für einen MD-basierten Ansatz zur Beschreibung des Polymerverhaltens spricht.Berechnet man θPrice jedoch für alle Konjugate (43), bei denen der WT zur Winkel-messung genutzt werden kann, so zeigt sich, dass die Leistung deutlich abfällt (Spalte4). Diese Berechnung schließt also alle WW-Konjugate des Datensatzes 6 ein, beidenen keine zusätzlichen Mutationen eingefügt wurden und der WT somit die un-PEGylierte Variante ohne Asparagin-Linker darstellt (dies bedeutet: z. B. nicht dieKonjugate 1PIN.19(S16A)-1, 1PIN.19(Y23F)-1 oder 1PIN.23(E12L:R14L)-Z). Wer-den auch entsprechende Simulationen zu WT-Varianten dieser Mutanten (vgl. Kap.3.1.1) zur Winkelmessung genutzt, können die statistischen Werte für den gesamtenDatensatz 6 evaluiert werden. Unter Nutzung dieser mutierten Kristallstrukturen alszusätzliche Referenzen ist die Leistung jedoch noch niedriger (Spalte 5). Zu Price etal. vergleichbare Winkel, bei dem der Vektor b (siehe Kap. 1.1.6.2) nicht zur nächstge-legenen Hydroxylgruppe, sondern zur nächstgelegenen polaren, apolaren oder aroma-tischen Aminosäure gezogen wird, wurden auch berechnet, führten jedoch ebenfallszu keinen vielversprechenden, univariaten Korrelationen.Insgesamt zeigen diese Beobachtungen, dass die strukturbasierte Richtlinie von Priceet al. bei einer kleinen Auswahl an Konjugaten auf Basis eines Asparagin-Linkers mitin der Nähe verorteten Hydroxylgruppen durchaus eine gute Korrelation ergibt, sie je-doch keineswegs auf eine größere Vielfalt an Strukturen anwendbar ist (auch nicht auferweiterte Datensätze, welche lediglich Konjugate mit Asparagin-Linkern enthalten).Das Modell A2-R6 erzielt im Hinblick auf den zugrundeliegenden, größeren Daten-satz mit verschiedenen Linkertypen eine bessere Leistung. Price et al. postuliertenin den nachfolgenden Studien viele zusätzliche Effekte einer PEGylierung, die durch4.1. QSPR-Vorhersagemodelle105Tabelle 4.6: Vergleichende Gegenüberstellung der Modellleistung von A2-R6 mit der vonPrice et al. entwickelten, strukturbasierten Richtlinie. Die Leistung von A2-R6 bzgl. des TSund des HS ist in der ersten Spalte gegeben (HS/TS). Es folgt die Leistung der struktur-basierten Richtlinie, wie sie in der Referenzpublikation [100] (Ref.) aufgeführt ist. In Spalte3 (WT MD) sind die Werte unter Verwendung der aus der WT-Simulation entnommenenWinkel aufgeführt. In Spalte 4 (erweitert) wird die Leistung bei Nutzung aller Konjugate, beidenen die WT-Kristallstruktur als Referenz genutzt werden kann, aufgeführt. In der letztenSpalte (D6) ist schließlich die Leistung für den Datensatz 6 (HS/TS) aufgelistet. Die Anzahlverwendeter Konjugate ist jeweils in Klammern angegeben.ParameterA2-R6 (61/11)Ref. (10)WT MD (10)Erweitert (43)D6 (61/11)RMSE0.25/0.370.220.190.470.43/0.45CCC0.81/0.740.780.850.230.06/0.43τB(TS)0.45/0.820.780.690.230.04/0.21R²0.68/0.760.640.740.230.03/0.27diesen Winkel nicht abgebildet werden (Desolvatation apolarer Flächen [103], Stär-kung intramolekularer Salzbrücken [102] und NH-π-Bindungen [104]). Insofern ist esnicht verwunderlich, dass diese strukturbasierte Richtlinie keine zufriedenstellendenVorhersagen für erweiterte Datensätze liefert. Durch die Verwendung verschiedenerDeskriptoren können im angewandten QSPR-Modellierungsansatz dagegen prinzipiellunterschiedliche Aspekte berücksichtigt werden, was aus den ausgewählten Modell-deskriptoren ersichtlich ist (vide infra).Die einzige zur Validierung zusätzlich von Price et al. untersuchte Struktur auf Basisdieses Winkels stellt 1SRL.20-1 dar. Dieses Konjugat wurde erfolgreich als stabilisiertklassifiziert, der vorhergesagte ∆∆G0f-Wert weicht mit -0.49 kcal/mol jedoch sehrweit vom experimentell bestimmten (-1.20 ± 0.10 kcal/mol) ab. Das Modell A2-R6liefert dagegen eine Vorhersage von -0.85 ± 0.25 kcal/mol und liegt damit näher amMesswert.4.1.2.2.5ModelldeskriptorenDen wichtigsten Deskriptor stellt molbur dar, dieser war bereits für die Modelle inAnsatz 1 sehr bedeutend. Es zeigt sich somit eine Stabilisierung durch eine erhöhteAbdeckung der Proteinoberfläche mit den Polymeratomen. Dieser Deskriptor korre-liert ebenso mit protpeghbonds (Anzahl an Protein-PEG-HB). Univariate Regressionenergaben im Vergleich zu den übrigen Deskriptoren relativ gute R2-Werte (siehe Abb.4.6 für Regressionen aus Ansatz 1). In allen Fällen jedoch sind die Trends maßgeb-lich durch die stark stabilisierten Varianten 16-2, 16+26-1 und 19+26-1 bestimmt.Weitere univariate Regressionen der anderen Modelldeskriptoren sind in Abb. D.3gezeigt. Der Deskriptor pgbrg beschreibt die Anzahl an überbrückten HB zwischenPEG-Atomen, die über dazwischenliegende Wassermoleküle ausgebildet werden. Ent-sprechende Konformationen der PEG-Ketten, bei denen sich diese um ein einzelnesWassermolekül unmittelbar auf der Proteinoberfläche winden, wurden in einigen Si-mulationen beobachtet. Eine erniedrigte Anzahl dieser HB korreliert mit einer grö-ßeren Stabilisierung, was damit erklärt werden könnte, dass das Polymer stattdessenmit dem Protein HB ausbilden kann.106Kapitel 4. Ergebnisse: ModellproteineMit Ausnahme des am wenigsten bedeutenden Deskriptors pegangle beschreiben al-le übrigen Deskriptoren die Solvatation des Konjugats oder der Polymerkette. DieDeskriptoren rdf_200ratio, rdf_400ratio und shellpego beschreiben die Solvatation inder ersten oder zweiten Hydrathülle des Konjugats bzw. der PEG-Kette, rdfpeg_500,rdf_500ratio und rdf_700ratio dagegen berücksichtigen Solvenseffekte von noch wei-ter entfernt gelegenen Regionen. Price et al. stellten die Hypothese auf, dass dieModulation äußerer Hydrathüllen für den stabilisierenden Effekt ausschlaggebendsein könnte [100]. Es ist jedoch anzumerken, dass die Koeffizienten der Deskripto-ren rdf_200ratio, rdf_400ratio, rdfpeg_500 und rdf_700ratio negativ sind und somiteine erhöhte Hydratation favorisiert wird. Die Deskriptoren rdf_XXXratio beschrei-ben jedoch die Solvatation um alle Protein- und Polymeratome herum, im Gegensatzzu den Deskriptoren rdfwopeg_XXXratio. Letztere wurden jedoch größtenteils durchden ersten Filterschritt im Modellgenerierungsansatz entfernt. Im Gegensatz dazuwird rdf_500ratio ein sehr großer Koeffizient zugeordnet. Beim Blick auf die univa-riaten Regressionen in Abb. D.3 zeigen all diese RDF-Deskriptoren einen leichtenTrend hin zu einer Desolvatation bei negativeren ∆∆G0f-Werten. Dies unterstreichtdie Desolvatationseffekte, die von Price et al. beschrieben wurden. Die entsprechendenKorrelationen sind jedoch sehr schwach ausgeprägt.Der Deskriptor shellpolratio zeigt ebenso eine leichte Korrelation der Proteinstabili-sation mit der Desolvatation polarer Aminosäuren, was die durch den Winkel θPricebeschriebene, entropisch günstige Wasserverdrängung wiedergeben könnte. Die Be-deutung dieses Deskriptors für das Modell ist jedoch gering. Der Deskriptor peganglestellt den Winkel zwischen den Massenzentren des Proteins, des Linkers und der Po-lymerkette dar (angelehnt an [35]) und kann somit als Maß dafür dienen, wie sehr dasPolymer zur Proteinoberfläche hin orientiert ist. Mit einem Koeffizienten von -0.08(geringste relative Gewichtung) hat dieser jedoch keine große Bedeutung.Insgesamt werden durch die Modelldeskriptoren direkte Protein-PEG-Interaktionen(molbur) als auch Solvenseffekte über zahlreiche RDF-/Hydrathüllen-Variablen be-schrieben. Letzteres könnte die in der Literatur beschriebenen, entropisch günstigenWasserverdrängungen wiedergeben, der wichtigste Parameter molbur ebenso enthal-pisch günstige Protein-PEG-Interaktionen. Die ausgewählten Variablen bzgl. der Sol-vatation berücksichtigen Effekte in vielen verschiedenen Hydrathüllen. Dies macht esschwierig, die wichtigsten Areale einer Desolvatation genauer zu spezifizieren.Eine exemplarische Untersuchung des wichtigsten Deskriptors molbur wurde bereitsfür die Modelle des ersten Ansatzes anhand eines besonders stabiliserten (16-1) undeines destabilisierten Konjugats (27-1) durchgeführt (siehe Kap. 4.1.1.7).4.1.2.2.6Alternative PolymerartenDie Deskriptoren des Modells A2-R6 sind ebenso für andere Polymerarten berechen-bar. Daher wurden für die in Kap. 4.3 beschriebenen Pin 1-Varianten mit LPG,4.1. QSPR-Vorhersagemodelle107PMeOx und PEtOx analoge Vorhersagen erstellt (siehe Tab. D.8). Für die allermeis-ten Konjugate werden jedoch sehr starke Destabilisationen vorhergesagt (Ausnahmebildet das Konjugat 19-ψ mit -1.25 kcal/mol). Dies legt die Vermutung nahe, dass fürdiese Polymere separate Vorhersagemodelle generiert werden müssen, was erst nachzukünftigen, experimentellen Charakterisierungen erfolgen kann. Vergleicht man dieMittelwerte der Modelldeskriptoren für die PEG-Alternativen mit jenen der Konjuga-te des TS, so zeigt sich, dass die Vorhersage einer Destabilisierung für LPG, PMeOxund PEtOx besonders durch größere Werte von pgbrg bestimmt wird (siehe auch Abb.D.5). Vor allem LPG kann aufgrund der Hydroxymethylseitenketten sehr viel mehrüberbrückende HB ausbilden, weshalb dieser Deskriptor größere Werte zeigt.4.1.2.2.7Modell A2-R6v2Im Anschluss an die Analyse des Modells A2-R6 wurde schließlich die Leistung vonModellen mit einer reduzierten Anzahl an Deskriptoren untersucht. Hierzu wurdensukzessive die für das Modell A2-R6 unbedeutendsten Deskriptoren entfernt und diefinale Modellgenerierung wiederholt. Durch Entfernung der beiden Deskriptoren pe-gangle und rdf_200ratio, die eine sehr geringe Bedeutung für das Modell A2-R6zeigten, wurde das abgewandelte Modell A2-R6v2 erhalten (Tab. 4.7, Abb. 4.14),nur geringfügig schlechtere statistische Werte zeigt, jedoch 8 statt 10 Deskriptorennutzt. Die adjustierten Bestimmtheitsmaße dieses Modells, die die Anzahl verwende-ter Deskriptoren berücksichtigen, sind dennoch niedriger als die des Modells A2-R6(R²adj(TS) = 0.67 und R²adj(HS) = 0.73 für Modell A2-R6; R²adj(TS) = 0.63 undR²adj(HS) = 0.63 für Modell A2-R6v2).Tabelle 4.7: Statistische Validierungsparameter für das Modell A2-R6v2, sowie dessen ver-wendete Deskriptoren (geordnet nach der Bedeutung für das Modell, mit Angabe der Regres-sionskoeffizienten für den skalierten Datensatz). Die Bedeutung basiert auf den t-Werten dereinzelnen Koeffizienten, die aus dem Verhältnis des jeweiligen Koeffizienten und des dazuge-hörigen Standardfehlers resultieren.Modell A2-R6v2RMSETSτB(TS)CCCTSR2TSRMSEHSτB(HS)CCCHSR2HS0.260.400.780.640.400.670.690.66DeskriptorKoeffizientmolbur-0.94rdfpeg_500-0.65rdf_500ratio1.37pgbrg0.20rdf_700ratio-0.61shellpego0.52rdf_400ratio-0.91shellpolratio0.08108Kapitel 4. Ergebnisse: ModellproteineAbbildung 4.14: Auftragung der durch das Modell A2-R6v2 vorhergesagten Differenzen derfreien Faltungsenergien gegen die experimentellen Werte der (A) TS- und (B) HS-Konjugate.4.1.2.2.8FazitDie besten Modelle für jeden Datensatz wurden in Ansatz 2 im Gegensatz zu An-satz 1 in Kap. 4.1.1 nicht auf Basis der Leistung bzgl. des externen HS, sondern imHinblick auf die aus dem TS gebildeten Testsets während des Generierungsprozessesausgewählt. Mithilfe dieser erweiterten Kreuzvalidierung fungieren die HS in diesemAnsatz als externe Validierungssets, gegenüber jenen die Modelle unvoreingenommensind. Dies spiegelt sich auch in den, je nach Datensatz, geringen Leistungen gemäßTab. D.6 und D.7 wieder. Modell A2-R6 konnte die beste Leistung für das HS erzie-len, wobei die Leistung für das TS vergleichbar mit den Modellen anderer Datensätzeist.Modell A2-R6 konnte besonders stabilisierte und destabilisierte Konjugate des ent-sprechenden HS erfolgreich einordnen, hat jedoch Probleme, die Rangfolge von Kon-jugaten in einem engen Bereich zwischen etwa -0.70 bis -0.20 kcal/mol korrekt wieder-zugeben. Es erzielt für 1SRL.20-1 eine bessere Vorhersage als die publizierte, kristall-strukturbasierte Richtlinie von Price et al. [100]. Im Vergleich zu dieser schneidet dasModell bzgl. des untersuchten Datensatzes auch insgesamt deutlich besser ab (vgl.Tab. 4.6).Wie auch beim Modell E in Ansatz 1 sind hierbei sowohl Modulationen der Hydrat-hüllen als auch direkte Protein-PEG-Interaktionen bedeutend. Letztere werden durchden Deskriptor molbur wiedergegeben, der auch für das Modell E bedeutend ist. Wiejedoch auch schon für die Modelle in Ansatz 1 erwähnt zeigen sich hier einige Limi-tierungen. Eine breitere Verteilung an ∆∆G0f-Werten und größere Datensätze sindmit einer vergleichbaren Verteilung an Linker- und Proteinstrukturen notwendig, umeine verlässliche Modellierung zu gewährleisten.4.1. QSPR-Vorhersagemodelle1094.1.2.3Klassifikationsmodell A2-C1Zur Unterscheidung stabilisierter und destabilisierter Varianten bietet sich alternativein klassifikationsbasierter Ansatz an. Hierbei stellt sich natürlich die Frage nacheinem geeigneten Schwellenwert zur Einteilung, da die meisten Konjugate eine leichteStabilisierung aufweisen. Während die Konjugate im ersten Modellgenerierungsansatzin stark stabilisierte (< 0.40 kcal/mol) und schwach stabilisierte / unbeeinflusste /destabilisierte (≥ 0.40 kcal/mol) Varianten eingeteilt wurden, wurde für den zweitenAnsatz die Trennung bei 0 kcal/mol vorgenommen (stabilisiert vs. unbeeinflusst /destabilisiert). Hier soll das Klassifikationsmodell für den ersten Datensatz, genanntA2-C1, näher beschrieben werden.4.1.2.3.1Allgemeine LeistungDas Klassifikationsmodell A2-C1 erreicht eine gewichtete Genauigkeit BAccTS von0.84 und einen AUCTS-Wert von 0.93 (siehe Tab. 4.8). Auch bzgl. des HS sind dieWerte besser als das anfangs vorgestellte Referenzmodell: BAccHS beträgt 0.77 undAUCHS 0.83. ROC-Kurven sind in Abb. 4.15 dargestellt.Tabelle 4.8: Statistische Validierungsparameter für das Modell A2-C1, sowie dessen verwen-dete Deskriptoren (geordnet nach der Bedeutung für das Modell, mit Angabe der Regressi-onskoeffizienten für den skalierten Datensatz). Die Bedeutung basiert auf den z-Werten dereinzelnen Koeffizienten, die aus dem Verhältnis des jeweiligen Koeffizienten und des dazuge-hörigen Standardfehlers resultieren.Modell A2-C1AUROCTSAUPRCTSCCRTSBAccTSAUROCHSAUPRCHSCCRHSBAccHS0.930.960.930.840.830.810.810.77DeskriptorKoeffizientpegwathbonds-5.16rdfpeg_500-13.69rdfpeg_400-14.17rdfwopeg_400ratio-2.07molbur5.05rdfpeg_2002.99rdfpeg_3002.09Abbildung 4.15: ROC-Kurven des Modells A2-C1 für das TS (blau) und das HS (rot).110Kapitel 4. Ergebnisse: Modellproteine4.1.2.3.2Einzelne KonjugateDie Wahrscheinlichkeit, der Klasse der stabilisierten Konjugate zugewiesen zu werden,ist in Tab. 4.9 wiederum für jedes Konjugat des HS separat aufgelistet. Es lässt sich er-kennen, dass die meisten Wahrscheinlichkeiten nahe oder genau bei 100 % liegen undes wenige Fälle nahe des Schwellenwerts von 0.50 gibt. Den deutlich destabillisiertenKonjugaten des HS (27-9 und 27-1) werden jedoch korrekterweise Wahrscheinlich-keiten von unter 50 % zugewiesen. Das Konjugat 16-Z liegt, unter Berücksichtigungder Fehlergrenze, sehr nahe an der Grenze um 0 kcal/mol. Diesem wird jedoch ei-ne Wahrscheinlichkeit von nahezu 100 % zugewiesen. Falsch ist ebenso die Zuweisungdes Konjugats 17-1 zur destabilisierten Klasse. Diese Variante zeigt auch lediglich einerelativ schwache Stabilisierung von -0.18 ± 0.05 kcal/mol. Das Konjugat 1SRL.20-1wird aber korrekterweise als stabilisiert eingestuft.Tabelle 4.9: Voraussagen (Wahrscheinlichkeiten, der Klasse der stabilisierten Konjugatezugewiesen zu werden) bzgl. des HS für jedes einzelne Konjugat. Dabei ist in Klammernangegeben, ob das Konjugat auf Basis dieser Werte als stabilisiert (S) oder nicht stabilisiert(N) einzustufen ist.KonjugatExperimenteller Wert[kcal/mol]Vorhersage17-1-0.18 ± 0.05 (S)0.26 (N)27-10.38 ± 0.04 (N)0.01 (N)16(S32A)-1-0.58 ± 0.03 (S)0.96 (S)27-90.27 ± 0.02 (N)0.43 (N)16-Z0.07 ± 0.05 (N)1.00 (S)19-2-1.19 ± 0.02 (S)1.00 (S)23(E12L:R14L)-Z-0.89 ± 0.04 (S)1.00 (S)23(E12L:R14A)-Z-0.31 ± 0.02 (S)1.00 (S)19(S16F)-1-1.13 ± 0.07 (S)1.00 (S)16+26-1-1.38 ± 0.03 (S)1.00 (S)1SRL.20-1-1.20 ± 0.10 (S)0.94 (S)4.1.2.3.3ModelldeskriptorenDer wichtigste Deskriptor ist pegwathbonds. Eine geringere Anzahl an HB des Poly-mers zum Solvens korreliert somit mit einer größeren Wahrscheinlichkeit einer Stabi-lisierung. Ist diese Anzahl geringer, so ist zu erwarten, dass das Polymer stattdessenmit dem Protein wechselwirkt. Ebenso ist abermals der Deskriptor molbur in derAuswahl der Variablen vorhanden.Interessanterweise beschreiben vier der fünf übrigen Deskriptoren die Solvatation derPEG-Kette in verschiedenen Hydrathüllen. Wie auch im Falle der Deskriptoren vonModell A2-R6, die eine Modulation der Solvatation beschreiben, zeigen sich hier teil-weise gegensätzliche Koeffizienten. Dem Deskriptor rdfpeg_400 wird ein negativerKoeffizient zugewiesen, den Deskriptoren rdfpeg_200, rdfpeg_300, rdfpeg_500 ein po-sitiver. Zusätzlich zu Variablen, die die Solvatation des Polymers beschreiben, be-schreibt der Deskriptor rdfwopeg_400ratio das Verhältnis der Solvatation des Prote-inanteils der Struktur zwischen PEGylierter und unPEGylierter Variante. Diese ist4.1. QSPR-Vorhersagemodelle111im Falle der stabilisierten Varianten niedriger als bei destabilisierten Konjugaten. Ei-ne Darstellung der univariaten Regressionen ist in Abb. D.4 gezeigt, wobei sich keineeindeutigen Trends erkennen lassen.4.1.2.3.4Alternative PolymerartenWie auch schon für das Regressionsmodell A2-R6 vorgestellt können Vorhersagen fürdie alternativen Polymere erhalten werden. Diese sind in Tab D.9 gezeigt. Ähnlich zuModell A2-R6 werden die allermeisten Konjugate hierbei als destabilisiert eingestuft.Lediglich die Konjugate 17-λ, 19-φ und 23-φ werden als stabilisiert klassifiziert. Dievom Modell ausgegebenen Wahrscheinlichkeiten, der Klasse an stabilisierten Konju-gaten zugewiesen zu werden, liegen zumeist nahe oder genau bei 0 % oder 100 %, mitAusnahme von 17-λ (67 %) und 32-λ (43 %). Wie bei Modell A2-R6 zeigt ein Blick aufdie Verhältnisse zwischen den Mittelwerten der PEG-Alternativen und der Konjugatedes TS, weshalb die meisten LPG-/POx-Konjugate als destabilisiert eingestuft wer-den (siehe Abb. D.6). Diese PEG-Alternativen weisen deutlich mehr HB zum Solvensauf, wodurch größere Werte für den Deskriptor pegwathbonds erhalten werden. Hierwäre evtl. eine Normierung dieses Deskriptowerts auf die Polymeroberfläche geeignet(ähnlich zu den Analysen in Kap. 4.3).4.1.2.3.5FazitDas Klassifikationsmodell A2-C1 zeigt eine gute Leistung bzgl. des TS und des HS.Der Deskriptor molbur ist abermals einer der bedeutendsten. Im Hinblick auf die Mo-dulation der Hydratation spielen jedoch, im Gegensatz zu Modell A2-R6, Deskripto-ren, die die Solvenshülle des Polymers beschreiben, eine größere Rolle. Das von Priceet al. zur Validierung genutzte Konjugat (1SRL.20-1) wird korrekterweise als stabili-siert eingestuft. Die deutlich destabilisierten Konjugate 27-1 und 27-9 wurden ebensoerfolgreich klassifiziert.4.1.2.4Zusätzliche Modelle4.1.2.4.1ParametervergleichDa zur Simulation der unterschiedlichen Konjugate letztlich drei verschiedene Para-metrisierungsmethoden zum Einsatz kamen (PyRED, antechamber, monomerbasiert,siehe Kap. 3.1.3), stellt sich die Frage, ob diese zu einer unterschiedlichen Dynamik desPolymers führten und damit die Werte der verwendeten Deskriptoren beeinflussten.Zur Untersuchung dessen wurden diese Parametrisierungen exemplarisch am Beispielder Konjugationsstelle 16, ausgehend von einer identischen Startkonformation desPEG-Tetramer-Konjugats, verglichen (Simulationsprotokoll analog zu Kap. 3.1.5).Die per PyRED parametrisierte Struktur, deren Simulation in dieser Arbeit zur Ge-nerierung von Vorhersagemodellen diente, wurde hierfür ein weiteres Mal simuliert,um ebenfalls die Fluktuationen zwischen verschiedenen Replika mit identischen Pa-rametern zu untersuchen.112Kapitel 4. Ergebnisse: ModellproteineAbb. 4.16-A und 4.16-B zeigen, dass die Dynamik der Tetramere zum großen Teil ver-gleichbar ist, die Polymerdichte der antechamber-Variante ist identisch mit jener derPyRED-Referenzsimulation. Lediglich die monomerbasierte Parametrisierung scheinteine geringere Tendenz aufzuzeigen, die Aminosäuren der ersten Schleife im Bereichum S18 - G20 zu adressieren. Dagegen ist eine etwas größere Dichte im Bereich desdritten β-Faltblatts um S32 zu finden.Zusätzlich wurden relevante Pin 1-unspezifische Deskriptoren aus Ansatz 2 berech-net. Abb. 4.16-C verdeutlicht, dass die Unterschiede zwischen der PyRED- und derantechamber-Methode minimal sind und sich im Bereich der Standardfehler der De-skriptoren und / oder der Unterschiede zwischen zwei PyRED-Replika befinden. KeinDeskriptor zeigt eine Abweichung von ≥25 % in Relation zum Wert der PyRED-Referenztrajektorie. Im Gegensatz hierzu zeigt die Simulation mit monomerbasiertenParametern für einzelne Deskriptoren etwas stärker abweichende Werte. Hier findensich insgesamt vier Parameter (vgl. Namen in Tab. D.1) mit Abweichungen von über25 %: intoh (134 %), pegwathbonds (152 %), rmsfp (155 %) und pgbrg (243 %). Diemonomerbasierten Parameter wurden allerdings hauptsächlich für die größeren Syste-me in Kap. 5 verwendet. Im Hinblick auf die generierten Vorhersagemodelle wurdensie für die Konjugate 19-B (-0.34 ± 0.04 kcal/mol), 19-J (-0.37 ± 0.05 kcal/mol),19-O (-0.72 ± 0.07 kcal/mol) und 19-U (-0.61 ± 0.04 kcal/mol) genutzt. Es handeltsich somit um einen sehr geringen Anteil an Konjugaten und Deskriptoren des gesam-ten Datensatzes und diese Varianten stellen keine besonders stark (de-)stabilisiertenMutanten dar, sodass kein entscheidender Einfluss der Deskriptorwerte dieser vierKonjugate bei der Modellgenerierung anzunehmen ist.Zur weiteren Untersuchung wurden diese vier genannten Konjugate im Anschluss auchnochmal mit der PyRED-Methode neu parametrisiert und simuliert. Im Zuge einererneuten Deskriptorberechnung wurden ebenfalls die in Kap. 4.1.2.2 zu Modell A2-R6beschriebenen, verschiedenen Abwandlungen des Price-Winkels (zu nächstgelegenen(a)polaren oder aromatischen Aminosäuren), die in univariaten Regressionen keineüberzeugenden Korrelationen zeigten, als zusätzliche Deskriptoren in diesen aktuali-sierten Datensatz integriert (vgl. Tab. D.1). Mit letzterem wurden dann gemäß Ansatz2 nochmal neue Modelle generiert. Für den Datensatz 1 gemäß Tab. 3.3 konnten zuden Modellen A2-R6 und A2-C1 vergleichbar gute Regressions- als auch Klassifikati-onsmodelle erhalten werden. Diese wurden als A2v2-R1 und A2v2-C1 bezeichnet undsollen im Folgenden kurz beschrieben werden.4.1. QSPR-Vorhersagemodelle113Abbildung 4.16: (A) Illustration der Polymerschweratome (alle 5 ns) und des Proteinrück-grats (alle 20 ns) für vier verschiedene Simulationen des Konjugats 16-1: per Parametrisie-rung über die PyRED-Methode (dunkelblau, Referenzsimulation), per Parametrisierung überdie antechamber-Methode (rot), per fragmentbasierter Parametrisierung mittels antechamber(grün), sowie ein weiteres Replikon der mit PyRED parametrisierten Struktur (cyan). (B)Dichten der Polymerschweratome (Durchschnitt aller Momentaufnahmen, Isowert: 0.04). (C)Prozentuale Abweichung der Deskriptorwerte (relevante Deskriptoren des Modells A2-R6, sie-he Kap. 4.1.2.2) vom Wert, der für die Referenzsimulation berechnet wurde. Dabei sind dieMittelwerte der vier verschiedenen Simulationen als Punkte gezeigt, die dazugehörigen Stan-dardfehler sind durch entsprechend eingefärbte Bereiche gekennzeichnet.4.1.2.4.2Modell A2v2-R1Die statistischen Werte von Modell A2v2-R1 sind in Tab. 4.10 aufgeführt und Re-gressionen sind in Abb. 4.17 gezeigt. Für das Konjugat 1SRL.20-1 wird ein Wertvon -0.98 ± 0.22 kcal/mol berechnet (lediglich 27-1, 19(S16F)-1 und 16+26-1 werdenin höhere ADAN-Kategorien eingestuft). Interessanterweise werden vier verschiede-ne Price-Winkel als Deskriptoren verwendet. Die Koeffizienten von price1 und price3(Varianten des ursprünglichen Winkels) deuten, ähnlich wie bei vorherigen Model-len, auf gegeneinander ausgleichende Effekte hin. Dabei können die Koeffizienten derDeskriptoren price3pol und price2arom auf eine Orientierung des Polymers weg vonpolaren hin zu apolaren Aminosäureseitenketten gedeutet werden, was wiederum dieHypothese einer Desolvatation apolarer Proteinoberflächen unterstützt [103]. Der Ko-effizient des zusätzlichen Deskriptors pegwathbonds deutet, ähnlich wie zuvor molbur,abermals auf eine Stabilisierung durch eine Bedeckung des Proteins durch das Polymerhin. Während univariate Regressionen der Price-Winkel also keine guten Korrelatio-nen zeigten, führte eine Kombination zu einem relativ guten multivariaten Modell.114Kapitel 4. Ergebnisse: ModellproteineTabelle 4.10: Statistische Validierungsparameter für das Modell A2v2-R1, sowie dessen ver-wendete Deskriptoren (geordnet nach der Bedeutung für das Modell, mit Angabe der Regres-sionskoeffizienten für den skalierten Datensatz). Die Bedeutung basiert auf den t-Werten dereinzelnen Koeffizienten, die aus dem Verhältnis des jeweiligen Koeffizienten und des dazuge-hörigen Standardfehlers resultieren.Modell A2v2-R1RMSETSτB(TS)CCCTSR2TSRMSEHSτB(HS)CCCHSR2HS0.220.580.840.720.350.750.850.77DeskriptorKoeffizientprice3pol-0.33price2arom0.34pegwathbonds0.35price30.44totalint0.22shellpego-0.30price1-0.33rdfoh_300ratio-0.11Abbildung 4.17: Auftragung der durch das Modell A2v2-R1 vorhergesagten Differenzen derfreien Faltungsenergien gegen die experimentellen Werte der (A) TS- und (B) HS-Konjugate.4.1.2.4.3Modell A2v2-C1Die Statistiken und Deskriptoren zu Modell A2v2-C1 sind in Tab. 4.11 und die ROC-Kurven in Abb. 4.18 gezeigt. Auch in diesem Falle kommt der Winkel price3pol zurAnwendung, zusätzlich zu verschiedenen RDF-Deskriptoren und, wie bei Modell A2-R1, dem Deskriptor pegwathbonds. Für 1SRL.20-1 wird eine Wahrscheinlichkeit von99 % der Zugehörigkeit zur Klasse stabilisierter Konjugate vorausgesagt.4.1. QSPR-Vorhersagemodelle115Tabelle 4.11: Statistische Validierungsparameter für das Modell A2v2-C1, sowie dessen ver-wendete Deskriptoren (geordnet nach der Bedeutung für das Modell, mit Angabe der Regres-sionskoeffizienten für den skalierten Datensatz). Die Bedeutung basiert auf den z-Werten dereinzelnen Koeffizienten, die aus dem Verhältnis des jeweiligen Koeffizienten und des dazuge-hörigen Standardfehlers resultieren.Modell A2v2-C1AUROCTSAUPRCTSCCRTSBAccTSAUROCHSAUPRCHSCCRHSBAccHS0.910.950.930.800.880.830.820.67DeskriptorKoeffizientpegwathbonds-2.15price3pol-1.36rdfpeg_2001.26rdfpeg_300-14.86rdfpeg_40014.23rdfwopeg_400ratio-2.52Abbildung 4.18: ROC-Kurven des Modells A2v2-C1 für das TS (blau) und das HS (rot).4.1.2.5FazitDer abgewandelte Modellgenerierungsansatz 2 führte zu den finalen Modellen A2-R6,A2-C1, A2v2-R1 und A2v2-C1. Auch wenn diese auf den ersten Blick leicht ver-schlechterte statistische Werte im Vergleich zu den komplexeren Modellen E und Fdes ersten Ansatzes zeigen, so basieren sie auf für kleinere Datensätze besser geeig-neten Modellarten und wurden mittels eines erweiterten Kreuzvalidierungsverfahrensund besseren Holdout-Sets genauer auf ihre prädiktive Leistung in Bezug auf externeKonjugate hin untersucht. Abermals wurden v. a. molbur sowie viele verschiedeneSolvatationsdeskriptoren zur Modellierung genutzt.Alle vier genannten Modelle des Ansatzes 2 stuften die von Price et al. zur Vali-dierung genutzte Src SH3-Domäne als stabilisiert ein, die Regressionsmodelle A2-R6und A2v2-R1 zeigten zudem im Vergleich zur strukturbasierten Richtlinie verbesserteVorhersagen. Eine bessere Leistung des Modells A2-R6 gegenüber letzterer Richtliniewurde durch analoge Winkelmessungen für den gesamten Datensatz 6 verdeutlicht.116Kapitel 4. Ergebnisse: ModellproteineAus diesen Ergebnissen lässt sich schließen, dass die von Price et al. publizierte Richt-linie lediglich für einen kleinen Teil der Asn-Konjugate geeignet ist, die in dieser Ar-beit erstellten Modelle jedoch bessere Leistungen für deutlich erweiterte Datensätzezeigen. Vergleichbare Leistungen zwischen den Modellen A2-R6 und A2v2-R1 deu-ten darauf hin, dass einerseits die monomerbasierte Parametrisierung einiger wenigerKonjugate für das Modell A2-R6 keinen entscheidenden Einfluss auf die Leistungder finalen Modelle zeigte, und dass andererseits eine Kombination verschiedener Ab-wandlungen des Price-Winkels zu einem recht guten Modell führen kann. Durch solcheAbwandlungen (bspw. price2arom) können von Price et al. in späteren Publikatio-nen erwähnte Hypothesen abgebildet werden (so z. B. eine PEG-Orientierung hin zuapolaren Bereichen). Für die PEG-Alternativen LPG, PMeOx und PEtOx könntendurch weitere Messungen analoge Modelle erstellt werden.Für die zusätzlich simulierten WW-Konjugate, zu denen von Price et al. noch keineexperimentellen Charakterisierungen vorgenommen wurden (siehe Kap. 3.1.2.2), sinddie Vorhersagen der Modelle aus Ansatz 2 in Tab. 4.12 gezeigt. Zukünftige experi-mentelle Charakterisierungen dieser Varianten könnten zur weiteren Evaluation allerin dieser Arbeit generierten Modelle herangezogen werden.Tabelle 4.12: Vorhersagen der Modelle A2-R6, A2-C1, A2v2-R1 und A2v2-C1 zu den zusätz-lich simulierten, von Price et al. nicht experimentell charakterisierten Konjugaten. Im Falleder Regressionsmodelle sind die Vorhersagen mit den ADAN-Konfidenzintervallen aufgelistet(NA = außerhalb der AD), für die Klassifikationsmodelle sind die Klassifizierungen mit derWahrscheinlichkeit zur Zuordnung zur Klasse der stabilisierten Konjugate gezeigt.KonjugatA2-R6 [kcal/mol]A2-C1A2v2-R1 [kcal/mol]A2v2-C17-1-1.01 ± 0.25S (1.00)NAS (1.00)11-1-0.93 ± 0.25S (1.00)-0.92 ± 0.22S (0.98)12-1-0.29 ± 0.25S (0.96)-0.38 ± 0.22S (0.99)13-1-0.24 ± 0.25S (0.84)-0.89 ± 0.22S (0.69)15-1-0.31 ± 0.25N (0.35)-0.17 ± 0.22S (0.67)21-1-0.33 ± 0.25S (0.71)-0.65 ± 0.22S (0.87)22-1-1.10 ± 0.25S (1.00)-0.84 ± 0.22S (1.00)24-1-0.77 ± 0.25N (0.01)1.07 ± 0.44N (0.03)25-1-0.79 ± 0.25S (0.99)-0.62 ± 0.22S (0.99)31-1-0.40 ± 0.25S (0.37)-0.38 ± 0.44S (0.98)33-1-0.15 ± 0.25N (0.24)0.29 ± 0.22S (0.81)35-1-0.77 ± 0.25S (0.99)-0.64 ± 0.22S (0.98)4.2. Metadynamik-Simulationen1174.2Metadynamik-SimulationenWohltemperierte Metadynamik-Simulationen mit jeweils acht „Walkern“, bei denenDistanzen zwischen den β-Faltblättern als CV dienten, wurden für die exemplarischenKonjugate 16-1 (-0.90 ± 0.03 kcal/mol) und 27-1 (0.38 ± 0.04 kcal/mol) sowie derenunkonjugierte Varianten durchgeführt. Hierdurch sollte untersucht werden, ob dervon Price et al. gemessene energetische Unterschied in silico durch eine (partielle)Entfaltung wiedergegeben werden kann. CV 1 stellt dabei die Entfernung zwischenden β-Faltblättern 1 und 2 dar, CV 2 dagegen jene zwischen den β-Faltblättern 2und 3.4.2.1EnergiediagrammeIn Abb. 4.19 ist die freie Energie als Funktion der beiden CV für alle vier untersuchtenSysteme dargestellt. Dabei sind die Energieminima eingezeichnet, die durch das R-Paket metadynminer erkannt wurden. Das 2D-Energiedagramm wird standardmäßigin 8 x 8 Bins unterteilt und in jedem dieser Abschnitte wird das Energieminimumbestimmt. Alle gefundenen Minima werden dann nach ihrer Energie sortiert undmit Buchstaben gekennzeichnet. Die Anzahl an Bins zur Unterteilung ist prinzipiellveränderbar.Diese Diagramme zeigen zunächst, dass eine volle Entfaltung entlang der CV 2 für 16-1 wahrscheinlicher als im Falle der unkonjugierten Variante ist (vgl. Minimum C von16 mit Minimum F von 16-1). Dagegen zeigt sich bei dem destabilisierten Konjugat27-1 eine höhere Entfaltungstendenz im Vergleich zur unkonjugierten Mutante (vgl.fehlendes Minimum für 27 mit Minimum D von 27-1). Im Falle des Konjugats 27-1 istebenso eine größere Entfaltungstendenz entlang CV 1 zu beobachten (Minimum E von27-1). Allerdings lässt sich dies auch beim stabilisierten Konjugat 16-1 beobachten(Minimum D von 16-1). Wie in Kap. 1.5.1.1 erwähnt ist eine Entfaltung entlangder CV 2 gemäß [255] für eine strukturell ähnliche WW-Domäne wahrscheinlicherals entlang der CV 1, da ein Übergang entlang CV 2 eine geringere Energiebarrieredarstellt. Es kann daher angenommen werden, dass die CV 2 initiale Vorgänge einerkompletten Proteinentfaltung beschreibt. In keiner der Energiediagramme ist einekomplette Entfaltung zu beobachten.4.2.2KonvergenzIm Zuge einer Berechnung energetischer Differenzen sollte die Konvergenz der Sys-teme evaluiert werden, da die oben beschriebenen Unterschiede in den erkanntenMinima von einem unzureichenden Sampling herrühren können. Für Metadynamik-Simulationen kann hierzu die Energiedifferenz der registrierten Minima zum globalenMinimum über die Simulationszeit hinweg untersucht werden. Ändert sich diese Diffe-renz gegen Ende der Stichprobennahme nicht mehr deutlich, so kann eine Konvergenzangenommen werden.118Kapitel 4. Ergebnisse: ModellproteineAbbildung 4.19: Energiediagramme für die Simulationen von 16, 16-1, 27 und 27-1. Diedurch das metadynminer-Paket erfassten Energieminima sind durch Buchstaben gekennzeich-net. Zusätzlich sind die weiter unten beschriebenen Übergänge als rote Linien eingezeichnet(siehe Kap. 4.2.3).4.2. Metadynamik-Simulationen119In Abb. 4.20 sind die Energiedifferenzen der Minima über die Zeit hinweg illustriert.Einige detektierte Minima zeigen gegen Ende der Simulationen weiterhin große Fluk-tuationen, so z. B. Minimum E der Variante 16 oder Minimum F der Variante 16-1.Für einen energetischen Vergleich (vide infra) wurde eine partielle Entfaltung entlangder CV 2 (bis 1.5 - 1.6 Å) untersucht, da diese Minima bzgl. der Energiedifferenzengegen Ende relativ geringe Fluktuationen zeigen.Abbildung 4.20: Verlauf der energetischen Differenzen zwischen den registrierten Minimaund den globalen Minima über die Simulationszeit hinweg.4.2.3EnergiedifferenzenEs stellt sich die Frage, ob die von Price et al. erkannten (De-)stabilisierungen überdie registrierten Energieminima wiedergegeben werden können. Eine vollständige Ent-faltung wurde nicht beobachtet, und die Energielandschaften zeigen teils unterschied-liche Minima, was einen Vergleich erschwert. In allen vier Simulationen jedoch lässtsich ein Energieminimum im Hinblick auf CV 2 zwischen 1.5 und 1.6 Å erkennen.120Kapitel 4. Ergebnisse: ModellproteineIn Tab. 4.13 sind daher die Differenzen dieser Minima zum jeweils globalen Mini-mum aufgeführt. Die dazu angegebenen ∆∆G-Werte vergleichen die energetischenUnterschiede der Differenzen in der unPEGylierten und der PEGylierten Variante. InÜbereinstimmung mit einer Destabilisierung zeigt sich im Falle des Konjugats 27-1 eingeringerer energetischer Unterschied des untersuchten Übergangszustands zur nativenStruktur im Vergleich zur unPEGylierten Variante. Das stabilisierte Konjugat 16-1zeigt dagegen eine leicht höhere Differenz. Die angegebenen Fehler basieren auf einemVergleich der minimalen und maximalen Energiedifferenzen zum globalen Minimumin den letzten 10, 15 und 20 ns der „Walker“. Unter Berücksichtigung dieser letztenNanosekunden lassen sich relativ stabile Energiewerte mit geringen Fehlerbereichenermitteln.Tabelle 4.13: Berechnung der energetischen Differenzen der Übergangszustände zum glo-balen Minimum (∆G). Der angegebene Fehler stellt die Differenz des größten Energieunter-schieds zum niedrigsten Unterschied innerhalb der letzten X ns dar (für X = 10 / 15 / 20 ns).Der Wert ∆∆G stellt den Unterschied dieser Differenzen dar (EnergieunterschiedunPEGyliert- EnergieunterschiedPEGyliert). Dazugehörige Fehler wurden über das Gaußsche Fehlerfort-pflanzungsgesetz ermittelt.VarianteÜbergang∆G[kcal/mol]Fehler [kcal/mol]∆∆G[kcal/mol]Fehler [kcal/mol]16A ->D9.960.06 / 0.10 / 0.20-0.110.07 / 0.16 / 0.2616-1A ->E10.070.04 / 0.12 / 0.1627A ->C10.850.01 / 0.06 / 0.111.020.13 / 0.19 / 0.2127-1A ->B9.830.13 / 0.18 / 0.18Die berechneten energetischen Differenzen berücksichtigen lediglich die Unterschie-de zwischen den Minima, für eine weitere Evaluation der Energiebarrieren zwischendiesen Zuständen wurde die NEB (Nudged elastic band)-Methode genutzt, die denminimalen Pfad (Abb. 4.21) zwischen zwei Minima berechnet [321]. Die ermitteltenEnergiebarrieren (Aktivierungsenergien ∆G̸=) für die in Tab. 4.13 aufgeführten Über-gänge sind in Tab. 4.14 aufgelistet. Die Energiebarriere ist im Falle der destabilisiertenVariante 27-1 geringer als bei der unkonjugierten Mutante.4.2.4PolymerdynamikUngeachtet energetischer Differenzen können diese Simulationen dazu dienen, die Dy-namik der Polymere während des Faltungsprozesses zu interpretieren. Generell scheintdie Entfaltung entlang der CV 1 zunächst über einen Bindungsbruch der HB in derNähe der ersten Schleife zu geschehen (also beginnend bei R21O-S16H), während dieEntfaltung entlang der CV 2 nicht von der zweiten Schleife aus sondern mit einerAuflösung der HB am Ende des dritten β-Faltblatts (Q33O-Y24H) beginnt. Wie kannPEG an der Konjugationsstelle 16 zu einer Stabilisierung führen? In Übereinstim-mung mit der in Kap. 4.1.1.7 beschriebenen höheren Tendenz zur Bedeckung derProteinoberfläche im Vergleich zu 27-1 kann die Polymerkette auch während des Ent-faltungsprozesses mit den Aminosäuren wechselwirken (Abb. 4.22).4.2. Metadynamik-Simulationen121Abbildung 4.21: Schematische Darstellung der geeignetsten Pfade auf Basis der NEB-Methode zwischen den globalen Minima und den in Tab. 4.13 beschriebenen Energieminima.Das Kennzeichen # gibt die dabei registrierte Energiebarriere an. Die Pfade sind ebenso inAbb. 4.19 als rote Linien gekennzeichnet.Tabelle 4.14: Berechnung der Energiebarrieren zum Zeitpunkt der jeweils letzten Moment-aufnahme in Bezug auf die aufgelisteten Übergänge. Ebenso wird die Differenz dieser ∆G̸=-Werte zwischen der unPEGylierten und der PEGylierten Variante gezeigt.VarianteÜbergang∆G̸= [kcal/mol]∆∆G̸= [kcal/mol]16A ->D11.11-0.0416-1A ->E11.1527A ->C12.411.9727-1A ->B10.44Das Polymer wechselwirkt mit S19 und R21 und legt sich dabei auf das bei derEntfaltung dem Solvens zugänglich werdende Areal. Diese Dynamik könnte den Ent-faltungsprozess modulieren. Auch bei einer Loslösung des dritten β-Faltblatts könntedie Entfaltung durch PEG an dieser Konjugationsstelle erschwert werden, da dasPolymer, wie in Kap. 4.3.1 weiter unten beschrieben, mit S32 dieses Faltblatts inter-agiert. Im Gegensatz hierzu zeigt sich bei dem destabilisierten Konjugat 27-1, dassdie Polymerkette keine Areale des Entfaltungsprozesses adressiert. Wie auch im Fal-le der entsprechenden klassischen MD (vgl. Kap. 4.1.1.7) liegt PEG hauptsächlichsolvensexponiert vor. Die relativ stabile Wechselwirkung des Linker-Amids mit derSeitenkette von E12 wird beim Entfaltungsprozess aufgehoben.122Kapitel 4. Ergebnisse: ModellproteineDiese Beobachtungen deuten also darauf hin, dass sich die PEG-Ketten bei Entfal-tungsprozessen auf die freiwerdenden Areale legen und so zu einer Stabilisierung bei-tragen können. Dies könnte die Hypothese aus [96] stützen. Hierin wurde, in Analogiezu Chaperonen, eine polymerbasierte erleichterte Rückfaltung zur nativen Strukturdurch eine Etablierung stabilerer partiell entfalteter Strukturen postuliert.4.2.5FazitDie erhaltenen Energiediagramme und die daraus berechneten Differenzen spiegelndie Stabilisierung von 16-1 und die Destabilisierung von 27-1 rein qualitativ wider,jedoch deuten die Fluktuationen der meisten Energiedifferenzen in Abb. 4.20 auchnach 2.4 µs Simulationszeit auf eine mangelnde Konvergenz hin, weshalb der Fokusauf eine partielle Entfaltung entlang CV 2 gelegt wurde. Es wäre dahingehend inter-essant, deutlich längere Simulationen der Konjugate 16-2 (-1.70 ± 0.03 kcal/mol) und27-2 (1.26 ± 0.06 kcal/mol) durchzuführen, da hierbei der experimentell ermittelteenergetische Unterschied noch größer ist.Unabhängig von der Eignung der Simulationen zur Berechnung der Energiedifferenzenkonnte beobachtet werden, dass PEG im Falle der destabilisierten Variante entfalteteAreale adressiert. Dies könnte den weiteren Entfaltungsprozess verhindern. Im Ge-gensatz dazu interagierte die Polymerkette im Falle der destablisierten Variante nichtmit diesen Bereichen.Abbildung 4.22: Exemplarische PEG-Konformationen im Falle einer partiellen Entfaltung.Gezeigt sind Entfaltungen entlang (A) der CV 1 und (B) der CV 2 für 16-1, sowie entlang(C) der CV 1 und (D) der CV 2 für 27-1. Im Falle des Konjugats 16-1 streckt sich diePEG-Kette entlang des entfalteten Areals, während bei 27-1 keine Interaktionen beobachtetwerden können.4.3. Vergleich der Polymere1234.3Vergleich der PolymereIm Folgenden wird die Dynamik der vier Polymervarianten PEG, LPG, PMeOx undPEtOx in den Modellprotein-Konjugaten beschrieben und vergleichend gegenüberge-stellt. Jede Konjugationsstelle wird auf einer Doppelseite mithilfe der in Kap. 3.1.8 be-schriebenen Eigenschaften vorgestellt. Diese Untersuchungen umfassen Interaktions-energien (mit -1 multipliziert), sowie Informationen zum Bedeckungsgrad der Protein-und Polymeroberflächen, zu der Anzahl an Protein-Polymer HB, der Solvatation derPolymerkette und den Fluktuationen des Proteinrückgrats und des Polymers. Zusätz-lich wurden die Simulationen visuell analysiert. Exemplarische Momentaufnahmen,die auf Basis dieser Analysen als charakteristisch für die Trajektorien mit Fokus aufmögliche Unterschiede hinsichtlich der Protein-Polymer-Wechselwirkungen angesehenwurden, werden ergänzend zu einer entsprechenden Beschreibung der Polymerdyna-mik präsentiert.Es werden 13 Konjugationsstellen der Pin 1 WW-Domäne mit Asparagin-Linker vor-gestellt: 14, 16, 17, 18, 19, 21, 23, 26, 27, 28, 29, 30 und 32. Dabei handelt es sichum Positionen aus den verschiedenen Strukturelmenten (β-Faltblätter und Schleifen 1und 2), welche von Price et al. bereits in Bezug zu PEG experimentell charakterisiertwurden und für die entsprechende zusätzliche Simulationen zu den alternativen Poly-meren durchgeführt wurden (vgl. Kap. 3.1.2.2). Im Falle des Her2-Affibodys werdeninsgesamt 15 Konjugationsstellen mit Cystein-Maleimid-Linker vorgestellt: 7, 8, 15,21, 23, 25, 29, 33, 39, 43, 46, 47, 49, 50 und 53. Diese Auswahl basierte auf bei der Bin-dung des Affibodys an Her2 freiliegenden Proteinarealen, sowie den strukturbasiertenRichtlinien von Price et al. (vgl. Kap. 3.1.2.2).Ein Einschub dieser systematisch dargestellten Analysen soll hier gezeigt werden, dasie als Datengrundlage für die im Anschluss für beide Proteine aufgeführten, zusam-menfassenden Vergleiche (siehe Kap. 4.3.3) dienen. Hierbei werden sowohl Unterschie-de zwischen den Polymerarten als auch den Konjugationsstellen beleuchtet.124Kapitel 4. Ergebnisse: Modellproteine4.3.1Pin 1-Konjugate4.3.1.1Konjugationsstelle 14Die Interaktionsprofile der Polymere sind vergleichbar (Abb. 4.24), wobei die POx-Varianten stärkere Interaktionen mit Y23 und F25 zeigen, PMeOx speziell auch mitS32 und F34. Die größeren POx-Polymere führen zur Bedeckung einer größeren Prote-inoberfläche. LPG ist am flexibelsten, kann aber im Vergleich zu den POx-Variantenaufgrund der Hydroxylgruppen mehr HB zum Protein aufbauen. PEG orientiert sichvornehmlich zum dritten β-Faltblatt hin und wechselwirkt dadurch mit Y23, F25und S32 (Abb. 4.23-A). Es ist zwischen den aromatischen Gruppen von Y23 und F25positioniert, teils auch in der Nähe von R17. Andererseits kann es sich auch in dieandere Richtung umorientieren und mit K13 und M15 wechselwirken. Die Hydro-xylgruppen von LPG interagieren v. a. mit dem Solvens und dem Polymerrückgrat.Die Struktur befindet sich hauptsächlich im selben Areal wie PEG, Interaktionen mitR17 sind aber häufiger zu beobachten (Abb. 4.23-B). Weiterhin sind Konformationenzu registrieren, in denen sich LPG entlang des ersten und zweiten β-Faltblatts inRichtung H27 streckt. PMeOx verbleibt sehr stabil zwischen den Aromaten von Y23und F25, wechselwirkt hierbei auch mit den Seitenketten von R17 und S32 (Abb.4.23-C). Während der Simulation kommt es auch zur Umlagerung der endständigenMonomere hin zum Aromaten von F34, wobei die Seitenkette des Monomers zumeistparallel auf dem aromatischen Ring aufliegt. Im Falle von PEtOx kommt es seltenzur Ausstreckung zum dritten β-Faltblatt, es interagiert v. a. mit dem Rückgrat vonR17 und dem Aromaten von Y23 (Abb. 4.23-D). Teilweise befindet es sich auch überF25 Richtung H27 und N30.Abbildung 4.23: Beispielhafte Momentaufnahmen von (A) PEG, (B) LPG, (C) PMeOxund (D) PEtOx an der Konjugationsstelle 14 von Pin 1 WW. PEG ist orange, LPG grün,PMeOx blau und PEtOx violett gefärbt. Ausgewählte zusätzliche Aminosäuren in Polymernä-he sind ebenso dargestellt.4.3. Vergleich der Polymere125Abbildung 4.24: Protein-Polymer-Dynamiken an Konjugationsstelle 14 von Pin 1 WW. Dieoberste Reihe zeigt Momentaufnahmen aus den Trajektorien jedes Polymers im Intervall von10 ns, die erste Momentaufnahme des Proteins ist als Cartoon dargestellt. PEG ist orange,LPG grün, PMeOx blau und PEtOx violett gefärbt. Im letzten Bild der obersten Reihe sinddiese Momentaufnahmen als Sphären übereinandergelagert. Darunter sind alle Eigenschaf-ten dargestellt, die in Kap. 3.1.8 geschildert wurden. Die Konjugationsstelle ist im Falle deraminosäurespezifischen Interaktionsenergien (mit -1 multipliziert) als K auf rotem Grund ge-kennzeichnet.126Kapitel 4. Ergebnisse: Modellproteine4.3.1.2Konjugationsstelle 16Im Gegensatz zur Konjugationsstelle 14 zeigen die POx-Polymere größere RMSF-Werte (Abb. 4.26), da die Monomere nicht zwischen den Aromaten von Y23, F25und F34 positioniert sind. PEG windet sich um die erste Schleife zwischen den Sei-tenketten von R17 und R21 zum Rückgrat von G20 hin (Abb. 4.25-A). Teilweiseorientiert es sich auch zu S32 oder auf die andere Seite der ersten Schleife zu M15und R17. LPG befindet sich ebenso hauptsächlich neben R17 in unmittelbarer Näheder ersten Schleife, in einer geknäuelten oder gestreckten Konformation (Abb. 4.25-B).Mit dieser Aminosäure zeigt es die im Vergleich zu allen Polymerarten größte Inter-aktionsenergie. Die ersten Monomere wechselwirken auch mit Y23 und F34. Währenddie PMeOx-Monomere initial auch in der Nähe von R17 vorzufinden sind, wechselnsie ihre Ausrichtung hin zu den Aminosäuren S32 und F34 (Abb. 4.25-C). Dabei sindwiederum zu Y23 und F34 parallel ausgerichtete Polymerseitenketten zu beobach-ten. PEtOx bildet eine ausgestreckte Konformation um die erste Schleife. Alternativwechselwirkt es mit Y23, N30, S32 und F34 (Abb. 4.25-D).Abbildung 4.25: Beispielhafte Momentaufnahmen von (A) PEG, (B) LPG, (C) PMeOxund (D) PEtOx an der Konjugationsstelle 16 von Pin 1 WW. PEG ist orange, LPG grün,PMeOx blau und PEtOx violett gefärbt. Ausgewählte zusätzliche Aminosäuren in Polymernä-he sind ebenso dargestellt.4.3. Vergleich der Polymere127Abbildung 4.26: Protein-Polymer-Dynamiken an Konjugationsstelle 16 von Pin 1 WW. Dieoberste Reihe zeigt Momentaufnahmen aus den Trajektorien jedes Polymers im Intervall von10 ns, die erste Momentaufnahme des Proteins ist als Cartoon dargestellt. PEG ist orange,LPG grün, PMeOx blau und PEtOx violett gefärbt. Im letzten Bild der obersten Reihe sinddiese Momentaufnahmen als Sphären übereinandergelagert. Darunter sind alle Eigenschaf-ten dargestellt, die in Kap. 3.1.8 geschildert wurden. Die Konjugationsstelle ist im Falle deraminosäurespezifischen Interaktionsenergien (mit -1 multipliziert) als K auf rotem Grund ge-kennzeichnet.128Kapitel 4. Ergebnisse: Modellproteine4.3.1.3Konjugationsstelle 17Nur PMeOx verweilt hauptsächlich auf der Proteinoberfläche, was in einer deutlichgrößeren Bedeckung des Proteins und stärkeren Interaktionen resultiert (Abb. 4.28).PEG ändert seine Konformation im Laufe der Simulation sehr häufig, einzig zwischenS32 und F34 entlang der Oberfläche von Y23 lassen sich relativ stabile Wechselwir-kungen beobachten (Abb. 4.27-A). Ähnlich zu PEG ist auch LPG recht flexibel undinteragiert vornehmlich mit sich selbst oder dem Solvens (Abb. 4.27-B). Daher kannkeiner stabile Konformation hin zu F34 kann beobachtet werden. PMeOx fluktuierthingegen nicht im Bereich der Konjugationsstelle um die erste Schleife, sondern bil-det sehr schnell stabile Wechselwirkungen zu Y23, F25, sowie mit den terminalenMonomeren zu den Rückgraten von N30 und A31 (Abb. 4.27-C) und den initialenMonomeren zur Seitenkette von R14. PEtOx zeigt andere Dynamiken, da es, ähnlichwie PEG und LPG, im Bereich der ersten Schleife fluktuiert. Hier wechselwirkt esteilweise mit der Carbonylgruppe von S18 (Abb. 4.27-D). Entsprechend weist es imVergleich zu den anderen Konjugationsstellen recht hohe RMSF-Werte auf.Abbildung 4.27: Beispielhafte Momentaufnahmen von (A) PEG, (B) LPG, (C) PMeOxund (D) PEtOx an der Konjugationsstelle 17 von Pin 1 WW. PEG ist orange, LPG grün,PMeOx blau und PEtOx violett gefärbt. Ausgewählte zusätzliche Aminosäuren in Polymernä-he sind ebenso dargestellt.4.3. Vergleich der Polymere129Abbildung 4.28: Protein-Polymer-Dynamiken an Konjugationsstelle 17 von Pin 1 WW. Dieoberste Reihe zeigt Momentaufnahmen aus den Trajektorien jedes Polymers im Intervall von10 ns, die erste Momentaufnahme des Proteins ist als Cartoon dargestellt. PEG ist orange,LPG grün, PMeOx blau und PEtOx violett gefärbt. Im letzten Bild der obersten Reihe sinddiese Momentaufnahmen als Sphären übereinandergelagert. Darunter sind alle Eigenschaf-ten dargestellt, die in Kap. 3.1.8 geschildert wurden. Die Konjugationsstelle ist im Falle deraminosäurespezifischen Interaktionsenergien (mit -1 multipliziert) als K auf rotem Grund ge-kennzeichnet.130Kapitel 4. Ergebnisse: Modellproteine4.3.1.4Konjugationsstelle 18Position 18 zeigt die größten Unterschiede hinsichtlich der Polymerdynamik, wodurchauch die Interaktionsprofile deutliche Differenzen aufweisen (Abb. 4.30). PMeOx istdas einzige Polymer mit Wechselwirkungen zu Y23, S32, Q33 und F34, PEtOx da-gegen zeigt Interaktionen mit K13, R14 und M15. LPG adressiert lediglich R17. DiePEG-Alternativen zeigen wesentlich höhere RMSF-Werte. PEG orientiert sich zu M15und R17 hin, mit Wechselwirkungen zum eigenen Asparagin-Linker. Öfters aber zeigtes Konformationen um die andere Seite der Schleife herum, abgewandt von M15 (Abb.4.29-A). Zwischenzeitlich fluktuiert das Polymer im Solvens. LPG adressiert nur sel-ten die Seite der ersten Schleife mit M15. Es ist hauptsächlich solvensexponiert undinteragiert mit sich selbst oder der Seitenkette von R17 (Abb. 4.29-B). In manchenMomentaufnahmen zeigt es in Richtung des zweiten und dritten β-Faltblatts. PMe-Ox liegt nicht solvensexponiert vor, sondern streckt sich hin zum Aromaten von Y23,ebenso zu F34 und S32 (Abb. 4.29-C). Im Gegensatz hierzu orientiert sich PEtOx v.a. auf die proteinabgewandte Seite der Schleife und interagiert mit dem hydrophobenAreal von K13, M15 und V22 (Abb. 4.29-D). Die terminale Carbonylgruppe kannmit der positiv geladenen Seitenktte von K13 wechselwirken. Zu Beginn fluktuiertdas Polymer im Solvens, abgewandt vom Protein.Abbildung 4.29: Beispielhafte Momentaufnahmen von (A) PEG, (B) LPG, (C) PMeOxund (D) PEtOx an der Konjugationsstelle 18 von Pin 1 WW. PEG ist orange, LPG grün,PMeOx blau und PEtOx violett gefärbt. Ausgewählte zusätzliche Aminosäuren in Polymernä-he sind ebenso dargestellt.4.3. Vergleich der Polymere131Abbildung 4.30: Protein-Polymer-Dynamiken an Konjugationsstelle 18 von Pin 1 WW. Dieoberste Reihe zeigt Momentaufnahmen aus den Trajektorien jedes Polymers im Intervall von10 ns, die erste Momentaufnahme des Proteins ist als Cartoon dargestellt. PEG ist orange,LPG grün, PMeOx blau und PEtOx violett gefärbt. Im letzten Bild der obersten Reihe sinddiese Momentaufnahmen als Sphären übereinandergelagert. Darunter sind alle Eigenschaf-ten dargestellt, die in Kap. 3.1.8 geschildert wurden. Die Konjugationsstelle ist im Falle deraminosäurespezifischen Interaktionsenergien (mit -1 multipliziert) als K auf rotem Grund ge-kennzeichnet.132Kapitel 4. Ergebnisse: Modellproteine4.3.1.5Konjugationsstelle 19An dieser Stelle zeigt PEtOx die größtmögliche Oberflächenbedeckung, den gerings-ten RMSF-Wert und die stärkste Interaktionsenergie (Abb. 4.32), hauptsächlich mitAminosäuren der ersten Schleife. PEG windet sich um die erste Schleife und inter-agiert mit dem Amidrückgrat von G20, sowie mit R21 mithilfe des Asparagin-Linkers(Abb. 4.31-A). Eine Orientierung zu S32 und F34 ist nur selten zu beobachten. LPGliegt solvensexponierter, abgewandt vom Protein, vor. Es interagiert mit der Seiten-kette von R21 und, für kürzere Zeit, mit der entfernt gelegenen Aminosäure E35 (Abb.4.31-B). PMeOx ist zur Rückseite der ersten Schleife orientiert und wechselwirkt mitdem Rückgrat von R17, sowie mit M15 (Abb. 4.31-C). Zu Beginn ist das Polymer inder Nähe von S32 und F34 positioniert. PEtOx verbleibt auf dem Aromaten von Y23und interagiert mit S32 über die terminale Carbonylgruppe, während die Seitenkettevon R21 ebenfalls in der Nähe des Polymers verbleibt (Abb. 4.31-D).Abbildung 4.31: Beispielhafte Momentaufnahmen von (A) PEG, (B) LPG, (C) PMeOxund (D) PEtOx an der Konjugationsstelle 19 von Pin 1 WW. PEG ist orange, LPG grün,PMeOx blau und PEtOx violett gefärbt. Ausgewählte zusätzliche Aminosäuren in Polymernä-he sind ebenso dargestellt.4.3. Vergleich der Polymere133Abbildung 4.32: Protein-Polymer-Dynamiken an Konjugationsstelle 19 von Pin 1 WW. Dieoberste Reihe zeigt Momentaufnahmen aus den Trajektorien jedes Polymers im Intervall von10 ns, die erste Momentaufnahme des Proteins ist als Cartoon dargestellt. PEG ist orange,LPG grün, PMeOx blau und PEtOx violett gefärbt. Im letzten Bild der obersten Reihe sinddiese Momentaufnahmen als Sphären übereinandergelagert. Darunter sind alle Eigenschaf-ten dargestellt, die in Kap. 3.1.8 geschildert wurden. Die Konjugationsstelle ist im Falle deraminosäurespezifischen Interaktionsenergien (mit -1 multipliziert) als K auf rotem Grund ge-kennzeichnet.134Kapitel 4. Ergebnisse: Modellproteine4.3.1.6Konjugationsstelle 21Das Interaktionsprofil zeigt wiederum starke Interaktionen von PMeOx mit Y23 undPEtOx mit F34 (Abb. 4.34). PEG verbleibt abermals zwischen S32 und F34 (Abb.4.33-A) und wechselwirkt mit S16 über seinen Linker. Alternativ windet es sich umR17. LPG orientiert sich ebenso zum zweiten und dritten β-Faltblatt hin (Abb. 4.33-B), manchmal auch zur ersten Schleife. PMeOx verbleibt auf dem Aromaten vonY23 und erreicht mit seinem terminalen Monomer A31 (Abb. 4.33-C). Im Gegensatzhierzu interagiert PEtOx mit F34 in der kompletten Simulation (Abb. 4.33-D). DerRest des Polymers ist vom Protein abgekehrt, die terminale Carbonylgruppe kannmit der Seitenkette von S19 wechselwirken.Abbildung 4.33: Beispielhafte Momentaufnahmen von (A) PEG, (B) LPG, (C) PMeOxund (D) PEtOx an der Konjugationsstelle 21 von Pin 1 WW. PEG ist orange, LPG grün,PMeOx blau und PEtOx violett gefärbt. Ausgewählte zusätzliche Aminosäuren in Polymernä-he sind ebenso dargestellt.4.3. Vergleich der Polymere135Abbildung 4.34: Protein-Polymer-Dynamiken an Konjugationsstelle 21 von Pin 1 WW. Dieoberste Reihe zeigt Momentaufnahmen aus den Trajektorien jedes Polymers im Intervall von10 ns, die erste Momentaufnahme des Proteins ist als Cartoon dargestellt. PEG ist orange,LPG grün, PMeOx blau und PEtOx violett gefärbt. Im letzten Bild der obersten Reihe sinddiese Momentaufnahmen als Sphären übereinandergelagert. Darunter sind alle Eigenschaf-ten dargestellt, die in Kap. 3.1.8 geschildert wurden. Die Konjugationsstelle ist im Falle deraminosäurespezifischen Interaktionsenergien (mit -1 multipliziert) als K auf rotem Grund ge-kennzeichnet.136Kapitel 4. Ergebnisse: Modellproteine4.3.1.7Konjugationsstelle 23PEG interagiert hier mit R14, R17 und S32 (Abb. 4.35-A). LPG zeigt ähnliche Wech-selwirkungen, die Konjugation führt hier jedoch zu einer stärkeren Desolvatation umF25 (Abb. 4.35-B). PMeOx fluktuiert stark, was ebenso in einem recht großen RMSF-Wert resultiert (Abb. 4.36). Zu Beginn ist es zwischen R14 und R17 positioniert undkann M15 mit seinen hydrophoben Strukturen erreichen (Abb. 4.35-C). Im zweitenTeil der Simulation streckt es sich ins Solvens aus, direkt über dem dritten β-Faltblatt.PEtOx ist nicht zwischen R14 und R17 vorzufinden, sondern verbleibt auf dem zwei-ten und dritten β-Faltblatt und wechselwirkt mit F25, N30, S32 und F34 (Abb.4.35-D). Es zeigt dadurch auch die größte Proteinoberflächenbedeckung und einensehr geringen RMSF-Wert (Abb. 4.36).Abbildung 4.35: Beispielhafte Momentaufnahmen von (A) PEG, (B) LPG, (C) PMeOxund (D) PEtOx an der Konjugationsstelle 23 von Pin 1 WW. PEG ist orange, LPG grün,PMeOx blau und PEtOx violett gefärbt. Ausgewählte zusätzliche Aminosäuren in Polymernä-he sind ebenso dargestellt.4.3. Vergleich der Polymere137Abbildung 4.36: Protein-Polymer-Dynamiken an Konjugationsstelle 23 von Pin 1 WW. Dieoberste Reihe zeigt Momentaufnahmen aus den Trajektorien jedes Polymers im Intervall von10 ns, die erste Momentaufnahme des Proteins ist als Cartoon dargestellt. PEG ist orange,LPG grün, PMeOx blau und PEtOx violett gefärbt. Im letzten Bild der obersten Reihe sinddiese Momentaufnahmen als Sphären übereinandergelagert. Darunter sind alle Eigenschaf-ten dargestellt, die in Kap. 3.1.8 geschildert wurden. Die Konjugationsstelle ist im Falle deraminosäurespezifischen Interaktionsenergien (mit -1 multipliziert) als K auf rotem Grund ge-kennzeichnet.138Kapitel 4. Ergebnisse: Modellproteine4.3.1.8Konjugationsstelle 26Mit Ausnahme einiger weniger Momentaufnahmen verharrt PEG in der Nähe deshydrophoben Zentrums um W11. Stabile HB mit den Seitenktten von W11 und Q33sind registrierbar (Abb. 4.37-A). Auch wenn LPG ein flexibleres Verhalten zeigt, sonimmt es doch vergleichbare Konformationen ein. HB zu W11, T29, Q33, S38 undzur C-terminalen Carboxylgruppe sind zu beobachten (Abb. 4.37-B). PMeOx ist nichtso nah zu Q33 wie PEG und LPG positioniert, kann jedoch auch W11 desolvatisie-ren (Abb. 4.37-C). Der Amidwasserstoff des Linkers bildet eine stabile Orientierungzum Aromaten von W11, während die entsprechende Carbonylgruppe mit dem Rück-grat von I28 und T29 wechselwirkt. Das Polymer bleibt hingegen relativ flexibel undstreckt sich manchmal zum C-Terminus aus, teilweise auch zu den hydrophoben Ami-nosäuren P8, P9 und I28. An dieser Konjugationsstelle verhält sich PEtOx identisch(Abb. 4.37-D).Abbildung 4.37: Beispielhafte Momentaufnahmen von (A) PEG, (B) LPG, (C) PMeOxund (D) PEtOx an der Konjugationsstelle 26 von Pin 1 WW. PEG ist orange, LPG grün,PMeOx blau und PEtOx violett gefärbt. Ausgewählte zusätzliche Aminosäuren in Polymernä-he sind ebenso dargestellt.4.3. Vergleich der Polymere139Abbildung 4.38: Protein-Polymer-Dynamiken an Konjugationsstelle 26 von Pin 1 WW. Dieoberste Reihe zeigt Momentaufnahmen aus den Trajektorien jedes Polymers im Intervall von10 ns, die erste Momentaufnahme des Proteins ist als Cartoon dargestellt. PEG ist orange,LPG grün, PMeOx blau und PEtOx violett gefärbt. Im letzten Bild der obersten Reihe sinddiese Momentaufnahmen als Sphären übereinandergelagert. Darunter sind alle Eigenschaf-ten dargestellt, die in Kap. 3.1.8 geschildert wurden. Die Konjugationsstelle ist im Falle deraminosäurespezifischen Interaktionsenergien (mit -1 multipliziert) als K auf rotem Grund ge-kennzeichnet.140Kapitel 4. Ergebnisse: Modellproteine4.3.1.9Konjugationsstelle 27PEG interagiert zunächst mit R14, anschließend über seinen Linker mit der Carboxyl-gruppe von E12 und über die Ethersauerstoffatome mit dem Rückgratamid von P9(Abb. 4.39-A). Die Wechselwirkung mit E12 ist über den Großteil der Zeit hinwegzu beobachten, während die Polymerkette sehr flexibel bleibt und teilweise I28 adres-siert. Im Falle von LPG etabliert der Linker dieselbe Interaktion, während sich diePolymerkette zum N-Terminus ausstreckt und hauptsächlich solvensexponiert bleibt(Abb. 4.39-B). Teilweise sind jedoch auch HB zwischen den Hydroxylgruppen vonLPG und E12 zu erkennen. Der Linker wechselwirkt auch im Falle von PMeOx mitE12 (Abb. 4.39-C). Das Polymer adressiert zum Teil P9. Im Gegensatz zu allen ande-ren Polymeren scheint PEtOx mit seiner Länge Y23 und F25 des oberen hydrophobenZentrums erreichen zu können. Dabei wechselwirkt es auch mit R14 und R17 (Abb.4.39-D). Die Bedeckung der Proteinoberfläche ist somit bei diesem Polymer am größ-ten (Abb. 4.40).Abbildung 4.39: Beispielhafte Momentaufnahmen von (A) PEG, (B) LPG, (C) PMeOxund (D) PEtOx an der Konjugationsstelle 27 von Pin 1 WW. PEG ist orange, LPG grün,PMeOx blau und PEtOx violett gefärbt. Ausgewählte zusätzliche Aminosäuren in Polymernä-he sind ebenso dargestellt.4.3. Vergleich der Polymere141Abbildung 4.40: Protein-Polymer-Dynamiken an Konjugationsstelle 27 von Pin 1 WW. Dieoberste Reihe zeigt Momentaufnahmen aus den Trajektorien jedes Polymers im Intervall von10 ns, die erste Momentaufnahme des Proteins ist als Cartoon dargestellt. PEG ist orange,LPG grün, PMeOx blau und PEtOx violett gefärbt. Im letzten Bild der obersten Reihe sinddiese Momentaufnahmen als Sphären übereinandergelagert. Darunter sind alle Eigenschaf-ten dargestellt, die in Kap. 3.1.8 geschildert wurden. Die Konjugationsstelle ist im Falle deraminosäurespezifischen Interaktionsenergien (mit -1 multipliziert) als K auf rotem Grund ge-kennzeichnet.142Kapitel 4. Ergebnisse: Modellproteine4.3.1.10Konjugationsstelle 28PEG orientiert sich entweder zum Rückgratamid von P9 oder, was häufiger der Fallist, in Richtung W11 und N26 (Abb. 4.41-A). LPG verhält sich ähnlich (Abb. 4.41-B),kann jedoch stärkere Interaktionen mit H27 etablieren. PMeOx bildet keine Wech-selwirkungen mit dem Protein aus, mit Ausnahme der Seitenkette der benachbar-ten Aminosäure T29 (Abb. 4.41-C). PEtOx verbleibt solvensexponiert und sehr fle-xibel (Abb. 4.41-D). Das Interaktionsprofil verdeutlicht stärkere Wechselwirkungenvon PEG und LPG mit den N-terminalen Aminosäuren im Vergleich zu den POx-Polymeren.Abbildung 4.41: Beispielhafte Momentaufnahmen von (A) PEG, (B) LPG, (C) PMeOxund (D) PEtOx an der Konjugationsstelle 28 von Pin 1 WW. PEG ist orange, LPG grün,PMeOx blau und PEtOx violett gefärbt. Ausgewählte zusätzliche Aminosäuren in Polymernä-he sind ebenso dargestellt.4.3. Vergleich der Polymere143Abbildung 4.42: Protein-Polymer-Dynamiken an Konjugationsstelle 28 von Pin 1 WW. Dieoberste Reihe zeigt Momentaufnahmen aus den Trajektorien jedes Polymers im Intervall von10 ns, die erste Momentaufnahme des Proteins ist als Cartoon dargestellt. PEG ist orange,LPG grün, PMeOx blau und PEtOx violett gefärbt. Im letzten Bild der obersten Reihe sinddiese Momentaufnahmen als Sphären übereinandergelagert. Darunter sind alle Eigenschaf-ten dargestellt, die in Kap. 3.1.8 geschildert wurden. Die Konjugationsstelle ist im Falle deraminosäurespezifischen Interaktionsenergien (mit -1 multipliziert) als K auf rotem Grund ge-kennzeichnet.144Kapitel 4. Ergebnisse: Modellproteine4.3.1.11Konjugationsstelle 29Mit Ausnahme einiger weniger Momentaufnahmen etabliert PEG abermals Interak-tionen mit W11 und Q33 des unteren hydrophoben Zentrums (Abb. 4.43-A). LPGzeigt ein ähnliches Verhalten (Abb. 4.43-B). PMeOx ist flexibler und wechselwirktseltener mit W11 und Q33. Dagegen adressiert es oft die hydrophoben SeitenkettenP9 und I28 (Abb. 4.43-C). Im Gegensatz hierzu verbleibt PEtOx in einer stabilen,geknäuelten Konformation in der Nähe von W11 und N26 (Abb. 4.43-D), was aucham RMSF-Wert erkennbar ist (Abb. 4.44).Abbildung 4.43: Beispielhafte Momentaufnahmen von (A) PEG, (B) LPG, (C) PMeOxund (D) PEtOx an der Konjugationsstelle 29 von Pin 1 WW. PEG ist orange, LPG grün,PMeOx blau und PEtOx violett gefärbt. Ausgewählte zusätzliche Aminosäuren in Polymernä-he sind ebenso dargestellt.4.3. Vergleich der Polymere145Abbildung 4.44: Protein-Polymer-Dynamiken an Konjugationsstelle 29 von Pin 1 WW. Dieoberste Reihe zeigt Momentaufnahmen aus den Trajektorien jedes Polymers im Intervall von10 ns, die erste Momentaufnahme des Proteins ist als Cartoon dargestellt. PEG ist orange,LPG grün, PMeOx blau und PEtOx violett gefärbt. Im letzten Bild der obersten Reihe sinddiese Momentaufnahmen als Sphären übereinandergelagert. Darunter sind alle Eigenschaf-ten dargestellt, die in Kap. 3.1.8 geschildert wurden. Die Konjugationsstelle ist im Falle deraminosäurespezifischen Interaktionsenergien (mit -1 multipliziert) als K auf rotem Grund ge-kennzeichnet.146Kapitel 4. Ergebnisse: Modellproteine4.3.1.12Konjugationsstelle 30PEG zeigt ein äußerst flexibles Verhalten und etabliert lediglich mit S32 länger an-dauernde Interaktionen (Abb. 4.45-A). LPG fluktuiert ebenso, bevor es Wechselwir-kungen mit S16, R17, Y23 und S32 in der zweiten Hälfte der Trajektorie zeigt (Abb.4.45-B). PMeOx verbleibt auf der Oberfläche von Y23 und F25, interagiert jedoch,im Gegensatz zu LPG, nicht mit Aminosäuren der ersten Schleife (Abb. 4.45-C). Zu-sätzlich zu Y23 und Y25 interagiert es auch über eine Carbonylgruppe mit R14. ImGegensatz zu den anderen Polymeren richtet sich PEtOx nicht entlang der Protein-oberfläche aus, sondern um I28 und den Aromaten von H27 (Abb. 4.45-D). PMeOxzeigt, basierend auf diesen Beschreibungen, die größte Oberflächenbedeckung (Abb.4.46).Abbildung 4.45: Beispielhafte Momentaufnahmen von (A) PEG, (B) LPG, (C) PMeOxund (D) PEtOx an der Konjugationsstelle 30 von Pin 1 WW. PEG ist orange, LPG grün,PMeOx blau und PEtOx violett gefärbt. Ausgewählte zusätzliche Aminosäuren in Polymernä-he sind ebenso dargestellt.4.3. Vergleich der Polymere147Abbildung 4.46: Protein-Polymer-Dynamiken an Konjugationsstelle 30 von Pin 1 WW. Dieoberste Reihe zeigt Momentaufnahmen aus den Trajektorien jedes Polymers im Intervall von10 ns, die erste Momentaufnahme des Proteins ist als Cartoon dargestellt. PEG ist orange,LPG grün, PMeOx blau und PEtOx violett gefärbt. Im letzten Bild der obersten Reihe sinddiese Momentaufnahmen als Sphären übereinandergelagert. Darunter sind alle Eigenschaf-ten dargestellt, die in Kap. 3.1.8 geschildert wurden. Die Konjugationsstelle ist im Falle deraminosäurespezifischen Interaktionsenergien (mit -1 multipliziert) als K auf rotem Grund ge-kennzeichnet.148Kapitel 4. Ergebnisse: Modellproteine4.3.1.13Konjugationsstelle 32PEG ist zwischen den Aminosäuren Y23 und F34 positioniert und streckt sich zuS16 und R17 aus (Abb. 4.47-A). In manchen Momentaufnahmen wechselt es seineKonformation in Richtung des dritten β-Faltblatts und interagiert mit dem Amid-rückgrat von F34. Ähnliche Konformationen hin zur ersten Schleife können im Fallevon LPG beobachtet werden (Abb. 4.47-B), wobei auch Momentaufnahmen mit kom-pakten Polymerkonformationen zu registrieren sind. PMeOx streckt sich zu R17 ausund ist imstande, die Aromaten von Y23 und F34 über die gesamte Simulationszeithinweg zu adressieren (Abb. 4.47-C). PEtOx ist rigider, etabliert aber ähnliche Inter-aktionen (Abb. 4.47-D). Beide POx-Varianten befinden sich auf der Oberfläche deszweiten und dritten β-Faltblatts und führen somit zu einer starken Bedeckung derProteinoberfläche (Abb. 4.48).Abbildung 4.47: Beispielhafte Momentaufnahmen von (A) PEG, (B) LPG, (C) PMeOxund (D) PEtOx an der Konjugationsstelle 32 von Pin 1 WW. PEG ist orange, LPG grün,PMeOx blau und PEtOx violett gefärbt. Ausgewählte zusätzliche Aminosäuren in Polymernä-he sind ebenso dargestellt.4.3. Vergleich der Polymere149Abbildung 4.48: Protein-Polymer-Dynamiken an Konjugationsstelle 32 von Pin 1 WW. Dieoberste Reihe zeigt Momentaufnahmen aus den Trajektorien jedes Polymers im Intervall von10 ns, die erste Momentaufnahme des Proteins ist als Cartoon dargestellt. PEG ist orange,LPG grün, PMeOx blau und PEtOx violett gefärbt. Im letzten Bild der obersten Reihe sinddiese Momentaufnahmen als Sphären übereinandergelagert. Darunter sind alle Eigenschaf-ten dargestellt, die in Kap. 3.1.8 geschildert wurden. Die Konjugationsstelle ist im Falle deraminosäurespezifischen Interaktionsenergien (mit -1 multipliziert) als K auf rotem Grund ge-kennzeichnet.150Kapitel 4. Ergebnisse: Modellproteine4.3.2Her2-Konjugate4.3.2.1Konjugationsstelle 7PEG orientiert sich zunächst um R10 herum und interagiert ebenso mit Y13, W14und Y35 (Abb. 4.49-A). In manchen Momentaufnahmen verbleibt das Polymer sol-vensexponiert oder adressiert V1 und N3 am N-Terminus. LPG ist nicht um R10herum positioniert, stattdessen wechselwirkt es mit der Seitenkette von E8 und derRückgratcarbonylgruppe von S39 (Abb. 4.49-B). Während der zweiten Hälfte der Tra-jektorie wird R10 durch eine Carbonylgruppe des Linkers adressiert, während sich dasPolymer zu N11 und W14 hin umorientiert. PMeOx ist zunächst oberhalb der drittenα-Helix zu N43 hin ausgestreckt. Danach findet es sich relativ solvensexponiert imBereich des N-Terminus wieder, bevor es, ähnlich wie PEG, wiederum R10, Y13, W14und Y35 adressiert (Abb. 4.49-C). PEtOx verbleibt über die gesamte Simulationszeithinweg in der Nähe des Aromaten von W14 parallel zur α-Helix, während das Amiddes Linkers teilweise mit der Carbonylgruppe von N11 wechselwirkt (Abb. 4.49-D).Mit W14 zeigt es die mit Abstand größte Interaktionsenergie.Abbildung 4.49: Beispielhafte Momentaufnahmen von (A) PEG, (B) LPG, (C) PMeOxund (D) PEtOx an der Konjugationsstelle 7 des Her2-Affibodys. PEG ist orange, LPG grün,PMeOx blau und PEtOx violett gefärbt. Ausgewählte zusätzliche Aminosäuren in Polymernä-he sind ebenso dargestellt.4.3. Vergleich der Polymere151Abbildung 4.50: Protein-Polymer-Dynamiken an Konjugationsstelle 7 des Her2 Affibody.Die oberste Reihe zeigt Momentaufnahmen aus den Trajektorien jedes Polymers im Intervallvon 10 ns, die erste Momentaufnahme des Proteins ist als Cartoon dargestellt. PEG ist oran-ge, LPG grün, PMeOx blau und PEtOx violett gefärbt. Im letzten Bild der obersten Reihesind diese Momentaufnahmen als Sphären übereinandergelagert. Darunter sind alle Eigen-schaften dargestellt, die in Kap. 3.1.8 geschildert wurden. Die Konjugationsstelle ist im Falleder aminosäurespezifischen Interaktionsenergien (mit -1 multipliziert) als K auf rotem Grundgekennzeichnet.152Kapitel 4. Ergebnisse: Modellproteine4.3.2.2Konjugationsstelle 8Zunächst liegt PEG geknäuelt vor und adressiert die Seitenkette von N11 mit demterminalen Monomer. Danach streckt es sich zu K49 und K50 aus (Abb. 4.51-A).Letzteres interagiert entweder mit dem Polymer oder dem Linker des Konjugats.LPG ist sehr flexibel und adressiert entweder eigene Strukturbereiche oder Amino-säuren in der Nähe des N-Terminus, v. a. V1, D2 und K4 (Abb. 4.51-B). Gegen Endeder Simulation streckt es sich nach unten zur Helix aus, in die Nähe von N11, E15und K49. Es wechselt dann die Richtung und wechselwirkt mit der Seitenkette vonD37 über seinen Linker und dem Rückgrat von Y35 über eine der Hydroxylgruppen.PMeOx verbleibt zunächst nicht zielgerichtet auf der Proteinoberfläche, hauptsächlichnahe dem N-Terminus. Später jedoch etabliert es eine relativ stabile Konformationum die Seitenkette von N43 herum und adressiert mit den terminalen Monomerenhydrophobe Areale von A29, F30 und L44 (Abb. 4.51-C). PEtOx wechselwirkt mitdem Seitenkettenamid von N11 in einer geknäuelten Konformation. Über einige aus-gestreckte Orientierungen hinweg windet es sich anschließend um den Aromaten vonF5 und wechselwirkt mit N3 über seine terminale Carbonylgruppe, was zu einer Defor-mation der ersten Helix führt (Abb. 4.51-D). Entsprechend ist die Interaktionsenergiemit F5 an dieser Stelle am größten.Abbildung 4.51: Beispielhafte Momentaufnahmen von (A) PEG, (B) LPG, (C) PMeOxund (D) PEtOx an der Konjugationsstelle 8 des Her2-Affibodys. PEG ist orange, LPG grün,PMeOx blau und PEtOx violett gefärbt. Ausgewählte zusätzliche Aminosäuren in Polymernä-he sind ebenso dargestellt.4.3. Vergleich der Polymere153Abbildung 4.52: Protein-Polymer-Dynamiken an Konjugationsstelle 8 des Her2 Affibody.Die oberste Reihe zeigt Momentaufnahmen aus den Trajektorien jedes Polymers im Intervallvon 10 ns, die erste Momentaufnahme des Proteins ist als Cartoon dargestellt. PEG ist oran-ge, LPG grün, PMeOx blau und PEtOx violett gefärbt. Im letzten Bild der obersten Reihesind diese Momentaufnahmen als Sphären übereinandergelagert. Darunter sind alle Eigen-schaften dargestellt, die in Kap. 3.1.8 geschildert wurden. Die Konjugationsstelle ist im Falleder aminosäurespezifischen Interaktionsenergien (mit -1 multipliziert) als K auf rotem Grundgekennzeichnet.154Kapitel 4. Ergebnisse: Modellproteine4.3.2.3Konjugationsstelle 15PEG wechselwirkt zunächst mit K7 und R10 (Abb. 4.53-A). In der zweiten Hälfteder Trajektorie verbleibt das Polymer vornehmlich in einer geknäuelten oder solvens-exponierten Konformation, interagiert aber auch teilweise mit den Seitenketten vonK49 und N52. LPG ist zunächst in der Nähe des C-Terminus um die AminosäurenN21, N52, D53 und K58 positioniert. Dann adressiert es K7, R10, N11 und vor allemE8 (Abb. 4.53-B). Am Ende orientiert es sich wieder zum C-Terminus hin. PMeOxverbleibt zumeist in der Nähe von N11 mit seinem Linker, sowie um R10 und W14mit den Monomeren (Abb. 4.53-C). Gegen Ende streckt es sich aus zu einem hy-drophoben Areal, wobei es mit K7, M9, Y13 und Y35 adressiert. In ähnlicher Weiseinteragiert auch der Linker von PEtOx mit N11, während der Polymer-Part oberhalbder Aromaten von Y13 und W14 liegt, dabei wechselwirkt es auch mit Y35 (Abb.4.53-D).Abbildung 4.53: Beispielhafte Momentaufnahmen von (A) PEG, (B) LPG, (C) PMeOxund (D) PEtOx an der Konjugationsstelle 15 des Her2-Affibodys. PEG ist orange, LPG grün,PMeOx blau und PEtOx violett gefärbt. Ausgewählte zusätzliche Aminosäuren in Polymernä-he sind ebenso dargestellt.4.3. Vergleich der Polymere155Abbildung 4.54: Protein-Polymer-Dynamiken an Konjugationsstelle 15 des Her2 Affibody.Die oberste Reihe zeigt Momentaufnahmen aus den Trajektorien jedes Polymers im Intervallvon 10 ns, die erste Momentaufnahme des Proteins ist als Cartoon dargestellt. PEG ist oran-ge, LPG grün, PMeOx blau und PEtOx violett gefärbt. Im letzten Bild der obersten Reihesind diese Momentaufnahmen als Sphären übereinandergelagert. Darunter sind alle Eigen-schaften dargestellt, die in Kap. 3.1.8 geschildert wurden. Die Konjugationsstelle ist im Falleder aminosäurespezifischen Interaktionsenergien (mit -1 multipliziert) als K auf rotem Grundgekennzeichnet.156Kapitel 4. Ergebnisse: Modellproteine4.3.2.4Konjugationsstelle 21PEG windet sich zunächst um K49. Nach nur 30 ns dreht sich der Linker um, waszu einer Polymerkonformation fernab der Helices führt. Interaktionen mit den Sei-tenketten von N23, Q26 und N55 können beobachtet werden (Abb. 4.55-A). Ähnlichzu PEG hält sich auch LPG hauptsächlich unterhalb des helikalen Parts des Proteinsauf und adressiert dabei N23, Q26, N24 und das Rückgrat von L22 (Abb. 4.55-B).Es streckt sich auch entlang des C-Terminus aus und wechselwirkt mit K58. PMe-Ox scheint über lange Zeit hinweg keine spezifische Aminosäure zu adressieren (Abb.4.55-C). Es hält sich vornehmlich in einer geknäuelten Konformation unterhalb deshelikalen Parts des Proteins auf. Manchmal streckt sich das Polymer zu den hydro-phoben Aminosäuren L18 und L19 sowie zur Seitenkette von N52 aus. PEtOx zeigtebenso geknäuelte Konformationen nahe N23, Q26 und Q55 (Abb. 4.55-D).Abbildung 4.55: Beispielhafte Momentaufnahmen von (A) PEG, (B) LPG, (C) PMeOxund (D) PEtOx an der Konjugationsstelle 21 des Her2-Affibodys. PEG ist orange, LPG grün,PMeOx blau und PEtOx violett gefärbt. Ausgewählte zusätzliche Aminosäuren in Polymernä-he sind ebenso dargestellt.4.3. Vergleich der Polymere157Abbildung 4.56: Protein-Polymer-Dynamiken an Konjugationsstelle 21 des Her2 Affibody.Die oberste Reihe zeigt Momentaufnahmen aus den Trajektorien jedes Polymers im Intervallvon 10 ns, die erste Momentaufnahme des Proteins ist als Cartoon dargestellt. PEG ist oran-ge, LPG grün, PMeOx blau und PEtOx violett gefärbt. Im letzten Bild der obersten Reihesind diese Momentaufnahmen als Sphären übereinandergelagert. Darunter sind alle Eigen-schaften dargestellt, die in Kap. 3.1.8 geschildert wurden. Die Konjugationsstelle ist im Falleder aminosäurespezifischen Interaktionsenergien (mit -1 multipliziert) als K auf rotem Grundgekennzeichnet.158Kapitel 4. Ergebnisse: Modellproteine4.3.2.5Konjugationsstelle 23PEG schützt Q25, Q26 und Q55 in einer geknäuelten Konformation (Abb. 4.57-A).Erst gegen Ende der Simulation adressiert es F30 und L44. LPG wechselwirkt miteiner Vielzahl von Aminosäuren: A17, Q25, Q26 und S33 (Abb. 4.57-B), gegen Endeauch mit Q47. Im Falle von PMeOx werden zunächst auch Aminosäuren in der Nähedes C-Terminus adressiert (Abb. 4.57-C), am Ende die hydrophoben AminosäurenF30 und L44. Das größere PEtOx hält sich im selben Areal in der Nähe von N23, Q26und Q55 auf (Abb. 4.57-D) und zeigt starke Wechselwirkungen mit den hydrophobenAminosäuren A29 und F30.Abbildung 4.57: Beispielhafte Momentaufnahmen von (A) PEG, (B) LPG, (C) PMeOxund (D) PEtOx an der Konjugationsstelle 23 des Her2-Affibodys. PEG ist orange, LPG grün,PMeOx blau und PEtOx violett gefärbt. Ausgewählte zusätzliche Aminosäuren in Polymernä-he sind ebenso dargestellt.4.3. Vergleich der Polymere159Abbildung 4.58: Protein-Polymer-Dynamiken an Konjugationsstelle 23 des Her2 Affibody.Die oberste Reihe zeigt Momentaufnahmen aus den Trajektorien jedes Polymers im Intervallvon 10 ns, die erste Momentaufnahme des Proteins ist als Cartoon dargestellt. PEG ist oran-ge, LPG grün, PMeOx blau und PEtOx violett gefärbt. Im letzten Bild der obersten Reihesind diese Momentaufnahmen als Sphären übereinandergelagert. Darunter sind alle Eigen-schaften dargestellt, die in Kap. 3.1.8 geschildert wurden. Die Konjugationsstelle ist im Falleder aminosäurespezifischen Interaktionsenergien (mit -1 multipliziert) als K auf rotem Grundgekennzeichnet.160Kapitel 4. Ergebnisse: Modellproteine4.3.2.6Konjugationsstelle 25PEG hält sich im selben hydrophoben Areal wie bei Konjugationsstelle 23 auf: überQ26 hin zu F30 und L44 (Abb. 4.59-A). LPG verbleibt unterhalb der Helices undschützt N23, Q26 und Q55 (Abb. 4.59-B), streckt sich aber auch manchmal aus undinteragiert mit K58. Genau wie PEG verbleibt auch PMeOx nahe F30 und L44 (Abb.4.59-C). Im Gegensatz hierzu streckt sich das größere PEtOx über diese Aminosäurenhinweg zu einem weiteren hydrophoben Areal aus: M9, Y13, I31 und Y35 (Abb. 4.59-D).Abbildung 4.59: Beispielhafte Momentaufnahmen von (A) PEG, (B) LPG, (C) PMeOxund (D) PEtOx an der Konjugationsstelle 25 des Her2-Affibodys. PEG ist orange, LPG grün,PMeOx blau und PEtOx violett gefärbt. Ausgewählte zusätzliche Aminosäuren in Polymernä-he sind ebenso dargestellt.4.3. Vergleich der Polymere161Abbildung 4.60: Protein-Polymer-Dynamiken an Konjugationsstelle 25 des Her2 Affibody.Die oberste Reihe zeigt Momentaufnahmen aus den Trajektorien jedes Polymers im Intervallvon 10 ns, die erste Momentaufnahme des Proteins ist als Cartoon dargestellt. PEG ist oran-ge, LPG grün, PMeOx blau und PEtOx violett gefärbt. Im letzten Bild der obersten Reihesind diese Momentaufnahmen als Sphären übereinandergelagert. Darunter sind alle Eigen-schaften dargestellt, die in Kap. 3.1.8 geschildert wurden. Die Konjugationsstelle ist im Falleder aminosäurespezifischen Interaktionsenergien (mit -1 multipliziert) als K auf rotem Grundgekennzeichnet.162Kapitel 4. Ergebnisse: Modellproteine4.3.2.7Konjugationsstelle 29PEG verbleibt in einer geknäuelten Konformation und interagiert mit R28 und R32(Abb. 4.61-A). Am Ende der Simulation adressiert es F30 und L44. LPG befindetsich ebenso auf den Aminosäuren F30 und L44, wobei die terminalen Monomere HBmit den Seitenketten von D36 und D37 ausbilden können (Abb. 4.61-B). PMeOxverbleibt ebenso auf F30, L44, sowie L51 (Abb. 4.61-C). PEtOx zeigt ausgestrecktereKonformationen in der Nähe von N23, Q25, Q55 und A56 (Abb. 4.61-D). Ähnlich zuallen anderen Polymeren zeigt es jedoch auch Konformationen nahe F30 und L44.Abbildung 4.61: Beispielhafte Momentaufnahmen von (A) PEG, (B) LPG, (C) PMeOxund (D) PEtOx an der Konjugationsstelle 29 des Her2-Affibodys. PEG ist orange, LPG grün,PMeOx blau und PEtOx violett gefärbt. Ausgewählte zusätzliche Aminosäuren in Polymernä-he sind ebenso dargestellt.4.3. Vergleich der Polymere163Abbildung 4.62: Protein-Polymer-Dynamiken an Konjugationsstelle 29 des Her2 Affibody.Die oberste Reihe zeigt Momentaufnahmen aus den Trajektorien jedes Polymers im Intervallvon 10 ns, die erste Momentaufnahme des Proteins ist als Cartoon dargestellt. PEG ist oran-ge, LPG grün, PMeOx blau und PEtOx violett gefärbt. Im letzten Bild der obersten Reihesind diese Momentaufnahmen als Sphären übereinandergelagert. Darunter sind alle Eigen-schaften dargestellt, die in Kap. 3.1.8 geschildert wurden. Die Konjugationsstelle ist im Falleder aminosäurespezifischen Interaktionsenergien (mit -1 multipliziert) als K auf rotem Grundgekennzeichnet.164Kapitel 4. Ergebnisse: Modellproteine4.3.2.8Konjugationsstelle 33PEG interagiert entweder mit N23, Q26, F30, K50, L51 und Q55 (Abb. 4.63-A)oder windet sich um A29. LPG wechselwirkt mit den Aminosäuren F30, L51 undA54, sowie mit E47 mit seinen ersten Hydroxylgruppen (Abb. 4.63-B). PMeOx undPEtOx verhalten sich ähnlich zu PEG und strecken sich nach unten zur Helix (N23,Q26, F30, K50, L51 und Q55) oder verbleiben in einer geknäuelten Konformation(Abb. 4.63-C und Abb. 4.63-D).Abbildung 4.63: Beispielhafte Momentaufnahmen von (A) PEG, (B) LPG, (C) PMeOxund (D) PEtOx an der Konjugationsstelle 33 des Her2-Affibodys. PEG ist orange, LPG grün,PMeOx blau und PEtOx violett gefärbt. Ausgewählte zusätzliche Aminosäuren in Polymernä-he sind ebenso dargestellt.4.3. Vergleich der Polymere165Abbildung 4.64: Protein-Polymer-Dynamiken an Konjugationsstelle 33 des Her2 Affibody.Die oberste Reihe zeigt Momentaufnahmen aus den Trajektorien jedes Polymers im Intervallvon 10 ns, die erste Momentaufnahme des Proteins ist als Cartoon dargestellt. PEG ist oran-ge, LPG grün, PMeOx blau und PEtOx violett gefärbt. Im letzten Bild der obersten Reihesind diese Momentaufnahmen als Sphären übereinandergelagert. Darunter sind alle Eigen-schaften dargestellt, die in Kap. 3.1.8 geschildert wurden. Die Konjugationsstelle ist im Falleder aminosäurespezifischen Interaktionsenergien (mit -1 multipliziert) als K auf rotem Grundgekennzeichnet.166Kapitel 4. Ergebnisse: Modellproteine4.3.2.9Konjugationsstelle 39Während PEG zunächst zwischen Y15 und Y35 positioniert ist, streckt es sich relativfrüh ins Solvens aus. Es verbleibt dann in der Nähe von F5 und windet sich zum Teilum K4 (Abb. 4.65-A). LPG verbleibt ebenso solvensexponiert und abgewandt vonden Helices. Oft interagiert es mit der Seitenkette von D2 über eine HB (Abb. 4.65-B). PMeOx hält sich ebenso in der Nähe von F5 des C-Terminus auf (Abb. 4.65-C),richtet sich jedoch auch teilweise nach unten zu Y15 und Y35 aus. PEtOx ist sehrflexibel, Interaktionen mit F5 sind jedoch auch zu beobachten (Abb. 4.65-D). Dabeihält sich das Polymer auf einer Seite der Helices auf, die von den anderen Polymerennicht adressiert wird (nahe von A42 und L45).Abbildung 4.65: Beispielhafte Momentaufnahmen von (A) PEG, (B) LPG, (C) PMeOxund (D) PEtOx an der Konjugationsstelle 39 des Her2-Affibodys. PEG ist orange, LPG grün,PMeOx blau und PEtOx violett gefärbt. Ausgewählte zusätzliche Aminosäuren in Polymernä-he sind ebenso dargestellt.4.3. Vergleich der Polymere167Abbildung 4.66: Protein-Polymer-Dynamiken an Konjugationsstelle 39 des Her2 Affibody.Die oberste Reihe zeigt Momentaufnahmen aus den Trajektorien jedes Polymers im Intervallvon 10 ns, die erste Momentaufnahme des Proteins ist als Cartoon dargestellt. PEG ist oran-ge, LPG grün, PMeOx blau und PEtOx violett gefärbt. Im letzten Bild der obersten Reihesind diese Momentaufnahmen als Sphären übereinandergelagert. Darunter sind alle Eigen-schaften dargestellt, die in Kap. 3.1.8 geschildert wurden. Die Konjugationsstelle ist im Falleder aminosäurespezifischen Interaktionsenergien (mit -1 multipliziert) als K auf rotem Grundgekennzeichnet.168Kapitel 4. Ergebnisse: Modellproteine4.3.2.10Konjugationsstelle 43Nachdem es zunächst nach unten zu E47, K50 und L51 gerichtet ist (Abb. 4.67-A),etabliert PEG eine geknäuelte Konformation in der zweiten Hälfte der Trajektorie.LPG wechselwirkt über seinen Linker und die ersten Hydroxylgruppen mit E47, überdie terminalen Monomere mit K49. Danach etabliert es zum Teil HB mit D36 (Abb.4.67-B) oder richtet sich nach unten zum hydrophoben Areal von F30 und L51. PMe-Ox fluktuiert zunächst stark, adressiert später jedoch dieselben Aminosäuren (Abb.4.67-C). Dieser Bereich wird auch von PEtOx geschützt (Abb. 4.67-D). Ähnlich zurKonjugationsstelle 39 wechselwirkt das Polymer später auch mit der anderen Seiteder Helix (F5, A42 und L45).Abbildung 4.67: Beispielhafte Momentaufnahmen von (A) PEG, (B) LPG, (C) PMeOxund (D) PEtOx an der Konjugationsstelle 43 des Her2-Affibodys. PEG ist orange, LPG grün,PMeOx blau und PEtOx violett gefärbt. Ausgewählte zusätzliche Aminosäuren in Polymernä-he sind ebenso dargestellt.4.3. Vergleich der Polymere169Abbildung 4.68: Protein-Polymer-Dynamiken an Konjugationsstelle 43 des Her2 Affibody.Die oberste Reihe zeigt Momentaufnahmen aus den Trajektorien jedes Polymers im Intervallvon 10 ns, die erste Momentaufnahme des Proteins ist als Cartoon dargestellt. PEG ist oran-ge, LPG grün, PMeOx blau und PEtOx violett gefärbt. Im letzten Bild der obersten Reihesind diese Momentaufnahmen als Sphären übereinandergelagert. Darunter sind alle Eigen-schaften dargestellt, die in Kap. 3.1.8 geschildert wurden. Die Konjugationsstelle ist im Falleder aminosäurespezifischen Interaktionsenergien (mit -1 multipliziert) als K auf rotem Grundgekennzeichnet.170Kapitel 4. Ergebnisse: Modellproteine4.3.2.11Konjugationsstelle 46Nach einiger Zeit in einer geknäuelten Form und Wechselwirkungen mit N43 richtetsich die PEG-Kette in Richtung F30 aus, während die Amid-Struktur des Linkersmit E47 und die Carbonylgruppe des Linkers mit K50 interagieren kann (Abb. 4.69-A) verbleibt PEG, abgewandt vom Protein, in einer geknäuelten Form. LPG scheintebenso geringe Fluktuationen an dieser Stelle aufzuweisen. Es wechselwirkt durch-gängig mit K49 und D53 (Abb. 4.69-B). Im Gegensatz hierzu wechselt PMeOx häufigdie Position. Nur nahe zu F30 und L44 lassen sich teilweise stabile Konformationenbeobachten (Abb. 4.69-C). Ähnlich zu den Dynamiken an den Konjugationsstellen 39und 43 verbleibt PEtOx in der Nähe von F5, A42 und L45 zwischen den Helices 1und 3 (Abb. 4.69-D). Gegen Ende streckt es sich auch zu N11 von Helix 1 aus.Abbildung 4.69: Beispielhafte Momentaufnahmen von (A) PEG, (B) LPG, (C) PMeOxund (D) PEtOx an der Konjugationsstelle 46 des Her2-Affibodys. PEG ist orange, LPG grün,PMeOx blau und PEtOx violett gefärbt. Ausgewählte zusätzliche Aminosäuren in Polymernä-he sind ebenso dargestellt.4.3. Vergleich der Polymere171Abbildung 4.70: Protein-Polymer-Dynamiken an Konjugationsstelle 46 des Her2 Affibody.Die oberste Reihe zeigt Momentaufnahmen aus den Trajektorien jedes Polymers im Intervallvon 10 ns, die erste Momentaufnahme des Proteins ist als Cartoon dargestellt. PEG ist oran-ge, LPG grün, PMeOx blau und PEtOx violett gefärbt. Im letzten Bild der obersten Reihesind diese Momentaufnahmen als Sphären übereinandergelagert. Darunter sind alle Eigen-schaften dargestellt, die in Kap. 3.1.8 geschildert wurden. Die Konjugationsstelle ist im Falleder aminosäurespezifischen Interaktionsenergien (mit -1 multipliziert) als K auf rotem Grundgekennzeichnet.172Kapitel 4. Ergebnisse: Modellproteine4.3.2.12Konjugationsstelle 47PEG adressiert den hydrophoben Bereich von F30 und L44 (Abb. 4.71-A). LPGverbleibt oberhalb von K49 und bildet, wie bei Konjugationsstelle 46, HB zu D53 aus(Abb. 4.71-B). PMeOx interagiert zunächst über seine terminale Carbonylgruppemit N23 (Abb. 4.71-C) und wechselwirkt dann gegen Ende mit Q26, F30 und L54.PEtOx interagiert mit ähnlichen Aminosäuren. Nach der Hälfte der Simulationszeitlassen sich Konformationen beobachten, bei denen das Polymer zwischen R32 undD36 positioniert ist (Abb. 4.71-D).Abbildung 4.71: Beispielhafte Momentaufnahmen von (A) PEG, (B) LPG, (C) PMeOxund (D) PEtOx an der Konjugationsstelle 47 des Her2-Affibodys. PEG ist orange, LPG grün,PMeOx blau und PEtOx violett gefärbt. Ausgewählte zusätzliche Aminosäuren in Polymernä-he sind ebenso dargestellt.4.3. Vergleich der Polymere173Abbildung 4.72: Protein-Polymer-Dynamiken an Konjugationsstelle 47 des Her2 Affibody.Die oberste Reihe zeigt Momentaufnahmen aus den Trajektorien jedes Polymers im Intervallvon 10 ns, die erste Momentaufnahme des Proteins ist als Cartoon dargestellt. PEG ist oran-ge, LPG grün, PMeOx blau und PEtOx violett gefärbt. Im letzten Bild der obersten Reihesind diese Momentaufnahmen als Sphären übereinandergelagert. Darunter sind alle Eigen-schaften dargestellt, die in Kap. 3.1.8 geschildert wurden. Die Konjugationsstelle ist im Falleder aminosäurespezifischen Interaktionsenergien (mit -1 multipliziert) als K auf rotem Grundgekennzeichnet.174Kapitel 4. Ergebnisse: Modellproteine4.3.2.13Konjugationsstelle 49Neben einigen geknäuelten Konformationen hält sich PEG entweder neben N43 oderoberhalb von A42 und L45 auf (Abb. 4.73-A). LPG bildet mit seinen Hydroxylgruppenzumeist HB mit E47, während sich die Ethersauerstoffatome um K50 winden. Späteradressiert das Polymer die Carboxylgruppen von E8 und E15 (Abb. 4.73-B). PMeOxhält sich zunächst zwischen den Aminosäuren R14, E15, S41 und L45 auf, bevor eseine stabile Konformation über eine Carbonylgruppe des Linkers mit N3 etabliert.Ebenso bildet eine Carbonylgruppe des Polymers Interaktionen mit dem Rückgratvon N43 aus (Abb. 4.73-C). Im Falle von PEtOx ist die Proteinkonformation deutlichmodifiziert, weil das Polymer sich um den N-Terminus windet. HB zu N3 und F5können beobachtet werden. Die Seitenkette von F5 wird durch das Polymer ebensogeschützt (Abb. 4.73-D).Abbildung 4.73: Beispielhafte Momentaufnahmen von (A) PEG, (B) LPG, (C) PMeOxund (D) PEtOx an der Konjugationsstelle 49 des Her2-Affibodys. PEG ist orange, LPG grün,PMeOx blau und PEtOx violett gefärbt. Ausgewählte zusätzliche Aminosäuren in Polymernä-he sind ebenso dargestellt.4.3. Vergleich der Polymere175Abbildung 4.74: Protein-Polymer-Dynamiken an Konjugationsstelle 49 des Her2 Affibody.Die oberste Reihe zeigt Momentaufnahmen aus den Trajektorien jedes Polymers im Intervallvon 10 ns, die erste Momentaufnahme des Proteins ist als Cartoon dargestellt. PEG ist oran-ge, LPG grün, PMeOx blau und PEtOx violett gefärbt. Im letzten Bild der obersten Reihesind diese Momentaufnahmen als Sphären übereinandergelagert. Darunter sind alle Eigen-schaften dargestellt, die in Kap. 3.1.8 geschildert wurden. Die Konjugationsstelle ist im Falleder aminosäurespezifischen Interaktionsenergien (mit -1 multipliziert) als K auf rotem Grundgekennzeichnet.176Kapitel 4. Ergebnisse: Modellproteine4.3.2.14Konjugationsstelle 50PEG fluktuiert stark, wobei stabile Konformationen in der Nähe von A29, F30, S33und L51 (Abb. 4.75-A), sowie von Q40 und N43 beobachtet werden können. LPGwechselwirkt mit einer Vielzahl an Seitenketten in der Nähe des hydrophoben Be-reichs: S33, E36, N40 und D47 (Abb. 4.75-B). Am Ende der Simulation streckt sichdas Polymer entlang der Helix aus, ähnlich zu PEG. Im Falle von PMeOx werdenzunächst A42 und N43 adressiert (Abb. 4.75-C), bevor sich das Polymer, wie auchdie anderen Varianten, nach unten zu F30 ausrichtet. Letztere Aminosäure wird auchvon PEtOx zu Beginn der Simulation adressiert. Danach orientiert sich das Polymerparallel zur Helix aus. Dabei kann K58 mit dem Linker wechselwirken (Abb. 4.75-D).Abbildung 4.75: Beispielhafte Momentaufnahmen von (A) PEG, (B) LPG, (C) PMeOxund (D) PEtOx an der Konjugationsstelle 50 des Her2-Affibodys. PEG ist orange, LPG grün,PMeOx blau und PEtOx violett gefärbt. Ausgewählte zusätzliche Aminosäuren in Polymernä-he sind ebenso dargestellt.4.3. Vergleich der Polymere177Abbildung 4.76: Protein-Polymer-Dynamiken an Konjugationsstelle 50 des Her2 Affibody.Die oberste Reihe zeigt Momentaufnahmen aus den Trajektorien jedes Polymers im Intervallvon 10 ns, die erste Momentaufnahme des Proteins ist als Cartoon dargestellt. PEG ist oran-ge, LPG grün, PMeOx blau und PEtOx violett gefärbt. Im letzten Bild der obersten Reihesind diese Momentaufnahmen als Sphären übereinandergelagert. Darunter sind alle Eigen-schaften dargestellt, die in Kap. 3.1.8 geschildert wurden. Die Konjugationsstelle ist im Falleder aminosäurespezifischen Interaktionsenergien (mit -1 multipliziert) als K auf rotem Grundgekennzeichnet.178Kapitel 4. Ergebnisse: Modellproteine4.3.2.15Konjugationsstelle 53Nach wenigen Nanosekunden windet sich PEG um K49 und verändert seine Kon-formation nicht mehr (Abb. 4.77-A). In vergleichbarer Weise interagiert LPG mitK49 über seine Ethergruppen, während die Hydroxylgruppen E15 adressieren (Abb.4.77-B). Gegen Ende streckt sich das Polymer nach oben und interagiert mit einerCarbonylgruppe des Linkers mit K50 und den Hydroxylgruppen mit E8. Beide POx-Varianten bleiben solvensexponiert und richten sich teilweise parallel zur Helix aus,wobei F5 adressiert werden kann (Abb. 4.77-C und 4.77-D).Abbildung 4.77: Beispielhafte Momentaufnahmen von (A) PEG, (B) LPG, (C) PMeOxund (D) PEtOx an der Konjugationsstelle 53 des Her2-Affibodys. PEG ist orange, LPG grün,PMeOx blau und PEtOx violett gefärbt. Ausgewählte zusätzliche Aminosäuren in Polymernä-he sind ebenso dargestellt.4.3. Vergleich der Polymere179Abbildung 4.78: Protein-Polymer-Dynamiken an Konjugationsstelle 53 des Her2 Affibody.Die oberste Reihe zeigt Momentaufnahmen aus den Trajektorien jedes Polymers im Intervallvon 10 ns, die erste Momentaufnahme des Proteins ist als Cartoon dargestellt. PEG ist oran-ge, LPG grün, PMeOx blau und PEtOx violett gefärbt. Im letzten Bild der obersten Reihesind diese Momentaufnahmen als Sphären übereinandergelagert. Darunter sind alle Eigen-schaften dargestellt, die in Kap. 3.1.8 geschildert wurden. Die Konjugationsstelle ist im Falleder aminosäurespezifischen Interaktionsenergien (mit -1 multipliziert) als K auf rotem Grundgekennzeichnet.180Kapitel 4. Ergebnisse: Modellproteine4.3.3Zusammenfassender Vergleich4.3.3.1Pin 1 WW-Domäne4.3.3.1.1Vergleich der berechneten EigenschaftenFür eine zusammenfassende Übersicht der ermittelten Eigenschaften sind die Mittel-werte aller Konjugationsstellen in Abb. 4.79 für die Pin 1 WW-Domäne dargestellt.Zum Vergleich einzelner Konjugate sind die in Kap. 4.3.1 gezeigten Durchschnitts-werte weiterhin in Abb. 4.80 aufgeführt. Eine tabellarische Übersicht der ermitteltenInteraktionsenergien ist für jede Simulation in der Abb. 4.81 zu finden.Beim Blick auf die Mittelwerte in Abb. 4.79 lässt sich zunächst anmerken, dass einigeEigenschaften wie bspw. die Interaktionsenergie pro Polymeroberfläche erwartungsge-mäß größere Werte für das kürzere PEG-Polymer zeigen. Durch die geringere Längeist ein größerer Anteil innerhalb der Proteinnähe zu verorten, während die Wahr-scheinlichkeit proteinferner Strukturen bei den anderen Polymeren mit jeweils fünfMonomeren deutlich größer ist. Insofern stellt sich grundsätzlich die Frage, in welcherForm die Polymerketten verschiedener Länge vergleichend gegenübergestellt werdensollten. Im Falle größerer Konjugate (siehe Kap. 5) werden in der Regel Polymere ähn-licher Molekulargewichte miteinander verglichen, die aufgrund der unterschiedlichenKonstitutionen ebenso verschiedene Längen besitzen.Ungeachtet dessen zeigt PEG auf Basis dieser Berechnungen die insgesamt größteSolvatation (Anzahl der in der Nähe befindlichen Wassermoleküle). Der Anteil anbedeckter Proteinoberfläche ist am größten im Falle der sterisch anspruchsvollerenPOx-Polymere. Diese führen ebenso insgesamt zu einer leichten Erniedrigung derRückgratflexibilität des Proteins. Dagegen zeigen die etherbasierten PolymervariantenPEG und LPG deutlich mehr HB zum Protein. PEG zeigt weiterhin deutlich geringereRMSF-Werte als die übrigen Polymerarten. LPG stellt dagegen das Polymer mit dengrößten Fluktuationen dar.Beim Vergleich der einzelnen Konjugationsstellen (Abb. 4.80) zeigt sich, dass alle Po-lymerarten bei den in den Faltblättern zu verortenden Stellen (14, 23, 26, 32) einegrößere Bedeckung aufweisen, als es bei den Konjugationen an den flexiblen Schleifender Fall ist (Abb. 4.80-A, 4.80-B). Für PEG und LPG ist die Anzahl an HB zumProtein besonders für die im zweiten Faltblatt gelegenen Konjugationsstellen 23 und26 relativ groß (Abb. 4.80-C). Im Gegensatz zu den Varianten, die an den Faltblätternkonjugiert sind, weisen jene mit Konjugationen an der ersten und zweiten Schleife einestärkere Solvatation von PEG und LPG auf (Abb. 4.80-D). Erhöhte Fluktuationendes Proteinrückgrats im Gegensatz zu den unkonjugierten Varianten sind besondersfür Konjugate an den Stellen 16, 17 und 18 zu finden (Abb. 4.80-E). PEG mit dengeringsten Fluktuationen zeigt keine großen Unterschiede für die RMSF-Werte derverschiedenen Konjugationsstellen (Abb. 4.80-F). Besonders große Unterschiede in4.3. Vergleich der Polymere181Abhängigkeit der Konjugationsstelle sind dagegen für PEtOx zu finden. Dieses Po-lymer zeigt für die Stellen 19, 21, 23 und 29 relativ geringe, für die Stellen 16, 17,26, 27 und 28 relativ hohe Fluktuationen. Übereinstimmend mit den Beobachtungenzur Bedeckung der Polymere bzw. des Proteins sind die Interaktionsnergien zwischenProtein und Polymer für die Konjugationsstellen an den Faltblättern relativ groß(Abb. 4.80-G).Im Hinblick auf die favorisierten Wechselwirkungsbereiche der Polymere (Abb. 4.81)interagiert PEG bei den Konjugationsstellen 14, 16, 17, 18, 19 und 21 mit Amino-säuren der ersten Schleife und des dritten Faltblatts (vgl. Beschreibungen in Kap.4.3.1). Konjugationen im zweiten Faltblatt und der zweiten Schleife führen dagegenzu Wechselwirkungen im Bereich von W11 und E12. Das Profil von LPG ähnelt je-nem von PEG. Die POx-Varianten zeigen starke Wechselwirkungen mit Aminosäurendes zweiten und dritten Faltblatts (Y23, F25, S32, F34), nur Konjugate der zweitenSchleife zeigen, wie bei PEG und LPG, erhöhte Wechselwirkungen mit Bereichendes N-Terminus und des ersten Faltblatts (W11 und E12). Die Protein-Polymer-Wechselwirkungen und die Dynamik der einzelnen Polymerarten sollen im Folgendennoch näher charakterisiert werden.Abbildung 4.79: Mittelwerte untersuchter Eigenschaften für die Pin 1-Konjugate (aus Wer-ten der Abbildungen in Kap. 4.3.1). Daten für PEG sind orange, für LPG grün, für PMeOxblau und für PEtOx violett gefärbt. Fehlerbalken zeigen den Standardfehler des Mittelwerts.182Kapitel 4. Ergebnisse: ModellproteineAbbildung 4.80: Übersicht über die Mittelwerte der berechneten Eigenschaften der Pin 1WW-Konjugate, aufgetragen für alle vier Polymere für jede Konjugationsstelle. Fehlerbalkenzeigen den Standardfehler des Mittelwerts.4.3. Vergleich der Polymere183Abbildung 4.81: Durchschnittliche Interaktionsenergien (mit -1 multipliziert, gefärbt vonweiß nach rot) zwischen den Polymeren an jeder Konjugationsstelle (Reihen) und jeder ein-zelnen Aminosäure der Pin 1 WW-Domäne (Spalten). Sekundärstrukturelemente sind blaugekennzeichnet. Die Interaktionsenergie für die Konjugationsstelle wurde jeweils auf Null ge-setzt.184Kapitel 4. Ergebnisse: Modellproteine4.3.3.1.2PEG und LPGDie Polymere PEG und LPG bestehen aus denselben Etherrückgraten und zeigen ins-gesamt sehr ähnliche Wechselwirkungsprofile. An vielen Konjugationsstellen (14, 17und 21) zeigen sich, besonders bei PEG, bevorzugte Konformationen in der Nähe derhydrophoben Aminosäuren Y23 und F25 unter gleichzeitiger Ausbildung von HB mitder Seitenkette von S32 (Abb. 4.82-A). Durch das Etherrückgrat werden somit hydro-phobe Areale adressiert, in denen gleichzeitig einzelne HB ausgebildet werden können(mit Y23 oder S32). Dies entspricht den Beobachtungen aus [107]. Sind die Polymerenicht in diesem Bereich vorzufinden, so lassen sich Interaktionen mit Argininen (R17,R21) oder Serinen der ersten Schleife (S18, S19) beobachten (Abb. 4.82-B).Abbildung 4.82: (A) Ausrichtung der PEG-Kette an der Konjugationsstelle 14 zwischenhydrophoben Aminosäuren hin zu S32. (B) Interaktion von PEG mit benachbartem Argininund Serin an Konjugationsstelle 16.Im Falle des hydrophileren LPG, welches über eine größere Anzahl an HB ausbil-denden Strukturelementen verfügt, sind die Interaktionen mit R17 teils etwas ausge-prägter. Auch stärkere intrapolymere Wechselwirkungen und Interaktionen mit demSolvens sind in vielen Simulationen über die zusätzlichen Hydroxylgruppen beob-achtbar (Abb. 4.83-A). Aufgrund des gleichen Polymerrückgrats ist jedoch nur seltenein deutlich unterschiedliches Verhalten von PEG zu LPG zu erkennen. So strecktsich letzteres im Falle der Konjugationsstelle 19 über F34 hinaus zur Carboxylgruppevon E35, ersteres hingegen adressiert diesen Bereich nicht. Über seine endständigenHydroxylgruppen kann LPG, wenn auch nur sehr kurz, mit dieser Aminosäure wech-selwirken (Abb. 4.83-B). Dies stellt jedoch, neben E12, die einzige Carboxylgruppeinnerhalb der WW-Domäne dar. Letztere wird an der benachbarten Konjugationsstel-le 27 ebenfalls adressiert. Im Falle der Her2-Simulationen sind solche Interaktionenhäufiger beobachtbar (vide infra).4.3. Vergleich der Polymere185Abbildung 4.83: (A) Interaktionen der LPG-Kette mit den eigenen Monomeren (Konjuga-tionsstelle 17). (B) Interaktionen von LPG mit R21 und E35 an Konjugationsstelle 19.4.3.3.1.3Effekt des LinkersAn einigen Konjugationsstellen wird auch deutlich, inwiefern die Dynamik des jeweili-gen Polymers von den Interaktionen des Linkers bestimmt wird. So ist der Asparagin-Linker entscheidend an Wechselwirkungen mit E12 (Abb. 4.84-A) und W11 (Abb.4.84-B) an den Positionen 26 bzw. 27 beteiligt und bestimmt somit die initiale Aus-richtung der Polymerkette. Die Auswahl des Linkers kann also durchaus einen großenEffekt auf die Polymerdynamik haben. Im Falle der Konjugationsstelle 26 ist dieNH-π-Wechselwirkung des Linkers mit W11 für alle Polymervarianten beobachtbar.Diese Interaktion wurde von Price et al. in deren neuester Studie näher beleuchtet[104] und konnte in der hier vorliegenden Arbeit somit wiedergefunden werden, wo-bei eine polymerbasierte Desolvatation in diesem Areal den Autoren zufolge zu einerStärkung der Wechselwirkung führen sollte.Abbildung 4.84: (A) Interaktionen des Asparagin-Linkers mit der Seitenkette von E12(PEG-Konjugationsstelle 27). (B) Interaktionen des Asparagin-Linkers mit W11 an LPG-Konjugationsstelle 19.186Kapitel 4. Ergebnisse: Modellproteine4.3.3.1.4PMeOx und PEtOxDie POx-Varianten zeigen eine teils deutlich von PEG und LPG abweichende Dy-namik, da sie strukturell grundlegend verschieden sind. Es zeigt sich insgesamt ei-ne stärkere Tendenz zu Wechselwirkungen mit aromatischen Strukturen, was auchdem Interaktionsprofil in Abb. 4.79 zu entnehmen ist. Dabei sind Konformationenzu erkennen, bei denen die Seitenketten der POx-Polymere auf den aromatischenRingen, besonders von Y23, F25 und F34, aufliegen (Abb. 4.85). Diese Wechselwir-kungen basieren höchstwahrscheinlich auf den hydrophoben Alkylgruppen der Sei-tenketten. Ähnliche Konformationen für PMeOx wurden kürzlich auch in einer Si-mulation einer Poly(2-methyl-2-oxazolin)-b-poly(2-phenyl-2-oxazin)-b-poly(2-methyl-2-oxazolin)-basierten Wurmmizelle beobachtet, wobei hierbei ebenfalls die Hypothesemöglicher nAmid→π*Aromat- und πAmid···πAromat-Wechselwirkungen aufgestellt wur-de [322]; diese spielen jedoch vermutlich, im Vergleich zu hydrophoben Effekten, beidiesen Oligomeren und einer geringen Anzahl möglicher aromatischer Wechselwir-kungspartner nur eine untergeordnete Rolle. Nichtkovalente n→π*-Wechselwirkungenwurden in der Vergangenheit bereits bzgl. strukturell ähnlicher Peptoide und des pro-linreichen Proteins Kollagen beschrieben [323, 324].Abbildung 4.85: Beispiele für POx-Konformationen, bei denen die Monomere mit aromati-schen Aminosäuren wechselwirken. (A) Interaktionen von PMeOx an der Konjugationsstelle16. (B) Interaktionen von PEtOx an Konjugationsstelle 19.Auch wenn diese Interaktionen an einigen Konjugationsstellen für beide POx-Variantenbeobachtet werden (14, 16, 21, 32), so zeigen PMeOx und PEtOx teilweise auch einedeutlich unterschiedliche Dynamik. Exemplarisch sollten hier die Stellen 18 und 19hervorgehoben werden (Abb. 4.86-A und 4.86-B). Während sich PMeOx an der Stelle18 zu F34 hin orientiert, wendet sich PEtOx zu den Aminosäuren K13 und M15 aufder Rückseite der ersten Schleife hin. An der Stelle 19 ist teilweise das Gegenteil zubeobachten: PMeOx zeigt Konformationen in der Nähe von M15, dagegen werden hierY23 und F34 deutlich öfter von PEtOx adressiert. Weiterhin kann PMeOx das oberehydrophobe Zentrum im Falle der Konjugationsstelle 30 erreichen, PEtOx verbleibtim Bereich der zweiten Schleife in der Nähe von H27. Dagegen adressiert PEtOx das4.3. Vergleich der Polymere187hydrophobe Zentrum an der Konjugationsstelle 27, PMeOx hingegen verbleibt hier,durch Wechselwirkungen des Linkers mit E12, in der Nähe der zweiten Schleife (Abb.4.86-C und 4.86-D).Diese beiden Gegenüberstellungen (Konjugationsstelle 18 vs. 19, 27 vs. 30) deutendarauf hin, dass die Distanz der Konjugationsstelle zu favorisierten Wechselwirkungs-partnern das Verhalten beider POx-Derivate, zumindest im Falle kürzerer Oligomere,entscheidend beeinflussen kann. Die Rückgratlängen beider Varianten sind gleich, diePEtOx-Seitenketten sind jedoch hydrophober und sterisch anspruchsvoller. Im Falleder Stellen 18 und 30 könnte die Konjugationsstelle für das größere PEtOx-Oligomeraufgrund sterischer Effekte zu nah an den Aminosäuren Y23 und F34 positioniert sein,während diese im Falle der etwas weiter entfernt gelegenen Stellen 19 und 27 durchdie größeren PEtOx-Seitenketten besser erreicht werden können als es bei PMeOxder Fall ist.Neben Wechselwirkungen mit aromatischen Gruppen zeigen sich auch viele Interak-tionen über die Carbonylgruppen der POx-Derivate, die prinzipiell als HB-Akzeptorenzur Verfügung stehen (z. B. bei PMeOx für R17 an den Konjugationsstellen 14 und 23,bei PEtOx für R14, R17 und N30 an Konjugationsstelle 27). So lassen sich auch HBzu Hydroxylgruppen von bspw. S32 und Y23 beobachten. Das Wechselwirkungspo-tenzial der eher polaren Amid-Strukturelemente der Poly(2-alkyl-2-oxazoline) solltedaher nicht vernachlässigt werden. Die Relevanz dieser Interaktionen wurde in derLiteratur im Falle POx-basierter, wirkstoffbeladener Mizellen bereits erwähnt [325].4.3.3.2Her2-Affibody4.3.3.2.1Vergleich der berechneten EigenschaftenAnalog zu den Konjugaten der Pin 1 WW-Domäne sind die Mittelwerte der Eigen-schaften in Abb. 4.87 gezeigt, jene für die einzelnen Konjugationsstellen in Abb. 4.88,sowie die Interaktionsenergien der Polymere mit den Aminosäuren in einer tabellari-schen Form in der Abb. 4.89.Wie bei der Pin 1 WW-Domäne zeigen die POx-Varianten im Hinblick auf die Mittel-werte der Konjugationsstellen (Abb. 4.87) die größte Proteinbedeckung und Wechsel-wirkungen mit aromatischen Aminosäuren (F5, W14, F30). Die kürzere PEG-Kettezeigt abermals die größte prozentuale Bedeckung. PEG und LPG zeigen mehr HB mitdem Protein und eine im Vergleich zu den POx-Varianten höhere Solvatation. AllePolymervarianten zeigen bei diesem Peptid größere RMSF-Werte im Verhältnis zuden Simulationen unkonjugierter Varianten. Auch hier zeigt PEG deutlich geringereRMSF-Werte im Vergleich zu den Polymeralternativen und die größte Interaktions-energie in Relation zur verfügbaren Polymeroberfläche.188Kapitel 4. Ergebnisse: ModellproteineAbbildung 4.86: Vergleich der POx-Interaktionen mit aromatischen Aminosäuren an denKonjugationsstellen (A) 18, (B) 19, (C) 27 und (D) 30. PMeOx ist türkis gefärbt, PEtOxviolett.4.3. Vergleich der Polymere189Vergleicht man die einzelnen Konjugationsstellen miteinander (Abb. 4.88) , so lässtsich, in Relation zur Pin 1 WW-Domäne (vgl. Abb. 4.80), keine Abhängigkeit derEigenschaften von der Sekundärstruktur des Konjugationsbereichs erkennen. DerBedeckungsgrad des Proteins und des jeweiligen Polymers (Abb. 4.88-A und 4.88-B), sowie die entsprechende Anzahl an HB (Abb. 4.88-C) und die Protein-Polymer-Interaktionsenergie (Abb. 4.88-G) sind aber im Falle der Stelle 39 zwischen Helices2 und 3 relativ gering. Bei PEG kommt es vermehrt zu erhöhten Fluktuationen desProteinrückgrats (Abb. 4.88-E, Stellen 21, 33, 46, 53), dies lässt sich für die anderenPolymere nur vereinzelt beobachten (für LPG an Stelle 50, für PMeOx an Stelle 49,für PEtOx an den Stellen 46 und 50).Die Tabelle zu den Protein-Polymer-Wechselwirkungen (Abb. 4.89) zeigt hauptsäch-lich erhöhte Interaktionen mit Aminosäuren in der Nähe der Konjugationsstelle (d. h.Konjugationen an Helix 1 führen bspw. zu Wechselwirkungen mit der ersten Helix).Die POx-Varianten zeigen teilweise starke Wechselwirkungen mit W14, sowie favori-sierte Interaktionen mit F5, auch bei Konjugationen an der dritten Helix. Auch fürden Affibody soll das Verhalten der einzelnen Polymere im Folgenden näher beschrie-ben werden.Abbildung 4.87: Mittelwerte untersuchter Eigenschaften für die Her2-Konjugate (aus Wer-ten der Abbildungen in Kap. 4.3.2). Daten für PEG sind orange, für LPG grün, für PMeOxblau und für PEtOx violett gefärbt. Fehlerbalken zeigen den Standardfehler des Mittelwerts.190Kapitel 4. Ergebnisse: ModellproteineAbbildung 4.88: Übersicht über die Mittelwerte der berechneten Eigenschaften derAffibody-Konjugate, aufgetragen für alle vier Polymere für jede Konjugationsstelle. Fehler-balken zeigen den Standardfehler des Mittelwerts.4.3. Vergleich der Polymere191Abbildung 4.89: Durchschnittliche Interaktionsenergien (mit -1 multipliziert, gefärbt vonweiß nach rot) zwischen den Polymeren an jeder Konjugationsstelle (Reihen) und jeder einzel-nen Aminosäure des Her2-Affibody (Spalten). Sekundärstrukturelemente sind blau gekenn-zeichnet. Die Interaktionsenergie für die Konjugationsstelle wurde jeweils auf Null gesetzt.192Kapitel 4. Ergebnisse: Modellproteine4.3.3.2.2PEGDurch die deutlich größere strukturelle Vielfalt des Her2-Affibodys im Vergleich zurkleineren WW-Domäne ergeben sich teils andere Wechselwirkungen. Können, in An-lehnung an Beobachtungen aus [105–108], Interaktionen der PEG-Kette und / oderdes Maleimid-Linkers mit Lysinen, besonders benachbarter an den Positionen i + 3und i + 4, beobachtet werden? Wechselwirkungen von PEG mit Lysinen lassen sichin der Tat sehr viel öfter als im Falle der WW-Domäne beobachten. In dieser lassensich auch nur zwei Lysine finden, von denen eines am N-Terminus (K6) und das an-dere auf der dem oberen hydrophoben Zentrum abgewandten Seite positioniert ist.Interaktionen mit diesen wurden daher erwartungsgemäß wenige gefunden. Im Falledes Her2-Affibodys jedoch lassen sich an vielen Konjugationsstellen entsprechendeWechselwirkungen finden. So interagiert PEG bspw. mit K49 und K50 an den Stellen7 und 8 (Abb. 4.90-A), mit K4 an Position 39 und mit K50 an Position 43.Abbildung 4.90: (A) PEG-Interaktion mit K49 an Konjugationsstelle 8. (B) Interaktionenvon PEG und den Carbonylgruppen des Maleimid-Linkers mit K49 an Konjugationsstelle 53.(C) Bedeckung des hydrophoben Areals um F30 und L44 durch die PEG-Kette an Konju-gationsstelle 47. (D) Exemplarische PEG-Konformation in der Nähe der Aminosäuren Q25,Q26 und Q55 im Bereich der Schleife zwischen Helices 1 und 2.4.3. Vergleich der Polymere193Lysine an den Stellen i + 3 und i + 4 sind im Falle der Konjugationsstellen 46 und53 vorzufinden (K49 und K50). An ersterer sind keine starken Wechselwirkungen derPEG-Kette mit Lysinen zu beobachten, an letzterer jedoch interagiert das Polymermit K49. Interaktionen der Carbonylgruppen des Maleimid-Linkers, wie von Price etal. postuliert, werden sowohl an Konjugationsstelle 46 als auch an Stelle 53 beobach-tet (Abb. 4.90-B). Auch an Position 39 interagiert eine Carbonylgruppe des Linkersteilweise mit einem nahegelegenen Lysin (K4). Inwiefern diese Interaktionen durcheine PEGylierung beeinträchtigt werden, müsste durch zusätzliche Simulationen derjeweiligen Cys-Maleimid-Strukturen untersucht werden. Dies zeigt jedoch, dass derCys-Maleimid-Linker, wie auch der Asparagin-Linker der WW-Domäne, bedeutendeWechselwirkungen mit dem Peptid ausbilden kann, die bei einer Konjugation berück-sichtigt werden müssen.An besonders vielen Stellen wird das Areal um F30 und L44 adressiert (Abb. 4.90-C).Ähnlich zu PEG-Konformationen nahe Y23 und F34 der WW-Domäne werden somiteinige apolare Proteinbereiche desolvatisiert. Es existieren weiterhin auch viele Inter-aktionen mit positiv geladenen Argininen (z. B. R10 an Position 7, R28 und R32 anPosition 29), ähnlich zu jenen mit R17 und R21 der WW-Domäne. Viele Konjugati-onsstellen (21, 23, 25) befinden sich nahe der Schleife zwischen der ersten und zweitenHelix. Hier finden sich eine Vielzahl an Glutaminen und Asparaginen (N23, N24, Q25,Q26, Q55). Diese können als HB-Donatoren fungieren, entsprechend sind starke In-teraktionsenergien registrierbar (Abb. 4.90-D). Gleichzeitig könnte die Stabilität desPeptids ebenso durch eine desolvatationsbasierte Stärkung der intramolekularen HB,bspw. zwischen Q26 und Q55, moduliert werden, ähnlich wie es Price et al. im Fallevon Salzbrücken postulierten [102].4.3.3.2.3LPGAuch im Falle von LPG lassen sich einige Interaktionen mit Lysinen finden, so z.B. an Position 46 mit K49 (wobei hier ebenso der Maleimid-Linker mit dieser Ami-nosäure wechselwirkt) oder an Position 8 mit K4 (Abb. 4.91-A). Ebenso wird dashydrophobe Areal F30 / L44 adressiert. Am deutlichsten jedoch lassen sich Wechsel-wirkungen der Hydroxylgruppen mit Aspartaten und Glutamaten finden. Die Anzahlan Carboxylgruppen dieser Aminosäuren ist in der WW-Domäne sehr begrenzt. ImFalle des Her2-Affibodys lassen sich jedoch zahlreiche Interaktionen an verschiedenenKonjugationsstellen beobachten: so z. B. an Position 8 mit D2 und E15, an Position33 mit E47 und an Position 43 mit D36 und E47 (Abb. 4.91-B). Die funktionellenGruppen der Aminosäuren N23, N24, Q25, Q26 und Q55 können, ähnlich wie beiPEG, erwartungsgemäß auch als Wechselwirkungspartner dienen.194Kapitel 4. Ergebnisse: ModellproteineAbbildung 4.91: (A) LPG-Interaktionen mit K4 an Konjugationsstelle 8. (B) LPG-Interaktionen mit D36 und E47 an Position 43.4.3.3.2.4PMeOx und PEtOxWie bei der WW-Domäne lassen sich zahlreiche Interaktionen mit aromatischenGruppen beobachten, was auch dem Interaktionsprofil in Abb. 4.87 zu entnehmenist. So werden bspw. F5 durch Konjugation an Position 8, Y13 und W14 durch Kon-jugation an Position 7, sowie F30 und Y35 durch Konjugation an der Position 25adressiert (Abb. 4.92). Es lassen sich, wie bei den anderen Polymervarianten, vieleKonformationen in der Nähe des solvensexponierten hydrophoben Areals von F30 /L44 / L45 finden. Analog zu den Konjugaten der WW-Domäne können die Polymereüber ihre Carbonylgruppen mit HB-ausbildenden funktionellen Gruppen des Peptidswechselwirken, so z. B. mit N23 an Position 33 oder Q55 an Position 47.Abbildung 4.92: (A) PMeOx- und (B) PEtOx-Interaktionen an Konjugationsstelle 15.4.4. Zusammenfassung1954.4ZusammenfassungIn Kap. 4.1.1 wurden verschiedene Modellarten des maschinellen Lernens zur Gene-rierung von Vorhersagemodellen zur PEG-basierten thermodynamischen Stabilisie-rung getestet. Eine fehlende Limitierung der Anzahl genutzter Deskriptoren führtebei Pin 1-spezifischen Modellen zum kompletten Overfitting. Auch wenn die Pin 1-unspezifischen Modelle mit einer deutlich geringeren Anzahl an Deskriptoren einesehr gute Leistung bzgl. eines Validierungssets zeigten, so deuten die Beobachtun-gen auch hier teilweise auf ein Overfitting hin. Weiterhin wurde das jeweils besteModell auf Basis des Validierungssets ausgewählt, sodass dieses nicht als unvorein-genommen genug angesehen werden kann. Der entwickelte Ansatz diente jedoch alsVorarbeit zu den Modellen in Kap. 4.1.2 und konnte aufzeigen, dass die Nutzbarkeitsolcher komplexerer Modellarten bei der vorliegenden Datensatzgröße sehr begrenztist. Modelldeskriptoren beschreiben sowohl spezifische Protein-PEG-Interaktionen alsauch Modulationen der Solvatation in verschiedenen Hydrathüllen. Verschiedene Mo-delldeskriptoren wurden zum exemplarischen Vergleich strukturell unterschiedlicherKonjugate (z. B. mit Linkern verschiedener Rigidität) mithilfe visueller Analysender Polymerdynamik und der Proteinsolvatation näher untersucht. Dies signalisiertdie Eignung von MDs als Methode zur Abbildung des Verhaltens der Konjugate aufmolekularer Ebene.Der Modellgenerierungsansatz in Kap. 4.1.2, welcher sich auf simplere Modelltypenbegrenzte, führte schließlich zu den Modellen A2-R6 und A2-C1. Modell A2-R6 zeigteeine bei einem deutlich erweiterten Datensatz im Vergleich zur kristallstrukturbasier-ten Richtlinie von Price et al. allgemein verbesserte Leistung. Auch für ein Konjugatder strukturell der WW-Domäne ähnlichen Src SH3-Domäne konnte eine optimierteVorhersage erhalten werden. Modell A2-C1 stufte dieses Konjugate ebenfalls als stabi-lisiert ein. Während Modell A2-R6 ähnliche Deskriptoren zu den Pin 1-unspezifischenRegressionsmodellen in Kap. 4.1.1 nutzte, legte Modell A2-C1 den Fokus auf die Sol-vatation der PEG-Kette. Die im Anschluss zusätzlich generierten Modelle A2v2-R1und A2v2-C1 mit einem aktualisierten Datensatz 1 zeigten vergleichbar gute Statisti-ken und nutzten verschiedene Abwandlungen des Price-Winkels als Deskriptoren. Innahezu allen Modellen aus Kap. 4.1 wurde der Deskriptor molbur, der die durch PEGbedeckte Proteinoberfläche beschreibt, als besonders bedeutend eingestuft. Es stelltsich jedoch die Frage, ob gewisse Einstellungen im Modellgenerierungsansatz (z. B.die Filterschritte bzgl. der Deskriptoren) weiter optimiert werden können. Vorhersa-gen zu alternativen Polymerarten deuten darauf hin, dass für diese separate Modelleentwickelt werden müssten.Generell stellt die für QSPR-Ansätze extrem kleine Anzahl an Konjugaten in ei-nem äußerst engen Energieintervall die für diese Studie entscheidende Limitation dar.Deutlich mehr experimentelle Untersuchungen sind notwendig, um optimierte Model-le zu entwickeln. Eine QSPR-Modellgenerierung wurde dennoch vorgenommen, da die196Kapitel 4. Ergebnisse: ModellproteineBedeutung verschiedener Deskriptoren anhand der Modelle interpretiert werden kannund die energetischen Differenzen für eine in silico-basierte Berechnung mithilfe deut-lich rechenintensiverer, erweiterter Simulationsmethoden (siehe Kap. 4.2) sehr geringsind bzw. im Bereich der für solche Methoden grundsätzlich angenommenen Fehler-grenzen liegen. Ein Fokus auf andere, prinzipiell aus Simulationen sehr viel leichterzugängliche Eigenschaften, wie bspw. die sterische Abschirmung der Binderegionenvon Proteinen durch PEG, wäre denkbar. Allerdings standen für diese Studie lediglichdie von Price et al. publizierten ∆∆G0f-Werte und keine Messwerte zu Bioaktivitätenzur Verfügung.In Kap. 4.2 wurde ein zu den QSPR-Modellen grundlegend anderer Ansatz gewählt,um den Effekt auf die konformative Stabilität zu untersuchen: Es wurden für zweiexemplarische, besonders stabilisierter und destabilisierter Konjugate Metadynamik-Simulationen durchgeführt. Die berechneten Energiedifferenzen stimmten qualitativmit den experimentellen Beobachtungen überein, für eine quantitative Aussage sinddie Profile jedoch nicht ausreichend konvergiert. Ebenso wurde beobachtet, dass sichdie PEG-Ketten im Falle von stabilisierten Varianten auf partiell entfaltete Arealelegen, wodurch der Entfaltungsprozess möglicherweise beeinflusst wird. Eine größereAnzahl an Konjugaten mit deutlich längeren Simulationszeiten müsste untersuchtwerden, um die Eignung dieses Metadynamik-Ansatzes zur Abschätzung der PEG-basierten Stabilisierung evaluieren zu können.Ein systematischer Vergleich der Dynamik verschiedener Polymervarianten (PEG,LPG, PMeOx und PEtOx) zeigte im Falle von PEG bevorzugte Konformationen inder Nähe solvensexponierter hydrophober Areale (in Übereinstimmung mit [103]), diegleichzeitig HB-Donatoren zur Etablierung von Wechselwirkungen mit den Ethersau-erstoffatomen des Polymers bieten. Starke Interaktionen mit Lysinen und Argininenwurden im Falle der strukturell deutlich komplexeren Her2-Affibody-Konjugate re-gistriert. Eine bevorzugte Wechselwirkung von PEG mit Lysinen wurde bereits inder Literatur beschrieben [107]. Ebenso konnten in diesen Simulationen Interaktio-nen des Maleimid-Linkers mit Lysinen beobachtet werden, in Übereinstimmung mitden Beschreibungen aus [108]. LPG besitzt das selbe Etherrückgrat wie PEG, aberzusätzliche Hydroxymethylseitenketten. Im Falle der WW-Konjugate wurden zu PEGsehr ähnliche Interaktionsprofile registriert. In den Affibody-Strukturen konnten je-doch deutlich mehr Wechselwirkungen mit den hier vermehrt vorkommenden Car-boxylgruppen von Aspartaten und Glutamaten beobachtet werden. Desolvatationenhydrophober Areale wurden ebenso beobachtet. Die POx-Varianten zeigten deutlichandere Wechselwirkungsprofile. Die hydrophoben Alkylseitenketten interagierten mitaromatischen Gruppen von Tyrosinen und Phenylalaninen. Zusätzlich konnten dieCarbonylgruppen der tertiären Amid-Strukturen als HB-Akzeptoren mit verschiede-nen Aminosäuren wechselwirken, bspw. Asparaginen oder Glutaminen. An einzel-nen Konjugationsstellen ließen sich deutliche Unterschiede zwischen dem Verhalten4.4. Zusammenfassung197von PMeOx und PEtOx erkennen, die mit der sterisch anspruchsvolleren Beschaffen-heit des letzteren Polymers erklärt werden können. Im Vergleich zum Her2-Affibodykonnte für die Pin 1 WW-Domäne eine größere Abhängigkeit der untersuchten Ei-genschaften von der Sekundärstruktur des Konjugationsbereichs beobachtet werden.Konjugationsstellen in den unstrukturierten Bereichen führten bei der WW-Domänezu stärker solvensexponierten Polymeren, geringeren Wechselwirkungen zum Proteinund größeren Fluktuationen des Proteinrückgrats.Insgesamt konnten in diesem Kapitel QSPR-Modelle zur PEG-basierten Stabilisie-rung generiert werden, die eine größere strukturelle Vielfalt abdeckten und im Ver-gleich zur von Price et al. publizierten kristallstrukturbasierten Richtlinie deutlichbessere Vorhersagen lieferten. Für eine weitere Optimierung stellt die vorliegende Da-tensatzgröße jedoch die entscheidende Limitierung dar. Exemplarische Metadynamik-Simulationen konnten die experimentell ermittelten energetischen Differenzen qua-litativ erfassen und zeigten eine mögliche Dynamik der PEG-Ketten beim Entfal-tungsprozess auf. Systematische Vergleiche der Dynamik verschiedener Polymerartenzeigten favorisierte Interaktionspartner auf. 199Kapitel 5Ergebnisse: Biologika5.1Gaußbeschleunigte Molekulardynamik-SimulationenDurch gaußbeschleunigte Molekulardynamik-Simulationen zehn verschiedener Start-konformationen (insgesamt 1.5 µs pro Konjugat) konnten lokale Wechselwirkungenmit allen Proteinbereichen näher studiert werden. Zur Analyse der Interaktionen wur-den Polymerdichten und entsprechende Kontaktprofile berechnet. Diese Aspekte sol-len im vorliegenden Kapitel interpretiert werden, zusätzliche Analysen zur experi-mentell bestimmten thermischen Stabilität sind in Kap. 5.3, jene zur Bioaktivität inKap. 5.4 aufgeführt.5.1.1PolymerdichtenEs wurden Polymerdichten berechnet, die die Gitterpunkte repräsentieren, für dieein Wert von mindestens 30 % der in der Simulation höchsten gefundenen Aufent-haltswahrscheinlichkeit der Polymeratome registriert wurde (Abb. 5.1). Diese Ana-lysen gewähren einen ersten Eindruck von der Verteilung der Polymeratome auf derProteinoberfläche und der bevorzugten Wechselwirkungsareale. Es zeigt sich, dass inallen Fällen die Binderegionen für die Rezeptoren IFNAR1 und IFNAR2 teilweiseoder vollständig von den Dichten bedeckt werden, was für eine generell verringerteBioaktivität im Vergleich zum Wildtyp spricht (vgl. Kap. 5.4). Die Dichten nehmenbei größeren Polymerketten (20 - 50 kDa) ein vergleichbar großes Volumen um dasProtein ein, auch bei den größten Konjugaten (30 - 50 kDa) sind jedoch unbedeck-te Proteinbereiche zu finden. Vor allem im Falle der PEtOx-Konjugate zeigen sichgrößere Areale mit freiliegender Proteinoberfläche, was darauf hindeutet, dass die In-teraktionen des Proteins mit dieser Polymerart am geringsten ausgeprägt sind. DieAufenthaltspunkte der Polymere der DBCO-Konjugate sind im Vergleich zu jenender BCN-10-kDa-Varianten leicht verändert. So zeigt sich bspw. für PEG eine grö-ßere Dichte im oberen Bereich unmittelbar um die Konjugationsstelle. PEtOx zeigtbeim DBCO-Konjugat eine geringere Tendenz der Adressierung dieses Areals. Für ei-nen detaillierteren Blick auf mögliche Protein-Polymer-Wechselwirkungen sollen dienachfolgend beschriebenen Kontaktprofile dienen.200Kapitel 5. Ergebnisse: BiologikaAbbildung 5.1: Polymerdichten für die Simulationen der Konjugate mit BCN-Linker fürPEG (10 kDa: orange, 20 kDa: rot, 40 kDa: pink), LPG (10 kDa: hellgrün, 20 kDa: grün, 50kDa: cyan) und PEtOx (10 kDa: schwarz, 20 kDa: sandfarben, 30 kDa: braun), sowie ganzrechts für jene mit DBCO-Linker (PEG: rot, LPG: grün, PEtOx: grau). Diese sind aus jeweilszwei Blickrichtungen dargestellt und repräsentieren alle Gitterpunkte mit mindestens 30 % derhöchsten ermittelten Polymerdichte in der jeweiligen Simulation. Die erste Momentaufnahmeist als Cartoon dargestellt, in welchem die Bindungsstellen von IFNAR1 (rot) und IFNAR2(grün) koloriert sind. Die Konjugationsstelle ist zusätzlich mit schwarzen Kugeln markiert(Nachdruck mit Genehmigung aus [304], Copyright 2021 American Chemical Society).5.1. Gaußbeschleunigte Molekulardynamik-Simulationen2015.1.2KontaktprofileZur Bestimmung favorisierter Wechselwirkungsbereiche auf der Proteinoberfläche wur-de zunächst der durchschnittliche Anteil jeder Aminosäure an der solvenszugänglichenOberfläche des Proteins in den zehn Simulationen des Wildtyps über den sasa-Befehlin VMD [326] mit einer Solvenssphäre von 1.4 Å bestimmt (Abb. 5.2). Anschließendwurde jeweils der prozentuale Anteil jeder einzelnen Aminosäure an der gesamtenBedeckung der Polymeroberfläche für alle 0.1 ns in den Simulationen bestimmt. DieProzente wurden ebenso für alle 20 verschiedenen proteinogenen Aminosäuretypenaufsummiert. Im Falle einer komplett ungerichteten Interaktion mit dem Protein sollteder ermittelte Anteil mit dem prozentualen Anteil der Aminosäure an der solvenszu-gänglichen Proteinoberfläche übereinstimmen, da sich das Polymer gleichmäßig überdas Protein verteilen würde und lang genug ist, um jeden Bereich des Proteins errei-chen zu können. Liegt der Anteil an der Bedeckung des Polymers aber sehr viel höher,so spricht dies für eine favorisierte Wechselwirkung. Bei dieser Analyse wurden dieWerte für die Konjugationsstelle K31 auf Null gesetzt bzw. herausgefiltert.Abbildung 5.2: Illustration der IFN-α2a-Struktur, gefärbt nach den per VMD ermitteltenAbsolutwerten der solvenszugänglichen Oberflächen (die Konjugationsstelle ist zusätzlich alsKugeln dargestellt). Für die Kontaktprofile wurden die prozentualen Anteile genutzt, wobeider Wert für K31 auf Null gesetzt wurde.Es ist zu erwarten, dass bei einer solchen Analyse hydrophobe Interaktionen, bei wel-chen die Polymeratome große Flächen der Aminosäuren (bspw. F, Y, L) adressieren,zu größeren Anteilen an bedeckter Polymeroberfläche führen. Interagiert das Polymerhingegen über eine HB spezifisch mit lediglich einem Bereich der Aminosäureseiten-kette (z. B. einem Amid-Stickstoffatom von N/Q), so wird die restliche Oberflächedieser Aminosäure nicht zwingend adressiert. Weiterhin fließen in die Ermittlung dermittleren Anteile an der bedeckten Polymeroberfläche auch wiederkehrende transi-ente Annäherungen des Polymers an das Protein ein, die über die Zeit hinweg nichtunbedingt lange bestehen bleiben.202Kapitel 5. Ergebnisse: BiologikaAls eine ergänzende Analyse der stabilsten, spezifischen Wechselwirkungen wurdendaher ebenfalls die längsten konsekutiven Aufenthaltsdauern des Polymers um dieAminosäuren herum ermittelt (sozusagen die maximalen Überlebensdauern der Kon-takte). Dies entspricht der Vorgehensweise aus [107]. Hierfür wurde ein Kontakt ineiner Momentaufnahme registriert, sobald sich ein Polymeratom (Wasserstoffatomeeingeschlossen) innerhalb von 3 Å eines spezifischen Seitenkettenschweratoms einerAminosäure befand (im Falle von Glycin das Cα-Atom). Es wurde dann für jedeAminosäure die größte Anzahl an konsekutiven Momentaufnahmen ermittelt, in de-nen der Kontakt bestand (max. 150 ns, was der Länge eines Replikons entspricht).Auch hier wurden die Kontakte mit der Konjugationsstelle K31, die zu jeder Zeitbestehen, herausgefiltert.5.1.2.1IFN-α2a-K31-BCN-10-kDa5.1.2.1.1BCN-PEG-10-kDaPEG adressiert v. a. Aminosäuren im Bereich zwischen R120 und Y135 (Abb. 5.3-A).Dieses Areal der vierten Helix besteht aus einer Vielzahl positiv geladener Seiten-ketten (R120, K121, R125, K131, K134). Beim Vergleich des Anteils der Bedeckungder Polymeroberfläche mit der solvenszugänglichen Oberfläche zeigen die hydropho-ben Aminosäuren L, Y, F, A und die positiv geladenen Aminosäuren R und K diegrößte Präferenz (Abb. 5.3-B und 5.3-C). Die teils recht großen Standardfehler sinddarauf zurückzuführen, dass unterschiedliche Areale des Proteins in den verschiede-nen Replika adressiert werden. Eine favorisierte Interaktion von PEG mit Lysinenund Argininen zeigt sich auch im Hinblick auf die maximalen Kontaktzeiten (Abb.5.3-D). Bei visueller Betrachtung der Simulationen lässt sich eine gehäufte Windungder Polymerketten um die positiv geladenen Seitenketten von R und K beobachten(Abb. 5.4-A und 5.4-B). Die drei Lysine mit der größten maximalen Kontaktzeit (K70,K131, K134) liegen dicht beieinander in unmittelbarer Nähe der Konjugationsstelle(Abb. 5.4-C).Eine bevorzugte Windung von PEG-Ketten um positiv geladene Lysine herum wurdein einer neueren Publikation, in der Simulationen von PEG-Biokonjugaten durchge-führt wurden, bereits beschrieben [107]. In Übereinstimmung mit dieser Publikationdeutet die bevorzugte Bedeckung der Aminosäuren L, Y, F und A darauf hin, dassbei der Interaktion mit Lysinen und Argininen gleichzeitig nahegelegene hydrophobeAreale adressiert werden (z. B. L117 und L128 in Helix 4). Die Tendenz zur Interak-tion von PEG über Ethersauerstoffatome mit Kationen wurde bzgl. positiv geladenerIonen bereits früh in der Literatur beschrieben [327, 328] und ebenso im Falle vonPEGylierten Lipidmembranen beobachtet, wobei ein Kation in der Regel durch fünfEthersauerstoffatome koordiniert wurde [23]. Der häufige Einsatz von PEG bei Prote-inkristallisationen resultiert weiterhin in einer Vielzahl an PDB-Strukturen, in denenPEG-Oligomere in der Nähe von Lysinen zu finden sind [107].5.1. Gaußbeschleunigte Molekulardynamik-Simulationen203Abbildung 5.3: (A) Anteil jeder Aminosäure am bedeckten Polymerbereich des 10-kDa-BCN-PEG-Konjugats in Relation zum Anteil an der solvenszugänglichen Oberfläche des Wild-typs (Berechnung der Flächen für alle 0.1 ns). Dazu ist die IFN-α2a-Struktur aus zwei Blick-richtungen gezeigt, mit schwarzen Kugeln an der Konjugationsstelle K31. Die Färbung zeigtdie Differenz der Prozentanteile an (Polymeroberfläche - solvenszugängliche Oberfläche, vonblau (+2) über weiß bis rot (-2)). (B) Top 20 Aminosäuren mit der positivsten Oberflächen-differenz aus (A). (C) Für die 20 proteinogenen Aminosäuren aufsummierte Werte aus (A).(D) Top 20 der längsten konsekutiven Kontaktzeiten mit dem Polymer. In (A), (B) und (C)sind die Standardfehler des Mittelwerts (Mittelwert + Fehler) gekennzeichnet.5.1.2.1.2BCN-LPG-10-kDaIm Falle von LPG, welches sich aus dem selben Rückgrat mit Ethersauerstoffatomenzusammensetzt, werden ebenso einige Arginine adressiert (R22, R125, R149), es lässtsich jedoch eine geringere Tendenz zur Adressierung der vierten Helix beobachten(Abb. 5.5-A). Insgesamt zeigt sich eine Verteilung der Anteile an der bedeckten Poly-meroberfläche, die eher jener der solvenszugänglichen Oberflächen entspricht. Wech-selwirkungen mit Aminosäuren im Bereich um die erste Helix (M16, R22) sind häufi-ger zu beobachten. Die aufsummierten Werte zeigen wiederum Wechselwirkungen mithydrophoben Aminosäuren wie Y, F, L, A, aber in diesem Falle auch Methioninenund Serinen (Abb. 5.5-C). Arginine zeigen eine stark positive Differenz, dagegen ste-hen Lysine in der Rangfolge im Vergleich zu den Resultaten der PEG-Simulationendeutlich weiter hinten. Erhöhte Bedeckungen des unteren unstrukturierten Bereichszwischen Helix 1 und Helix 2 resultieren weiterhin in stärkeren Interaktionen mitGlutaminen (v. a. Q40).204Kapitel 5. Ergebnisse: BiologikaAbbildung 5.4: Darstellungen exemplarischer Interaktionen mit (A) Lysinen und (B) Ar-gininen in der Simulation der 10-kDa-PEG-Variante. Eine wie im Text beschriebene Koordi-nation über fünf Ethersauerstoffatome ist hier gegeben. Wasserstoffatome sind ausgeblendet.In (C) ist die Verteilung der am stärksten adressierten Lysine illustriert. Dabei ist die Kon-jugationsstelle mit schwarzen Kugeln gekennzeichnet.Die maximalen Kontaktzeiten zeichnen ein anderes Bild. Neben Argininen und Lysi-nen sind hier einige Aspartate und Glutamate vorzufinden (E51, D71, D82, E87), dieüber ihre Carboxylgruppen die HB-Donatoren der LPG-Hydroxymethylseitenkettenadressieren können (Abb. 5.6). Aspartat steht in der Rangfolge in Abb. 5.5-C imVergleich zu den PEG-Simulationen entsprechend weiter vorne, Glutamat dagegenverbleibt hier an letzter Stelle (die Differenz ist jedoch nicht so negativ wie in Abb.5.3-C). Dies ist v. a. darauf zurückzuführen, dass E165 den C-Terminus und damiteinen sehr großen Anteil der solvenszugänglichen Oberfläche darstellt. Der Anteil anbedeckter Polymeroberfläche ist für diese Aminosäure im Vergleich zu Abb. 5.3-Csehr viel höher, liegt aber dennoch unterhalb der entsprechenden SASA. Generellsind die maximalen Kontaktzeiten in Abb. 5.5-D im Vergleich zu Abb. 5.3-D deutlichlänger, sodass davon ausgegangen werden kann, dass LPG in der Lage ist, stabilereWechselwirkungen mit dem Protein zu etablieren.5.1. Gaußbeschleunigte Molekulardynamik-Simulationen205Es gibt aufgrund des gemeinsamen Polymerrückgrats also insgesamt große Über-schneidungen mit dem Verhalten von PEG (Wechselwirkungen mit hydrophoben Ami-nosäuren und Argininen), allerdings führen die Seitenketten von LPG zu einem erwei-terten Interaktionsprofil, indem v. a. zusätzliche HB-Akzeptoren (Carboxylgruppenvon D/E) adressiert werden können. Weiterhin scheint die Tendenz der Bedeckungvon Serinen (z. B. S163) und Methionen (M16, M148) etwas erhöht zu sein.Abbildung 5.5: (A) Anteil jeder Aminosäure am bedeckten Polymerbereich des 10-kDa-BCN-LPG-Konjugats in Relation zum Anteil an der solvenszugänglichen Oberfläche des Wild-typs (Berechnung der Flächen für alle 0.1 ns). Dazu ist die IFN-α2a-Struktur aus zwei Blick-richtungen gezeigt, mit schwarzen Kugeln an der Konjugationsstelle K31. Die Färbung zeigtdie Differenz der Prozentanteile an (Polymeroberfläche - solvenszugängliche Oberfläche, vonblau (+2) über weiß bis rot (-2)). (B) Top 20 Aminosäuren mit der positivsten Oberflächen-differenz aus (A). (C) Für die 20 proteinogenen Aminosäuren aufsummierte Werte aus (A).(D) Top 20 der längsten konsekutiven Kontaktzeiten mit dem Polymer. In (A), (B) und (C)sind die Standardfehler des Mittelwerts (Mittelwert + Fehler) gekennzeichnet.Abbildung 5.6: Darstellungen exemplarischer Interaktionen mit (A) Aspartaten und (B)Glutamaten in der Simulation der 10-kDa-LPG-Variante. Wasserstoffatome sind ausgeblen-det.206Kapitel 5. Ergebnisse: Biologika5.1.2.1.3BCN-PEtOx-10-kDaPEtOx mit seinen größeren, teils hydrophoben Seitenketten zeigt wiederum ein deut-lich anderes Kontaktprofil. Es werden v. a. die Aminosäuren des flexiblen Bereichszwischen der ersten und der zweiten Helix adressiert, in welchem sich auch die Kon-jugationsstelle befindet. Die Carbonylgruppe der Seitenkette kann als potenziellerHB-Akzeptor fungieren. Dies ist aus Interaktionen mit Hydroxylgruppen von Serinen(z. B. S72, S152), sowie Amiden von Glutaminen (z. B. Q40, Q46, Q61) und Aspa-raginen (z. B. N45, N65, N156) ersichtlich (Abb. 5.7 und 5.8-A). Weiterhin scheinenv. a. Wechselwirkungen mit Tyrosinen (Y85, Y89, Y135) und Phenylalaninen (z. B.F27, F64) vorzukommen. Hierbei lassen sich Konformationen erkennen, bei denen diehydrophobe Seitenkette auf diesen Arealen aufliegt (Abb. 5.8-B).Diese Beobachtungen resultieren in einer im Vergleich zu den PEG/LPG-Simulationenanderen Rangfolge der Flächendifferenzen (Abb. 5.7-C). F und Y stehen in dieserweiter vorne. Auch die Differenzen der Serine und Asparagine sind sehr viel positi-ver als bei PEG und LPG. Wie auch im Falle des LPG werden ebenfalls Methionineadressiert (M16 und M148). Insgesamt ist das Wechselwirkungsprofil also von Inter-aktionen mit aromatischen Aminosäuren (Y, F) und HB-Donatoren (S, N, Q, R) fürdie Carbonylgruppe der Polymerseitenketten geprägt. Die Absolutwerte der längstenKontaktzeiten sind vergleichbar mit jenen in den PEG-Simulationen.5.1.2.2IFN-α2a-K31-DBCO-10-kDa5.1.2.2.1DBCO-PEG-10-kDaWie beeinflusst die Nutzung des DBCO-Linkers das Interaktionsprofil? Analoge Kon-taktprofile zu den BCN-Konjugaten sind in den Abb. 5.9, 5.10 und 5.11 gezeigt. ImFalle von PEG zeigen sich gehäufte Wechselwirkungen mit Aminosäuren in der Näheder Konjugationsstelle (R22, K23, R33) und dem C-terminal gelegenen R162. Inter-aktionen mit den Aminosäuren der vierten Helix (K121, R125, K134) sind weiterhinzu beobachten, jedoch weitaus seltener (Abb. 5.9-A). Wie auch im Falle der BCN-Variante zeigen sich v. a. Wechselwirkungen mit Argininen und Lysinen, sowie hydro-phoberen Aminosäuren (L, F und A). Ähnlich zu den BCN-LPG- und BCN-PEtOx-Varianten, die eine Tendenz zur Adressierung des unstrukturierten Bereichs um dieKonjugationsstelle aufweisen, sind hier Polymerbedeckungen durch M16 erkennbar.Weiterhin werden Serine deutlich öfter adressiert (S25, S136, S152). Während sichalso das bevorzugte Proteinareal in diesen Simulationen des DBCO-PEG-Konjugatsvon jenem des BCN-PEG-Konjugats unterscheidet, handelt es sich, mit Ausnahmeder Serine, zum größten Teil um die selben Aminosäuretypen.5.1. Gaußbeschleunigte Molekulardynamik-Simulationen207Abbildung 5.7: (A) Anteil jeder Aminosäure am bedeckten Polymerbereich des 10-kDa-BCN-PEtOx-Konjugats in Relation zum Anteil an der solvenszugänglichen Oberfläche desWildtyps (Berechnung der Flächen für alle 0.1 ns). Dazu ist die IFN-α2a-Struktur aus zweiBlickrichtungen gezeigt, mit schwarzen Kugeln an der Konjugationsstelle K31. Die Färbungzeigt die Differenz der Prozentanteile an (Polymeroberfläche - solvenszugängliche Oberfläche,von blau (+2) über weiß bis rot (-2)). (B) Top 20 Aminosäuren mit der positivsten Oberflä-chendifferenz aus (A). (C) Für die 20 proteinogenen Aminosäuren aufsummierte Werte aus(A). (D) Top 20 der längsten konsekutiven Kontaktzeiten mit dem Polymer. In (A), (B) und(C) sind die Standardfehler des Mittelwerts (Mittelwert + Fehler) gekennzeichnet.Abbildung 5.8: Darstellungen exemplarischer Interaktionen mit (A) Serinen und (B) Ty-rosinen in der Simulation der 10-kDa-PEtOx-Variante. Wasserstoffatome sind ausgeblendet.208Kapitel 5. Ergebnisse: BiologikaAbbildung 5.9: (A) Anteil jeder Aminosäure am bedeckten Polymerbereich des 10-kDa-DBCO-PEG-Konjugats in Relation zum Anteil an der solvenszugänglichen Oberfläche desWildtyps (Berechnung der Flächen für alle 0.1 ns). Dazu ist die IFN-α2a-Struktur aus zweiBlickrichtungen gezeigt, mit schwarzen Kugeln an der Konjugationsstelle K31. Die Färbungzeigt die Differenz der Prozentanteile an (Polymeroberfläche - solvenszugängliche Oberfläche,von blau (+2) über weiß bis rot (-2)). (B) Top 20 Aminosäuren mit der positivsten Oberflä-chendifferenz aus (A). (C) Für die 20 proteinogenen Aminosäuren aufsummierte Werte aus(A). (D) Top 20 der längsten konsekutiven Kontaktzeiten mit dem Polymer. In (A), (B) und(C) sind die Standardfehler des Mittelwerts (Mittelwert + Fehler) gekennzeichnet.5.1.2.2.2DBCO-LPG-10-kDaAuch die DBCO-LPG-Variante zeigt eine im Vergleich zur BCN-LPG-Variante ge-ringere Tendenz, die vierte Helix zu adressieren (Abb. 5.10-A). Stattdessen zeigt dasPolymer eine höhere Wahrscheinlichkeit, mit dem unteren Bereich um Q40 - Q48 zwi-schen Helix 1 und Helix 2 zu wechselwirken. Im Falle des BCN-Derivats wurden eherAminosäuren am Beginn dieses unstrukturierten Bereichs oder innerhalb der erstenHelix adressiert (L15, M16, R22). Somit wird auch bei LPG durch den Austausch desLinkers ein anderes Areal des Proteins adressiert, im Gegensatz zu BCN-LPG schei-nen hierbei Asparagine, Glutamine und Serine öfter adressiert zu werden (vgl. Abb.5.5-C und Abb. 5.10-C). Wie auch bei BCN-LPG zeigen die maximalen Kontaktzeitendagegen eine Präferenz mit Aspartaten und Glutamaten. Mit Ausnahme der stabilenWechselwirkungen der BCN-LPG-Variante mit D82 und K83 sind die Absolutwerteder Kontaktzeiten der DBCO- vergleichbar mit jener der BCN-Variante.5.1. Gaußbeschleunigte Molekulardynamik-Simulationen209Abbildung 5.10: (A) Anteil jeder Aminosäure am bedeckten Polymerbereich des 10-kDa-DBCO-LPG-Konjugats in Relation zum Anteil an der solvenszugänglichen Oberfläche desWildtyps (Berechnung der Flächen für alle 0.1 ns). Dazu ist die IFN-α2a-Struktur aus zweiBlickrichtungen gezeigt, mit schwarzen Kugeln an der Konjugationsstelle K31. Die Färbungzeigt die Differenz der Prozentanteile an (Polymeroberfläche - solvenszugängliche Oberfläche,von blau (+2) über weiß bis rot (-2)). (B) Top 20 Aminosäuren mit der positivsten Oberflä-chendifferenz aus (A). (C) Für die 20 proteinogenen Aminosäuren aufsummierte Werte aus(A). (D) Top 20 der längsten konsekutiven Kontaktzeiten mit dem Polymer. In (A), (B) und(C) sind die Standardfehler des Mittelwerts (Mittelwert + Fehler) gekennzeichnet.5.1.2.2.3DBCO-PEtOx-10-kDaSchließlich kann auch im Falle des DBCO-PEtOx-Derivats eine präferierte Interak-tion des Polymers mit Aminosäuren der Helix 1 und des unstrukturierten Bereichszwischen Helix 1 und Helix 2 nahe der Konjugationsstelle erkannt werden (Abb. 5.11-A). Im Vergleich zur BCN-Variante werden die unteren Bereiche vor Helix 2 jedochdeutlich seltener adressiert. Asparagine und Serine fallen in der Rangfolge der bevor-zugten Aminosäuretypen zurück, das Polymer bedeckt bevorzugt F27 und umliegendehydrophobe Aminosäuren (L30) und Arginine (R22, R33). Leucine werden insgesamtstärker bedeckt. Das Profil der maximalen Kontaktzeiten zeigt im Vergleich zumBCN-Derivat insgesamt geringere Absolutwerte und eine größere Anzahl an Glutami-nen in den Top 20. In beiden Varianten bildet Y89 die stabilsten Kontakte mit demPolymer aus.210Kapitel 5. Ergebnisse: BiologikaAbbildung 5.11: (A) Anteil jeder Aminosäure am bedeckten Polymerbereich des 10-kDa-DBCO-PEtOx-Konjugats in Relation zum Anteil an der solvenszugänglichen Oberfläche desWildtyps (Berechnung der Flächen für alle 0.1 ns). Dazu ist die IFN-α2a-Struktur aus zweiBlickrichtungen gezeigt, mit schwarzen Kugeln an der Konjugationsstelle K31. Die Färbungzeigt die Differenz der Prozentanteile an (Polymeroberfläche - solvenszugängliche Oberfläche,von blau (+2) über weiß bis rot (-2)). (B) Top 20 Aminosäuren mit der positivsten Oberflä-chendifferenz aus (A). (C) Für die 20 proteinogenen Aminosäuren aufsummierte Werte aus(A). (D) Top 20 der längsten konsekutiven Kontaktzeiten mit dem Polymer. In (A), (B) und(C) sind die Standardfehler des Mittelwerts (Mittelwert + Fehler) gekennzeichnet.5.1.2.3Acyl-PEG-10-kDaZusätzlich zum Vergleich zwischen BCN- und DBCO-Linker wurde ein Replikon derBCN-PEG-10-kDa-Variante für 150 ns simuliert, in der der Linker in der Startkonfor-mation durch eine einfache Peptidbindung bzw. Acylgruppe ausgetauscht wurde. Ei-ne solche wird oft bei klassischen Kupplungsreaktionen ohne Click-Chemie-Verfahrenunter Nutzung eines Succinimidylesters erhalten (siehe erstes Beispiel in Tab. 1.3).Sehr ähnliche Proteinoberflächen werden durch PEG mit dieser Linkerart, verglichenmit BCN, adressiert (Abb. 5.12). Während das acylierte Lysin teilweise ins Solvensgestreckt ist (im Gegensatz zum BCN-Linker, welcher zumeist direkt auf der Protein-oberfläche aufliegt), wechselwirken die ersten PEG-Monomere mit den selben Amino-säuren (K121 und R149). Die von PEG bedeckte Proteinoberfläche beträgt im Falleder K31-BCN-Variante 21.98 ± 0.05 % und beim acylierten Derivat 20.84 ± 0.10 %.Dieser exemplarische Vergleich zeigt damit auf, dass der in diesen Studien genutz-te BCN-Linker trotz seiner Größe nicht zu einer drastischen Verringerung der vomPolymer adressierten Oberfläche führt [304].5.1. Gaußbeschleunigte Molekulardynamik-Simulationen211Abbildung 5.12: Polymerdichten (Berechnung analog zu Abb. 5.1) für die Simulationen derK31-BCN-PEG- (blaues Gitter) und K31-Acyl-PEG-Varianten (rotes Gitter). Rückgratatomeder Konjugationsstelle sind als schwarze Kugeln dargestellt (Nachdruck mit Genehmigung aus[304], Copyright 2021 American Chemical Society).5.1.2.4Größere BCN-Konjugate (20 - 50 kDa)Werden die Kontaktprofile durch Nutzung längerer Polymerketten beeinflusst? ImFolgenden sollen auch die Kontaktprofile der BCN-Konjugate mit längeren Polymer-ketten (20 - 50 kDa) kurz beschrieben werden. Dabei sollte angemerkt werden, dassdiese Varianten aus Zeitgründen genauso lange wie die 10-kDa-Varianten simuliertwurden, obwohl diese Systeme deutlich mehr Freiheitsgrade beinhalten.5.1.2.4.120 kDaIm Falle von BCN-PEG-20-kDa zeigt sich ein zur 10-kDa-Variante nahezu identischesProfil: Das Polymer wechselwirkt bevorzugt mit Aminosäuren der Helix 4, v. a. Ar-gininen und Lysinen (Abb. 5.13). Abermals zeigen sich positive Flächendifferenzenfür die hydrophoben Aminosäuren Tyrosin, Phenylalanin und Alanin. R22, K23 undR33 in der Nähe der Konjugationsstelle werden allerdings häufiger adressiert. Auchdie maximalen Kontaktzeiten zeigen generell etwas größere Werte.BCN-LPG-20-kDa zeigt ein Profil, welches auf eine gleichmäßigere Bedeckung derProteinoberfläche hindeutet, als es bei der 10-kDa-Variante der Fall ist (Abb. 5.14-A). So werden im Vergleich zu letzterer auch vermehrt Aminosäuren im Bereich derzweiten Helix und zwischen Helix 3 und 4 adressiert. Nur noch die Termini stel-len größere Bereiche mit negativen Flächendifferenzen dar. Viele Arginine werdenadressiert (R22, R120, R125), diese fallen in der Rangfolge dennoch zurück (Abb.5.14-C). Dagegen rückt Cystein (C29, C138) nach vorne. Weiterhin kommt es auchzur Bedeckung von I24, welches sich in unmittelbarer Nähe zu R22 und K23 befindet.Dies führt zu insgesamt sehr positiven Flächendifferenzen für Isoleucin. Serine zeigendie positivste Flächendifferenz. Weiterhin zeigen sich (noch ausgeprägter als bei der212Kapitel 5. Ergebnisse: Biologika10-kDa-Variante) stabile Wechselwirkungen mit Glutamaten und Aspartaten (z. B.D32, E58, D77, E78, E132, Abb. 5.14-D), die Absolutwerte der Kontaktzeiten sindvergleichbar mit jenen aus Abb. 5.5-D.BCN-PEtOx-20-kDa zeigt eine im Vergleich zum 10-kDa-Derivat geringere Tendenzzur Bedeckung der Aminosäuren im unteren Areal des unstrukturierten Bereichs zwi-schen Helix 1 und 2 (Q40 - A50, Abb. 5.15). Entsprechend rücken Asparagine undGlutamine, die in diesem Bereich vermehrt vorzufinden sind, in der Rangfolge bzgl.der Flächendifferenzen nach hinten. Aminosäuren um F27, die auch schon von der10-kDa-Variante adressiert wurden, zeigen hingegen einen noch größeren Anteil ander Polymerbedeckung.5.1.2.4.230 - 50 kDaAuch für das BCN-PEG-40-kDa-Konjugat zeigt sich eine Präferenz für Arginine undLysine, sowie hydrophobe Aminosäuren Leucin, Methionin, Alanin und Phenylalanin.Tyrosin rückt in der Rangfolge nach hinten. Im Gegensatz zur 10-kDa- und 20-kDa-Variante werden die Termini stärker bedeckt, v. a. das C-terminal gelegene R162 wirdöfter adressiert.Im Falle von BCN-LPG-50-kDa werden Aminosäuren der flexiblen Termini und in un-mittelbarer Nähe zur Konjugationsstelle im Vergleich zur 10-kDa-Variante vermehrtadressiert. Wie bei der LPG-20-kDa-Variante auch zeigt sich eine Bedeckung derPolymeroberfläche, die sehr der Verteilung der solvenszugänglichen Oberfläche äh-nelt. Die Absolutwerte der maximalen Kontaktzeiten sind für diese Größenordnungder Konjugate wiederum für LPG am größten. Diese signalisieren abermals stabileWechselwirkungen mit Aspartaten (D2, D71) und vielen Glutamaten (E107, E113,E146).Wie auch im Falle der 20-kDa-Variante sind Bedeckungen beim BCN-PEtOx-30-kDa-Konjugat im unteren unstruktierten Bereich zwischen Helix 1 und 2 (Q40 - A50) imGegensatz zur 10-kDa-Variante erniedrigt. Analog zu den größten PEG- und LPG-Konjugaten zeigen sich leicht erhöhte Bedeckungen im Bereich der beiden Termini.Wechselwirkungen mit Argininen rücken in der Rangfolge der Flächendifferenzen nachhinten. Im Hinblick auf die maximalen Kontaktzeiten zeigen sich ähnliche Präferen-zen wie bei der 10-kDa-Variante (Interaktionen mit Tyrosinen und Serinen). DurchInteraktionen mit Q61 rückt Glutamin an die Spitze der Rangfolge der Aminosäurenmit den positivsten Flächendifferenzen.5.1. Gaußbeschleunigte Molekulardynamik-Simulationen213Abbildung 5.13: (A) Anteil jeder Aminosäure am bedeckten Polymerbereich des 20-kDa-BCN-PEG-Konjugats in Relation zum Anteil an der solvenszugänglichen Oberfläche des Wild-typs (Berechnung der Flächen für alle 0.1 ns). Dazu ist die IFN-α2a-Struktur aus zwei Blick-richtungen gezeigt, mit schwarzen Kugeln an der Konjugationsstelle K31. Die Färbung zeigtdie Differenz der Prozentanteile an (Polymeroberfläche - solvenszugängliche Oberfläche, vonblau (+2) über weiß bis rot (-2)). (B) Top 20 Aminosäuren mit der positivsten Oberflächen-differenz aus (A). (C) Für die 20 proteinogenen Aminosäuren aufsummierte Werte aus (A).(D) Top 20 der längsten konsekutiven Kontaktzeiten mit dem Polymer. In (A), (B) und (C)sind die Standardfehler des Mittelwerts (Mittelwert + Fehler) gekennzeichnet.Abbildung 5.14: (A) Anteil jeder Aminosäure am bedeckten Polymerbereich des 20-kDa-BCN-LPG-Konjugats in Relation zum Anteil an der solvenszugänglichen Oberfläche des Wild-typs (Berechnung der Flächen für alle 0.1 ns). Dazu ist die IFN-α2a-Struktur aus zwei Blick-richtungen gezeigt, mit schwarzen Kugeln an der Konjugationsstelle K31. Die Färbung zeigtdie Differenz der Prozentanteile an (Polymeroberfläche - solvenszugängliche Oberfläche, vonblau (+2) über weiß bis rot (-2)). (B) Top 20 Aminosäuren mit der positivsten Oberflächen-differenz aus (A). (C) Für die 20 proteinogenen Aminosäuren aufsummierte Werte aus (A).(D) Top 20 der längsten konsekutiven Kontaktzeiten mit dem Polymer. In (A), (B) und (C)sind die Standardfehler des Mittelwerts (Mittelwert + Fehler) gekennzeichnet.214Kapitel 5. Ergebnisse: BiologikaAbbildung 5.15: (A) Anteil jeder Aminosäure am bedeckten Polymerbereich des 20-kDa-BCN-PEtOx-Konjugats in Relation zum Anteil an der solvenszugänglichen Oberfläche desWildtyps (Berechnung der Flächen für alle 0.1 ns). Dazu ist die IFN-α2a-Struktur aus zweiBlickrichtungen gezeigt, mit schwarzen Kugeln an der Konjugationsstelle K31. Die Färbungzeigt die Differenz der Prozentanteile an (Polymeroberfläche - solvenszugängliche Oberfläche,von blau (+2) über weiß bis rot (-2)). (B) Top 20 Aminosäuren mit der positivsten Oberflä-chendifferenz aus (A). (C) Für die 20 proteinogenen Aminosäuren aufsummierte Werte aus(A). (D) Top 20 der längsten konsekutiven Kontaktzeiten mit dem Polymer. In (A), (B) und(C) sind die Standardfehler des Mittelwerts (Mittelwert + Fehler) gekennzeichnet.Abbildung 5.16: (A) Anteil jeder Aminosäure am bedeckten Polymerbereich des 40-kDa-BCN-PEG-Konjugats in Relation zum Anteil an der solvenszugänglichen Oberfläche des Wild-typs (Berechnung der Flächen für alle 0.1 ns). Dazu ist die IFN-α2a-Struktur aus zwei Blick-richtungen gezeigt, mit schwarzen Kugeln an der Konjugationsstelle K31. Die Färbung zeigtdie Differenz der Prozentanteile an (Polymeroberfläche - solvenszugängliche Oberfläche, vonblau (+2) über weiß bis rot (-2)). (B) Top 20 Aminosäuren mit der positivsten Oberflächen-differenz aus (A). (C) Für die 20 proteinogenen Aminosäuren aufsummierte Werte aus (A).(D) Top 20 der längsten konsekutiven Kontaktzeiten mit dem Polymer. In (A), (B) und (C)sind die Standardfehler des Mittelwerts (Mittelwert + Fehler) gekennzeichnet.5.1. Gaußbeschleunigte Molekulardynamik-Simulationen215Abbildung 5.17: (A) Anteil jeder Aminosäure am bedeckten Polymerbereich des 50-kDa-BCN-LPG-Konjugats in Relation zum Anteil an der solvenszugänglichen Oberfläche des Wild-typs (Berechnung der Flächen für alle 0.1 ns). Dazu ist die IFN-α2a-Struktur aus zwei Blick-richtungen gezeigt, mit schwarzen Kugeln an der Konjugationsstelle K31. Die Färbung zeigtdie Differenz der Prozentanteile an (Polymeroberfläche - solvenszugängliche Oberfläche, vonblau (+2) über weiß bis rot (-2)). (B) Top 20 Aminosäuren mit der positivsten Oberflächen-differenz aus (A). (C) Für die 20 proteinogenen Aminosäuren aufsummierte Werte aus (A).(D) Top 20 der längsten konsekutiven Kontaktzeiten mit dem Polymer. In (A), (B) und (C)sind die Standardfehler des Mittelwerts (Mittelwert + Fehler) gekennzeichnet.Abbildung 5.18: (A) Anteil jeder Aminosäure am bedeckten Polymerbereich des 30-kDa-BCN-PEtOx-Konjugats in Relation zum Anteil an der solvenszugänglichen Oberfläche desWildtyps (Berechnung der Flächen für alle 0.1 ns). Dazu ist die IFN-α2a-Struktur aus zweiBlickrichtungen gezeigt, mit schwarzen Kugeln an der Konjugationsstelle K31. Die Färbungzeigt die Differenz der Prozentanteile an (Polymeroberfläche - solvenszugängliche Oberfläche,von blau (+2) über weiß bis rot (-2)). (B) Top 20 Aminosäuren mit der positivsten Oberflä-chendifferenz aus (A). (C) Für die 20 proteinogenen Aminosäuren aufsummierte Werte aus(A). (D) Top 20 der längsten konsekutiven Kontaktzeiten mit dem Polymer. In (A), (B) und(C) sind die Standardfehler des Mittelwerts (Mittelwert + Fehler) gekennzeichnet.216Kapitel 5. Ergebnisse: Biologika5.1.3FazitDie berechneten Kontaktprofile der 10-kDa-BCN-Varianten deuten auf eine bevorzug-te Interaktion von PEG mit positiven Argininen und Lysinen hin, um die es sich mitseinen Ethersauerstoffatomen windet. Diese Aminosäuren sind v. a. im Bereich dervierten Helix vorzufinden. Ergänzend hierzu werden hydrophobe Aminosäuren (F, Y,L, A) adressiert, was die Beobachtungen aus [107] stützt. LPG mit zusätzlichen Hy-droxymethylseitenketten zeigt ein erweitertes Interaktionsprofil mit geringerer Bede-ckung von Lysinen. Bedeckungen von Serinen, Methioninen und Glutaminen werdenvermehrt registriert. Der größte Unterschied zu PEG findet sich in den maximalenKontaktzeiten des Polymers mit dem Protein. LPG zeigt hierbei die stabilsten Inter-aktionen aller drei Polymerarten und eine Präferenz zu den Aminosäuren Aspartatund Glutamat, mit denen es HB ausbilden kann. PEtOx zeigt bevorzugte Interaktio-nen mit den aromatischen Aminosäuren Tyrosin (v. a. Y89) und Phenylalanin. Zu-sätzlich können über die Carbonylgruppe der Seitenkette verschiedene HB-Donatoren,v. a. Glutamine, Asparagine und Serine, adressiert werden.Insgesamt bewirkt der Austausch des BCN- durch den DBCO-Linker durchaus eineVeränderung des vom Polymer bedeckten Proteinareals. Interaktionen mit Aminosäu-ren der Helix 4 werden verringert, dagegen zeigen sich größere Wechselwirkungen mitAminosäuren, die sich in der Helix 1 oder im unstrukturierten Bereich zwischen derersten und zweiten Helix befinden. Die zur Interaktion bevorzugten Aminosäurety-pen ändern sich jedoch nur marginal. Im Falle des DBCO-LPG-Derivats, welches daseinzige DBCO-Konjugat darstellt, das eine im Vergleich zur BCN-Variante deutlicherniedrigte Schmelztemperatur aufweist (vgl. Kap. 5.3.1), zeigen sich u. a. geringereBedeckungen von Argininen und eine erhöhte Tendenz zu Interaktionen mit Asparagi-nen und Serinen. Weiterhin scheinen stabile Wechselwirkungen mit D82 und K83, wiesie bei der BCN-Variante gefunden wurden, beim DBCO-Derivat nicht ausgebildetzu werden.Im Hinblick auf die größeren Konjugate (20 - 50 kDa) zeigen die Polymere generelleine höhere Tendenz, Aminosäuren in der Nähe der Konjugationsstelle zu bedecken.Die größten Derivate (30 - 50 kDa) weißen zudem deutlich mehr Wechselwirkungenmit den beiden flexiblen Termini des Proteins auf.5.2. Coarse-grained-Simulation2175.2Coarse-grained-Simulation5.2.1Vergleich mit GaMD-SimulationenDie CG-Simulation einer IFN-α2a-K31-PEG-10-kDa-Variante ohne Linker zeigt ei-ne im Vergleich zu den GaMD-Ansätzen deutlich andere Polymerdynamik. Durchdie längere Simulationszeit von 10 µs mit einer sehr viel größeren Solvensbox undeiner starken Reduktion an Freiheitsgraden kommt es zum Sampling vieler Konfor-mationen mit solvensexponierten Polymerketten. Diese Verteilung ist aus den Punkt-wolken in Abb. 5.19 ersichtlich. Eine Polymerdichte in Analogie zu jenen aus denGaMD-Ansätzen zeigt jedoch lediglich im Bereich um R120 - R125 erhöhte Aufent-haltswahrscheinlichkeiten (Abb. 5.20). Dies entspricht dem Areal der Helix 4, welchesauch in den GaMDs durch PEG am häufigsten adressiert wird.Abbildung 5.19: Darstellung von Momentaufnahmen der CG-Simulation aus zwei Blickrich-tungen (Rückgratkugeln lachsfarben, Seitenkettenkugeln gelb, Konjugationsstelle schwarz).Der Polymer-Anteil des Konjugats ist für alle 100 ns als kleinere, transparente, sowie für jedeMikrosekunde zusätzlich als opake Kugeln gezeigt. Diese Kugeln sind von rot über weiß hinzu blau, basierend auf dem Zeitpunkt der Momentaufnahme, koloriert.Abbildung 5.20: (A) Polymerdichte analog zu Abb. 5.1 aus der CG-MD. (B) Polymerdichteaus Abb. 5.1 zur GaMD des BCN-PEG-10-kDa-Konjugats (Nachdruck mit Genehmigung aus[304], Copyright 2021 American Chemical Society).218Kapitel 5. Ergebnisse: BiologikaEin entsprechendes Kontaktprofil wurde ebenfalls erstellt und ist in Abb. 5.21 einseh-bar. Auch wenn hier wie bei der GaMD der BCN-PEG-10-kDa-Variante viele Arginineund Lysine bedeckt werden (R120, K121, R125, K133), so scheinen Bedeckungen derhydrophoben Aminosäuren Leucin und v. a. Phenylalanin deutlich ausgeprägter zusein. Im Gegensatz zur GaMD zeigt sich eine geringe Adressierung von R33, K70 undR149, dagegen interagiert das Polymer mit F27 und F36 zwischen Helix 1 und 2. Einedeutliche Präferenz für Phenylalanine zeigt sich auch beim Blick auf die maximalenKontaktzeiten. Die Absolutwerte dieser Zeiten sind im Vergleich zur GaMD deutlichkleiner. Hierbei wurde ein Kontakt registriert, sofern eine Polymerkugel innerhalb von8 Å der äußersten Aminosäureseitenkettenkugel zu finden war. Ein größerer Grenz-wert im Vergleich zu den GaMDs (3 Å) wurde gewählt, da die CG-Kugeln größereRadii als die Atome in den GaMDs aufweisen. Geringere Grenzwerte als 8 Å wurdengetestet, führten jedoch zu kaum registrierbaren Kontakten.Insgesamt zeigt sich, dass PEG analog zur GaMD der BCN-PEG-10-kDa-VarianteBereiche der vierten Helix adressiert. Im Unterschied zur GaMD zeigen sich jedochschwächere Wechselwirkungen mit Argininen und Lysinen und stärkere mit Leucinenund Phenylalaninen. Diese Unterschiede resultieren vermutlich aus dem MARTINI-2-Kraftfeld, möglicherweise werden PEG-Wechselwirkungen mit polaren Proteinbe-reichen in diesem unterschätzt (vide infra). Andererseits kommt es zum Samplingproteinferner Polymerkonformationen, was bei den GaMDs nicht beobachtet wurde.Abbildung 5.21: (A) Anteil jeder Aminosäure am bedeckten Polymerbereich des CG-10-kDa-BCN-PEG-Konjugats in Relation zum Anteil an der solvenszugänglichen Oberfläche desWildtyps (Berechnung der Flächen für alle 0.1 ns). Dazu ist die IFN-α2a-Struktur aus zweiBlickrichtungen gezeigt, mit schwarzen Kugeln an der Konjugationsstelle K31. Die Färbungzeigt die Differenz der Prozentanteile an (Polymeroberfläche - solvenszugängliche Oberfläche,von blau (+2) über weiß bis rot (-2)). (B) Top 20 Aminosäuren mit der positivsten Oberflä-chendifferenz aus (A). (C) Für die 20 proteinogenen Aminosäuren aufsummierte Werte aus(A). (D) Top 20 der längsten konsekutiven Kontaktzeiten mit dem Polymer. In (A), (B) und(C) sind die Standardfehler des Mittelwerts (Mittelwert + Fehler) gekennzeichnet.5.2. Coarse-grained-Simulation2195.2.2Verwendetes KraftfeldEs wurden die PEG-Parameter des MARTINI-2-Kraftfelds verwendet, welche auchim PolyPly-Programm implementiert sind [313]. Diese wurden mit Berücksichtigungder Dynamik reiner PEG-Ketten, PEGylierter Lipid-Bilayer, PEG-basierter nichtio-nischer Tenside und Polystyren-PEG-Blockcopolymere entwickelt. Die Fähigkeit äl-terer MARTINI-Parameter für PEG, Protein-Interaktionen richtig wiederzugeben,wurde in einer Publikation kürzlich infrage gestellt und in diesem Zuge ein abge-wandeltes Parameterset entwickelt [329]. Unter diesen Gesichtspunkten könnte eineImplementierung dieser modifizierten MARTINI-2-Parameter oder eine Nutzung desMARTINI-3-Kraftfelds entlang eines kürzlich publizierten Protokolls [163] zu einergrößeren Übereinstimmung mit den Ergebnissen aus den atombasierten Simulationenführen.Dennoch zeigen die Ergebnisse dieser Simulation exemplarisch, dass die GaMDs ledig-lich proteinnahe Konformationen der Polymere abbilden, die aber unter Umständennur einen sehr geringen Anteil des in der Wirklichkeit besuchten Phasenraums dar-stellen. Im Hinblick auf PEG wird hierbei zwischen einem Shroud (Hülle) und einemDumbbell (Hantel)-Modell unterschieden (Abb. 5.22). Im Dumbbell-Modell bildetdie PEG-Kette eine eigenständige Entität (random coil) fernab der Proteinoberflä-che, während das Shroud-Modell ein Umwinden des Polymers um diese herum be-schreibt. Je nach System und Polymerlänge werden beide Vorstellungen als möglichePolymeranordnungen in Betracht gezogen [107, 330]. Entsprechende weiterführendeCG-Simulationen könnten daher dabei helfen, den Anteil an Konformationen, in de-nen sich das Polymer vom Protein abgewandt befindet, richtig wiederzugeben unddamit die Relevanz der in den GaMD-Simulationen beobachteten Interaktionen bes-ser einstufen zu können. Eine solche Gewichtung wäre auch im Hinblick auf die inden Kap. 5.3 und 5.4 durchgeführten Berechnungen zur Bestimmung der thermischenStabilität und Bioaktivität von Bedeutung.Abbildung 5.22: Schematische Darstellungen des Shroud (links) und des Dumbbell-Modells(rechts). Das Polymer kann das Protein, je nach Größe, umhüllen und mit diesem wechsel-wirken oder als eher unabhängige Entität im Solvens vorliegen. Die durchgeführten GaMD-Simulationen können lediglich proteinnahe Konformationen wiedergeben.220Kapitel 5. Ergebnisse: Biologika5.3Untersuchungen zur thermischen StabilitätIn den letzten Kapiteln wurden die bevorzugten Wechselwirkungspartner der ver-schiedenen Polymerarten näher analysiert. Hier soll nun untersucht werden, ob dieexperimentell mittels DSF-Messungen (siehe Kap. 3.2.7) ermittelten Schmelztempe-raturen verschiedener Konjugate mit aus den Simulationen ableitbaren Parameternin Verbindung gebracht werden können, um somit den Effekt der Polymere auf diethermische Stabilität interpretieren zu können.5.3.110-kDa-VariantenFolgende Schmelzpunkte wurden experimentell zunächst ermittelt:Abbildung 5.23: Per DSF ermittelte Schmelztemperaturen (95%-Konfidenzintervalle; n = 5;signifikante Unterschiede mit p-Werten von <0.05 durch Sterne angedeutet). Gelistete Massenbasieren auf MALDI-TOF (Matrix-assistierte Laser-Desorption-Ionisierung-Time-of-flight)-Messungen. Die Polydispersität (PDI) wurde über eine Größenausschlusschromatographieermittelt (Reproduzierung mit Genehmigung aus [304], Copyright 2021 American ChemicalSociety).Es handelt sich um einen äußerst engen Temperaturbereich. Der Wildtyp weist einegegenüber allen anderen aufgelisteten Strukturen signifikant höhere thermische Sta-bilität auf. Alle Konjugate weisen dementsprechend eine Destabilisierung auf, wobeidiese im Falle des BCN-LPG-Derivats am geringsten ausfällt. Dieses Konjugat ist daseinzige, welches sich signifikant von allen anderen unterscheidet und eine vergleichbareSchmelztemperatur zur unkonjugierten Struktur mit Azidgruppe aufweist (hier mitN3 gekennzeichnet). Es handelt sich auch um den einzigen Fall, in dem der Austauschdes BCN- mit einem DBCO-Linker einen deutlichen Effekt auf die Schmelztemperaturhat. Die PEG-Derivate zeigen im Vergleich zu den LPG-Konjugaten etwas niedrigereSchmelztemperaturen, die PEtOx-Konjugate sind die instabilsten PBK.Zunächst wurden aus den entsprechenden GaMD-Simulationen der BCN-10-kDa-Varianten Parameter ermittelt, die Protein-Polymer-Wechselwirkungen beschreiben,z. B. die Anzahl an Protein-Polymer-HB, die durch das jeweilige Polymer bedeckte5.3. Untersuchungen zur thermischen Stabilität221Proteinoberfläche oder die Interaktionsenergie mit allen Aminosäuren des Proteins.Diese Untersuchungen zeigten jedoch zunächst keine Korrelation mit den experimen-tellen Befunden. Es wurden schließlich drei Parameter gefunden, die alle in gewis-ser Weise die Proteinflexibilität beschreiben: (1) die durch eine CNA vorhergesagteSchmelztemperatur, (2) die Anzahl an nativen Kontakten, (3) sowie das Verhältnisder RMSF-Werte zum Wildtyp. Die nachträglich simulierten DBCO-Varianten wur-den mithilfe dieser Parameter korrekt bzgl. ihrer thermischen Stabilität eingestuft.Entsprechende Korrelationen sind in den Abb. 5.24 gezeigt und sollen im Folgendennäher beschrieben werden.Abbildung 5.24: (A) Korrelation zwischen den CNA-basierten und den experimentell be-stimmten Schmelztemperaturen. Farben entsprechen den Dichtedarstellungen der Abb. 5.1:WT = blau, BCN-LPG = grün, DBCO-LPG = dunkelgrün, BCN-PEG = orange, DBCO-PEG = rot, BCN-PEtOx = dunkelgrau, DBCO-PEtOx = grau. (B) Korrelation zwischenden experimentell bestimmten Schmelztemperaturen und der Anzahl nativer Kontakte. (C)Korrelation zwischen den experimentell bestimmten Schmelztemperaturen und den RMSF-Verhältnissen, die in (D) oben rechts aufgelistet sind. (D) Aminosäurespezifische RMSF-Verhältnisse zwischen den Biokonjugaten und dem Wildtyp (Verhältnis der Durchschnittswer-te aller Replika). Dargestellt sind die Aminosäuren G10 bis T155 ohne die flexiblen Termini.Die größten Unterschiede sind für die Aminosäuren der Schleifen und einige Bereiche der He-lices 2 und 3 zu erkennen (L30 - T52, Q62 - T79 und L95 - E113). Diese sind als rote Kugelnin der dargestellten Struktur markiert. Letztere ist gelb in Arealen mit großen Unterschiedenbzgl. der in Abb. 5.27 weiter unten dargestellten Perkolationsindizes gefärbt (Reproduzierungbzw. Nachdruck mit Genehmigung aus [304], Copyright 2021 American Chemical Society).5.3.1.1CN-AnalysenZur Durchführung der CN-Analysen der in Kap. 5 behandelten IFN-α2a-Konjugatemit dem Programm CNA 2.0 wurden für alle 0.1 ns PDB-Strukturen der Simulationengespeichert. Dieses Analysenprogramm wurde bisher nur für Proteine oder Systeme222Kapitel 5. Ergebnisse: Biologikamit kleineren, nichtkovalenten Liganden verwendet. Die Monomere wurden entspre-chend als Ligand-Heteroatome definiert, während der BCN- bzw. DBCO-Linker alsTeil der Aminosäure K31 beschrieben wurde. Somit konnte der Einfluss nichtkovalen-ter Wechselwirkungen des Polymers mit dem Protein auf die Rigidität des Proteinsuntersucht werden.Ein Energiebereich zwischen -10 und 0 kcal/mol mit Schritten von 0.05 kcal/molwurde analysiert (E = -10 0 0.05). Zur Beurteilung hydrophober Effekte wurde derStandardgrenzwert von 0.25 Å (c = 0.25) verwendet. Befanden sich zwei Kohlenstoff-,ein Kohlenstoff- und ein Schwefel-, oder zwei Schwefelatome innerhalb dieses Grenz-werts (inklusive der jeweiligen VdW-Radien) wurde dies als hydrophober Kontakteingestuft (H = 1). Deren Stärke wurde mit zunehmenden Temperaturen bei derEntfaltungssimulation des Programms nicht moduliert (TUS = 1). Diese Einstellun-gen in Kombination mit einem auf der Konfigurationsentropie vom Typ 2 basierendenSpline-Fit ergab die beste Korrelation (Abb. 5.24-A).Der Wildtyp wurde korrekt als stabilste Variante eingestuft, gefolgt vom BCN-LPG-Konjugat. Auch die übrige Rangfolge stimmt mit den experimentellen Befunden über-ein, wobei beide PEtOx-Varianten als deutlich destabilisiert beurteilt wurden. Dievom Programm ausgegebene Schmelztemperatur basiert auf einer Umrechnung derjeweiligen Energie-Grenzwerte, bei denen eine Entfaltung anzunehmen ist. Im Falleder Konfigurationsentropie vom Typ 2 ist dies an jenem Grenzwert gegeben, an demsich der größte Rigiditätscluster des Netzwerks anfängt aufzulösen und das Systemnicht mehr dominiert. Die Konfigurationsentropie H vom Typ 2 basiert auf der In-formationstheorie nach Shannon und beschreibt den Grad der Unordnung in Bezugzur Realisation eines bestimmten Zustands:1H = −�sws ln ws(5.1)ws =s2 ns�s s2 ns(5.2)ns =Anzahl Cluster der Größe sAnzahl Atome N(5.3)Die finale Entropie wird hierbei aus der Summe der Werte für alle im System ge-fundenen Clustergrößen s gebildet. Dabei beschreibt ws die Wahrscheinlichkeit einesAtoms, sich in einem Cluster der Größe s zu befinden. Diese Wahrscheinlichkeit re-sultiert wiederum aus einer Verrechnung mit der normierten Clustergröße ns, bei derdie Anzahl der Cluster mit der Größe s in Relation zur Anzahl an Atomen N imgesamten Netzwerk gesetzt wird. Zu Beginn besteht lediglich ein rigider Cluster imSystem, die Entropie besitzt daher den Wert Null. Fällt der größte Cluster in sichzusammen, so kommt es zum Anstieg möglicher Systemzustände (verschiedene Clus-tergrößen s) und damit der Entropie H. Verändert sich die Entropie also an einem1Die eher unübliche Nutzung des Kennzeichens H für eine Entropie richtet sich hierbei nach denpublizierten Arbeiten zu CNA-basierten Deskriptoren [234, 235].5.3. Untersuchungen zur thermischen Stabilität223gegebenen Energie-Grenzwert sehr stark, so wird ein Phasenübergang vermutet. Dieausgegebenen absoluten Temperaturwerte sollten allerdings lediglich relativ zwischenvergleichbaren Systemen beurteilt werden [231, 234, 235, 331].5.3.1.1.1Generierte NetzwerkeDie Stabilität der PEG- und LPG-Varianten scheint bei der thermalen Entfaltungssi-mulation durch das CNA-Programm überbewertet zu werden. Dies könnte daraus re-sultieren, dass diese Strukturelemente Ethersauerstoffatome besitzen, deren Fähigkeitzur Ausbildung von HB vom Programm überschätzt wird. Eine Visualisierung exem-plarischer Ensemble-Netzwerke per visualCNA zeigt, dass LPG besonders viele HBausbildet, PEtOx das Protein dagegen hauptsächlich über hydrophobe Constraintsbeeinflusst (Abb. 5.25). PEG zeigt deutlich weniger HB als LPG. Die Ethersauer-stoffatome scheinen bei beiden Varianten jedoch maßgeblich zur Rigidität beizutra-gen und viele HB-Constraints, v. a. über Lysine und Arginine, ausbilden zu können(Abb. 5.26).5.3.1.1.2PerkolationsindizesNeben globalen Rigiditätsindizes können mithilfe des CNA-Programms auch loka-le Effekte über verschiedene Parameter analysiert werden (siehe Kap. 1.4). Einenähere Inspektion der lokalen Indizes zeigt auf, dass die destabilisierten PEtOx-Konjugate besonders im Falle der Perkolationsindizes vom Typ 2 deutlich positi-vere Energiegrenzwerte in bestimmten Bereichen aufweisen (um V55 - N65 und R125- T155, entspricht Bereichen der Helices 2, 4 und 5), verglichen mit allen übrigenStrukturen (Abb. 5.27-A). Diese Indizes geben die Grenzwerte an, bei denen sichdie Aminosäuren vom aktuell größten rigiden Cluster loslösen. Geringe Absolutwertedeuten auf eine während der Entfaltung frühere Flexibilisierung der entsprechendenBereiche hin. Beim Blick auf die instabilen Areale in Kombination mit den gefundenenPolymerdichten (Abb. 5.27-B und 5.27-C) zeigt sich, dass diese Aminosäuren durchPEG und LPG bedeckt sind und damit rigidisierende Wechselwirkungen ermöglichtwerden, während sie im Falle der PEtOx-Konjugate zu einem stärkeren Grad solvens-exponiert bleiben.5.3.1.2Native KontakteDie Anzahl nativer Kontakte wird oft zur Untersuchung von Entfaltungssimulationengenutzt [332]. In diesem Falle wurden alle Aminosäurekontakte (für die individuellenAtome berechnet), basierend auf der jeweiligen NMR-Struktur mit dem standard-mäßigen Grenzwert von 4.5 Å, berücksichtigt. Ausgenommen wurden die Kontaktevon K31 und der sehr flexiblen Termini C1 - L9 und N156 - E165. Daraus wurde derdurchschnittliche Anteil an Momentaufnahmen berechnet, in denen diese Kontakte in224Kapitel 5. Ergebnisse: Biologikaden Simulationen registriert wurden. Trotz der äußerst geringen prozentualen Unter-schiede kann eine Korrelation im Hinblick auf die untersuchten Konjugate beobachtetwerden (Abb. 5.24-B).Abbildung5.25:VisualisierungexemplarischerCNA-Netzwerkeder10-kDa-BCN-Konjugate von (A) PEG, (B) LPG und (C) PEtOx mittels visualCNA. Rote Stäbe illustrie-ren HB, violette Salzbrücken und grüne hydrophobe Wechselwirkungen.5.3. Untersuchungen zur thermischen Stabilität225Abbildung 5.26:Visualisierung exemplarischer CNA-Interaktionen der 10-kDa-BCN-Konjugate. Die betreffenden Aminosäuren sind violett gefärbt, die Polymerkohlenstoffatomeweiß. Constraints sind analog zu Abb. 5.25 illustriert. (A) PEG bildet viele HB-basierte Cons-traints über seine Ethersauerstoffatome aus, z. B. mit Lysinen. (B) LPG bildet ebenso vieleHB-Constraints über sein Rückgrat aus, aber auch über die zusätzlichen Hydroxymethylsei-tenketten. (C) PEtOx bildet überwiegend hydrophobe Constraints mit dem Protein aus.Abbildung 5.27: (A) Energie-Grenzwerte für die aminosäurespezifischen Perkolationsindi-zes vom Typ 2, mit großen Differenzen für die PEtOx-Konjugate in den Bereichen um V55 -N65 und R125 - T155. (B) Darstellung der Polymerdichten der BCN-Konjugate aus Abb. 5.1,mit den relevanten Bereichen aus (A) in gelb markiert. (C) Darstellung der Polymerdichtender DBCO-Konjugate aus Abb. 5.1, mit den relevanten Bereichen aus (A) in gelb markiert(Reproduzierung bzw. Nachdruck mit Genehmigung aus [304], Copyright 2021 American Che-mical Society).Korrekterweise wird bei dieser Kontakte-Analyse auch lediglich im Falle der LPG-Derivate ein signifikanter Unterschied zwischen der BCN- und der DBCO-Varianteerkannt. Es ist denkbar, dass der Anteil durch die Polymere auf verschiedene Weisenbeeinflusst werden kann: Durch Protein-Polymer-Interaktionen können üblicherweiseim Protein vorliegende Wechselwirkungen ersetzt und damit die Prozentzahlen ver-ringert werden. Dagegen kann ein Aufliegen des Polymers auf der Proteinoberflächeauch zum Erhalt bzw. zur Stärkung solcher intramolekularen Interaktionen beitragen226Kapitel 5. Ergebnisse: Biologika(vgl. [102]), was besonders im Falle der sehr flexiblen Schleifenbereiche vorstellbarist, die sich ohne Gegenwart der Polymere im Solvens leichter voneinander abwendenkönnen. Daher ist es durchaus auch nachvollziehbar, dass der Wildtyp eine niedrigereAnzahl an Kontakten aufweisen kann.Bei dieser Analyse kann ebenso eine relative „Kontaktstärke“ für jedes Atom ausge-geben werden, wobei der stärkste native Kontakt (ein Kontakt, der durchweg erhaltenbleibt) den Wert 100 erhält (somit handelt es sich um die Prozentzahl an Moment-aufnahmen, in denen der Kontakt gefunden wird). Für den exemplarischen Vergleichzwischen der BCN-LPG- und der DBCO-LPG-Variante wurde die Differenz dieser„Kontaktstärken“ ermittelt und visualisiert (Abb. 5.28-A). Es zeigt sich, dass Un-terschiede in der Anzahl der nativen Kontakte vermehrt aus Bereichen der unterenunstrukturierten Bereiche und der Termini resultieren und nicht ausschließlich amOrt der Konjugationsstelle zu verorten sind. Um letztere zeigt sich lediglich eine grö-ßere Kontaktstärke für das benachbarte I24 im Falle des DBCO-Konjugats, erhöhteKontaktstärken der BCN-Variante sind v. a. im Bereich um R125 zu finden. Dersterisch anspruchsvollere DBCO-Linker zeigt eine im Vergleich zum BCN-Linker grö-ßere Anzahl an Atomen, die sich weiter entfernt von der Proteinoberfläche befinden(Abb. 5.28-B), was die Ausrichtung der Polymerkette beeinflussen kann. Insgesamtexistieren sowohl Bereiche, in denen ein Erhalt der nativen Kontakte im Falle derBCN-Variante wahrscheinlicher ist, als auch Areale, in denen das DBCO-Konjugatgrößere „Kontaktstärken“ zeigt. Somit lässt sich kein einzelner Proteinbereich bestim-men, der für den Unterschied beider Konjugate hinsichtlich ihrer nativen Kontakteals entscheidend eingestuft werden kann.Abbildung 5.28: (A) Atombasierte Differenzen der relativen „Kontaktstärken“ zwischen derBCN-LPG- und der DBCO-LPG-Variante (KontaktstärkeBCN - KontaktstärkeDBCO). RoteAreale zeigen Bereiche, in denen in den DBCO-Simulationen native Kontakte öfter erhaltenblieben, blaue repräsentieren jene Atome, die ihre nativen Kontakte in den BCN-Simulationenlänger erhalten konnten (die Konjugationsstelle ist mit schwarzen Kugeln gekennzeichnet).(B) Verteilung der BCN- (blau) und DBCO-Atome (rot) in den Simulationen (Illustrationaller 5 ns als Punkte).5.3. Untersuchungen zur thermischen Stabilität2275.3.1.3RMSF-VerhältnisRMSF-Werte der Rückgratschweratome aller Aminosäuren wurden berechnet. Dabeiwurde die Fluktuation um die durchschnittliche Struktur der jeweiligen Simulationberechnet und zum Erhalt von Pseudo-B-Faktoren mit 8π23multipliziert. Der Durch-schnittswert aller Replika wurde jeweils mit jenem der Wildtyp-Simulationen in Re-lation gesetzt (Abb. 5.24-C und 5.24-D). Größere Einflüsse der Polymere sind v. a.im Bereich der flexiblen Schleifen und einiger Aminosäuren der Helices 2 und 3 zubeobachten (L30 - T52, Q62 - T79 und L95 - E113). Die Änderungen in diesen Berei-chen korrelieren mit den experimentell bestimmten Schmelztemperaturen, wobei eineDestabilisierung durch eine größere Fluktuation dieser Bereiche gekennzeichnet ist.5.3.1.4Vergleich mit LiP-AssaysWie in Kap. 3.2.7 erwähnt wurden sogenannte LiP-Assays durchgeführt, bei denendie Konjugate, die einer Proteolyse durch die Proteinase K unterzogen wurden, nachbestimmten Zeitintervallen massenspektrometrisch charakterisiert wurden. Die de-tektierten Peptidfragmente gaben Rückschlüsse auf die Schnittstellen durch die Pro-teinase und damit auf mögliche Schutzmechanismen der konjugierten Polymere. DieErgebnisse zu den 10-kDa-BCN-Varianten sind in Abb. 5.29 grafisch zusammenge-fasst. Es lassen sich keine bedeutenden Unterschiede zwischen den Schnittmusternder unterschiedlichen Polymervarianten erkennen, auch nicht im Vergleich zum Wild-typ. Dies deutet darauf hin, dass die Polymere lediglich schwache Interaktionen zumProtein ausbilden und somit proteinabgewandte Konformationen (Dumbbell-Modell,siehe Kap. 5.2) dominieren.Die Resultate sollten jedoch nicht direkt mit den Ergebnissen aus Kap. 5.1 in Verbin-dung gebracht werden. Während hierbei die Proteolysen zu bestimmten Zeitpunktenim Minutenbereich untersucht wurden, konnten durch die Simulationen lediglich je-weils 1.5 µs jedes Biokonjugats simuliert werden. Dadurch konnten ausschließlich pro-teinnahe Konformationen wiedergegeben werden. Die Beobachtungen unterstreichenjedoch die in Kap. 5.2 bereits erwähnte Notwendigkeit eines alternativen bzw. er-gänzenden Multiscale-Modeling-Simulationsansatzes, der die proteinfernen Polymer-konformationen in geeigneter Weise berücksichtigen kann. Weiterhin ist zu erwähnen,dass der Proteinverdau natürlich ebenso von den gewählten Zeitpunkten und v. a. derverwendeten Proteinase abhängt. Eine Analyse mit verschiedenen Proteasen, welchemehr Schnittstellen abdecken, könnte ein differenzierteres Bild liefern.228Kapitel 5. Ergebnisse: BiologikaAbbildung 5.29: Ergebnisse der limitierten Proteolyse der 10-kDa-BCN-Konjugate unddes Wildtyps mit Proteinase K. Die Aminosäuren an Position P1 und P1’ des detektiertenVerdaus sind zu gegebenen Zeitpunkten farblich markiert. Dabei wurden in (A) die Frag-mente des Wildtyps (blau), sowie der BCN-PEG- (orange) und BCN-LPG-Konjugate (grün)zur Detektion mit Ethylacetat extrahiert; in (B) wurden jene des Wildtyps (blau) und desBCN-PEtOx-Konjugats (violett) mit Chloroform extrahiert (Nachdruck mit Genehmigungaus [304], Copyright 2021 American Chemical Society).5.3.2Größere KonjugateVon den Verbundpartnern wurden für die größeren Varianten (20 - 50 kDa) zusätz-liche Schmelztemperaturmessungen vorgenommen (Abb. 5.30). Die für die 10-kDa-Konjugate beschriebenen Parameter wurden analog in den Simulationen dieser Struk-turen (siehe Kap. 5.1.2.4) untersucht. Price et al. konnten bei einer Vergrößerungeiner konjugierten PEG-Kette an der WW-Domäne über vier Monomere hinaus hinzu einer 2 kDa-PEG-Kette (entspricht ca. 33 % Massenanteil am Konjugat, ähnlichzu therapeutisch verwendeten Strukturen) keine bedeutenden Unterschiede in derthermischen Stabilität registrieren [100]. Auch bei den hier vorliegenden IFN-α2a-PEG-Konjugaten ist keine Änderung der Schmelztemperaturen mit ansteigender Po-lymermasse (10, 20, 40 kDa) erkennbar.5.3. Untersuchungen zur thermischen Stabilität229Abbildung 5.30: Per DSF ermittelte Schmelztemperaturbereiche (Mittelwert ± Standard-abweichung; n ≥ 4; signifikante Unterschiede mit p-Werten von <0.05 durch Sterne ange-deutet). Die gelisteten molaren Massen basieren auf MALDI-TOF-Messungen, dynamischerLichtstreuung (°) oder einer Größenausschlusschromatographie (*). Die Polydispersität (PDI)wurde über eine Größenausschlusschromatographie ermittelt (PEG/LPG-Daten publiziert un-ter Creative Commons Attribution License [333], PEtOx-Daten eingereicht [334]).Die Ergebnisse analoger Parameterberechnungen zu Abb. 5.24 und 5.27 sind in Abb.5.31 zusammengefasst. Im Falle der größeren PEtOx-Konjugate wird abermals eineDestabilisierung mit relativ positiven Perkolationsindizes in den Bereichen V55 - N65und R125 - T155 durch das CNA-Programm vorhergesagt. Hier kann jedoch fürkeinen der oben beschriebenen Parameter eine überzeugende Korrelation mit derSchmelztemperatur festgestellt werden, besonders das größte LPG-Konjugat zeigtbei den RMSF-Verhältnissen große Abweichungen zu den übrigen Varianten.Da diese größeren Konjugate genauso lange wie die 10-kDa-Varianten simuliert wur-den, jedoch noch weitaus mehr Freiheitsgrade besitzen und somit deutliche Konforma-tionsänderungen der Polymere vermutlich auch nur in weitaus größeren Solvensboxenrealisierbar sind, ist die Annahme eines unzureichenden Samplings als primäre Ursa-che für diese Diskrepanz gerechtfertigt. Zu diesem Aspekt kommt hinzu, dass sich dieMolekulargewichte der simulierten Ketten im Falle der größeren Konjugate deutlichervon den experimentell untersuchten Strukturen unterscheiden, als es bei den 10-kDa-Varianten der Fall ist (die modellierten Molekulargewichte stellen die anfangs experi-mentell geplanten Zielgrößen dar). So macht es vermutlich einen Unterschied, ob einIFN-α2a-K31-BCN-PEtOx25 kDa- oder ein IFN-α2a-K31-BCN-PEtOx30 kDa-Derivatmodelliert wird. Generell stellt die gegebene Polydispersität dieser Konjugate, welchebesonders im Falle der LPG-Derivate relativ stark vom Wert 1 abweicht, eine Varietätan tatsächlich vorliegenden Kettenlängen dar, welche durch die hier vorgenommenenModellierungen nicht wiedergegeben werden kann.230Kapitel 5. Ergebnisse: BiologikaAbbildung 5.31: Analoge Darstellungen der Berechnungen aus Abb. 5.24 und 5.27 mit denzusätzlich experimentell bestimmten Schmelztemperaturen aller IFN-α2a-BCN-Konjugateaus Abb. 5.23 und 5.30.5.4. Untersuchungen zur Bioaktivität2315.4Untersuchungen zur BioaktivitätZusätzlich zu DSF-Messungen wurden die Bioaktivitäten der BCN-Konjugate experi-mentell über SEAP-Assays (siehe Kap. 3.2.7) bestimmt (Abb. 5.32). Diese Aktivitätenkorrelierten mit den per dynamischer Lichtstreuung (DLS) ermittelten hydrodynami-schen Durchmessern der Konjugate (5.33-A). Ein größerer hydrodynamischer Durch-messer kann die Bindung des Proteins an IFNAR1 und IFNAR2 beeinträchtigen, dadie Polymere somit eine Annäherung des Proteins an die Rezeptoren unterbinden.Dabei zeigt das PEtOx-20-kDa-Konjugat experimentell einen Durchmesser, der eherjenem der 10-kDa-Konjugaten ähnelt, und die entsprechende 30-kDa-Variante einenDurchmesser, der jenem der 20-kDa-PEG/LPG-Konjugaten ähnelt.Da dieser Durchmesser mit einem bestimmten Faktor mit dem jeweiligen Gyrations-durchmesser korrelieren sollte, wurde dieser für alle BCN-Konjugate in silico ermitteltund mit den experimentellen Daten in Verbindung gesetzt (Abb. 5.33-B und 5.33-C).Analog zu den hydrodynamischen Durchmessern zeigt sich eine Korrelation, sodassder Gyrationsdurchmesser zur Abschätzung der Bioaktivität genutzt werden kann.Abbildung 5.32: Per SEAP-Assay und DLS ermittelte Bioaktivitäten und hydrodynamischeDurchmesser (Mittelwert ± Standardabweichung; n ≥ 3; signifikante Unterschiede mit p-Werte von <0.05 angedeutet; PEG/LPG-Daten unter Creative Commons Attribution License[333], PEtOx-Daten eingereicht [334]).232Kapitel 5. Ergebnisse: BiologikaAbbildung 5.33: (A) Korrelation der Bioaktivitäten mit den hydrodynamischen Durch-messern der Konjugate (R2 = 0.92, τB = 0.96, PEG/LPG-Daten unter Creative CommonsAttribution License [333], PEtOx-Daten eingereicht [334]). (B) Korrelation der hydrodyna-mischen Durchmesser mit berechneten Gyrationsdurchmessern (R2 = 0.85, τB = 0.78). (C)Korrelation der Gyrationsdurchmesser mit der Bioaktivität (R2 = 0.88, τB = 0.82).Sofern die Bioaktivität direkt durch die Länge des ausgestreckten Polymers bestimmtwerden kann, sollte bei selbem Molekulargewicht (z. B. 10 kDa) und vergleichbarerRückgratlänge von jeweils drei Schweratomen ein ansteigender Durchmesser in derReihenfolge PEtOx (98.13 Da pro Monomer), LPG (74.08 Da pro Monomer) undPEG (44.05 Da pro Monomer) zu erwarten sein. Die PEtOx-20-kDa-Variante zeigterwartungsgemäß einen im Vergleich zu PEG und LPG mit selbem Molekularge-wicht signifikant geringeren hydrodynamischen Durchmesser, die hydrodynamischenDurchmesser der PEG- und LPG-Varianten gleicher Größenordnung unterscheidensich jedoch nicht signifikant.Aufgrund der Beobachtungen stellt sich die Frage, ob analoge Korrelationen mit derBioaktivität auch mithilfe einfacher ermittelbarer Eigenschaften wie der Monomeran-zahl und / oder der Länge des ausgestreckten Polymers gefunden werden können, ohnehydrodynamische Durchmesser experimentell bestimmen oder längere Simulationendurchführen zu müssen. In Abb. 5.34-A und 5.34-B sind entsprechende Korrelationengezeigt, die verdeutlichen, dass eine geringfügig schlechtere Korrelation direkt mithil-fe der Monomeranzahlen erhalten wird. Zusätzlich wurde die Korrelation aus Abb.5.33-C lediglich mit den 10 Startkonformationen jedes Konjugats wiederholt, welchenach den initialen, 1 ns langen Simulationen in implizitem Solvens erhalten wurden(Abb. 5.34-C). Hier wurden zu den GaMD-Analysen nahezu identische Mittelwer-te erhalten, deren Korrelation leicht verbesserte statistische Werte aufweist. Somitstellt die Generierung dieser Startkonformationen eine ausreichende und gleichzeitigim Vergleich zu den GaMDs und experimentellen Messungen der hydrodynamischenDurchmessern deutlich schnellere Methode zur ersten Abschätzung der Bioaktivitätdar.5.4. Untersuchungen zur Bioaktivität233Abbildung 5.34: (A) Korrelation der Bioaktivitäten mit den für die Simulationen model-lierten Anzahlen an Monomeren (R2 = 0.83, τB = 0.78) und (B) den in MOE vermessenenLängen der ausgestreckten Polymerketten (R2 = 0.81, τB = 0.73). (C) Korrelation der Bio-aktivität mit den Gyrationsdurchmessern der Startkonformationen der GaMDs (R2 = 0.93,τB = 0.87). Farbschema siehe 5.33-A.234Kapitel 5. Ergebnisse: Biologika5.5ZusammenfassungIn Kap. 5.1 wurden GaMD-Simulationen für eine Vielzahl von Konjugaten unter-schiedlicher Größe und Zusammensetzung durchgeführt. Durch nähere Analysen derProtein-Polymer-Kontaktprofile wurden bevorzugte Wechselwirkungspartner der Po-lymerarten ermittelt. Im Falle der BCN-PEG-10-kDa-Variante wurden bevorzugteInteraktionen des Polymers mit Argininen und Lysinen in der Nähe hydrophober Ami-nosäuren (F, Y, L, A) detektiert, was den Beobachtungen aus der Literatur entspricht[107]. LPG zeigte ein erweitertes Interaktionsprofil, indem die zusätzlichen Hydro-xymethylseitenketten HB-Akzeptoren adressierten, Lysine wurden geringer bedeckt.LPG wies die stabilsten Wechselwirkungen aller Polymerarten auf, vornehmlich mitAspartaten und Glutamaten. PEtOx wechselwirkte über die HB-Akzeptorfähigkeitder Carbonylsauerstoffatome der Seitenketten mit HB-Donatoren wie Serinen, Glut-aminen und Asparaginen, sowie mit den hydrophoben Alkylseitenketten auch überaromatische Aminosäuren, v. a. Tyrosine. Diese Beobachtungen decken sich mit densystematischen Analysen aus Kap. 4.3. DBCO-10-kDa-Varianten zeigten verstärkteBedeckungen der Aminosäuren in der Nähe der Konjugationsstelle, favorisierte Ami-nosäuretypen stimmten allerdings zu einem großen Teil mit den Beobachtungen zuden BCN-Konjugaten überein. Größere BCN-Konjugate (20 - 50 kDa) zeigten ebensoverstärkte Wechselwirkungen mit Aminosäuren der ersten Helix und des unstruktu-rierten Bereichs zwischen Helix 1 und 2. Die größten Varianten (30 - 50 kDa) wiesenzudem stärkere Bedeckungen beider Termini auf.Die CG-Simulation eines PEG-Konjugats in Kap. 5.2 verdeutlicht, dass die Polymeredurchaus auch zum großen Teil solvensexponierte, proteinabgewandte Konformatio-nen einnehmen können, die durch die Limitationen der GaMD-Simulationen nichtbeobachtet werden können. Eine passende Gewichtung zwischen proteinnahen und-fernen Konformationen durch geeignete CG-Simulationen könnte die Ergebnisse inKap. 5.3 verbessern, da hierbei lediglich erstere berücksichtigt werden konnten.In Kap. 5.3 konnten mögliche Ursachen für eine Destabilisierung der 10-kDa-Konjugatevon IFN-α2a ermittelt werden. Die Parameter beschreiben in verschiedener Weise dieFlexibilität unterschiedlicher Proteinbereiche. Ein Effekt von PEG auf die Stabilitätvon Proteinen über eine Modulation der Flexibilität wurde bereits in der Literatur be-schrieben [92]. Vermutlich aufgrund mangelnder Stichprobengrößen wurden im Falleder größeren IFN-α2a-Konjugate keine analogen Korrelationen gefunden. Effekte, diein einer (De-)stabilisierung resultieren, repräsentieren vermutlich proteinindividuelle,multifaktorielle Phänomene. Es wären voraussichtlich, ähnlich zu den Modellprote-instudien in Kap. 4, weiterführende multivariate Korrelationen notwendig, um dieseausreichend abbilden zu können.5.5. Zusammenfassung235Kap. 5.4 zeigte, analog zu experimentell ermittelten hydrodynamischen Durchmes-sern, eine Korrelation der Gyrationsdurchmesser mit den Bioaktivitäten der Konju-gate auf. Dabei konnte eine vergleichbare Korrelation mithilfe der Startkonformatio-nen der GaMDs erhalten werden. Diese können relativ zeiteffizient modelliert werden,sodass die Ermittlung dieser Gyrationsdurchmesser eine wertvolle Alternative zur ex-perimentellen Bestimmung hydrodynamischer Durchmesser darstellen kann. Da Kor-relationen mit der Anzahl an Monomeren oder den Polymerlängen zu etwas schlech-teren Ergebnissen führten, hängt die erniedrigte Bioaktivität nicht ausschließlich vonder Länge der ausgestreckten Polymerketten ab.Insgesamt konnte durch dieses Kapitel demonstriert werden, dass Simulationen grö-ßerer Konjugate auf atomarer Ebene durchaus wertvolle Beiträge zur Interpretationexperimenteller Befunde liefern können. Sie können weiterhin dazu dienen, favorisierteProtein-Polymer-Wechselwirkungspartner zu detektieren. Die genannten Limitierun-gen machen es jedoch unabdingbar, für weitergehende Analysen solcher Systeme inZukunft deutlich zeitintensivere Simulationen durchzuführen und / oder zur Abde-ckung des möglichen Konformationsraums geeignete Parameter für CG-Kraftfelder zuentwickeln. Letztere können jedoch spezifische Wechselwirkungen mit den Biologikaunter Umständen nicht akkurat wiedergeben. Nur mit einem geeigneten Multiscale-Modeling-Ansatz [335] könnten also detaillierte Polymer-Protein-Interaktionen wie inKap. 5.1.2 unter gleichzeitiger Berücksichtigung proteinabgewandter Polymerkonfor-mationen über längere Simulationszeiten hinweg getreu abgebildet werden. 2376.Fazit und Ausblick6.1FazitPBK stellen im Vergleich zu klassischen Proteinsystemen aufgrund der kovalent ge-bundenen, in der Regel sehr großen Polymerketten äußerst flexible Strukturen dar.Computergestützte Verfahren zur Evaluation von Protein-Protein-Interaktionen, bspw.über eigens dafür entwickelte Docking-Programme mit Scoring-Funktionen, erhieltenin den letzten Jahren große Aufmerksamkeit [336–339]. Proteine verfügen über charak-teristische Sekundär-/Tertiär-/Quartärstrukturen, aufgebaut aus den 20 bekanntenproteinogenen Aminosäuren mit unterschiedlichen physikochemischen Eigenschaften,die die Entwicklung solcher generalisierter Ansätze erlauben. Dagegen bilden die zurBiokonjugation genutzten Polymere, vornehmlich PEG, in der Regel keine geordnetenStrukturbereiche aus. Die simplen Etherstrukturen von PEG lassen die Vermutungeher unspezifischer, schwacher Interaktionen mit Proteinen zu. Entsprechend ist ne-ben der Nutzung von Docking-Ansätzen auch die Anwendung experimenteller Me-thoden, bspw. der Röntgenstrukturanalyse, zur Strukturaufklärung dieser Systemeextrem limitiert.Die Charakterisierung von PBK profitierte in der Vergangenheit jedoch von moleku-lardynamischen Simulationsmethoden, die die Dynamik dieser Systeme abbilden kön-nen. Weithin etablierte Kraftfelder zur Abbildung solcher hybrider Systeme stehenjedoch noch nicht zur Verfügung, bisher wurden unterschiedlichste Parametersätzezur Simulation genutzt. Dies macht einen Vergleich dieser Studien schwierig. Vergrö-berte Kraftfelder erlauben das Abbilden von Konformationsänderungen über größereZeitintervalle hinweg und können hier sinnvoll sein, resultieren jedoch auch in einemdeutlichen Detailverlust, der ein Studium spezifischer Protein-Polymer- oder Solvens-Polymer-Wechselwirkungen verhindert. Es stellt sich die Frage, ob charakteristischeWechselwirkungen über deutlich kleinere Oligomer-Modellprotein-Konjugate studiertwerden können. Diese stellen Systeme dar, die auch in annehmbarer Zeit mit ato-marem Detail simuliert werden können. Systematische Studien aus dem Labor vonJ. L. Price deuten darauf hin, dass die thermische Stabilität größerer Konjugate vonden Eigenschaften solcher Systeme abgeleitet werden kann. Hinweise auf den Ein-fluss entropisch günstiger Wasserverdrängungen untermauern die Vermutung, dassvergröberte Kraftfelder hierfür unzulänglich sein könnten.238Kapitel 6. Fazit und AusblickIn Kap. 4 wurde daher untersucht, ob für die thermische Stabilität entscheidendeEigenschaften aus MD-Simulationen dieser Systeme extrahiert werden können. Dievon Price et al. experimentell determinierten ∆∆G0f-Werte wurden in multivariatenQSPR-Vorhersagemodellen als Antwortvariablen genutzt, verschiedenste Eigenschaf-ten aus den MD-Simulationen als Deskriptoren. Die ersten Modellgenerierungen inKap. 4.1.1 zeigten zu optimistische Leistungen, da die besten Modelle auf Basis derLeistung des Holdout-Sets ausgewählt wurden, sowie sehr komplexe Modelltypen miteiner Vielzahl an Deskriptoren für einen relativ kleinen Datensatz verwendet wurden.Ansätze in Kap. 4.1.2 wurden auf lineare Regressions- und logistische Klassifikations-modelle beschränkt, mit einem Holdout-Set, welches nicht zur Auswahl des bestenModells diente. Dies führte zu Modellen, die eine zur kristallstrukturbasierten Richt-linie von Price et al. verbesserte Leistung zeigten und dabei sehr viel mehr Konju-gate mit größerer struktureller Vielfalt abdeckten. Es wurden Deskriptoren genutzt,die sowohl spezifische Protein-PEG-Wechselwirkungen als auch die Solvatation derStrukturen beschreiben.Die Modelle untermauern das Gesamtbild, welches sich aus der bisherigen Literaturergibt: PEG kann über verschiedene Mechanismen zur Erhöhung der konformativenStabilität führen. Hierzu können spezifische Protein-Polymer-Wechselwirkungen [99],Stärkungen intramolekularer Salzbrücken [102] oder NH-π-Bindungen [104] oder auchDesolvatationen um Hydroxylgruppen [100] oder apolare Flächen [103] beitragen.Ein multivariates Vorhersagemodell sollte daher zur Ermittlung der PEG-basiertenStabilisierung geeignet sein. Die entscheidende Limitierung dieser Studie stellte jedochder zur Verfügung stehende Datensatz dar. Zur Etablierung generalisierter Modellemüsste dieser in Zukunft noch deutlich erweitert werden. Die in Kap. 4.2 gezeigtenMetadynamik-Simulationen können als Startpunkt für weiterführende Studien dienenund zeigen einen möglichen Mechanismus einer PEG-basierten Stabilisierung währenddes Entfaltungsprozesses auf.Ungeachtet einer Abschätzung der thermodynamischen Stabilität können die Analy-sen zu favorisierten Wechselwirkungspartnern aus den Kap. 4.3 und 5.1 dabei helfen,die Dynamik der PEG-Alternativen LPG, PMeOx und PEtOx abzuschätzen. PEG in-teragierte, in Übereinstimmung mit bisherigen Publikationen, besonders mit Lysinenund Argininen in der Nähe apolarer Aminosäuren. LPG adressierte über zusätzlicheHydroxylgruppen auch Aspartate und Glutamate. Die POx-Varianten wechselwirk-ten einerseits mit HB-Donatoren über ihre Carbonylgruppen und andererseits mitden aromatischen Aminosäuren Tyrosin und Phenylalanin. Die Simulationen zu Kon-jugaten von IFN-α2a in Kombination mit Untersuchungen zur thermischen Stabilitätund Bioaktivität demonstrieren, dass in silico-Ansätze bei der Interpretation expe-rimenteller Beobachtungen helfen können. Die CG-Simulation verdeutlicht schließ-lich: Zur Untersuchung therapierelevanter Biokonjugate ist ein Multiscale-Modeling-Ansatz notwendig. So könnten spezifische Interaktionen evaluiert und gleichzeitig inRelation zu proteinabgewandten Polymerkonformationen gewichtet werden.6.1. Fazit239Abbildung 6.1: Zusammenfassende Illustration der Arbeit.240Kapitel 6. Fazit und Ausblick6.2AusblickAusgehend von den Ergebnissen dieser Arbeit lassen sich eine Vielzahl an denkba-ren Fortführungen der Studien zu PBK festhalten. Ein Teil soll im Folgenden kurzaufgeführt werden.6.2.1Studien zu Modellproteinen6.2.1.1Vergrößerung des DatensatzesDie größte Limitierung bzgl. der Generierung geeigneter Vorhersagemodelle in Kap.4 stellte die geringe Anzahl an Konjugaten mit geeigneten experimentellen Datendar. Eine deutlich größere Menge an charakterisierten Varianten mit einer breite-ren Verteilung der ∆∆G0f-Werte wäre essenziell, möglichst mit einem gleichmäßigenVorkommen der verwendeten Linkertypen und Konjugationsbereiche (β-Faltblätter,α-Helices, unstrukturierte Elemente). Zur Erstellung generalisierter Modelle wärendaher ebenso mehr Konjugate anderer Proteine nötig.Zusätzlich könnten die simulierten Oligomer-Konjugate der Polymeralternativen ana-log charakterisiert werden und so deren Vorhersagen überprüft werden. Für diesekönnten zusätzliche Modelle entwickelt werden. Durch einen anschließenden Ver-gleich bedeutender Modelldeskriptoren könnte überprüft werden, ob für diese PEG-Alternativen andere Eigenschaften einen wesentlich bedeutenderen Einfluss auf diethermische Stabilisierung zeigen als es bei PEG der Fall ist. Bspw. könnten, ausge-hend von den Beobachtungen zur Polymerdynamik, die Anzahl an Protein-Polymer-HB im Falle von LPG oder hydrophobe Interaktionen im Falle der POx-Varianteneinen stärkeren Effekt haben.6.2.1.2Weitere QSPR-ModelleBei der Vielzahl an gewonnenen Deskriptoren wäre eine Nutzung von Modellartendenkbar, die mehrere kollineare Parameter in einigen wenigen latenten Variablen ver-einen. Der entwickelte Modellgenerierungsansatz in Kap. 4.1.2 könnte mit Modellar-ten auf Basis der PLS-Methode getestet werden. Die Generierung von PLS-Modellenkönnte dabei weiterhin von zusätzlichen 3D-Deskriptoren profitieren. Diese Informa-tionen könnten aus Dichtekarten gewonnen werden. So könnte z. B. die Übereinstim-mung des Aufenthalts der Polymerketten oder bestimmter Desolvatations-Hotspotsneu simulierter Konjugate mit jenen stark stabilisierter Varianten verglichen wer-den. Differenzdichtekarten zwischen den Simulationen PEGylierter und unPEGylier-ter Strukturen wurden bspw. für das Solvens erstellt und darauf aufbauend die Anzahlan Gitterpunkten zwischen den beiden Konjugaten 16-1 und 27-1 verglichen, an denenstarke Desolvatationen registriert wurden. Detailliertere Arbeiten mit vergleichbarenDichtekarten könnten zur Gewinnung geeigneterer Deskriptoren führen.6.2. Ausblick2416.2.1.3Sampling der ungefalteten VariantenWie einige in Kap. 1.1.6.1 aufgeführte Publikationen suggerieren [95], könnte einePEGylierung ebenso einen starken Einfluss auf die ungefaltete Variante eines Pro-teins haben. In der hier vorliegenden Arbeit wurden solche Konformationen nichtuntersucht. Ein großes Ensemble an jeweils 50 verschiedenen, ungefalteten Konfor-mationen der unPEGylierten Varianten von 16-1 und 23-1 wurde testweise über denProtSA-Webserver generiert [340] und, mitsamt PEGylierter Versionen, in kurzenReplika (je 2 ns) simuliert, um die solvenszugänglichen Oberflächen mit jenen der ge-falteten Varianten zu vergleichen. Die Idee hierbei war, dass der abschirmende Effektdes Polymers bei stabilisierten Varianten deutlich größer für die gefaltete Varianteim Vergleich zur ungefalteten ist (und somit die gefaltete Struktur im Vergleich zurungefalteten stärker stabilisiert wird). Die Ergebnisse zeigten jedoch keine Unterschie-de zwischen den Konjugaten. Eventuell könnten weitere Deskriptoren zwischen den(un)gefalteten Konformationen verglichen werden. Solche Ensembles könnten ebensodazu dienen, mittels erweiterter Sampling-Methoden die Faltungsgeschwindigkeit desKonjugats zu ermitteln.6.2.1.4Analyse anderer EigenschaftenVergleichbare Simulationsstudien zu Affinitäts- oder Proteolysedaten kleinerer Mo-dellproteine sind denkbar. Solche Daten standen für diese Arbeit nicht zur Verfügungund konnten nur in sehr begrenztem Maße in Bezug auf größere Konjugate untersuchtwerden. Während die thermodynamische Stabilität von Konjugaten nicht direkt ausden beobachteten Wechselwirkungen des Polymers mit dem Protein ableitbar ist, wä-ren Schlussfolgerungen zu Affinitäten der Konjugate bei bekannter Binderegion aufBasis solcher Simulationen einfacher. Eine verringerte Affinität sollte mit dem Aus-maß der sterischen Abschirmung der Binderegion in den Simulationen korrelieren. Inähnlicher Weise sollte die Abschirmung bekannter Schnittstellen von Proteasen mitProteolysedaten übereinstimmen.Eine weitere pharmazeutisch wichtige Eigenschaft stellt die Aggregationstendenz vonBiologika dar. Durch den AGGRESCAN3D-Webserver wurden bspw. potenzielle Hot-spots der untersuchten Proteine detektiert, die maßgeblich zu einer Aggregation füh-ren könnten. Hilfsstoffe in pharmazeutischen Formulierungen können dazu beitragen,diese Areale zu maskieren [341–343]. In ähnlicher Weise könnte eine in silico-Studiezur polymerbasierten Abschirmung dieser Proteinbereiche zur Abschätzung der Ag-gregationstendenz von Konjugaten beitragen.Im Hinblick auf die thermische Stabilität könnte der Fokus bspw. auch von den∆∆G0f-Werten weg, hin zu möglichen Entfaltungszentren der Proteine gerichtet wer-den. So wurden bspw. mit dem CNA-Programm mögliche Weak spots der Proteineidentifiziert. Es könnte näher untersucht werden, inwiefern eine Abschirmung dieserBereiche mit einer Stabilisierung der Konjugate einhergeht.242Kapitel 6. Fazit und Ausblick6.2.1.5Experimentelle AnsätzeDie hier präsentierten Modellierungsarbeiten basierten auf publizierten Stabilitäts-daten, die über eine temperaturabhängige CD-Spektroskopie ermittelt wurden. ZurEtablierung von Datensätzen wären auch andere Messmethoden zur Charakterisie-rung der Proteinstabilität denkbar, bspw. über eine chemisch induzierte Entfaltungmittels Guanidiniumchlorid [95] oder DSF-Messungen analog zu jenen der größerenKonjugate in dieser Arbeit. Eine Alternative zum beschriebenen LiP-Assay zum Stu-dium der Protein-Polymer-Interaktionen könnten NMR-Messungen darstellen. Erstkürzlich wurde ein solcher Ansatz mit 1H/15N-HSQC (Heteronuclear Single Quan-tum Coherence)-Messungen beschrieben [15], wobei jedoch fraglich ist, inwiefern die2,2,6,6-Tetramethylpiperidinyloxyl (TEMPO)-Strukturen, die hierbei als Spin labelseingesetzt wurden, die Polymerdynamik beeinflussen.6.2.1.6Alternative Simulationsansätze6.2.1.6.1Metadynamik-SimulationenDer in dieser Arbeit entwickelte Ansatz zu Metadynamik-basierten Entfaltungssi-mulationen könnte auf eine größere Auswahl an Konjugaten mit deutlich längerenSimulationszeiten erweitert werden. So könnte die Eignung zur Abschätzung PEG-basierter Stabilisierungen evaluiert werden. Die engen Energieintervalle erlauben abervermutlich keinen verlässlichen quantitativen Vergleich.6.2.1.6.2Erhöhte TemperaturenNeben klassischen MD-Simulationen bei 300 K wurden zu Beginn der Arbeit einigewenige Konjugationsstellen ebenso bei erhöhten Temperaturen simuliert (330 K und360 K). Dieser Ansatz wurde jedoch nicht weiter verfolgt, da die Anzahl benötigterSimulationen bei analoger Untersuchung des gesamten Datensatzes immens ansteigenwürde. Dennoch wäre es retrospektiv gesehen interessant zu studieren, ob bestimmteberechnete Eigenschaften (z. B. der Sekundärstrukturanteil und die Anzahl nativerKontakte) bei erhöhten Temperaturen im Falle besonders destabilisierter Variantendeutlich stärker moduliert werden als es bei stabilisierten Konjugaten der Fall ist. Inden Anfängen dieser Arbeit wurde für einige wenige Varianten bspw. die Temperatur-abhängigkeit der Rückgratorderparameter untersucht, da diese in einer Studie überRibonucleasen bereits sehr gut mit den experimentell bestimmten Schmelztemperatu-ren korrelierten [344]. Nach zusätzlichen Simulationen weiterer Konjugate ergab sichaber keine vielversprechende Korrelation mehr. Es müsste im Einzelfall überprüftwerden, in welcher Art und Weise die jeweiligen Eigenschaften eine Temperaturab-hängigkeit zeigen. Vergleichbare Simulationen könnten auch zur Beobachtung einerkompletten Entfaltung des Proteins dienen.6.2. Ausblick2436.2.1.6.3Gerichtete MD-SimulationenNeben Metadynamik-Simulationen existieren andere Simulationsansätze zur Untersu-chung von Nichtequilibriumsbedingungen. Gerichtete MD-Simulationen können bspw.genutzt werden, um mithilfe der Jarzynski-Gleichung die Bindungskinetiken von Li-ganden zu studieren [345, 346]. Durch Anlegen einer konstanten Kraft zur Entfaltungeiner einfachen, PEGylierten Helix wurde bereits der schützende Effekt des Polymersauf die Rückgrat-HB untersucht [94]. Ein solcher Ansatz wurde in dieser Arbeit nichtverfolgt, da die zugrundeliegenden Modellproteine wesentlich komplexere Struktu-ren aufweisen und die Auswahl eines möglichen Angriffspunkts zur Entfaltung somitkeineswegs trivial ist.6.2.1.6.4Entfaltung über Gō-ModelleZum Studium der Proteinentfaltung können ebenso extrem vereinfachte Gō-Modelledienen, die eine Art der Vergröberung des Kraftfelds repräsentieren [253, 256, 347].Dabei werden Aminosäuren als einzelne Kugeln dargestellt. So wurde bspw. ein Gō-Modell für einen Replica Exchange-Ansatz zum Studium der nativen Kontakte vonT4-Lysozym-PEG-Konjugaten mit DBCO-Linker verwendet [348]. Vergleichbare Gō-basierte Simulationsansätze wurden in dieser Arbeit mit dem MMTSB-Toolset [349]für CHARMM [350], sowie mit dem Programm GENESIS [351] für unPEGylierteVarianten getestet. Hier besteht jedoch grundsätzlich das Problem einer geeignetenParameterentwicklung für den Polymer- und Linker-Part der Konjugate.6.2.1.7MD-AlternativenWeiterhin existieren weitaus recheneffizientere Alternativen zu MDs zur schnellen Be-urteilung der Stabilität und Dynamik biomolekularer Strukturen. Verschiedene An-sätze sind bspw. über das Python-Paket PolyDy [352] möglich: Analysen der Normal-modi, sowie anisotropische oder auch Gaußsche Netzwerkmodelle. Diese Methodenkönnten direkt zur Beurteilung von Stabilitäten dienen oder aber zur Gewinnungmöglicher Deskriptoren für die Modellgenerierung. Dabei stellt sich aber die Fragenach einer geeigneten Implementierung der PEGylierten Aminosäure im Netzwerk.6.2.2Größere Konjugate6.2.2.1Verbesserung des Samplings6.2.2.1.1Abbildung des SolvensEine entscheidende Limitierung dieser Studie stellte die begrenzte Stichprobennahmeder größeren Konjugate dar, die, gerade aufgrund des expliziten Solvens, Systememit zahlreichen Freiheitsgraden darstellen. Denkbar wären hierzu Simulationen mitimplizitem Solvens, Hybrid-Simulationen (z. B. mit dem PACE-Kraftfeld [353]) odereine Nutzung des neu entwickelten Solvation Shell-Modells [354].244Kapitel 6. Fazit und Ausblick6.2.2.1.2GaMD-AlternativenIn dieser Studie wurden zur Unterstützung des Samplings GaMDs durchgeführt. Auchwenn die generierten Daten grundsätzlich zur energetischen Regewichtung entlangausgewählter Reaktionskoordinaten (z. B. Gyrationsradien, SASA) zur Verfügungstehen, könnten klassische beschleunigte MD-Simulationen oder auch andere Ansät-ze wie Replica Exchange-Methoden das Sampling weiter verbessern, auch wenn einegenaue Regewichtung entlang bestimmter Variablen dadurch evtl. nicht mehr gewähr-leistet ist.6.2.2.1.3Vergröberte KraftfelderDie Etablierung geeigneter CG-Parameter könnte die Analyse größerer Konjugatemithilfe längerer Simulationen erleichtern. Ein Protokoll zur Simulation PEGylier-ter Biologika mit dem neuen MARTINI 3-Kraftfeld wurde kürzlich publiziert [163].Im Falle der alternativen Polymere müssten geeignete Parameter vorerst entwickeltwerden. In einer kürzlich publizierten Studie zu CG-Modellen von Polyethyleniminenwurde darauf hingewiesen, dass diese als Startpunkt zur Entwicklung von Parameternfür Poly(2-alkyl-2-oxazoline) dienen könnten [355]. Diese wurden mit einer zusätzli-chen Kugel für die Seitenkette testweise im PolyPly-Programm implementiert. Ebensowurden für LPG die zur Verfügung stehenden PEG-Parameter mit einer kleinen Ku-gel ähnlich jener der Seitenkette von Serin erweitert. Die generierten Modelle führtenin ersten CG-Simulationen jedoch zur vollständigen Aggregation der Polymere aufdie Proteinoberfläche, was auf unzureichende Parameter hindeutet. WeiterführendeStudien unter iterativem Abgleich mit Ergebnissen aus klassischen MD-Simulationenund Experimenten wären notwendig. Zuletzt stellt aber auch die korrekte Parametri-sierung der komplexen Cyclooctin-Linker eine große Herausforderung dar.Wären vergröberte Kraftfelder für die entsprechenden Systeme etabliert, so könnteweiterhin ein Backmapping-Ansatz eingesetzt werden, um geeignete Konformationenfür weitergehende Analysen gemäß Kap. 5.3.1 durchzuführen [356]. So könnten dieErgebnisse der CN-Analysen eventuell verbessert werden, da der Anteil an protein-abgewandten Konformationen berücksichtigt werden würde.6.2.2.2MD-AlternativenNeben molekulardynamischen Ansätzen bieten bspw. auch Methoden, die auf demPrinzip des maschinellen Lernens basieren, Möglichkeiten, den Konformationsraumvon Systemen zu untersuchen, so z. B. der Ansatz des Generative Topographic Map-ping [357, 358] oder jener der Boltzmann-Generatoren [359]. Auch rigiditätsbasierte,geometrische Simulationsmethoden existieren, um den Konformationsraum größererProteinsysteme zu studieren [360].245ZusammenfassungPolymer-Biokonjugationen, vornehmlich mit dem Goldstandard PEG, führen zu einerverbesserten Pharmakokinetik, beeinflussen aber auch die konformative Stabilität vonProteinen. Bisherige Mutationsstudien, in denen überwiegend (Asn)PEG4-Konjugateder β-faltblattstrukturreichen, humanen Pin 1 WW-Domäne untersucht wurden, pos-tulieren auf einer Proteindesolvatation beruhende Stabilisierungsmechanismen: eineStärkung intramolekularer Salzbrücken und NH-π-Bindungen, sowie entropisch güns-tige Wasserverdrängungen um apolare Aminosäuren und Hydroxylgruppen.Ziel dieser Arbeit ist es, die Protein-Polymer-Dynamik auf molekularer Ebene zu cha-rakterisieren, um damit rationale Ansätze zum Design neuer Biokonjugate voranzu-treiben und mögliche PEG-Alternativen zu etablieren. Hierzu wurde eine Vielzahl anDeskriptoren mittels Molekulardynamik-Simulationen der WW-Konjugate gewonnenund mit publizierten Stabilitätsdaten (∆∆G0f-Werte) in multivariaten Regressions-und logistischen Klassifikationsmodellen korreliert. Die gewonnenen QSPR-Modelledecken im Vergleich zu einer bereits publizierten, kristallstrukturbasierten Richtli-nie einen größeren und strukturell vielfältigeren Datensatz an Konjugaten ab undzeigen gleichzeitig, auch für ein Konjugat der Src SH3-Domäne, eine deutlich verbes-serte Leistung. Die Modelldeskriptoren beschreiben sowohl eine Modulation der Sol-vatation als auch Protein-Polymer-Interaktionen. Metadynamik-Simulationen zeigtenzudem die Polymerdynamik während einer partiellen Proteinentfaltung auf. Mithil-fe weiterer Simulationen von Konjugaten des α-helikalen Her2-Affibodys wurde dieDynamik von PEG und verschiedener Alternativen (LPG, PEtOx, PMeOx) syste-matisch studiert. PEG interagierte mit positiv geladenen Lysinen und Argininen inder Nähe hydrophober Aminosäuren. LPG zeigte zusätzliche Wechselwirkungen derHydroxylgruppen mit Aspartaten und Glutamaten. POx-Polymere interagierten mitPhenylalaninen, Tyrosinen und über Carbonylgruppen mit HB-Donatoren.Größere Konjugate (10 - 50 kDa PEG/LPG/PEtOx) des antiviralen BiologikumsInterferon-α2a wurden mittels gaußbeschleunigter MDs und einer CG-Simulation ana-lysiert. Charakteristische Wechselwirkungspartner stimmten mit den Beobachtungenzu Oligomer-Konjugaten überein. In Einklang mit experimentellen Daten der Koope-rationspartner zu den 10-kDa-Varianten deuteten zusätzliche Constrained-Network-Analysen, welche die Proteinflexibilität evaluieren, auf eine thermische Destabilisie-rung hin. Die Bioaktivität der untersuchten Konjugate wurde weiterhin erfolgreichmit den Gyrationsdurchmessern der modellierten Strukturen korreliert. 247SummaryBioconjugation of polymers, mainly the gold standard PEG, can improve pharmako-kinetic properties but also modulate conformational stability of proteins. Mutationstudies on (Asn)PEG4 conjugates of the β-sheet rich human Pin 1 WW domain sug-gest various desolvation effects playing a crucial role: strengthening of intramolecularsalt-bridges and NH-π bonds, as well as entropically favorable water expulsion aroundhydrophobic patches and hydroxyl groups.The goal of this study is to characterize protein-polymer dynamics on a molecularlevel to drive forward rational design of new bioconjugates and establish viable PEGalternatives. A variety of descriptors was calculated from molecular dynamics simula-tions of WW conjugates and correlated with published stability data (∆∆G0f values)generating multivariate regression and logistic classification models. Compared to apreviously published crystal structure-based guideline, QSPR models covered a struc-turally more diverse and bigger dataset and showed significantly improved predictions,including for a conjugate of the Src SH3 domain. Model descriptors captured modula-tions of solvation as well as protein-polymer interactions. Metadynamics simulationsdepicted PEG dynamics upon partial protein unfolding. Combined with simulationsfor conjugates of the α-helical Her2 affibody, data was further used to systematical-ly dissect the dynamics of PEG and its alternatives LPG, PEtOx and PMeOx. PEGinteracted with lysines and arginines near hydrophobic patches. LPG additionally ad-dressed aspartates and glutamates via its hydroxyl groups. POx variants interactedwith phenylalanines, tyrosines, as well as hydrogen bond donors via carbonyl groups.Larger conjugates (10 - 50 kDa PEG/LPG/PEtOx) of antiviral biologic Interferon-α2awere analyzed via Gaussian accelerated MDs and an exemplary CG simulation. In-teraction patterns agreed with observations for oligomer conjugates. In accordancewith experimental data of collaboration partners for 10 kDa variants, constrained net-work analyses, assessing protein flexibility, suggested a thermal destabilization uponbioconjugation. Bioactivity of conjugates was further successfully correlated with dia-meters of gyration of modeled structures. 249Anhang 251A.AusgleichungsrechnungenA.1Zugrundeliegende GleichungenDie Elliptizität [θ]λ steht, unter Annahme einer klassischen Zwei-Phasen-Kinetik derEntfaltung, mit der temperaturabhängigen Faltungskonstante Kf, die das Gleich-gewicht zwischen der gefalteten ([F] mit [θf]) und der ungefalteten ([U] mit [θu])Struktur bzw. den Anteil des ersteren (f) beschreibt, in folgendem Zusammenhang[100, 102, 104]:[θ] = (1 − f) [θu] + f[θf](A.1)[θf] = N0 + N1T(A.2)[θu] = D0 + D1T(A.3)Kf = [F][U](A.4)=f1 − f(A.5)[θ] = (D0 + D1T) + Kf(N0 + N1T)1 + Kf(A.6)wobei D0 und N0 die Ordinatenabschnitte, sowie D1 und N1 die Steigungen derRegressionsgeraden für das denaturierte (D, Basislinie nach Entfaltung) bzw. dasnative (N, Basislinie vor Entfaltung) Protein darstellen. Aus der Faltungskonstantekann wiederum die freie Faltungsenergie ∆G0f ermittelt werden:1Kf = e−∆G0fRT(A.7)∆G0f steht unmittelbar in Zusammenhang mit der Enthalpieänderung ∆Hf(Tm) undder Änderung der Wärmekapazität ∆Cp:∆G0f = ∆Hf(Tm)Tm(Tm − T) + ∆Cp [T − Tm − T ln( TTm)](A.8)1In den Referenzpublikationen von Price et al. wird die Angabe zu Standardbedingungen teilweisenicht aufgeführt (Bezeichnung als ∆Gf anstatt ∆G0f).252Anhang A. AusgleichungsrechnungenDurch Kombination der Glg. A.1 - A.8 ergibt sich folgende Bedingung:[θ] = (D0 + D1T) + e−[∆Hf (Tm)Tm(Tm−T )+∆Cp [T −Tm−T ln( TTm )]]RT(N0 + N1T)1 + e−[∆Hf (Tm)Tm(Tm−T )+∆Cp [T −Tm−T ln( TTm )]]RT(A.9)Eine Ausgleichungsrechnung (Fitting) kann somit mit den experimentell ermitteltenWerten für [θ] basierend auf dieser Beziehung durchgeführt werden. Hierbei sind D0,N0, D1, N1, ∆Hf(Tm), ∆Cp und Tm variable Parameter. Glg. A.8 basiert auf folgen-den Beziehungen:∆G0 = ∆H0 − T∆S0(A.10)∆H0 = ∆Hf(Tm) + ∆Cp (T − Tm)(A.11)∆S0 = ∆Sf(Tm) + ∆Cp ln( TTm)(A.12)∆Sf(Tm) = ∆Hf(Tm)Tm(A.13)Die Ergebnisse der Fitting-Parameter, besonders ∆Hf(Tm), hängen bei dieser Me-thode bei einer einzigen Denaturierungskurve in der Regel stark von der gewähltenAnzahl an Punkten zur Erstellung der Regressionslinien der nativen und denaturier-ten Varianten ab [361]. Durch Kombination aller drei Messreihen einer Variante ineinem einzigen Fitting-Prozess können die Variablen aber mit einer ausreichendenSicherheit bestimmt werden. Dies wird erreicht, indem die Messreihen [θ]1, [θ]2 und[θ]3 einer Triplika-Messung in einem einzigen Vektor [θ]ges vereinigt werden und fürdie rechte Seite der Glg. A.9 dann jeweils die für den Messwert passenden Basislinien-Variablen verwendet werden (d. h., für Messwerte aus [θ]1 werden z. B. N01, N11,D01 und D11 genutzt). Die Variablen Tm, ∆Hf(Tm) und ∆Cp kommen dagegen beijedem Messwert von [θ]ges zum Einsatz. Die Ausgleichungsrechnung basiert also aufder Minimierung der Residuen aller Messpunkte aus [θ]ges (Glg. A.15). Es handeltsich somit, bei Triplika-Messungen, um einen Fitting-Prozess mit insgesamt 15 Va-riablen: Tm, ∆Hf(Tm), ∆Cp, N01, N11, D01, D11, N02, N12, D02, D12, N03, N13,D03 und D13 [104].[θ]ges =�[θ]1, [θ]2, [θ]3�(A.14)A.2. Fehlerfortpflanzung253[θ] =(D01+D11 T)+e−[∆Hf (Tm)Tm(Tm−T )+∆Cp [T −Tm−T ln( TTm )]]RT(N01+N11 T)1+e−[∆Hf (Tm)Tm(Tm−T )+∆Cp [T −Tm−T ln( TTm )]]RT[θ] ∈ [θ]1(D02+D12 T)+e−[∆Hf (Tm)Tm(Tm−T )+∆Cp [T −Tm−T ln( TTm )]]RT(N02+N12 T)1+e−[∆Hf (Tm)Tm(Tm−T )+∆Cp [T −Tm−T ln( TTm )]]RT[θ] ∈ [θ]2(D03+D13 T)+e−[∆Hf (Tm)Tm(Tm−T )+∆Cp [T −Tm−T ln( TTm )]]RT(N03+N13 T)1+e−[∆Hf (Tm)Tm(Tm−T )+∆Cp [T −Tm−T ln( TTm )]]RT[θ] ∈ [θ]3(A.15)Erhalten werden schließlich die relevanten Parameter Tm, ∆Hf(Tm) und ∆Cp für diePEGylierte und die unPEGylierte Variante, mit dazugehörigen Standardfehlern desFits. Aus diesen wird mit Glg. A.8 zunächst ∆G0f bei der Schmelztemperatur Tmund daraus abgeleitet −T∆S0f der unPEGylierten Variante berechnet. ∆G0f der un-PEGylierten Struktur beträgt bei dieser Temperatur 0 kcal/mol, Kf beträgt 1 undf 0.5. Anschließend werden ∆G0f, ∆H0f und −T∆S0f der PEGylierten Variante beider Schmelztemperatur Tm der unPEGylierten Variante berechnet und somit die ge-suchten energetischen Differenzen zwischen der PEGylierten und der unPEGyliertenVariante erhalten, die als Indikator für eine veränderte thermodynamische Stabilitätdienen: die Differenz der freien Faltungsenthalpien ∆∆G0f und den damit zusammen-hängenden Faltungsenthalpien ∆∆H0f und -entropien −T∆∆S0f [86, 88, 100, 104].Für die nichtlinearen Kurven werden je nach verwendetem Analysenprogramm (z.B. GraphPad, SigmaPlot oder Psi-Plot) unterschiedliche Optimierungsalgorithmenzur Minimerung der Residuen genutzt, wobei zumeist der Levenberg-Marquardt-Algorithmus genutzt wird [86, 88, 362].A.2FehlerfortpflanzungZum Erhalt der Standardabweichung einer Variable x, die über eine Funktion f mitden Parametern a, b und c in Zusammenhang steht, werden die Fehler sa, sb, sc gemäßGaußscher Fehlerfortpflanzung mit den partiellen Ableitungen verrechnet.sf =�(δf(x)δasa)2 + (δf(x)δbsb)2 + (δf(x)δcsc)2(A.16)254Anhang A. AusgleichungsrechnungenAuf Basis der Glg. A.8 gilt für den Fehler von ∆G0f somit [102]:d∆G0fd∆Hf(Tm) = Tm − TTm(A.17)d∆G0fd∆Cp= T − Tm − T ln[ TTm](A.18)d∆G0fd∆Tm= ∆Cp [ TTm− 1] − ∆Hf(Tm)Tm[Tm − TTm− 1](A.19)σ∆G0f =�(d∆G0fd∆Hf(Tm)s∆Hf(Tm))2 + (d∆G0fd∆Cps∆Cp))2 + (d∆G0fd∆TmsTm))2(A.20)255B.DeskriptorenViele verschiedene Eigenschaften, die für Vorhersagemodelle als Deskriptoren genutztwerden können, können aus MD-Trajektorien gewonnen werden. Im Folgenden sol-len einige zentrale Konzepte kurz erklärt werden, die in der vorliegenden Arbeit eineAnwendung finden. Fast alle genannten Parameter können mithilfe von cpptraj vonAmberTools18 berechnet werden [282, 363]. Die zu verwendenden Befehle sind mitaufgeführt. Wenn nicht anders referenziert stammen die entsprechenden Informatio-nen aus dem Handbuch von AMBER18 [282].B.1Mittlere Abweichung (RMSD)cpptraj: rmsdDer RMSD-Wert (Root-mean-square deviation) beschreibt die mittlere Abweichungeiner Struktur von einer Referenzstruktur. In der Regel wird als Referenz die zugrun-deliegende Kristallstruktur genutzt. Der Wert lässt sich nach folgender Gleichungberechnen, wobei N die Anzahl der Atome, xi die Position des i-ten Atoms im Raumund xrefidie Position dieses Atoms in der Referenzstruktur darstellen [364]:RMSD =���� 1NN�i=1(xi − xrefi)2(B.1)B.2Mittlere Fluktuation (RMSF)cpptraj: rmsfWährend der RMSD-Wert einen Mittelwert zu einem bestimmten Zeitpunkt angibt,stellt der RMSF-Wert (Root-mean-square fluctuation) die Fluktuation über ein ge-wisses Zeitintervall dar [365]:RMSF =���� 1TT�j=1(xj − xrefj)2(B.2)Hierbei wird die Abweichung eines einzigen Partikels von der Referenzstruktur alsoüber das Zeitintervall j bis T gemittelt. Als Referenz dient hier in der Regel die256Anhang B. Deskriptorengemittelte Position des jeweiligen Partikels. Der RMSF-Wert für eine Gruppe vonAtomen kann massengewichtet über folgende Gleichung berechnet werden:RMSFGesamt =� Atomfluktuationi Massei� Massei(B.3)Die Fluktuationen können weitergehend auch quadriert und mit dem Faktor 83π2 ge-wichtet werden, um einen zu den B-Faktoren von Kristallstrukturanalysen vergleich-baren Wert zu erhalten.B.3Solvenszugängliche Oberfläche (SASA)cpptraj: surf, molsurfÜber den Befehl surf kann der Beitrag einer Atomauswahl zur Gesamtoberflächedes jeweiligen Moleküls gemäß des LCPO-Algorithmus (Linear combination of pair-wise overlaps) berechnet werden [366]. Hierbei wird der Van-der-Waals-Radius miteinem Faktor von 1.4 multipliziert. Im Gegensatz hierzu berechnet der molsurf -Befehldie absolute Connolly-Oberfläche (standardmäßiger Sondenradius: 1.4 Å) [367] einerAuswahl ohne Berücksichtigung der Bindung zu übrigen Atomen. Die Auswahl wirdhierbei also als isoliertes System betrachtet.B.4Radiale Verteilungsfunktion (RDF)cpptraj: rdfÜber die radiale Verteilungsfunktion (Radial distribution function) wird die Wahr-scheinlichkeit wiedergegeben, mit der eine Atomauswahl A innerhalb eines bestimm-ten Abstands R von der Atomauswahl B lokalisiert ist. Dieser Wert wird aus einemnormalisierten Histogramm berechnet, mit der Formel:RDF = Dichte ∗ ([4π3 (R + dR)3] − [4π3 dR3])(B.4)Dabei entspricht dR der Länge eines Intervalls des Histogramms. In der Regel wirddiese Berechnung für Solvensmoleküle um bestimmte Proteinatome genutzt, wobeicharakteristische Maxima für die unterschiedlichen Hydrathüllen erkennbar werden.In diesem Falle sollte die RDF-Funktion bei größeren Distanzen im Bereich des Bulk-Wassers einen Wert von 1 annehmen.B.5Gyrationsradiuscpptraj: radgyrDer Gyrationsradius Rg (oder Streumassenradius) ist eine bekannte Größe in derPolymerforschung und stellt die mittlere Distanz aller Atome zum Massen- oder geo-metrischen Zentrum der Auswahl dar. Nach IUPAP (Internationale Union für reineB.6. HB/Lebensdauer-Analyse257und angewandte Physik) ist der Radius definiert nach [368]:Rg =��ni=1 mir2i�ni=1 mi(B.5)Dabei stellen mi die Masse und ri die Distanz des i-ten Atoms zum Massenzentrumder gesamten Auswahl dar. Für bestimmte Formen (z. B. Zylinder oder Kugeln)existieren auch Umrechnungsmethoden zur Ermittlung des hydrodynamischen RadiusRh der Struktur [369].B.6HB/Lebensdauer-Analysecpptraj: hbond, lifetimeÜber den hbond-Befehl können die Wasserstoffbrückenbindungen (HB) pro Moment-aufnahme, mit definierten Grenzen bzgl. Distanz und Winkel, aus einer Trajektorieerhalten werden. Diese Daten können anschließend mit dem lifetime-Befehl genutztwerden, um die durchschnittlichen und maximalen Zeitspannen zu berechnen, überdie eine HB beibehalten wird.B.7Interaktionsenergiecpptraj: lieDie Interaktionsenergie zwischen zwei Atomauswahlen kann über einen vereinfachtenAnsatz mithilfe von lie mit einem standardmäßigen Cut-off-Wert von 12 Å für Van-der-Waals- und elektrostatische Wechselwirkungen berechnet werden. Es wird einesimple Shifting-Funktion genutzt, wobei die Coulomb-Interaktionen über folgendeGleichung berechnet werden:E = kqiqjrij(1 −r2ijr2cut)2(B.6)Dabei wird die Konstante k durch die vorher definierbare Dielektrizitätskonstantebestimmt (normalerweise jene des verwendeten Wassermodels), qi und qj stellen diePotentialladungen der Atome, rij die Distanz und rcut die Cut-off-Distanz dar. Eshandelt sich somit um eine Berechnung ohne Berücksichtigung expliziter Solvensmo-leküle.B.8Rückgrat-Diederwinkelcpptraj: multidihedralÜber multidihedral können die Φ- und Ψ-Winkel jeder Aminosäure bestimmt werden,über die die Konformation des Proteins beschrieben werden kann. Diese Diederwinkelsind bekannt aus Ramachandran-Plots [370]. Der Diederwinkel Φ wird über die AtomeCi-1, Ni, Cαi und Ci definiert, Ψ über die Atome Ni, Cαi, Ci und Ni+1 (Abb. B.1).258Anhang B. DeskriptorenAbbildung B.1: Illustration der Winkel Φ und Ψ.B.9Ordnungsparametercpptraj: ired, matrix, diagmatrix, vectorDie Bestimmung generalisierter Ordnungsparameter S2 beruht im Zuge des iRED-Ansatzes (Isotropic reorientational eigenmode dynamics) auf einer Diagonalisierungeiner aus einer Hauptkomponentenanalyse erhaltenen Kovarianzmatrix [371]. AusMDs können diese Parameter auf Basis der x-, y- und z-Komponenten der Einheits-vektoren entlang der zu untersuchenden Bindung über folgende Gleichung erhaltenwerden [372]:S2 = 32[(x2)2 + (y2)2 + (z2)2 + 2(xy2)2 + 2(xz2)2 + 2(yz2)2] − 12(B.7)Dieser Parameter beschreibt die interne Dynamik der Bindung und bewegt sich imBereich zwischen 0 und 1. Kleine Werte signalisieren eine sehr flexible, hohe einerigide Bindung. Diese stehen mit der molekularen Korrelationsfunktion CM(t) in Zu-sammenhang, die eine im Raum frei bewegliche Bindung beschreibt. Um die Rota-tionskorrelationsfunktion CT (t) zu definieren, ist weiterhin eine von der molekularenKorrelation unabhängige, zusätzliche Korrelationsfunktion CE(t) notwendig, die diepartielle Rigidität aufgrund der kovalenten Bindung zum Makromolekül berücksich-tigt [373–375]:CT (t) =CM(t) CE(t)(B.8)=e−tτM (S2 + (1 − S2) e−tτE )(B.9)=S2 e−tτM + (1 − S2) e−tτT(B.10)Hierbei stellen τM, τE und τT die entsprechenden Korrelationszeiten dar. Experimen-tell können Orderparameter über NMR-Messungen der Spektraldichtefunktion J(ω)ermittelt werden. Werden S2-Werte für bestimmte Seitenkettenbindungen oder fürdie N-H-Bindungen des Proteinrückgrats ermittelt, so können über diese die Konfor-mationsentropien S auf Basis der Konstanten M, A und B berechnet werden:S = kBM[A + B(1 − S2)](B.11)B.10. Native Kontakte259M stellt dabei die Anzahl an Diederwinkeln dar und A und B sind Fitting-Parameter,die sich je nach zu untersuchender Aminosäure unterscheiden. Diese Fits wurdenauf Basis einer Korrelation von S2NMR-Messwerten mit Konformationsentropien Spubliziert, wobei letztere aus MD-Simulationen mithilfe von Diederwinkeln über einenvon-Mises-Kernel-Ansatz stammen [376].B.10Native Kontaktecpptraj: nativecontactsDie Stabilität der nativen Konformation kann über eine Messung ihrer Atom-Atom-Kontakte erfolgen. Diese Methode wird bspw. in Studien zur thermischen Stabilitätvon Proteinen [332] oder Faltungen von Go-Modellen [258] genutzt. Zunächst werdenalle Atom-Atom-Kontakte in der Referenzstruktur (üblicherweise der Kristallstruk-tur) über einen Cut-off-Wert von 4.5 Å definiert [377]. Für jede Momentaufnahmeder Simulation mit einer Distanz unterhalb dieses Werts wird ein Kontakt registriert.So kann der Anteil an Momentaufnahmen bestimmt werden, in der die Kontakte derReferenzstruktur wiedergefunden wurden.B.11Diffusionskoeffizientcpptraj: diffusionDer Diffusionskoeffizient D entlang der Dimensionen n steht mit der mittleren qua-dratischen Verschiebung (MSD) über die Zeit t über die Einstein-Beziehung in Zu-sammenhang:2nD = limt→∞MSDt(B.12)Diese Kalkulation sollte für ein statistisch valides Ergebnis über eine Vielzahl an Ato-men durchgeführt werden und eignet sich daher in der Regel nur für Solvensmoleküle.B.12HB-AutokorrelationAls Alternative zur weiter oben genannten Lebensdauer-Analyse in cpptraj kann mit-hilfe des mdanalysis-Moduls [378, 379] die Autokorrelationsfunktion C(t) und darausdie Überlebenszeit τ von HB hij zwischen den Atomen i und j ab dem Zeitpunkt t0berechnet werden [380]:τ =� ∞0C(t) = ⟨hij(t0)hij(t0 + t)hij(t0)2⟩(B.13)Hier kann die kontinuierliche oder die intermittierende Überlebenszeit berechnet wer-den. Bei letzterer sind zwischenzeitliche Bindungsbrüche erlaubt.260Anhang B. DeskriptorenB.13CNA-basierte DeskriptorenMit dem CNA-Programm kann die Entfaltungstemperatur über einen sigmoidalenoder Spline-Fit bestimmt werden. Hierzu eignen sich die verschiedenen globalen Pa-rameter (die Clusterkonfigurationsentropie H, der Rigiditätsordnungsparameter P∞,die Floppy Mode-Dichte Φ, die durchschnittliche Rigiditätsclustergröße S) zur Detek-tierung unterschiedlicher Entfaltungsmomente. Die Entfaltung eines Protein erfolgtin der Regel in mehreren Schritten. So detektiert die Clusterkonfigurationsentropievom Typ 1 (CCE1) bspw. besonders den ersten Übergang von einer rigiden zu ei-ner flexiblen Struktur, während jene vom Typ 2 (CCE2) eine höhere Sensitivität fürbeim Entfaltungsprozess später ablaufende Flexibilisierungen aufweist (eine genaueBeschreibung aller Parameter ist im SI-Material von [235] zu finden). Die Entfal-tungstemperatur wird somit durch den Übergang der Struktur von einem überwie-gend rigiden zu einem flexiblen Netzwerk definiert. Da sich die Temperatur auf Basiseiner bisherigen Studie zu 20 Proteinen aus dem Cut-off-Wert der Energie ableitet(siehe Kap. 1.4.2), sollte diese Temperatur nicht als absoluter Wert beurteilt werden,sie kann jedoch als Vergleich der Stabilität zweier vergleichbarer Systeme dienen.Zusätzlich zu diesen globalen Parametern berechnet das Programm auch lokale Para-meter (Perkolationsindizes pi, Rigiditätsindizes ri und Stabilitätskarten rcij), die dieRigidität einzelner Aminosäuren beschreiben. Schließlich können daraus sogenannteWeak spots erfasst werden. Diese stellen Aminosäuren dar, die sich beim Erreichen derEntfaltungstemperatur vom größten Cluster lösen und somit potenzielle Mutations-orte zur Verbesserung der thermischen Stabilität des Makromoleküls repräsentieren.Eine Beschreibung aller berechenbarer CNA-Parameter ist in Tab. B.1 gegeben.B.13. CNA-basierte Deskriptoren261Tabelle B.1: Bei der CN-Analyse berechenbare lokale und globale Parameter, die die Rigi-dität des Biomakromoleküls beschreiben. Ebenso sind die vier verschiedenen Arten von Weakspots aufgeführt, die auf Basis der Phasenübergangstemperatur erfasst werden [235].ParameterArtBeschreibungRigiditätsordnungsparameter P∞globalAnteil des Netzwerks, welches Teil des aktuell größten Clusters (Typ 1)bzw. des ursprünglich größten Clusters ist (Typ 2).Clusterkonfigurationsentropie HglobalWahrscheinlichkeit, dass sich ein Atom innerhalb eines Clustersmit einer bestimmten Größe befindet (Grad der Unordnung des Netzwerks).Bei CCE2 fließt die Clustergröße mit ihrem Quadrat in die Berechnung ein.Durchschnittliche Rigiditätsclustergröße SglobalDurchschnittliche Clustergröße, wobei der größteCluster nicht in die Berechnung einfließt.Floppy Mode-Dichte ΦglobalAnzahl an Floppy Modes (Freiheitsgraden) in Relation zur Anzahl an Atomen.Perkolationsindizes pilokalLokales Analog zu P∞. Beschreibt die Wasserstoffbrückenbindungsenergie,bei der sich eine Bindung vom größten ursprünglichen Cluster (Typ 1)oder vom aktuell größten Cluster (Typ 2) löst.Rigiditätsindizes rilokalÄhnlich wie pi, wobei hier die Energie bestimmt wird, bei der sich die Bindungvon jeglichem Cluster löst.Seitenkettenrigiditätsindizes rsclokalEnergie, bei der sich Cα und Cβ nicht mehr im selben Cluster befinden.Stabilitätskarten rcijlokalEnergie, bei der sich die Cα-Atomezweier Aminosäuren nicht mehr im selben Cluster befinden.Weak spot Typ 1lokalDie Aminosäure, die Teil des größten Clusters ist, wird gebräuchlichnach derPhasenübergangstemperatur flexibel.Weak spot Typ 2lokalDie Aminosäure jeglichen Clusters (mind. 30 Atome) wird nach derPhasenübergangstemperatur flexibel.Weak spot Typ 3lokalAminosäuren, die Wasserstoffbrückenbindungen bilden, welche nach derPhasenübergangstemperatur gelöst werden.Weak spot Typ 4lokalAminosäuren innerhalb von 5 Å der Aminosäuren vom Weak spot Typ 3. 263C.KonjugateTabelle C.1: Chemische Strukturformeln aller in den Modellproteinen untersuchten, kon-jugierten Aminosäuren. In der rechten Spalte ist jeweils ein Kennzeichen für die jeweiligeStruktur gegeben, welches in Tabelle C.2 verwendet wird.unkonjugiertkonjugiertKennzeichen123BJOU$ϵλ264Anhang C. Konjugateψ4567φγω89Z0Anhang C. Konjugate265X§Tabelle C.2: Liste aller simulierten Konjugate der Modellproteine. In der ersten Spalte wur-de jedem Biokonjugat ein Name nach einem einheitlichen Schema AAAA.XX(YY)-Z gegeben.Der Name beginnt mit dem vierstelligen PDB-Code AAAA der jeweiligen Struktur. Danachsteht XX für die Konjugationsstelle (im Falle doppelt konjugierter Varianten ergänzt durch+XX). Der in manchen Fällen enthaltene Term (YY) beschreibt im Vergleich zum Wildtypzusätzlich vorkommende Punktmutationen. Z entspricht einem Kennzeichen aus Tabelle C.1und beschreibt die chemische Struktur der konjugierten Aminosäure. Unkonjugierte Varian-ten werden im Haupttext lediglich mit der Konjugationsstelle XX bezeichnet. Der in derjeweiligen originalen Publikation für das entsprechende Konjugat verwendete Name ist inder dritten Spalte gegeben. In der vierten Spalte ist die Aminosäuresequenz gegeben, wobeidie konjugierte Aminosäure mit dem Kennzeichen aus C.1 in Blau dargestellt ist. Etwaigezusätzliche Mutationen (YY) sind orange gefärbt. Die jeweils experimentell ermittelte freieFaltungsenergie ist in der fünften Spalte aufgeführt. Die sechste Spalte R gibt die Referenz-publikation an. In der letzten Spalte A ist gekennzeichnet, für welchen QSPR-Modellansatzdas betreffende Konjugat verwendet wurde (1, 2 oder 1+2). Varianten, welche aus keinerReferenzpublikation stammen, enthalten die Kennzeichnungen NA (not available). TerminaleAcetylgruppen und Amide sind jeweils mit A- bzw. -N in der Sequenz gegeben. Die Sequenzist für Di- und Trimere jeweils nur für ein Monomer dargestellt. Der Buchstabe χ symbolisierteine Cyclohexylalanin-Mutation, ν eine Naphthylalanin-Mutation.Nr.NameName inReferenzSequenz∆∆G0f[kcal/mol]RA11PIN.19-B6PEG1KLPPGWEKRMSRSBGRVYYFNHITNASQFERPSG-0.34 ± 0.04[97]1+221PIN.19-J6PEG2KLPPGWEKRMSRSJGRVYYFNHITNASQFERPSG-0.37 ± 0.05[97]1+231PIN.19-O6PEG3KLPPGWEKRMSRSOGRVYYFNHITNASQFERPSG-0.72 ± 0.07[97]1+241PIN.19-U6PEG8KLPPGWEKRMSRSUGRVYYFNHITNASQFERPSG-0.61 ± 0.04[97]1+251PIN.14-114pKLPPGWEK1MSRSSGRVYYFNHITNASQFERPSG0.00 ± 0.40[100]1+261PIN.16-116pKLPPGWEKRM1RSSGRVYYFNHITNASQFERPSG-0.90 ± 0.03[100]1+271PIN.17-117pKLPPGWEKRMS1SSGRVYYFNHITNASQFERPSG-0.18 ± 0.05[100]1+281PIN.18-118pKLPPGWEKRMSR1SGRVYYFNHITNASQFERPSG0.00 ± 0.07[100]1+291PIN.19-119pKLPPGWEKRMSRS1GRVYYFNHITNASQFERPSG-0.70 ± 0.04[100]1+2101PIN.23-123pKLPPGWEKRMSRSSGRV1YFNHITNASQFERPSG0.40 ± 0.10[100]1+2111PIN.26-126pKLPPGWEKRMSRSSGRVYYF1HITNASQFERPSG-0.58 ± 0.06[100]1+2121PIN.27-127pKLPPGWEKRMSRSSGRVYYFN1ITNASQFERPSG0.38 ± 0.04[100]1+2131PIN.28-128pKLPPGWEKRMSRSSGRVYYFNH1TNASQFERPSG0.00 ± 0.07[100]1+2141PIN.29-129pKLPPGWEKRMSRSSGRVYYFNHI1NASQFERPSG-0.36 ± 0.04[100]1+2151PIN.30-130pKLPPGWEKRMSRSSGRVYYFNHIT1ASQFERPSG0.00 ± 0.07[100]1+2161PIN.32-132pKLPPGWEKRMSRSSGRVYYFNHITNA1QFERPSG-0.45 ± 0.02[100]1+2171PIN.19-8D-19pKLPPGWEKRMSRS8GRVYYFNHITNASQFERPSG0.01 ± 0.04[100]1+2181PIN.19(S16A)-119p-S16AKLPPGWEKRMARS1GRVYYFNHITNASQFERPSG-0.51 ± 0.02[100]1+2191PIN.19(Y23F)-119p-Y23FKLPPGWEKRMSRS1GRVFYFNHITNASQFERPSG-0.43 ± 0.03[100]1+2201PIN.19(S32A)-119p-S32AKLPPGWEKRMSRS1GRVYYFNHITNAAQFERPSG-0.71 ± 0.03[100]1+2211PIN.19(S16A:Y23F)-119p-S16A,Y23FKLPPGWEKRMARS1GRVFYFNHITNASQFERPSG-0.72 ± 0.08[100]1+2266Anhang C. Konjugate221PIN.16(Y23F)-116p-Y23FKLPPGWEKRM1RSSGRVFYFNHITNASQFERPSG-0.45 ± 0.06[100]1+2231PIN.16(S32A)-116p-S32AKLPPGWEKRM1RSSGRVYYFNHITNAAQFERPSG-0.58 ± 0.03[100]1+2241PIN.26(T29A)-126p-T29AKLPPGWEKRMSRSSGRVYYF1HIANASQFERPSG-0.32 ± 0.06[100]1+2251PIN.29(S32A)-129p-S32AKLPPGWEKRMSRSSGRVYYFNHI1NAAQFERPSG-0.88 ± 0.06[100]1+2261PIN.32(Y23F)-132p-Y23FKLPPGWEKRMSRSSGRVFYFNHITNA1QFERPSG-0.61 ± 0.11[100]1+2271PIN.16+26-116p/26pKLPPGWEKRM1RSSGRVYYF1HITNASQFERPSG-1.38 ± 0.03[100]1+2281PIN.19+26-119p/26pKLPPGWEKRMSRS1GRVYYF1HITNASQFERPSG-1.26 ± 0.02[100]1+2291PIN.26+29-126p/29pKLPPGWEKRMSRSSGRVYYF1HI1NASQFERPSG-0.56 ± 0.04[100]1+2301PIN.16+19-116p/19pKLPPGWEKRM1RS1GRVYYFNHITNASQFERPSG-0.61 ± 0.11[100]1+2311PIN.16-$16p45KLPPGWEKRM$RSSGRVYYFNHITNASQFERPSG-0.39 ± 0.03[100]1+2321PIN.18-$18p45KLPPGWEKRMSR$SGRVYYFNHITNASQFERPSG0.02 ± 0.08[100]1+2331PIN.19-$19p45KLPPGWEKRMSRS$GRVYYFNHITNASQFERPSG-0.67 ± 0.05[100]1+2341PIN.26-$26p45KLPPGWEKRMSRSSGRVYYF$HITNASQFERPSG-0.27 ± 0.06[100]1+2351PIN.27-$27p45KLPPGWEKRMSRSSGRVYYFN$ITNASQFERPSG0.65 ± 0.05[100]1+2361PIN.28-$28p45KLPPGWEKRMSRSSGRVYYFNH$TNASQFERPSG0.36 ± 0.05[100]1+2371PIN.29-$29p45KLPPGWEKRMSRSSGRVYYFNHI$NASQFERPSG-0.36 ± 0.04[100]1+2381PIN.19-21-nbpKLPPGWEKRMSRS2GRVYYFNHITNASQFERPSG-1.19 ± 0.02[101]1+2391PIN.19-31-cbpKLPPGWEKRMSRS3GRVYYFNHITNASQFERPSG-1.13 ± 0.04[101]1+2401PIN.19-0m1pKLPPGWEKRMSRS0GRVYYFNHITNASQFERPSG-0.30 ± 0.01[101]1+2411PIN.19-72p-1KLPPGWEKRMSRS7GRVYYFNHITNASQFERPSG-0.16 ± 0.02[101]1+2421PIN.19-62p-2KLPPGWEKRMSRS6GRVYYFNHITNASQFERPSG-0.08 ± 0.02[101]1+2431PIN.19-52p-3KLPPGWEKRMSRS5GRVYYFNHITNASQFERPSG0.01 ± 0.02[101]1+2441PIN.19-42p-4KLPPGWEKRMSRS4GRVYYFNHITNASQFERPSG0.24 ± 0.02[101]1+2451PIN.19-Z4pKLPPGWEKRMSRSZGRVYYFNHITNASQFERPSG-0.04 ± 0.02[101]1+2461PIN.19-95pKLPPGWEKRMSRS9GRVYYFNHITNASQFERPSG-0.37 ± 0.01[101]1+2471PIN.19-X6pKLPPGWEKRMSRSXGRVYYFNHITNASQFERPSG-0.59 ± 0.03[101]1+2481PIN.19-§3pKLPPGWEKRMSRS§GRVYYFNHITNASQFERPSG-0.36 ± 0.01[101]2491PIN.18(S16D:R17A)-1pβ18KLPPGWEKRMDA1SGRVYYFNHITNASQFERPSG-0.25 ± 0.02[102]1+2501PIN.18(S16D:R17A:R21A)-1pβ18-DAKLPPGWEKRMDA1SGAVYYFNHITNASQFERPSG-0.28 ± 0.01[102]1+2511PIN.18(R17A)-1pβ18-SRKLPPGWEKRMSA1SGRVYYFNHITNASQFERPSG-0.51 ± 0.01[102]1+2521PIN.18(R17A:R21A)-1pβ18-SAKLPPGWEKRMSA1SGAVYYFNHITNASQFERPSG-0.57 ± 0.02[102]1+2531PIN.23(R14A)-Zpβ23-EAKLPPGWEKAMSRSSGRVZYFNHITNASQFERPSG-0.11 ± 0.02[102]1+2541PIN.23(E12A)-Zpβ23-ARKLPPGWAKRMSRSSGRVZYFNHITNASQFERPSG-0.05 ± 0.02[102]1+2551PIN.23(E12A:R14A)-Zpβ23-AAKLPPGWAKAMSRSSGRVZYFNHITNASQFERPSG-0.30 ± 0.02[102]1+2561PIN.23(E12χ:R14χ)-ZpXXKLPPGWχKχMSRSSGRVZYFNHITNASQFERPSG-1.30 ± 0.02[102]2571PIN.23(E12χ:R14A)-ZpXAKLPPGWχKAMSRSSGRVZYFNHITNASQFERPSG-0.40 ± 0.02[102]2581PIN.23(E12χ:R14L)-ZpXLKLPPGWχKLMSRSSGRVZYFNHITNASQFERPSG-0.90 ± 0.02[102]2591PIN.23(E12A:R14χ)-ZpAXKLPPGWAKχMSRSSGRVZYFNHITNASQFERPSG-0.51 ± 0.02[102]2601PIN.23(E12L:R14L)-ZpLLKLPPGWLKLMSRSSGRVZYFNHITNASQFERPSG-0.89 ± 0.04[103]1+2611PIN.23(E12L:R14A)-ZpLAKLPPGWLKAMSRSSGRVZYFNHITNASQFERPSG-0.31 ± 0.02[103]1+2621PIN.23(E12A:R14L)-ZpALKLPPGWAKLMSRSSGRVZYFNHITNASQFERPSG-0.56 ± 0.04[103]1+2631PIN.23(E12F:R14F)-ZpFFKLPPGWFKFMSRSSGRVZYFNHITNASQFERPSG-0.33 ± 0.02[103]1+2641PIN.23(E12F:R14A)-ZpFAKLPPGWFKAMSRSSGRVZYFNHITNASQFERPSG-0.36 ± 0.02[103]1+2651PIN.23(E12A:R14F)-ZpAFKLPPGWAKFMSRSSGRVZYFNHITNASQFERPSG-0.37 ± 0.02[103]1+2661PIN.23(E12F:R14L)-ZpFLKLPPGWFKLMSRSSGRVZYFNHITNASQFERPSG-0.68 ± 0.02[103]1+2671PIN.23(E12L:R14F)-ZpLFKLPPGWLKFMSRSSGRVZYFNHITNASQFERPSG-0.56 ± 0.02[103]1+2681PIN.16-916QpKLPPGWEKRM9RSSGRVYYFNHITNASQFERPSG-0.67 ± 0.02[118]1+2691PIN.16-Z16ZpKLPPGWEKRMZRSSGRVYYFNHITNASQFERPSG0.07 ± 0.05[118]1+2701PIN.16-216NbpKLPPGWEKRM2RSSGRVYYFNHITNASQFERPSG-1.70 ± 0.03[118]1+2Anhang C. Konjugate267711PIN.16-§16XpKLPPGWEKRM§RSSGRVYYFNHITNASQFERPSG-0.39 ± 0.04[118]2721PIN.18-918QpKLPPGWEKRMSR9SGRVYYFNHITNASQFERPSG-0.12 ± 0.02[118]1+2731PIN.18-Z18ZpKLPPGWEKRMSRZSGRVYYFNHITNASQFERPSG-0.05 ± 0.04[118]1+2741PIN.18-218NbpKLPPGWEKRMSR2SGRVYYFNHITNASQFERPSG0.03 ± 0.01[118]1+2751PIN.18-§18XpKLPPGWEKRMSR§SGRVYYFNHITNASQFERPSG0.00 ± 0.02[118]2761PIN.23-923QpKLPPGWEKRMSRSSGRV9YFNHITNASQFERPSG0.23 ± 0.04[118]1+2771PIN.23-Z23ZpKLPPGWEKRMSRSSGRVZYFNHITNASQFERPSG-0.29 ± 0.03[118]1+2781PIN.23-223NbpKLPPGWEKRMSRSSGRV2YFNHITNASQFERPSG0.18 ± 0.10[118]1+2791PIN.23-§23XpKLPPGWEKRMSRSSGRV§YFNHITNASQFERPSG0.43 ± 0.16[118]2801PIN.27-927QpKLPPGWEKRMSRSSGRVYYFN9ITNASQFERPSG0.27 ± 0.02[118]1+2811PIN.27-Z27ZpKLPPGWEKRMSRSSGRVYYFNZITNASQFERPSG-0.62 ± 0.03[118]1+2821PIN.27-227NbpKLPPGWEKRMSRSSGRVYYFN2ITNASQFERPSG1.26 ± 0.06[118]1+2831PIN.27-§27XpKLPPGWEKRMSRSSGRVYYFN§ITNASQFERPSG0.68 ± 0.04[118]2841PIN.29-929QpKLPPGWEKRMSRSSGRVYYFNHI9NASQFERPSG0.16 ± 0.06[118]1+2851PIN.29-Z29ZpKLPPGWEKRMSRSSGRVYYFNHIZNASQFERPSG-0.17 ± 0.06[118]1+2861PIN.29-229NbpKLPPGWEKRMSRSSGRVYYFNHI2NASQFERPSG-0.67 ± 0.02[118]1+2871PIN.29-§29XpKLPPGWEKRMSRSSGRVYYFNHI§NASQFERPSG0.32 ± 0.07[118]2881PIN.32-932QpKLPPGWEKRMSRSSGRVYYFNHITNA9QFERPSG-0.44 ± 0.02[118]1+2891PIN.32-Z32ZpKLPPGWEKRMSRSSGRVYYFNHITNAZQFERPSG-0.33 ± 0.03[118]1+2901PIN.32-232NbpKLPPGWEKRMSRSSGRVYYFNHITNA2QFERPSG-0.93 ± 0.03[118]1+2911PIN.32-§32XpKLPPGWEKRMSRSSGRVYYFNHITNA§QFERPSG0.01 ± 0.03[118]2921PIN.19(S16F:R21T)-16PEG4-F,TKLPPGWEKRMFRS1GTVYYFNHITNASQFERPSG-0.81 ± 0.10[98]1+2931PIN.19(S16F)-16PEG4-FKLPPGWEKRMFRS1GRVYYFNHITNASQFERPSG-1.13 ± 0.07[98]1+2941PIN.19(R21T)-16PEG4-TKLPPGWEKRMSRS1GTVYYFNHITNASQFERPSG-0.55 ± 0.06[98]1+2951PIN.19(R21A)-16PEG4-AKLPPGWEKRMSRS1GAVYYFNHITNASQFERPSG-0.96 ± 0.12[98]1+2961PIN.19(R21L)-16PEG4-LKLPPGWEKRMSRS1GLVYYFNHITNASQFERPSG-0.68 ± 0.11[98]1+2971SRL.20-1SH3T20NPEGTFVALYDYESR1ETDLSFKKGERLQIVNNTEGDWWLAHSLTTGQTGYIPSNYVAPS-1.20 ± 0.10[100]2982ZTA.1-1p2α1A-1MKQLEDKVEELLSKNYHLENEVARLKKLVG-N0.16 ± 0.04[102]2992ZTA.3-1p2α3A-RM1QLEDKVEELLSKNYHLENEVARLKKLVG-N0.24 ± 0.04[102]21002ZTA.4-1p2α4A-RMK1LEDKVEELLSKNYHLENEVARLKKLVG-N0.70 ± 0.04[102]21012ZTA.6-1p2α6A-RMKQL1DKVEELLSKNYHLENEVARLKKLVG-N0.46 ± 0.02[102]21022ZTA.7-1p2α7A-RMKQLE1KVEELLSKNYHLENEVARLKKLVG-N0.72 ± 0.02[102]21032ZTA.10-1p2α10A-RMKQLEDKV1ELLSKNYHLENEVARLKKLVG-N1.34 ± 0.04[102]21042ZTA.14-1p2α14A-RMKQLEDKVEELL1KNYHLENEVARLKKLVG-N-0.14 ± 0.06[102]21052ZTA.18-1p2α18A-RMKQLEDKVEELLSKNY1LENEVARLKKLVG-N-0.22 ± 0.02[102]21062ZTA.21-1p2α21A-RMKQLEDKVEELLSKNYHLE1EVARLKKLVG-N0.48 ± 0.02[102]21072ZTA.25-1p2α25A-RMKQLEDKVEELLSKNYHLENEVA1LKKLVG-N0.34 ± 0.02[102]21082ZTA.28-1p2α28A-RMKQLEDKVEELLSKNYHLENEVARLK1LVG-N-0.26 ± 0.02[102]21092ZTA.4(E11A)-1p2α4-KAA-RMK1LEDKVEALLSKNYHLENEVARLKKLVG-N-0.04 ± 0.01[102]21102ZTA.4(K8A)-1p2α4-AEA-RMK1LEDAVEELLSKNYHLENEVARLKKLVG-N-0.12 ± 0.02[102]21112ZTA.4(K8A:E11A)-1p2α4-AAA-RMK1LEDAVEALLSKNYHLENEVARLKKLVG-N-0.70 ± 0.02[102]21122ZTA.18(R25A)-1p2α18-EAA-RMKQLEDKVEELLSKNY1LENEVAALKKLVG-N0.56 ± 0.02[102]21132ZTA.18(E22A)-1p2α18-ARA-RMKQLEDKVEELLSKNY1LENAVARLKKLVG-N0.28 ± 0.02[102]21142ZTA.18(E22A:R25A)-1p2α18-AAA-RMKQLEDKVEELLSKNY1LENAVAALKKLVG-N-0.32 ± 0.01[102]21151COI.1-1p3α1Ace-9VEALEKKVAALESKVQALEKKVEALEY-N0.48 ± 0.01[102]21161COI.1(E3A)-1p3α1-AKA-9VAALEKKVAALESKVQALEKKVEALEY-N0.45 ± 0.01[102]21171COI.1(K7A)-1p3α1-EAA-9VEALEAKVAALESKVQALEKKVEALEY-N0.51 ± 0.01[102]21181COI.1(E3A:K7A)-1p3α1-AAA-9VAALEAKVAALESKVQALEKKVEALEY-N0.33 ± 0.01[102]2268Anhang C. Konjugate1191COI.6-1p3α6A-EVEAL9KKVAALESKVQALEKKVEALEY-N0.96 ± 0.01[102]21201COI.6(K8A)-1p3α6-AEA-EVEAL9KAVAALESKVQALEKKVEALEY-N0.42 ± 0.01[102]21211COI.6(E13A)-1p3α6-KAA-EVEAL9KKVAALASKVQALEKKVEALEY-N0.39 ± 0.01[102]21221COI.6(K8A:E13A)-1p3α6-AAA-EVEAL9KAVAALASKVQALEKKVEALEY-N-0.51 ± 0.01[102]21235UXT.14-1NpA-EVEALEKKVEALE1KVQKLEKKVEALEHGWDGR-N0.50 ± 0.02[264]1245UXT.14(E10A)-1Np1A-EVEALEKKVAALE1KVQKLEKKVEALEHGWDGR-N-0.24 ± 0.02[264]1255UXT.14(K18A)-1Np2A-EVEALEKKVEALE1KVQALEKKVEALEHGWDGR-N0.22 ± 0.02[264]1265UXT.14(E10A:K18A)-1Np3A-EVEALEKKVAALE1KVQALEKKVEALEHGWDGR-N-0.62 ± 0.02[264]1271PIN.26(W11F)-1FNpKLPPGFEKRMSRSSGRVYYF1HITNASQFERPSG-0.34 ± 0.02[104]1281PIN.26-(W11ν)-1ZNpKLPPGνEKRMSRSSGRVYYF1HITNASQFERPSG-0.62 ± 0.03[104]1301PIN.26(W11Y)-1YNpKLPPGYEKRMSRSSGRVYYF1HITNASQFERPSG-0.47 ± 0.05[104]1311PIN.26-9WQpKLPPGWEKRMSRSSGRVYYF9HITNASQFERPSG0.14 ± 0.02[104]1321PIN.7-1NAK1PPGWEKRSMSRSSGRVYYFNHITNASQFERPSGNANA1331PIN.11-1NAKLPPG1EKRSMSRSSGRVYYFNHITNASQFERPSGNANA1341PIN.12-1NAKLPPGW1KRSMSRSSGRVYYFNHITNASQFERPSGNANA1351PIN.13-1NAKLPPGWE1RSMSRSSGRVYYFNHITNASQFERPSGNANA1361PIN.15-1NAKLPPGWEKRS1SRSSGRVYYFNHITNASQFERPSGNANA1371PIN.21-1NAKLPPGWEKRSMSRSSG1VYYFNHITNASQFERPSGNANA1381PIN.22-1NAKLPPGWEKRSMSRSSGR1YYFNHITNASQFERPSGNANA1391PIN.24-1NAKLPPGWEKRSMSRSSGRVY1FNHITNASQFERPSGNANA1401PIN.25-1NAKLPPGWEKRSMSRSSGRVYY1NHITNASQFERPSGNANA1411PIN.31-1NAKLPPGWEKRSMSRSSGRVYYFNHITN1SQFERPSGNANA1421PIN.33-1NAKLPPGWEKRSMSRSSGRVYYFNHITNAS1FERPSGNANA1431PIN.35-1NAKLPPGWEKRSMSRSSGRVYYFNHITNASQF1RPSGNANA1441PIN.14-ϵNAKLPPGWEKϵMSRSSGRVYYFNHITNASQFERPSGNANA1451PIN.16-ϵNAKLPPGWEKRMϵRSSGRVYYFNHITNASQFERPSGNANA1461PIN.17-ϵNAKLPPGWEKRMSϵSSGRVYYFNHITNASQFERPSGNANA1471PIN.18-ϵNAKLPPGWEKRMSRϵSGRVYYFNHITNASQFERPSGNANA1481PIN.19-ϵNAKLPPGWEKRMSRSϵGRVYYFNHITNASQFERPSGNANA1491PIN.21-ϵNAKLPPGWEKRMSRSSGϵVYYFNHITNASQFERPSGNANA1501PIN.23-ϵNAKLPPGWEKRMSRSSGRVϵYFNHITNASQFERPSGNANA1511PIN.26-ϵNAKLPPGWEKRMSRSSGRVYYFϵHITNASQFERPSGNANA1521PIN.27-ϵNAKLPPGWEKRMSRSSGRVYYFNϵITNASQFERPSGNANA1531PIN.28-ϵNAKLPPGWEKRMSRSSGRVYYFNHϵTNASQFERPSGNANA1541PIN.29-ϵNAKLPPGWEKRMSRSSGRVYYFNHIϵNASQFERPSGNANA1551PIN.30-ϵNAKLPPGWEKRMSRSSGRVYYFNHITϵASQFERPSGNANA1561PIN.32-ϵNAKLPPGWEKRMSRSSGRVYYFNHITNAϵQFERPSGNANA1571PIN.14-λNAKLPPGWEKλMSRSSGRVYYFNHITNASQFERPSGNANA1581PIN.16-λNAKLPPGWEKRMλRSSGRVYYFNHITNASQFERPSGNANA1591PIN.17-λNAKLPPGWEKRMSλSSGRVYYFNHITNASQFERPSGNANA1601PIN.18-λNAKLPPGWEKRMSRλSGRVYYFNHITNASQFERPSGNANA1611PIN.19-λNAKLPPGWEKRMSRSλGRVYYFNHITNASQFERPSGNANA1621PIN.21-λNAKLPPGWEKRMSRSSGλVYYFNHITNASQFERPSGNANA1631PIN.23-λNAKLPPGWEKRMSRSSGRVλYFNHITNASQFERPSGNANA1641PIN.26-λNAKLPPGWEKRMSRSSGRVYYFλHITNASQFERPSGNANA1651PIN.27-λNAKLPPGWEKRMSRSSGRVYYFNλITNASQFERPSGNANA1661PIN.28-λNAKLPPGWEKRMSRSSGRVYYFNHλTNASQFERPSGNANA1671PIN.29-λNAKLPPGWEKRMSRSSGRVYYFNHIλNASQFERPSGNANA1681PIN.30-λNAKLPPGWEKRMSRSSGRVYYFNHITλASQFERPSGNANAAnhang C. Konjugate2691691PIN.32-λNAKLPPGWEKRMSRSSGRVYYFNHITNAλQFERPSGNANA1701PIN.14-ψNAKLPPGWEKψMSRSSGRVYYFNHITNASQFERPSGNANA1711PIN.16-ψNAKLPPGWEKRMψRSSGRVYYFNHITNASQFERPSGNANA1721PIN.17-ψNAKLPPGWEKRMSψSSGRVYYFNHITNASQFERPSGNANA1731PIN.18-ψNAKLPPGWEKRMSRψSGRVYYFNHITNASQFERPSGNANA1741PIN.19-ψNAKLPPGWEKRMSRSψGRVYYFNHITNASQFERPSGNANA1751PIN.21-ψNAKLPPGWEKRMSRSSGψVYYFNHITNASQFERPSGNANA1761PIN.23-ψNAKLPPGWEKRMSRSSGRVψYFNHITNASQFERPSGNANA1771PIN.26-ψNAKLPPGWEKRMSRSSGRVYYFψHITNASQFERPSGNANA1781PIN.27-ψNAKLPPGWEKRMSRSSGRVYYFNψITNASQFERPSGNANA1791PIN.28-ψNAKLPPGWEKRMSRSSGRVYYFNHψTNASQFERPSGNANA1801PIN.29-ψNAKLPPGWEKRMSRSSGRVYYFNHIψNASQFERPSGNANA1811PIN.30-ψNAKLPPGWEKRMSRSSGRVYYFNHITψASQFERPSGNANA1821PIN.32-ψNAKLPPGWEKRMSRSSGRVYYFNHITNAψQFERPSGNANA1832KZJ.7-4NAVDNKFN4EMRNAYWEIALLPNLNNQQKRAFIRSLYDDPSQSANLLAEAKKLNDAQAPKNANA1842KZJ.8-4NAVDNKFNK4MRNAYWEIALLPNLNNQQKRAFIRSLYDDPSQSANLLAEAKKLNDAQAPKNANA1852KZJ.15-4NAVDNKFNKEMRNAYW4IALLPNLNNQQKRAFIRSLYDDPSQSANLLAEAKKLNDAQAPKNANA1862KZJ.21-4NAVDNKFNKEMRNAYWEIALLP4LNNQQKRAFIRSLYDDPSQSANLLAEAKKLNDAQAPKNANA1872KZJ.23-4NAVDNKFNKEMRNAYWEIALLPNL4NQQKRAFIRSLYDDPSQSANLLAEAKKLNDAQAPKNANA1882KZJ.25-4NAVDNKFNKEMRNAYWEIALLPNLNN4QKRAFIRSLYDDPSQSANLLAEAKKLNDAQAPKNANA1892KZJ.29-4NAVDNKFNKEMRNAYWEIALLPNLNNQQKR4FIRSLYDDPSQSANLLAEAKKLNDAQAPKNANA1902KZJ.33-4NAVDNKFNKEMRNAYWEIALLPNLNNQQKRAFIR4LYDDPSQSANLLAEAKKLNDAQAPKNANA1912KZJ.39-4NAVDNKFNKEMRNAYWEIALLPNLNNQQKRAFIRSLYDDP4QSANLLAEAKKLNDAQAPKNANA1922KZJ.43-4NAVDNKFNKEMRNAYWEIALLPNLNNQQKRAFIRSLYDDPSQSA4LLAEAKKLNDAQAPKNANA1932KZJ.46-4NAVDNKFNKEMRNAYWEIALLPNLNNQQKRAFIRSLYDDPSQSANLL4EAKKLNDAQAPKNANA1942KZJ.47-4NAVDNKFNKEMRNAYWEIALLPNLNNQQKRAFIRSLYDDPSQSANLLA4AKKLNDAQAPKNANA1952KZJ.49-4NAVDNKFNKEMRNAYWEIALLPNLNNQQKRAFIRSLYDDPSQSANLLAEA4KLNDAQAPKNANA1962KZJ.50-4NAVDNKFNKEMRNAYWEIALLPNLNNQQKRAFIRSLYDDPSQSANLLAEAK4LNDAQAPKNANA1972KZJ.53-4NAVDNKFNKEMRNAYWEIALLPNLNNQQKRAFIRSLYDDPSQSANLLAEAKKLN4AQAPKNANA1982KZJ.7-φNAVDNKFNφEMRNAYWEIALLPNLNNQQKRAFIRSLYDDPSQSANLLAEAKKLNDAQAPKNANA1992KZJ.8-φNAVDNKFNKφMRNAYWEIALLPNLNNQQKRAFIRSLYDDPSQSANLLAEAKKLNDAQAPKNANA2002KZJ.15-φNAVDNKFNKEMRNAYWφIALLPNLNNQQKRAFIRSLYDDPSQSANLLAEAKKLNDAQAPKNANA2012KZJ.21-φNAVDNKFNKEMRNAYWEIALLPφLNNQQKRAFIRSLYDDPSQSANLLAEAKKLNDAQAPKNANA2022KZJ.23-φNAVDNKFNKEMRNAYWEIALLPNLφNQQKRAFIRSLYDDPSQSANLLAEAKKLNDAQAPKNANA2032KZJ.25-φNAVDNKFNKEMRNAYWEIALLPNLNNφQKRAFIRSLYDDPSQSANLLAEAKKLNDAQAPKNANA2042KZJ.29-φNAVDNKFNKEMRNAYWEIALLPNLNNQQKRφFIRSLYDDPSQSANLLAEAKKLNDAQAPKNANA270Anhang C. Konjugate2052KZJ.33-φNAVDNKFNKEMRNAYWEIALLPNLNNQQKRAFIRφLYDDPSQSANLLAEAKKLNDAQAPKNANA2062KZJ.39-φNAVDNKFNKEMRNAYWEIALLPNLNNQQKRAFIRSLYDDPφQSANLLAEAKKLNDAQAPKNANA2072KZJ.43-φNAVDNKFNKEMRNAYWEIALLPNLNNQQKRAFIRSLYDDPSQSAφLLAEAKKLNDAQAPKNANA2082KZJ.46-φNAVDNKFNKEMRNAYWEIALLPNLNNQQKRAFIRSLYDDPSQSANLLφEAKKLNDAQAPKNANA2092KZJ.47-φNAVDNKFNKEMRNAYWEIALLPNLNNQQKRAFIRSLYDDPSQSANLLAφAKKLNDAQAPKNANA2102KZJ.49-φNAVDNKFNKEMRNAYWEIALLPNLNNQQKRAFIRSLYDDPSQSANLLAEAφKLNDAQAPKNANA2112KZJ.50-φNAVDNKFNKEMRNAYWEIALLPNLNNQQKRAFIRSLYDDPSQSANLLAEAKφLNDAQAPKNANA2122KZJ.53-φNAVDNKFNKEMRNAYWEIALLPNLNNQQKRAFIRSLYDDPSQSANLLAEAKKLNφAQAPKNANA2132KZJ.7-γNAVDNKFNγEMRNAYWEIALLPNLNNQQKRAFIRSLYDDPSQSANLLAEAKKLNDAQAPKNANA2142KZJ.8-γNAVDNKFNKγMRNAYWEIALLPNLNNQQKRAFIRSLYDDPSQSANLLAEAKKLNDAQAPKNANA2152KZJ.15-γNAVDNKFNKEMRNAYWγIALLPNLNNQQKRAFIRSLYDDPSQSANLLAEAKKLNDAQAPKNANA2162KZJ.21-γNAVDNKFNKEMRNAYWEIALLPγLNNQQKRAFIRSLYDDPSQSANLLAEAKKLNDAQAPKNANA2172KZJ.23-γNAVDNKFNKEMRNAYWEIALLPNLγNQQKRAFIRSLYDDPSQSANLLAEAKKLNDAQAPKNANA2182KZJ.25-γNAVDNKFNKEMRNAYWEIALLPNLNNγQKRAFIRSLYDDPSQSANLLAEAKKLNDAQAPKNANA2192KZJ.29-γNAVDNKFNKEMRNAYWEIALLPNLNNQQKRγFIRSLYDDPSQSANLLAEAKKLNDAQAPKNANA2202KZJ.33-γNAVDNKFNKEMRNAYWEIALLPNLNNQQKRAFIRγLYDDPSQSANLLAEAKKLNDAQAPKNANA2212KZJ.39-γNAVDNKFNKEMRNAYWEIALLPNLNNQQKRAFIRSLYDDPγQSANLLAEAKKLNDAQAPKNANA2222KZJ.43-γNAVDNKFNKEMRNAYWEIALLPNLNNQQKRAFIRSLYDDPSQSAγLLAEAKKLNDAQAPKNANA2232KZJ.46-γNAVDNKFNKEMRNAYWEIALLPNLNNQQKRAFIRSLYDDPSQSANLLγEAKKLNDAQAPKNANA2242KZJ.47-γNAVDNKFNKEMRNAYWEIALLPNLNNQQKRAFIRSLYDDPSQSANLLAγAKKLNDAQAPKNANA2252KZJ.49-γNAVDNKFNKEMRNAYWEIALLPNLNNQQKRAFIRSLYDDPSQSANLLAEAγKLNDAQAPKNANA2262KZJ.50-γNAVDNKFNKEMRNAYWEIALLPNLNNQQKRAFIRSLYDDPSQSANLLAEAKγLNDAQAPKNANA2272KZJ.53-γNAVDNKFNKEMRNAYWEIALLPNLNNQQKRAFIRSLYDDPSQSANLLAEAKKLNγAQAPKNANA2282KZJ.7-ωNAVDNKFNωEMRNAYWEIALLPNLNNQQKRAFIRSLYDDPSQSANLLAEAKKLNDAQAPKNANA2292KZJ.8-ωNAVDNKFNKωMRNAYWEIALLPNLNNQQKRAFIRSLYDDPSQSANLLAEAKKLNDAQAPKNANA2302KZJ.15-ωNAVDNKFNKEMRNAYWωIALLPNLNNQQKRAFIRSLYDDPSQSANLLAEAKKLNDAQAPKNANA2312KZJ.21-ωNAVDNKFNKEMRNAYWEIALLPωLNNQQKRAFIRSLYDDPSQSANLLAEAKKLNDAQAPKNANA2322KZJ.23-ωNAVDNKFNKEMRNAYWEIALLPNLωNQQKRAFIRSLYDDPSQSANLLAEAKKLNDAQAPKNANA2332KZJ.25-ωNAVDNKFNKEMRNAYWEIALLPNLNNωQKRAFIRSLYDDPSQSANLLAEAKKLNDAQAPKNANA2342KZJ.29-ωNAVDNKFNKEMRNAYWEIALLPNLNNQQKRωFIRSLYDDPSQSANLLAEAKKLNDAQAPKNANA2352KZJ.33-ωNAVDNKFNKEMRNAYWEIALLPNLNNQQKRAFIRωLYDDPSQSANLLAEAKKLNDAQAPKNANAAnhang C. Konjugate2712362KZJ.39-ωNAVDNKFNKEMRNAYWEIALLPNLNNQQKRAFIRSLYDDPωQSANLLAEAKKLNDAQAPKNANA2372KZJ.43-ωNAVDNKFNKEMRNAYWEIALLPNLNNQQKRAFIRSLYDDPSQSAωLLAEAKKLNDAQAPKNANA2382KZJ.46-ωNAVDNKFNKEMRNAYWEIALLPNLNNQQKRAFIRSLYDDPSQSANLLωEAKKLNDAQAPKNANA2392KZJ.47-ωNAVDNKFNKEMRNAYWEIALLPNLNNQQKRAFIRSLYDDPSQSANLLAωAKKLNDAQAPKNANA2402KZJ.49-ωNAVDNKFNKEMRNAYWEIALLPNLNNQQKRAFIRSLYDDPSQSANLLAEAωKLNDAQAPKNANA2412KZJ.50-ωNAVDNKFNKEMRNAYWEIALLPNLNNQQKRAFIRSLYDDPSQSANLLAEAKωLNDAQAPKNANA2422KZJ.53-ωNAVDNKFNKEMRNAYWEIALLPNLNNQQKRAFIRSLYDDPSQSANLLAEAKKLNωAQAPKNANA 273D.VorhersagemodelleD.1Berechnete DeskriptorenTabelle D.1: Übersicht über die untersuchten Deskriptoren. In der ersten Spalte ist derjeweilige bei der Modellgenerierung verwendete Name dargestellt (bei zwei verschiedenen Na-men wurde der jeweilige Deskriptor im zweiten Modellansatz umbenannt). Die zweite Spaltestellt jeweils eine Beschreibung des Deskriptors bereit. Die Variablen wurden für alle einzelnenAminosäuren (Rs = residues), für die unterschiedlichen Proteinbereiche (Rg = regional) oderauch das gesamte Protein (Pr = protein) berechnet. Variablen, die für die fünf unterschiedli-chen Regionen kalkuliert wurden, wurde der Suffix 1 für β1, 2 für β2, 3 für β3, 4 für Schleife 1und 5 für Schleife 2 zugeschrieben (z. B. beschreibt sscontent1 den durchschnittlichen Sekun-därstrukturanteil für alle Aminosäuren des ersten β-Faltblatts). In ähnlicher Weise erhaltenDeskriptoren, die für alle einzelnen Aminosäuren berechnet wurden, den jeweiligen Suffix 6– 39 (z. B. beschreibt int14 die Interaktionsenergie des Polymers mit R14). Im Falle des An-satzes 1 wurden, wo möglich Medianwerte berechnet, im Falle von Ansatz 2 Mittelwerte. Dieletzte Spalte gibt an, bei welchem Modellansatz der jeweilige Deskriptor verwendet wurde (1,2 oder 1+2).NameBeschreibungRsRg PrAbbhbondsAnzahl an Proteinrückgrat-HB pro Momentaufnahme (Cutoff: 3.5Å und 150°).X1+2bbwathbondsAnzahl an Proteinrückgrat-HB zu Wassermolekülen pro Moment-aufnahme.X1+2contactsAnzahl an nativen Kontakten des Proteinrückgrats (Cutoff: 4.5Å).X1+2contacts2Anzahl an nativen Kontakten aller Proteinatome (Cutoff: 4.5 Å).X2crigidityRigide-Kontakte-Index (CNA-Parameter) für jede Aminosäure X(crigidityX) und Region X (crigiditymeanX), sowie für das ganzeProtein (crigiditymean).XXX1+2entropysideKonfigurationsentropie, berechnet nach den Seitenkettenorderpa-rametern gemäß Brüschweiler (kein Wert für nicht proteinogeneAminosäuren kalkulierbar) [376].XXX1glycinflipProzent an Momentaufnahmen, in denen G20 innerhalb der erstenSchleife einen positiven φ-Wert zeigt, was einer Drehung des Pro-teinrückgrats entspricht. Dieses Phänomen wurde in den erstenSimulationen beobachtet (kein Median / Durchschnitt berechen-bar).X1+2intInteraktionsenergie der Polymeratome zu jeder Aminosäure.X1+2intohInteraktionsenergie der Polymeratome zu jedem Serin, Threoninund Tyrosin.X1+2274Anhang D. VorhersagemodelleintXInteraktionsenergie der Polymeratome zu bestimmten proteinoge-nen Aminosäure (X = Gly, Ser, Tyr, Phe, Asn, Glu oder Pro).X1+2intX_YInteraktionsenergie jeder Aminosäure X zu jeder anderen Amino-säure Y im Protein.X1+2lifetime / lifeavgDurchschnittliche Überlebenszeit der HB zwischen Proteinrück-gratatomen und Wasser.X1+2lifenumTotale Überlebenszeit der HB zwischen Proteinrückgratatomenund Wasser.X2lifepegavgDurchschnittliche Überlebenszeit der HB des Polymers zum Sol-vens.X2lifepegnumTotale Überlebenszeit der HB des Polymers zum Solvens.X2lifescavgDurchschnittliche Überlebenszeit der HB des Polymers zum Sol-vens.X2lifescnumTotale Überlebenszeit der HB des Polymers zum Solvens.X2molburProteinoberfläche, welche durch Polymeratome bedeckt ist.X1+2orderbackProteinrückgratamid-Orderparameter. Der Durchschnittswert füralle 10 ns-Trajektorienblöcke wurde hierzu berechnet, angelehntan die Methoden aus [381].XXX1+2ordersideSeitenkettenorderparameter(keineWertefürnichtprotein-ogene Aminosäuren). Der Durchschnittswert für alle 10 ns-Trajektorienblöcke wurde hierzu berechnet.XXX1+2p1Perkolationsindex vom Typ 1 (CNA-Parameter) für jede Amino-säure X (p1_X) und Region X (p1mean_X), sowie für das ganzeProtein (p1mean).XXX1+2pegangleWinkel zwischen dem Massenzentrum des Protein, dem Massen-zentrum des Aminosäurebereichs der Konjugationsstelle und derPolymerkette (beschrieben in [35]) als ein Indikator für die Aus-richtung des Polymers zur Proteinoberfläche.X1+2pegdcomDistanz zwischen dem Massenzentrum des Proteins und dem Mas-senzentrum der konjugierten Aminosäure (beschrieben in [35]).X1+2pegohhbondsAnzahl an HB der Polymerkette zu Hydroxylseitenketten (S, Tund Y).X2pegwathbondsAnzahl an HB der Polymerkette zu Wassermolekülen pro Moment-aufnahme.X1+2pg4distAnsatz 1: Minimaldistanz des letzen Polymer-Sauerstoffatoms zujeglicher oder einer bestimmten Aminosäure. Ansatz 2: Maximal-distanz jeglicher Polymeratome zu Proteinatomen.XX1+2pgbrgAnzahl an HB pro Momentaufnahme, die zwischen Polymerato-men über überbrückende Wassermoleküle gebildet wurden. DiesesPhänomen wurde in den ersten Simulationen beobachtet.X1+2phiφ-Winkel aller Aminosäuren X.X1+2price1Winkel θP rice in der Kristallstruktur, gemäß [100], zwischem demCα-Atom der konjugierten Aminosäure, dem Massenzentrum derentsprechenden Seitenkette (vor Konjugation) und dem nächst-gelegenen Serin, Threonin oder Tyrosin (kein Durchschnitt bere-chenbar).X1+2price2Winkel analog zu price1, jedoch handelt es sich um den Durch-schnitt aus der Simulation der entsprechenden unkonjugierten Mu-tante.X1+2D.1. Berechnete Deskriptoren275price3Winkel analog zu price2, jedoch handelt es sich um den Durch-schnitt aus der Simulation der entsprechenden konjugierten Mu-tante.X1+2price4Winkel analog zu price3, jedoch wird bei der Berechnung des Mas-senzentrums der Seitenkette der Polymerbereich ignoriert.X1+2price1apolWie price1 für apolare Aminosäureseitenketten. Nur im Falle derModelle A2v2-R1 und A2v2-C1 genutzt.X(2)price2apolWie price2 für apolare Aminosäureseitenketten. Nur im Falle derModelle A2v2-R1 und A2v2-C1 genutzt.X(2)price3apolWie price3 für apolare Aminosäureseitenketten. Nur im Falle derModelle A2v2-R1 und A2v2-C1 genutzt.X(2)price4apolWie price4 für apolare Aminosäureseitenketten. Nur im Falle derModelle A2v2-R1 und A2v2-C1 genutzt.X(2)price1aromWie price1 für aromatische Aminosäureseitenketten. Nur im Falleder Modelle A2v2-R1 und A2v2-C1 genutzt.X(2)price2aromWie price2 für aromatische Aminosäureseitenketten. Nur im Falleder Modelle A2v2-R1 und A2v2-C1 genutzt.X(2)price3aromWie price3 für aromatische Aminosäureseitenketten. Nur im Falleder Modelle A2v2-R1 und A2v2-C1 genutzt.X(2)price4aromWie price4 für aromatische Aminosäureseitenketten. Nur im Falleder Modelle A2v2-R1 und A2v2-C1 genutzt.X(2)price1polWie price1 für polare Aminosäureseitenketten. Nur im Falle derModelle A2v2-R1 und A2v2-C1 genutzt.X(2)price2polWie price2 für polare Aminosäureseitenketten. Nur im Falle derModelle A2v2-R1 und A2v2-C1 genutzt.X(2)price3polWie price3 für polare Aminosäureseitenketten. Nur im Falle derModelle A2v2-R1 und A2v2-C1 genutzt.X(2)price4polWie price4 für polare Aminosäureseitenketten. Nur im Falle derModelle A2v2-R1 und A2v2-C1 genutzt.X(2)protpeghbAnzahl an HB des Polymers pro Momentaufnahme zu jeder ein-zelnen Aminosäure X.X1+2protpeghbondsDurchschnittliche Anzahl an HB des Polymers pro Momentauf-nahme zum gesamten Protein.X1+2protprothbondsDurchschnittliche Anzahl an Protein-Protein-HB pro Momentauf-nahme.X1+2protwathbondsDurchschnittliche Anzahl an Protein-Wasser-HB pro Momentauf-nahme.X1+2psiψ-Winkel aller Aminosäuren X.X1+2radgyrpeg / rg-pegGyrationsradius der Polymer-Schweratome.X1+2rdf_YRDF-Wert von Wassermolekülen um Aminosäuren herum in ei-nem Bereich von 2-4 Å, 4-6 Åund 6-8 Å(benannt als Y = 200,400, 600). Im Ansatz 2 wurden dagegen die Bereiche 2-3 Å, 3-4Å, 4-5 Å, 5-6 Å, 6-7 Å, 7-8 Åund 9-10 Å berechnet.XXX1+2rdfapol_YRDF-Wert von Wassermolekülen um apolare Aminosäuren herumin einem Bereich von 2-4 Å, 4-6 Åund 6-8 Å(benannt als Y = 200,400, 600). Im Ansatz 2 wurden dagegen die Bereiche 2-3 Å, 3-4Å, 4-5 Å, 5-6 Å, 6-7 Å, 7-8 Åund 9-10 Å berechnet.X1+2276Anhang D. Vorhersagemodellerdfoh_YRDF-Wert von Wassermolekülen um die Aminosäuren Serin, Th-reonin und Tyrosin herum in einem Bereich von 2-4 Å, 4-6 Åund6-8 Å(benannt als Y = 200, 400, 600). Im Ansatz 2 wurden dage-gen die Bereiche 2-3 Å, 3-4 Å, 4-5 Å, 5-6 Å, 6-7 Å, 7-8 Åund 9-10Å berechnet.X1+2rdfpeg_YRDF-Wert von Wassermolekülen um die Polymeratome herum ineinem Bereich von 2-4 Å, 4-6 Åund 6-8 Å(benannt als Y = 200,400, 600). Im Ansatz 2 wurden dagegen die Bereiche 2-3 Å, 3-4Å, 4-5 Å, 5-6 Å, 6-7 Å, 7-8 Åund 9-10 Å berechnet.X1+2rdfpegtores_YRDF-Wert von Polymeratomen um das Protein herum in einemBereich von 2-4 Å, 4-6 Åund 6-8 Å(benannt als Y = 200, 400,600). Im Ansatz 2 wurden dagegen die Bereiche 2-3 Å, 3-4 Å, 4-5Å, 5-6 Å, 6-7 Å, 7-8 Åund 9-10 Å berechnet.XX1+2rdfpol_YRDF-Wert von Wassermolekülen um polare Aminosäuren herumin einem Bereich von 2-4 Å, 4-6 Åund 6-8 Å(benannt als Y = 200,400, 600). Im Ansatz 2 wurden dagegen die Bereiche 2-3 Å, 3-4Å, 4-5 Å, 5-6 Å, 6-7 Å, 7-8 Åund 9-10 Å berechnet.X1+2rdfwopeg_YRDF-Wert von Wassermolekülen um das Protein (ohne Polymer)herum in einem Bereich von 2-4 Å, 4-6 Åund 6-8 Å(benannt als Y= 200, 400, 600). Im Ansatz 2 wurden dagegen die Bereiche 2-3Å, 3-4 Å, 4-5 Å, 5-6 Å, 6-7 Å, 7-8 Åund 9-10 Å berechnet.X1+2rdfX_YRDF-Wert von Wassermolekülen um bestimmte Aminosäuren her-um in einem Bereich von 2-4 Å, 4-6 Åund 6-8 Å(benannt als Y =200, 400, 600). Dies wurde für X = Gly, Ser, Tyr, Phe, Asn, Gluund Pro berechnet. Im Ansatz 2 wurden dagegen die Bereiche 2-3Å, 3-4 Å, 4-5 Å, 5-6 Å, 6-7 Å, 7-8 Åund 9-10 Å berechnet.X1+2rdfX_YRDF-Wert von Wassermolekülen um alle Aminosäuren X herum,in einem Bereich von 2-3 Å, 3-4 Å, 4-5 Å, 5-6 Å, 6-7 Å, 7-8 Åund9-10 Å.X2rgGyrationsradius der Protein- und Polymer-Schweratome.XXX1+2rgbbGyrationsradius des Proteinrückgrats.X1+2rgwopegGyrationsradius der Protein-Schweratome (ohne Polymer).XX1+2rhbAnzahl an HB jeder Aminosäure X zu Wassermolekülen pro Mo-mentaufnahme.X1+2rigidityRigiditätsindex (CNA-Parameter) für jede Aminosäure X (crigidi-tyX) und Region X (rigiditymeanX), sowie für das ganze Protein(rigiditymean).XXX1+2rmsdRMSD-Wert der Schweratome des Proteinrückgrats (Rg und Pr-Kalkulationen), bzw. der Schweratome der gesamten Aminosäure(Rs). Als Referenz diente die aufbereitete Kristallstruktur.XXX1+2rmsdpRMSD-Wert der Polymer-Schweratome (Startstruktur als Refe-renz).X1rmsfMassengewichtete RMSF-Werte der Schweratome des Protein-rückgrats pro Aminosäure (Rg und Pr-Kalkulationen), bzw. derSchweratome der gesamten Aminosäure (Rs). Als Referenz dientedie durchschnittliche Struktur der Simulation. Kein Durchschnittkalkulierbar.XXX1+2rmsfpMassengewichtete RMSF-Werte der Schweratome des Polymers.Als Referenz diente die durchschnittliche Struktur der Simulation.Kein Durchschnitt kalkulierbar.X1+2shellAnzahl an Wassermolekülen in der ersten Hydrathülle (<3 Å),Polymeratome eingeschlossen.XXX1+2D.1. Berechnete Deskriptoren277shellapolAnzahl an Wassermolekülen in der ersten Hydrathülle (<3 Å) vonapolaren Aminosäuren.X1+2shellapoloAnzahl an Wassermolekülen in der ersten und zweiten Hydrathülle(<5 Å) von apolaren Aminosäuren.X1+2shelloAnzahl an Wassermolekülen in der ersten und zweiten Hydrathülle(<5 Å), Polymeratome eingeschlossen.XXX1+2shellohAnzahl an Wassermolekülen in der ersten Hydrathülle (<3 Å) derAminosäuren Serin, Threonin und Tyrosin.X1+2shellohoAnzahl an Wassermolekülen in der ersten und zweiten Hydrathülle(<5 Å) der Aminosäuren Serin, Threonin und Tyrosin.X1+2shellpegAnzahl an Wassermolekülen in der ersten Hydrathülle (<3 Å) derPolymeratome.X1+2shellpegapolAnzahl an Polymeratomen in der ersten Hydrathülle (<3 Å) apo-larer Proteinatome.X2shellpegoAnzahl an Wassermolekülen in der zweiten Hydrathülle (<5 Å)der Polymeratome.X1+2shellpegpolAnzahl an Polymeratomen in der ersten Hydrathülle (<3 Å) po-larer Proteinatome.X2shellpegohAnzahl an Polymeratomen in der ersten Hydrathülle (<3 Å) vonHydroxylseitenketten (S, T und Y).X2shellpegapoloAnzahl an Polymeratomen in der zweiten Hydrathülle (<5 Å) apo-larer Proteinatome.X2shellpegpoloAnzahl an Polymeratomen in der zweiten Hydrathülle (<5 Å) po-larer Proteinatome.X2shellpegohoAnzahl an Polymeratomen in der zweiten Hydrathülle (<5 Å) vonHydroxylseitenketten (S, T und Y).X2shellpolAnzahl an Wassermolekülen in der ersten Hydrathülle (<3 Å) vonpolaren Aminosäuren.X1+2shellpoloAnzahl an Wassermolekülen in der ersten und zweiten Hydrathülle(<5 Å) von polaren Aminosäuren.X1+2shellwopegAnzahl an Wassermolekülen in der ersten Hydrathülle (<3 Å),Polymeratome ausgenommen.X1+2shellwopegoAnzahl an Wassermolekülen in der ersten Hydrathülle (<5 Å),Polymeratome ausgenommen.X1+2srigiditySeitenkettenrigiditätsindex (CNA-Parameter) für jede Aminosäu-re X (srigidityX) und Region X (srigiditymedianX), sowie für dasganze Protein (srigiditymean).XXX1+2sscontentDurchschnittlicher Sekundärstrukturanteil des Proteins.XXX1+2surfSolvenszugängliche Oberfläche.XXX1+2surfapolSolvenszugängliche Oberfläche apolarer Atome.XXX1+2surfaprVerhältnis zwischen SASA apolarer und polarer Atome.X1+2surfpegSolvenszugängliche Oberfläche der Polymeratome.X1+2surfpolSolvenszugängliche Oberfläche polarer Atome.XXX1+2tau1Kontinuierliche HB-Überlebenszeit der Seitenketten zu Wasser-molekülen, berechnet mit dem mdanalysis-Modul.X2tau2Intermittierende HB-Überlebenszeit der Seitenketten zu Wasser-molekülen, berechnet mit dem mdanalysis-Modul.X2278Anhang D. VorhersagemodelletotalintInteraktionsenergie zwischen Polymer- und Proteinatomen.X1+2trphcce1 / trphc-cex1Phasenübergangstemperatur (CNA-Parameter), basierend auf derClusterkonfigurationsentropie vom Typ 1.X1+2trphcce2 / trphc-cex2Phasenübergangstemperatur (CNA-Parameter), basierend auf derClusterkonfigurationsentropie vom Typ 2.X1+2trphcces / trphc-cexsPhasenübergangstemperatur (CNA-Parameter), basierend auf derClusterkonfigurationsentropie vom Typ 2 mit doppelt sigmoida-lem Fit.X1+2trphrop1/trphropx1Phasenübergangstemperatur(CNA-Parameter),basierendaufdem Rigiditätsorderparameter vom Typ 1.X1+2trphrop2/trphropx2Phasenübergangstemperatur(CNA-Parameter),basierendaufdem Rigiditätsorderparameter vom Typ 2.X1+2D.2Ansatz 1D.2.1ModelltypenTabelle D.2: Übersicht über die getesteten Modelltypen im ersten QSPR-Ansatz (Kap.3.1.6.2) gemäß der caret-Webseite [382]. Der Name der Modellart und eine kurze Beschreibungsind aufgeführt. Zusätzlich sind die zugrundeliegenden R-Pakete aufgelistet. Die letzte Spaltegibt an, ob hierbei Regressions- (R) und / oder Klassifikationsmodelle (C) generiert wurden.NameBeschreibungR-PaketTypavNNetModel averaged neural networknnetCC5.0RulesSingle C5.0 rulesetC50CC5.0TreeSingle C5.0 treeC50CcubistCubist modelcubistRgaussprLinearGaussian process modelkernlabRgaussprPolyGaussian process model with polynomial kernelkernlabRgplsGeneralized partial least squares modelgplsCmlpMulti-layer perceptronRSNNSCmlpMLMulti-layer perceptron with multiple layersRSNNSCmlpWeightDecayMulti-layer perceptron using weight decayRSNNSCmlpWeightDe-cayMLMulti-layer perceptron using weight decay, with mul-tiple layersRSNNSCnnetNeural networknnetC, RORFlogOblique random forest (using logistic regression)obliqueRFCORFplsOblique random forest (using partial least squares)obliqueRFCORFridgeOblique random forest (linear combinations createdusing L2 regularization)obliqueRFCORFsvmOblique random forest (using support vector machi-nes)obliqueRFCD.2. Ansatz 1279parRFParallel random foreste1071,randomFo-rest,foreach,im-portC, RpcaNNetNeural networks with feature extractionnnetCpcrPrincipal component analysisplsRplsPartial least squaresplsRregLogisticRegularized logistic regressionLiblineaRCrfRandom forestrfC, RridgeRidge regressionelasticnetRrqlassoQuantile regression with LASSO penaltyrqPenRrqncNon-convex penalized quantile regressionrqPenRRRFglobalRegularized random forestRRFRrvmRadialRelevance vector machines with radial basis functionkernelkernlabRsvmLinear3Support vector machines with linear kernelLiblineaRRsvmPolySupport vector machines with polynomial kernelkernlabRsvmRadialSupport vector machines with radial basis functionkernelkernlabRsvmRadialSigmaSupport vector machines with radial basis functionkernel (tunes over cost and sigma)kernlabRxgbLineareXtreme gradient boosting (linear learners)xgboostRxgbTreeeXtreme gradient boosting (tree learners)xgboost, plyrR280Anhang D. VorhersagemodelleD.2.2ModellleistungenTabelle D.3: Validierungsparameter für die besten Modelle aus den Kategorien A bis Dim ersten Modellgenerierungsansatz. Der Modelltyp ist angegeben (vgl. Tab. D.2), dazu zujedem Datenset RMSE, MAE, CCC und R2 (für Regressionsmodelle), bzw. AUROC und CCR(Klassifikationsmodelle). Die jeweilige Anzahl genutzter Deskriptoren (Des.) ist in der letztenSpalte angegeben. Die Modelle sind auf Basis der RMSEHS- bzw. CCRHS-Werte geordnet.Die Kategorien A und B weisen ausschließlich Modelle auf, die, in Relation zum Datensatz,zu viele Deskriptoren verwenden und damit ein komplettes Overfitting aufweisen.Pin 1-spezifische Regressionsmodelle (Kategorie A)#ArtRMSETSMAETSCCCTSR2TSRMSEHSMAEHSCCCHSR2HSDes.A1xgbLinear0.020.001.001.000.250.190.720.59166A2xgbLinear0.020.001.001.000.260.210.700.59526A3cubist0.040.031.001.000.280.230.690.5276Pin 1-spezifische Klassifikationsmodelle (Kategorie B)#ArtAUROCTSCCRTSAUROCHSCCRHSDes.B1pcaNNet0.990.991.000.8749B2avNNet1.001.001.000.87190B3ORFsvm1.001.001.000.8049Pin 1-unspezifische Regressionsmodelle (Kategorie C)#ArtRMSETSMAETSCCCTSR2TSRMSEHSMAEHSCCCHSR2HSDes.C1xgbTree0.140.100.950.920.240.200.750.6311C2svmLinear30.360.260.550.400.250.200.760.6112C3gaussprLinear0.340.280.620.460.260.200.740.5612Pin 1-unspezifische Klassifikationsmodelle (Kategorie D)#ArtAUROCTSCCRTSAUROCHSCCRHSDes.D1nnet0.950.930.910.808D2regLogistic0.760.710.910.738D3mlpWeightDecayML1.001.000.940.6723D.2. Ansatz 1281D.2.3ADAN-AnalysenTabelle D.4: Kategorisierung der HS-Konjugate in eine von sechs ADAN-Kategorien, dar-gestellt für die Modelle A1 bis A3, C1 bis C3 und E. Je höher die Kategorie, desto schlechterist das Modell zur Einschätzung der Faltungsenergie für das jeweilige HS-Konjugat geeignet(Kategorien 0 - 1 in grün = innerhalb der Applikationsdomäne, 2 - 3 in orange = unsichereVorhersage, 4 - 6 in rot = außerhalb der Applikationsdomäne).ModelleKonjugatA1A2A3C1C2C3E29-1312000032-1312000019(S32A)-1313000019-4312232216-9414000016-Z313344229-Z411000132-Z313000119-2424000332-2413000023(E12F:R14L)-Z313111319-J314222018(S16D:R17A)-1412000019-5312000316+19-131322201SRL.20-1---3332D.2.4Pin 1-unspezifische Deskriptorenprotpeghbonds, pgbrg, totalint, pg4dist, rmsdp, molbur, surfpeg, radgyrpeg, rdfpeg_200,price2, pegdcom, pegangle, protprothbondsratio, bbhbondsratio, bbwathbondsratio, life-timeratio, rmsdratio, shelloratio, shellwopegoratio, shellohoratio, shellpoloratio, shell-apoloratio,rdfoh_200ratio,rdfoh_400ratio,rdfoh_600ratio,rdfpol_200ratio,rdfpol_400ratio, rdfpol_600ratio, rdfapol_200ratio, rdfapol_400ratio, rdfapol_600ra-tio, entropysideratio, p1medianratio282Anhang D. VorhersagemodelleAbbildung D.1: RDF-Differenz-Plots einer Auswahl stark destabilisierter (linke Spalte) undstabilisierter Varianten (rechte Spalte).D.3. Ansatz 2283Abbildung D.2: Illustration des beobachteten Glycin-Flips. In grün ist die Startstrukturvon 27-1 dargestellt, während die Konformation in cyan die Umorientierung beschreibt.D.3Ansatz 2D.3.1Pin 1-unspezifische Deskriptorenprotpeghbonds, pgbrg, pegwathbonds, totalint, intoh, pg4dist, molbur, surfpeg, rmsfp, rg-peg, shellpeg, shellpego, shellpegapolo, shellpegpolo, rdfpeg_200, rdfpeg_300, rdfpeg_400,rdfpeg_500, rdfpeg_900, price1, price2, price3, price4, pegdcom, pegangle, lifepe-gnum, trphccex1, trphccex2, trphropx1, trphccexs, bbhbondsratio, rmsdratio, surfaprra-tio, rgratio, rgbbratio, rgwopegratio, shellohratio, shellpolratio, shellapolratio, shel-lohoratio, shellpoloratio, shellapoloratio, rdf_200ratio, rdf_400ratio, rdf_500ratio,rdf_600ratio, rdf_700ratio, rdf_900ratio, rdfwopeg_200ratio, rdfwopeg_400ratio, rdfwo-peg_500ratio, rdfwopeg_600ratio, rdfwopeg_700ratio, rdfwopeg_900ratio, rdfoh_200ra-tio, rdfoh_300ratio, rdfoh_400ratio, rdfoh_500ratio, rdfoh_600ratio, rdfoh_700ratio,rdfoh_800ratio, rdfoh_900ratio, rdfpol_200ratio, rdfpol_300ratio, rdfpol_400ratio,rdfpol_500ratio, rdfpol_600ratio, rdfpol_700ratio, rdfpol_800ratio, rdfpol_900ratio,rdfapol_200ratio, rdfapol_300ratio, rdfapol_400ratio, rdfapol_500ratio, rdfapol_600ra-tio, rdfapol_700ratio, rdfapol_800ratio, rdfapol_900ratio, orderbackratio, tau1ratio,tau2ratio284Anhang D. VorhersagemodelleD.3.2ModelltypenTabelle D.5: Übersicht über die getesteten Modelltypen im zweiten QSPR-Ansatz (Kap.3.1.6.3) gemäß der caret-Webseite [382]. Der Name der Modellart und eine kurze Beschreibungsind aufgeführt. Zusätzlich sind die zugrundeliegenden R-Pakete aufgelistet. Die letzte Spaltegibt an, ob hierbei Regressions- (R) oder Klassifikationsmodelle (C) generiert wurden.NameBeschreibungR-PaketTyplmLinear regressionBasis-R-PaketRglm (binomial)Generalized linear model (logistic regression)Basis-R-PaketCD.3.3ModellleistungenTabelle D.6: Validierungsparameter für die erstellten Regressionsmodelle des zweiten Mo-dellierungsansatzes für die verschiedenen Datensätze (DS). Werte für das Trainings- und dasHoldout-Set sind für das standardmäßige Interkorrelationslimit von 0.99 aufgelistet. Nachge-schaltete Filterungen des finalen Deskriptorsets mit einem zusätzlichen Limit von 0.75 führtenzu Modellen mit den Werten in der unteren Hälfte der Tabelle.Korrelationslimit: 0.99DSRMSETSCCCTSτB(TS)R2TSRMSEHSCCCHSτB(HS)R2HS10.210.850.610.730.590.580.350.3420.680.02-0.050.000.360.510.340.3430.320.710.530.550.810.300.450.1840.260.870.690.770.940.03-0.050.0050.200.650.460.480.390.720.670.9660.250.810.450.680.370.740.820.7670.280.760.590.620.650.270.200.0980.320.590.450.420.76-0.19-0.110.0690.240.830.530.710.580.200.090.04Korrelationslimit: 0.75DSRMSETSCCCTSτB(TS)R2TSRMSEHSCCCHSτB(HS)R2HS10.260.760.560.610.560.530.350.2920.680.02-0.050.000.360.510.340.3430.770.280.450.390.360.610.420.4440.260.870.690.770.940.03-0.050.0050.210.560.380.380.660.180.670.7960.340.580.320.410.400.630.710.6670.310.690.510.530.660.270.200.0980.320.590.450.420.76-0.19-0.110.0690.300.720.460.560.530.130.270.02D.3. Ansatz 2285Tabelle D.7: Validierungsparameter für die erstellten Klassifikationsmodelle des zweiten Mo-dellierungsansatzes für die verschiedenen Datensätze (DS). Werte für das Trainings- und dasHoldout-Set sind für das standardmäßige Interkorrelationslimit von 0.99 aufgelistet. Nachge-schaltete Filterungen des finalen Deskriptorsets mit einem zusätzlichen Limit von 0.75 führtenzu Modellen mit den Werten in der unteren Hälfte der Tabelle.Korrelationslimit: 0.99DSBAccTSAUCTSAUPRCTSBAccHSAUCHSAUPRCHS10.840.930.960.770.830.8120.800.810.900.600.670.7130.750.830.900.710.690.5940.850.930.930.460.670.4551.001.000.360.500.500.0060.850.940.970.480.600.4270.730.890.950.290.430.5781.001.000.080.830.830.0090.810.910.960.500.710.67Korrelationslimit: 0.75DSBAccTSAUCTSAUPRCTSBAccHSAUCHSAUPRCHS10.630.860.950.500.830.8020.560.730.880.500.540.5930.500.690.760.570.630.5340.820.930.940.460.420.3950.500.820.930.501.000.5060.770.930.970.400.600.4270.670.850.940.520.570.6181.001.000.080.830.830.0090.600.850.950.610.680.61286Anhang D. VorhersagemodelleD.3.4AnalysenAbbildung D.3: Univariate Regressionen weiterer Modelldeskriptoren von Modell A2-R6.D.3. Ansatz 2287Abbildung D.4: Univariate Regressionen der Modelldeskriptoren von A2-C1, analog zu Abb.D.3.288Anhang D. VorhersagemodelleTabelle D.8: Vorhersagen des Modells A2-R6 zu den Pin 1-Konjugaten der Polymeralterna-tiven (vgl. Kennzeichen in Tab. C.1).KonjugatVorhersage [kcal/mol]14-λ-0.1416-λ0.3917-λ0.1018-λ0.4519-λ0.8621-λ0.2323-λ-0.0526-λ0.9227-λ1.7028-λ1.1029-λ1.2430-λ0.3332-λ0.0014-ψ0.6716-ψ0.8717-ψ1.0518-ψ1.5519-ψ-1.2521-ψ1.4723-ψ0.0526-ψ1.1727-ψ0.8428-ψ2.0529-ψ1.1430-ψ1.2432-ψ0.6714-ϵ1.1816-ϵ0.7117-ϵ1.5818-ϵ1.1819-ϵ1.3221-ϵ1.0523-ϵ1.0726-ϵ0.9027-ϵ2.0828-ϵ1.3129-ϵ1.5630-ϵ1.0132-ϵ0.92molburrdf_500ratiordfpeg_500rdf_700ratiopgbrgshellpegoshellpolratiordf_400ratiordf_200ratiopegangleMittelwert LPG/PMeOx/PEtOx−KonjugateMittelwert PEG−Konjugate0.00.51.01.52.0Abbildung D.5: Verhältnis zwischen den Mittelwerten der Polymere LPG, PMeOx undPEtOx und dem Mittelwert der PEG-Konjugate des TS für die relevanten Deskriptoren desModells A2-R6.D.3. Ansatz 2289Tabelle D.9: Vorhersagen des Modells A2-C1 zu den Pin 1-Konjugaten der Polymeralterna-tiven (vgl. Kennzeichen in Tab. C.1). Dabei steht S für stabilisiert und N für nicht stabilisiert.In Klammern ist die Wahrscheinlichkeit aufgeführt, das jeweilige Konjugat der Klasse 1 (sta-bilisiert) zuzuordnen.KonjugatVorhersage14-λN (0.00)16-λN (0.00)17-λS (0.67)18-λN (0.00)19-λN (0.00)21-λN (0.09)23-λN (0.00)26-λN (0.00)27-λN (0.00)28-λN (0.00)29-λN (0.00)30-λN (0.03)32-λN (0.43)14-ψN (0.02)16-ψN (0.00)17-ψN (0.00)18-ψN (0.00)19-ψS (0.99)21-ψN (0.00)23-ψS (0.99)26-ψN (0.00)27-ψN (0.00)28-ψN (0.00)29-ψN (0.00)30-ψN (0.00)32-ψN (0.00)14-ϵN (0.08)16-ϵN (0.00)17-ϵN (0.00)18-ϵN (0.00)19-ϵN (0.00)21-ϵN (0.00)23-ϵN (0.00)26-ϵN (0.00)27-ϵN (0.00)28-ϵN (0.00)29-ϵN (0.00)30-ϵN (0.00)32-ϵN (0.00)pegwathbondsrdfpeg_500rdfpeg_400rdfwopeg_400ratiomolburrdfpeg_200rdfpeg_300Mittelwert LPG/PMeOx/PEtOx−KonjugateMittelwert PEG−Konjugate0.00.51.01.52.02.53.03.5Abbildung D.6: Verhältnis zwischen den Mittelwerten der Polymere LPG, PMeOx undPEtOx und dem Mittelwert der PEG-Konjugate des TS für die relevanten Deskriptoren desModells A2-C1. 291Literatur[1]I. S. Johnson, „Human Insulin from Recombinant DNA Technology“, Science,Jg. 219, Nr. 4585, S. 632–637, 1983.[2]D. S. Dimitrov, „Therapeutic Proteins“, in Therapeutic Proteins: Methods andProtocols, Ser. Methods in Molecular Biology, V. Voynov und J. A. Caravella,Bd. 899, Humana Press, 2012, S. 1–26.[3]M. Kesik‐Brodacka, „Progress in biopharmaceutical development“, Biotechno-logy and Applied Biochemistry, Jg. 65, Nr. 3, S. 306–322, 2018.[4]L. Urquhart, „Top companies and drugs by sales in 2019“, Nature ReviewsDrug Discovery, Jg. 19, Nr. 4, S. 228–228, 2020.[5]S. Mitragotri, P. A. Burke und R. Langer, „Overcoming the challenges in ad-ministering biopharmaceuticals: Formulation and delivery strategies“, NatureReviews Drug Discovery, Jg. 13, Nr. 9, S. 655–672, 2014.[6]M. Morishita und N. A. Peppas, „Is the oral route possible for peptide andprotein drug delivery?“, Drug Discovery Today, Jg. 11, Nr. 19-20, S. 905–910,2006.[7]P. J. Carter, „Introduction to current and future protein therapeutics: A pro-tein engineering perspective“, Experimental Cell Research, Jg. 317, Nr. 9,S. 1261–1269, 2011.[8]M. Kessler, D. Goldsmith und H. Schellekens, „Immunogenicity of biophar-maceuticals“, Nephrology Dialysis Transplantation, Jg. 21, S. v9–v12, suppl_52006.[9]B. Kaupbayeva und A. J. Russell, „Polymer-enhanced biomacromolecules“,Progress in Polymer Science, Jg. 101, S. 101 194, 2020.[10]Y.-H. Bi, Z.-Q. Duan, X.-Q. Li, Z.-Y. Wang und X.-R. Zhao, „IntroducingBiobased Ionic Liquids as the Nonaqueous Media for Enzymatic Synthesis ofPhosphatidylserine“, Journal of Agricultural and Food Chemistry, Jg. 63, Nr. 5,S. 1558–1561, 2015.[11]A. Sharma, N. Kumar, B. D. Kuppermann, F. Bandello und A. Loewenstein,„Biologics, biosilimars, and biobetters: Different terms or different drugs?“,Eye, Jg. 33, Nr. 7, S. 1032–1034, 2019.[12]Y. Wang und C. Wu, „Site-Specific Conjugation of Polymers to Proteins“,Biomacromolecules, Jg. 19, Nr. 6, S. 1804–1825, 2018.292Literatur[13]M. A. Stahmann und R. R. Becker, „A New Method for Adding Amino Acidsand Peptides to Proteins“, Journal of the American Chemical Society, Jg. 74,Nr. 10, S. 2695–2696, 1952.[14]R. R. Becker und A. Mark, „Protein Modification by Reaction with N-Carboxy-Amino Acid Anhydrides“, Journal of Biological Chemistry, Jg. 204, Nr. 2,S. 745–752, 1953.[15]K. M. Burridge, B. A. Shurina, C. T. Kozuszek, R. F. Parnell, J. S. Mont-gomery, J. L. VanPelt, N. M. Daman, R. M. McCarrick, T. A. Ramelot, D.Konkolewicz und R. C. Page, „Mapping protein–polymer conformations in bio-conjugates with atomic precision“, Chemical Science, Jg. 11, Nr. 24, S. 6160–6166, 2020.[16]A. Abuchowski, T. Van Es, N. C. Palczuk und F. F. Davis, „Alteration ofImmunological Properties of Bovine Serum Albumin by Covalent Attachmentof Polyethylene Glycol“, The Journal of Biological Chemistry, Jg. 252, Nr. 11,S. 3578–3581, 1977.[17]A. Abuchowski, J. R. McCoy, N. C. Palczuk, T. Van Es und F. F. Davis, „Effectof Covalent Attachment of Polyethylene Glycol on Immunogenicity and Cir-culating Life of Bovine Liver Catalase“, The Journal of Biological Chemistry,Jg. 252, Nr. 11, S. 3582–3586, 1977.[18]F. M. Veronese und A. Mero, „The Impact of PEGylation on Biological The-rapies“, BioDrugs, Jg. 22, Nr. 5, S. 315–329, 2008.[19]P. L. Turecek, M. J. Bossard, F. Schoetens und I. A. Ivens, „PEGylation ofBiopharmaceuticals: A Review of Chemistry and Nonclinical Safety Informa-tion of Approved Drugs“, Journal of Pharmaceutical Sciences, Jg. 105, Nr. 2,S. 460–475, 2016.[20]G. Kozma, T. Shimizu, T. Ishida und J. Szebeni, „Anti-PEG antibodies: Pro-perties, formation and role in adverse immune reactions to PEGylated nano-biopharmaceuticals“, Advanced Drug Delivery Reviews, 2020.[21]W. D. Chey, L. Webster, M. Sostek, J. Lappalainen, P. N. Barker und J. Tack,„Naloxegol for Opioid-Induced Constipation in Patients with Noncancer Pain“,New England Journal of Medicine, Jg. 370, Nr. 25, S. 2387–2396, 2014.[22]K. Kopeckova, T. Eckschlager, J. Sirc, R. Hobzova, J. Plch, J. Hrabeta und J.Michalek, „Nanodrugs Used in Cancer Therapy“, Biomedical Papers, Jg. 163,Nr. 2, S. 122–131, 2019.[23]A. Bunker, A. Magarkar und T. Viitala, „Rational design of liposomal drug de-livery systems, a review: Combined experimental and computational studies oflipid membranes, liposomes and their PEGylation“, Biochimica et BiophysicaActa (BBA) - Biomembranes, Jg. 1858, Nr. 10, S. 2334–2352, 2016.[24]P. C. Ke, S. Lin, W. J. Parak, T. P. Davis und F. Caruso, „A Decade of theProtein Corona“, ACS Nano, Jg. 11, Nr. 12, S. 11 773–11 776, 2017.[25]T. Okusaka, S. Okada, H. Ishii, M. Ikeda, H. Nakasuka, H. Nagahama, R.Iwata, H. Furukawa, K. Takayasu, Y. Nakanishi, M. Sakamoto, S. HirohashiLiteratur293und M. Yoshimori, „Transarterial Chemotherapy with Zinostatin Stimalamerfor Hepatocellular Carcinoma“, Oncology, Jg. 55, Nr. 4, S. 276–283, 1998.[26]E. M. Pelegri-O’Day, E.-W. Lin und H. D. Maynard, „Therapeutic Protein–Po-lymer Conjugates: Advancing Beyond PEGylation“, Journal of the AmericanChemical Society, Jg. 136, Nr. 41, S. 14 323–14 332, 2014.[27]P. Kiran, A. Khan, S. Neekhra, S. Pallod und R. Srivastava, „Nanohybrids asProtein-Polymer Conjugate Multimodal Therapeutics“, Frontiers in MedicalTechnology, Jg. 3, S. 676 025, 2021.[28]D. E. Koshland, „Application of a Theory of Enzyme Specificity to ProteinSynthesis“, Proceedings of the National Academy of Sciences, Jg. 44, Nr. 2,S. 98–104, 1958.[29]C. A. Lipinski, F. Lombardo, B. W. Dominy und P. J. Feeney, „Experimen-tal and computational approaches to estimate solubility and permeabilityin drug discovery and development settings 1PII of original article: S0169-409X(96)00423-1. The article was originally published in Advanced Drug De-livery Reviews 23 (1997) 3–25. 1“, Advanced Drug Delivery Reviews, Jg. 46,Nr. 1-3, S. 3–26, 2001.[30]C. A. Lipinski, „Lead- and drug-like compounds: The rule-of-five revolution“,Drug Discovery Today: Technologies, Jg. 1, Nr. 4, S. 337–341, 2004.[31]A. Bunker und T. Róg, „Mechanistic Understanding From Molecular Dyna-mics Simulation in Pharmaceutical Research 1: Drug Delivery“, Frontiers inMolecular Biosciences, Jg. 7, 2020.[32]A. J. Russell, S. L. Baker, C. M. Colina, C. A. Figg, J. L. Kaar, K. Matyja-szewski, A. Simakova und B. S. Sumerlin, „Next generation protein-polymerconjugates“, AIChE Journal, Jg. 64, Nr. 9, S. 3230–3245, 2018.[33]P. B. Lawrence und J. L. Price, „How PEGylation influences protein confor-mational stability“, Current Opinion in Chemical Biology, Jg. 34, S. 88–94,2016.[34]S. A. McNelles, V. M. Marando und A. Adronov, „Globular Polymer GraftsRequire a Critical Size for Efficient Molecular Sieving of Enzyme Substrates“,Angewandte Chemie International Edition, Jg. 58, Nr. 25, S. 8448–8453, 2019.[35]A. Munasinghe, A. Mathavan, A. Mathavan, P. Lin und C. M. Colina, „PEGy-lation within a confined hydrophobic cavity of a protein“, Physical ChemistryChemical Physics, Jg. 21, Nr. 46, S. 25 584–25 596, 2019.[36]A. Zaghmi, E. Mendez-Villuendas, A. Greschner, J. Liu, H. de Haan und M.Gauthier, „Mechanisms of activity loss for a multi-PEGylated protein by ex-periment and simulation“, Materials Today Chemistry, Jg. 12, S. 121–131,2019.[37]T. T. Hoang Thi, E. H. Pilkington, D. H. Nguyen, J. S. Lee, K. D. Parkund N. P. Truong, „The Importance of Poly(ethylene glycol) Alternatives forOvercoming PEG Immunogenicity in Drug Delivery and Bioconjugation“, Po-lymers, Jg. 12, Nr. 2, S. 298, 2020.294Literatur[38]M. S. Messina, K. M. Messina, A. Bhattacharya, H. R. Montgomery undH. D. Maynard, „Preparation of biomolecule-polymer conjugates by grafting-from using ATRP, RAFT, or ROMP“, Progress in Polymer Science, Jg. 100,S. 101 186, 2020.[39]G. Pasut und F. M. Veronese, „State of the art in PEGylation: The great ver-satility achieved after forty years of research“, Journal of Controlled Release,Jg. 161, Nr. 2, S. 461–472, 2012.[40]G. Wandrey, J. Wurzel, K. Hoffmann, T. Ladner, J. Büchs, L. Meinel undT. Lühmann, „Probing unnatural amino acid integration into enhanced greenfluorescent protein by genetic code expansion with a high-throughput screeningplatform“, Journal of Biological Engineering, Jg. 10, Nr. 1, S. 11, 2016.[41]A. C. Braun, M. Gutmann, T. Lühmann und L. Meinel, „Bioorthogonal stra-tegies for site-directed decoration of biomaterials with therapeutic proteins“,Journal of Controlled Release, Jg. 273, S. 68–85, 2018.[42]S. Kim, W. Ko, B. H. Sung, S. C. Kim und H. S. Lee, „Direct protein–proteinconjugation by genetically introducing bioorthogonal functional groups intoproteins“, Bioorganic & Medicinal Chemistry, Jg. 24, Nr. 22, S. 5816–5822,2016.[43]T. Lühmann, M. Schmidt, M. N. Leiske, V. Spieler, T. C. Majdanski, M.Grube, M. Hartlieb, I. Nischang, S. Schubert, U. S. Schubert und L. Mei-nel, „Site-Specific POxylation of Interleukin-4“, ACS Biomaterials Science &Engineering, Jg. 3, Nr. 3, S. 304–312, 2017.[44]F. Xie, X. Jia, Z. Zhu, Y. Wu, H. Jiang, H. Yang, Y. Cao, R. Zhu, B. Zhou, J.Du und Y. Tang, „Chemical trigger-enabled bioconjugation reaction“, Organic& Biomolecular Chemistry, 10.1039.D1OB01177D, 2021.[45]R. Falatach, S. Li, S. Sloane, C. McGlone, J. A. Berberich, R. C. Page, S.Averick und D. Konkolewicz, „Why synthesize protein–polymer conjugates?The stability and activity of chymotrypsin-polymer bioconjugates synthesizedby RAFT“, Polymer, Jg. 72, S. 382–386, 2015.[46]G. Pasut, F. Caboi, R. Schrepfer, G. Tonon, O. Schiavon und F. Veronese,„New active poly(ethylene glycol) derivative for amino coupling“, Reactiveand Functional Polymers, Jg. 67, Nr. 6, S. 529–539, 2007.[47]N. Vanparijs, S. Maji, B. Louage, L. Voorhaar, D. Laplace, Q. Zhang, Y. Shi,W. E. Hennink, R. Hoogenboom und B. G. De Geest, „Polymer-protein conju-gation via a ‘grafting to’ approach – a comparative study of the performanceof protein-reactive RAFT chain transfer agents“, Polymer Chemistry, Jg. 6,Nr. 31, S. 5602–5614, 2015.[48]A. L. Klibanov, M. A. Sllnkin und V. P. Torchilin, „Conjugation of proteinswith chelating polymers via water-soluble carbodiimide and N-hydroxysulfo-succinimide“, Applied Biochemistry and Biotechnology, Jg. 22, Nr. 1, S. 45–58,1989.Literatur295[49]G. Pound, J. M. McKenzie, R. F. M. Lange und B. Klumperman, „Poly-mer–protein conjugates from �-aldehyde endfunctional poly(N-vinylpyrrolidone)synthesised via xanthate-mediated living radical polymerisation“, ChemicalCommunications, Nr. 27, S. 3193–3195, 2008.[50]L. Tao, G. Mantovani, F. Lecolley und D. M. Haddleton, „Α-Aldehyde Termi-nally Functional Methacrylic Polymers from Living Radical Polymerization:Application in Protein Conjugation “Pegylation”“, Journal of the AmericanChemical Society, Jg. 126, Nr. 41, S. 13 220–13 221, 2004.[51]I. Dovgan, S. Kolodych, O. Koniev und A. Wagner, „2-(Maleimidomethyl)-1,3-Dioxanes (MD): A Serum-Stable Self-hydrolysable Hydrophilic Alternative toClassical Maleimide Conjugation“, Scientific Reports, Jg. 6, Nr. 1, S. 30 835,2016.[52]Y. Wang und C. Wu, „Quantitative Study of the Oligomerization of YeastPrion Sup35NM Proteins“, Biochemistry, Jg. 56, Nr. 50, S. 6575–6584, 2017.[53]Z. Li, R. Huang, H. Xu, J. Chen, Y. Zhan, X. Zhou, H. Chen und B. Jiang,„Divinylsulfonamides as Specific Linkers for Stapling Disulfide Bonds in Pep-tides“, Organic Letters, Jg. 19, Nr. 18, S. 4972–4975, 2017.[54]A. Foettinger, M. Melmer, A. Leitner und W. Lindner, „Reaction of the IndoleGroup with Malondialdehyde: Application for the Derivatization of Tryptop-han Residues in Peptides“, Bioconjugate Chemistry, Jg. 18, Nr. 5, S. 1678–1683, 2007.[55]M. A. Gauthier und H.-A. Klok, „Arginine-Specific Modification of Proteinswith Polyethylene Glycol“, Biomacromolecules, Jg. 12, Nr. 2, S. 482–493, 2011.[56]Y. Cong, E. Pawlisz, P. Bryant, S. Balan, E. Laurine, R. Tommasi, R. Singh,S. Dubey, K. Peciak, M. Bird, A. Sivasankar, J. Swierkosz, M. Muroni, S.Heidelberger, M. Farys, F. Khayrzad, J. Edwards, G. Badescu, I. Hodgson, C.Heise, S. Somavarapu, J. Liddell, K. Powell, M. Zloh, J.-w. Choi, A. Godwinund S. Brocchini, „Site-Specific PEGylation at Histidine Tags“, BioconjugateChemistry, Jg. 23, Nr. 2, S. 248–263, 2012.[57]S. Yamashita, H. Katsumi, N. Hibino, Y. Isobe, Y. Yagi, Y. Tanaka, S. Yama-da, C. Naito und A. Yamamoto, „Development of PEGylated aspartic acid-modified liposome as a bone-targeting carrier for the delivery of paclitaxel andtreatment of bone metastasis“, Biomaterials, Jg. 154, S. 74–85, 2018.[58]Y.-A. Lu und A. M. Felix, „Pegylated peptides III. Solid-phase synthesis withpegylating reagents of varying molecular weight: Synthesis of multiply pegy-lated peptides“, Reactive Polymers, Jg. 22, Nr. 3, S. 221–229, 1994.[59]M. J. Lobba, C. Fellmann, A. M. Marmelstein, J. C. Maza, E. N. Kissman,S. A. Robinson, B. T. Staahl, C. Urnes, R. J. Lew, C. S. Mogilevsky, J. A.Doudna und M. B. Francis, „Site-Specific Bioconjugation through Enzyme-Catalyzed Tyrosine–Cysteine Bond Formation“, ACS Central Science, Jg. 6,Nr. 9, S. 1564–1571, 2020.296Literatur[60]S. C. Reddington und M. Howarth, „Secrets of a covalent interaction for bio-materials and biotechnology: SpyTag and SpyCatcher“, Current Opinion inChemical Biology, Jg. 29, S. 94–99, 2015.[61]H. Schiffter-Weinle, „PEG - Das Multitalent“, Deutsche Apotheker Zeitung,Nr. 26, 2015.[62]L. Radi, M. Fach, M. Montigny, E. Berger-Nicoletti, W. Tremel und P. R.Wich, „Methods of protein surface PEGylation under structure preservationfor the emulsion-based formation of stable nanoparticles“, MedChemComm,Jg. 7, Nr. 9, S. 1738–1744, 2016.[63]Y. Qi und A. Chilkoti, „Protein–polymer conjugation — moving beyond PEGy-lation“, Current Opinion in Chemical Biology, Jg. 28, S. 181–193, 2015.[64]J. T. Huckaby, T. M. Jacobs, Z. Li, R. J. Perna, A. Wang, N. I. Nicely undS. K. Lai, „Structure of an anti-PEG antibody reveals an open ring that cap-tures highly flexible PEG polymers“, Communications Chemistry, Jg. 3, Nr. 1,S. 124, 2020.[65]C. Hohmann-Jeddi. „PEG Als Ein Anaphylaxie-Auslöser Nach Covid-ImpfungBestätigt“. (2021), Adresse: https://www.pharmazeutische-zeitung.de/peg-als-ein-anaphylaxie-ausloeser-nach-covid-impfung-bestaetigt-124884/ (besucht am 31. 01. 2021).[66]M. D. McSweeney, L. Shen, A. C. DeWalle, J. B. Joiner, E. C. Ciociola, D.Raghuwanshi, M. S. Macauley und S. K. Lai, „Pre-treatment with high mole-cular weight free PEG effectively suppresses anti-PEG antibody induction byPEG-liposomes in mice“, Journal of Controlled Release, Jg. 329, S. 774–781,2021.[67]Y. Kaneda, Y. Tsutsumi, Y. Yoshioka, H. Kamada, Y. Yamamoto, H. Kodaira,S.-i. Tsunoda, T. Okamoto, Y. Mukai, H. Shibata, S. Nakagawa und T. Mayu-mi, „The use of PVP as a polymeric carrier to improve the plasma half-life ofdrugs“, Biomaterials, Jg. 25, Nr. 16, S. 3259–3266, 2004.[68]P. Caliceti, O. Schiavon und F. M. Veronese, „Biopharmaceutical Propertiesof Uricase Conjugated to Neutral and Amphiphilic Polymers“, BioconjugateChemistry, Jg. 10, Nr. 4, S. 638–646, 1999.[69]P. Zhang, F. Sun, S. Liu und S. Jiang, „Anti-PEG antibodies in the clinic: Cur-rent issues and beyond PEGylation“, Journal of Controlled Release, Jg. 244,S. 184–193, 2016.[70]A. Thomas, S. S. Müller und H. Frey, „Beyond Poly(ethylene glycol): LinearPolyglycerol as a Multifunctional Polyether for Biomedical and Pharmaceuti-cal Applications“, Biomacromolecules, Jg. 15, Nr. 6, S. 1935–1954, 2014.[71]M. Imran ul-haq, B. F. Lai, R. Chapanian und J. N. Kizhakkedathu, „Influenceof architecture of high molecular weight linear and branched polyglycerolson their biocompatibility and biodistribution“, Biomaterials, Jg. 33, Nr. 35,S. 9135–9147, 2012.Literatur297[72]R. K. Kainthan und D. E. Brooks, „In vivo biological evaluation of highmolecular weight hyperbranched polyglycerols“, Biomaterials, Jg. 28, Nr. 32,S. 4779–4787, 2007.[73]R. K. Kainthan, S. R. Hester, E. Levin, D. V. Devine und D. E. Brooks, „Invitro biological evaluation of high molecular weight hyperbranched polyglyce-rols“, Biomaterials, Jg. 28, Nr. 31, S. 4581–4590, 2007.[74]K. P. Luef, R. Hoogenboom, U. S. Schubert und F. Wiesbrock, „Microwave-Assisted Cationic Ring-Opening Polymerization of 2-Oxazolines“, in Microwave-Assisted Polymer Synthesis, R. Hoogenboom, U. S. Schubert und F. Wiesbrock,Bd. 274, Springer International Publishing, 2015, S. 183–208.[75]T. Lorson, M. M. Lübtow, E. Wegener, M. S. Haider, S. Borova, D. Nahm,R. Jordan, M. Sokolski-Papkov, A. V. Kabanov und R. Luxenhofer, „Poly(2-oxazoline)s based biomaterials: A comprehensive and critical update“, Bioma-terials, Jg. 178, S. 204–280, 2018.[76]O. Sedlacek, V. R. de la Rosa und R. Hoogenboom, „Poly(2-oxazoline)–proteinconjugates“, in Polymer-Protein Conjugates, Elsevier, 2020, S. 407–420.[77]F. C. Gaertner, R. Luxenhofer, B. Blechert, R. Jordan und M. Essler, „Synthe-sis, biodistribution and excretion of radiolabeled poly(2-alkyl-2-oxazoline)s“,Journal of Controlled Release, Jg. 119, Nr. 3, S. 291–300, 2007.[78]T. X. Viegas, M. D. Bentley, J. M. Harris, Z. Fang, K. Yoon, B. Dizman,R. Weimer, A. Mero, G. Pasut und F. M. Veronese, „Polyoxazoline: Chemis-try, Properties, and Applications in Drug Delivery“, Bioconjugate Chemistry,Jg. 22, Nr. 5, S. 976–986, 2011.[79]A. Magarkar, T. Róg und A. Bunker, „A computational study suggests thatreplacing PEG with PMOZ may increase exposure of hydrophobic targetingmoiety“, European Journal of Pharmaceutical Sciences, Jg. 103, S. 128–135,2017.[80]Y. Qi, A. Simakova, N. J. Ganson, X. Li, K. M. Luginbuhl, I. Ozer, W. Liu,M. S. Hershfield, K. Matyjaszewski und A. Chilkoti, „A brush-polymer/exendin-4 conjugate reduces blood glucose levels for up to five days and eliminates po-ly(ethylene glycol) antigenicity“, Nature Biomedical Engineering, Jg. 1, Nr. 1,S. 0002, 2017.[81]C. J. Wiedermann und M. Joannidis, „Accumulation of hydroxyethyl starchin human and animal tissues: A systematic review“, Intensive Care Medicine,Jg. 40, Nr. 2, S. 160–170, 2014.[82]G. J. Lonergan, D. B. Cline und S. L. Abbondanzo, „Sickle Cell Anemia“,RadioGraphics, Jg. 21, Nr. 4, S. 971–994, 2001.[83]K. A. Luke, C. L. Higgins und P. Wittung-Stafshede, „Thermodynamic stabili-ty and folding of proteins from hyperthermophilic organisms: Thermodynamicstability and folding of proteins“, FEBS Journal, Jg. 274, Nr. 16, S. 4023–4033,2007.298Literatur[84]C. L. Araya, D. M. Fowler, W. Chen, I. Muniez, J. W. Kelly und S. Fields, „Afundamental protein property, thermodynamic stability, revealed solely fromlarge-scale measurements of protein function“, Proceedings of the NationalAcademy of Sciences, Jg. 109, Nr. 42, S. 16 858–16 863, 2012.[85]B. Li, Y. T. Yang, J. A. Capra und M. B. Gerstein, „Predicting changes inprotein thermodynamic stability upon point mutation with deep 3D convo-lutional neural networks“, PLOS Computational Biology, Jg. 16, Nr. 11, P.Fariselli, e1008291, 2020.[86]S. M. Kelly, T. J. Jess und N. C. Price, „How to study proteins by circulardichroism“, Biochimica et Biophysica Acta (BBA) - Proteins and Proteomics,Jg. 1751, Nr. 2, S. 119–139, 2005.[87]A. Rodger und B. Nordén, Circular Dichroism and Linear Dichroism, Ser. Ox-ford Chemistry Masters 1. Oxford ; New York: Oxford University Press, 1997,150 S.[88]N. J. Greenfield, „Using circular dichroism collected as a function of tem-perature to determine the thermodynamics of protein unfolding and bindinginteractions“, Nature Protocols, Jg. 1, Nr. 6, S. 2527–2535, 2007.[89]J. K. Kranz und C. Schalk-Hihi, „Protein Thermal Shifts to Identify LowMolecular Weight Fragments“, in Methods in Enzymology, Bd. 493, Elsevier,2011, S. 277–298.[90]K. Huynh und C. L. Partch, „Analysis of Protein Stability and Ligand Inter-actions by Thermal Shift Assay“, Current Protocols in Protein Science, Jg. 79,Nr. 1, S. 28.9.1–28.9.14, 2015.[91]A. O. Magnusson, A. Szekrenyi, H.-J. Joosten, J. Finnigan, S. Charnock undW.-D. Fessner, „nanoDSF as screening tool for enzyme libraries and biotech-nology development“, The FEBS Journal, Jg. 286, Nr. 1, S. 184–204, 2019.[92]J. A. Rodríguez-Martínez, R. J. Solá, B. Castillo, H. R. Cintrón-Colón, I.Rivera-Rivera, G. Barletta und K. Griebenow, „Stabilization of �-chymotrypsinupon PEGylation correlates with reduced structural dynamics“, Biotechnologyand Bioengineering, Jg. 101, Nr. 6, S. 1142–1149, 2008.[93]C. Yang, D. Lu und Z. Liu, „How PEGylation Enhances the Stability andPotency of Insulin: A Molecular Dynamics Simulation“, Biochemistry, Jg. 50,Nr. 13, S. 2585–2593, 2011.[94]E. P. DeBenedictis, E. Hamed und S. Keten, „Mechanical Reinforcement ofProteins with Polymer Conjugation“, ACS Nano, Jg. 10, Nr. 2, S. 2259–2267,2016.[95]W. Meng, X. Guo, M. Qin, H. Pan, Y. Cao und W. Wang, „MechanisticInsights into the Stabilization of srcSH3 by PEGylation“, Langmuir, Jg. 28,Nr. 46, S. 16 133–16 140, 2012.Literatur299[96]S. L. Baker, A. Munasinghe, H. Murata, P. Lin, K. Matyjaszewski, C. M.Colina und A. J. Russell, „Intramolecular Interactions of Conjugated Poly-mers Mimic Molecular Chaperones to Stabilize Protein–Polymer Conjugates“,Biomacromolecules, Jg. 19, Nr. 9, S. 3798–3813, 2018.[97]B. K. Pandey, M. S. Smith, C. Torgerson, P. B. Lawrence, S. S. Matthews,E. Watkins, M. L. Groves, M. B. Prigozhin und J. L. Price, „Impact of Site-Specific PEGylation on the Conformational Stability and Folding Rate of thePin WW Domain Depends Strongly on PEG Oligomer Length“, BioconjugateChemistry, Jg. 24, Nr. 5, S. 796–802, 2013.[98]J. L. Price, E. T. Powers und J. W. Kelly, „N-PEGylation of a Reverse TurnIs Stabilizing in Multiple Sequence Contexts, unlike N-GlcNAcylation“, ACSChemical Biology, Jg. 6, Nr. 11, S. 1188–1192, 2011.[99]S.-H. Chao, S. S. Matthews, R. Paxman, A. Aksimentiev, M. Gruebele undJ. L. Price, „Two Structural Scenarios for Protein Stabilization by PEG“, TheJournal of Physical Chemistry B, Jg. 118, Nr. 28, S. 8388–8395, 2014.[100]P. B. Lawrence, Y. Gavrilov, S. S. Matthews, M. I. Langlois, D. Shental-Bechor, H. M. Greenblatt, B. K. Pandey, M. S. Smith, R. Paxman, C. D.Torgerson, J. P. Merrell, C. C. Ritz, M. B. Prigozhin, Y. Levy und J. L. Price,„Criteria for Selecting PEGylation Sites on Proteins for Higher Thermody-namic and Proteolytic Stability“, Journal of the American Chemical Society,Jg. 136, Nr. 50, S. 17 547–17 560, 2014.[101]P. B. Lawrence, W. M. Billings, M. B. Miller, B. K. Pandey, A. R. Stephens,M. I. Langlois und J. L. Price, „Conjugation Strategy Strongly Impacts theConformational Stability of a PEG-Protein Conjugate“, ACS Chemical Biolo-gy, Jg. 11, Nr. 7, S. 1805–1809, 2016.[102]Q. Xiao, S. R. E. Draper, M. S. Smith, N. Brown, N. A. B. Pugmire, D. S.Ashton, A. J. Carter, E. E. K. Lawrence und J. L. Price, „Influence of PEGyla-tion on the Strength of Protein Surface Salt Bridges“, ACS Chemical Biology,Jg. 14, Nr. 7, S. 1652–1659, 2019.[103]S. R. E. Draper, D. S. Ashton, B. M. Conover, A. J. Carter, K. L. Stern, Q.Xiao und J. L. Price, „PEGylation near a Patch of Nonpolar Surface ResiduesIncreases the Conformational Stability of the WW Domain“, The Journal ofOrganic Chemistry, Jg. 85, Nr. 3, S. 1725–1730, 2020.[104]S. R. E. Draper, Z. B. Jones, S. O. Earl, N. A. Dalley, D. S. Ashton, A. J.Carter, B. M. Conover und J. L. Price, „PEGylation Increases the Strengthof a Nearby NH-� Hydrogen Bond in the WW Domain“, Biochemistry, Jg. 60,Nr. 26, S. 2064–2070, 2021.[105]E. Hamed, T. Xu und S. Keten, „Poly(ethylene glycol) Conjugation Stabilizesthe Secondary Structure of �-Helices by Reducing Peptide Solvent AccessibleSurface Area“, Biomacromolecules, Jg. 14, Nr. 11, S. 4053–4060, 2013.[106]C.-Y. Liu, X. Li, W.-Y. Chen, L.-C. Chang, Y.-F. Chen, H.-L. Chen, Y.-S.Sun, H.-Y. Lai und E.-W. Huang, „PEGylation Site-Dependent Structural300LiteraturHeterogeneity Study of MonoPEGylated Human Parathyroid Hormone Frag-ment hPTH(1–34)“, Langmuir, Jg. 30, Nr. 38, S. 11 421–11 427, 2014.[107]A. Munasinghe, A. Mathavan, A. Mathavan, P. Lin und C. M. Colina, „Mole-cular Insight into the Protein–Polymer Interactions in N-Terminal PEGylatedBovine Serum Albumin“, The Journal of Physical Chemistry B, Jg. 123, Nr. 25,S. 5196–5205, 2019.[108]B. K. Pandey, M. S. Smith und J. L. Price, „Cysi –Lysi +3 –Lysi +4Triad: A General Approach for PEG-Based Stabilization of �-Helical Proteins“,Biomacromolecules, Jg. 15, Nr. 12, S. 4643–4647, 2014.[109]T. A. Wright, M. Lucius Dougherty, B. Schmitz, K. M. Burridge, K. Makaroff,J. M. Stewart, H. D. Fischesser, J. T. Shepherd, J. A. Berberich, D. Konkole-wicz und R. C. Page, „Polymer Conjugation to Enhance Cellulase Activity andPreserve Thermal and Functional Stability“, Bioconjugate Chemistry, Jg. 28,Nr. 10, S. 2638–2645, 2017.[110]S. Farhadian, B. Shareghi, A. A. Saboury, A. K. Babaheydari, F. Raisi und E.heidari, „Molecular aspects of the interaction of spermidine and �-chymotrypsin“,International Journal of Biological Macromolecules, Jg. 92, S. 523–532, 2016.[111]C. S. Cummings, A. S. Campbell, S. L. Baker, S. Carmali, H. Murata undA. J. Russell, „Design of Stomach Acid-Stable and Mucin-Binding EnzymePolymer Conjugates“, Biomacromolecules, Jg. 18, Nr. 2, S. 576–586, 2017.[112]H. García-Arellano, B. Valderrama, G. Saab-Rincón und R. Vazquez-Duhalt,„High Temperature Biocatalysis by Chemically Modified Cytochrome c“, Bio-conjugate Chemistry, Jg. 13, Nr. 6, S. 1336–1344, 2002.[113]B. Plesner, P. Westh und A. D. Nielsen, „Biophysical characterisation of Gly-coPEGylated recombinant human factor VIIa“, International Journal of Phar-maceutics, Jg. 406, Nr. 1-2, S. 62–68, 2011.[114]Y. Nodake und N. Yamasaki, „Some Properties of a Macromolecular Conju-gate of Lysozyme Prepared by Modification with a MonomethoxypolyethyleneGlycol Derivative“, Bioscience, Biotechnology, and Biochemistry, Jg. 64, Nr. 4,S. 767–774, 2000.[115]J. Morgenstern, P. Baumann, C. Brunner und J. Hubbuch, „Effect of PEGmolecular weight and PEGylation degree on the physical stability of PEGy-lated lysozyme“, International Journal of Pharmaceutics, Jg. 519, Nr. 1-2,S. 408–417, 2017.[116]M. Lucius, R. Falatach, C. McGlone, K. Makaroff, A. Danielson, C. Williams,J. C. Nix, D. Konkolewicz, R. C. Page und J. A. Berberich, „Investigatingthe Impact of Polymer Functional Groups on the Stability and Activity ofLysozyme–Polymer Conjugates“, Biomacromolecules, Jg. 17, Nr. 3, S. 1123–1134, 2016.Literatur301[117]A. Natalello, D. Ami, M. Collini, L. D’Alfonso, G. Chirico, G. Tonon, S. Scara-muzza, R. Schrepfer und S. M. Doglia, „Biophysical Characterization of Met-G-CSF: Effects of Different Site-Specific Mono-Pegylations on Protein Stabilityand Aggregation“, PLoS ONE, Jg. 7, Nr. 8, e42511, 2012.[118]S. R. E. Draper, P. B. Lawrence, W. M. Billings, Q. Xiao, N. P. Brown, N. A.Bécar, D. J. Matheson, A. R. Stephens und J. L. Price, „Polyethylene Gly-col Based Changes to �-Sheet Protein Conformational and Proteolytic Stabili-ty Depend on Conjugation Strategy and Location“, Bioconjugate Chemistry,Jg. 28, Nr. 10, S. 2507–2513, 2017.[119]L. Cao, X. Shi, Y. Cui, W. Yang, G. Chen, L. Yuan und H. Chen, „Pro-tein–polymer conjugates prepared via host–guest interactions: Effects of theconjugation site, polymer type and molecular weight on protein activity“, Po-lymer Chemistry, Jg. 7, Nr. 32, S. 5139–5146, 2016.[120]W. Yang, L. Zhu, Y. Cui, H. Wang, Y. Wang, L. Yuan und H. Chen, „Impro-vement of Site-Directed Protein–Polymer Conjugates: High Bioactivity andStability Using a Soft Chain-Transfer Agent“, ACS Applied Materials & In-terfaces, Jg. 8, Nr. 25, S. 15 967–15 974, 2016.[121]Q. Mu, T. Hu und J. Yu, „Molecular Insight into the Steric Shielding Effectof PEG on the Conjugated Staphylokinase: Biochemical Characterization andMolecular Dynamics Simulation“, PLoS ONE, Jg. 8, Nr. 7, e68559, 2013.[122]H. F. Gaertner und A. J. Puigserver, „Increased activity and stability ofpoly(ethylene glycol)-modified trypsin“, Enzyme and Microbial Technology,Jg. 14, Nr. 2, S. 150–155, 1992.[123]B. Treetharnmathurot, L. Dieudonné, E. L. Ferguson, D. Schmaljohann, R.Duncan und R. Wiwattanapatapee, „Dextrin–trypsin and ST-HPMA–trypsinconjugates: Enzyme activity, autolysis and thermal stability“, InternationalJournal of Pharmaceutics, Jg. 373, Nr. 1-2, S. 68–76, 2009.[124]B. J. Alder und T. E. Wainwright, „Phase Transition for a Hard Sphere Sys-tem“, The Journal of Chemical Physics, Jg. 27, Nr. 5, S. 1208–1209, 1957.[125]B. J. Alder und T. E. Wainwright, „Studies in Molecular Dynamics. I. GeneralMethod“, The Journal of Chemical Physics, Jg. 31, Nr. 2, S. 459–466, 1959.[126]J. A. McCammon, B. R. Gelin und M. Karplus, „Dynamics of folded proteins“,Nature, Jg. 267, Nr. 5612, S. 585–590, 1977.[127]S. A. Hollingsworth und R. O. Dror, „Molecular Dynamics Simulation for All“,Neuron, Jg. 99, Nr. 6, S. 1129–1143, 2018.[128]W. F. van Gunsteren, D. Bakowies, R. Baron, I. Chandrasekhar, M. Chris-ten, X. Daura, P. Gee, D. P. Geerke, A. Glättli, P. H. Hünenberger, M. A.Kastenholz, C. Oostenbrink, M. Schenk, D. Trzesniak, N. F. A. van der Vegtund H. B. Yu, „Biomolekulare Modellierung: Ziele, Probleme, Perspektiven“,Angewandte Chemie, Jg. 118, Nr. 25, S. 4168–4198, 2006.[129]J. D. Durrant und J. A. McCammon, „Molecular dynamics simulations anddrug discovery“, BMC Biology, Jg. 9, Nr. 1, S. 71, 2011.302Literatur[130]E. Braun, J. Gilmer, H. B. Mayes, D. L. Mobley, J. I. Monroe, S. Prasad undD. M. Zuckerman, „Best Practices for Foundations in Molecular Simulations[Article v1.0]“, Living Journal of Computational Molecular Science, Jg. 1,Nr. 1, S. 5957, 2019.[131]S. Riniker, „Fixed-Charge Atomistic Force Fields for Molecular Dynamics Si-mulations in the Condensed Phase: An Overview“, Journal of Chemical Infor-mation and Modeling, Jg. 58, Nr. 3, S. 565–578, 2018.[132]J. D. Hirst, D. R. Glowacki und M. Baaden, „Molecular simulations and vi-sualization: Introduction and overview“, Faraday Discuss., Jg. 169, S. 9–22,2014.[133]J. A. Maier, C. Martinez, K. Kasavajhala, L. Wickstrom, K. E. Hauser und C.Simmerling, „ff14SB: Improving the Accuracy of Protein Side Chain and Back-bone Parameters from ff99SB“, Journal of Chemical Theory and Computation,Jg. 11, Nr. 8, S. 3696–3713, 2015.[134]A. Leach, Molecular Modeling: Principles and Applications. Pearson EducationLimited, 2001.[135]D. Frenkel und B. Smit, Understanding Molecular Simulation: From Algorith-ms to Applications, 2nd ed, Ser. Computational Science Series 1. San Diego:Academic Press, 2002, 638 S.[136]C. J. Cramer, Essentials of Computational Chemistry: Theories and Models.Wiley, 2004.[137]J.-P. Ryckaert, G. Ciccotti und H. J. Berendsen, „Numerical integration of thecartesian equations of motion of a system with constraints: Molecular dynamicsof n-alkanes“, Journal of Computational Physics, Jg. 23, Nr. 3, S. 327–341,1977.[138]H. C. Andersen, „Rattle: A “velocity” version of the shake algorithm for mole-cular dynamics calculations“, Journal of Computational Physics, Jg. 52, Nr. 1,S. 24–34, 1983.[139]H. M. Berman, „The Protein Data Bank“, Nucleic Acids Research, Jg. 28,Nr. 1, S. 235–242, 2000.[140]T. Darden, D. York und L. Pedersen, „Particle mesh Ewald: An N log(N)method for Ewald sums in large systems“, The Journal of Chemical Physics,Jg. 98, Nr. 12, S. 10 089–10 092, 1993.[141]B. N. Manjula, A. Tsai, R. Upadhya, K. Perumalsamy, P. K. Smith, A. Mala-valli, K. Vandegriff, R. M. Winslow, M. Intaglietta, M. Prabhakaran, J. M.Friedman und A. S. Acharya, „Site-Specific PEGylation of Hemoglobin atCys-93(Beta): Correlation between the Colligative Properties of the PEGy-lated Protein and the Length of the Conjugated PEG Chain“, BioconjugateChemistry, Jg. 14, Nr. 2, S. 464–472, 2003.[142]H. Lee, „Molecular Simulations of PEGylated Biomolecules, Liposomes, andNanoparticles for Drug Delivery Applications“, Pharmaceutics, Jg. 12, Nr. 6,S. 533, 2020.Literatur303[143]R. Woods und R. Chappelle, „Restrained electrostatic potential atomic partialcharges for condensed-phase simulations of carbohydrates“, Journal of Mole-cular Structure: THEOCHEM, Jg. 527, Nr. 1-3, S. 149–156, 2000.[144]R. Galindo-Murillo, J. C. Robertson, M. Zgarbová, J. Šponer, M. Otyepka, P.Jurečka und T. E. Cheatham, „Assessing the Current State of Amber ForceField Modifications for DNA“, Journal of Chemical Theory and Computation,Jg. 12, Nr. 8, S. 4114–4127, 2016.[145]K. N. Kirschner, A. B. Yongye, S. M. Tschampel, J. González-Outeiriño, C. R.Daniels, B. L. Foley und R. J. Woods, „GLYCAM06: A generalizable biomole-cular force field. Carbohydrates: GLYCAM06“, Journal of Computational Che-mistry, Jg. 29, Nr. 4, S. 622–655, 2008.[146]C. J. Dickson, B. D. Madej, Å. A. Skjevik, R. M. Betz, K. Teigen, I. R.Gould und R. C. Walker, „Lipid14: The Amber Lipid Force Field“, Journal ofChemical Theory and Computation, Jg. 10, Nr. 2, S. 865–879, 2014.[147]J. Wang, R. M. Wolf, J. W. Caldwell, P. A. Kollman und D. A. Case, „Deve-lopment and testing of a general amber force field“, Journal of ComputationalChemistry, Jg. 25, Nr. 9, S. 1157–1174, 2004.[148]S. F. Sousa, J. Peres, M. Coelho und T. F. Vieira, „Analyzing PEGylationthrough Molecular Dynamics Simulations“, ChemistrySelect, Jg. 3, Nr. 29,S. 8415–8427, 2018.[149]H. Lee, R. M. Venable, A. D. MacKerell und R. W. Pastor, „Molecular Dyna-mics Studies of Polyethylene Oxide and Polyethylene Glycol: HydrodynamicRadius and Shape Anisotropy“, Biophysical Journal, Jg. 95, Nr. 4, S. 1590–1599, 2008.[150]J. Fischer, D. Paschek, A. Geiger und G. Sadowski, „Modeling of Aqueous Po-ly(oxyethylene) Solutions: 1. Atomistic Simulations“, The Journal of PhysicalChemistry B, Jg. 112, Nr. 8, S. 2388–2398, 2008.[151]N. S. V. Barbosa, Y. Zhang, E. R. A. Lima, F. W. Tavares und E. J. Maginn,„Development of an AMBER-compatible transferable force field for poly(ethy-lene glycol) ethers (glymes)“, Journal of Molecular Modeling, Jg. 23, Nr. 6,S. 1–14, 2017.[152]T. Ottallah, S. A. Parandian und S. W. Rick, „Analysis of Atomistic Potentialsfor Poly(ethylene glycol) Ethers“, Journal of Chemical Theory and Computa-tion, Jg. 17, Nr. 1, S. 315–321, 2021.[153]K. Lindorff-Larsen, S. Piana, K. Palmo, P. Maragakis, J. L. Klepeis, R. O.Dror und D. E. Shaw, „Improved side-chain torsion potentials for the Amberff99SB protein force field: Improved Protein Side-Chain Potentials“, Proteins:Structure, Function, and Bioinformatics, Jg. 78, Nr. 8, S. 1950–1958, 2010.[154]W. F. van Gunsteren, S. Billeter, A. Eising, P. Hünenberger, P. Krüger, A.Mark, W. Scott und I. Tironi, „Biomolecular Simulation: The GROMOS96Manual and User Guide“, Vdf Hochschulverlag AG an der ETH Zürich, Zürich,Jg. 86, S. 1–1044, 1996.304Literatur[155]B. Isralewitz, M. Gao und K. Schulten, „Steered molecular dynamics and me-chanical functions of proteins“, Current Opinion in Structural Biology, Jg. 11,Nr. 2, S. 224–230, 2001.[156]D. Hamelberg, J. Mongan und J. A. McCammon, „Accelerated molecular dy-namics: A promising and efficient simulation method for biomolecules“, TheJournal of Chemical Physics, Jg. 120, Nr. 24, S. 11 919–11 929, 2004.[157]S. J. Marrink, A. H. de Vries und A. E. Mark, „Coarse Grained Model forSemiquantitative Lipid Simulations“, The Journal of Physical Chemistry B,Jg. 108, Nr. 2, S. 750–760, 2004.[158]L. Monticelli, S. K. Kandasamy, X. Periole, R. G. Larson, D. P. Tielemanund S.-J. Marrink, „The MARTINI Coarse-Grained Force Field: Extension toProteins“, Journal of Chemical Theory and Computation, Jg. 4, Nr. 5, S. 819–834, 2008.[159]C. A. López, A. J. Rzepiela, A. H. de Vries, L. Dijkhuizen, P. H. Hünenbergerund S. J. Marrink, „Martini Coarse-Grained Force Field: Extension to Car-bohydrates“, Journal of Chemical Theory and Computation, Jg. 5, Nr. 12,S. 3195–3210, 2009.[160]J. J. Uusitalo, H. I. Ingólfsson, P. Akhshi, D. P. Tieleman und S. J. Marrink,„Martini Coarse-Grained Force Field: Extension to DNA“, Journal of ChemicalTheory and Computation, Jg. 11, Nr. 8, S. 3932–3945, 2015.[161]J. J. Uusitalo, H. I. Ingólfsson, S. J. Marrink und I. Faustino, „Martini Coarse-Grained Force Field: Extension to RNA“, Biophysical Journal, Jg. 113, Nr. 2,S. 246–256, 2017.[162]S. Kmiecik, D. Gront, M. Kolinski, L. Wieteska, A. E. Dawid und A. Kolinski,„Coarse-Grained Protein Models and Their Applications“, Chemical Reviews,Jg. 116, Nr. 14, S. 7898–7936, 2016.[163]F. Grünewald, P. C. Kroon, P. C. T. Souza und S. J. Marrink, „Protocol forSimulations of PEGylated Proteins with Martini 3“, in Structural Genomics,Y. W. Chen und C.-P. B. Yiu, Bd. 2199, New York, NY: Springer US, 2021,S. 315–335.[164]P. V. Coveney und S. Wan, „On the calculation of equilibrium thermodynamicproperties from molecular dynamics“, Physical Chemistry Chemical Physics,Jg. 18, Nr. 44, S. 30 236–30 240, 2016.[165]A. Grossfield und D. M. Zuckerman, „Chapter 2 Quantifying Uncertainty andSampling Quality in Biomolecular Simulations“, in Annual Reports in Com-putational Chemistry, Bd. 5, Elsevier, 2009, S. 23–48.[166]A. Grossfield, P. N. Patrone, D. R. Roe, A. J. Schultz, D. Siderius und D. M.Zuckerman, „Best Practices for Quantification of Uncertainty and SamplingQuality in Molecular Simulations [Article v1.0]“, Living Journal of Computa-tional Molecular Science, Jg. 1, Nr. 1, S. 5067, 2019.Literatur305[167]L. Sawle und K. Ghosh, „Convergence of Molecular Dynamics Simulation ofProtein Native States: Feasibility vs Self-Consistency Dilemma“, Journal ofChemical Theory and Computation, Jg. 12, Nr. 2, S. 861–869, 2016.[168]L. J. Smith, X. Daura und W. F. van Gunsteren, „Assessing equilibration andconvergence in biomolecular simulations“, Proteins: Structure, Function, andGenetics, Jg. 48, Nr. 3, S. 487–496, 2002.[169]T. D. Romo und A. Grossfield, „Block Covariance Overlap Method and Con-vergence in Molecular Dynamics Simulation“, Journal of Chemical Theory andComputation, Jg. 7, Nr. 8, S. 2464–2472, 2011.[170]W. K. Hastings, „Monte Carlo sampling methods using Markov chains andtheir applications“, Biometrika, Jg. 57, Nr. 1, S. 97–109, 1970.[171]U. Krengel, Einführung in die Wahrscheinlichkeitstheorie und Statistik. View-eg+Teubner Verlag, 2005.[172]P. Heidelberger und P. D. Welch, „A spectral method for confidence intervalgeneration and run length control in simulations“, Communications of theACM, Jg. 24, Nr. 4, S. 233–245, 1981.[173]P. Heidelberger und P. D. Welch, „Simulation Run Length Control in thePresence of an Initial Transient“, Operations Research, Jg. 31, Nr. 6, S. 1109–1144, 1983.[174]M. Plummer, N. Best, K. Cowles und K. Vines, „CODA: Convergence Dia-gnosis and Output Analysis for MCMC“, R News, Jg. 6, Nr. 1, S. 7–11, 2006.[175]J. Geweke, „Evaluating the Accuracy of Sampling-Based Approaches to theCalculation of Posterior Moments“, Federal Reserve Bank of Minneapolis, Re-search Department Staff Report 148, 1991.[176]Y. Miao, V. A. Feher und J. A. McCammon, „Gaussian Accelerated MolecularDynamics: Unconstrained Enhanced Sampling and Free Energy Calculation“,Journal of Chemical Theory and Computation, Jg. 11, Nr. 8, S. 3584–3595,2015.[177]Y. T. Pang, Y. Miao, Y. Wang und J. A. McCammon, „Gaussian AcceleratedMolecular Dynamics in NAMD“, Journal of Chemical Theory and Computa-tion, Jg. 13, Nr. 1, S. 9–19, 2017.[178]Y. Miao, W. Sinko, L. Pierce, D. Bucher, R. C. Walker und J. A. McCam-mon, „Improved Reweighting of Accelerated Molecular Dynamics Simulationsfor Free Energy Calculation“, Journal of Chemical Theory and Computation,Jg. 10, Nr. 7, S. 2677–2689, 2014.[179]O. F. Lange und H. Grubmüller, „Full correlation analysis of conformationalprotein dynamics“, Proteins: Structure, Function, and Bioinformatics, Jg. 70,Nr. 4, S. 1294–1312, 2007.[180]Y. Miao, A. Bhattarai und J. Wang, „Ligand Gaussian Accelerated MolecularDynamics (LiGaMD): Characterization of Ligand Binding Thermodynamicsand Kinetics“, Journal of Chemical Theory and Computation, Jg. 16, Nr. 9,S. 5526–5547, 2020.306Literatur[181]A. Laio und F. L. Gervasio, „Metadynamics: A method to simulate rare eventsand reconstruct the free energy in biophysics, chemistry and material science“,Reports on Progress in Physics, Jg. 71, Nr. 12, S. 126 601, 2008.[182]A. Barducci, M. Bonomi und M. Parrinello, „Metadynamics“, Wiley Interdi-sciplinary Reviews: Computational Molecular Science, Jg. 1, Nr. 5, S. 826–843,2011.[183]L. Sutto, S. Marsili und F. L. Gervasio, „New advances in metadynamics: Newadvances in metadynamics“, Wiley Interdisciplinary Reviews: ComputationalMolecular Science, Jg. 2, Nr. 5, S. 771–779, 2012.[184]A. Barducci, G. Bussi und M. Parrinello, „Well-Tempered Metadynamics: ASmoothly Converging and Tunable Free-Energy Method“, Physical Review Let-ters, Jg. 100, Nr. 2, S. 020 603, 2008.[185]P. Raiteri, A. Laio, F. L. Gervasio, C. Micheletti und M. Parrinello, „EfficientReconstruction of Complex Free Energy Landscapes by Multiple Walkers Me-tadynamics“, The Journal of Physical Chemistry B, Jg. 110, Nr. 8, S. 3533–3539, 2006.[186]G. Bussi und D. Branduardi, „Free-Energy Calculations with Metadynamics:Theory and Practice“, in Reviews in Computational Chemistry, A. L. Parrillund K. B. Lipkowitz, John Wiley & Sons, Inc, 2015, S. 1–49.[187]M. Bonomi, D. Branduardi, G. Bussi, C. Camilloni, D. Provasi, P. Raite-ri, D. Donadio, F. Marinelli, F. Pietrucci, R. A. Broglia und M. Parrinello,„PLUMED: A portable plugin for free-energy calculations with molecular dy-namics“, Computer Physics Communications, Jg. 180, Nr. 10, S. 1961–1972,2009.[188]G. A. Tribello, M. Bonomi, D. Branduardi, C. Camilloni und G. Bussi, „PLU-MED 2: New feathers for an old bird“, Computer Physics Communications,Jg. 185, Nr. 2, S. 604–613, 2014.[189]A. Cherkasov, E. N. Muratov, D. Fourches, A. Varnek, I. I. Baskin, M. Cronin,J. Dearden, P. Gramatica, Y. C. Martin, R. Todeschini, V. Consonni, V. E.Kuz’min, R. Cramer, R. Benigni, C. Yang, J. Rathman, L. Terfloth, J. Gastei-ger, A. Richard und A. Tropsha, „QSAR Modeling: Where Have You Been?Where Are You Going To?“, Journal of Medicinal Chemistry, Jg. 57, Nr. 12,S. 4977–5010, 2014.[190]C. Hansch, P. P. Maloney, T. Fujita und R. M. Muir, „Correlation of BiologicalActivity of Phenoxyacetic Acids with Hammett Substituent Constants andPartition Coefficients“, Nature, Jg. 194, Nr. 4824, S. 178–180, 1962.[191]C. Hansch, „Quantitative approach to biochemical structure-activity relati-onships“, Accounts of Chemical Research, Jg. 2, Nr. 8, S. 232–239, 1969.[192]C. Hansch, R. M. Muir, T. Fujita, P. P. Maloney, F. Geiger und M. Streich,„The Correlation of Biological Activity of Plant Growth Regulators and Chlo-romycetin Derivatives with Hammett Constants and Partition Coefficients“,Journal of the American Chemical Society, Jg. 85, Nr. 18, S. 2817–2824, 1963.Literatur307[193]C. Hansch und T. Fujita, „P-sigma-pi Analysis. A Method for the Correlati-on of Biological Activity and Chemical Structure“, Journal of the AmericanChemical Society, Jg. 86, Nr. 8, S. 1616–1626, 1964.[194]S. Wold, „Chemometrics; what do we mean with it, and what do we wantfrom it?“, Chemometrics and Intelligent Laboratory Systems, Jg. 30, Nr. 1,S. 109–115, 1995.[195]E. Byvatov, U. Fechner, J. Sadowski und G. Schneider, „Comparison of Sup-port Vector Machine and Artificial Neural Network Systems for Drug/Non-drug Classification“, Journal of Chemical Information and Computer Sciences,Jg. 43, Nr. 6, S. 1882–1889, 2003.[196]M. Damale, S. Harke, F. Kalam Khan, D. Shinde und J. Sangshetti, „Re-cent Advances in Multidimensional QSAR (4D-6D): A Critical Review“, Mini-Reviews in Medicinal Chemistry, Jg. 14, Nr. 1, S. 35–55, 2014.[197]S. Wold, M. Sjöström und L. Eriksson, „PLS-regression: A basic tool of che-mometrics“, Chemometrics and Intelligent Laboratory Systems, Jg. 58, Nr. 2,S. 109–130, 2001.[198]R. D. Cramer, D. E. Patterson und J. D. Bunce, „Comparative molecular fieldanalysis (CoMFA). 1. Effect of shape on binding of steroids to carrier proteins“,Journal of the American Chemical Society, Jg. 110, Nr. 18, S. 5959–5967, 1988.[199]P. Gramatica und A. Sangion, „A Historical Excursus on the Statistical Vali-dation Parameters for QSAR Models: A Clarification Concerning Metrics andTerminology“, Journal of Chemical Information and Modeling, Jg. 56, Nr. 6,S. 1127–1131, 2016.[200]P. Yin und X. Fan, „Estimating R 2 Shrinkage in Multiple Regression: AComparison of Different Analytical Methods“, The Journal of ExperimentalEducation, Jg. 69, Nr. 2, S. 203–224, 2001.[201]L. M. Shi, H. Fang, W. Tong, J. Wu, R. Perkins, R. M. Blair, W. S. Branham,S. L. Dial, C. L. Moland und D. M. Sheehan, „QSAR Models Using a LargeDiverse Set of Estrogens“, Journal of Chemical Information and ComputerSciences, Jg. 41, Nr. 1, S. 186–195, 2001.[202]G. Schüürmann, R.-U. Ebert, J. Chen, B. Wang und R. Kühne, „ExternalValidation and Prediction Employing the Predictive Squared Correlation Co-efficient — Test Set Activity Mean vs Training Set Activity Mean“, Journalof Chemical Information and Modeling, Jg. 48, Nr. 11, S. 2140–2145, 2008.[203]V. Consonni, D. Ballabio und R. Todeschini, „Comments on the Definition ofthe Q 2 Parameter for QSAR Validation“, Journal of Chemical Informationand Modeling, Jg. 49, Nr. 7, S. 1669–1678, 2009.[204]L. I.-K. Lin, „A Concordance Correlation Coefficient to Evaluate Reproduci-bility“, Biometrics, Jg. 45, Nr. 1, S. 255, 1989.[205]L. I.-K. Lin, „Assay Validation Using the Concordance Correlation Coeffici-ent“, Biometrics, Jg. 48, Nr. 2, S. 599–604, 1992.308Literatur[206]M. G. Kendall, „A new measure of rank correlation“, Biometrika, Jg. 30, Nr. 1-2, S. 81–93, 1938.[207]M. G. Kendall, „The treatment of ties in ranking problems“, Biometrika,Jg. 33, Nr. 3, S. 239–251, 1945.[208]L. Puka, „Kendall’s Tau“, in International Encyclopedia of Statistical Science,M. Lovric, Springer Berlin Heidelberg, 2011, S. 713–715.[209]T. Engel und J. Gasteiger, Chemoinformatics: Basic Concepts and Methods,Wiley-VCH, 2018.[210]T. Fawcett, „An introduction to ROC analysis“, Pattern Recognition Letters,Jg. 27, Nr. 8, S. 861–874, 2006.[211]M. Kuhn und K. Johnson, Applied Predictive Modeling. Springer New York,2013.[212]C. M. Florkowski, „Sensitivity, Specificity, Receiver-Operating Characteristic(ROC) Curves and Likelihood Ratios: Communicating the Performance of Dia-gnostic Tests“, The Clinical Biochemist Reviews, Jg. 29, S83–S87, Suppl 12008.[213]J. Davis und M. Goadrich, „The relationship between Precision-Recall andROC curves“, in Proceedings of the 23rd International Conference on MachineLearning - ICML ’06, ACM Press, 2006, S. 233–240.[214]T. Saito und M. Rehmsmeier, „The Precision-Recall Plot Is More Informa-tive than the ROC Plot When Evaluating Binary Classifiers on ImbalancedDatasets“, PLOS ONE, Jg. 10, Nr. 3, G. Brock, e0118432, 2015.[215]A. Gaulton, L. J. Bellis, A. P. Bento, J. Chambers, M. Davies, A. Hersey, Y.Light, S. McGlinchey, D. Michalovich, B. Al-Lazikani und J. P. Overington,„ChEMBL: A large-scale bioactivity database for drug discovery“, NucleicAcids Research, Jg. 40, Nr. D1, S. D1100–D1107, 2012.[216]E. N. Muratov, J. Bajorath, R. P. Sheridan, I. V. Tetko, D. Filimonov, V.Poroikov, T. I. Oprea, I. I. Baskin, A. Varnek, A. Roitberg, O. Isayev, S. Cur-talolo, D. Fourches, Y. Cohen, A. Aspuru-Guzik, D. A. Winkler, D. Agrafiotis,A. Cherkasov und A. Tropsha, „QSAR without borders“, Chemical Society Re-views, Jg. 49, Nr. 11, S. 3525–3564, 2020.[217]H. A. Martens und P. Dardenne, „Validation and verification of regression insmall data sets“, Chemometrics and Intelligent Laboratory Systems, Jg. 44,Nr. 1-2, S. 99–121, 1998.[218]M. Kokabi, M. Donnelly und G. Xu, „Benchmarking Small-Dataset Structure-Activity-Relationship Models for Prediction of Wnt Signaling Inhibition“, IEEEAccess, Jg. 8, S. 228 831–228 840, 2020.[219]G. James, D. Witten, T. Hastie und R. Tibshirani, An Introduction to Sta-tistical Learning, Ser. Springer Texts in Statistics. Springer New York, 2013,Bd. 103.[220]S. Tibshirani und H. Friedman, The Elements of Statistical Learning, Ser. Sprin-ger Texts in Statistics. Springer New York, 2015.Literatur309[221]R. Febrianti, Y. Widyaningsih und S. Soemartojo, „The Parameter Estimationof Logistic Regression with Maximum Likelihood Method and Score FunctionModification“, Journal of Physics: Conference Series, Jg. 1725, S. 012 014,2021.[222]T. Chen und C. Guestrin, „XGBoost: A Scalable Tree Boosting System“, inProceedings of the 22nd ACM SIGKDD International Conference on Know-ledge Discovery and Data Mining, ACM Press, 2016, S. 785–794.[223]Distributed (Deep) Machine Learning Community. „Machine Learning Chall-enge Winning Solutions“. (2021), Adresse: https : / / github . com / dmlc /xgboost / tree / master / demo # machine - learning - challenge - winning -solutions (besucht am 31. 01. 2021).[224]R. P. Sheridan, W. M. Wang, A. Liaw, J. Ma und E. M. Gifford, „Extre-me Gradient Boosting as a Method for Quantitative Structure–Activity Re-lationships“, Journal of Chemical Information and Modeling, Jg. 56, Nr. 12,S. 2353–2360, 2016.[225]I. H. Rekand und R. Brenk, „DrugPred_RNA—A Tool for Structure-BasedDruggability Predictions for RNA Binding Sites“, Journal of Chemical Infor-mation and Modeling, Jg. 61, Nr. 8, S. 4068–4081, 2021.[226]XGBoost developers. „Introduction to Boosted Trees“. (2020), Adresse: https://xgboost.readthedocs.io/en/latest/tutorials/model.html (besuchtam 31. 01. 2021).[227]S. H. Unger und C. Hansch, „Model building in structure-activity relations. Re-examination of adrenergic blocking activity of Beta-Halo-Beta-arylalkylamines“,Journal of Medicinal Chemistry, Jg. 16, Nr. 7, S. 745–749, 1973.[228]A. Tropsha, „Best Practices for QSAR Model Development, Validation, andExploitation“, Molecular Informatics, Jg. 29, Nr. 6-7, S. 476–488, 2010.[229]OECD. „OECD Principles for the Validation, for Regulatory Purposes, of(Quantitative) Structure-Activity Relationship Models“. (2004), Adresse: https://www.oecd.org/chemicalsafety/risk-assessment/37849783.pdf (be-sucht am 30. 01. 2021).[230]W. L. Jorgensen, „QSAR/QSPR and Proprietary Data“, Journal of ChemicalInformation and Modeling, Jg. 46, Nr. 3, S. 937–937, 2006.[231]S. Radestock und H. Gohlke, „Exploiting the Link between Protein Rigidityand Thermostability for Data-Driven Protein Engineering“, Engineering inLife Sciences, Jg. 8, Nr. 5, S. 507–522, 2008.[232]S. Fulle und H. Gohlke, „Analyzing the Flexibility of RNA Structures by Cons-traint Counting“, Biophysical Journal, Jg. 94, Nr. 11, S. 4202–4219, 2008.[233]S. Radestock und H. Gohlke, „Protein rigidity and thermophilic adaptation“,Proteins: Structure, Function, and Bioinformatics, Jg. 79, Nr. 4, S. 1089–1108,2011.[234]C. Pfleger, P. C. Rathi, D. L. Klein, S. Radestock und H. Gohlke, „ConstraintNetwork Analysis (CNA): A Python Software Package for Efficiently Linking310LiteraturBiomacromolecular Structure, Flexibility, (Thermo-)Stability, and Function“,Journal of Chemical Information and Modeling, Jg. 53, Nr. 4, S. 1007–1015,2013.[235]C. Pfleger, S. Radestock, E. Schmidt und H. Gohlke, „Global and local indicesfor characterizing biomolecular flexibility and rigidity“, Journal of Computa-tional Chemistry, Jg. 34, Nr. 3, S. 220–233, 2013.[236]D. J. Jacobs, A. Rader, L. A. Kuhn und M. Thorpe, „Protein flexibility predic-tions using graph theory“, Proteins: Structure, Function, and Genetics, Jg. 44,Nr. 2, S. 150–165, 2001.[237]H. Gohlke, L. A. Kuhn und D. A. Case, „Change in protein flexibility uponcomplex formation: Analysis of Ras-Raf using molecular dynamics and a mole-cular framework approach“, Proteins: Structure, Function, and Bioinforma-tics, Jg. 56, Nr. 2, S. 322–337, 2004.[238]C. Pfleger und H. Gohlke, „Efficient and Robust Analysis of Biomacromolecu-lar Flexibility Using Ensembles of Network Topologies Based on Fuzzy Nonc-ovalent Constraints“, Structure, Jg. 21, Nr. 10, S. 1725–1734, 2013.[239]P. C. Rathi, S. Radestock und H. Gohlke, „Thermostabilizing mutations pre-ferentially occur at structural weak spots with a high mutation ratio“, Journalof Biotechnology, Jg. 159, Nr. 3, S. 135–144, 2012.[240]J. C. Maxwell, „L. On the calculation of the equilibrium and stiffness offrames“, The London, Edinburgh, and Dublin Philosophical Magazine andJournal of Science, Jg. 27, Nr. 182, S. 294–299, 1864.[241]B. Hendrickson, „Conditions for Unique Graph Realizations“, SIAM Journalon Computing, Jg. 21, Nr. 1, S. 65–84, 1992.[242]A. Sljoka, „Counting for rigidity, flexibility and extensions via the pebble gamealgorithm“, Dissertation, York University, 2006.[243]S. M. Hermans, C. Pfleger, C. Nutschel, C. A. Hanke und H. Gohlke, „Rigiditytheory for biomolecules: Concepts, software, and applications: Rigidity theoryfor biomolecules“, Wiley Interdisciplinary Reviews: Computational MolecularScience, Jg. 7, Nr. 4, e1311, 2017.[244]A. Lee, I. Streinu und L. Theran, „Graded Sparse Graphs and Matroids“,Journal of Universal Computer Science, Jg. 13, Nr. 11, S. 1671–1679, 2007.[245]N. M. O’Boyle, M. Banck, C. A. James, C. Morley, T. Vandermeersch undG. R. Hutchison, „Open Babel: An open chemical toolbox“, Journal of Chem-informatics, Jg. 3, Nr. 1, S. 1–14, 2011.[246]N. M. O’Boyle, C. Morley und G. R. Hutchison, „Pybel: A Python wrapperfor the OpenBabel cheminformatics toolkit“, Chemistry Central Journal, Jg. 2,Nr. 1, S. 1–7, 2008.[247]R. Ranganathan, K. P. Lu, T. Hunter und J. P. Noel, „Structural and Func-tional Analysis of the Mitotic Rotamase Pin1 Suggests Substrate RecognitionIs Phosphorylation Dependent“, Cell, Jg. 89, Nr. 6, S. 875–886, 1997.Literatur311[248]E. Bayer, S. Goettsch, J. W. Mueller, B. Griewel, E. Guiberman, L. M. Mayrund P. Bayer, „Structural Analysis of the Mitotic Regulator hPin1 in Solu-tion: Insights into Domain Architecture and Substrate Binding“, Journal ofBiological Chemistry, Jg. 278, Nr. 28, S. 26 183–26 193, 2003.[249]M. Jäger, H. Nguyen, J. C. Crane, J. W. Kelly und M. Gruebele, „The foldingmechanism of a beta-sheet: The WW domain“, Journal of Molecular Biology,Jg. 311, Nr. 2, S. 373–393, 2001.[250]C. M. Davis und R. B. Dyer, „WW Domain Folding Complexity Revealed byInfrared Spectroscopy“, Biochemistry, Jg. 53, Nr. 34, S. 5476–5484, 2014.[251]G. T. Ibragimova und R. C. Wade, „Stability of the Beta-Sheet of the WW Do-main: A Molecular Dynamics Simulation Study“, Biophysical Journal, Jg. 77,Nr. 4, S. 2191–2198, 1999.[252]H. Nguyen, M. Jager, A. Moretto, M. Gruebele und J. W. Kelly, „Tuningthe free-energy landscape of a WW domain by temperature, mutation, andtruncation“, Proceedings of the National Academy of Sciences, Jg. 100, Nr. 7,S. 3948–3953, 2003.[253]J. Karanicolas und C. L. Brooks, „Integrating folding kinetics and proteinfunction: Biphasic kinetics and dual binding specificity in a WW domain“,Proceedings of the National Academy of Sciences, Jg. 101, Nr. 10, S. 3432–3437, 2004.[254]M. Jager, Y. Zhang, J. Bieschke, H. Nguyen, M. Dendle, M. E. Bowman, J. P.Noel, M. Gruebele und J. W. Kelly, „Structure-function-folding relationshipin a WW domain“, Proceedings of the National Academy of Sciences, Jg. 103,Nr. 28, S. 10 648–10 653, 2006.[255]J. Juraszek und P. G. Bolhuis, „(Un)Folding Mechanisms of the FBP28 WWDomain in Explicit Solvent Revealed by Multiple Rare Event Simulation Me-thods“, Biophysical Journal, Jg. 98, Nr. 4, S. 646–656, 2010.[256]J. Karanicolas und C. L. Brooks, „The structural basis for biphasic kineticsin the folding of the WW domain from a formin-binding protein: Lessons forprotein design?“, Proceedings of the National Academy of Sciences, Jg. 100,Nr. 7, S. 3954–3959, 2003.[257]T. Sharpe, A. L. Jonsson, T. J. Rutherford, V. Daggett und A. R. Fersht, „Therole of the turn in �-hairpin formation during WW domain folding“, ProteinScience, Jg. 16, Nr. 10, S. 2233–2239, 2007.[258]Z. Luo, J. Ding und Y. Zhou, „Folding mechanisms of individual �-hairpins in aGō model of Pin1 WW domain by all-atom molecular dynamics simulations“,The Journal of Chemical Physics, Jg. 128, Nr. 22, S. 225 103, 2008.[259]Z. Luo, J. Ding und Y. Zhou, „Temperature-Dependent Folding Pathways ofPin1 WW Domain: An All-Atom Molecular Dynamics Simulation of a GōModel“, Biophysical Journal, Jg. 93, Nr. 6, S. 2152–2161, 2007.[260]H. Gelman und M. Gruebele, „Fast protein folding kinetics“, Quarterly Reviewsof Biophysics, Jg. 47, Nr. 2, S. 95–142, 2014.312Literatur[261]J. Löfblom, J. Feldwisch, V. Tolmachev, J. Carlsson, S. Ståhl und F. Frejd,„Affibody molecules: Engineered proteins for therapeutic, diagnostic and bio-technological applications“, FEBS Letters, Jg. 584, Nr. 12, S. 2670–2680, 2010.[262]C. Eigenbrot, M. Ultsch, A. Dubnovitsky, L. Abrahmsen und T. Hard, „Struc-tural basis for high-affinity HER2 receptor binding by an engineered protein“,Proceedings of the National Academy of Sciences, Jg. 107, Nr. 34, S. 15 039–15 044, 2010.[263]H. Yu, M. K. Rosen und S. L. Schreiber, „1H and 15N assignments and secon-dary structure of the Src SH3 domain“, FEBS Letters, Jg. 324, Nr. 1, S. 87–92,1993.[264]K. L. Stern, M. S. Smith, W. M. Billings, T. J. Loftus, B. M. Conover, D. Del-la Corte und J. L. Price, „Context-Dependent Stabilizing Interactions amongSolvent-Exposed Residues along the Surface of a Trimeric Helix Bundle“, Bio-chemistry, Jg. 59, Nr. 17, S. 1672–1679, 2020.[265]E. O’Shea, J. Klemm, P. Kim und T. Alber, „X-ray structure of the GCN4leucine zipper, a two-stranded, parallel coiled coil“, Science, Jg. 254, Nr. 5031,S. 539–544, 1991.[266]M. S. Weiss und D. Eisenberg’, „The crystal structure of the designed trimericcoiled coil coil-V,Ld: Implications for engineering crystals and supramolecularassemblies“, Protein Science, Jg. 6, Nr. 1, S. 80–88, 1997.[267]M. S. Smith, W. M. Billings, F. G. Whitby, M. B. Miller und J. L. Price,„Enhancing a long-range salt bridge with intermediate aromatic and nonpolaramino acids“, Organic & Biomolecular Chemistry, Jg. 15, Nr. 28, S. 5882–5886,2017.[268]K. Aktories, U. Förstermann, F. Hofmann und K. Starke, Allgemeine UndSpezielle Pharmakologie Und Toxikologie. Elsevier, 2018.[269]L. Zinser und F. Antwerpes. „Peginterferon Alfa - DocCheck Flexikon“. (2021),Adresse: https://flexikon.doccheck.com/de/Peginterferon_alfa (be-sucht am 31. 01. 2021).[270]C. Sarrazin, T. Zimmermann, T. Berg, U. Neumann, P. Schirmacher, H. Schmidt,U. Spengler, J. Timm, H. Wedemeyer, S. Wirth, S. Zeuzem, Deutsche Gesell-schaft für Gastroenterologie, Verdauungs- und Stoffwechselkrankheiten (DGVS),Deutsche Gesellschaft für Pathologie e.V. (DGP), Deutsche Leberstiftung, Ge-sellschaft für Virologie e.V. (GfV), Gesellschaft für Pädiatrische Gastroente-rologie und Ernährung (GPGE), Österreichische Gesellschaft für Gastroente-rologie und Hepatologie (ÖGGH), Schweizerische Gesellschaft für Gastroente-rologie (SGG), Deutsche Transplantationsgesellschaft e.V. (DTG), DeutscheLeberhilfe e.V., Deutsche Gesellschaft für Infektiologie e.V. (DGI), DeutscheGesellschaft für Suchtmedizin e.V. (DGS), Deutsche AIDS-Gesellschaft e.V.(DAIG), Deutsche Arbeitsgemeinschaft niedergelassener Ärzte für die Versor-gung HIV-Infizierter (DAGNÄ) und Robert Koch-Institut, „S3-Leitlinie „Pro-phylaxe, Diagnostik und Therapie der Hepatitis-C-Virus (HCV) -Infektion“:Literatur313AWMF-Register-Nr.: 021/012“, Zeitschrift für Gastroenterologie, Jg. 56, Nr. 07,S. 756–838, 2018.[271]M. T. Dill, Z. Makowska, G. Trincucci, A. J. Gruber, J. E. Vogt, M. Filipowicz,D. Calabrese, I. Krol, D. T. Lau, L. Terracciano, E. van Nimwegen, V. Rothund M. H. Heim, „Pegylated IFN-� regulates hepatic gene expression throughtransient Jak/STAT activation“, Journal of Clinical Investigation, Jg. 124,Nr. 4, S. 1568–1581, 2014.[272]C. Thomas, I. Moraga, D. Levin, P. O. Krutzik, Y. Podoplelova, A. Trejo, C.Lee, G. Yarden, S. E. Vleck, J. S. Glenn, G. P. Nolan, J. Piehler, G. Schreiberund K. C. Garcia, „Structural Linkage between Ligand Discrimination andReceptor Activation by Type I Interferons“, Cell, Jg. 146, Nr. 4, S. 621–632,2011.[273]W. Klaus, B. Gsell, A. M. Labhardt, B. Wipf und H. Senn, „The three-dimensional high resolution structure of human interferon �-2a determined byheteronuclear NMR spectroscopy in solution“, Journal of Molecular Biology,Jg. 274, Nr. 4, S. 661–675, 1997.[274]MOE, Molecular Operating Environment (MOE), 2020.01; Chemical Compu-ting Group ULC, 1010 Sherbooke St. West, Suite #910, Montreal, QC, Canada,H3A 2R7, 2020.[275]C. I. Bayly, P. Cieplak, W. Cornell und P. A. Kollman, „A well-behaved elec-trostatic potential based method using charge restraints for deriving atomiccharges: The RESP model“, The Journal of Physical Chemistry, Jg. 97, Nr. 40,S. 10 269–10 280, 1993.[276]F.-Y. Dupradeau, A. Pigache, T. Zaffran, C. Savineau, R. Lelong, N. Grivel,D. Lelong, W. Rosanski und P. Cieplak, „The R.E.D. tools: Advances in RESPand ESP charge derivation and force field library building“, Physical ChemistryChemical Physics, Jg. 12, Nr. 28, S. 7821–7839, 2010.[277]E. Vanquelef, S. Simon, G. Marquant, E. Garcia, G. Klimerak, J. C. Delepine,P. Cieplak und F.-Y. Dupradeau, „R.E.D. Server: A web service for derivingRESP and ESP charges and building force field libraries for new molecules andmolecular fragments“, Nucleic Acids Research, Jg. 39, W511–W517, suppl_22011.[278]F. Wang, J.-P. Becker, P. Cieplak und F.-Y. Dupradeau, „R.E.D. Python:Object Oriented Programming for Amber Force Fields“, in Abstracts of Papersof the American Chemical Society, Bd. 247, American Chemical Society, 2014.[279]P. R. Gerber und K. Müller, „MAB, a generally applicable molecular force fieldfor structure modelling in medicinal chemistry“, Journal of Computer-AidedMolecular Design, Jg. 9, Nr. 3, S. 251–268, 1995.[280]M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R.Cheeseman, G. Scalmani, V. Barone, G. A. Petersson, H. Nakatsuji, X. Li, M.Caricato, A. V. Marenich, J. Bloino, B. G. Janesko, R. Gomperts, B. Mennucci,H. P. Hratchian, J. V. Ortiz, A. F. Izmaylov, J. L. Sonnenberg, Williams, F.314LiteraturDing, F. Lipparini, F. Egidi, J. Goings, B. Peng, A. Petrone, T. Henderson,D. Ranasinghe, V. G. Zakrzewski, J. Gao, N. Rega, G. Zheng, W. Liang, M.Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima,Y. Honda, O. Kitao, H. Nakai, T. Vreven, K. Throssell, J. A. MontgomeryJr., J. E. Peralta, F. Ogliaro, M. J. Bearpark, J. J. Heyd, E. N. Brothers,K. N. Kudin, V. N. Staroverov, T. A. Keith, R. Kobayashi, J. Normand, K.Raghavachari, A. P. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi,J. M. Millam, M. Klene, C. Adamo, R. Cammi, J. W. Ochterski, R. L. Martin,K. Morokuma, O. Farkas, J. B. Foresman und D. J. Fox, Gaussian 16 Rev.B.01, Wallingford, CT, 2016.[281]M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R.Cheeseman, G. Scalmani, V. Barone, G. A. Petersson, H. Nakatsuji, X. Li, M.Caricato, A. V. Marenich, J. Bloino, B. G. Janesko, R. Gomperts, B. Mennucci,H. P. Hratchian, J. V. Ortiz, A. F. Izmaylov, J. L. Sonnenberg, Williams, F.Ding, F. Lipparini, F. Egidi, J. Goings, B. Peng, A. Petrone, T. Henderson,D. Ranasinghe, V. G. Zakrzewski, J. Gao, N. Rega, G. Zheng, W. Liang, M.Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima,Y. Honda, O. Kitao, H. Nakai, T. Vreven, K. Throssell, J. A. MontgomeryJr., J. E. Peralta, F. Ogliaro, M. J. Bearpark, J. J. Heyd, E. N. Brothers,K. N. Kudin, V. N. Staroverov, T. A. Keith, R. Kobayashi, J. Normand, K.Raghavachari, A. P. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi,J. M. Millam, M. Klene, C. Adamo, R. Cammi, J. W. Ochterski, R. L. Martin,K. Morokuma, O. Farkas, J. B. Foresman und D. J. Fox, Gaussian 09 Rev.C.01, Wallingford, CT, 2009.[282]D.A. Case, I.Y. Ben-Shalom, S.R. Brozell, D.S. Cerutti, T.E. Cheatham, III,V.W.D. Cruzeiro, T.A. Darden, R.E. Duke, D. Ghoreishi, M.K. Gilson, H.Gohlke, A.W. Goetz, D. Greene, R Harris, N. Homeyer, S. Izadi, A. Kovalenko,T. Kurtzman, T.S. Lee, S. LeGrand, P. Li, C. Lin, J. Liu, T. Luchko, R. Luo,D.J. Mermelstein, K.M. Merz, Y. Miao, G. Monard, C. Nguyen, H. Nguyen,I. Omelyan, A. Onufriev, R. Salomon-Ferrer, J. Swails, R.C. Walker, J. Wang,H. Wei, R.M. Wolf, X. Wu, L. Xiao, D.M. York und P.A. Kollman, AMBER2018, University of California, San Francisco, 2018.[283]J. Wang, W. Wang, P. A. Kollman und D. A. Case, „Antechamber, An Acces-sory Software Package For Molecular Mechanical Calculations“, Submitted toJournal of Chemical Information and Computer Sciences, 2000.[284]A. T. P. Carvalho, P. A. Fernandes und M. J. Ramos, „Parameterization ofAZT—A widely used nucleoside inhibitor of HIV-1 reverse transcriptase“, In-ternational Journal of Quantum Chemistry, Jg. 107, Nr. 2, S. 292–298, 2007.[285]G. Pieffet und P. A. Petukhov, „Parameterization of aromatic azido groups:Application as photoaffinity probes in molecular dynamics studies“, Journalof Molecular Modeling, Jg. 15, Nr. 11, S. 1291–1297, 2009.Literatur315[286]R. S. Mulliken, „Electronic Population Analysis on LCAO–MO Molecular Wa-ve Functions. I“, The Journal of Chemical Physics, Jg. 23, Nr. 10, S. 1833–1840, 1955.[287]C. Tian, K. Kasavajhala, K. A. A. Belfon, L. Raguette, H. Huang, A. N.Migues, J. Bickel, Y. Wang, J. Pincay, Q. Wu und C. Simmerling, „ff19SB:Amino-Acid-Specific Protein Backbone Parameters Trained against QuantumMechanics Energy Surfaces in Solution“, Journal of Chemical Theory andComputation, Jg. 16, Nr. 1, S. 528–552, 2020.[288]W. L. Jorgensen, J. Chandrasekhar, J. D. Madura, R. W. Impey und M. L.Klein, „Comparison of simple potential functions for simulating liquid water“,The Journal of Chemical Physics, Jg. 79, Nr. 2, S. 926–935, 1983.[289]J. C. Phillips, R. Braun, W. Wang, J. Gumbart, E. Tajkhorshid, E. Villa, C.Chipot, R. D. Skeel, L. Kalé und K. Schulten, „Scalable molecular dynamicswith NAMD“, Journal of Computational Chemistry, Jg. 26, Nr. 16, S. 1781–1802, 2005.[290]T. D. Romo, N. Leioatts und A. Grossfield, „Lightweight object oriented struc-ture analysis: Tools for building tools to analyze molecular dynamics simula-tions“, Journal of Computational Chemistry, Jg. 35, Nr. 32, S. 2305–2318,2014.[291]T. Romo und A. Grossfield, „LOOS: An Extensible Platform for the StructuralAnalysis of Simulations“, in 2009 Annual International Conference of the IEEEEngineering in Medicine and Biology Society, Minneapolis, MN: IEEE, 2009,S. 2332–2335.[292]M. Kuhn, „Building Predictive Models in R using the caret package“, Journalof Statistical Software, Jg. 28, Nr. 1, S. 1–26, 2008.[293]RStudio. „Max Kuhn - Profil“. (2021), Adresse: https://www.rstudio.com/speakers/max-kuhn/ (besucht am 31. 01. 2021).[294]I. Guyon, J. Weston und S. Barnhill, „Gene Selection for Cancer Classificationusing Support Vector Machines“, Machine Learning, Jg. 46, Nr. 1, S. 389–422,2002.[295]S. Kirkpatrick, C. D. Gelatt Jr. und M. P. Vecchi, „Optimization by SimulatedAnnealing“, Science, Jg. 220, Nr. 4598, S. 671–680, 1983.[296]L. Scrucca, „GA : A Package for Genetic Algorithms in R“, Journal of Statis-tical Software, Jg. 53, Nr. 1, S. 1–137, 2013.[297]P. Carrió, M. Pinto, G. Ecker, F. Sanz und M. Pastor, „Applicability DomainAnalysis (ADAN): A Robust Method for Assessing the Reliability of DrugProperty Predictions“, Journal of Chemical Information and Modeling, Jg. 54,Nr. 5, S. 1500–1511, 2014.[298]P. Filzmoser, B. Liebmann und K. Varmuza, „Repeated double cross validati-on“, Journal of Chemometrics, Jg. 23, Nr. 4, S. 160–171, 2009.[299]K. Varmuza und P. Filzmoser, Introduction to Multivariate Statistical Analysisin Chemometrics. CRC Press, 2009.316Literatur[300]R. L. Thorndike, „Who belongs in the family?“, Psychometrika, Jg. 18, Nr. 4,S. 267–276, 1953.[301]PLUMED Manual. „PLUMED: UPPER_WALLS“. (2021), Adresse: https://www.plumed.org/doc-v2.7/user-doc/html/_u_p_p_e_r__w_a_l_l_s.html(besucht am 26. 11. 2021).[302]D. Trapl und V. Spiwok, „Analysis of the Results of Metadynamics Simulationsby Metadynminer and Metadynminer3d“, arXiv Preprint (arXiv:2009.02241),2020.[303]S. Foser, A. Schacher, K. A. Weyer, D. Brugger, E. Dietel, S. Marti und T.Schreitmüller, „Isolation, structural characterization, and antiviral activity ofpositional isomers of monopegylated interferon �-2a (PEGASYS)“, Protein Ex-pression and Purification, Jg. 30, Nr. 1, S. 78–87, 2003.[304]N. Hauptstein, P. Pouyan, J. Kehrein, M. Dirauf, M. D. Driessen, M. Raschig,K. Licha, M. Gottschaldt, U. S. Schubert, R. Haag, L. Meinel, C. Sotriffer undT. Lühmann, „Molecular Insights into Site-Specific Interferon-�2a Bioconju-gates Originated from PEG, LPG, and PEtOx“, Biomacromolecules, Jg. 22,Nr. 11, S. 4521–4534, 2021.[305]Schrödinger Release 2018-1: Epik, Schrödinger, LLC, New York, NY, 2018.[306]S. D. Bond, B. J. Leimkuhler und B. B. Laird, „The Nosé–Poincaré Methodfor Constant Temperature Molecular Dynamics“, Journal of ComputationalPhysics, Jg. 151, Nr. 1, S. 114–134, 1999.[307]J. B. Sturgeon und B. B. Laird, „Symplectic algorithm for constant-pressuremolecular dynamics using a Nosé–Poincaré thermostat“, The Journal of Che-mical Physics, Jg. 112, Nr. 8, S. 3474–3482, 2000.[308]W. C. Still, A. Tempczyk, R. C. Hawley und T. Hendrickson, „Semianalyticaltreatment of solvation for molecular mechanics and dynamics“, Journal of theAmerican Chemical Society, Jg. 112, Nr. 16, S. 6127–6129, 1990.[309]A. Munasinghe, A. Mathavan, A. Mathavan, P. Lin und C. M. Colina, „Ato-mistic insight towards the impact of polymer architecture and grafting densityon structure-dynamics of PEGylated bovine serum albumin and their appli-cations“, The Journal of Chemical Physics, Jg. 154, Nr. 7, S. 075 101, 2021.[310]M. E. Fortunato und C. M. Colina, „Pysimm: A python package for simulationof molecular systems“, SoftwareX, Jg. 6, S. 7–12, 2017.[311]M. E. Fortunato und C. M. Colina. „Pysimm“. (2021), Adresse: https://github.com/polysimtools/pysimm. (besucht am 31. 01. 2021).[312]S. J. Marrink, H. J. Risselada, S. Yefimov, D. P. Tieleman und A. H. deVries, „The MARTINI Force Field: Coarse Grained Model for BiomolecularSimulations“, The Journal of Physical Chemistry B, Jg. 111, Nr. 27, S. 7812–7824, 2007.[313]F. Grunewald, G. Rossi, A. H. de Vries, S. J. Marrink und L. Monticelli,„Transferable MARTINI Model of Poly(ethylene Oxide)“, The Journal of Phy-sical Chemistry B, Jg. 122, Nr. 29, S. 7436–7449, 2018.Literatur317[314]F. Grünewald, R. Alessandri, P. C. Kroon, L. Monticelli, P. C. T. Souza undS. J. Marrink, „Polyply: A Python Suite for Facilitating Simulations of (Bio-)Macromolecules and Nanomaterials“, arXiv Preprint (arXiv:2105.05890), 2021.[315]Invivogen. „IFN-Alpha/Beta Reporter HEK 293 Cells“. (2021), Adresse: https://www.invivogen.com/hek-blue-ifn-ab (besucht am 19. 08. 2021).[316]Y. Feng, G. De Franceschi, A. Kahraman, M. Soste, A. Melnik, P. J. Boersema,P. P. de Laureto, Y. Nikolaev, A. P. Oliveira und P. Picotti, „Global analysisof protein structural changes in complex proteomes“, Nature Biotechnology,Jg. 32, Nr. 10, S. 1036–1044, 2014.[317]K. M. Merz, „Limits of Free Energy Computation for Protein−Ligand Inter-actions“, Journal of Chemical Theory and Computation, Jg. 6, Nr. 5, S. 1769–1776, 2010.[318]B. K. Shoichet, W. P. Walters, H. Jiang und J. Bajorath, „Advances in Compu-tational Medicinal Chemistry: A Reflection on the Evolution of the Field andPerspective Going Forward“, Journal of Medicinal Chemistry, Jg. 59, Nr. 9,S. 4033–4034, 2016.[319]M. Gevrey, I. Dimopoulos und S. Lek, „Review and comparison of methodsto study the contribution of variables in artificial neural network models“,Ecological Modelling, Jg. 160, Nr. 3, S. 249–264, 2003.[320]J. Heaton, Introduction to Neural Networks with Java. 2008.[321]G. Henkelman und H. Jónsson, „Improved tangent estimate in the nudgedelastic band method for finding minimum energy paths and saddle points“,The Journal of Chemical Physics, Jg. 113, Nr. 22, S. 9978–9985, 2000.[322]L. Hahn, T. Zorn, J. Kehrein, T. Kielholz, A.-L. Ziegler, B. Sochor, E. S. Lisit-syna, N. A. Durandin, T. Laaksonen, V. Aseyev, C. Sotriffer, M. Windbergs,A.-C. Pöppler und R. Luxenhofer, „Unravelling a novel mechanism in polymerself-assemblies: An order-order transition based on molecular interactions bet-ween hydrophilic and hydrophobic polymer blocks“, ChemRxiv Preprint (DOI:10.33774/chemrxiv-2021-mjd8n-v2), 2021.[323]R. W. Newberry und R. T. Raines, „The n →�* Interaction“, Accounts ofChemical Research, Jg. 50, Nr. 8, S. 1838–1846, 2017.[324]B. C. Gorske, B. L. Bastian, G. D. Geske und H. E. Blackwell, „Local andTunable n→�* Interactions Regulate Amide Isomerism in the Peptoid Back-bone“, Journal of the American Chemical Society, Jg. 129, Nr. 29, S. 8928–8929, 2007.[325]A.-C. Pöppler, M. M. Lübtow, J. Schlauersbach, J. Wiest, L. Meinel und R.Luxenhofer, „Loading dependent Structural Model of Polymeric Micelles En-capsulating Curcumin by Solid-State NMR Spectroscopy“, Angewandte Che-mie International Edition, 2019.[326]W. Humphrey, A. Dalke und K. Schulten, „VMD: Visual Molecular Dyna-mics“, J. Mol. Graphics, Jg. 14, S. 33–36, 1996.318Literatur[327]F. Müller‐Plathe und W. F. van Gunsteren, „Computer simulation of a po-lymer electrolyte: Lithium iodide in amorphous poly(ethylene oxide)“, TheJournal of Chemical Physics, Jg. 103, Nr. 11, S. 4745–4756, 1995.[328]K. Laasonen und M. L. Klein, „Molecular dynamics simulations of the structu-re and ion diffusion in poly(ethylene oxide)“, Journal of the Chemical Society,Faraday Transactions, Jg. 91, Nr. 16, S. 2633, 1995.[329]F. Ramezanghorbani, P. Lin und C. M. Colina, „Optimizing Protein–PolymerInteractions in a Poly(ethylene glycol) Coarse-Grained Model“, The Journalof Physical Chemistry B, Jg. 122, Nr. 33, S. 7997–8005, 2018.[330]S. S. Pai, B. Hammouda, K. Hong, D. C. Pozzo, T. M. Przybycien undR. D. Tilton, „The Conformation of the Poly(ethylene glycol) Chain in Mono-PEGylated Lysozyme and Mono-PEGylated Human Growth Hormone“, Bio-conjugate Chemistry, Jg. 22, Nr. 11, S. 2317–2323, 2011.[331]C. Andraud, A. Beghdadi und J. Lafait, „Entropic analysis of random morpho-logies“, Physica A: Statistical Mechanics and its Applications, Jg. 207, Nr. 1-3,S. 208–212, 1994.[332]G.-J. Bekker, B. Ma und N. Kamiya, „Thermal stability of single‐domain anti-bodies estimated by molecular dynamics simulations“, Protein Science, Jg. 28,Nr. 2, S. 429–438, 2019.[333]N. Hauptstein, P. Pouyan, K. Wittwer, G. Cinar, O. Scherf-Clavel, K. Licha,T. Lühmann, I. Nischang, U. S. Schubert, C. Pfaller, R. Haag und L. Meinel,„Polymer selection impacts the pharmaceutical profile of site specifically con-jugated Interferon-alpha2a“, ChemRxiv Preprint (DOI: 10.26434/chemrxiv-2022-j85br), 2022.[334]N. Hauptstein, M. Dirauf, K. Wittwer, B. Sawatsky, G. Cinar, O. Scherf-Clavel, M. Raschig, T. Lühmann, I. Nischang, U. S. Schubert, C. Pfaller undL. Meinel, „Site Specific PEtOxylation of Interferon-Alpha2a“, unpubliziert(eingereicht), 2022.[335]T. Murtola, A. Bunker, I. Vattulainen, M. Deserno und M. Karttunen, „Mul-tiscale modeling of emergent materials: Biological and soft matter“, PhysicalChemistry Chemical Physics, Jg. 11, Nr. 12, S. 1869, 2009.[336]D. Kozakov, D. R. Hall, B. Xia, K. A. Porter, D. Padhorny, C. Yueh, D. Beglovund S. Vajda, „The ClusPro web server for protein–protein docking“, NatureProtocols, Jg. 12, Nr. 2, S. 255–278, 2017.[337]K. A. Porter, I. Desta, D. Kozakov und S. Vajda, „What method to use forprotein–protein docking?“, Current Opinion in Structural Biology, Jg. 55, S. 1–7, 2019.[338]I. A. Vakser, „Protein-Protein Docking: From Interaction to Interactome“,Biophysical Journal, Jg. 107, Nr. 8, S. 1785–1793, 2014.[339]Y. Yan, H. Tao, J. He und S.-Y. Huang, „The HDOCK server for integratedprotein–protein docking“, Nature Protocols, Jg. 15, Nr. 5, S. 1829–1852, 2020.Literatur319[340]J. Estrada, P. Bernado, M. Blackledge und J. Sancho, „ProtSA: A web ap-plication for calculating sequence specific protein solvent accessibilities in theunfolded ensemble“, BMC Bioinformatics, Jg. 10, Nr. 1, S. 104, 2009.[341]O. Conchillo-Solé, N. S. de Groot, F. X. Avilés, J. Vendrell, X. Daura und S.Ventura, „AGGRESCAN: A server for the prediction and evaluation of ”hotspotsöf aggregation in polypeptides“, BMC Bioinformatics, Jg. 8, Nr. 1, S. 65,2007.[342]R. Zambrano, M. Jamroz, A. Szczasiuk, J. Pujols, S. Kmiecik und S. Ventura,„AGGRESCAN3D (A3D): Server for prediction of aggregation properties ofprotein structures“, Nucleic Acids Research, Jg. 43, Nr. W1, W306–W313,2015.[343]M. Gil-Garcia, M. Bañó-Polo, N. Varejão, M. Jamroz, A. Kuriata, M. Díaz-Caballero, J. Lascorz, B. Morel, S. Navarro, D. Reverter, S. Kmiecik und S.Ventura, „Combining Structural Aggregation Propensity and Stability Predic-tions To Redesign Protein Solubility“, Molecular Pharmaceutics, Jg. 15, Nr. 9,S. 3846–3859, 2018.[344]T. Zeiske, K. A. Stafford und A. G. Palmer, „Thermostability of Enzymes fromMolecular Dynamics Simulations“, Journal of Chemical Theory and Compu-tation, Jg. 12, Nr. 6, S. 2489–2492, 2016.[345]S. Park, F. Khalili-Araghi, E. Tajkhorshid und K. Schulten, „Free energy cal-culation from steered molecular dynamics simulations using Jarzynski’s equa-lity“, The Journal of Chemical Physics, Jg. 119, Nr. 6, S. 3559–3566, 2003.[346]S. Decherchi und A. Cavalli, „Thermodynamics and Kinetics of Drug-TargetBinding by Molecular Simulation“, Chemical Reviews, Jg. 120, Nr. 23, S. 12 788–12 833, 2020.[347]S. Takada, „Gō model revisited“, Biophysics and Physicobiology, Jg. 16, Nr. 0,S. 248–255, 2019.[348]K. M. Wilding, A. K. Smith, J. W. Wilkerson, D. B. Bush, T. A. Knotts undB. C. Bundy, „The Locational Impact of Site-Specific PEGylation: StreamlinedScreening with Cell-Free Protein Expression and Coarse-Grain Simulation“,ACS Synthetic Biology, Jg. 7, Nr. 2, S. 510–521, 2018.[349]M. Feig, J. Karanicolas und C. L. Brooks, „MMTSB Tool Set: Enhanced sam-pling and multiscale modeling methods for applications in structural biology“,Journal of Molecular Graphics and Modelling, Jg. 22, Nr. 5, S. 377–395, 2004.[350]B. R. Brooks, C. L. Brooks, A. D. Mackerell, L. Nilsson, R. J. Petrella, B.Roux, Y. Won, G. Archontis, C. Bartels, S. Boresch, A. Caflisch, L. Caves,Q. Cui, A. R. Dinner, M. Feig, S. Fischer, J. Gao, M. Hodoscek, W. Im, K.Kuczera, T. Lazaridis, J. Ma, V. Ovchinnikov, E. Paci, R. W. Pastor, C. B.Post, J. Z. Pu, M. Schaefer, B. Tidor, R. M. Venable, H. L. Woodcock, X.Wu, W. Yang, D. M. York und M. Karplus, „CHARMM: The biomolecularsimulation program“, Journal of Computational Chemistry, Jg. 30, Nr. 10,S. 1545–1614, 2009.320Literatur[351]C. Kobayashi, J. Jung, Y. Matsunaga, T. Mori, T. Ando, K. Tamura, M. Ka-miya und Y. Sugita, „GENESIS 1.1: A hybrid-parallel molecular dynamicssimulator with enhanced sampling algorithms on multiple computational plat-forms“, Journal of Computational Chemistry, Jg. 38, Nr. 25, S. 2193–2206,2017.[352]A. Bakan, L. M. Meireles und I. Bahar, „ProDy: Protein Dynamics Inferredfrom Theory and Experiments“, Bioinformatics, Jg. 27, Nr. 11, S. 1575–1577,2011.[353]M. D. Ward, S. Nangia und E. R. May, „Evaluation of the hybrid resolutionPACE model for the study of folding, insertion, and pore formation of membra-ne associated peptides“, Journal of Computational Chemistry, Jg. 38, Nr. 16,S. 1462–1471, 2017.[354]V. Ovchinnikov, S. Conti, E. Y. Lau, F. C. Lightstone und M. Karplus, „Micro-second Molecular Dynamics Simulations of Proteins Using a Quasi-EquilibriumSolvation Shell Model“, Journal of Chemical Theory and Computation, Jg. 16,Nr. 3, S. 1866–1881, 2020.[355]T. A. Beu, A.-E. Ailenei und R.-I. Costinaş, „Martini Force Field for Proto-nated Polyethyleneimine“, Journal of Computational Chemistry, Jg. 41, Nr. 4,S. 349–361, 2020.[356]T. A. Wassenaar, K. Pluhackova, R. A. Böckmann, S. J. Marrink und D. P. Tie-leman, „Going Backward: A Flexible Geometric Approach to Reverse Trans-formation from Coarse Grained to Atomistic Models“, Journal of ChemicalTheory and Computation, Jg. 10, Nr. 2, S. 676–690, 2014.[357]Horvath, Marcou und Varnek, „Generative Topographic Mapping of the Do-cking Conformational Space“, Molecules, Jg. 24, Nr. 12, S. 2269, 2019.[358]D. Horvath, I. Baskin, G. Marcou und A. Varnek, „Generative TopographicMapping of Conformational Space“, Molecular Informatics, Jg. 36, Nr. 10,S. 1 700 036, 2017.[359]F. Noé, S. Olsson, J. Köhler und H. Wu, „Boltzmann generators: Samplingequilibrium states of many-body systems with deep learning“, Science, Jg. 365,Nr. 6457, eaaw1147, 2019.[360]A. Tucs, K. Tsuda und A. Sljoka, „Probing conformational dynamics of an-tibodies with geometric simulations“, arXiv Preprint (arXiv:2109.14724v1),S. 13, 2021.[361]D. L. Allen und G. J. Pielak, „Baseline length and automated fitting of dena-turation data“, Protein Science, Jg. 7, Nr. 5, S. 1262–1263, 1998.[362]D. W. Marquardt, „An Algorithm for Least-Squares Estimation of NonlinearParameters“, Journal of the Society for Industrial and Applied Mathematics,Jg. 11, Nr. 2, S. 431–441, 1963.[363]D. R. Roe und T. E. Cheatham, „PTRAJ and CPPTRAJ: Software for Pro-cessing and Analysis of Molecular Dynamics Trajectory Data“, Journal ofChemical Theory and Computation, Jg. 9, Nr. 7, S. 3084–3095, 2013.Literatur321[364]B. Knapp, S. Frantal, M. Cibena, W. Schreiner und P. Bauer, „Is an Intui-tive Convergence Definition of Molecular Dynamics Simulations Solely Basedon the Root Mean Square Deviation Possible?“, Journal of ComputationalBiology, Jg. 18, Nr. 8, S. 997–1005, 2011.[365]B. Knapp, N. Lederer, U. Omasits und W. Schreiner, „vmdICE: A plug-in forrapid evaluation of molecular dynamics simulations using VMD“, Journal ofComputational Chemistry, Jg. 31, Nr. 16, S. 2868–2873, 2010.[366]R. Weiser, P. S. Shenkin und W. C. Still, „Approximate atomic surfaces fromlinear combinations of pairwise overlaps (LCPO)“, Journal of ComputationalChemistry, Jg. 20, Nr. 2, S. 217–230, 1999.[367]M. L. Connolly, „Analytical Molecular Surface Calculation“, Journal of AppliedCrystallography, Jg. 16, Nr. 5, S. 548–558, 1983.[368]R. Stepto, T. Chang, P. Kratochvíl, M. Hess, K. Horie, T. Sato und J. Vohlídal,„Definitions of Terms Relating to Individual Macromolecules, MacromolecularAssemblies, Polymer Solutions, and Amorphous Bulk Polymers (IUPAC Re-commendations 2014)“, Pure and Applied Chemistry, Jg. 87, Nr. 1, S. 71–120,2015.[369]Z. Xinyi, Synchrotron Radiation Applications. World Scientific, 2018.[370]G. Ramachandran, C. Ramakrishnan und V. Sasisekharan, „Stereochemistryof polypeptide chain configurations“, Journal of Molecular Biology, Jg. 7, Nr. 1,S. 95–99, 1963.[371]J. J. Prompers und R. Brüschweiler, „General Framework for Studying theDynamics of Folded and Nonfolded Proteins by NMR Relaxation Spectroscopyand MD Simulation“, Journal of the American Chemical Society, Jg. 124,Nr. 16, S. 4522–4534, 2002.[372]C. A. MacRaild, A. H. Daranas, A. Bronowska und S. W. Homans, „GlobalChanges in Local Protein Dynamics Reduce the Entropic Cost of Carbohydra-te Binding in the Arabinose-binding Protein“, Journal of Molecular Biology,Jg. 368, Nr. 3, S. 822–832, 2007.[373]G. Lipari und A. Szabo, „Model-free approach to the interpretation of nuclearmagnetic resonance relaxation in macromolecules. 1. Theory and range of va-lidity“, Journal of the American Chemical Society, Jg. 104, Nr. 17, S. 4546–4559, 1981.[374]M. Clarkson. „The Model-Free Dynamics Formalism of Lipari and Szabo“.(2007), Adresse: http://conflux.mwclarkson.com/2007/11/the-model-free-dynamics-formalism-of-lipari-and-szabo/ (besucht am 12. 12. 2020).[375]C. Kersten, „Selectivity Determining Features in N-Myristoyltransferases – AModel System for Drug Targets with Conserved Binding Sites“, Dissertation,Johannes Gutenberg-Universität Mainz, 2017.[376]D.-W. Li und R. Brüschweiler, „A Dictionary for Protein Side-Chain Entropiesfrom NMR Order Parameters“, Journal of the American Chemical Society,Jg. 131, Nr. 21, S. 7226–7227, 2009.[377]R. B. Best, G. Hummer und W. A. Eaton, „Native contacts determine prote-in folding mechanisms in atomistic simulations“, Proceedings of the NationalAcademy of Sciences, Jg. 110, Nr. 44, S. 17 874–17 879, 2013.[378]N. Michaud-Agrawal, E. J. Denning, T. B. Woolf und O. Beckstein, „MDAna-lysis: A toolkit for the analysis of molecular dynamics simulations“, Journalof Computational Chemistry, Jg. 32, Nr. 10, S. 2319–2327, 2011.[379]R. Gowers, M. Linke, J. Barnoud, T. Reddy, M. Melo, S. Seyler, J. Domański,D. Dotson, S. Buchoux, I. Kenney und O. Beckstein, „MDAnalysis: A PythonPackage for the Rapid Analysis of Molecular Dynamics Simulations“, Pythonin Science Conference, Austin, Texas, 2016, S. 98–105.[380]R. J. Gowers und P. Carbone, „A multiscale approach to model hydrogenbonding: The case of polyamide“, The Journal of Chemical Physics, Jg. 142,Nr. 22, S. 224 907, 2015.[381]S. Genheden, M. Akke und U. Ryde, „Conformational Entropies and OrderParameters: Convergence, Reproducibility, and Transferability“, Journal ofChemical Theory and Computation, Jg. 10, Nr. 1, S. 432–438, 2014.[382]M. Kuhn. „Available Models“. (2021), Adresse: https://topepo.github.io/caret/available-models.html (besucht am 27. 08. 2021). 323Abbildungsverzeichnis1.1Grafting-Methoden . . . . . . . . . . . . . . . . . . . . . . . . . . . . .61.2Cyclooctin-Konjugationen . . . . . . . . . . . . . . . . . . . . . . . . .91.3Polymere. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .121.4Faltungsprozess . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .151.5Price-Winkel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .181.6PEG-Lysin-Wechselwirkung . . . . . . . . . . . . . . . . . . . . . . . .191.7Kraftfeld-Terme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .221.8Simulationsansätze . . . . . . . . . . . . . . . . . . . . . . . . . . . . .261.9Konfusionsmatrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .361.10 Pin 1-Struktur. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .451.11 Her2-Affibody-Struktur. . . . . . . . . . . . . . . . . . . . . . . . . .461.12 Src SH3-Struktur . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .471.13 GCN4/coilVaLd-Struktur. . . . . . . . . . . . . . . . . . . . . . . . .471.14 IFN-α2a-Struktur . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .483.1Pin 1-Konjugate. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .533.2Zusätzliche Pin 1-Konjugate . . . . . . . . . . . . . . . . . . . . . . . .543.3PyRED-Methode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .553.4antechamber-Methode . . . . . . . . . . . . . . . . . . . . . . . . . . .563.5Monomerbasierte Parametrisierung . . . . . . . . . . . . . . . . . . . .573.6Interaktionsprofil-Vergleich 1. . . . . . . . . . . . . . . . . . . . . . .593.7Interaktionsprofil-Vergleich 2. . . . . . . . . . . . . . . . . . . . . . .603.8Interaktionsprofil-Vergleich 3. . . . . . . . . . . . . . . . . . . . . . .613.9RDF-Vergleich. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .623.10 SCC-Kriterien . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .653.11 QSPR-Modellgenerierung Ansatz 1 . . . . . . . . . . . . . . . . . . . .673.12 Datensatzverteilung Ansatz 1 . . . . . . . . . . . . . . . . . . . . . . .683.13 k-Means-Clustering . . . . . . . . . . . . . . . . . . . . . . . . . . . . .733.14 Clustering-Beispiel . . . . . . . . . . . . . . . . . . . . . . . . . . . . .743.15 QSPR-Modellgenerierung Ansatz 2 . . . . . . . . . . . . . . . . . . . .753.16 Kollektivvariablen. . . . . . . . . . . . . . . . . . . . . . . . . . . . .773.17 Positionierung von K31 und K134. . . . . . . . . . . . . . . . . . . .793.18 Cyclooctin-Linker-Strukturen . . . . . . . . . . . . . . . . . . . . . . .803.19 GaMDs - Modellierung . . . . . . . . . . . . . . . . . . . . . . . . . . .813.20 CG-Startstruktur . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .83324Abbildungsverzeichnis3.21 Experimentelle Arbeiten zu Interferon-α2a . . . . . . . . . . . . . . . .844.1QSPR-Ansatz 1 - Modell C1 . . . . . . . . . . . . . . . . . . . . . . . .884.2QSPR-Ansatz 1 - Modell D1. . . . . . . . . . . . . . . . . . . . . . .894.3QSPR-Ansatz 1 - Modell E. . . . . . . . . . . . . . . . . . . . . . . .904.4QSPR-Ansatz 1 - Modell F. . . . . . . . . . . . . . . . . . . . . . . .914.5QSPR-Ansatz 1 - Modell G . . . . . . . . . . . . . . . . . . . . . . . .924.6Ansatz 1 - Univariate Regressionen . . . . . . . . . . . . . . . . . . . .934.7MD-Analyse: Dynamik von 16-1 und 27-1 . . . . . . . . . . . . . . . .954.8MD-Analyse: Vergleich zwischen 16-1 und 27-1. . . . . . . . . . . . .964.9MD-Analyse: Vergleich zwischen 29-1 und 29-9. . . . . . . . . . . . .974.10 MD-Analyse: Vergleich zwischen 19-1 und 19-8. . . . . . . . . . . . .984.11 MD-Analyse: Vergleich zwischen 19-1 und 19-4. . . . . . . . . . . . .984.12 MD-Analyse: Vergleich zwischen Mutanten von 23-Z . . . . . . . . . .994.13 Ansatz 2 - Modell A2-R6. . . . . . . . . . . . . . . . . . . . . . . . . 1024.14 Ansatz 2 - Modell A2-R6v2 . . . . . . . . . . . . . . . . . . . . . . . . 1084.15 Ansatz 2 - ROC-Kurve Modell A2-C1. . . . . . . . . . . . . . . . . . 1094.16 Parametervergleich . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1134.17 Ansatz 2 - Modell A2v2-R1 . . . . . . . . . . . . . . . . . . . . . . . . 1144.18 Ansatz 2 - ROC-Kurve Modell A2v2-C1 . . . . . . . . . . . . . . . . . 1154.19 Energiediagramme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1184.20 Energiedifferenzen. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1194.21 NEB-Pfade. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1214.22 Entfaltungsdynamik. . . . . . . . . . . . . . . . . . . . . . . . . . . . 1224.23 Momentaufnahmen Pin 1 WW - 14 . . . . . . . . . . . . . . . . . . . . 1244.24 Eigenschaften Pin 1 WW - 14 . . . . . . . . . . . . . . . . . . . . . . . 1254.25 Momentaufnahmen Pin 1 WW - 16 . . . . . . . . . . . . . . . . . . . . 1264.26 Eigenschaften Pin 1 WW - 16 . . . . . . . . . . . . . . . . . . . . . . . 1274.27 Momentaufnahmen Pin 1 WW - 17 . . . . . . . . . . . . . . . . . . . . 1284.28 Eigenschaften Pin 1 WW - 17 . . . . . . . . . . . . . . . . . . . . . . . 1294.29 Momentaufnahmen Pin 1 WW - 18 . . . . . . . . . . . . . . . . . . . . 1304.30 Eigenschaften Pin 1 WW - 18 . . . . . . . . . . . . . . . . . . . . . . . 1314.31 Momentaufnahmen Pin 1 WW - 19 . . . . . . . . . . . . . . . . . . . . 1324.32 Eigenschaften Pin 1 WW - 19 . . . . . . . . . . . . . . . . . . . . . . . 1334.33 Momentaufnahmen Pin 1 WW - 21 . . . . . . . . . . . . . . . . . . . . 1344.34 Eigenschaften Pin 1 WW - 21 . . . . . . . . . . . . . . . . . . . . . . . 1354.35 Momentaufnahmen Pin 1 WW - 23 . . . . . . . . . . . . . . . . . . . . 1364.36 Eigenschaften Pin 1 WW - 23 . . . . . . . . . . . . . . . . . . . . . . . 1374.37 Momentaufnahmen Pin 1 WW - 26 . . . . . . . . . . . . . . . . . . . . 1384.38 Eigenschaften Pin 1 WW - 26 . . . . . . . . . . . . . . . . . . . . . . . 1394.39 Momentaufnahmen Pin 1 WW - 27 . . . . . . . . . . . . . . . . . . . . 1404.40 Eigenschaften Pin 1 WW - 27 . . . . . . . . . . . . . . . . . . . . . . . 141Abbildungsverzeichnis3254.41 Momentaufnahmen Pin 1 WW - 28 . . . . . . . . . . . . . . . . . . . . 1424.42 Eigenschaften Pin 1 WW - 28 . . . . . . . . . . . . . . . . . . . . . . . 1434.43 Momentaufnahmen Pin 1 WW - 29 . . . . . . . . . . . . . . . . . . . . 1444.44 Eigenschaften Pin 1 WW - 29 . . . . . . . . . . . . . . . . . . . . . . . 1454.45 Momentaufnahmen Pin 1 WW - 30 . . . . . . . . . . . . . . . . . . . . 1464.46 Eigenschaften Pin 1 WW - 30 . . . . . . . . . . . . . . . . . . . . . . . 1474.47 Momentaufnahmen Pin 1 WW - 32 . . . . . . . . . . . . . . . . . . . . 1484.48 Eigenschaften Pin 1 WW - 32 . . . . . . . . . . . . . . . . . . . . . . . 1494.49 Momentaufnahmen Her2 - 7 . . . . . . . . . . . . . . . . . . . . . . . . 1504.50 Eigenschaften Her2 - 7 . . . . . . . . . . . . . . . . . . . . . . . . . . . 1514.51 Momentaufnahmen Her2 - 8 . . . . . . . . . . . . . . . . . . . . . . . . 1524.52 Eigenschaften Her2 - 8 . . . . . . . . . . . . . . . . . . . . . . . . . . . 1534.53 Momentaufnahmen Her2 - 15 . . . . . . . . . . . . . . . . . . . . . . . 1544.54 Eigenschaften Her2 - 15 . . . . . . . . . . . . . . . . . . . . . . . . . . 1554.55 Momentaufnahmen Her2 - 21 . . . . . . . . . . . . . . . . . . . . . . . 1564.56 Eigenschaften Her2 - 21 . . . . . . . . . . . . . . . . . . . . . . . . . . 1574.57 Momentaufnahmen Her2 - 23 . . . . . . . . . . . . . . . . . . . . . . . 1584.58 Eigenschaften Her2 - 23 . . . . . . . . . . . . . . . . . . . . . . . . . . 1594.59 Momentaufnahmen Her2 - 25 . . . . . . . . . . . . . . . . . . . . . . . 1604.60 Eigenschaften Her2 - 25 . . . . . . . . . . . . . . . . . . . . . . . . . . 1614.61 Momentaufnahmen Her2 - 29 . . . . . . . . . . . . . . . . . . . . . . . 1624.62 Eigenschaften Her2 - 29 . . . . . . . . . . . . . . . . . . . . . . . . . . 1634.63 Momentaufnahmen Her2 - 33 . . . . . . . . . . . . . . . . . . . . . . . 1644.64 Eigenschaften Her2 - 33 . . . . . . . . . . . . . . . . . . . . . . . . . . 1654.65 Momentaufnahmen Her2 - 39 . . . . . . . . . . . . . . . . . . . . . . . 1664.66 Eigenschaften Her2 - 39 . . . . . . . . . . . . . . . . . . . . . . . . . . 1674.67 Momentaufnahmen Her2 - 43 . . . . . . . . . . . . . . . . . . . . . . . 1684.68 Eigenschaften Her2 - 43 . . . . . . . . . . . . . . . . . . . . . . . . . . 1694.69 Momentaufnahmen Her2 - 46 . . . . . . . . . . . . . . . . . . . . . . . 1704.70 Eigenschaften Her2 - 46 . . . . . . . . . . . . . . . . . . . . . . . . . . 1714.71 Momentaufnahmen Her2 - 47 . . . . . . . . . . . . . . . . . . . . . . . 1724.72 Eigenschaften Her2 - 47 . . . . . . . . . . . . . . . . . . . . . . . . . . 1734.73 Momentaufnahmen Her2 - 49 . . . . . . . . . . . . . . . . . . . . . . . 1744.74 Eigenschaften Her2 - 49 . . . . . . . . . . . . . . . . . . . . . . . . . . 1754.75 Momentaufnahmen Her2 - 50 . . . . . . . . . . . . . . . . . . . . . . . 1764.76 Eigenschaften Her2 - 50 . . . . . . . . . . . . . . . . . . . . . . . . . . 1774.77 Momentaufnahmen Her2 - 53 . . . . . . . . . . . . . . . . . . . . . . . 1784.78 Eigenschaften Her2 - 53 . . . . . . . . . . . . . . . . . . . . . . . . . . 1794.79 Zusammenfassender Vergleich der Pin 1-Konjugate . . . . . . . . . . . 1814.80 Pin 1 - Übersicht Konjugationsstellen . . . . . . . . . . . . . . . . . . . 1824.81 Interaktionsenergien - Pin 1 WW-Domäne . . . . . . . . . . . . . . . . 1834.82 Pin 1 - PEG-Interaktionen . . . . . . . . . . . . . . . . . . . . . . . . . 184326Abbildungsverzeichnis4.83 Pin 1 - LPG-Interaktionen . . . . . . . . . . . . . . . . . . . . . . . . . 1854.84 Pin 1 - LPG-Interaktionen . . . . . . . . . . . . . . . . . . . . . . . . . 1854.85 Pin 1 - POx-Interaktionen . . . . . . . . . . . . . . . . . . . . . . . . . 1864.86 Pin 1 - POx-Vergleich. . . . . . . . . . . . . . . . . . . . . . . . . . . 1884.87 Zusammenfassender Vergleich der Her2-Konjugate. . . . . . . . . . . 1894.88 Her2 - Übersicht Konjugationsstellen . . . . . . . . . . . . . . . . . . . 1904.89 Interaktionsenergien - Her2-Affibody . . . . . . . . . . . . . . . . . . . 1914.90 Her2 - PEG-Interaktionen . . . . . . . . . . . . . . . . . . . . . . . . . 1924.91 Her2 - LPG-Interaktionen . . . . . . . . . . . . . . . . . . . . . . . . . 1944.92 Her2 - POx-Interaktionen . . . . . . . . . . . . . . . . . . . . . . . . . 1945.1Polymerdichten . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2005.2SASA IFN-α2a . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2015.3Kontaktprofil 10-kDa-BCN-PEG. . . . . . . . . . . . . . . . . . . . . 2035.4PEG-Interaktionen mit K/R . . . . . . . . . . . . . . . . . . . . . . . . 2045.5Kontaktprofil 10-kDa-BCN-LPG. . . . . . . . . . . . . . . . . . . . . 2055.6LPG-Interaktionen mit D/A . . . . . . . . . . . . . . . . . . . . . . . . 2055.7Kontaktprofil 10-kDa-BCN-PEtOx . . . . . . . . . . . . . . . . . . . . 2075.8PEtOx-Interaktionen mit S/Y . . . . . . . . . . . . . . . . . . . . . . . 2075.9Kontaktprofil 10-kDa-DBCO-PEG . . . . . . . . . . . . . . . . . . . . 2085.10 Kontaktprofil 10-kDa-DBCO-LPG. . . . . . . . . . . . . . . . . . . . 2095.11 Kontaktprofil 10-kDa-DBCO-PEtOx . . . . . . . . . . . . . . . . . . . 2105.12 Vergleich BCN und Peptidbindung . . . . . . . . . . . . . . . . . . . . 2115.13 Kontaktprofil 20-kDa-BCN-PEG. . . . . . . . . . . . . . . . . . . . . 2135.14 Kontaktprofil 20-kDa-BCN-LPG. . . . . . . . . . . . . . . . . . . . . 2135.15 Kontaktprofil 20-kDa-BCN-PEtOx . . . . . . . . . . . . . . . . . . . . 2145.16 Kontaktprofil 40-kDa-BCN-PEG. . . . . . . . . . . . . . . . . . . . . 2145.17 Kontaktprofil 50-kDa-BCN-LPG. . . . . . . . . . . . . . . . . . . . . 2155.18 Kontaktprofil 30-kDa-BCN-PEtOx . . . . . . . . . . . . . . . . . . . . 2155.19 Verteilung von PEG in der CG-Simulation . . . . . . . . . . . . . . . . 2175.20 Dichtevergleich CG MD - GaMD . . . . . . . . . . . . . . . . . . . . . 2175.21 Kontaktprofil CG-MD PEG-10-kDa. . . . . . . . . . . . . . . . . . . 2185.22 Shroud / Dumbbell-Modell. . . . . . . . . . . . . . . . . . . . . . . . 2195.23 Schmelzpunkte - 10 kDa . . . . . . . . . . . . . . . . . . . . . . . . . . 2205.24 In silico-Interpretation der Schmelztemperaturen - 10 kDa . . . . . . . 2215.25 CNA-Netzerke 10 kDa . . . . . . . . . . . . . . . . . . . . . . . . . . . 2245.26 CNA-Interaktionen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2255.27 Perkolationsindizes IFN-α2a . . . . . . . . . . . . . . . . . . . . . . . . 2255.28 Native Kontakte LPG-Konjugate . . . . . . . . . . . . . . . . . . . . . 2265.29 LiP-Assay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2285.30 Schmelzpunkte - 20 - 50 kDa. . . . . . . . . . . . . . . . . . . . . . . 2295.31 In silico-Interpretation der Schmelztemperaturen - 10 - 50 kDa . . . . 230Abbildungsverzeichnis3275.32 Bioaktivitäten . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2315.33 Korrelation Bioaktivität . . . . . . . . . . . . . . . . . . . . . . . . . . 2325.34 Korrelation Bioaktivität 2 . . . . . . . . . . . . . . . . . . . . . . . . . 2336.1Übersicht der Arbeit . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239B.1Ramachandran-Winkel . . . . . . . . . . . . . . . . . . . . . . . . . . . 258D.1 RDF-Differenz-Plots . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282D.2 Umorientierung der ersten Schleife . . . . . . . . . . . . . . . . . . . . 283D.3 Korrelationen - Modell A2-R6 . . . . . . . . . . . . . . . . . . . . . . . 286D.4 Korrelationen - Modell A2-C1 . . . . . . . . . . . . . . . . . . . . . . . 287D.5 Verhältnis Modell A2-R6 - Polymeralternativen / PEG . . . . . . . . . 288D.6 Verhältnis Modell A2-C1 - Polymeralternativen / PEG . . . . . . . . . 289 329Tabellenverzeichnis1.1Vermarktete Biokonjugate . . . . . . . . . . . . . . . . . . . . . . . . .31.2Vor- / Nachteile der Biokonjugation. . . . . . . . . . . . . . . . . . .51.3Konjugationsbeispiele. . . . . . . . . . . . . . . . . . . . . . . . . . .81.4Stabilisierungsmechanismen . . . . . . . . . . . . . . . . . . . . . . . .203.1PyRED-Parameter . . . . . . . . . . . . . . . . . . . . . . . . . . . . .553.2antechamber-Parameter . . . . . . . . . . . . . . . . . . . . . . . . . .563.3Datensätze. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .723.4Simulierte Konjugate von IFN-α2a . . . . . . . . . . . . . . . . . . . .794.1Statistische Parameter für Modelle C1 und D1. . . . . . . . . . . . .884.2Statistische Parameter für Modelle E und F . . . . . . . . . . . . . . .904.3Statistische Parameter für Modelle G und H . . . . . . . . . . . . . . .924.4Statistische Parameter für Modell A2-R6 . . . . . . . . . . . . . . . . . 1024.5Konjugate - Modell A2-R6 . . . . . . . . . . . . . . . . . . . . . . . . . 1034.6Strukturbasierte Richtlinie . . . . . . . . . . . . . . . . . . . . . . . . . 1054.7Statistische Parameter für Modell A2-R6v2. . . . . . . . . . . . . . . 1074.8Statistische Parameter für Modell A2-C1 . . . . . . . . . . . . . . . . . 1094.9Konjugate - Modell A2-C1 . . . . . . . . . . . . . . . . . . . . . . . . . 1104.10 Statistische Parameter für Modell A2v2-R1. . . . . . . . . . . . . . . 1144.11 Statistische Parameter für Modell A2v2-C1. . . . . . . . . . . . . . . 1154.12 Zusätzliche Konjugate - Vorhersagen . . . . . . . . . . . . . . . . . . . 1164.13 Energiedifferenzen. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1204.14 Energiebarrieren. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121B.1CNA-basierte Deskriptoren. . . . . . . . . . . . . . . . . . . . . . . . 261C.1 Chemische Strukturen . . . . . . . . . . . . . . . . . . . . . . . . . . . 263C.2 Simulierte Modellprotein-Konjugate. . . . . . . . . . . . . . . . . . . 265D.1 Berechnete Deskriptoren . . . . . . . . . . . . . . . . . . . . . . . . . . 273D.2 QSPR-Ansatz 1 - Modelltypen. . . . . . . . . . . . . . . . . . . . . . 278D.3 QSPR-Ansatz 1 - Statistische Parameter . . . . . . . . . . . . . . . . . 280D.4 QSPR-Ansatz 1 - ADAN-Analyse . . . . . . . . . . . . . . . . . . . . . 281D.5 QSPR-Ansatz 2 - Modelltypen. . . . . . . . . . . . . . . . . . . . . . 284D.6 QSPR-Ansatz 2 - Statistische Parameter der Regressionsmodelle. . . 284330TabellenverzeichnisD.7 QSPR-Ansatz 2 - Statistische Parameter der Klassifikationsmodelle . . 285D.8 Polymeralternativen - Vorhersagen Modell A2-R6 . . . . . . . . . . . . 288D.9 Polymeralternativen - Vorhersagen Modell A2-C1 . . . . . . . . . . . . 289331AbkürzungsverzeichnisAAAcrylsäureABCAccelerated blood clearanceAccAccuracyADApplicability domainADANApplicability domain analysisAGAN-Acryloyl-D-glucosaminAmAcrylamidaMDAccelerated MDATRPAtomtransferradikalpolymerisationAUPRCArea under the precision-recall curveAUROCArea under the receiver operating characteristic curveAZKNε-[(2-Azidoethoxy)carbonyl]-L-lysinBAccBalanced accuracyBBCOMBootstrapped block covariance overlap methodBCNBicyclononinBCOMBlock covariance overlap methodCCCKonkordanzkorrelationskoeffizientCCEConstancy of cluster entropyCCE1/2Clusterkonfigurationsentropie Typ 1/2CCRCorrect classification rateCDCirculardichroismusCGCoarse-grainedcMDConventional MDCNAConstrained Network-AnalyseCoMFAComparative molecular field analysisCPOLinear combination of pairwise overlapsCROPKationische RingöffnungspolymerisationCRPKontrolliert radikalische PolymerisationCTAKettentransferreagenzCVCollective variablesDAADirect acting agentsDBCODibenzocyclooctinDLSDynamische LichtstreuungDMAEMADimethylaminoethoxymethacrylatDMAmDimethylacrylamidDSFDifferential Scanning FluorimetryDUDummy atomEPREnhanced permeability and retentionFDAFood and Drug AdministrationFIRSTFloppy inclusions and rigid substructure topographyFNFalse negativesFPFalse positives332AbkürzungsverzeichnisFPRFalse positive rateG-CSFGranulozyten-Kolonie-stimulierender FaktorGaMDGaussian accelerated MDHBWasserstoffbrückenbindungHEKHuman embryonic kidneyHer2Human epidermal growth factor receptor 2HESHydroxyethylstärkeHSHoldout-SetHSQCHeteronuclear Single Quantum CoherenceIFN-α2aInterferon-α2aIFNARInterferon-α-RezeptoriREDIsotropic reorientational eigenmode dynamicsIUPAPInternationale Union für reine und angewandte PhysikLiPLimitierte ProteolyseLPGLineares PolyglycerolMAEMean absolute errorMALDI-TOFMatrix-assistierte Laser-Desorption-Ionisierung-Time-of-flightMCMCMarkow-Chain-Monte-CarloMDMolekulardynamikMetaMDMetadynamik-SimulationMOEMolecular Operating EnvironmentmPEGMonomethoxy-PEGMRWMean residue weightMSDMittlere quadratische VerschiebungMWMultiple walkersMW-wtMetaMDsMultiple walkers well-tempered metadynamics simulationNANot availableNAEKNε-[(2-Azidoethoxy)carbonyl]-L-lysinNHSN-HydroxysuccinimidNMRNuklearmagnetresonanzNPTIsothermal-isobarisches EnsembleNPVNegative predictive valueNVTMikrokanonisches EnsembleOECDOrganisation for Economic Co-operation and DevelopmentOEOAOligo(ethylenoxid)methyletheracrylatPAMPoly(ammoniummethacrylat)PBCPeriodic boundary conditionsPBKPolymer-BiokonjugatPCAPrincipal component analysisPCBAmPoly(carboxybetainacrylamid)PCMAPhosphorylcholinmethacrylatPDBProteindatenbankPDIPolydispersitätPEGPolyethylenglykolPEtOxPoly(2-ethyl-2-oxazolin)PGPolyglycerolPLSPartial least squaresPMeOxPoly(2-methyl-2-oxazolin)PNIPAAmPoly(N-isopropylacrylamid)POEGMAPoly(oligo(ethylenglykol)methacrylat)POxPoly(2-alkyl-2-oxazolin)Abkürzungsverzeichnis333PPVPositive predictive valuePRESSPredicted residual sum of squaresPSMAPoly(sulfonatmethacrylat)PVPPolyvinylpyrrolidonQMQuantenmechanikQSARQuantitative structure-activity relationshipsQSPRQuantitative structure-property relationshipsRAFTReversible Additions-Fragmentierungs-Kettenübertragungs-PolymerisationrdCVRepeated double cross validationRDFRadial distribution functionRESPRestrained electrostatic potentialRFERekursive Feature-EliminierungRMSDRoot-mean-square deviationRMSERoot-mean-square errorRMSFRoot-mean-square fluctuationROCGrenzwertoptimierungskurveRSSResidual sum of squaresSASASolvenszugängliche OberflächeSCCSelf-consistency checkSCIDSchwerer kombinierter ImmundefektSDVGSmith, Daura und van GunsterenSEStandard errorSEAPSecreting alkaline phosphatase assaySPAACStrain-promoted azide-alkyne click chemistry reactionSSESum of squared errorsST-HPMASemitelechelisches Poly[N-(2-hydroxypropyl)-methacrylamidTEMPO2,2,6,6-TetramethylpiperidinyloxylTNTrue negativesTNFTumor-Nekrose-FaktorTNRTrue negative rateTPTrue positivesTPMATris(2-pyridylmethyl)aminTPRTrue positive rateTSTrainingsdatensatzTSAThermal Shift AssayTSSTotal sum of squaresVdWVan-der-WaalsVEGFVaskulärer endothelialer WachstumsfaktorwtWell-temperedWTWildtyp | RESEARCH GATE |
Active impact on the ionosphere and variations of velocity char-acteristics of field-aligned irregularitiesVladimir Sivokon’1,⋆1Institute of Cosmophysical Research and Radio Wave Propagation FEB RAS, Kamchatskiy kray, Paratunka,RussiaAbstract. Ionospheric irregularities make significant impact on the efficiency of radio-engineering system operation. Our investigations carried out on Kamchatka peninsulashowed that field-aligned irregularities have particularly destructive influence on decame-ter radars. We pin some definite expectations with them to increase the efficiency of radio-engineering systems by active impact on the ionosphere. One of the important parametersof such irregularities is their velocity characteristics. When determining irregularity ve-locity in the impact area, different approaches are used, for example, aspect scatteringmethod which specifies velocity on the whole but does not allow one to determine itsvertical and horizontal components, direction of its motion. We have developed an ap-proach which allows us to solve this problem. As an example, results of observationsover EISCAT experiments in October 2019 are presented.1 IntroductionOur investigations [1] showed that presence of a special class of ionospheric irregularities, field-aligned ones, is acute not only for beyond-the-horizon radio location. This type of irregularitiesis of interest during active impact on the ionosphere [2]. Consequently, research of field-alignedirregularities is a topical task.In order to estimate the parameters of artificial irregularities of the ionosphere, different methodsare applied, for example, test wave [3], aspect [4] and backward scattering [5] methods. To real-ize these methods one needs complex equipment unpurchasable for the majority of researches. Wesuggested a method [6] which included generally acceptable software-defined SDR receivers (Soft-DefinedRadio). Results of the observation over the experiments taking place in October 2019 aredescribed in the paper.2 Method justificationA network of SDR receivers installed in Europe, in particular, make it possible to record electromag-netic radiation from the impact area within the line-of-sight, that minimizes the ionosphere impact⋆Corresponding author: nina@ikir.ruThe paper was carried out within the framework on the subject «Dynamics of physical processes in active zones of near spaceand geospheres» (AAAA-A17-117080110043-4)© The Authors, published by EDP Sciences. This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0 (http://creativecommons.org/licenses/by/4.0/).E3S Web of Conferences 196, 02002 (2020) https://doi.org/10.1051/e3sconf/202019602002STRPEP 2020on radiation propagation. An analysis of the distribution of the amplitudes of the heating radiationscattered by ionospheric irregularities showed it’s bimodal and, in some cases, polymodal character.When estimating the capacity of the short-wave communication channel (ionosphere heating is car-ried out just in this wave length range) one take Rayleigh, Rice or Nakahami distributions [7], as arule. However, they are not bimodal, i.e. the observed distributions are anomalous with respect tothose distributions. Based on the velocities of the observed fading, Fig.1, they refer to fast ones.Figure 1. Scattered signal record on October 18, 2019 at 13:14 UT at AltaFast fading is divided into interference and polarization ones. It is known that polarization fadingis determined by magneto-ionic splitting of the wave incident to the ionosphere and, in a generalcase, by the formation of elliptically polarization signal the polarization characteristics of which donot coincide with the same characteristics of a receiving antenna. However, under active impact onthe ionosphere, when the direction is towards the magnetic zenith, radiation with circular polarizationis used and the possibility of polarization fading manifestation is low as long as in this case therewill always be a position of electric field strength vector which will coincide with receiving antennapolarization plane [8, 9]. Interference fading is, as a rule, determined by ray path differences, i.e.,phase ratios. In our case, it is necessary to take into account frequency Doppler shift as long as field-aligned irregularities move [1]. Then, in a general case, the resultant field can be represented as asumE(t) = E1 · cos(ω1t + ϕ1) + E2 · cos(ω2t + ϕ2) + ... + ξ(t),where E1, E2 are amplitudes; ω1, ω2 are frequencies; ϕ1, ϕ2 are phases of separate scatteredcomponents; ξ(t) is the noise component.For the case when ω1, ω2 differ insignificantly and noise dispersion is small, in the paper [10] anapproximated expression for the density of process envelope possibility for two-ray distribution wasobtainedW(R) �R12σ2 exp�−12σ2 (R2 + E21 + E22)�×�PI0� E2σ2 (R − E1)�+ 1PI0� E2σ2 (R + E1)�++2�PI0� E2σ2�R2 − RE1 + E21�+ 1PI0� E2σ2�R2 + RE1 + E21���2E3S Web of Conferences 196, 02002 (2020) https://doi.org/10.1051/e3sconf/202019602002STRPEP 2020where P = exp�− RE12σ2�.When choosing the parameters E1, E2 and σ we can obtain bimodal distribution and componentcorrelation is not necessary in this case. The main requirement is the difference of frequencies ω1and ω2. As long as this difference is determined by frequency Doppler shift, it cannot be significantcompared to the heating frequency, then quasi-beats and quasi-periodic processes should take place.In Fig. 1, features of such a process are indicated by arrows.3 Observation resultsBesides the features of beats in a wave form, we detected spectrum splitting, Fig.2. This phenomenonis described in the papers [10, 11]. In the paper [10] it is assumed that this phenomenon is determinedby field-aligned irregularity drift.Figure 2. Scattered signal record on October 18, 2019 at 13:14 UT at AltaSpectrum splitting is an interesting phenomenon but during automatic processing of records itcauses difficulties in detection of real value of frequency Doppler shift. As long as at the current stageof investigations the experimental data volume is not large and our resources are limited, we madesome simplification, the software determines frequency Doppler shift for the maximum amplitude.At the initial stage of investigations when estimating velocity and direction of irregularity motion,we applied ten-second averaging that is, probably, unjustified for the irregularity parameters describedabove. Thus, when we processed the data from October 2019, another approach was used. In softwarewe defined the FFT maximum and the maximum time resolution.Based on the initial data, frequency Doppler shift distributions were drawn and velocity distri-butions were calculated from them. As long as it is not possible to present the data over the wholeheating cycle, we divided the cycle into periods of power increase and decrease. Results for ALTAsite are illustrated in Fig.3-4. Markers show velocity discrete values, the lines connecting the markersare applied in order to trace easier velocity changes at the heating stage. It is clear from the graphsthat when the impact power increases, the velocity value and sign change. Under minimum impact,the irregularities moved away from an observer. As the power grew, they began moved in an opposite3E3S Web of Conferences 196, 02002 (2020) https://doi.org/10.1051/e3sconf/202019602002STRPEP 2020Figure 3. Velocity distribution during impact power increasedirection. When impact power decreased, we expected an opposite picture, that occurred, however,the process was slow that may indicate the inertial process in ionospheric plasma.Figure 4. Velocity distribution during impact power decreaseWe use multi-position diagnosis of ionosphere modification, i.e. the area is observed at differentangles in vertical and horizontal planes, that allows us to determine the components of irregularityvelocity and direction of their primary motion, Fig.5.4E3S Web of Conferences 196, 02002 (2020) https://doi.org/10.1051/e3sconf/202019602002STRPEP 2020Figure 5. Distribution of irregularity motion directions4 Conclusions• We have shown the possibility to investigate velocities and directions of ionospheric field-alignedirregularity motion applying SDR technology.• Observations showed the dependence of motion velocity and direction on heating wave power.• The motion determined by active impact on the ionosphere has primary southern direction.References[1] V.P.Sivokon,GeomagnetismandAeronomy,57(3),317–325(2017)DOI:10.1134/S001679321702013X[2] N.F. Blagoveshchenskaya, T.D. Borisova, A.S. Kalishin, T.K. Yeoman, Artificial Small-scaleField-aligned Irregularities in the High Latitude Ionospheric F-region, AGU Fall Meeting 10 –14 December (Washington, DC, USA, 2018)[3] S.M. Grach, G.P. Kompakov, M.M. Shvarts, M.A. Yurishchev, Radiophysics and Quantum Elec-tronics, 41(8), 653–660 (1998) DOI: 10.1007/BF02676473[4] N.F. Blagoveshchenskaya, T.D. Borisova, V.A. Kornienko, T.R. Robinzon,T.K. Yoman, V.L.Frolov, M.T Ritvel’d, Solnechno-zemnaya fizika [Solar-Terrestrial Physics], 12(2), 206–209 (2008)(in Russian)[5] N. A. Mityakov, V. A. Alimov, V. A. Zinichev, G. P. Komrakov, S. N. Mityakov, Radiophysicsand Quantum Electronics, 53(5-6), 297–304 (2010) DOI: 10.1007/s11141-010-9228-6[6] V.P. Sivokon’, R.V. Kolesnikov, I.V. Demichev, Elektromagnitnye volny i elektronnye sis-temy [Electromagnetic waves and electronic systems], 24(8), 23–28 (2019) (in Russian) DOI:10.18127/j15604128-201908-03[7] D.D. Klovskiy Peredacha diskretnykh soobshcheniy po radiokanalam [Transmission of discretemessages via radio channels], (Moscow: Radio i svyaz’, 1982), –304 p. (in Russian)5E3S Web of Conferences 196, 02002 (2020) https://doi.org/10.1051/e3sconf/202019602002STRPEP 2020[8] V.P. Sivokon’, Elektrosvyaz’ [Telecommunications and Radio Engineering], 7, 55–58 (2007) (inRussian)[9] V.P. Sivokon, G.I. Druzhin, Geomagnetism and Aeronomy, 46(4), 492–495 (2006) DOI:10.1134/S0016793206040116[10] A.V. Koloskov, V.S. Beley, T.B. Leyzer, Yu.M. Yampol’skiy, Radiophysics and Radioastronomy,4(3), 247–260 (1999) (in Russian) http://rpra-journal.org.ua/index.php/ra/article/view/978[11] T.D. Borisova, N.F. Blagoveshchenskaya, V.A. Kornienko, V.L. Frolov, V.V. Vertogradov, V.G.Vertogradov, Geomagnetism and Aeronomy, 49(4), 535–544 (2009) (in Russian)6E3S Web of Conferences 196, 02002 (2020) https://doi.org/10.1051/e3sconf/202019602002STRPEP 2020 | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/306265547Relationship of each anterior pituitary hormone deficiency to the size of non-functioning pituitary adenoma in the hospitalized patientsArticle in Endocrine Journal · August 2016DOI: 10.1507/endocrj.EJ16-0168CITATIONS14READS506 authors, including:Some of the authors of this publication are also working on these related projects:PRIMARY ALDOSTERONISM View projectMichio OtsukiTokyo Women's Medical University177 PUBLICATIONS 3,150 CITATIONS SEE PROFILEAll content following this page was uploaded by Michio Otsuki on 24 March 2017.The user has requested enhancement of the downloaded file.2016, 63 (1), 1-10NON-FUNCTIONING PITUITARY ADENOMA (NFPA) accounts for more than half of pituitary ade-nomas [1]. The mass effect of NFPA causes headache, visual field defects and hypopituitarism. Most patients with macro-NFPA present with hypopituitarism [1]. Early diagnosis and treatment are necessary since NFPA-hypopituitarism is associated with poor progno-sis [2-4] and low quality of life (QOL). The clinical features of hypopituitarism (e.g. tiredness, weakness, hair loss, cognitive slowing and mood impairment), are not specific for accurate diagnosis [5, 6]. Therefore, Relationship of each anterior pituitary hormone deficiency to the size of non-functioning pituitary adenoma in the hospitalized patientsKosuke Mukai, Tetsuhiro Kitamura, Daisuke Tamada, Masahiko Murata, Michio Otsuki and Iichiro ShimomuraDepartment of Metabolic Medicine, Osaka University Graduate School of Medicine, Suita, JapanAbstract. Non-functioning pituitary adenoma (NFPA) is often associated with hypopituitarism. Diagnosis of hypopituitarism is important because of its poor prognosis and low quality of life. Among hypopituitarism, it is difficult to diagnose secondary adrenocortical insufficiency and GH deficiency without hormone stimulation test. Therefore, the aim of our study was to identify patients with NFPA who require more careful endocrinological examination. We examined the relationship between NFPA size and the prevalence of each hypopituitarism or the response of each anterior pituitary hormone by insulin tolerance test, LHRH test and TRH test. We studied 63 patients with NFPA admitted for evaluation of pituitary function and surgical indication. They were classified three groups by tumor diameter. The prevalence of GH deficiency, male secondary hypogonadism, secondary hypothyroidism and PRL deficiency were higher in the group of larger tumor diameter (p<0.0001, p<0.05, p<0.05 and p<0.05, respectively). However, that of secondary adrenocortical insufficiency only tended to be higher (p=0.07). In the group with small NFPA (less than 20 mm), the prevalence of secondary adrenocortical insufficiency was 38% although those of GH deficiency, male secondary hypogonadism, secondary hypothyroidism and PRL deficiency were 0%, 0% and 8% and 9%, respectively. Anterior pituitary hormone responses except TSH had significantly negative correlation with tumor diameter (ACTH: r=-0.40, GH: r=-0.57, LH: r=-0.69, FSH: r=-0.46, PRL: r=-0.36). The results suggested physicians should proactively suspect GH deficiency, male secondary hypogonadism and secondary hypothyroidism in patients with larger NFPA. On the other hand, adrenocortical function should be examined even in patients with small NFPA. Key words: Non-functioning pituitary adenoma, Hypopituitarism, Pituitary tumor diameter, Secondary adrenocortical insufficiencypituitary function should be assessed based on not only the clinical features but also on endocrinological exam-ination. The latter includes measurement of basal hor-mone levels and hormone stimulation tests. Usually it is difficult to perform hormone stimulation test because it requires higher cost and more specialized experience of endocrinologist compared with the measurement of basal hormone level. Moreover, severe complications associated with hormone stimulation test, such as pitu-itary apoplexy, have been reported [7, 8]. However, in order to properly evaluate secondary adrenocortical insufficiency and GH deficiency, the insulin tolerance test (ITT) and hormone stimulation test must be per-formed because the basal levels of ACTH, cortisol and GH do not reflect their secretion capacity [9-11]. Although hypopituitarism can only be diagnosed by endocrinological examination, only one study showed Advance Publicationdoi: 10.1507/endocrj.EJ16-0168Original©The Japan Endocrine SocietySubmitted Apr. 3, 2016; Accepted Jul. 7, 2016 as EJ16-0168Released online in J-STAGE as advance publication Aug. 18, 2016Correspondence to: Tetsuhiro Kitamura, M.D., Ph.D., Department of Metabolic Medicine, Osaka University Graduate School of Medicine, 2-2 Yamada-oka, Suita, Osaka 565-0871, Japan. E-mail: tkitamura@endmet.med.osaka-u.ac.jp2Mukai et al.range [13]. There were no patients of hyperthyroid-ism. In the patients with hyperprolactinemia, we ruled out prolactinoma based on their serum prolactin level relative to their tumor size [14]. The exclusion crite-ria were the followings: conditions that could affect pituitary function (primary hyperthyroidism [n=1], ste-roid use [n=3], levothyroxine sodium use [n=2]), and conditions affecting measurement of pituitary tumor size (history of pituitary surgery [n=9]). After appli-cation of the above exclusion criteria, the study sub-jects included 63 patients. None had pituitary cystic lesion that could possibly affect pituitary function and tumor size. Surgical indication was based on visual field defect, risk of optic nerve compression and ten-dency to tumor development. Neurosurgeons decided not to perform surgery in 9 of 63 patients. Therefore, these 9 patients were only evaluated pituitary function but not operated. The study protocol was approved by the Human Ethics Committee of Osaka University (No.15310) and performed according to the Helsinki declaration. Clinical diagnosis of each pituitary deficiencySecondary adrenocortical insufficiency and GH defi-ciency were diagnosed by ITT (secondary adrenocor-tical insufficiency; peak serum cortisol level: <18 μg/dL, GH deficiency; peak serum GH level: <3 ng/mL) [15-17]. As for secondary hypogonadism, the diagno-sis was limited to men because the study included only a few female subjects and the majority was old in age (only 10 subjects were <50 years of age) and had no menstrual information. Male secondary hypogonadism was diagnosed on the basis of total testosterone level that the likelihood of hypopituitarism increased in larger NFPAs [12], suggesting that the size of pitu-itary adenoma is a potential marker of hypopituita-rism. However, this study defined hypopituitarism as at least one anterior pituitary hormone deficien-cies, such as secondary adrenocortical insufficiency, GH deficiency, secondary hypogonadism and second-ary hypothyroidism, and did not indicate each defi-ciency. Therefore, it should be evaluated the relation-ship between the size of pituitary adenoma and each pituitary hormone deficiency, respectively, in particu-lar secondary adrenocortical insufficiency and second-ary hypothyroidism, because these two hormone defi-ciencies are critical for survival. The aim of the present study was to identify patients with NFPA who should require more careful endocri-nological examination, by elucidating the relationship between NFPA size and prevalence of each pituitary hormone deficiency. Subjects and MethodsSubjectsPatients’ selection criteria in this study were sum-marized in Fig. 1. The subjects were recruited from NFPA patients (n=78) previously admitted to Osaka University Hospital between January 2009 and April 2015 for evaluation of pituitary functions and surgi-cal indication. All patients had no symptom such as cushingoid or acromegalic features. They also had no endocrinological data suggesting Cushing disease and acromegaly such as elevated serum or urinary corti-sol and serum IGF-1 over the upper limit of normal 78 admitted patients with NFPA63 patients with NFPAPatients with appropriate endcrinologicalevaluation∙ With free thyroxine (n=60)∙ With total testosterone (n=29)∙ With insulin tolerance test (n=44)∙ With LHRH test (n=27 men)∙ With TRH test (assessment for TSH [n=52] and PRL secretion [n=37])Exclude the followings.∙ Pituitary hormone excess (n=0)∙ Conditions affecting pituitary function (Primary hyperthyroidism: n=1, Levothyroxine sodium use: n=2, Steroid use: n=3)∙ Diffculty measuring pituitary tumor size (With surgical history: n=9)∙ Pituitary cystic lesion infuencing the evaluation of both pituitary function and tumor size (n=0)Fig. 1 Selection criteria of patients in this study Grey area represents studied subjects. 3Adrenocortical dysfunction in small NFPAdopamine effect from the hypothalamus. The dysreg-ulation altered the correlation between size and PRL secretion. Thus, we evaluated PRL secretion in 43 patients excluding these patients. We also evaluated the relative PRL (peak PRL/basal PRL) [17, 20]. None of the female subjects were assessed for LH and FSH secretion based on the reasons mentioned above. The 11 male subjects with elevated basal LH and FSH were diagnosed as primary hypogonadism and 4 men were not examined with LHRH test. Thus, we evaluated LH and FSH secretion in 27 male patients. In these patients, we evaluated the relative LH and FSH (peak LH or FSH/basal LH or FSH) [17]. We also analyzed ΔLH and ΔFSH (peak LH or FSH minus basal LH or FSH) and got the same conclusion as relative LH and FSH (data not shown) [22]. Blood samples were obtained before, and 15, 30, 45, 60, 90, 120 minutes after intravenous administration of insulin, 100 μg LHRH or 250 μg TRH. Measurement of tumor size, tumor volume, Knosp grade and pituitary stalk angleBy tracing the tumor, the maximum tumor diameter and tumor volume were measured using an image anal-ysis software (Image J, National Institutes of Health, Bethesda, MD, http://rsbweb.nih.gov/ij/), as described previously [23]. We classified the tumors according to Knosp grade [24] and evaluated the angle between pituitary stalk and craniocaudal axis as ‘pituitary stalk angle’. In 19 patients, this angle could not be detected because of suprasellar tumor development. Evaluation of visual field defectVisual field was evaluated by Goldmann perimeter. One patient with 17.5 mm NFPA was excluded from the subjects by glaucoma. Laboratory testsBlood samples in fasting and in stimulation tests were obtained in the morning while the patient rested in supine position. Plasma ACTH levels were measured by electrochemiluminescence Immunoassay (ECLusys ACTH kit, Roche Diagnostics, Tokyo, Japan) with an inter-assay coefficient of variation (CV) of <25% and intra-assay CV of <10%. Serum cortisol levels were measured by chemiluminescent enzyme immunoassay (Access cortisol kit, Beckman Coulter, Tokyo). This kit has a sensitivity of 0.4 μg/dL, with an intra-assay CV of <4.3% and inter-assay CV of <5.9%. Serum GH lev-below the lower limit of the normal range (below nor-mal limit of normal) [18]. The 11 male subjects with elevated basal LH and FSH were diagnosed as primary hypogonadism and 2 men didn’t have the data of total testosterone. Thus, we evaluated secondary hypogo-nadism in 29 male patients. Diagnosis of secondary hypothyroidism was based on low serum free T4 level (below normal limit of normal) with inappropriately normal or low TSH level (less than upper limit of nor-mal in our hospital [<3.8 μU/mL]) [19]. PRL deficiency was diagnosed based on relative PRL (peak PRL/basal PRL) by TRH test. Although two cut-off value of rela-tive PRL, 2.5 fold and 2 fold, were referred, PRL defi-ciency was diagnosed by cut-off value of 2 fold because we got the same conclusion using each cut-off [17, 20]. Hyperprolactinemia (PRL >20 ng/mL) was regarded as preservation of PRL secretion capacity. In general, hypopituitarism was defined as pituitary hormone defi-ciency with at least one among secondary adrenocorti-cal insufficiency, GH deficiency, secondary hypogonad-ism and secondary hypothyroidism. Evaluation of secretory capacity of each anterior pituitary hormoneITT was used to evaluate ACTH and GH secre-tion. None of the participating patients had high basal plasma ACTH or high basal serum GH. Because ITT have not been performed for patients with the risk for seizure or ischemic attack by hypoglycemia, only 45 patients were evaluated. Hypoglycemia in ITT was defined as blood glucose (BG) level <40 mg/dL (n=37) and/or half of baseline BG level (n=7). Because hypo-glycemia by ITT was not induced in one patient, we evaluated 44 patients with respect to secondary adre-nocortical insufficiency and GH deficiency. We evalu-ated the peak ACTH and GH in this test. Thyroid stim-ulating hormone (TSH)/prolactin (PRL) secretion and leutenizing hormone (LH)/follicular stimulating hor-mone (FSH) secretion were evaluated by the thyrotro-pin releasing hormone (TRH) test and luteinizing hor-mone–releasing hormone (LHRH) test, respectively. Three patients with high basal thyroid stimulating hor-mone (TSH) were diagnosed as subclinical primary hypothyroidism and excluded. Thus, we evaluated TSH secretion in 52 patients after the exclusion of the above three patients. We evaluated the ΔTSH (peak TSH minus basal TSH) [17, 21]. Basal PRL was also elevated in 14 of the patients performed TRH test, sug-gesting abnormal regulation through inhibition of the 4Mukai et al.28.1 degree, second tertile 28.2-44.9 degree, third ter-tile 45-69.7 degree). Nineteen patients with undetect-able ‘pituitary stalk angle’ were categorized as N.D. group. Prevalence of each anterior pituitary hormone deficiency according to hyperprolactinemia and visual field defect were evaluated by Fisher’s exact test. The cut-off value, sensitivity, and specificity were derived from the receiver operating characteristic (ROC) anal-ysis. Pearson’s correlation coefficient was used to test the correlation between tumor diameter and the results of each hormone during ITT, LHRH test and TRH test. JMP Pro software (ver. 10.0.2 for Window, SAS Institute, Cary, NC) was used for all analyses. ResultsClinical characteristicsThe study included 63 patients (21 females and 42 males) with a median age of 58 years (IQR 47-70), median body mass index (BMI) of 23.6 kg/m2 (IQR 21.7-25.5) and median tumor diameter of 25.6 mm (IQR 20.6-31.6). Except one patient with glaucoma, 61% (38/62) of the patients had visual field defect. Headache and general malaise were recorded in 30% (19/63) and 4.8% (3/63) of all patients. Ten percent (3/29) of men suffered with hypogonadal symptoms. All patients except 3 harbored macroadenoma mea-suring ≥10 mm in diameter. In all study subjects, the prevalence of secondary adrenocortical insufficiency and GH deficiency was 55% (24/44) and 59% (26/44), respectively, among the subjects evaluated by ITT. The prevalence of secondary hypogonadism in men was 41% (12/29). The prevalence of secondary hypo-thyroidism and PRL deficiency was 30% (18/60) and 24% (14/58), respectively. The prevalence of hypopi-tuitarism with at least one anterior pituitary hormone deficiency was 64% (40/63). The clinical characteris-tics by groups based on NFPA size are summarized in Table 1. Sex and BMI were not significantly differ-ent according to tumor diameter. The group of tumor diameter smaller than 20 mm was younger than the group of 20-29 mm tumor (p<0.05). Prevalence of visual field defect was significantly increased in the group of larger NFPA (p<0.0001). The group of larger NFPA have significantly higher prevalence of hypo-pituitarism (with at least one pituitary hormone defi-ciency, p<0.01). In addition, the group of larger NFPA was associated with more pituitary hormone deficien-cies (p<0.005, data not shown). els were measured using a chemiluminescent enzyme immunoassay kit (Access hGH kit, Beckman Coulter, Tokyo) with a sensitivity of 0.002 ng/mL and CV val-ues of inter- and intra-assays of <20% and <15%, respectively. Serum LH and FSH levels were deter-mined by chemiluminescent enzyme immunoassay (Access LH kit and Access FSH kit, Beckman Coulter, Tokyo). Both kits had a sensitivity of 0.2 IU/mL, with inter-assay CV of <25% and intra-assay CV of <10%. Serum total testosterone levels were measured by chemiluminescent enzyme immunoassay (Access Testosterone kit, Beckman Coulter, Tokyo), with a sensitivity of 0.1 ng/mL, intra- and inter-assay CV of <10%. Serum TSH levels were assessed by immunoen-zymometric assay method (TOSOH-II ST AIA-PACK TSH, Tosoh Bioscience Inc.), with intra- and inter-assay CV of <5.0%. Serum FT4 levels were measured by enzyme immunoassay (TOSOH-II ST AIA-PACK FT4, Tosoh Bioscience Inc., Tokyo) with CV values of intra- and inter-assays of <3.9%. Serum PRL was measured by chemiluminescent enzyme immunoassay (SphereLight Prolactin kit, Wako, Tokyo, from January 2006 to May 2012 and Access Prolactin kit, Beckman Coulter, Tokyo, from April 2012 to November 2013). The former kit has a sensitivity of 0.5 ng/mL, with intra-assay CV of <10% and inter-assay CV of <15%, while the latter kit has a sensitivity of 0.25 ng/mL, with intra-assay CV of <1.54% and inter-assay CV of <4.23%. The PRL level (A) measured by the former kit was converted to PRL values (B) for the latter using the formula; B=(0.92×A)+2.1 (r=0.98). Statistical analysisData of clinical characteristics are presented as median with interquartile range (IQR, first and third quartiles). Age, body mass index (BMI) were com-pared between each group based on tumor diameter by Kruskal-Wallis test and Steel-Dwass test. Sex and prevalence (visual field defect, hypopituitarism and each pituitary hormone deficiency) were analyzed by Cochran-Armitage test. Because of a few Patients with NFPA ≤9 mm (n=3) and ≥40 mm (n=2), clinical characteristics were com-pared between patients with tumor diameter of 1-19 mm, 20-29 mm and ≥30 mm. Prevalence according to tumor volume were analyzed by inter-tertile range (first tertile 123-3,308 mm3, second tertile 3,309-6,350 mm3, third tertile 6,351-30,535 mm3). Pituitary stalk angle was sorted by inter-tertile range (first tertile 5.4-5Adrenocortical dysfunction in small NFPAnosed as secondary hypothyroidism. The sensitivity, specificity and optimal cut-off value of tumor diam-eter were 88.9%, 73.1% and 25.9 mm, respectively, for GH deficiency (AUC 0.888), 92.3%, 75.0% and 24.1 mm, respectively, for male secondary hypogo-nadism (AUC 0.817), 83.3%, 47.7% and 24.1 mm, respectively, for secondary hypothyroidism (AUC 0.654) and 71.4%, 72.7% and 29.4 mm, respectively, for PRL deficiency (AUC 0.693). The cut-off value not to miss the patient with GH deficiency, male sec-Prevalence of hypopituitarism according to tumor diameterSignificant high prevalence of GH deficiency, male secondary hypogonadism, secondary hypothy-roidism and PRL deficiency in the group of larger tumor diameter were demonstrated (Fig. 2, p<0.0001, p<0.05, p<0.05, and p<0.05, respectively). Among patients with NFPA smaller than 20 mm, none of them was diagnosed as GH deficiency and male sec-ondary hypogonadism. Only one of them was diag-Table 1 Clinical characteristics of participating patients according to tumor diameterTumor diameter (mm)1-1920-29≥30p valueSex (F/M)3/1010/188/14NSAge (year)49 (38.5-54.5)64.5 (56.0-70.8)56.5 (46.5-72.3)* p<0.05BMI (kg/m2)24.1 (22.1-25.9)23.3 (21.4-24.6)24.2 (21.4-27.2)NSVisual field defect8%54%100%p<0.0001Hypopituitarism38%61%82%p<0.01Data are median values (1st and 3rd quartiles) or percentage of patients (affected/all evaluated patients). BMI, body mass index; NS, not significant. Hypopituitarim represent at least one anterior pituitary hormone deficiency such as secondary adrenocortical deficiency, growth hormone deficiency, secondary hypogonadism and secondary thyroidism. A patient with 17.5 mm NFPA were excluded from subjects of visual defect by glaucoma. * With respect to age, the group of tumor diameter smaller than 20 mm was younger than the group of 20-29 mm tumor (p<0.05). Secondary adrenocortical insufficiencyGH deficiency Male secondary hypogonadismSecondary hypothyroidismPRL deficiency0204060801001-19 20-291-19 20-291-19 20-291-19 20-291-19 20-29Prevalence (%)Tumor diameter (mm)≥30≥30≥30≥30≥3043%(9/21)15%(4/26)10%(1/10)47%(9/19)29%(8/28)8%(1/12)57%(4/7)60%(9/15)0%(0/6)94%(16/17)53%(10/19)0%(0/8)71%(12/17)47%(9/19)38%(3/8)Fig. 2 Prevalence of hypopituitarism in NFPA patients with different tumor size (secondary adrenocortical insufficiency [white bar]: n=44, GH deficiency [gray bar]: n=44, male secondary hypogonadism [black bar]: n=29, secondary hypothyroidism [dotted bar]: n=60, PRL deficiency [horizontal-striped bar]: n=58) Data are percentage of patients and affected/all evaluated patients. Significant high prevalence of GH deficiency, male secondary hypogonadism, secondary hypothyroidism and PRL deficiency in the group of larger tumor diameter were demonstrated by Cochran-Armitage test (p<0.0001, p<0.05, p<0.05 and p<0.05, respectively). 6Mukai et al.ondary hypogonadism, secondary hypothyroidism and PRL deficiency was 20.6 mm (sensitivity 100%, specificity 50.0%), 21.1 mm (sensitivity 100%, spec-ificity 62.5%), 19.3 mm (sensitivity 100%, specific-ity 28.5%) and 16.9 mm (sensitivity 100%, specific-ity 18.1%), respectively. In contrast, prevalence of secondary adrenocortical insufficiency tended to be increased in the group of larger NFPA but did not significantly depend on tumor diameter (p=0.07). Thirty-eight percent of patients with NFPA smaller than 20 mm were diagnosed as secondary adreno-cortical insufficiency. The sensitivity, specificity and optimal cut-off value of tumor diameter were 66.7%, 75.0% and 25.9 mm, respectively, for secondary adre-nocortical insufficiency (AUC 0.666). The cut-off value not to miss the patient with secondary adreno-cortical insufficiency was 9.0 mm (sensitivity 100%, specificity 0%). Prevalence of hypopituitarism according to other clinical features and image parametersSignificant high prevalence of GH deficiency, sec-ondary hypothyroidism and PRL deficiency in NFPA group of larger tumor volume were demonstrated (Fig. 3A, p<0.0001, p<0.05 and p<0.05, respectively). Lateral tumor development (Knosp grade) was related with GH deficiency and male secondary hypogo-nadism (Fig. 3B, p<0.01 and p<0.01, respectively). Pituitary stalk angle was only related with GH defi-ciency (Fig. 3C, p<0.01). Hyperprolactinemia had no relationship with any pituitary hormone deficiency (Fig. 4A). Visual field defect was only related to GH deficiency (Fig. 4B, p<0.005). Secondary adrenocor-tical insufficiency was not significantly related with any clinical features and image parameters (tumor volume: p=0.20, Knosp grade: p=0.20, pituitary stalk angle: p=0.29, hyperprolactinemia: p=1.00, visual field defect: p=0.33) (Fig. 3, 4). Secretory capacity of pituitary hormone according to tumor diameterFig. 5 shows the correlation between pituitary hor-mones secretion capacity and tumor diameter. Peak plasma ACTH and serum GH in the ITT (n=44), rela-tive serum LH and serum FSH in the LHRH test (n=27 men), and relative serum PRL in the TRH test (n=43) correlated negatively with tumor diameter (ACTH: r=-0.40, p<0.01, GH: r=-0.57, p<0.0001, LH: r=-0.69, p<0.0001, FSH: r=-0.46, p<0.05, PRL: r=-0.36, p<0.05). In contrast, there was no correlation between ΔTSH and tumor diameter (n=52, p=0.52). ΔTSH related negatively with peak cortisol in the ITT (n=41, r=-0.39, p<0.05: Fig. 6). DiscussionIn the present study, prevalence of GH deficiency, male secondary hypogonadism, secondary hypothy-roidism and PRL deficiency were significantly higher in NFPA group of larger tumor diameter (Fig. 2: p<0.0001, p<0.05, p<0.05 and p<0.05, respectively). However, that of secondary adrenocortical insuffi-ciency only tended to be higher without statistical sig-nificance by the diameter of NFPA (Fig. 2: p=0.07). We also demonstrated the correlation between the secretion capacity of each anterior pituitary hormone and tumor diameter, based on the hormone stimulat-ing tests (Fig. 5, ACTH: r=-0.40, p<0.01, GH: r=-0.57, p<0.0001, LH: r=-0.69, p<0.0001, FSH: r=-0.46, p<0.05, PRL: r=-0.36, p<0.05). Our study analyzed each anterior pituitary hormone deficiency according to NFPA size for the first time. Differently from a pre-vious study that analyzed the prevalence of hypopitu-itarism in NFPA [12], we appropriately confirmed the diagnosis of secondary adrenocortical insufficiency and GH deficiency by ITT, which should be the golden standard method for diagnosis. Prevalence of secondary adrenocortical insuf-ficiency only tended to be increased in the group of larger tumor diameter although peak plasma ACTH in the ITT was correlated with tumor diameter. Moreover, as much as 38% of patient were diagnosed as secondary adrenocortical insufficiency among patients with NFPA smaller than 20 mm. In addition, the specificity was 0% when we set the cut-off value not to miss the patient with secondary adrenocorti-cal insufficiency (the cut-off value of NFPA size: 9.0 mm). This result indicated that it was difficult to iden-tify secondary adrenocortical insufficiency by NFPA size. Hypothalamic-pituitary-adrenocortical axis plays an important role for survival, so missing diag-nosis of secondary adrenocortical insufficiency should be avoided. Our results suggested that ITT should be performed even in patients with small NFPA to the diagnosis of secondary adrenocortical insufficiency. On the other hand, all with NFPA smaller than 20 mm reached peak cortisol levels more than 15 μg/dL although the levels did not fulfill the diagnostic cut-7Adrenocortical dysfunction in small NFPA020406080100123ND123ND123ND123ND123NDPrevalence (%)Angle020406080100123123123123123Prevalence (%)Tumor volume0204060801000IIIIII IV0IIIIII IV0IIIIII IV0IIIIII IV0IIIIII IVPrevalence (%)Knosp gradeABCSecondary adrenocortical insufficiencyGH deficiencyMale secondary hypogonadismSecondary hypothyroidismPRL deficiencyp=0.20p<0.0001p=0.27p<0.05p<0.05p=0.20p<0.01p<0.01p=0.23p=0.06p=0.29p<0.01p=0.15p=0.20p=0.06Fig. 3 Relationship between pituitary tumor image and prevalence of secondary adrenocortical insufficiency (white bar, n=44), GH deficiency (gray bar, n=44), male secondary hypogonadism (black bar, n=29), secondary hypothyroidism (dotted bar, n=60) and PRL deficiency (horizontal-striped bar, n=58) The trends of prevalence according to pituitary tumor image were analyzed by Cochran-Armitage test. (A) Significant high prevalence of GH deficiency, secondary hypothyroidism, PRL deficiency in NFPA group of larger tumor volume (p<0.0001, p<0.05 and p<0.05, respectively). 1: first tertile 123-3,308 mm3, 2: second tertile 3,309-6,350 mm3, 3: third tertile 6,351-30,535 mm3. (B) Significant high prevalence of GH deficiency and male secondary hypogonadism in NFPA group with more lateral tumor development (Knosp grade) (p<0.01 and p<0.01, respectively). (C) Significant high prevalence of GH deficiency with larger pituitary stalk angle (p<0.01). 1: first tertile 5.4-28.1 degree, 2: second tertile 28.2-44.9 degree, 3: third tertile 45-69.7 degree. ND: not detectable pituitary stalk angle. 8Mukai et al.off of secondary adrenocortical insufficiency (peak cortisol ≥18 μg/dL). A previous study on prolacti-noma but not NFPA showed that the group of larger tumor had significantly high prevalence of second-ary adrenocortical insufficiency which was defined as peak cortisol after ACTH stimulation test <16.3 μg/dL (450 nmol/L) [25]. Similarly in this study, the pro-portion of subjects with cortisol peak less than 15 μg/dL was significantly high in the group of larger NFPA (Supplemental Fig. 1: p<0.05). This result corre-sponded to negative correlation between peak plasma ACTH and tumor diameter. Thus, secondary adre-nocortical insufficiency associated with small NFPA might be considered as mild secondary adrenocortical insufficiency whose peak cortisol was under 18 μg/dL but over 15 μg/dL. AB020406080100GH deficiencyPrevalence (%)Non-hyperprolactinemiaHyperprolactinemia020406080100GH deficiencyPRL deficiencyPrevalence (%)Normal visual fieldVisual field defectp=1.00Secondaryadrenocorticalinsufficiencyp=1.00Male secondaryhypogonadismp=0.63Secondaryhypothyroidismp=0.32p=0.33Secondaryadrenocorticalinsufficiencyp<0.005Male secondaryhypogonadismp=0.34Secondaryhypothyroidismp=0.15p=0.11Fig. 4 Difference in prevalence of anterior pituitary hormone deficiency according to hyperprolactinemia and visual field defect (A) Hyperprolactinemia had no relationship with any pituitary hormone deficiency (Secondary adrenocortical insufficiency: n=44, p=1.00, GH deficiency: n=44, p=1.00, male secondary hypogonadism: n=29, p=0.63, secondary hypothyroidism: n=60, p=0.32). Non-hyperprolactinemia: white bar, hyperprolactinemia: left hatched bar. (B) Visual field defect was related to GH deficiency (Secondary adrenocortical insufficiency: n=44, p=0.33, GH deficiency: n=44, p<0.005, male secondary hypogonadism: n=28, p=0.34, secondary hypothyroidism: n=59, p=0.15, PRL deficiency: n=57, p=0.11). Normal visual field: white bar, visual field defect: right hatched bar. 9Adrenocortical dysfunction in small NFPAAFCDEB1101000102030405060Relative LH (fold)Tumor diameter (mm)1101000102030405060Tumor diameter (mm)0.11101000102030405060Tumor diameter (mm)101001,0000102030405060Peak ACTH (pg/mL)Tumor diameter (mm)0.010.11101000102030405060Peak GH (ng/mL)Tumor diameter (mm)1100102030405060Relative FSH (fold)Tumor diameter (mm)r=-0.40p<0.01r=-0.57p<0.0001r=-0.69p<0.0001r=-0.46p<0.05ΔTSH (μU/mL)p=0.52Relative PRL (fold)r=-0.36p<0.05Fig. 5 Correlations between tumor diameter and response of pituitary hormone in the insulin tolerance test (ITT), LHRH test and TRH test Regression line: the best fit for the relation between the two variables. (A) Tumor diameter and peak ACTH level in ITT (n=44, r=-0.40, p<0.01). (B) Tumor diameter and peak GH level in ITT (n=44, r=-0.57, p<0.0005). (C), (D) Tumor diameter and relative level of LH and FSH in the LHRH test (n=27 men, LH: r=-0.69, p<0.0001, FSH: r=-0.46, p<0.05). (E) Tumor diameter and ΔTSH in the TRH test (n=52, p=0.52). (F) Tumor diameter and relative PRL level in the TRH test (n=43, r=-0.36, p<0.05). 10Mukai et al.In respect of GH and LH/FSH, there was negative correlation between their secretion capacity and tumor diameter. Correspondingly, we also confirmed a sig-nificantly high prevalence of GH deficiency and male secondary hypogonadism in the group of larger NFPA. These results indicated hypopituitarism of these axes was byproduct of tumor mass effect. Among patients with NFPA smaller than 20 mm, none of them was diagnosed as GH deficiency and male secondary hypo-gonadism. In addition, the cut-off value of tumor diam-eter not to miss the patient with GH deficiency and male secondary hypogonadism was about 20 mm. The findings might lead us to a proactive examination of GH deficiency and secondary hypogonadism in NFPA larger than 20 mm. This would be useful not only at first image of pituitary in all NFPA patents but also at follow-up one in case of NFPA patents who were observed without pituitary operation. The present study also showed that the prevalence of secondary hypothyroidism significantly increased in the group of larger NFPA. Only one patient, whose tumor diameter was 19.3 mm, was diagnosed as sec-ondary hypothyroidism among patients with NFPA smaller than 20 mm. Thus, the cut-off value of tumor diameter not to miss the patient with secondary hypo-thyroidism was almost 20 mm. Therefore, NFPA larger than 20 mm might be proactively examined for sec-ondary hypothyroidism. Differently from GH and LH/FSH secretion capacity, there was no correlation between ΔTSH and tumor diameter. On the other hand, ΔTSH correlated negatively with peak cortisol (Fig. 6). Previous studies reported an important role of glu-cocorticoid in TSH secretion [26, 27]. One previous study found high TSH levels in Addison’s disease and that the levels returned to normal after replacement of glucocorticoids [28]. TSH response to TRH might be influenced by adrenocortical functional status, which might be the reason why there was the discrepancy between prevalence of secondary hypothyroidism and TSH response according to tumor diameter. We also analyzed the prevalence of each anterior pituitary hormone deficiency with clinical features and image parameters in addition to tumor diameter (Fig 3, 4). Similar to tumor diameter, tumor volume related with prevalence of anterior pituitary hormone defi-ciency except male secondary hypogonadism. In our patients, there were strong correlation between tumor diameters and volumes (r=-0.78, p<0.0001, data not shown). Theoretically, the tumor volumes but not the tumor diameter is more proper considering that tumor progression is three-dimensional. However our study investigated mainly the relation between tumor diam-eter and prevalence of each anterior pituitary defi-ciency because the tumor diameter is easier to mea-sure in clinical practice. On the other hand, other parameters were related with only a few anterior pitu-itary deficiencies. Knosp grade, hyperprolactinemia, visual field defect and pituitary stalk angle were asso-ciated with suprasellar or lateral tumor progression. These features and parameters described above might not fully reflect the damage of anterior pituitary gland because these parameters evaluated tumor position but not the position of anterior pituitary gland. All of our clinical features and image parameters had no signif-icant relation with secondary adrenocortical insuffi-ciency. It was suggested that this insufficiency was not predictable based on our clinical features or image parameters. Therefore, ITT should be performed for the exact diagnosis. The prevalence of hypopituitarism in our study, which was defined as at least one anterior pituitary defi-ciency, was less than those reported in previous studies (69-89%) [12, 29]. It is possible that the difference is due to differences in the methods used for evaluation of hypopituitarism. For example, we diagnosed sec-ondary adrenocortical insufficiency or GH deficiency by ITT, the golden standard test, while previous stud-ies used various other methods including basal serum cortisol level, CRH test and ACTH test or low serum IGF-1 level. However, the prevalence of hypopituita-110100051015202530ΔTSH (μU/mL)Peak cortisol (μg/dL)r=-0.39p<0.05Fig. 6 Correlations between serum peak cortisol during insulin tolerance test (ITT) and ΔTSH in the TRH test Regression line: the best fit for the relation between the two variables (n=41, r=-0.39, p<0.05). 11Adrenocortical dysfunction in small NFPArism in our study was significantly lower in patients with smaller NFPA, in agreement with the finding of a previous study [12]. The present study has certain limitations. We did not investigate secondary hypogonadism in women in the present study due to the lack of information on the menstrual state for all patients and the small number of premenopausal women. Further studies are neces-sary to determine the relation between female second-ary hypogonadism and NFPA size. Our study subjects who have been recruited for evaluation of pituitary functions and surgical indication were another limita-tions. It was reported that patients with microadenoma are rarely evaluated in terms of hypopituitarism [30]. Accordingly, the number of microadenomas was small in this study (only three patients). In addition, as the limitation of recruitment, the group of tumor diame-ter smaller than 20 mm was younger than the group of 20-29 mm tumor (Table 1). Because younger patients have long lifetime than older patients, the former has more probability of tumor progression. Therefore, pituitary surgery might be recommended for the former [30]. Also, there are significant risks of surgical proce-dure in older patients. These might affect recruitment of the subjects. Although there were these limitations, the correlation between GH, LH/FSH secretion capac-ity and tumor diameter, that is, the etiology of mass effect mentioned above, may suggest less prevalence of GH deficiency and secondary hypogonadism in pitu-itary microadenoma. In conclusion, physicians should proactively exam-ine pituitary function in patients with NFPA measur-ing ≥20 mm, in particular, GH deficiency, male sec-ondary hypogonadism and secondary hypothyroidism. Differently from GH deficiency, male secondary hypo-gonadism and secondary hypothyroidism, secondary adrenocortical insufficiency did not significantly cor-relate with NFPA size. Our results also suggest that examination of adrenocortical function is warranted even in patients with small NFPA. DisclosureThe authors declare no conflict of interest. 0204060801001-1920-29Prevalence (%)Tumor size (mm)0%(0/8)28%(5/19)≥3050%(8/17)Supplemental Fig. 1 Proportion of NFPA patients with peak cortisol less than 15 μg/dL Data are percentage of patients and affected/all evaluated patients (n=44). Significantly higher proportion in the group of larger NFPA were demonstrated by Cochran-Armitage test (p<0.05). References 1. Dekkers OM, Pereira AM, Romijn JA (2008) Treatment and follow-up of clinically nonfunctioning pituitary mac-roadenomas. J Clin Endocrinol Metab 93: 3717-3726. 2. Rosén T, Bengtsson BÅ (1990) Premature mortality due to cardiovascular disease in hypopituitarism. Lancet 336: 285-288. 3. Bülow B, Hagmar L, Mikoczy Z, Nordström CH, Erfurth EM (1997) Increased cerebrovascular mortality in patients with hypopituitarism. Clin Endocrinol (Oxf) 46: 75-81. 4. Tomlinson JW, Holden N, Hills RK, Wheatley K, Clayton RN, et al. (2001) Association between premature mortal-ity and hypopituitarism. Lancet 357: 425-431. 5. Schneider HJ, Aimaretti G, Kreitschmann-Andermahr I, Stalla GK, Ghigo E (2007) Hypopituitarism. Lancet 369: 1461-1470. 6. Ascoli P, Cavagnini F (2006) Hypopituitarism. Pituitary 9: 335-342. 7. Matsuura I, Saeki N, Kubota M, Murai H, Yamaura A (2001) Infarction followed by hemorrhage to endocrine stimulation test. Endocr J 48: 493-498. 8. Lee DH, Chung MY, Chung DJ, Kim JM, Lee TH, et al. (2000) Apoplexy of pituitary macroadenoma after com-bined test of anterior pituitary function. Endocr J 47: 329-333. 9. Erturk E, Jaffe CA, Barkan AL (1998) Evaluation of the 12Mukai et al.integrity of the hypothalamic-pituitary-adrenal axis by insulin hypoglycemia test. J Clin Endocrinol Metab 83: 2350-2354.10. Grossman AB (2010) The diagnosis and management of central hypoadrenalism. J Clin Endocrinol Metab 95: 4855-4863.11. Molitch ME, Clemmons DR, Malozowski S, Merriam GR, Shalet SM, et al. (2006) Evaluation and treatment of adult growth hormone deficiency: An Endocrine Society Clinical Practice Guideline. J Clin Endocrinol Metab 91: 1621-1634.12. Nomikos P, Ladar C, Fahlbusch R, Buchfelder M (2004) Impact of primary surgery on pituitary function in patients with non-functioning pituitary adenomas - a study on 721 patients. Acta Neurochir (Wien) 146: 27-35.13. Isojima T, Shimatsu A, Yokoya S, Chihara K, Tanaka T, et al. (2012) Standardized centile curves and reference intervals of serum insulin-like growth factor-I (IGF-I) levels in a normal Japanese population using the LMS method. Endocr J 59: 771-780.14. Delgrange E, Trouillas J, Maiter D, Donckier J, Tourniaire J (1997) Sex-related difference in the growth of prolactinomas: a clinical and proliferation marker study. J Clin Endocrinol Metab 82: 2102-2107.15. Arlt W, Allolio B (2003) Adrenal insufficiency. Lancet 361: 1881-1893.16. Hoffman DM, O’Sullivan AJ, Baxter RC, Ho KK (1994) Diagnosis of growth hormone deficiency in adults. Lancet 343: 1065-1068.17. Melmed S, Kleinberg D, Ho K (2011) Pituitary phys-iology and diagnostic evaluation. In: Melmed S (ed) Williams Textbook of Endocrinology (12th). Saunders, Philadelphia, USA: 218.18. Silveira LF, Latronico AC (2013) Approach to the patient with hypogonadotropic hypogonadism. J Clin Endocrinol Metab 98: 1781-1788.19. Persani L (2012) Central hypothyroidism: Pathogenic, diagnostic, and therapeutic challenges. J Clin Endocrinol Metab 97: 3068-3078.20. Arafah BM, Manni A, Brodkey JS, Kaufman B, Velasco M, et al. (1981) Cure of hypogonadism after removal of prolactin-secreting adenomas in men. J Clin Endocrinol Metab 52: 91-94.21. Bassiri RM, Utiger RD (1973) Metabolism and excre-tion of exogenous thyrotropin-releasing hormone in humans. J Clin Invest 52: 1616-1619.22. Kelch RP, Markovs M, Huss J (1976) LH and FSH responsiveness to intravenous gonadotropin-releas-ing hormone (GnRH) in children with hypothalamic or pituitary disorders: lack of effect of replacement therapy with human growth hormone. J Clin Endocrinol Metab 42: 11049-1113.23. Arita H, Kinoshita M, Oshino S, Kitamura T, Otsuki M, et al. (2012) Biological characteristics of growth hormone-producing pituitary adenomas are different according to responsiveness to thyrotropin-releasing hormone. J Clin Endocrinol Metab 97: 2741-2747.24. Knosp E, Steiner E, Kitz K, Matula C (1993) Pituitary Adenomas with Invasion of the Cavernous Sinus Space: A Magnetic Resonance Imaging Classification Compared with Surgical Findings. Neurosurgery 33: 610-618.25. Tirosh A, Benbassat C, Lifshitz A, Shimon I (2015) Hypopituitarism patterns and prevalence among men with macroprolactinomas. Pituitary 18: 108-115.26. Nicoloff JT, Fisher DA, Appleman MD Jr (1970) The role of glucocorticoids in the regulation of thyroid func-tion in man. J Clin Invest 49: 1922-1929.27. Re RN, Kourides IA, Ridgway EC, Weintraub BD, Maloof F (1976) The effect of glucocorticoid adminis-tration on human pituitary secretion of thyrotropin and prolactin. J Clin Endocrinol Metab 43: 338-346.28. Ismail AA, Burr WA, Walker PL (1989) Acute changes in serum thyrotrophin in treated Addison’s disease. Clin Endocrinol (Oxf) 30: 225-230.29. Dekkers OM, Pereira AM, Roelfsema F, Voormolen JHC, Neelis KJ, et al. (2006) Observation alone after transs-phenoidal surgery for nonfunctioning pituitary macroad-enoma. J Clin Endocrinol Metab 91: 1796-1801.30. Freda PU, Beckers AM, Katznelson L, Molitch ME, Montori VM, et al. (2011) Pituitary incidentaloma: An endocrine society clinical practice guideline. J Clin Endocrinol Metab 96: 894-904.View publication stats | RESEARCH GATE |
Full Terms & Conditions of access and use can be found athttps://www.tandfonline.com/action/journalInformation?journalCode=idrd20Drug DeliveryISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/idrd20Intra-articular injection PLGA blends sustained-release microspheres loaded with meloxicam:preparation, optimization, evaluation in vitro andin vivoZheng Sun, Xuejing Gu, Teng Hao, Jiali Liu, Rongrong Gao, Yanli Li, Bin Yu &Hui XuTo cite this article: Zheng Sun, Xuejing Gu, Teng Hao, Jiali Liu, Rongrong Gao, Yanli Li, BinYu & Hui Xu (2022) Intra-articular injection PLGA blends sustained-release microspheres loadedwith meloxicam: preparation, optimization, evaluation invitro and invivo, Drug Delivery, 29:1,3317-3327, DOI: 10.1080/10717544.2022.2144545To link to this article: https://doi.org/10.1080/10717544.2022.2144545© 2022 The Author(s). Published by InformaUK Limited, trading as Taylor & FrancisGroup.Published online: 11 Nov 2022.Submit your article to this journal View related articles View Crossmark dataRESEARCH ARTICLEDrug Delivery2022, vOl. 29, NO. 1, 3317–3327Intra-articular injection PLGA blends sustained-release microspheres loaded with meloxicam: preparation, optimization, evaluation in vitro and in vivoZheng Sun , Xuejing Gu, Teng Hao, Jiali Liu , Rongrong Gao, Yanli Li , Bin Yu and Hui Xu School of Pharmacy, Collaborative innovation Center of Advanced Drug Delivery System and Biotech Drugs in universities of Shandong, Key laboratory of Molecular Pharmacology and Drug evaluation (yantai university), Ministry of education, yantai university, yantai, ChinaABSTRACTMeloxicam (MLX) is a commonly used drug in the clinical treatment of osteoarthritis, but it is associated with gastrointestinal adverse reactions. Therefore, in this study, we developed a sustained-release microsphere formulation of MLX for topical administration of knee joint. The MLX-loaded PLGA microspheres (MLX-MS) were prepared by emulsion solvent evaporation method with optimization of formulation using orthogonal experimental design. Physicochemical characterization results show MLX-MS were spherical with a smooth surface, the particle size was about 100 μm, drug loading was 30%, and encapsulation efficiency was 76.8%. In addition, the in vivo pharmacokinetics, tissue distribution, and pharmacodynamics were evaluated in rats by intra-articular administration of MLX. The microspheres showed a typical long-term sustained release pattern with a low initial burst release. In contrast to oral administration, local injection of MLX-MS produced a much higher value of elimination half-life time(T1/2) and peak time (Tmax) in plasma, while the intestinal drug distribution was significantly decreased. MLX-MS could also cause a greater reduction in the body level of IL-6 and TNF-α, which was positively correlated with R2=0.981. A good linear relationship (R2 = 0.9945) between the in vitro and in vivo drug release from MLX-MS could be observed, bivariate correlation analysis. All the findings demonstrated that local administration of MLX-MS can prolong the action time of MLX and reduce side effects, thus would be a promising preparation for the treatment of arthritis.1. IntroductionOsteoarthritis (OA) is a common complex and multi-etiologic degenerative joint disease, which is characterized by a series of degenerative diseases such as joint local inflammation, cartilage degradation, and subchondral bone sclerosis. The global incidence rate is more than 3%, which is a serious threat to human health (Cross et al., 2014; Hanafy, & El-Ganainy, 2020). At present, there is still a lack of radical treatment for OA. In addition to surgical intervention, non-steroidal anti-inflammatory drugs (NSAIDs) are often used in clinical practice to alleviate arthritis-related swelling, stiff-ness, and pain. However, gastrointestinal reactions and other side effects are widespread and may increase the risk of cardiovascular disease (Laavola et al., 2018; Tang, 2019). Among them, meloxicam (MLX) is a selective inhibitor of cyclooxygenase (Cox)-2. Its intensity of action on Cox-2 in vitro is more than 10 times of Cox-1, which is lower than © 2022 The Author(s). Published by informa uK limited, trading as Taylor & Francis group.CONTACT Hui Xu xuhui33@sina.com School of Pharmacy, yantai university, No. 32 Qingquan road, 264005 yantai, Shandong, Pr China.https://doi.org/10.1080/10717544.2022.2144545This is an Open Access article distributed under the terms of the Creative Commons Attribution license (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.ARTICLE HISTORYReceived 27 September 2022Revised 31 October 2022Accepted 1 November 2022KEYWORDSMeloxicam; microsphere; PLGA blends; topical administration; osteoarthritis; anti-inflammatory3318Z. SUN ET AL.the risk of gastrointestinal mucosal injury caused by other NSAIDs and is commonly used in clinical symptoms of OA. Nevertheless, the conventional preparations of MLX still have serious adverse reactions such as gastric bleeding, headache, and rash when they are administered orally with high doses and long-term treatment (Patoia et al., 1996).Compared with oral administration, local administration can not only reduce the side effects on the gastrointestinal tract but also avoid the first-pass effect and improve the bioavailability of the system. Given the use of NSAIDs in the treatment of mild to moderate OA pain, the European Union against Rheumatism (EULAR) and the National Institute for Health and Nursing Excellence (NICE) recommended the treat-ment plan of local administration before oral administration (Rodriguez-Merchan, 2018). Therefore, NSAID long-acting sustained-release drug delivery system for local administra-tion has attracted extensive research attention.Microspheres, as a new drug delivery system, usually refer to a particle dispersion system formed by drug adsorption or dispersion in the polymer matrix, with a diameter range of 1–1000 μm (Karan et al., 2020). Because of its rich structure and function, it has many advantages. For example, it can be administered in various ways, including subcutaneous injection, oral administration, and lung administration as an inhalation (Su et al., 2021; Chen & Fang, 2000). Moreover, as a commonly used drug delivery system, it can prolong the action time of drugs in the body, reduce the number of administrations, and improve the compliance of patients (Singh Malik et al., 2016). Poly (lactic-co-glycolic acid) (PLGA) is one of the most widely used drug delivery carriers, which has good biocompatibility and biodegradability. Therefore, the preparation of meloxicam sustained-release microspheres with PLGA carrier for local administration in the treatment of arthritis has good safety, which not only can achieve the long-term anti-inflammatory effect but also can reduce the gastrointestinal adverse reactions of NASID.In this study, the S/O/W emulsion-solvent evaporation method was used to prepare MLX long-term sustained-release microspheres (MLX-MS) suitable for local injection with PLGA as the carrier. The key prescription factors and preparation process were screened and optimized, and the obtained microspheres were systematically characterized in vitro and evaluated in drug release characteristics, providing an exper-imental basis for the development of new preparations suit-able for OA clinical practice.2. Material and methods2.1. Reagent and animalsMLX was purchased in Shandong Chenghui Shuangda Pharmaceutical Co., Ltd.; PLGA was provided by Shandong Lvye Pharmaceutical Co., Ltd. Polyvinyl alcohol (PVA) was pur-chased from German EMD Company; other reagents are from the national pharmaceutical chemical reagent company.Male Wistar rats (200 ± 10 g, SCK20190003) were supplied by Jinan Pengyue Laboratory Animal Breeding Co., Ltd. All animals are strictly handled following the Care and Use of Laboratory Animals (NIH Publication No. 85-23, revised 1996) and approved by the Animal Experimentation Ethics Committee of Yantai University (No. 20220302).2.2. Preparation of microspheresMicrospheres were prepared by an emulsification-solvent evaporation method. A certain amount of PLGA was dissolved in dichloromethane, added to the appropriate amount of MLX, ultrasonic 3 min, obtain drug-containing suspension; then added drop by drop to the water phase 1% PVA, and stirred at 1200 rpm for 5 min. Add 0.1% PVA, solidify at 500 rpm/min for 5 h, make the solvent fully volatilize, and screen, the microspheres were collected and washed with distilled water. The resulting microspheres were finally freeze-dried and stored at 4 °C.2.3. Characterization of microspheres2.3.1. Morphology and particle sizeThe surface morphology of MLX-MS was observed by SEM (EM-30PLUS, Coxem, Germany). Installation of freeze-dried microspheres on metal columns with sticky carbon strips. The average particle size and size distribution of microspheres were measured by Malvern 3000 laser particle size analyzer (Malvern Instruments Ltd., Malvern, UK), and the dispersion index (Span) was calculated to evaluate the size distribution.2.3.2. Drug loading and entrapment efficiencyThe drug loading (DL) and encapsulation efficiency (EE) of MLX-MS were determined by HPLC. Briefly, accurately weighed 20 mg of microspheres were dissolved in 10 mL of methylene chloride. The sample was diluted 10-fold with mobile phase (methanol: 0.1% formic acid water = 70:30), and 6000 rpm/min centrifugal 5 min, collect supernatant to determine the concentration of MLX with waters e2695 chromatographic system (Waters Corporation, Massachusetts, USA). Agilent TC-C18 column was used with a column temperature of 35 °C, the flow rate of 1 mL/min. The detection wavelength was set at 362 nm, the injection volume is 10 μL. DL%amount of drug in microspheresamount of microspheres��100 EE%amount of drug in microspherestheoretical mass of drug in micro�shperes �100 2.3.3. In vitro release studyThe in vitro release was determined by dialysis. 5 mg MLX-MS were transferred to a dialysis bag (8–14 kDa) and placed in a 50 mL centrifuge tube, adding release medium (0.1% Tween 80 pH 7.4 PBS) 20 mL. The microsphere suspension was shaken horizontally at 60 rpm in a shaker maintained 37 ± 0.5 °C, remove 500ul release medium at different times and immediately add fresh, isothermal, and equal volume release medium. The drug concentration in the release DRUG DELIVERY3319medium was determined by HPLC, and the cumulative drug release (Q) was calculated.To investigate the mechanism of drug release, the in vitro release data was analyzed by fitting zero-order, first-order, Higuchi, and Ritger-Peppas equation models (Brown et al., 2006). Polymer degradation during microsphere release was also monitored by using SEM observation.2.3.4. Fourier transform infrared (FTIR) and X-ray powder diffraction (XRPD) assaysFTIR and XRPD assays were performed for MLX, PLGA, MLX-MS and the mixture of MLX and PLGA. Fourier transform infrared spectrometer (NICOLET iS10, Thermo Fisher Scientific, USA) was used and the samples were prepared by tableting with KBr at a weight ratio of 1: 100, and the scanning range was 400–4000 cm−1. XRPD (D-MAX 2500 X, Rigaku Corporation, Japan) was performed at 10°–80° (2θ), the relative strength is read from the bar graph and corrected to fix the slit value.2.4. Animal model and experimental groupingMale Wistar rats were injected with sodium iodate (MIA) into the left joint to establish the osteoarthritis model (Hanafy & EI-Ganainy,, 2020). The MIA was dissolved with normal saline and the injection volume was 100 μL (30 mg/mL). After injec-tion, the knee joints of rats were massaged to ensure uniformity.The model rats were randomly divided into two groups (n = 4), namely MLX oral administration group and the MLX-MS intra-articular administration group. Meloxicam 0.5% sodium hydroxymethyl cellulose suspension was prepared and admin-istered by gavage at a dose of 1.35 mg/kg for seven consecutive days. For the oral group, blood was collected from the inner canthus venous plexus at 0.5, 1, 2, 4, 8, 12, 24 h after the first dose and 0.5, 1, 2, 4, 8, 12 h after the last administration. The microspheres group was administered by intra-articular injec-tion with a new needle for each injection at a dose of 9.45 mg/kg once (Conaghan et al., 2018; Li et al., 2021). Blood was collected at 0.5, 1, 2, 4, 8, 12, 24, 48, 96, 120, and 144 h after administration. The blood samples were centrifuged immedi-ately at 12,000 rpm at 4 °C for 10 min and the plasma samples were then obtained and stored at −80 °C. A portion of the samples were naturally solidified at room temperature for 20 min and centrifuged at 3000 rpm at 4 °C for 10 min. The supernatant was taken and IL-6 and TNF-α inflammatory factors were detected according to the instructions.2.5. Drug quantification of biological samplesUPLC-MS/MS assay was performed to determine drug content in biosamples such as plasma and tissues by using an AB Sciex Triple QuadTM 4500 system (AB SCIEX, USA) connected with Shimadzu LC-30AD via electrospray ionization (ESI) inter-face. The mobile phase was methanol (A) and 0.1 formic acid water (B) with the following gradient elution: 90%–55% (0–0.5 min); 55%–35% (0.5–3.5 min); 35%–10% (3.5–6.5 min), column temperature 40 °C, flow rate 0.4 mL/min, injection volume 5 μL. Mass spectrometry using electrospray ion source ESI, detection mode is positive ion mode, scanning mode is multiple reaction monitoring (MRM), meloxicam m/z:352.3 →115.2, piroxicam m/z: 332.1 → 94.8.2.6. In vitro and in vivo correlation (IVIVC) analysisAccording to previous research (D'Souza et al., 2014; Laavola et al., 2018), the AUCt was calculated by the trapezoidal area method (Formula 1), and then the in vivo absorption rate was obtained by the fractional AUC method (Formula 2). AUCCC(tt )tt1212212� ���������� (1) F ( )AUCAUCatt �� (2)Fa(t) is the percentage of drug absorbed within time t, AUCt is the area under the plasma concentration-time curve at time t, and AUC∞ is the total area under the plasma concentration-time curve. Linear regression analysis was applied to fit the IVIVC plot and R2 was calculated to evaluate the IVIVC.2.7. Biochemical and histological analysesELISA assays were performed using a microplate reader (Tecan Spark, Tecan, Switzerland) for the known markers of inflammation in osteoarthritis, such as TNF-α and IL-6 in serum. For histological examination, pathological anatomy was performed on rats. Anatomy of both sides of the joint, trim the surrounding muscles. Samples were fixed in 10% formalin for 24 h, rinsed, and joints were decalcified with 5% formic acid for 72 h. Then the samples were dehydrated, embedded in paraffin, sliced and stained with HE, and observed under light microscope for histological evaluation. Furthermore, referring to the Osteoarthritis Research Society International OARSI scoring criteria (Table 1), we evaluated the histology of osteoarthritis in rats (Gerwin et al., 2010).Table 1. OArSi histopathological evaluation criteria.Histological featureScoreCartilage degenerationNormal0Minimal degeneration1Mild degeneration2Moderate degeneration3Marked degeneration4Severe degeneration5Calcified cartilage and subchondral bone damageNormal0No fragmentation of tidemark; no marrow changes1Mesenchymal change in marrow2Mesenchymal change in marrow; areas of marrow chondrogenesis may be evident3Marked to severe fragmentation of calcified cartilage; marrow mesenchymal change involves up to 3/4 of area4Synovial membrane inflammationNormal0Slight proliferation of subsynovial tissue.1Proliferation of subsynovial tissue2Proliferation of subsynovial tissue; infiltration of few inflammatory cells3Proliferation of subsynovial tissue; infiltration of large number of inflammatory cells.43320Z. SUN ET AL.2.8. StatisticsOne-way analysis of variance (ANOVA) and two-tailed Student’s t-test was performed for significance tests. A p-value less than 0.05 was considered statistically significant.3. Results and discussion3.1. Optimization of microsphere formulation3.1.1. Primary screening of PLGAA single-factor test was used to screen the type of PLGA. According to the preparation method above, MLX-MS were prepared with different kinds of PLGA. The particle size, drug loading (DL), encapsulation efficiency (EE), and span of micro-spheres are shown in Table 2. Cluster analysis is carried out according to the particle size, drug loading, and encapsulation efficiency of microspheres by the hierarchical cluster method, the nine PLGAs were clustered into three categories, PLGA 50502E and 95052A were classified into one category, 50502A and 75254A were classified into one category, and the remain-ing five categories were classified into one category.In summary, considering the encapsulation efficiency, drug loading, 50502A, 50502E, and 75252A were finally selected from the three categories for further screening.3.1.2. Orthogonal testAn orthogonal experiment was used to optimize the formulation of MLX-MS. Referring to the results of the single factor experi-ment, PLGA type, PLGA/MLX, PLGA concentration, and O/W were selected to optimize the formulation by L9(3)4 orthogonal design.The factor level table of the orthogonal test is shown in Table 3, and the three-factor three-level orthogonal table is shown in Table 4. The optimal prescription was screened with the level of encapsulation efficiency as the index. The results showed that the type of PLGA, PLGA/MLX, PLGA concentration, and O/W had significant effects on encapsulation efficiency (p < .01). According to the effect of various factors on the encapsulation efficiency (Figure 1), the optimal prescription is A3, B1, C2, and D2. To obtain higher drug loading and combined with the results of intuitive analysis, A3, B3, C2, and D2 was the optimal prescription, namely PLGA type was 7525 2A, PLGA/MX was 2:1, PLGA concentration was 100 mg/mL, O/W was 1:80.The in vitro release characteristics of the nine prescriptions were also investigated, and the results were shown in Figure 2(A). Different types of PLGA have a great influence on the drug release characteristics of microspheres. For the micro-spheres prepared by PLGA 5050 2E, MLX was tightly wrapped inside the microspheres and could not be released, the rea-son was that MLX was a hydrophobic drug and had a strong binding force with PLGA 5050 2E ester-terminated polymer, so it was easier to combine with MLX; Moreover, the slower degradation rate of ester-terminated PLGA also limited the release of drugs in the microspheres (Panyam et al., 2004; Mao et al., 2007; Félix Lanao et al., 2011). However, about 40% of the drug was released in prescription 6, which was speculated the dissolution of some unencapsulated drug on the surface of the microspheres due to the high loadingThe PLGA 7525 2A MLX-MS release slowly, because the proportion of lactide (LA) increases, which reduces the deg-radation rate of PLGA itself, and the interaction between the drug and PLGA increases with increasing dose, resulting in a slower release rate (Schliecker et al., 2003; Siegel et al., 2006). From the results of the experiment, we can find that formulations 6, 8, and 9 have obvious burst release at the initial stage of release. It’s speculated that the drug adsorbed on the surface of the microspheres is dissolved in the release medium to form a burst release, or the drug is not uniformly dispersed, drug outside the microspheres triggers burst release (Leach et al., 2005; Lee et al., 2010).3.1.3. Microspheres based on PLGA blendsFrom the release curves of different formulations, the micro-spheres prepared by PLGA 50502A have sustained-release properties in vitro, but the encapsulation efficiency is low. To obtain microspheres with higher drug loading and improved drug release characteristics, we mixed PLGA 5050 2A and 7525 2A in a certain proportion (90/10, 80/20, 70/30), and other parameters following the above optimization results.The in vitro release results were shown in Figure 2(B), MLX-MS release behavior significantly improved after mixing two different PLGA. When the proportion of PLGA 7525 2A increased, the microspheres were not released after three days, and with the increase of 5050 2A, the speed of drug release is gradually accelerated, this is because PLGA blends affected the internal structure of the microspheres (Mao et al., 2007). In summary, the optimal process is PLGA type 5050 2A (90%)/7525 2A (10%), PLGA/MX 2:1, PLGA concen-tration 100 mg/mL, and O/W 1:80.3.2. Characterization3.2.1. Morphology and particle sizeTo further verify the reliability of the optimal prescription, six batches of MLX-MS were prepared with the optimal pre-scription. The drug loading was 36.77 ± 3.5%, the Table 2. Characterization of microspheres prepared by different kinds of PlgA.PlgAParticle size(μm)Dl(%)ee(%)Span5050 1A99.2 ± 5.0523.07 ± 0.0763.27 ± 0.190.5195050 2A104.9 ± 6.2513.59 ± 0.0855.77 ± 0.330.5425050 4A112.0 ± 6.3217.22 ± 0.4359.87 ± 1.540.4715050 2e102.0 ± 1.6723.84 ± 0.0799.53 ± 0.330.9127525 1A109.0 ± 0.0722.57 ± 0.0368.60 ± 0.080.6327525 2A106.0 ± 3.8219.86 ± 0.0168.10 ± 0.010.4087525 4A138.0 ± 6.9414.47 ± 0.2973.50 ± 1.500.5108515 2A108.0 ± 6.1019.54 ± 0.0560.10 ± 0.160.4739505 2A111.0 ± 2.4122.99 ± 0.09102.10 ± 0.410.448Table 3. Orthogonal design and experimental results.levelsFactorABCD15050 2A100:7.5501:6025050 2e100:15751:8037525 2A100:501001:100Factor A: PlgA, factor B: PlgA/MlX(m/m), factor C: Concentration of PlgA (mg/ml), factor D: O/W (v/v).DRUG DELIVERY3321Table 4. Orthogonal design and experimental results (n = 3).No.ABCDParticle size(μm)ee(%)Dl(%)Span11111111 ± 0.0152.71 ± 0.175.4 ± 0.020.43021222127 ± 1.9074.63 ± 0.6712.8 ± 0.120.46731333156 ± 4.8042.63 ± 0.0233.1 ± 0.020.54842123143 ± 5.9070.18 ± 0.496.6 ± 0.050.49352231142 ± 0.4446.11 ± 0.248.5 ± 0.040.47962312109 ± 1.5070.51 ± 0.2537.8 ± 0.130.44673132127 ± 8.5086.62 ± 0.777.8 ± 0.070.48983213111 ± 1.8056.08 ± 0.1112.3 ± 0.120.46793321125 ± 7.6079.19 ± 0.2737.6 ± 0.380.492Figure 1. The impact trend of PlgA (A), PlgA/MlX (B), PlgA concentration (C), and O/W(D) on encapsulation efficiency of microspheres.Figure 2. The in vitro release profiles of (A) microspheres prepared by different polymers, and (B) PlgA blends microspheres.3322Z. SUN ET AL.encapsulation efficiency was 76.81 ± 7.13%, the particle size was 102 ± 3.90 μm, and the particle distribution was uniform.Scanning electron microscopy (SEM) was used to investi-gate the morphology of MLX-MS. The results were shown in Figure 3(A-1). The microspheres were spherical in shape, smooth and non-porous on the surface, and there is no adhesion between the microspheres, indicating that the microspheres have a good distribution. The results of particle size determination of multiple batches of microspheres by dynamic light (DLS) scattering showed (Figure 3B) that the average particle size of the microspheres was between 100 and 110 μm, and the span was 0.5–0.6, indicating that the distribution of the microspheres was uniform.3.2.2. Spectral analysisMeloxicam is a polymorph, so during the preparation of microspheres, including dissolution and desolvation processes, the crystallinity of MLX will be affected. We used the MLX/PLGA physical mixture with the same material composition as a reference and analyzed the XRPD and IR spectra under the same conditions. The characteristic peaks in the spectrum were used as indicators to investigate whether the dissolution and solvent removal process during the preparation of the microspheres affected the crystal form of the raw material. The results were shown in Figure 4(A). The peak of MLX is about 25.8°, and the broad peak of PLGA is about 19.6°. It can be seen that the characteristic peaks of MLX exist in the Figure 3. SeM observation (A) and particle size distribution (B) of MlX-MS.Figure 4. XrPD (A) and ir (B) spectra (a: PlgA, b: MlX, c: PlgA/MlX, d: MlX-MS).DRUG DELIVERY3323microspheres and the mixture of PLGA/MLX, although the response is reduced, the peaks were not broadened or shifted. The FTIR results were shown in Figure 4(B). MLX shares char-acteristic absorption peaks at 3290.56 cm−1, and PLGA shares characteristic absorption peaks at 3002.41 cm−1 and 2954.78 cm−1, these characteristic peaks exist in the physical mixture of MLX-MS and PLGA. The presence of these peaks indicates that the drug did not react with the polymer and that the drug was encapsulated in the microspheres (Erdemli et al., 2014). Therefore, it was considered that MLX was encap-sulated in PLGA and the structure did not change.3.2.3. In vitro drug releaseThe in vitro cumulative release curves of MLX-MS were shown in Figure 5. The results showed that the release of MLX-MS was biphasic, MLX rapid release within 72 h, then entered the stage of slow drug release. The rapid release is due to the increased drug concentration near the surface caused by sol-vent convection during the preparation of the microspheres, so the drug is rapidly released in the form of diffusion (Allison, 2008). The second release stage is mainly controlled by the degradation of PLGA. Drug release is due to the gradual hydro-lysis of PLGA to oligomers, which leads to the formation of pores on the surface and inside of microspheres, providing conditions for drug release through diffusion and erosion within microspheres (Makadia & Siegel, 2011).The mechanism of MLX release from the microspheres was investigated by fitting the data obtained from in vitro release (Table 5). The in vitro release of MLX-MS follows the Ritger-Peppas model, (R2 = 0.9969). The drug release charac-teristic index ‘n’ in the release curve of the Ritger-Peppas equation is analyzed, when n ≤ 0.45, the drug release mechanism is Fick diffusion, and when 0.45 < n < 0.89, the drug release mechanism is the combined effect of drug dif-fusion and matrix erosion. When n ≥ 0.89, the drug release mechanism is skeletal erosion. Therefore, the drug release mechanism of MLX-MS would be the coexistence of drug diffusion and skeletal erosion (n = 0.46) (Siepmann et al., 2005). The degradation results of the microspheres were shown in Figure 3(A2–4). The microspheres were continuously degraded in the release system. In the initial stage of release (3 days), there were tiny holes on the surface. After 2 weeks, the size of the holes increased and the microspheres grad-ually dissolved. Around 1 month, the microspheres rupture. Thus, with the continuous degradation of PLGA, the structure of microspheres changed greatly. In this process, the micro-spheres continue to dissolve, the pores increase, and the drug is gradually released from the microspheres.3.3. Therapeutic effectsThe results of changes in the circumference of the knee joint of the rats were shown in Figure 6(A). The right leg of the rat was used as a control to monitor the change of the difference in the circumference of the left after the injection of MIA. The results showed that the swelling degree of the left joint of the rats increased significantly with time, and the difference in the circumference of the joints on both sides showed a continuous and significant change with time, indicating that the inflammation model was successful. To further study the efficacy of MLX-MS in vivo, this experiment also investigated the changes in inflammatory factors in rats.It has been reported that a variety of inflammatory factors were involved in the occurrence of OA, such as IL-1β, IL-6, TNF-α, etc., which were closely related to the degeneration of articular cartilage and the production of OA (Chu et al., 2006; Kapoor et al., 2011). We evaluated the changes of inflammatory cytokines IL-6 and TNF-α in rats after oral administration and microsphere injection. The results were shown in Figure 6(B). Compared with the blank control group, the levels of inflammatory factors IL-6 and TNF-α in the rats after MIA injection were significantly increased (p < .01), indi-cating that the experimental arthritis model was successful.Within seven days of oral administration and microsphere injection, the levels of IL-6 and TNF-α in rats gradually decreased (Figure 7). For the MLX-MS intraarticular injection, the drug is released in the joint cavity and quickly reaches the site of inflammation to exert its efficacy. Moreover, due to the slow release of the drug, the level of inflammatory factors in the rat body decreases at a lower rate after 72 hours than in the oral administration group. Feature-related phar-macodynamic relationships were more pronounced on IL-6, which could be further analyzed by PK-PD.3.4. Pharmacokinetics in plasma3.4.1. Drug concentration-time curve and pharmacokinetic parametersFrom the results in Figure 8, it can be seen that the plasma drug concentration during 0.5–12 h, plasma drug concentra-tion in the oral group at each time point was significantly higher than that in the microsphere group, and soon reached Figure 5. The in vitro release profile of MlX-MS obtained under optimal conditions (n = 3).Table 5. The kinetic models simulated for the release behavior of microspheres.Modelequationr2Zero-order modelQtt ��0 088713 8271..0.8020First-order modelQett���56 21600 0094..0.9522Higuchi modelQtt ��2 53122 51681 2../0.9495ritger-Peppas modelQtt = 3 2202 0 4683..0.99693324Z. SUN ET AL.the maximum value; after continuous oral administration, MLX reached a steady-state concentration in vivo. After the MLX-MS were injected with the drug solution into the joint cavity, the drug leaked through the joint cavity and was absorbed into the blood, so compared with the solution group, the blood drug concentration was lower.Its main pharmacokinetic parameters were shown in Table 6. After oral administration of MLX, the plasma concentration Figure 6. Changes of periarticular length difference (A) and inflammatory factors (B) in rats before and after modeling (**p < 0.01, compared with control).Figure 7. Changes of inflammatory factor il-6 (A), and TNF-α (B) in rats (n = 4).Figure 8. Mean plasma concentration-time profile of MlX in rats (A) oral administration, and (B) injection of microsphere (n = 4).DRUG DELIVERY3325reached its peak at 4 hours, and the Cmax was 2.60 ± 0.39 μg/mL. The time for the peak plasma concentration of the microspheres was 8 hours, and the Cmax was significantly decreased to 0.78 ± 0.76 μg/mL. Compared with oral administration, the T1/2 microspheres were prolonged, which was about 1.1 times that of oral administration; AUC0-∞ after oral administration was 5.14 times higher than microspheres administration. It can be seen that the blood drug concentration of microsphere administration is much lower than that of oral administration, thereby reducing the systemic side effects caused by high blood drug concentra-tion. The administration of microsphere injection can effectively achieve the effect of sustained drug release.3.4.2. IVIVCThe in vitro and in vivo correlation is always a crucial issue for the dosage forms in which the drug needs to be absorbed into the circulation, since it is important for both formulation development and quality control (Sun et al., 2008). In vitro and in vivo correlations can be divided into three categories, among which the point-to-point relation-ship is classified as an A-level correlation by the US Food and Drug Administration (Ma et al., 2020). In this study, a point-to-point relationship between in vitro drug release and in vivo drug absorption was observed for MLX-MS. There was a good linear relationship between the fraction of MLX absorbed in vivo and the percent released in vitro: y = 3.6043x − 16.8842 (R2 =0.9945). Therefore, we conclude that MLX-MS exhibited similar release profiles in vitro and in vivo, and it seems reasonable to predict in vivo drug absorption from in vitro release studies.3.4.3. PK-PD correlationComparison of MLX pharmacokinetics at corresponding time points with IL-6 levels. First, from the time-concentration-ef-fect curves we can see that when the blood concentration of MLX reaches Cmax, the inflammatory factors in the rat body decrease the fastest. Intra-articular injection allows MLX to quickly reach the site of inflammation to play a role, so although the plasma concentration decreases, the level of inflammatory factors remains decreased. Further bivariate correlation analysis was performed. The results show that a positive correlation between PD-PK and the correlation coef-ficient is 0.981 (p < .05).3.4.4. Tissue distribution and histopathologyAs shown in Figure 9, MLX is widely distributed in the heart, liver, spleen, lung, and kidney of rats after oral administration and joint injection of MLX-MS. In the heart, liver, spleen, lung, kidney, brain, the drug distribution of the MLX-MS group was lower than that of the MLX oral administration group (p < .05), while in joint tissue, the distribution of the MLX-MS group was significantly higher than MLX oral admin-istration group (p < .01). It can be seen that after intra-articular administration, the drug can be quickly concentrated in the joint and its surrounding tissues, and play efficacy faster. At the same time, the distribution of drug in duodenum decreased significantly (p < .01) after microsphere injection, which can effectively reduce the gastrointestinal side effects of MLX.The histopathological examination results of rats were shown in Figure 10. According to the histopathological scoring results shown in Table 7, the score of model group was higher than that of blank control group. The articular cavity of normal rats (A1–A4) maintained a steady state, the surface of the articular cartilage was smooth and flat, and the chondrocytes were evenly distributed and arranged neatly. Occasionally, chondrocyte hypertrophy was observed, and no obvious cell clustering was observed. In the Model group (B1–B4), there were different degrees of defects in the left articular cavity, soft tissue edema tissue in the articular cavity, focal necrosis and hemor-rhage, infiltration of inflammatory cells, formation of foam cells and multinucleated giant cells, and proliferation of synovial blood vessels. In addition, the articular cartilage of rats became thinner and distributed disorderly. A large number of hypertrophic chondrocytes were seen in the bottom of the cartilage, and some articular cartilage was hardened. The proliferation of bone marrow cells in the bone marrow cavity was active, showing obvious symp-toms of arthritis. As shown in Figure 10(C1–C4), Histological score of OA decreased significantly after MLX-MS inter-vention (p < .05), rat bone tissue cell regeneration, partial recovery of bone hyperplasia; there was no obvious tissue necrosis and hemorrhage in the joint cavity. The edema of the surrounding soft tissue disappeared. The fibrous tissue proliferated. There was no obvious inflammatory cell infiltration.Table 6. Pharmacokinetic parameters.ParameterOralMicrosphereSingle doseMultiple doseCmax (μg/ml)2.77 ± 0.212.60 ± 0.390.78 ± 0.76*Tmax (h)4 ± 0.004.00 ± 0.008.00 ± 0.00*AuC0–∞ (μg/l*h)117.28 ± 4.1095.73 ± 54.5622.81 ± 9.29*MrT0–∞ (h)52.40 ± 22.1548.12 ± 20.0056.11 ± 11.41T1/2 (h)32.64 ± 18.7230.80 ± 15.5033.28 ± 4.29AuC, area under the curve; MrT, mean residence time.*p < .05 compared with the oral administration group.Figure 9. Tissue distribution of MlX in rats (n = 4, **p < 0.01, compared with oral group).3326Z. SUN ET AL.4. ConclusionIn this study, a novel MLX-MS for local injection was devel-oped for the treatment of arthritis, and the microspheres with high drug loading and encapsulation efficiency were obtained by using PLGA blends (PLGA 5050 2A and PLGA 75525 2A). The in vitro evaluation results show that MLX-MS has good sustained-release behavior. The in vivo effect evaluation shows that the microspheres can reach the site of inflammation faster after being injected into the joint cavity, and reduce the distribution of the drug in the intestinal tract, effectively reducing the risk of intestinal side effects. Therefore, such new dosage forms were expected to prolong the treatment time, reduce the number of administrations, improve patient compliance, and have good application prospects.Disclosure statementThe author reports no conflicts of interest in this work.Ethical approval statementAll animals are strictly handled following the Care and Use of Laboratory Animals (NIH Publication No. 85-23, revised 1996) and approved by the Animal Experimentation Ethics Committee of Yantai University (No.20220302).FundingThis work was supported by the Natural Science Foundation of Shandong Province (ZR2019MB054), the Science and Technology Project of TCM in Shandong Province (No. 2020Z37, 2020M190).ORCIDZheng Sun http://orcid.org/0000-0001-9285-8618Jiali Liu http://orcid.org/0000-0001-9513-1086Yanli Li http://orcid.org/0000-0002-0863-0934Bin Yu http://orcid.org/0000-0002-0397-6610Hui Xu http://orcid.org/0000-0002-0136-4784ReferencesAllison SD. (2008). Effect of structural relaxation on the preparation and drug release behavior of poly(lactic-co-glycolic)acid microparticle drug delivery systems. J Pharm Sci 97:2022–35.Brown MB, Martin GP, Jones SA, Akomeah FK. (2006). Dermal and trans-dermal drug delivery systems: current and future prospects. Drug Deliv 13:175–87.Chen HY, Fang JY. (2000). Therapeutic patents for topical and transder-mal drug delivery systems. Expert Opin Ther Pat 10:1035–43.Chu DF, Fu XQ, Liu WH, et al. (2006). Pharmacokinetics and in vitro and in vivo correlation of huperzine A loaded poly(lactic-co-glycolic acid) microspheres in dogs. Int J Pharm 325:116–23.Conaghan PG, Hunter DJ, Cohen SB, et al. (2018). Effects of a single intra-articular injection of a microsphere formulation of triamcinolone acetonide on knee osteoarthritis pain: a double-blinded, randomized, placebo-controlled, multinational study. J Bone Joint Surg Am 100:666–77.Cross M, Smith E, Hoy D, et al. (2014). The global burden of hip and knee osteoarthritis: estimates from the global burden of disease 2010 study. Ann Rheum Dis 73:1323–30.D'Souza S, Faraj JA, Giovagnoli S, Deluca PP. (2014). IVIVC from long acting olanzapine microspheres. Int J Biomater 2014:407065.Erdemli Ö, Usanmaz A, Keskin D, Tezcaner A. (2014). Characteristics and release profiles of MPEG-PCL-MPEG microspheres containing immu-noglobulin G. Colloids Surf B Biointerfaces 117:487–96.Figure 10. Pathological sections (×100) in blank control group (A1-A4), osteoarthritis model group (B1-B4), and MlX-MS group (C1-C4).Table 7. rat histopathological score.groupHistological featureCartilage degenerationCalcified cartilage and subchondral bone damageSynovial reactionBlank control000Model444MlX-MS1*0*2*p < .05 compared with the Model group.DRUG DELIVERY3327Félix Lanao RP, Leeuwenburgh SC, Wolke JG, Jansen JA. (2011). In vitro degradation rate of apatitic calcium phosphate cement with incor-porated PLGA microspheres. Acta Biomater 7:3459–68.Gerwin N, Bendele AM, Glasson S, Carlson CS. (2010). The OARSI histo-pathology initiative - recommendations for histological assessments of osteoarthritis in the rat. Osteoarthr Cartil 18:S24–S34.Hanafy AS, El-Ganainy SO. (2020). Thermoresponsive Hyalomer intra-articular hydrogels improve monoiodoacetate-induced osteoar-thritis in rats. Int J Pharm 573:118859.Kapoor M, Martel-Pelletier J, Lajeunesse D, et al. (2011). Role of proin-flammatory cytokines in the pathophysiology of osteoarthritis. Nat Rev Rheumatol 7:33–42.Karan S, Debnath S, Kuotsu K, Chatterjee TK. (2020). In-vitro and in-vivo evaluation of polymeric microsphere formulation for colon targeted delivery of 5-fluorouracil using biocompatible natural gum katira. Int J Biol Macromol 158:922–36.Laavola M, Leppänen T, Hämäläinen M, et al. (2018). IL-6 in osteoarthri-tis: effects of pine stilbenoids. Molecules 24:109.Leach WT, Simpson DT, Val TN, et al. (2005). Uniform encapsulation of stable protein nanoparticles produced by spray freezing for the re-duction of burst release. J Pharm Sci 94:56–69.Lee YS, Lowe JP, Gilby E, et al. (2010). The initial release of cisplatin from poly(lactideco-glycolide) microspheres. Int J Pharm 383:244–54.Li J, Liu N, Huang Z, et al. (2021). Intra-articular injection of loaded sPL sustained-release microspheres inhibits osteoarthritis and promotes cartilaginous repairs. J Orthop Surg Res 16:646.Ma P, Chen J, Qu H, et al. (2020). Hypoxic targeting and activating TH-302 loaded transcatheter arterial embolization microsphere. Drug Deliv 27:1412–24.Makadia HK, Siegel SJ. (2011). Poly lactic-co-glycolic acid (PLGA) as bio-degradable controlled drug delivery carrier. Polymers (Basel) 3:1377–97.Mao S, Xu J, Cai C, et al. (2007). Effect of WOW process parameters on morphology and burst release of FITC-dextran loaded PLGA micro-spheres. Int J Pharm 334:137–48.Panyam J, Williams D, Dash A, et al. (2004). Solid-state solubility influ-ences encapsulation and release of hydrophobic drugs from PLGA/PLA nanoparticles. J Pharm Sci 93:1804–14.Patoia L, Santucci L, Furno P, et al. (1996). A 4-week, double-blind, parallel-group study to compare the gastrointestinal effects of meloxi-cam 7.5 mg, meloxicam 15 mg, piroxicam 20 mg and placebo by means of faecal blood loss, endoscopy and symptom evaluation in healthy volunteers. Br J Rheumatol 35:61–7.Rodriguez-Merchan EC. (2018). Topical therapies for knee osteoarthritis. Postgrad Med 130:607–12.Schliecker G, Schmidt C, Fuchs S, et al. (2003). Hydrolytic degradation of poly(lactide-co-glycolide) films: effect of oligomers on degradation rate and crystallinity. Int J Pharm 266:39–49.Siegel SJ, Kahn JB, Metzger K, et al. (2006). Effect of drug type on the degradation rate of PLGA matrices. Eur J Pharm Biopharm 64:287–93.Singh Malik D, Mital N, Kaur G. (2016). Topical drug delivery systems: a patent review. Expert Opin Ther Pat 26:213–28.Siepmann J, Elkharraz K, Siepmann F, Klose D. (2005). How autocatalysis accelerates drug release from PLGA-based microparticles: a quantita-tive treatment. Biomacromolecules 6:2312–9.Su Y, Zhang B, Sun R, et al. (2021). PLGA-based biodegradable micro-spheres in drug delivery: recent advances in research and application. Drug Deliv 28:1397–418.Sun Y, Wang J, Zhang X, et al. (2008). Synchronic release of two hor-monal contraceptives for about one month from the PLGA micro-spheres: in vitro and in vivo studies. J Control Release 129:192–9.Tang CH. (2019). Research of pathogenesis and novel therapeutics in arthritis. IJMS 20:1646. | RESEARCH GATE |
Citation: Lamb, L.M.; Gann, D.;Velazquez, J.T.; Troxler, T.G.Detecting Vegetation to Open WaterTransitions in a Subtropical WetlandLandscape from HistoricalPanchromatic Aerial Photographyand Multispectral Satellite Imagery.Remote Sens. 2022, 14, 3976. https://doi.org/10.3390/rs14163976Academic Editor: Mingming JiaReceived: 20 June 2022Accepted: 12 August 2022Published: 16 August 2022Publisher’s Note: MDPI stays neutralwith regard to jurisdictional claims inpublished maps and institutional affil-iations.Copyright:© 2022 by the authors.Licensee MDPI, Basel, Switzerland.This article is an open access articledistributedunderthetermsandconditions of the Creative CommonsAttribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).remote sensing ArticleDetecting Vegetation to Open Water Transitions in aSubtropical Wetland Landscape from Historical PanchromaticAerial Photography and Multispectral Satellite ImageryLukas M. Lamb 1,*, Daniel Gann 1, Jesse T. Velazquez 1and Tiffany G. Troxler 21Department of Biological Sciences, Florida International University, Miami, FL 33199, USA2Department of Earth and Environment, Florida International University, Miami, FL 33199, USA*Correspondence: llamb009@fiu.edu; Tel.: +1-305-348-1576Abstract: Over the last century, direct human modification has been a major driver of coastal wetlanddegradation, resulting in widespread losses of wetland vegetation and a transition to open water.High-resolution satellite imagery is widely available for monitoring changes in present-day wetlands;however, understanding the rates of wetland vegetation loss over the last century depends onthe use of historical panchromatic aerial photographs. In this study, we compared manual imagethresholding and an automated machine learning (ML) method in detecting wetland vegetationand open water from historical panchromatic photographs in the Florida Everglades, a subtropicalwetland landscape. We compared the same classes delineated in the historical photographs to 2012multispectral satellite imagery and assessed the accuracy of detecting vegetation loss over a 72 yeartimescale (1940 to 2012) for a range of minimum mapping units (MMUs). Overall, classificationaccuracies were >95% across the historical photographs and satellite imagery, regardless of theclassification method and MMUs. We detected a 2.3–2.7 ha increase in open water pixels across allchange maps (overall accuracies > 95%). Our analysis demonstrated that ML classification methodscan be used to delineate wetland vegetation from open water in low-quality, panchromatic aerialphotographs and that a combination of images with different resolutions is compatible with changedetection. The study also highlights how evaluating a range of MMUs can identify the effect of scaleon detection accuracy and change class estimates as well as in determining the most relevant scale ofanalysis for the process of interest.Keywords: change detection; historical photograph; supervised classification; land cover; coastalwetlands; Florida Coastal Everglades1. IntroductionCoastal wetlands around the world are an economically and ecologically importantecosystem type [1–3]. While only occupying ~15% of global natural wetland area, theecosystem services provided by coastal wetlands are estimated to have a global monetaryvalue of 20.4 trillion USD/per year [4]. However, coastal wetlands have experienceddisproportionately large amounts of degradation, primarily as a result of human-drivenland-use change [5,6]. As human activities continue to drive land-use change in coastalwetlands, a key concern is how coastal wetlands will cope with the coupled effect of humanactivities and 21st century climate change [6].Climate-change-related drivers of land-cover change in coastal wetlands can broadlybe attributed to changes in temperature, precipitation [7–9], and sea-level rise (SLR), withSLR including salinization and increases in inundation [6,10,11]. The impacts of SLR alonepresent major challenges for coastal wetlands and have the potential to drive significantchanges in coastal wetlands due to the disruption of ecogeomorphic feedback loops [11–13].When ecogeomorphic feedback loops are disrupted because of hydrological shifts imposedby SLR, plant growth may slow down and decrease the stability of the wetland system,Remote Sens. 2022, 14, 3976. https://doi.org/10.3390/rs14163976https://www.mdpi.com/journal/remotesensingRemote Sens. 2022, 14, 39762 of 17leading to a rapid collapse of plant communities [14–17]. Collapse of plant communitieshas been well studied in subtropical wetland landscapes, such as the Florida Everglades,and is a key concern of restoration efforts [16,18–20].Quantifying the conversion of vegetated wetlands to open water is crucial for un-derstanding and estimating the rate and extent of vegetation loss across coastal wet-lands [21–23]. Remote sensing offers low-cost methods for monitoring landscape changesacross broad spatial and temporal scales, relative to resource-intensive, field-based meth-ods [22,23]. Advances in satellite technology have increased the spatial, temporal, spectral,and radiometric resolution of satellite imagery, which allows for more detailed detectionof land-cover changes at very fine scales [23]. This aspect is particularly important forcapturing the process of vegetation loss, as the transition of vegetated wetlands to openwater occurs at a fine scale and requires a high resolution for early detection.The advantages of using remote sensing methods to classify and monitor coastalwetlands has long been recognized, with numerous reviews detailing advances in thedifferent aspects of remote sensing applications [21–28]. Initially, reviews focused on thetypes of sensors available for remote sensing applications in wetlands and their utilityin estimating biophysical parameters, such as vegetative biomass, with some discussionof specific classification methods [21,22,24]. By now, a wide range of classification meth-ods exist, with numerous machine learning algorithms having been used for supervisedclassification of wetland vegetation [28]. Some of the more popular algorithms used forsupervised classifications include K-nearest neighbor, maximum likelihood, support vectormachines, and random forest [28]. While no single algorithm has been identified as themost optimal for all remote sensing applications, the random forest algorithm has beenshown to have a high degree of prediction accuracy when applied to spectral data [29]. Therandom forest algorithm is a classification tree method that creates an ensemble of treesand combines predictions from each tree to arrive at a single decision tree [30], and it hasbeen used in multiple wetland monitoring studies [28,31–34].When very high-resolution, modern-day satellite imagery is paired with historicalaerial photography, robust analyses of landscape change across long temporal extents canbe obtained [22]. Aerial photographs are a valuable tool for assessing land-cover changeover time as they pre-date the satellite era (starting in the 1970s) by decades, enablinginsight into historic changes in land cover [35]. However, there are numerous challengesinvolved in the classification of historical photographs, as they are often low-resolution,panchromatic data with limited spectral information and have degraded image quality dueto the presence of artifacts in the digitization process that can increase spectral noise [35].Additionally, historical photographs often lack field reference data from the time of photoacquisition, which limits their use to the differentiation of coarse land-cover classes thatcan clearly be distinguished. Nonetheless, photographic interpreters can utilize the pattern,texture, shape, size, and color or tone of features, along with experience and knowledge ofthe region of interest to accurately classify land-cover types [35].Given the well-documented degradation of wetland ecosystems during the 20thcentury, the classification of historical photographs and their use in detecting land-coverchanges present an opportunity to quantify losses in wetland vegetation across broad spatialand temporal extents. Early uses of historical aerial photographs for change detection inwetlands focused on assessing increases in open water areal extents [36–38]. These studiesused differing methods to document transitions from vegetated wetland to open watersuch as histogram thresholding [22] and manual delineation of wetland boundaries usingstereoscopes and planimeters [37,38]. While rudimentary, these early studies recognized thedistinct spectral differences between open water and the surrounding vegetation, allowingthese two land-cover types to be delineated using image processing techniques [36].More recently, studies have used a mix of manual delineation, image segmentation,and hierarchical classification to classify historical wetland landscapes [15,39–41]. Thesestudies report highly accurate historical wetland classifications (i.e., >90% overall accura-cies), yet they use different methods for quantifying the accuracy and provide no discussionRemote Sens. 2022, 14, 39763 of 17on why they selected their minimum mapping unit (MMU) or on how the selected MMUaffected the detection accuracy. The MMU is defined as “the smallest size area entity to bemapped as a discrete area” [42]. The selection of the MMU affects not only the detectionaccuracy but also the degree of information loss, as the MMU increases because smallerobjects are removed [43,44]. Selecting the MMU size depends on the original image resolu-tion, the characteristics of the landscape (i.e., composition and configuration), and the sizeof the smallest, discrete feature of interest to be retained [43–45].Proper selection of an MMU can aid in the accurate detection of open water embeddedwithin a vegetated wetland matrix from panchromatic photography by reducing falsepositives. False positives can arise due to the variability in brightness values of open waterpixels in panchromatic photographs. This variability can be due to (1) mixed pixels, wherea pixel may contain a proportion of vegetation and open water, which obscures the openwater signal; (2) solar glint reflecting off the water’s surface and increasing the brightnessvalues in the photographs; (3) bright spectral noise introduced during the process ofscanning the photograph to convert analog data to digital data; (4) misinterpreting darkshadows as water.Our objectives were three-fold: (1) detect and quantify the transition of vegetated wet-land to open water in a coastal wetland landscape over a 72-year period using digital imageprocessing techniques; (2) estimate the accuracy differences for the detection methods whenclassifying historical panchromatic photographs; (3) determine how the MMU affects theaccuracy of open water pond detection in a coastal wetland landscape. To accomplish this,we detected the extent of open water ponds embedded in a coastal graminoid-dominatedmarsh from very high-resolution 2012 multispectral satellite data. We then assessed howaccurately we could detect open water ponds in a historical, panchromatic photographfrom 1940, by comparing two classification methods: (1) manual image thresholding and(2) automated classification using the random forest algorithm. We varied the MMU ofthe 1940 and the 2012 wetland vegetation maps to identify how it affected the accuracyand precision of detecting vegetation transitioning to open water ponds. We evaluated theextent of vegetation to open water transitions across a 72-year period and then quantifiedthe classification accuracy of the 1940 and 2012 wetland classifications and change maps,using a spatially explicit, design-based accuracy assessment (Figure 1). Our goal was torefine methods for combining aerial photographs and multispectral satellite images toassess coastal wetland degradation across multidecadal timescales.More recently, studies have used a mix of manual delineation, image segmentation, and hierarchical classification to classify historical wetland landscapes [15,39–41]. These studies report highly accurate historical wetland classifications (i.e., >90% overall accura-cies), yet they use different methods for quantifying the accuracy and provide no discus-sion on why they selected their minimum mapping unit (MMU) or on how the selected MMU affected the detection accuracy. The MMU is defined as “the smallest size area en-tity to be mapped as a discrete area” [42]. The selection of the MMU affects not only the detection accuracy but also the degree of information loss, as the MMU increases because smaller objects are removed [43,44]. Selecting the MMU size depends on the original im-age resolution, the characteristics of the landscape (i.e., composition and configuration), and the size of the smallest, discrete feature of interest to be retained [43–45]. Proper selection of an MMU can aid in the accurate detection of open water embed-ded within a vegetated wetland matrix from panchromatic photography by reducing false positives. False positives can arise due to the variability in brightness values of open water pixels in panchromatic photographs. This variability can be due to (1) mixed pixels, where a pixel may contain a proportion of vegetation and open water, which obscures the open water signal; (2) solar glint reflecting off the water’s surface and increasing the brightness values in the photographs; (3) bright spectral noise introduced during the process of scan-ning the photograph to convert analog data to digital data; (4) misinterpreting dark shad-ows as water. Our objectives were three-fold: (1) detect and quantify the transition of vegetated wetland to open water in a coastal wetland landscape over a 72-year period using digital image processing techniques; (2) estimate the accuracy differences for the detection meth-ods when classifying historical panchromatic photographs; (3) determine how the MMU affects the accuracy of open water pond detection in a coastal wetland landscape. To ac-complish this, we detected the extent of open water ponds embedded in a coastal grami-noid-dominated marsh from very high-resolution 2012 multispectral satellite data. We then assessed how accurately we could detect open water ponds in a historical, panchro-matic photograph from 1940, by comparing two classification methods: (1) manual image thresholding and (2) automated classification using the random forest algorithm. We var-ied the MMU of the 1940 and the 2012 wetland vegetation maps to identify how it affected the accuracy and precision of detecting vegetation transitioning to open water ponds. We evaluated the extent of vegetation to open water transitions across a 72-year period and then quantified the classification accuracy of the 1940 and 2012 wetland classifications and change maps, using a spatially explicit, design-based accuracy assessment (Figure 1). Our goal was to refine methods for combining aerial photographs and multispectral satellite images to assess coastal wetland degradation across multidecadal timescales. Figure 1. Generalized overview of the wetland classification and change detection methods.Remote Sens. 2022, 14, 39764 of 172. Methods2.1. Study AreaThe Florida Everglades, a World Heritage Site, International Biosphere Reserve, andWetland of International Importance, comprise hundreds of thousands of hectares of marsh,mangrove, and estuarine open water habitat [46]. The southernmost region is home to theEverglades National Park (ENP), where paleoecological studies indicate that some locationsare as old as 5700 cal years BP [47] and have seen considerable coastal transgression andregression, indicating the dynamic nature of this coastal environment [48–52]. Since the late19th century, anthropogenic activities have driven a decrease in hydrologic connectivitybetween the northern Everglades and the coastal regions [53–55], altering the abioticconditions necessary for peatland maintenance and growth, leading to a 50% reduction inpeat horizontal and vertical extents [55–57].For this analysis, we focused on a 1.1 × 0.66 km area of coastal marsh along thesouthwest coast of the ENP for detecting transitions from coastal graminoid marsh to openwater pond (Figure 2). The study area was located ~10 km upstream of the Gulf of Mexico,on the south side of the Harney River, a major estuarine channel in ENP (Figure 2). Thestudy area is dominated by two emergent, graminoid marsh species: Cladium jamaicenseand Juncus romerianus (Figure 3A). Vegetative cover of these graminoids can range fromhigh density (100% cover; Figure 3A) to low density (<25% cover; Figure 3B). Interspersedwithin the graminoid marsh matrix are permanently flooded, unvegetated open waterponds of varying size (Figure 3C) and dense clusters of trees, known as “tree islands”,dominated by mangrove and subtropical broad-leaved species (Figure 3A, background).Remote Sens. 2022, 14, x FOR PEER REVIEW 4 of 18 Figure 1. Generalized overview of the wetland classification and change detection methods. 2. Methods 2.1. Study Area The Florida Everglades, a World Heritage Site, International Biosphere Reserve, and Wetland of International Importance, comprise hundreds of thousands of hectares of marsh, mangrove, and estuarine open water habitat [46]. The southernmost region is home to the Everglades National Park (ENP), where paleoecological studies indicate that some locations are as old as 5700 cal years BP [47] and have seen considerable coastal transgression and regression, indicating the dynamic nature of this coastal environment [48–52]. Since the late 19th century, anthropogenic activities have driven a decrease in hydrologic connectivity between the northern Everglades and the coastal regions [53–55], altering the abiotic conditions necessary for peatland maintenance and growth, leading to a 50% reduction in peat horizontal and vertical extents [55–57]. For this analysis, we focused on a 1.1 × 0.66 km area of coastal marsh along the south-west coast of the ENP for detecting transitions from coastal graminoid marsh to open wa-ter pond (Figure 2). The study area was located ~10 km upstream of the Gulf of Mexico, on the south side of the Harney River, a major estuarine channel in ENP (Figure 2). The study area is dominated by two emergent, graminoid marsh species: Cladium jamaicense and Juncus romerianus (Figure 3A). Vegetative cover of these graminoids can range from high density (100% cover; Figure 3A) to low density (<25% cover; Figure 3B). Interspersed within the graminoid marsh matrix are permanently flooded, unvegetated open water ponds of varying size (Figure 3C) and dense clusters of trees, known as “tree islands”, dominated by mangrove and subtropical broad-leaved species (Figure 3A, background). Figure 2. Overview of study area indicating the location of (A) the study area (black star) within the state of Florida (grey) in the United States; (B) the study area (white rectangle) and its position rel-ative to the southwest coast of Florida with the distance (~10 km) to the Gulf of Mexico indicated; Figure 2. Overview of study area indicating the location of (A) the study area (black star) withinthe state of Florida (grey) in the United States; (B) the study area (white rectangle) and its positionrelative to the southwest coast of Florida with the distance (~10 km) to the Gulf of Mexico indicated;(C) the spatial extent (white box) of the study area. The aerial photography used an ArcGIS Pro (v.5.6)imagery base map obtained on 29 January 2021.Remote Sens. 2022, 14, 39765 of 17Remote Sens. 2022, 14, x FOR PEER REVIEW 5 of 18 (C) the spatial extent (white box) of the study area. The aerial photography used an ArcGIS Pro (v.5.6) imagery base map obtained on 29 January 2021. Figure 3. Photos from a graminoid dominated marsh in the coastal Everglades showing the major wetland types mapped in this study; (A) high-density, graminoid marsh dominated by Cladium ja-maicense with a tree island dominated by mangrove spp. in the background; (B) low-density grami-noid marsh interspersed with open water; (C) an unvegetated open water pond bordered by Cla-dium jamaicense and the scrub Rhizophora mangle. 2.2. Image Selection, Specifications, and Preprocessing Historical wetland vegetation was mapped from a digitized March 1940 panchro-matic aerial photograph. The panchromatic photograph was publicly available and stored in a digital image repository maintained by the United States Geological Survey for the Greater Everglades Ecosystem [58]. The digital copy of the photograph was an 8-bit gray-scale panchromatic image, georeferenced to UTM NAD83 zone 17 with a 1 m grid spacing (Foster 2004). Using the crop function in the R package raster [59], we clipped the photo-graph to a study area with a maximum east–west extent of 0.66 km and a maximum north–south extent of 1.1 km (Figure 2C). To align the grid spacing of the historical aerial photo-graph with the high-resolution satellite image, we used bilinear resampling to resample the aerial photograph to the same 2 m grid as the satellite image. Contemporary vegetation cover was mapped from one, cloud-free, 8-band WorldView-2 (WV-2) satellite image (Digital Globe, Inc., Westminster, CO, USA) obtained on 3 November 2012, which covered the area of the historical photograph. The 2012 WV-2 image was a 16-bit image. The image was geometrically corrected and projected to UTM NAD 83 zone 17 with a 2 m grid spacing using the Rigorous Orthorectification tool in ENVI (v.5.6). Radiometric and atmospheric corrections were also accomplished in ENVI using the radiometric calibration tool to convert digital numbers to radiance. The radiance image was then atmospherically corrected using the Fast Line-of-Sight Atmospheric Anal-ysis of Hypercubes (FLAASH) atmospheric correction module with a mid-latitude sum-mer atmospheric model and maritime aerosol model. A scale factor of 1.0 was set for the bands. Visibility was set at 100 km. The output scale was 10,000 to represent percent re-flectance values with 2 decimal precision as integers. Following corrections, the image was clipped to the extent of the study area (Figure 2C) using the crop function in the R package raster [59]. 2.3. Wetland Classification 2.3.1. Classification Scheme Vegetated wetland generally reflects greater magnitudes of electromagnetic radia-tion relative to open water bodies, which absorb the majority of incoming solar radiation [36]. This strong spectral separation was observed in both the panchromatic photograph (Figure 4A,C) and the satellite image (Figure 4B,D). Since the focus of our study was assessing the transition of vegetated wetland to open water, and open water to vegetated wetland, we mapped 1940 and 2012 wetland cover using a binary classification scheme of (1) vegetated wetland and (2) open water. For the 2012 classification, when the reference aerial photography indicated a mixed pixel (see Figure 3. Photos from a graminoid dominated marsh in the coastal Everglades showing the majorwetland types mapped in this study; (A) high-density, graminoid marsh dominated by Cladiumjamaicense with a tree island dominated by mangrove spp. in the background; (B) low-densitygraminoid marsh interspersed with open water; (C) an unvegetated open water pond bordered byCladium jamaicense and the scrub Rhizophora mangle.2.2. Image Selection, Specifications, and PreprocessingHistorical wetland vegetation was mapped from a digitized March 1940 panchromaticaerial photograph. The panchromatic photograph was publicly available and stored in adigital image repository maintained by the United States Geological Survey for the GreaterEverglades Ecosystem [58]. The digital copy of the photograph was an 8-bit grayscalepanchromatic image, georeferenced to UTM NAD83 zone 17 with a 1 m grid spacing(Foster 2004). Using the crop function in the R package raster [59], we clipped the photographto a study area with a maximum east–west extent of 0.66 km and a maximum north–southextent of 1.1 km (Figure 2C). To align the grid spacing of the historical aerial photographwith the high-resolution satellite image, we used bilinear resampling to resample the aerialphotograph to the same 2 m grid as the satellite image.Contemporary vegetation cover was mapped from one, cloud-free, 8-band WorldView-2(WV-2) satellite image (Digital Globe, Inc., Westminster, CO, USA) obtained on 3 November2012, which covered the area of the historical photograph. The 2012 WV-2 image wasa 16-bit image. The image was geometrically corrected and projected to UTM NAD 83zone 17 with a 2 m grid spacing using the Rigorous Orthorectification tool in ENVI (v.5.6).Radiometric and atmospheric corrections were also accomplished in ENVI using the ra-diometric calibration tool to convert digital numbers to radiance. The radiance imagewas then atmospherically corrected using the Fast Line-of-Sight Atmospheric Analysisof Hypercubes (FLAASH) atmospheric correction module with a mid-latitude summeratmospheric model and maritime aerosol model. A scale factor of 1.0 was set for the bands.Visibility was set at 100 km. The output scale was 10,000 to represent percent reflectancevalues with 2 decimal precision as integers. Following corrections, the image was clipped tothe extent of the study area (Figure 2C) using the crop function in the R package raster [59].2.3. Wetland Classification2.3.1. Classification SchemeVegetated wetland generally reflects greater magnitudes of electromagnetic radia-tion relative to open water bodies, which absorb the majority of incoming solar radia-tion [36]. This strong spectral separation was observed in both the panchromatic photo-graph (Figure 4A,C) and the satellite image (Figure 4B,D).Since the focus of our study was assessing the transition of vegetated wetland to openwater, and open water to vegetated wetland, we mapped 1940 and 2012 wetland coverusing a binary classification scheme of (1) vegetated wetland and (2) open water. For the2012 classification, when the reference aerial photography indicated a mixed pixel (seeFigure 3B, for example), we assigned a class label of vegetated wetland when the vegetatedcover was >50% within the pixel.Remote Sens. 2022, 14, 39766 of 17Remote Sens. 2022, 14, x FOR PEER REVIEW 6 of 18 Figure 3B, for example), we assigned a class label of vegetated wetland when the vege-tated cover was >50% within the pixel. Figure 4. Arbitrary transect across (A) the 1940 panchromatic photograph and (B) 2012 WorldView-2 image showing the brightness values of the (C) panchromatic photograph and the (D) WorldView-2 imagery bands. 2.3.2. 1940 Manual Threshold Classification For the manual threshold classification, visual interpretation of spectral values in the historical photograph determined the threshold value that separated open water from vegetated wetland (Figure 4A,C). Water had very low brightness values relative to the surrounding vegetated wetland, with panchromatic spectral brightness values ranging from 17 to 30 and wetland vegetation > 31 (Figure 4C). Pixels above this brightness value were classified as “vegetated” and below this brightness value as “water”. 2.3.3. 1940 and 2012 Automated Machine Learning Classification For the automated ML classification of the 1940 panchromatic photograph and the 2012 satellite image, we used the random forest algorithm with 5-fold cross-validation using the R package caret [60]. For both the 1940 and 2012 random forest classification, we developed training da-tasets to cover the range of brightness values for each class using on-screen point digiti-zation with samples labeled in ArcGIS Pro (v2.8.0). Characteristics, such as brightness, color, texture, pattern, and context, were used to identify and label samples as either veg-etation or open water. For classifying open water in the 1940 photograph, we focused on placing training points within identifiable, discrete open water ponds. The training datasets totaled 5362 points for the 1940 panchromatic photograph and 5606 points for the 2012 satellite image. The 1940 training dataset had 3082 samples for vegetated wetland and 2280 open water samples, while the 2012 training dataset consisted of 3582 vegetated wetland samples and 2024 open water samples. Figure 4. Arbitrary transect across (A) the 1940 panchromatic photograph and (B) 2012 WorldView-2image showing the brightness values of the (C) panchromatic photograph and the (D) WorldView-2imagery bands.2.3.2. 1940 Manual Threshold ClassificationFor the manual threshold classification, visual interpretation of spectral values in thehistorical photograph determined the threshold value that separated open water fromvegetated wetland (Figure 4A,C). Water had very low brightness values relative to thesurrounding vegetated wetland, with panchromatic spectral brightness values rangingfrom 17 to 30 and wetland vegetation > 31 (Figure 4C). Pixels above this brightness valuewere classified as “vegetated” and below this brightness value as “water”.2.3.3. 1940 and 2012 Automated Machine Learning ClassificationFor the automated ML classification of the 1940 panchromatic photograph and the2012 satellite image, we used the random forest algorithm with 5-fold cross-validationusing the R package caret [60].For both the 1940 and 2012 random forest classification, we developed training datasetsto cover the range of brightness values for each class using on-screen point digitization withsamples labeled in ArcGIS Pro (v2.8.0). Characteristics, such as brightness, color, texture,pattern, and context, were used to identify and label samples as either vegetation or openwater. For classifying open water in the 1940 photograph, we focused on placing trainingpoints within identifiable, discrete open water ponds.The training datasets totaled 5362 points for the 1940 panchromatic photograph and5606 points for the 2012 satellite image. The 1940 training dataset had 3082 samples forvegetated wetland and 2280 open water samples, while the 2012 training dataset consistedof 3582 vegetated wetland samples and 2024 open water samples.For the 1940 training points, brightness values were extracted from the panchromaticphotograph. For the 2012 training points, spectral reflectance values of the eight spectralbands and seven vegetation indices were extracted from the satellite image (Table 1). Arandom forest algorithm was then trained using each dataset. Training datasets weredeveloped using an iterative process where we began with an initial sample of trainingRemote Sens. 2022, 14, 39767 of 17points, then added points to locations that were incorrectly classified by the random forestalgorithm until the addition of new points did not improve the model’s accuracy.Table 1.WorldView-2 bands and vegetation indices used in the 2012 wetland classification.NIR = near-infrared; RE = red edge.Band/Vegetation IndexEquationReferenceRed, Blue, Green, Coastal, Red,NIR1, NRI2, RENoneN/ASimple Ratio (SR)Red/NIR1[61]Normalized DifferenceVegetation Index (NDVI)(NIR1 − Red)/(NIR1 + Red)[62]Normalized Difference Red-EdgeIndex (NDRE)(NIR1 − RE)/(NIR1 + RE)[63]Normalized Difference WaterIndex (NDWI)(Green-NIR2)/(Green + NIR2)[64]Green Normalized DifferenceVegetation Index (GNDVI)(NIR1-Green)/(NIR1 + Green)[65]Soil Adjusted VegetationIndex (SAVI)1.5 × ((NIR1 − Red)/(NIR1 + Red + 0.5))[66]Enhanced Vegetation Index (EVI)2.5 × ((NIR1 − Red)/(NIR1 + 2.4 × Red + 1))[67]2.4. Morphological Filtering and Minimum Mapping UnitFollowing image classification, we applied a morphological filter to remove vegetationand open water patches smaller than three specified MMUs: 12 m2 (3 pixels), 24 m2(6 pixels), and 36 m2 (9 pixels). Mathematical morphological frameworks have been usedby image analysts for decades as a tool to reduce error rates through an iterative processof dilation (opening) and erosion (closing) operations to remove noisy pixels based on aspecified window [68–70]. The morphological filter iteratively filled in patches smaller thanthe MMU from the edges of the patches to the center, replacing values of those patches withthe majority rule applied to a user-defined 3 × 3 kernel size. The morphological filteringalgorithm was scripted in R [71]. For the 1940 wetland classification, this resulted in sixmaps (two classification methods times three MMUs) and for the 2012 wetland classificationin three maps (one classification method times three MMUs). We selected the range ofMMUs to use 12 m2 as the smallest MMU to filter single- and two-pixel noise that oftenoccurs due to the vegetation’s shadows. The largest MMU we evaluated was 36 m2, beyondwhich the loss of open water ponds was considered unacceptable.Following morphological filtering, we converted each raster-based map to polygonsin ArcGIS Pro (v2.8.0) using the Raster to Polygon tool. We then calculated the total countand area of open water ponds to assess how the MMU affected the total number of detectedponds and their size distributions.2.5. Vegetation Change DetectionTo quantify vegetation transitions, we overlaid the filtered wetland maps from 1940and 2012 and assigned class-change labels to each pixel. The resulting classification hadthree classes: (1) no change, (2) vegetated wetland to open water, and (3) open water tovegetated wetland. We scripted change detection using R statistical software [71] andassessed per-pixel changes in wetland cover between 1940 and 2012 for each 1940 detectionmethod (n = 2) and each MMU (n = 3), for a total of six change maps.2.6. Design-Based Accuracy Assessments of 1940 and 2012 Wetland Classifications and Change MapsWe conducted a post-classification design-based accuracy assessment with a stratifiedrandom sample design to assess the accuracy of the 1940 and 2012 maps and each changeRemote Sens. 2022, 14, 39768 of 17map. For each accuracy assessment, the spatial assessment unit was the pixel, and agree-ment was defined as the predicted map label being the same as the reference label. Weestimated the sample size needed for each map based on a multinomial distribution witha desired level of confidence of 95% and a precision of 3% [72]. For vegetation maps thenumber of samples was calculated for the vegetated wetland and open water classes. Forthe change maps, we sampled the three change classes of no change, vegetated to water,and water to vegetated.For all accuracy assessments, we sampled a set of reference points for each map(Tables S1 and S2) using the Strata function in the R package sampling [73]. Samples werevisually evaluated and labeled using ArcGIS Pro (v.2.8.0) using the corresponding aerialphotograph for each map. Reference points were used to compute an area-based error ma-trix with adjusted area, adjusted user’s accuracy, producer’s accuracy, and overall accuracy.We estimated the upper and lower bounds for each accuracy metric with 95% confidenceintervals using error matrix reference samples and map ratios [74]. The equations we usedfor all accuracy estimates can be found in [74].3. Results3.1. 1940 Wetland ClassificationAcross both the thresholding and random forest classifications of the 1940 historicalphotograph, the overall accuracy and user’s accuracy were consistently above 95%, rangingfrom 96.2 ± 4.8% to 99.8 ± 0.1%, with the MMU only showing a marginal effect on accuracy(Table 2; Figure 5A). For the open water class, the only statistically lower user’s accuracywas observed for the random forest classification at an MMU of 24 m2 (Table 2). Theproducer’s accuracy was consistently above 99% for both classes, except for the thresholdclassification at an MMU of 24 m2, which showed a decrease of 50.1 ± 49.2%. The highconfidence interval for that accuracy, however, suggests that the estimated accuracy ishighly uncertain. The random forest classifier predicted ~50% more open water than thethreshold classifier, ranging from 1.9 to 2.2 ha for the random forest classification and from1.4 to 1.5 ha for the threshold classification.Table 2. Area-adjusted accuracy estimates (percent ± 95% confidence interval) as well as the totalmapped area for wetland classifications for the 1940 panchromatic photograph and 2012 satelliteimage. Estimates are presented for each 1940 classification method, the 2012 classification, and eachminimum mapping unit (MMU). Thres = threshold classifications; rF = random forest classification.Year andMethodMMU (m2)ClassMappedArea(ha)AdjustedArea(ha)AdjustedProducer’sAccuracyAdjustedUser’sAccuracyOverallAccuracy1940 Thres12Vegetated70.470.6 ± 0.199.8 ± 0.1100 ± 0.099.8 ± 0.1Open water1.51.4 ± 0.1100 ± 0.091.5 ± 7.224Vegetated70.569.4 ± 2.599.7 ± 0.296.4 ± 4.996.2 ± 4.8Open water1.52.6 ± 2.550.1 ± 49.287.5 ± 8.736Vegetated70.670.7 ± 0.199.8 ± 0.2100 ± 0.098.8 ± 0.1Open water1.41.3 ± 0.1100 ± 0.090.6 ± 0.41940 rF12Vegetated69.870.2 ± 0.199.7 ± 0.2100 ± 0.099.7 ± 0.1Open water2.21.8 ± 0.1100 ± 0.090.4 ± 6.824Vegetated69.970.2 ± 0.299.5 ± 0.2100 ± 0.099.5 ± 0.2Open water2.01.7 ± 0.2100 ± 0.083.1 ± 8.436Vegetated70.070.2 ± 0.199.7 ± 0.2100 ± 0.099.5 ± 0.2Open water1.91.8 ± 0.1100 ± 0.090.4 ± 6.8Remote Sens. 2022, 14, 39769 of 17Table 2. Cont.Year andMethodMMU (m2)ClassMappedArea(ha)AdjustedArea(ha)AdjustedProducer’sAccuracyAdjustedUser’sAccuracyOverallAccuracy2012 rF12Vegetated68.568.1 ± 1.199.8 ± 0.199.2 ± 1.599.1 ± 1.5Open water3.53.9 ± 1.186.3 ± 23.296.8 ± 3.024Vegetated68.667.1 ± 1.899.8 ± 0.197.6 ± 2.797.6 ± 2.6Open water3.44.9 ± 1.866.5 ± 25.096.8 ± 3.136Vegetated68.768.1 ± 1.199.9 ± 0.199.1 ± 1.699.2 ± 1.5Open water3.33.8 ± 1.185.1 ± 12.699.9 ± 1.636 98.8 ± 0.1 Open water 1.4 1.3 ± 0.1 100 ± 0.0 90.6 ± 0.4 1940 rF 12 Vegetated 69.8 70.2 ± 0.1 99.7 ± 0.2 100 ± 0.0 99.7 ± 0.1 Open water 2.2 1.8 ± 0.1 100 ± 0.0 90.4 ± 6.8 24 Vegetated 69.9 70.2 ± 0.2 99.5 ± 0.2 100 ± 0.0 99.5 ± 0.2 Open water 2.0 1.7 ± 0.2 100 ± 0.0 83.1 ± 8.4 36 Vegetated 70.0 70.2 ± 0.1 99.7 ± 0.2 100 ± 0.0 99.5 ± 0.2 Open water 1.9 1.8 ± 0.1 100 ± 0.0 90.4 ± 6.8 2012 rF 12 Vegetated 68.5 68.1 ± 1.1 99.8 ± 0.1 99.2 ± 1.5 99.1 ± 1.5 Open water 3.5 3.9 ± 1.1 86.3 ± 23.2 96.8 ± 3.0 24 Vegetated 68.6 67.1 ± 1.8 99.8 ± 0.1 97.6 ± 2.7 97.6 ± 2.6 Open water 3.4 4.9 ± 1.8 66.5 ± 25.0 96.8 ± 3.1 36 Vegetated 68.7 68.1 ± 1.1 99.9 ± 0.1 99.1 ± 1.6 99.2 ± 1.5 Open water 3.3 3.8 ± 1.1 85.1 ± 12.6 99.9 ± 1.6 Figure 5. (A) 1940 random forest land cover classification; (B) 2012 random forest land cover classi-fication. Both classifications are shown at a 12 m2 MMU. Figure 5. (A) 1940 random forest land cover classification; (B) 2012 random forest land coverclassification. Both classifications are shown at a 12 m2 MMU.3.2. 2012 Wetland ClassificationFor the 2012 random forest wetland land-cover classification, we obtained an overallaccuracy and user’s accuracy above 96% across all MMU’s (Table 2; Figure 5B). The user’saccuracy for the vegetated wetland and open water classes were similar across classifiedwetland maps (Table 2). For the open water class, the producer’s accuracy was low, rangingfrom 66.5 ± 25.0% to 86.3 ± 23.2%.3.3. Changes in Wetland Classes from 1940 to 2012Change maps (Figure 6) had overall accuracies above 96%, with the change mapcomparing the 1940 threshold classification and 2012 classification at a 12 m2 MMU havingthe greatest overall accuracy of 99.0 ± 0.3% (Table 3). The highest user’s accuracy ofRemote Sens. 2022, 14, 397610 of 17the vegetated wetland to water change class was 83.5 ± 7.2% for the 1940 random forestclassification and 2012 classification change map at a 12 m2 MMU (Table 3). However, the1940 threshold and 2012 classification at a 24 m2 MMU had a significantly lower user’saccuracy of 70.1 ± 9.7%. This corresponded to a low producer’s accuracy for the no-changeclass, which ranged from 58.9% to 68.3% across all change maps (Table 3). The user’saccuracy of the open water to vegetated wetland class was above 50% for all change maps,ranging from 53.2 ± 10.1% to 72.6 ± 9.0%.The total estimated area of 1940 vegetated wetland that was converted to open waterranged from 2.3 ha (3.2%) to 2.7 ha (3.7%) across all change maps, with a conversion fromwater to vegetation ranging from 0.6 ha (0.08%) to 1.2 ha (1.7%) across all change mapversions (Table 3). The “no-change” class was consistently estimated to have an area of68.3 ha (94.8%) to 68.7 ha (95.4%) across all change map versions (Table 3).4, x FOR PEER REVIEW 11 of 18 Figure 6. Maps of (A) wetland change between the 1940 manual threshold and the 2012 random forest classifications; (B) wetland change between the 1940 random forest and 2012 random forest classifications. Both maps are shown at a 12 m2 MMU. There are three wetland change classes: no change (semitransparent grey), vegetated wetland to open water (blue), and open water to vege-tated wetland (yellow). 3.4. Open Water Pond Detections: Total Count and Size For the 1940 wetland classification and across the MMUs, the median open water body size was similar across classification methods, but the random forest classification consistently identified more open water bodies than the threshold classification across each MMU (Table 4). For the 2012 wetland classification, the median open water body size increased with the MMU, from 55 m2 to 97 m2, while the total number of open water bod-ies decreased from 307 to 197 (Table 4). Figure 6. Maps of (A) wetland change between the 1940 manual threshold and the 2012 randomforest classifications; (B) wetland change between the 1940 random forest and 2012 random forestclassifications. Both maps are shown at a 12 m2 MMU. There are three wetland change classes: nochange (semitransparent grey), vegetated wetland to open water (blue), and open water to vegetatedwetland (yellow).Remote Sens. 2022, 14, 397611 of 17Table 3. Area-adjusted accuracy estimates for change maps (percent ± 95% confidence interval)including total mapped area. Thres = threshold classifications; rF = random forest classification.ChangeMapMMU(m2)ClassMappedArea (ha)AdjustedArea (ha)AdjustedProducer’sAccuracyAdjustedUser’sAccuracyOverallAccuracy1940 Thres,2012 rF12No Change68.669.3 ± 0.298.9 ± 0.3100 ± 0.099.0 ± 0.3Vegetated to Water2.72.2 ± 0.2100 ± 0.081.1 ± 7.9Water to Vegetated0.70.5 ± 0.1100 ± 0.072.6 ± 9.024No Change68.769.5 ± 0.298.9 ± 0.3100 ± 0.098.9 ± 0.3Vegetated to Water2.62.0 ± 0.2100 ± 0.080.2 ± 8.2Water to Vegetated0.70.4 ± 0.1100 ± 0.063.7 ± 9.936No Change68.969.0 ± 1.698.6 ± 0.398.8 ± 2.297.6 ± 2.2Vegetated to Water2.52.5 ± 1.668.7 ± 42.370.1 ± 9.7Water to Vegetated0.60.4 ± 0.1100 ± 0.068.9 ± 9.81940 rF,2012 rF12No Change68.367.9 ± 1.898.6 ± 0.398.1 ± 2.696.9 ± 2.5Vegetated to Water2.53.4 ± 1.861.1 ± 32.983.5 ± 7.2Water to Vegetated1.20.7 ± 0.1100 ± 0.057.3 ± 9.624No Change68.568.7 ± 1.498.6 ± 0.398.9 ± 1.998.7 ± 0.3Vegetated to Water2.42.0 ± 0.2100 ± 0.081.6 ± 7.7Water to Vegetated1.11.3 ± 1.4100 ± 0.056.1 ± 9.836No Change68.668.9 ± 1.498.6 ± 0.398.9 ± 2.197.6 ± 2.0Vegetated to Water2.42.5 ± 1.471.2 ± 40.277.6 ± 8.4Water to Vegetated1.00.5 ± 0.9100 ± 0.053.2 ± 10.13.4. Open Water Pond Detections: Total Count and SizeFor the 1940 wetland classification and across the MMUs, the median open waterbody size was similar across classification methods, but the random forest classificationconsistently identified more open water bodies than the threshold classification acrosseach MMU (Table 4). For the 2012 wetland classification, the median open water body sizeincreased with the MMU, from 55 m2 to 97 m2, while the total number of open water bodiesdecreased from 307 to 197 (Table 4).Table 4.Median open water body detection number and median size in square meters (m2)for 1940 and 2012 wetland maps across classification methods and MMUs of 12, 24, and 36 m2.Thres = threshold classifications; rF = random forest classification.Year,Classification-MMUTotal Open Water BodyDetectionsMedian Open WaterArea (m2)1940, Thres-1217243.01940, Thres-2413258.01940, Thres-3610380.01940, rF-1225132.01940, rF-2416170.01940, rF-3612587.02012, rF-1230755.02012, rF-2423284.52012, rF-3619797.0Remote Sens. 2022, 14, 397612 of 174. DiscussionThe results of this study demonstrate that manual thresholding and automated clas-sification methods can achieve similar classification accuracy when classifying wetlandlandscapes with panchromatic photographs (Table 2). High overall accuracies (>95%)were consistent for all wetland vegetation maps across classification methods and MMUs(Table 2). We obtained high user’s and producer’s accuracies for vegetated wetland andopen water classes; however, in two instances, the producer’s accuracy of the open waterdeclined significantly, decreasing to as low as 50.1 ± 49.2%. Further, while change mapoverall accuracy was consistently above 96%, we obtained multiple change maps with a lowuser’s and/or producer’s accuracy, in the range of 50 to 70%. Low class-specific accuraciescan be attributed to (1) large differences in proportional class area (Tables S1 and S2) and(2) misclassifications of tree islands in the panchromatic photograph.4.1. The Importance of Spatially Explicit Accuracy AssessmentsAccuracy assessments are critical for quantifying the error and uncertainty of land-cover classifications. The importance of accuracy assessments increases when spatiallyexplicit, detailed, class-specific changes are of interest, rather than just the spatially implicitcomparisons of land-cover classes between time points. For point-based accuracy assess-ments, the number of reference samples per class is determined by the desired confidenceand precision level, the proportion of the majority class, and total number of classes [72].While the proportional area of the majority class was similar across wetland and changemaps, the change maps contained three classes while the wetland maps contained twoclasses (Tables S1 and S2). The increase in class count led to more reference samples perclass in the change maps and the identification of misclassified water pixels, and a reduc-tion in the user’s and producer’s accuracy. Had an implicit estimate of change classesbeen relied upon, as conducted in other change detection studies [75], we would not haveidentified these misclassifications. This underscores the need for spatially explicit accuracyassessments of change classes when conducting digital change detection analyses.When evaluating change classes in a spatially explicit manner, it has also been rec-ommended, and demonstrated here, that estimates of accuracy and areal class cover needto be adjusted for misclassifications that are estimated from the reference dataset. Whenthere are few classes that have large differences in proportional area, small changes inthe reference sample error matrix can amount to large changes in estimated accuracy anduncertainty leading to 95% confidence intervals greater than ±40%, as occurred in multipleof our change maps (Table 3). This level of uncertainty makes it impossible to evaluatethe producer’s accuracy of a map. This demonstrates that even for seemingly trivial re-mote sensing problems (i.e., differentiating water from vegetation across multidecadaltimescales), the methods used to evaluate spatially explicit change need to account forlarge areal differences in mapped classes and consider adjustments using the error matrixproportions in order to build confidence in the location-specific mapped changes [74].4.2. Misclassifications and Improving AccuracyThe binary classification scheme of vegetated wetland and open water was predicatedon the large difference in spectral reflectance between graminoid marsh and open water inboth the panchromatic photograph and high-resolution satellite image (Figure 3). However,in the panchromatic photography, the brightness values of dense clusters of trees werelower and appeared darker than the graminoid vegetation and often resembled grey valuesof open water areas. The misclassification of tree islands in the 1940 photograph wasonly apparent when overlaying the photograph on modern-day aerial photographs. Thisobservation required a well-trained aerial photo-interpreter with familiarity of the region,highlighting a challenge of working with historical, panchromatic photographs and theneed for developing robust protocols for image processing.To improve the differentiation of dark pixels, such as open water or tree clusters,additional data sources should be incorporated into the image processing procedure. Tex-Remote Sens. 2022, 14, 397613 of 17tural analysis has been shown to improve the discrimination of components of vegetationstructure in infrared images [76] while also improving the classification accuracy of panchro-matic photographs [77,78]. Additionally, stereoscopic imaging can aid in the discriminationof dark pixels that contain vegetation, as stereoscopy allows for a three-dimensional con-struction of an area and the estimation of vegetation height [79].4.3. Impact of the Minimum Mapping Unit on Open Water DetectionAcross the 72-ha study area, an estimated 2.3 to 2.7 ha of vegetated wetland transi-tioned to open water across all change map versions (Table 4). This can be attributed toboth an increase in the median open water pond size and total count of open water pondsdetected (Table 3). Across MMUs, the increases in total open water ponds detected between1940 and 2012 ranged from 56 to 135 (Table 3), which corresponds to a 22% to 78% increase.The increase in median open water pond size ranged from 12 m2 to 26.5 m2. Our imageshad a spatial resolution of 4 m2, meaning an MMU of 12 m2 is a combination of 3 pixels.Based on these results, a 12 m2 MMU was large enough to filter out noisy pixels but smallenough to retain the detail needed to capture the conversion of vegetated wetland to openwater. Since conversion of vegetation to open water can occur at spatial scales <1 m2 inwetland landscapes, such as the Florida Everglades [19], retaining smaller water bodies isdesirable and suggests the MMU should be selected at the smallest size possible that stillretains high accuracy.4.4. Ecological ImplicationsOur finding of losses in vegetation and increases in open water have broader implica-tions for the coastal wetland landscape of the Florida Everglades. In a similar study, manualdelineation of historical images was used to assess changes in open water ponds between1953 and 2009 in the Ten Thousand Islands region of the ENP [16]. The results of this studyfound a similar increase in total open water ponds and pond size, ranging from 13 m2 to31 m2 across their study area, which had sites that were differentially affected by freshwatermanagement and sea-level rise [16]. Our study area was in a similar ecogeomorphic settingbut encompassed a smaller spatial extent and was in a remote setting ~74 km south of [16]and along the southwest coast of the ENP. Both sites have been subjected to hydrologicstresses, but sites within [16] have been more directly impacted by decreased freshwaterdelivery and saltwater intrusion as a result of the creation of a large canal that intersectstheir study area [80].In wetland landscapes, geomorphic processes, such as the collapse of highly organicpeat soils, have the potential to drive landscape-level transitions from vegetated wetland toopen water [17,81]. A process that can lead to the conversion of vegetated wetland to openwater that has recently received considerable attention is “peat collapse” [17,82–84]. Peatcollapse is the process of rapid subsidence of the soil surface that alters elevation profiles ofthe landscape, and it can drive a permanent loss in emergent vegetation, causing a transi-tion to an open water environment [17,85–88]. Across the Florida Everglades, decades offreshwater management has altered the timing, distribution, and volume of freshwater de-livery to the coastal estuary and prompted the creation of the largest freshwater restorationeffort in the world [20]. There is growing concern that by 2040–2050, the coastal wetlandecosystems will begin a widespread conversion to estuarine conditions [89]. Uncertaintyremains at the global scale as to whether coastal wetland spatial extent will increase ordecrease over the next century [90], but clearly more regional scale research is needed atbroad temporal extents to understand past rates of vegetation loss and better predict theirfuture condition.5. ConclusionsIn this study, a methodology was developed to detect losses in coastal wetland vege-tated land cover at the multidecadal scale using a combination of historical panchromaticaerial photography and high-resolution satellite imagery. This methodology was testedRemote Sens. 2022, 14, 397614 of 17in a coastal peat marsh in the subtropical Florida Everglades, where two methods werecompared for classifying historical panchromatic photographs across a range of minimummapping units to assess the classification accuracy of open water body detection. Theresults indicate that when classifying historical wetland land cover using panchromaticaerial photography, manual thresholding and automated, machine learning methods canobtain equally accurate classifications in coastal wetland environments, but automatedclassification methods should be preferentially used if additional data can be obtained thatwill increase performance. It was also shown how an image analyst’s selection of the MMUcan impact the classification results, indicating that an MMU should be selected at thesmallest size possible, based on spatial resolution and the phenomenon of interest whiledelivering high-accuracy maps that allow for location-specific analysis of landscape-scaleprocesses such as peat collapse. Further, the importance of conducting point-based accu-racy assessments on change maps when performing digital change detection analyses wasalso underscored, as misclassifications can go undetected in classification map accuracyassessments, and this error will propagate to the change maps. Including the confusionmatrix in accuracy assessments assists in the evaluation of uncertainty and areal coverestimation, even when the overall accuracy of the maps is high. This methodology can beapplied across broad spatial extents in order to assess degradation and vegetation loss incoastal wetland environments that may have occurred over the last century due to the factof anthropogenic- and climate-related impacts in order to better predict an uncertain future.Supplementary Materials: The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/rs14163976/s1, Table S1: classified wetland maps accuracy as-sessment; Table S2: Change maps accuracy assessment; Figure S1: Frequency distribution of openwater pond sizes for each 1940 and 2012 classification across minimum mapping units.Author Contributions: Conceptualization, L.M.L., D.G. and T.G.T.; methodology, L.M.L. and D.G.;formal analysis, L.M.L., J.T.V. and D.G.; original draft preparation, L.M.L. and D.G.; writing—reviewand editing, L.M.L., D.G., J.T.V. and T.G.T.; visualization, L.M.L.; supervision, T.G.T. All authors haveread and agreed to the published version of the manuscript.Funding: L.M.L. and J.T.V. are grateful for support through the Everglades Foundation FIU ForEver-glades Scholarship. L.M.L. was additionally supported by Florida Sea Grant (R/C-S-86), with thecollaborative cooperation and financial support of the South Florida Water Management District, andthe National Park Service. L.M.L. was also supported by the National Science Foundation award(DBI-1237517), to the Florida Coastal Everglades Long-Term Ecological Research (FCE-LTER) program.Data Availability Statement: Scripts for the R code, classification maps, change maps, and accuracyassessments will be made publicly available via the Environmental Data Initiative data repository andstored in the Florida Coastal Everglades Long-Term Ecological Research data archive upon publication.Acknowledgments: The authors thank Jennifer Richards and three anonymous reviewers for theirinsightful comments. This is contribution #1469 from the Institute of Environment at Florida Interna-tional University.Conflicts of Interest: The authors declare no conflict of interest.References1.Barbier, E.B.; Hacker, S.D.; Kennedy, C.; Koch, E.W.; Stier, A.C.; Silliman, B.R. The value of estuarine and coastal ecosystemservices. Ecol. Monogr. 2011, 81, 169–193. [CrossRef]2.Mitsch, W.J.; Gosselink, J.G. Wetlands; John Wiley & Sons: New York, NY, USA, 2015; p. 752.3.Davidson, N.C.; Fluet-Chouinard, E.; Finlayson, C.M. Global extent and distribution of wetlands: Trends and issues. Mar. Freshw.Res. 2018, 69, 620–627. [CrossRef]4.Davidson, N.; Van Dam, A.; Finlayson, C.; McInnes, R. Worth of wetlands: Revised global monetary values of coastal and inlandwetland ecosystem services. Mar. Freshw. Res. 2019, 70, 1189–1194. [CrossRef]5.Mcleod, E.; Chmura, G.L.; Bouillon, S.; Salm, R.; Björk, M.; Duarte, C.M.; Lovelock, C.E.; Schlesinger, W.H.; Silliman, B.R. Ablueprint for blue carbon: Toward an improved understanding of the role of vegetated coastal habitats in sequestering CO2.Front. Ecol. Environ. 2011, 9, 552–560. [CrossRef]Remote Sens. 2022, 14, 397615 of 176.Kirwan, M.L.; Megonigal, J.P. Tidal wetland stability in the face of human impacts and sea-level rise. Nature 2013, 504, 53–60.[CrossRef] [PubMed]7.Osland, M.J.; Enwright, N.M.; Day, R.H.; Gabler, C.A.; Stagg, C.L.; Grace, J.B. Beyond just sea-level rise: Considering macroclimaticdrivers within coastal wetland vulnerability assessments to climate change. Glob. Change Biol. 2016, 22, 1–11. [CrossRef]8.Osland, M.J.; Gabler, C.A.; Grace, J.B.; Day, R.H.; McCoy, M.L.; McLeod, J.L.; From, A.S.; Enwright, N.M.; Feher, L.C.; Stagg, C.L.Climate and plant controls on soil organic matter in coastal wetlands. Glob. Change Biol. 2018, 24, 5361–5379. [CrossRef]9.Osland, M.J.; Grace, J.B.; Guntenspergen, G.R.; Thorne, K.M.; Carr, J.A.; Feher, L.C. Climatic controls on the distribution offoundation plant species in coastal wetlands of the conterminous United States: Knowledge gaps and emerging research needs.Estuaries Coasts 2019, 42, 1991–2003. [CrossRef]10.Herbert, E.R.; Boon, P.; Burgin, A.J.; Neubauer, S.C.; Franklin, R.B.; Ardón, M.; Hopfensperger, K.N.; Lamers, L.P.; Gell, P. A globalperspective on wetland salinization: Ecological consequences of a growing threat to freshwater wetlands. Ecosphere 2015, 6, 1–43.[CrossRef]11.Cahoon, D.R.; McKee, K.L.; Morris, J.T. How Plants Influence Resilience of Salt Marsh and Mangrove Wetlands to Sea-Level Rise.Estuaries Coasts 2020, 44, 883–898. [CrossRef]12.Cahoon, D.R.; Hensel, P.F.; Spencer, T.; Reed, D.J.; McKee, K.L.; Saintilan, N. Coastal wetland vulnerability to relative sea-levelrise: Wetland elevation trends and process controls. In Wetlands and Natural Resource Management; Springer: Berlin/Heidelberg,Germany, 2006; pp. 271–292.13.Day, J.W.; Christian, R.R.; Boesch, D.M.; Yáñez-Arancibia, A.; Morris, J.; Twilley, R.R.; Naylor, L.; Schaffner, L. Consequences ofclimate change on the ecogeomorphology of coastal wetlands. Estuaries Coasts 2008, 31, 477–491. [CrossRef]14.Morris, J.T.; Sundareshwar, P.; Nietch, C.T.; Kjerfve, B.; Cahoon, D.R. Responses of coastal wetlands to rising sea level. Ecology2002, 83, 2869–2877. [CrossRef]15.Schepers, L.; Kirwan, M.; Guntenspergen, G.; Temmerman, S. Spatio-temporal development of vegetation die-off in a submergingcoastal marsh. Limnol. Oceanogr. 2017, 62, 137–150. [CrossRef]16.Andres, K.; Savarese, M.; Bovard, B.; Parsons, M. Coastal wetland geomorphic and vegetative change: Effects of Sea-level riseand water management on brackish marshes. Estuaries Coasts 2019, 42, 1308–1327. [CrossRef]17.Chambers, L.G.; Steinmuller, H.E.; Breithaupt, J.L. Toward a mechanistic understanding of “peat collapse” and its potentialcontribution to coastal wetland loss. Ecology 2019, 100, e02720. [CrossRef]18.Wilson, B.J.; Servais, S.; Charles, S.P.; Davis, S.E.; Gaiser, E.E.; Kominoski, J.S.; Richards, J.H.; Troxler, T.G. Declines in plantproductivity drive carbon loss from brackish coastal wetland mesocosms exposed to saltwater intrusion. Estuaries Coasts 2018, 41,2147–2158. [CrossRef]19.Charles, S.P.; Kominoski, J.S.; Troxler, T.G.; Gaiser, E.E.; Servais, S.; Wilson, B.J.; Davis, S.E.; Sklar, F.H.; Coronado-Molina, C.;Madden, C.J. Experimental Saltwater Intrusion Drives Rapid Soil Elevation and Carbon Loss in Freshwater and BrackishEverglades Marshes. Estuaries Coasts 2019, 42, 1868–1881. [CrossRef]20.National Academies of Sciences, Engineering, and Medicine. Progress Toward Restoring the Everglades: The Seventh BiennialReview—2018; National Academies Press: Cambridge, MA, USA, 2019.21.Hardisky, M.; Gross, M.; Klemas, V. Remote sensing of coastal wetlands. Bioscience 1986, 36, 453–460. [CrossRef]22.Ozesmi, S.L.; Bauer, M.E. Satellite remote sensing of wetlands. Wetl. Ecol. Manag. 2002, 10, 381–402. [CrossRef]23.Klemas, V. Remote sensing techniques for studying coastal ecosystems: An overview. J. Coast. Res. 2011, 27, 2–17.24.Rundquist, D.C.; Narumalani, S.; Narayanan, R.M. A review of wetlands remote sensing and defining new considerations. RemoteSens. Rev. 2001, 20, 207–226. [CrossRef]25.Adam, E.; Mutanga, O.; Rugege, D. Multispectral and hyperspectral remote sensing for identification and mapping of wetlandvegetation: A review. Wetl. Ecol. Manag. 2010, 18, 281–296. [CrossRef]26.Moffett, K.B.; Nardin, W.; Silvestri, S.; Wang, C.; Temmerman, S. Multiple stable states and catastrophic shifts in coastal wetlands:Progress, challenges, and opportunities in validating theory using remote sensing and other methods. Remote Sens. 2015, 7,10184–10226. [CrossRef]27.Guo, M.; Li, J.; Sheng, C.; Xu, J.; Wu, L. A review of wetland remote sensing. Sensors 2017, 17, 777. [CrossRef] [PubMed]28.Mahdavi, S.; Salehi, B.; Granger, J.; Amani, M.; Brisco, B.; Huang, W. Remote sensing for wetland classification: A comprehensivereview. GIScience Remote Sens. 2018, 55, 623–658. [CrossRef]29.Belgiu, M.; Drăgu¸t, L. Random forest in remote sensing: A review of applications and future directions. ISPRS J. Photogramm.Remote Sens. 2016, 114, 24–31. [CrossRef]30.Breiman, L. Random forests. Mach. Learn. 2001, 45, 5–32. [CrossRef]31.Whiteside, T.G.; Bartolo, R.E. Mapping aquatic vegetation in a tropical wetland using high spatial resolution multispectral satelliteimagery. Remote Sens. 2015, 7, 11664–11694. [CrossRef]32.Franklin, S.E.; Ahmed, O.S. Object-based wetland characterization using radarsat-2 quad-polarimetric SAR data, landsat-8 OLIimagery, and airborne lidar-derived geomorphometric variables. Photogramm. Eng. Remote Sens. 2017, 83, 27–36. [CrossRef]33.Mutanga, O.; Adam, E.; Cho, M.A. High density biomass estimation for wetland vegetation using WorldView-2 imagery andrandom forest regression algorithm. Int. J. Appl. Earth Obs. Geoinf. 2012, 18, 399–406. [CrossRef]34.Wendelberger, K.S.; Gann, D.; Richards, J.H. Using Bi-seasonal worldview-2 multi-spectral data and supervised random forestclassification to map coastal plant communities in Everglades National Park. Sensors 2018, 18, 829. [CrossRef] [PubMed]Remote Sens. 2022, 14, 397616 of 1735.Morgan, J.L.; Gergel, S.E.; Coops, N.C. Aerial photography: A rapidly evolving tool for ecological management. BioScience 2010,60, 47–59. [CrossRef]36.Work, E.A.; Gilmer, D.S. Utilization of satellite data for inventorying prairie ponds and lakes. Photogramm. Eng. Remote Sens. 1976,42, 685–694.37.Lyon, J.G.; Greene, R. Use of aerial photographs to measure the historical areal extent of Lake Erie coastal wetlands. Photogramm.Eng. Remote Sens. 1992, 58, 1355–1360.38.Williams, D.C.; Lyon, J.G. Historical aerial photographs and a geographic information system (GIS) to determine effects oflong-term water level fluctuations on wetlands along the St. Marys River, Michigan, USA. Aquat. Bot. 1997, 58, 363–378. [CrossRef]39.Cserhalmi, D.; Nagy, J.; Kristóf, D.; Neidert, D. Changes in a wetland ecosystem: A vegetation reconstruction study based onhistorical panchromatic aerial photographs and succession patterns. Folia Geobot. 2011, 46, 351–371. [CrossRef]40.Ballanti, L.; Byrd, K.B.; Woo, I.; Ellings, C. Remote sensing for wetland mapping and historical change detection at the NisquallyRiver Delta. Sustainability 2017, 9, 1919. [CrossRef]41.Rapinel, S.; Clément, B.; Dufour, S.; Hubert-Moy, L. Fine-scale monitoring of long-term wetland loss using LiDAR data andhistorical aerial photographs: The example of the couesnon floodplain, France. Wetlands 2018, 38, 423–435. [CrossRef]42.Lillesand, T.M.; Kiefer, R.W.; Chipman, J. Remote Sensing and Image Interpretation; John Willey & Sons. Inc.: Hoboken, NJ,USA, 1994.43.Saura, S. Effects of minimum mapping unit on land cover data spatial configuration and composition. Int. J. Remote Sens. 2002, 23,4853–4880. [CrossRef]44.Knight, J.F.; Lunetta, R.S. An experimental assessment of minimum mapping unit size. IEEE Trans. Geosci. Remote Sens. 2003, 41,2132–2134. [CrossRef]45.Rutchey, K.; Godin, J. Determining an appropriate minimum mapping unit in vegetation mapping for ecosystem restoration: Acase study from the Everglades, USA. Landsc. Ecol. 2009, 24, 1351–1362. [CrossRef]46.Lodge, T.E. The Everglades Handbook: Understanding the Ecosystem; CRC Press: Boca Raton, FL, USA, 2016.47.Yao, Q.; Liu, K.-B. Dynamics of marsh-mangrove ecotone since the mid-Holocene: A palynological study of mangrove encroach-ment and sea level rise in the Shark River Estuary, Florida. PLoS ONE 2017, 12, e0173670. [CrossRef] [PubMed]48.Scholl, D.W. Recent sedimentary record in mangrove swamps and rise in sea level over the southwestern coast of Florida: Part 2.Mar. Geol. 1964, 2, 343–364. [CrossRef]49.Smith, W.G. Sedimentary Environments and Environmental Change in the Peat-Forming Area of South Florida. Ph.D. Thesis,The Pennsylvania State University, State College, PA, USA, 1970.50.Parkinson, R.W. Decelerating Holocene sea-level rise and its influence on Southwest Florida coastal evolution; a transgres-sive/regressive stratigraphy. J. Sediment. Res. 1989, 59, 960–972.51.Kaplan, S.W. Peat Records of Late Holocene Climate and Sea Level Change in South Florida. Ph.D. Thesis, University ofWisconsin–Madison, Madison, WI, USA, 2004.52.Volety, A.K.; Savarese, M.; Hoye, B.; Loh, A.N. Landscape Pattern: Present and Past Distribution of Oysters in South Florida CoastalComplex (Whitewater Bay/Oyster Bay/Shark to Robert’s Rivers); South Florida Water Management District Final Technical Report;Florida Gulf Coast University: West Palm Beach, FL, USA, 2009.53.Davis, S.; Ogden, J.C. Everglades: The Ecosystem and Its Restoration; CRC Press: Boca Raton, FL, USA, 1994.54.Childers, D.L.; Gaiser, E.; Ogden, L.A. The Coastal Everglades: The Dynamics of Social-Ecological Transformation in the South FloridaLandscape; Oxford University Press: Oxford, UK, 2019.55.McVoy, C.; Said, W.P.; Obeysekera, J.; VanArman, J.A.; Dreschel, T.W. Landscapes and Hydrology of the Predrainage Everglades;University Press of Florida Gainesville: Gainesville, FL, USA, 2011.56.Aich, S.; McVoy, C.W.; Dreschel, T.W.; Santamaria, F. Estimating soil subsidence and carbon loss in the Everglades AgriculturalArea, Florida using geospatial techniques. Agric. Ecosyst. Environ. 2013, 171, 124–133. [CrossRef]57.Hohner, S.M.; Dreschel, T.W. Everglades peats: Using historical and recent data to estimate predrainage and current volumes,masses and carbon contents. Mires Peat. 2015, 16, 1–15.58.Foster, A. Historic Aerial Photography of the Greater Everglades Archive and Geodatabase Development; USGS Publications Warehouse:Restond, VA, USA, 2004; pp. 2327–6932.59.Hijmans, R.J.; Van Etten, J.; Cheng, J.; Mattiuzzi, M.; Sumner, M.; Greenberg, J.A.; Lamigueiro, O.P.; Bevan, A.; Racine, E.B.;Shortridge, A. Raster: Geographic Data Analysis and Modeling. R Package Version 2.8-4. 2015. Available online: https://CRAN.R-project.org/package=raster (accessed on 19 June 2022).60.Kuhn, M.; Wing, J.; Weston, S.; Williams, A.; Keefer, C.; Engelhardt, A.; Cooper, T.; Mayer, Z.; Kenkel, B.; Team, R.C. Caret:Classification and Regression Training. R Package Version 6.0-80. 2020. Available online: https://CRAN.R-project.org/package=caret (accessed on 19 June 2022).61.Jordan, C.F. Derivation of leaf-area index from quality of light on the forest floor. Ecology 1969, 50, 663–666. [CrossRef]62.Rousel, J.; Haas, R.; Schell, J.; Deering, D. Monitoring Vegetation Systems in the Great Plains with ERTS. In Proceedings ofthe Third Earth Resources Technology Satellite—1 Symposium, NASA SP-351, Washington, DC, USA, 10–14 December 1973;pp. 309–317.Remote Sens. 2022, 14, 397617 of 1763.Barnes, E.; Clarke, T.; Richards, S.; Colaizzi, P.; Haberland, J.; Kostrzewski, M.; Waller, P.; Choi, C.; Riley, E.; Thompson, T.Coincident detection of crop water stress, nitrogen status and canopy density using ground based multispectral data. InProceedings of the Fifth International Conference on Precision Agriculture, Bloomington, MN, USA, 16–19 July 2000; p. 6.64.McFeeters, S.K. The use of the Normalized Difference Water Index (NDWI) in the delineation of open water features. Int. J.Remote Sens. 1996, 17, 1425–1432. [CrossRef]65.Gitelson, A.A.; Kaufman, Y.J.; Merzlyak, M.N. Use of a green channel in remote sensing of global vegetation from EOS-MODIS.Remote Sens. Environ. 1996, 58, 289–298. [CrossRef]66.Huete, A.R. A soil-adjusted vegetation index (SAVI). Remote Sens. Environ. 1988, 25, 295–309. [CrossRef]67.Jiang, Z.; Huete, A.R.; Didan, K.; Miura, T. Development of a two-band enhanced vegetation index without a blue band. RemoteSens. Environ. 2008, 112, 3833–3845. [CrossRef]68.Serra, J. Introduction to mathematical morphology. Comput. Vis. Graph. Image Processing 1986, 35, 283–305. [CrossRef]69.Heijmans, H.J. Mathematical morphology: A modern approach in image processing based on algebra and geometry. SIAM Rev.1995, 37, 1–36. [CrossRef]70.Soille, P.; Pesaresi, M. Advances in mathematical morphology applied to geoscience and remote sensing. IEEE Trans. Geosci.Remote Sens. 2002, 40, 2042–2055. [CrossRef]71.R Core Team. R: A Language and Environment for Statistical Computing; R Foundation for Statistical Computing: Vienna, Austria, 2022.72.Jensen, J.R. Introductory Digital Image Processing: A Remote Sensing Perspective; Prentice-Hall Inc.: Hoboken, NJ, USA, 1996.73.Tillé, Y.; Matei, A.; Sampling: Survey Sampling. R Package Version 2.8. 2012. Available online: https://CRAN.R-project.org/package=sampling (accessed on 19 June 2022).74.Olofsson, P.; Foody, G.M.; Herold, M.; Stehman, S.V.; Woodcock, C.E.; Wulder, M.A. Good practices for estimating area andassessing accuracy of land change. Remote Sens. Environ. 2014, 148, 42–57. [CrossRef]75.McCarthy, M.J.; Jessen, B.; Barry, M.J.; Figueroa, M.; McIntosh, J.; Murray, T.; Schmid, J.; Muller-Karger, F.E. Automated high-resolution time series mapping of mangrove forests damaged by hurricane Irma in Southwest Florida. Remote Sens. 2020, 12, 1740.[CrossRef]76.Wood, E.M.; Pidgeon, A.M.; Radeloff, V.C.; Keuler, N.S. Image texture as a remotely sensed measure of vegetation structure.Remote Sens. Environ. 2012, 121, 516–526. [CrossRef]77.Hudak, A.T.; Wessman, C.A. Textural analysis of historical aerial photography to characterize woody plant encroachment inSouth African savanna. Remote Sens. Environ. 1998, 66, 317–330. [CrossRef]78.Caridade, C.; Marçal, A.R.; Mendonça, T. The use of texture for image classification of black & white air photographs. Int. J.Remote Sens. 2008, 29, 593–607.79.Biskup, B.; Scharr, H.; Schurr, U.; Rascher, U. A stereo imaging system for measuring structural parameters of plant canopies.Plant Cell Environ. 2007, 30, 1299–1308. [CrossRef]80.U.S. Army Corps of Engineers. Comprehensive Everglades Restoration Plan: Picayune Strand Restoration (Formally the South GoldenGate Estates Ecosystem Restoration); United States Army Corps of Engineers: Washington, DC, USA, 2004; p. 488.81.Sklar, F.H.; Meeder, J.F.; Troxler, T.G.; Dreschel, T.; Davis, S.E.; Ruiz, P.L. The Everglades: At the Forefront of Transition. In Coastsand Estuaries; Elsevier: Amsterdam, The Netherlands, 2019; pp. 277–292.82.DeLaune, R.D.; Nyman, J.A.; Patrick Jr, W. Peat collapse, ponding and wetland loss in a rapidly submerging coastal marsh.J. Coast. Res. 1994, 4, 1021–1030.83.Wilson, B.J.; Servais, S.; Mazzei, V.; Kominoski, J.S.; Hu, M.; Davis, S.E.; Gaiser, E.; Sklar, F.; Bauman, L.; Kelly, S. Salinity pulsesinteract with seasonal dry-down to increase ecosystem carbon loss in marshes of the Florida Everglades. Ecol. Appl. 2018, 28,2092–2108. [CrossRef] [PubMed]84.Solohin, E.; Widney, S.E.; Craft, C.B. Declines in plant productivity drive loss of soil elevation in a tidal freshwater marsh exposedto saltwater intrusion. Ecology 2020, 101, e03148. [CrossRef] [PubMed]85.Cahoon, D.R.; Hensel, P.; Rybczyk, J.; McKee, K.L.; Proffitt, C.E.; Perez, B.C. Mass tree mortality leads to mangrove peat collapseat Bay Islands, Honduras after Hurricane Mitch. J. Ecol. 2003, 91, 1093–1105. [CrossRef]86.Day, J.W.; Kemp, G.P.; Reed, D.J.; Cahoon, D.R.; Boumans, R.M.; Suhayda, J.M.; Gambrell, R. Vegetation death and rapid loss ofsurface elevation in two contrasting Mississippi delta salt marshes: The role of sedimentation, autocompaction and sea-level rise.Ecol. Eng. 2011, 37, 229–240. [CrossRef]87.Schepers, L.; Brennand, P.; Kirwan, M.L.; Guntenspergen, G.R.; Temmerman, S. Coastal Marsh Degradation Into Ponds InducesIrreversible Elevation Loss Relative to Sea Level in a Microtidal System. Geophys. Res. Lett. 2020, 47, e2020GL089121. [CrossRef]88.Wang, C.; Schepers, L.; Kirwan, M.L.; Belluco, E.; D’Alpaos, A.; Wang, Q.; Yin, S.; Temmerman, S. Different coastal marsh sitesreflect similar topographic conditions under which bare patches and vegetation recovery occur. Earth Surf. Dyn. 2021, 9, 71–88.[CrossRef]89.Parkinson, R.W.; Wdowinski, S. Accelerating sea-level rise and the fate of mangrove plant communities in South Florida, USA.Geomorphology 2022, 412, e108329. [CrossRef]90.Schuerch, M.; Spencer, T.; Temmerman, S.; Kirwan, M.L.; Wolff, C.; Lincke, D.; McOwen, C.J.; Pickering, M.D.; Reef, R.;Vafeidis, A.T. Future response of global coastal wetlands to sea-level rise. Nature 2018, 561, 231–234. [CrossRef] | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/362338523Evaluación de diferentes métodos para la estimación del PIB potencial: el caso deMéxicoArticle in Estudios Económicos de El Colegio de México · July 2022DOI: 10.24201/ee.v37i2.432CITATIONS0READS713 authors:Some of the authors of this publication are also working on these related projects:On the importance of the probabilistic model in identifying the most decisive games in a tournament View projectFrancisco CoronaInstituto Nacional de Estadística y Geografía20 PUBLICATIONS 82 CITATIONS SEE PROFILEPedro Paulo Orraca RomanoEl Colegio de la Frontera Norte39 PUBLICATIONS 188 CITATIONS SEE PROFILEJesus Lopez PerezInstituto Nacional de Estadística y Geografía9 PUBLICATIONS 1 CITATION SEE PROFILEAll content following this page was uploaded by Pedro Paulo Orraca Romano on 30 July 2022.The user has requested enhancement of the downloaded file.EVALUACI´ON DE DIFERENTES M´ETODOS PARALA ESTIMACI´ON DEL PIB POTENCIAL:EL CASO DE M´EXICOEVALUATING DIFFERENT METHODS OFPOTENTIAL GDP ESTIMATES:THE CASE OF MEXICOFrancisco CoronaInstituto Nacional de Estad´ıstica y Geograf´ıaPedro OrracaEl Colegio de la Frontera NorteJes´us L´opez-P´erezInstituto Nacional de Estad´ıstica y Geograf´ıaResumen: Con el objetivo de contribuir a la literatura que se centra en la esti-maci´on del Producto Interno Bruto (PIB) potencial, en este trabajose eval´ua el funcionamiento de cuatro m´etodos de estimaci´on del PIBpotencial trimestral de M´exico para el periodo 1998:T1-2020:T2. Losprocedimientos utilizados para su estimaci´on son: 1) m´etodos heur´ısti-cos, 2) el filtro de Hodrick Prescott, 3) el Modelo de Factores Din´amicosno estacionario y 4) la descomposici´on Permanente-Transitorio (PT) deGonzalo y Granger (1995). Se concluye que, por razones econom´etri-cas, los mejores resultados se obtienen al usar la descomposici´on PT.Abstract:With the purpose of contributing to the literature that focuses on theestimation of Potential Gross Domestic Product (GDP), this studyevaluates four different estimation methods of Potential GDP usingquarterly data from Mexico for the period 1998:Q1-2020:Q2. The pro-cedures used for its estimation are: 1) Heuristic methods, 2) The Ho-drick Prescott filter, 3) Non-Stationary Dynamic Factor Model, and 4)The Permanent-Transient (PT) decomposition of Gonzalo and Granger(1995). We conclude that, econometrically, the best results are ob-tained when using the PT decomposition.Clasificaci´on JEL/JEL Classification: C38, C52, E32Palabras clave/keywords:filtro HP; descomposici´on PT; m´etodos heur´ısticos;modelo de factores din´amicos; PIB potencialFecha de recepci´on: 28 VI 2021Fecha de aceptaci´on: 28 VII 2021https://doi.org/10.24201/ee.v37i2.432Estudios Econ´omicos,vol. 37, n´um. 2, julio-diciembre 2022,p´aginas 285-313286ESTUDIOS ECON ´OMICOShttps://doi.org/10.24201/ee.v37i2.4321. Introducci´onEs sabido que para la toma de decisiones de pol´ıtica econ´omica e-xiste la necesidad de tener elementos objetivos que permitan tomardecisiones informadas. Asimismo, es trascendental entender las for-talezas y debilidades de los m´etodos o herramientas que coadyuvana dicha toma de decisiones. Conocer el crecimiento potencial de unaeconom´ıa es muy relevante en temas de pol´ıtica econ´omica, ya que,por ejemplo, cuando el Producto Interno Bruto (PIB) de un pa´ıs seencuentra por debajo de lo que en teor´ıa pudiese producir, la im-plementaci´on de pol´ıticas fiscales expansivas pueden ser claves pararevertir el efecto de tal comportamiento. Desde la ´optica de la pol´ıticamonetaria, conocer la brecha de producci´on entre el PIB real y el PIBpotencial permite determinar si la econom´ıa requiere o no de est´ı-mulos monetarios. Mientras que a corto plazo las estimaciones de labrecha entre el PIB real y el PIB potencial proporcionan un punto dereferencia que permiten evaluar las presiones inflacionarias, a media-no plazo la tendencia estimada del PIB potencial ayuda a determinarel ritmo de crecimiento sostenible de la econom´ıa (De Masi, 1997).De esta forma, es claro que resulta necesario contar con los ele-mentos que permitan tener una correcta conceptualizaci´on y medici´onde este fen´omeno macroecon´omico, para as´ı, poder identificar e im-plementar las pol´ıticas econ´omicas apropiadas.Este trabajo tiene como objetivo contribuir a la literatura quese centra en la estimaci´on del PIB potencial al evaluar dos m´etodostradicionales y dos nuevas metodolog´ıas y, al realizar una aplicaci´onpara el caso de M´exico, proporcionar informaci´on que permita a loshacedores de pol´ıtica econ´omica del pa´ıs llegar a una conclusi´on ra-zonada sobre qu´e t´ecnica es la que brinda las mayores ventajas en lossentidos econom´etricos y estad´ısticos.El PIB potencial puede analizarse desde dos perspectivas: 1) lakeynesiana, relacionada a la evoluci´on del PIB dado el eficiente uso delos factores de la producci´on bajo un escenario de inflaci´on estable, odicho de otro modo, donde representa la producci´on m´axima que unaeconom´ıa puede sostener sin generar un aumento en la inflaci´on, locual implica que bajo este enfoque una brecha de producci´on negativa(positiva) debe tener como resultado un decremento (incremento) enla inflaci´on, y 2) desde un enfoque neocl´asico, que indica que el PIBpotencial est´a relacionado con el comportamiento permanente de laproducci´on, determinado por choques ex´ogenos de oferta y demanda.Para una discusi´on m´as amplia sobre este tema, v´eanse Acevedo-Fern´andez (2009), Mantescu y Lazar (2014), Kuttner (1994), Heath(2012) y Gavin (2012), entre otros.EVALUACI ´ON DE DIFERENTEShttps://doi.org/10.24201/ee.v37i2.432287Dado que el PIB potencial es un concepto te´orico y por lo tantono es observable, no puede cuantificarse como s´ı puede hacerse parael PIB real en el contexto de estad´ıstica oficial. En la pr´actica, el PIBpotencial suele estimarse construyendo medidas de la tendencia delPIB real que suavizan las fluctuaciones de los ciclos econ´omicos. Hasido com´un recurrir a la extracci´on de tendencias determin´ısticas, lautilizaci´on de filtros estad´ısticos y la estimaci´on de variables latentes,principalmente a trav´es del filtro de Kalman, para tener aproxima-ciones del PIB potencial (Ozbek y Ozlale, 2005).Algunos trabajos que se centran en estimar al PIB potencial yla brecha del producto son Kuttner (1994) e Islas y Guerrero (2019)para el caso de Estados Unidos, Theoduloz (2010) para la econom´ıade Uruguay, Kichian (1999) quien lo analiza para el caso de Canad´a,Scacciavillani y Swagel (2002) para Israel, y Hoffman y Tapia (2003)y M´endez et al. (2013) quienes se enfocan en los pa´ıses de Am´ericaLatina.Para el caso de M´exico, se encuentra el trabajo de Loria et al.(2008) quienes utilizan modelos estructurales de series de tiempo condatos de 1980 a 2006 para estimar el producto potencial, la brechadel producto y los ciclos econ´omicos para el PIB de M´exico, y el deFaal (2005) en el cual se hace una descomposici´on del crecimientoecon´omico para M´exico desde 1960 hasta 2003 en sus componentesde tendencia y ciclo utilizando un modelo de componentes no obser-vados. Tambi´en se encuentra la investigaci´on de Cervantes y Arenas(2004), quienes utilizan funciones Cobb-Douglas y el filtro de Hodrick-Prescott (HP) para estimar el producto potencial para los a˜nos 1980-2003, y m´as recientemente, Ventosa-Santaul`aria et al. (2021), dondeusan t´ecnicas de series de tiempo para identificar, de manera robusta,cambios y explicaciones a la tasa de crecimiento decreciente causadapor fuerzas de largo plazo en lugar de las crisis econ´omicas.1En el contexto de estad´ıstica oficial en M´exico, el Instituto Na-cional de Estad´ıstica y Geograf´ıa (INEGI) ha realizado y realiza inten-tos por esquematizar las crestas y valles de la econom´ıa mexicana. Sibien, esto es diferente a la estimaci´on del PIB potencial, intr´ınseca-mente algunos conceptos est´an relacionados, particularmente, en lossiguientes aspectos: i) concepci´on de composiciones latentes, puestoque se utilizan un grupo de variables relacionadas al PIB para construirindicadores compuestos que permitan identificar cu´ando la econom´ıaest´a en fase de recesi´on o expansi´on y, ii) visi´on del ciclo econ´omico,1 Para una descripci´on de los principales hechos estilizados del ciclo econ´omicoen M´exico, v´ease Cuadra (2008).288ESTUDIOS ECON ´OMICOShttps://doi.org/10.24201/ee.v37i2.432dado que existe una perspectiva que trabaja con los niveles de seriesde tiempo desestacionalizadas y otra que extrae el comportamientoc´ıclico de manera suavizada a las series de tiempo; a la primera sele conoce como “visi´on cl´asica” y a la segunda como “visi´on de cre-cimiento” (Heath, 2012).Vale la pena hacer aclarar que la estimaci´on del PIB potencial s´ıpermite identificar las fechas de inicio y fin de posibles recesiones yexpansiones; sin embargo, si el objetivo primordial es la identificaci´onde las crestas y valles, las t´ecnicas estad´ısticas y econom´etricas puedenir en otra direcci´on, sin ser necesario estimar el PIB potencial.En cualquier caso, y en sentido estrictamente metodol´ogico, n´o-tese que el PIB potencial puede ser abordado bajo una idea de esti-maci´on de variables latentes. Dado que el PIB real es generalmenteno estacionario, resulta id´oneo trabajar con los niveles de la serie,es decir, bajo el enfoque cl´asico de ciclos econ´omicos.Este hechono est´a necesariamente separado de la conceptualizaci´on keynesianadesde una perspectiva econom´etrica.Las contribuciones metodol´ogicas de esta investigaci´on se basanen proponer dos t´ecnicas alternativas de estimaci´on del PIB potencialde M´exico usando series de tiempo no estacionarias, las cuales nohan sido trabajadas por la literatura previa. Estas dos metodolog´ıasse comparan con dos procedimientos ampliamente difundidos en laliteratura, como lo son m´etodos heur´ısticos que asumen tendenciasdetermin´ısticas de crecimiento y la utilizaci´on del filtro de HP.La primera metodolog´ıa propuesta se basa en estimar las compo-nentes de un Modelo de Factores Din´amicos (MFD) usando series detiempo no estacionarias y correlacionadas con la actividad econ´omica.Lo anterior se hace siguiendo la propuesta de Corona et al. (2020),extrayendo un factor din´amico cointegrado con las observaciones, asu-miendo que el factor din´amico es no estacionario o I(1) y el erroridiosincr´atico es estacionario o I(0). La ventaja de esta t´ecnica esque, econom´etricamente, el factor din´amico es la tendencia com´un delas observaciones y, por ende, una estimaci´on del PIB potencial. Otraventaja es que se puede validar la significancia estad´ıstica del factorsobre las variables que componen al MFD y estimar los intervalos deconfianza del factor din´amico y, por ende, tener la estimaci´on de laincertidumbre asociada al c´alculo del PIB potencial.La segunda metodolog´ıa propuesta usa la descomposici´on Perma-nente-Transitoria (PT) de Gonzalo y Granger (1995). Para ello, se cal-culan las elasticidades de largo plazo de una funci´on Cobb-Douglas ysujeto a estos resultados, la componente permanente estimada se aso-cia al PIB potencial. Este procedimiento permite conocer y racionalizarEVALUACI ´ON DE DIFERENTEShttps://doi.org/10.24201/ee.v37i2.432289el orden de cointegraci´on entre la producci´on, y los factores trabajo ycapital para tener una descomposici´on PT v´alida que permite encon-trar una estimaci´on del PIB potencial que, de manera te´orica, subyacede la din´amica permanente y transitoria que existe entre los factoresde la producci´on.2En este trabajo no es prioridad analizar los resultados en unsentido econ´omico ni hacer recomendaciones en materia de pol´ıticaecon´omica, sino brindar a los tomadores de decisi´on y economistasaplicados, herramientas de ´ındole cuantitativo para los fines que seconsideren apropiados.El estudio muestra que los m´etodos de MFD no estacionarios yla descomposici´on PT tienen ventajas respecto a los m´etodos tradi-cionales de modelos heur´ısticos y filtros HP en el sentido que puedenevaluarse los supuestos que garantizan la consistencia estad´ıstica delos resultados. Adem´as, la descomposici´on PT es la que genera losmejores resultados debido a que su componente c´ıclico tiene unaaportaci´on importante en el comportamiento com´un de los ciclos ylas estimaciones son robustas a trav´es del tiempo.El trabajo se estructura de la siguiente forma. En la segundasecci´on se resumen los m´etodos para la estimaci´on del PIB potencial.En la tercera secci´on se presentan los datos, la aplicaci´on emp´ıricay un an´alisis de robustez. Finalmente, la cuarta secci´on describe lasconclusiones.2. M´etodosEn esta secci´on se introduce la notaci´on y se describen las t´ecnicasconsideradas en este trabajo para la estimaci´on del PIB potencial.2 Otros trabajos que se enfocan en M´exico han utilizado metodolog´ıas distintasa las aqu´ı presentadas para estimar la tendencia del PIB. Noriega y Rodr´ıguez-P´erez (2012), quienes examinan la evoluci´on del PIB real y el PIB per c´apita,modelan los cambios estructurales por medio de funciones log´ısticas de transici´onsuave entre reg´ımenes o tendencias lineales, donde los periodos de transici´on seestiman de manera end´ogena. Carrillo et al. (2018), quienes analizan los deter-minantes de la tasa de inter´es neutral o natural, estiman la brecha del productoy el crecimiento potencial mediante el modelo propuesto por Laubach y Williams(2003) y modelos de vectores autorregresivos bayesianos con interceptos que var´ıanen el tiempo.290ESTUDIOS ECON ´OMICOShttps://doi.org/10.24201/ee.v37i2.4322.1 M´etodos heur´ısticosConsidere al PIB observado como yt y al PIB potencial como yPt ambospara t = 1, . . .N. El objetivo es estimar yPt usando reglas heur´ısticas.Por ejemplo, para el caso de M´exico, Guerrero-de-Lizardi (2020) llamaa una alternativa de estimaci´on del PIB potencial como “la perspectivade la Secretar´ıa de Hacienda y Cr´edito P´ublico”, la cual surge a partirde lo establecido en el Reglamento de la Ley Federal de Presupuesto yResponsabilidad Hacendaria y se basa en asumir a yPt como la tasa decrecimiento media anual del PIB observado de manera constante paraun periodo determinado de tiempo. A la par de esta idea, podemosconsiderar la tasa de crecimiento media anual del PIB observado oreal en un periodo de inflaci´on constante, lo anterior, para acercarnosm´as a la definici´on original de PIB potencial comentada en la secci´onanterior.De esta forma, se considera la siguiente medida heur´ıstica del PIBpotencial:ˆyPt = ˆytP−1(1 + ∆j ) para j = 1, 2. Cuando t = 1, se asume que ˆyP0 = y1; es decir, la condici´on iniciales equivalente al valor de la primera observaci´on y ∆1 se refiere a latasa de crecimiento media anual que hay en todo el periodo de tiempoconsiderado, mientras que ∆2 se puede definir de la siguiente forma:∆2 = (∆1|ϖ = c)(2)es decir, la tasa de crecimiento media anual del PIB cuando latasa de inflaci´on anual, ϖt, es constante, es decir, ϖ = c. En este tra-bajo, asumimos una banda de inflaci´on como dos veces la desviaci´onalrededor de la mediana de la inflaci´on.3 Para ello, al existir depen-dencias temporales en la inflaci´on, se estima la varianza de la inflaci´oncon errores est´andar robustos ante la presencia de autocorrelaci´on yheteroscedasticidad (Newey y West, 1987), denotada como ˆσ2ϖ,HA:4ˆσ2ϖ,HA =1N(N−1)� N�t=1ϖ2t + 2q�v=1��1 −vq+1�N�t=v+1ϖtϖt−v��(3)3 Asumimos la mediana dado que, aunque suponemos estacionariedad en lainflaci´on anual, la distribuci´on de esta variable no es sim´etrica.4 HA hace referencia a siglas en ingl´es de heteroskedasticity y autocorrelation. (1)EVALUACI ´ON DE DIFERENTEShttps://doi.org/10.24201/ee.v37i2.432291donde v es el par´ametro iterativo que es funci´on de q, la constanteque aproxima el grado de dependencia temporal en los datos. En estetrabajo asumimos que q = 2, de forma similar a como lo proponen Bai(2003) y Corona et al. (2020). Estos estimadores o aproximaciones delPIB potencial son denotados como ˆyP,MH1ty ˆyP,MH2t, respectivamente.2.2 Filtro de Hodrick-PrescottEl filtro de HP es el m´etodo m´as convencional para extraer tendenciasdetermin´ısticas en series de tiempo econ´omicas dado que la vastaevidencia emp´ırica concluye que es un buen m´etodo para representarlas tendencias de largo plazo, y, en consecuencia, la parte c´ıclica delas series de tiempo. V´ease, por ejemplo, McElroy (2008) y Phillipsy Shi (2021).Consideremos que yt = yPt + ct, es decir, que el PIB puede des-componerse en un componente de tendencia o permanente, es decir,el patr´on de largo plazo que caracteriza a la serie de tiempo m´asun componente c´ıclico, ct, que es dictaminado por las fluctuacionesecon´omicas exhibidas como las desviaciones que existen entre lo obser-vado y el componente de tendencia. En su representaci´on estad´ıstica,Guerrero (2007) muestra que la estimaci´on de la tendencia se basa ensolucionar un problema de M´ınimos Cuadrados Penalizados; en otraspalabras, se soluciona una funci´on cuadr´atica M (λ), donde λ es unaconstante de penalizaci´on que tambi´en se le conoce como par´ametrode suavizamiento tal que cuando λ → 0, implica que ct → 0. Enconsecuencia, para estimar yPt se debe resolver el siguiente problemade optimizaci´on:mincM (λ) =�y − yP �′ �y − yP �+ λyP ′K′KyP ′(4)donde y, yP y c son las representaciones vectoriales de yt, yPty ct, respectivamente; mientras que K es una matriz de dimensi´on(N − 2) × N definida de la siguiente forma:1− 2100...000K =01 −210000(5)...00000... 1 −210[]292ESTUDIOS ECON ´OMICOShttps://doi.org/10.24201/ee.v37i2.432Puede mostrarse que la soluci´on para obtener ˆyP se obtiene alderivar M (λ) respecto a c e igualar a 0, de tal forma que la estimaci´onde la tendencia es:ˆyP = (IN + λK′K)−1y(6)donde IN es la matriz de identidad. N´otese que (6) depende de λ,el par´ametro de suavizaci´on. En Guerrero (2007) se presenta un estu-dio sobre la estimaci´on de λ, donde se propone fijar un porcentaje desuavidad deseado para la tendencia y de ah´ı deducir el valor de dichopar´ametro; sin embargo, en la pr´actica suele dar buenos resultado laregla emp´ırica que emplea la Organizaci´on para la Cooperaci´on y elDesarrollo Econ´omico (v´ease OCDE, 2012), donde la idea consiste enrepresentar fluctuaciones que ocurren cada N ∗ periodos, por ejemplo18, 25 o 30 trimestres (v´ease Guerrero y Corona, 2018). La expresi´onpara obtener λ viene dada de la siguiente manera:λ =14�1 − cos� 2πN∗��2(7)En este trabajo, dados los resultados emp´ıricos obtenidos porGuerrero y Corona (2018), quienes concluyen que el ciclo econ´omicomexicano se representa mejor considerando una campana de oscilaci´onde cada 6 a˜nos (72 meses), en consecuencia, consideramos N ∗ = 18trimestres. En este caso, denotamos al estimador de HP como ˆyP,HPt.2.3 Modelo de factores din´amicosUn m´etodo econom´etrico alternativo y poco utilizado para extraertendencias, se basa en la estimaci´on de las componentes de un MFD.Esta t´ecnica permite estimar, dado un amplio conjunto de variablesecon´omicas y financieras, factores comunes, los cuales resumen lasdin´amicas de las series de tiempo que conforman al sistema. El MFD,que permite representar la tendencia com´un de la econom´ıa, es escritocomo sigue:Xt = PFt + et,(8)Ft = Ft−1 + ηt,(9)et = Γet−1 + at(10)EVALUACI ´ON DE DIFERENTEShttps://doi.org/10.24201/ee.v37i2.432293donde Xt es un vector de η variables que, en principio, satisfacencor (Xt, yt) ̸= 0, es decir, que existe correlaci´on diferente de 0 entreel PIB y el vector de covariables, mientras que P es el vector quedefine la contribuci´on del factor din´amico no estacionario, Ft, sobrelas observaciones, Xt y et es el error idiosincr´atico estacionario. Enconsecuencia, Γ describe la din´amica del error idiosincr´atico, mientrasque finalmente ηt y at son los errores del factor y el componente idio-sincr´atico respectivamente, los cuales se asumen como ruido blanco.Las implicaciones econom´etricas de que Ft sea I(1) y et sea I(0)son muy importantes, ya que implica que las Xt sean no estacionarias,pero la combinaci´on lineal et = Xt−PFt s´ı lo sea, de tal forma que Ftest´e cointegrada con las Xt y tambi´en sea la tendencia com´un de lasobservaciones. Adem´as, implica que cada par de variables del vectorXt son cointegradas (Bai, 2004; Corona et al., 2017, 2020).La estimaci´on del factor com´un se realiza siguiendo el proce-dimiento de Corona et al. (2020), quienes adaptan el procedimientode Doz et al.(2011) para series de tiempo no estacionarias.Enresumen, el procedimiento es el siguiente:1. Estimar con la metodolog´ıa de Bai (2004) a ˜Ft, ˜P y ˜et, de tal formaque ˜P es √n veces el primer vector propio de X′X y ˜F = X ˜P/ny en consecuencia ˜et = Xt − ˜P˜Ft, donde X = (X1, . . ., XT).2. Fijar ˆΩ = diag [Σˆe], σ2η =�Nt=1 ∆ ˜F 2tN−1, X0 = X1 y σ2η0 = 107, dondeΣˆe es la matriz de covarianza del error idiosincr´atico.3. Usar los par´ametros del paso anterior y ˜P, ˜F y X en el suaviza-miento de Kalman para obtener finalmente ˆyPt= E(Ft|X1:N).En este caso, el estimador obtenido a trav´es de la representaci´onde factores comunes se renombra como ˆyP,MF Dt.2.4 Descomposici´on PT de Gonzalo y Granger (1995)Los m´etodos anteriores son de car´acter heur´ıstico, estad´ıstico o econo-m´etrico, sin embargo, se apartan del concepto econ´omico de PIB po-tencial, al menos, de la perspectiva keynesiana. Como vimos en laintroducci´on, existen algunos esfuerzos por estimar el PIB potenciala trav´es de la estimaci´on de funciones de producci´on tipo Cobb-Douglas, es decir, de la conceptualizaci´on te´orica y, posteriormente,usar diferentes procedimientos estad´ısticos para extraer tendencias ocomportamientos de largo plazo que representen la producci´on poten-cial, principalmente, a trav´es del filtro de Kalman.294ESTUDIOS ECON ´OMICOShttps://doi.org/10.24201/ee.v37i2.432En este trabajo procedemos a utilizar esta idea, pero se usa unametodolog´ıa econom´etrica que se adapta de manera natural a la ex-tracci´on de componentes de tendencia. En otras palabras, a trav´es delas posibles relaciones de cointegraci´on en el vector Yt = (yt, lt, kt),donde lt y kt son los factores trabajo y capital respectivamente, sepuede estimar un componente permanente (de ah´ı la P en PT) paracada variable, y, por ende, la estimaci´on del PIB potencial.Dado que cointegraci´on implica que Yt tiene una representaci´onde factores (v´ease Stock y Watson, 1989), definida ahora como Yt =Aft + εt, entonces ft ∼ I (1) puede ser asociada a los efectos perma-nentes o de largo plazo mientras que εt ∼ I (0) a los efectos transi-torios o de corto plazo (de ah´ı la T en PT). En este sentido, Gonzaloy Granger (1995) proponen una descomposici´on v´alida que permiteencontrar una descomposici´on PT.Supongamos que Yt ∼ I (1) y, adem´as, que las variables est´ancointegradas de tal forma que se puede expresar en su forma de Vectorde Correcci´on de Errores (VEC, por sus siglas en ingl´es):∆Yt = αβ′Yt−p + Γ1∆Yt−1 + Γ2∆Yt−2 + . . . + Γp−1∆Yt−p+1 + ut(11)donde βYt representa la ecuaci´on de cointegraci´on.De estaforma, 3 > rango (αβ′) > 0 = m, lo que quiere decir que hay mcombinaciones lineales de las variables que son estacionarias o I(0).Lo anterior implica que s = 3 − m combinaciones lineales que sonI(1). En este contexto, Gonzalo y Granger (1995) proponen que unadescomposici´on PT debe satisfacer lo siguiente:1. Yt = Pt + Tt.2. ∆Pt y Tt son estacionarias.3. Para el modelo H (L) Yt = ut:a) lims→∞∂Et(Yt+s )∂uPt̸= 0b) lims→∞∂Et(Yt+s )∂uTt= 0Es decir, Pt ∼ I (1) y Tt ∼ I (0) y las funciones de impulso-respuesta indican que ´unicamente los choques permanentes tienenefecto en el largo plazo en Yt. De esta manera, condicionado a losresultados de cointegraci´on, la descomposici´on PT puede escribirse dela siguiente forma:Yt = A1ft + A2β′Yt = Pt + Tt(12)EVALUACI ´ON DE DIFERENTEShttps://doi.org/10.24201/ee.v37i2.432295donde A1 = β⊥(α′⊥β⊥)−1, ft = α′⊥Yt y A2 = α(β′α)−1, siendoα⊥ y β⊥ los complementos ortogonales de α y β respectivamente.En Gonzalo y Granger (1995) se muestra el procedimiento paraestimar A1 y ft basado en m´axima verosimilitud. N´otese que A2 yZt = β′Yt pueden estimarse directamente del ejercicio tradicional deJohansen (1991). En consecuencia, el estimador obtenido por estem´etodo, ˆPt = ˆA1�ft = ( ˆP yt , ˆP λt , ˆP kt ), tomando la columna relacionadayt, se le renombra como ˆyP,P Tty tiene las caracter´ısticas de 1) serno estacionario y 2) representa los efectos de largo plazo del PIB, porende, es un estimador del PIB potencial.En otra aplicaci´on parael caso de M´exico, Corona y Orraca (2019) extraen el componentepermanente y transitorio de las remesas mexicanas.3. Aplicaci´on emp´ırica3.1 DatosLas variables utilizadas en este trabajo se basan en Corona et al.(2021) quienes construyen un MFD para estimar oportunamente al In-dicador Global de la Actividad Econ´omica (IGAE). Si bien es cierto queel IGAE no representa totalmente al PIB, s´ı representa su mejor proxymensual, ya que debido a sus caracter´ısticas se considera que satisfacerazonablemente los criterios necesarios para desagregar el PIB trimes-tral (Guerrero, 2004) y, adem´as, es la variable que com´unmente seutiliza en M´exico como una aproximaci´on del PIB mensual (Elizondo,2019). En este sentido, se espera que las variables que se relacionancon el IGAE tambi´en est´en relacionadas con el PIB. De hecho, tenerestimaciones oportunas del IGAE representan la mejor aproximaci´onpara tener tambi´en estimaciones oportunas del PIB (G´alvez-Soriano,2020; Corona et al., 2021).El periodo seleccionado es el 1998:01-2020:10, el cual conside-ramos apropiado dado que abarca periodos importantes de la econom´ıa,como la crisis financiera mundial acontecida en 2009 y el segundotrimestre del 2020, esta ´ultima que marca el inicio de la pandemia dela COVID-19, as´ı como periodos de estabilidad econ´omica y una faserecesiva ocurrida entre 2000 y 2003 (v´ease Guerrero y Corona, 2018).Cuadro 1Variables utilizadas (1998:01-2020:10)VariableDescripci´onFuenteFrecuenciaADFADF PrimerasNivelesdiferenciasANTADVentas totalesANTADM0.920.01BMV´Indice de precios y cotizaciones de la Bolsa Mexicana de ValoresBanco de M´exicoM0.750.01CONF_MANUFMomento adecuado para invertir de las manufacturasINEGIM0.130.01CONF_SERVMomento adecuado para invertir del sector serviciosINEGIM0.650.02GASOLINASDemanda de combustiblesSecretar´ıa de Energ´ıaM0.530.01IAI´Indice de producci´on industrialINEGIM0.290.01IGAEIndicador Global de la Actividad Econ´omicaINEGIM0.550.01IMSSAsegurados permanentes y eventuales del Seguro SocialInstituto Mexicano del Seguro SocialM0.950.01IPI_EUA´Indice de producci´on industrial de los Estados UnidosBureau of Economic AnalysisM0.130.01ISBSVMIngresos por suministros de bienes y servicios. Comercio al por menorINEGIM0.620.03KInversi´on totalINEGIQ0.01*0.10*MImportaciones totalesINEGIM0.60.01M4Agregado monetario M4Banco de M´exicoM0.990.01PIBProducto Interno BrutoINEGIQ0.01*0.10*PROD_VEHProducci´on de veh´ıculos automotoresINEGIM0.610.01REMESASRemesas familiaresBanco de M´exicoM0.970.01SP_500´Indice Standard & PoorsYahoo! FinancieroM0.990.01T_EMP_MANPersonal ocupado (serie de tenencia)INEGIM0.70.05TCTipo de cambio nominal promedioBanco de M´exicoM0.950.01TDUTasa de desocupaci´on en ´areas urbanasINEGIM0.470.01XExportaciones totalesINEGIM0.620.01Notas: M: Mensual, Q: Trimestral. *Para estas series se utiliz´o la prueba KPSS con constante.Fuente: Elaboraci´on propia.EVALUACI ´ON DE DIFERENTEShttps://doi.org/10.24201/ee.v37i2.432297Asimismo, partimos de 1998 porque cubre en su totalidad un insumomuy importante como lo es el n´umero de asegurados permanentes yeventuales del Seguro Social. Es decir, antes de 1998 esta variable noest´a oficialmente publicada. La totalidad de las variables utilizadasen este trabajo se describen en el cuadro 1.5N´otese que para obtener yPta trav´es del m´etodo heur´ıstico ydel filtro de HP s´olo se requiere el PIB, mientras que para el MFD seemplean las variables mensuales. En nuestro caso, y a diferencia deCorona et al. (2021), las variables no son transformadas para tenerm´axima correlaci´on con el IGAE, no consideramos t´opicos de GoogleTrends y s´olo consideramos los niveles de las series.6 Asimismo, en losmodelos econom´etricos, las series de tiempo se introducen de maneradesestacionalizada.Para el caso de la descomposici´on PT, se considera la variableIMSS7 trimestralizada promediando cada tres meses los valores dedicha variable.Las variables trimestrales, es decir, el PIB y K, setienen para el primer trimestre de 1998 hasta el segundo trimestredel 2020.3.2 Inflaci´on constante, MFD y elasticidades de factores de la pro-ducci´onPara encontrar la tasa de inflaci´on anual constante se computa lavarianza de la inflaci´on anual robusta a la heterocesdasticidad y au-tocorrelaci´on, considerando ±2 desviaciones est´andar alrededor de lamediana. Los resultados se muestran en la gr´afica 1.Se puede apreciar que la inflaci´on ronda entre 2.9 y 5.7, de talforma que si consideramos la tasa de crecimiento media anual delPIB ocurrida cuando se cumple esta restricci´on de inflaci´on, el valores ∆2 = 1.93, mientras que la de todo el periodo de muestra es de∆1 = 2.11.5 La prueba ADF se especific´o con constante. Se presentan los p valores dedichas pruebas.6 No se consideraron, adem´as, las variables de movilidad, ocupaci´on hotelera,confianza del sector empresarial (comercio y construcci´on) e ´ındices de pedidosmanufactureros porque los niveles de estas series presentaron correlaci´on negativarespecto a los niveles del PIB. Asimismo, s´olo se consideraron variables I(1).7 El nombre de la variable es la abreviatura del Instituto Mexicano del SeguroSocial.298ESTUDIOS ECON ´OMICOShttps://doi.org/10.24201/ee.v37i2.432Gr´afica 1Inflaci´on constanteNota: Las l´ıneas punteadas representan ±2 desviaciones est´andar alrededor dela mediana muestral.Fuente: Elaboraci´on propia.En relaci´on con el MFD, las contribuciones estimadas del factorsobre las variables y sus intervalos de confianza al 95%, estimados deacuerdo con lo propuesto por Bai (2003), se muestran en la gr´afica 2.Gr´afica 2Contribuci´on del factor sobre las variablese intervalo de confianza al 95%.Nota: Intervalos de confianza al 95%.Fuente: Elaboraci´on propia.EVALUACI ´ON DE DIFERENTEShttps://doi.org/10.24201/ee.v37i2.432299Se aprecia que el factor se relaciona positivamente con todas lasvariables a excepci´on de la tasa de desocupaci´on, que, aunque se rela-ciona de manera negativa, el intervalo de confianza incluye al cero,por lo que podr´ıa considerarse que esta variable no es significativa.Por otra parte, las variables del factor que m´as contribuyen son elIGAE, el n´umero de asegurados permanentes y eventuales del IMSS,importaciones, exportaciones y las ventas totales de la Asociaci´onNacional de Tiendas de Autoservicio y Departamentales (ANTAD). Encontraparte, las menos relevantes son la confianza del sector empre-sarial en el sector manufacturero y servicios. Vale la pena destacarque el factor din´amico estimado es no estacionario de acuerdo conla prueba ADF usando una constante y una tendencia en la especifi-caci´on de dicha prueba, obteniendo un estad´ıstico de prueba de -0.80que genera un valor p de 0.96. Finalmente, y muy importante, verifi-camos que ˆet ∼ I (0) mediante pruebas PANIC (Bai, 2004), estimandoun estad´ıstico de prueba de 5.21 que otorga un valor p de 0.00, esdecir, se rechaza hip´otesis nula de presencia de ra´ız unitaria.8Para la estimaci´on de las elasticidades del factor trabajo y capi-tal, primero hay que realizar una estimaci´on de los acervos de capitalde M´exico. Para lo anterior, seguimos la metodolog´ıa de Lor´ıa y DeJes´us (2007),9 que considera las siguientes expresiones:KSt = (1 − δ) KSt−1 + It,(13)Adjt = (1 − δ) Adjt−1 + 1,(14)Kt = KSt/Adjtδ,(15)donde KSt es el acervo del capital real, δ es la tasa de depre-ciaci´on, It es la inversi´on total y Adjt es un factor de ajuste. Lor´ıa y8 Esto fue obtenido asumiendo que el n´umero de factores estimados es ˆr = 2,aunque el factor analizado ´unicamente es el primero.El primer factor explicaalrededor del 69% de la variabilidad total, mientras que, con el segundo, se explicael 82%.9 La metodolog´ıa de Loria y De Jes´us (2007) sigue el m´etodo de inventariosperpetuos (MIP) ajustado, el cual resuelve uno de los problemas del MIP aliniciar con valor nulo de capital en la primera observaci´on. Entre otros estudiosque utilizan el MIP se encuentran, por ejemplo, Santaella (1998) y Bergoeninget al. (2002) en los cuales se supone de manera arbitraria que el dato inicial esigual a cero, y es igual a la inversi´on a partir de la segunda observaci´on y solo apartir de la tercera se empieza a acumular el capital estimado a la inversi´on y as´ısucesivamente.300ESTUDIOS ECON ´OMICOShttps://doi.org/10.24201/ee.v37i2.432De Jes´us (2007) afirman que suponer que KSt es cero en la primeraobservaci´on y que al aumentar muy r´apidamente hasta estabilizarse,representa una desventaja emp´ırica muy importante. En consecuen-cia, sugieren realizar dicho ajuste para tener una estimaci´on m´as fiabledel acervo del capital, Kt.Gr´afica 3Estimaciones del acervo de capital para δ = 0.10 (l´ınea s´olida),δ = 0.15 (l´ınea descontinua) y δ = 0.20 (l´ınea punteada)Fuente: Elaboraci´on propia.En nuestro caso, se asume una δj como una secuencia de es-calares que ronda entre el 8 y 20% anual, considerando una longitudde j = 1, . . ., 1000.De esta forma, en cada iteraci´on calculamoslas elasticidades de largo plazo a trav´es de la ecuaci´on de cointe-graci´on de Johansen (1991) normalizada para la variable de inter´es,10yt = β0 + β1lt + β2kt + ut, tomando el logaritmo tanto de los asegu-rados permanentes y eventuales del IMSS trimestral y como de cadaKt,j.11 En la gr´afica 3 se muestran unos ejemplos del comportamientode Kt,j para δ = (0.10, 0.15, 0.20) .10 En cada iteraci´on se estima el rango de la matriz de cointegraci´on de acuerdocon Johansen (1991). La descomposici´on PT se estima ´unicamente cuando 3 >rango (αβ′) > 0 = m. En cualquier otro caso, no se considera la descomposi-ci´on PT. Las estimaciones se realizaron con series de tiempo desestacionalizadas.11 Se asumi´o una estimaci´on restringida tal que ˜Xβ1 + ˜Xβ2 = 1, donde ˜X serefiere a la mediana. Es decir, se considera que la distribuci´on de las elasticidadesEVALUACI ´ON DE DIFERENTEShttps://doi.org/10.24201/ee.v37i2.432301Se puede apreciar que conforme δ aumenta, es decir, la tasa dedepreciaci´on, el acervo de capital disminuye. Esto tiene sentido puestoque conforme pasa el tiempo el capital se hace menos rentable conuna tasa de depreciaci´on m´as alta.La gr´afica 4 muestra las densidades de las elasticidades de largoplazo estimadas para el factor trabajo y capital.Gr´afica 4Densidad de las elasticidades estimadas para el factortrabajo y capital en las ecuaciones de cointegraci´onFuente: Elaboraci´on propia.Podemos apreciar que las elasticidades del factor trabajo rondanentre 0.81 y 0.86, mientras que las del factor capital entre 0.14 y 0.23.Considerando el intervalo de confianza al 95%, tenemos las siguienteselasticidades:β0 = (−1.33, −0.02, 0.57)(16)β1 = (0.74, 0.81, 0.81)(17)β2 = (0.14, 0.19, 0.24)(18)est´e centrada en el caso de rendimientos constantes a escala, similar a como lopropone Acevedo-Fern´andez (2009).302ESTUDIOS ECON ´OMICOShttps://doi.org/10.24201/ee.v37i2.432N´otese que, por construcci´on, la mediana representa los rendi-mientos constantes a escala, mientras el intervalo superior e inferiorson los casos de rendimientos crecientes y decrecientes respectiva-mente. La teor´ıa econ´omica y evidencia emp´ırica previa sugiere resul-tados similares, por ejemplo, Cervantes y Arenas (2004) para el casomexicano obtienen estimaciones de 0.65 y 0.49 para β1 y β2 respecti-vamente, mientas que Onalan y Basegmez (2018) encuentran para laeconom´ıa de los Estados Unidos β0= 0.031, β1=1.097 y β2=0.403; porsu parte, Fraser (2002) obtiene par´ametros β=(-0.191, 0.812, 0.231).Para validar estad´ısticamente el modelo VEC, en cada iteraci´onse estimaron pruebas de autocorrelaci´on serial Portmanteau-Breusch-Godfrey, obteniendo un cuantil al 95% para los valores p de (0.86,0.89, 0.99); es decir, no se rechaza hip´otesis nula de ausencia de au-tocorrelaci´on serial multivariada. Si bien es cierto, esto garantiza,entre otras cosas, estimaciones correctas de los errores est´andar aso-ciados a los coeficientes de las ecuaciones de cointegraci´on, en estetrabajo validamos la significancia con la amplitud de los intervalos delas ecuaciones (15)-(17).3.3 PIB potencialUna vez estimado el factor com´un no estacionario, se normaliza res-pecto al nivel de yt mediante una regresi´on lineal, de tal forma quees ´este el estimador ˆyP,MF Dt. Por otra parte, una vez estimadas laselasticidades del factor trabajo y capital, el estimador ˆyP,P Ttes en-tonces la mediana de las estimaciones de P y =�Py(1), . . ., Py(1000)�obtenidas en cada iteraci´on. Las estimaciones del PIB potencial semuestran en la gr´afica 5.1212 Se presentan los intervalos de confianza al 95% para los casos donde puedenser computados. Para el caso del filtro de HP se sigue la metodolog´ıa de Guerrero(2007), para el MFD se obtienen al computar la varianza otorgada al momento deusar el suavizamiento de Kalman y para la descomposici´on PT, se toman en cuentalos cuantiles 0.025 y 0.975 seg´un el proceso iterativo sugerido en este trabajo.EVALUACI ´ON DE DIFERENTEShttps://doi.org/10.24201/ee.v37i2.432303Gr´afica 5Estimaciones del PIB potencial.Fuente: Elaboraci´on propia.Se puede apreciar que los m´etodos heur´ısticos difieren ´unica-mente del nivel de la tasa media de crecimiento, mientras que losrestantes procedimientos se diferencian de la suavidad alrededor delPIB observado. En este sentido, ˆyP,MH1test´a por encima de ˆyP,MH2t.Por otra parte, es interesante notar que el MFD genera una tendenciam´as ajustada al comportamiento del PIB y esto tiene que ver con quelas variables con las que se estiman los componentes del MFD presen-tan movimientos comunes con la actividad econ´omica, como se˜nalanCorona et al.(2021).Asimismo, el filtro de HP y la descomposi-ci´on PT generan tendencias similares, donde la menos suavizada esla otorgada por ˆyP,P Tt, que conceptualmente representa la trayecto-ria de efectos permanentes sobre yt. Adem´as, enfoc´andonos en losintervalos de confianza, el filtro de HP tiene una mayor incertidum-bre asociada a la estimaci´on mientras que la descomposici´on PT espr´acticamente apenas notoria. Para el MFD n´otese que, aunque laamplitud de los intervalos es mayor respecto a la descomposici´on PT,304ESTUDIOS ECON ´OMICOShttps://doi.org/10.24201/ee.v37i2.432puede considerarse razonable puesto que dicha amplitud representaaproximadamente ±2.5% el valor puntual estimado.Es importante comentar que la principal ventaja de los m´etodosheur´ısticos es la simplicidad que permite otorgar una primera ideadescriptiva del comportamiento del PIB potencial, as´ı como que estepuede ser calculado a partir de 1980.13Asimismo, el filtro de HPpuede ser una estimaci´on estad´ıstica naive, la cual carece de supuestosque pueden garantizar su consistencia te´orica y estad´ıstica. Por otraparte, el MFD es un m´etodo econom´etrico que subyace de una vastacantidad de variables, que puede permitir incluso, ser una herramientaalternativa al Sistema de Indicadores C´ıclicos del INEGI (v´ease Gue-rrero y Corona, 2018). Entre otras ventajas, el MFD permite validarsu consistencia en un sentido estad´ıstico como, por ejemplo, evaluarla significancia de las variables y los intervalos de confianza del factordin´amico.14 Finalmente, la descomposici´on PT es quiz´a la alternativaconceptual, tanto econ´omica como estad´ıstica, m´as recomendada paraestimar el PIB potencial dado que abarca la postura keynesiana yneocl´asica a la vez, al partir de la modelaci´on del PIB a trav´es desus factores de producci´on, y, condicionado a la relaci´on de largoexistente, se estima la combinaci´on lineal que permite desentra˜narsus efectos permanentes, por ende, la estimaci´on del PIB potencial.Para visualizar de una mejor manera las similitudes y las dife-rencias de los componentes de las estimaciones, la gr´afica 6 presentalos componentes la diferencia que existe entre el PIB observado y cadauno de los PIB potenciales estimados, que, para fines pr´acticos, lla-maremos comportamiento c´ıclico estimado, denotado entonces comoˆct.Podemos apreciar de forma visual que todos los estimadorestienen relativamente, movimientos comunes. A simple vista, los m´eto-dos heur´ısticos se relacionan al principio del periodo de muestra; sinembargo, dado que la tasa media de crecimiento de ˆyP,MH1tes m´asgrande, la discrepancia respecto a ˆyP,MH2taumenta. Como lo comen-tamos anteriormente, ˆyP,P Ttest´a muy relacionado a ˆyP,MH1ty ˆyP,HPt.De hecho, se puede observar c´omo, en general, ˆyP,P Ttest´a en mediode las otras dos series. Finalmente, ˆyP,MF Dtes el m´etodo que est´am´as relacionado con la estimaci´on obtenida por el filtro de HP, aunqueapenas cae por debajo del 0 durante la crisis financiera de 2009, estoporque el factor com´un sintetiza las din´amicas exhibidas de todas las13 El INEGI publica resultados del PIB trimestral a partir de 1980.14 Se obtiene para series estandarizadas dado que la estimaci´on de las compo-nentes del MFD se realiza con series centradas en 0 y con varianza 1.EVALUACI ´ON DE DIFERENTEShttps://doi.org/10.24201/ee.v37i2.432305series de tiempo que componen al MFD. Durante el final del periodode muestra, es decir, el 2020, ´epoca de la pandemia de la COVID-19,podemos apreciar que todos los comportamientos c´ıclicos presentanla ca´ıda, es decir, tampoco las variables ex´ogenas utilizadas para lamodelaci´on del factor din´amico o tendencia com´un presentan en con-junto, la magnitud de la ca´ıda que s´ı presenta el PIB.Gr´afica 6ˆct = yt − ˆyPt para cada uno de los m´etodos de estimaci´onFuente: Elaboraci´on propia.Al realizar la comparaci´on con c´alculos recientes de Banco deM´exico, publicados en el Informe Trimestral Enero-Marzo 2021(Ban-xico, 2021), el cual emplea el filtro de HP con correcci´on de colas, seobserva mayor similitud entre las estimaciones con el m´etodo PT quecon el resto de los m´etodos.Para distinguir qu´e t´ecnica pudiese ser la mejor en t´erminos es-tad´ısticos, se propone estimar el vector de cargas de los componentesc´ıclicos dado que estos por definici´on muestran m´as variabilidad alser la diferencia entre el PIB observado y el PIB potencial. Es decir, seestima ˆP para el modelo ct = PFt+wt y se analiza qu´e variable es laque m´as aporta en dicha componente, estimando tambi´en, los interva-los de confianza asint´oticos al 95% (Bai, 2003). Interpretativamente,306ESTUDIOS ECON ´OMICOShttps://doi.org/10.24201/ee.v37i2.432esto nos permite visualizar la importancia de las variables dentro delos movimientos comunes que presentan los ciclos. El cuadro 2 pre-senta estos resultados.Cuadro 2Estimaci´on de la matriz de cargas de los componentes c´ıclicosEstimadorInferiorˆpiSuperiorˆcP,MHt1.031.081.12ˆcP,MHt0.901.061.22ˆcP,HPt0.790.921.05ˆcP,MFDt0.650.841.02ˆcP,PTt1.061.081.10Fuente: Elaboraci´on propia.Se puede apreciar que los ciclos obtenidos por los m´etodos heur´ıs-ticos y por la descomposici´on PT son los que m´as aportan en el fac-tor com´un, el cual explica el 79% de la variabilidad total observada.Sin embargo, n´otese que la amplitud de los intervalos es mayor enlos componentes c´ıclicos obtenidos por los m´etodos heur´ısticos. Eneste sentido, ˆcP,P Ttaporta menor incertidumbre en la estimaci´on delcomponente c´ıclico com´un generado por estos cinco m´etodos de esti-maci´on.Para ejemplificar el comportamiento de este ciclo com´un, ˆF ct , enla gr´afica 7 se presenta la estimaci´on de dicho factor, sombreando laszonas negativas que tienen -al menos- dos trimestres por debajo del0; es decir, cuando el PIB potencial es mayor al PIB observado.Se puede apreciar que el ciclo com´un presenta dos fases en lasque el PIB potencial estuvo por encima del PIB observado, ´estas son:entre el ´ultimo trimestre del 2001 y el ´ultimo trimestre del 2003, as´ıcomo entre el primer trimestre del 2009 y el segundo trimestre del2011. Comparando estos resultados con los obtenidos por Guerrero-de-Lizardi (2020), a lo que el autor llama “se˜nales extra´ıdas seg´unla perspectiva del Banco de M´exico”, podemos ver que se obtienenresultados similares, puesto que la brecha anual es negativa entre el2002 y 2003, as´ı como entre el 2009 y 2011.EVALUACI ´ON DE DIFERENTEShttps://doi.org/10.24201/ee.v37i2.432307Gr´afica 7Ciclo com´un y regiones negativasFuente: Elaboraci´on propia.En consecuencia, dados los resultados estad´ısticos, se consideraapropiado recomendar al estimador ˆyP,P Ttporque es una metodolog´ıaeconom´etrica que permite validar los supuestos que garantizan la con-sistencia de las estimaciones en un sentido estad´ıstico; a la par de quecontribuye en la explicaci´on de los movimientos comunes con respectoa los dem´as componentes c´ıclicos extra´ıdos con el resto de los proce-dimientos.3.4 An´alisis de robustez de las estimacionesEl fen´omeno de la pandemia de la COVID-19 se refleja en el PIB apartir del segundo trimestre de 2020, que en sentido univariado re-presenta un cambio estructural que puede ocasionar (por ejemplo, enmodelos de regresi´on) que los par´ametros estimados cambien radi-calmente antes y despu´es de dicho cambio estructural. Respecto alMFD, Stock y Watson (1989) mencionan que las estimaciones de losfactores a trav´es de componentes principales generan estimacionesrobustas en el sentido que las variables que integran al MFD suelen308ESTUDIOS ECON ´OMICOShttps://doi.org/10.24201/ee.v37i2.432tener movimientos comunes, incluidos cambios estructurales. En estetrabajo, usamos en una primera etapa, la metodolog´ıa de Bai (2004),la cual se basa en componentes principales, refinando la estimaci´ona trav´es del suavizamiento de Kalman. Sin embargo, la descomposi-ci´on PT subyace de ejercicios de cointegraci´on, los cuales pueden sersensibles a la presencia del efecto de la COVID-19. Para verificar larobustez de los resultados, se estimaron los ejercicios econom´etricospara el periodo 2018:T1-2020:T2, computando las distribuciones delas elasticidades del factor trabajo y capital, si es que puede corro-borarse la cointegraci´on. La gr´afica 8 muestra los resultados de lamediana de las distribuciones para ambas elasticidades.Gr´afica 8Mediana de las elasticidades del factor trabajo y capitalFuente: Elaboraci´on propia.Se puede observar que los resultados son robustos en t´erminostales que -aunque se han supuesto rendimientos constantes a escala-las elasticidades computadas est´an entre 0.80 y 0.88 para el factortrabajo y 0.12 y 0.20 para el factor capital; es decir, aunque se ob-serva cierta tendencia, no presentan variaciones dr´asticas atribuiblesa la COVID-19 que pudiesen suponer problemas en la especificaci´onEVALUACI ´ON DE DIFERENTEShttps://doi.org/10.24201/ee.v37i2.432309del modelo. En todos los casos, se puede denotar una distribuci´on bi-modal para el factor trabajo y colas inclinadas hacia la izquierda parael factor capital, similar a lo observado en la gr´afica 4. Es decir, dadoque se pudieron computar dichas distribuciones, se pudo verificar lacointegraci´on en todo el periodo de tiempo.15Finalmente, al replicar todos los ejercicios para este periodo detiempo, se pueden estimar los vectores de cargas de los componentesc´ıclicos en cada trimestre, evaluando la evoluci´on de la importanciaatribuible a cada m´etodo en el ciclo com´un. Los resultados de lasestimaciones puntuales se muestran en la gr´afica 9.Gr´afica 9Cargas atribuibles a cada m´etodo en el ciclo com´unFuente: Elaboraci´on propia.En todos los casos, los m´etodos heur´ısticos y la descomposici´onPT generan resultados similares. Sin embargo, a partir del segundotrimestre del 2019 aumenta la importancia del m´etodo heur´ıstico 1 yde la descomposici´on PT, mientras que baja el filtro de HP y el m´etodoheur´ıstico 2. N´otese que el MFD explica mayor variabilidad a partir15 Estos resultados est´an disponibles bajo petici´on.310ESTUDIOS ECON ´OMICOShttps://doi.org/10.24201/ee.v37i2.432del primer trimestre del 2020, pero en todos los casos, es el que tienemenor importancia en la explicaci´on del ciclo com´un.En este sentido, puede concluirse que las estimaciones presen-tadas aqu´ı son robustas ante la presencia del posible cambio estruc-tural aludido a la pandemia de la COVID-19.4. ConclusionesLa evoluci´on del PIB depende en gran parte de las pol´ıticas econ´omi-cas, principalmente las fiscales y monetarias. Adem´as, como argu-menta Acevedo-Fern´andez (2009), su dise˜no correcto e implementa-ci´on eficaz requieren de la utilizaci´on de diversas herramientas quepermitan un diagn´ostico puntual y minucioso sobre la situaci´on econ´o-mica estructural y coyuntural del pa´ıs.Dada la importancia en materia de pol´ıtica econ´omica con relaci´ona obtener estimaciones adecuadas del PIB potencial, en este trabajose utilizaron cuatro procedimientos para estimarlo y se realiz´o la apli-caci´on emp´ırica para M´exico con datos del periodo 1998:T1-2020:T2.Dos de estos procedimientos fueron propuestos en este trabajo y sub-yacen del uso del MFD y la descomposici´on PT de Gonzalo y Granger(1995), los cuales fueron comparados con dos metodolog´ıas tradi-cionales, espec´ıficamente los m´etodos heur´ısticos y el filtro de HP.Las conclusiones principales son que los dos m´etodos propuestostienen ventajas sobre las t´ecnicas tradicionales, en el sentido que sub-yacen de metodolog´ıas econom´etricas que tienen como objetivo esti-mar tendencias comunes o componentes permanentes, adem´as de quees posible validar los resultados en un sentido econom´etrico. A la parde esto, la evidencia emp´ırica corrobora que la descomposici´on PTgenera los mejores resultados en los siguientes sentidos: 1) su compo-nente c´ıclico tiene una contribuci´on importante en la estimaci´on delcomponente c´ıclico com´un y, 2) las estimaciones de las elasticidadesdel factor trabajo y capital son robustas a trav´es del tiempo, es decir,no hay diferencias abruptas en dichas estimaciones que puedan seraludidas al posible cambio estructural atribuible a la COVID-19.Francisco Corona: franciscoj.corona@inegi.org.mx; Pedro Orraca: porraca@colef.mx;Jes´us L´opez-P´erez: jesus.lopezp@inegi.org.mxEVALUACI ´ON DE DIFERENTEShttps://doi.org/10.24201/ee.v37i2.432311ReferenciasAcevedo-Fern´andez, E. 2009. PIB potencial y productividad total de los factores:recesiones y expansiones en M´exico, Econom´ıa Mexicana Nueva ´Epoca, 18(2):175-219.Bai, J. 2003. Inferential theory for factor models of large dimensions, Economet-rica, 71(1): 135-171.Bai, J. 2004. Estimating cross-section common stochastic trends in nonstationarypanel data, Journal of Econometrics, 122(1): 137-183.Banxico. 2021. Informe trimestral enero-marzo 2021, https://www.banxico.org.mx/publicaciones-y-prensa/informes-trimestrales/%7B49D9C039-CE93-FC5A-59A6-DFF7579FDB26%7D.pdf.Bergoeing, R., P. Kehoe y R. Soto. 2002. A decade lost and found: Mexico andChile in the 1980s, Review of Economic Dynamics, 5(1): 166-205.Carrillo, J., R. Elizondo, C.A. Rodr´ıguez-P´erez y J. Rold´an-Pe˜na. 2018. Whatdetermines the neutral rate of interest in an emerging economy?, Documentode Trabajo No. 2018-22, Banco de M´exico.Cervantes, M. y E.A. Arenas. 2004. Estimate of the potential product of M´ex-ico: An analysis of time series with the Hodrick Prescott filter, Revista deEstad´ıstica, Econometr´ıa y Finanzas Aplicadas, 2(2): 85-102.Corona, F. y P. Orraca. 2019. Remittances in Mexico and their unobserved com-ponents, The Journal of International Trade and Economic Development,28(8): 1047-1066.Corona, F., P. Poncela y E. Ruiz. 2017. Determining the number of factors afterstationary univariate transformations, Empirical Economics, 53(1):351-372.Corona, F., P. Poncela y E. Ruiz.2020.Estimating non-stationary commonfactors: Implications for risk sharing, Computational Economics, 55(1): 37-60.Corona, F., G. Gonz´alez-Far´ıas y J. L´opez-P´erez. 2021. A nowcasting approachto generate timely estimates of Mexican economic activity: An application tothe period of COVID-19, ArXiv preprint, https://arxiv.org/abs/2101.10383.Cuadra, G. 2008. Hechos estilizados del ciclo econ´omico en M´exico, Documentode Trabajo No. 2008-14, Banco de M´exico.De Masi, P.R. 1997. IMF estimates of potential output: Theory and practice,IMF Working Paper No. 97/177.Doz, C., D. Giannone y L. Reichlin. 2011. A two-step estimator for large ap-proximate dynamic factor models based on Kalman filtering, Journal ofEconometrics, 164(1): 188-205.Elizondo, R. 2019. Estimaciones del PIB mensual en M´exico basadas en el IGAE,Estudios Econ´omicos, 34(2): 197-241.Faal, E. 2005. GDP growth, potential output, and output gaps in Mexico, IMFWorking Paper No. 05/93.Fraser, I. 2002. The Cobb-Douglas production function: An antipodean defense?,Economic Issues, 7(1): 39-58.G´alvez-Soriano, O. 2020. Nowcasting Mexico’s quarterly GDP using factor mod-els and bridge equations, Estudios Econ´omicos, 35(2): 213-265.312ESTUDIOS ECON ´OMICOShttps://doi.org/10.24201/ee.v37i2.432Gavin, W. 2012. What is potential GDP and why does it matter?, EconomicSynopses, 11: 1-2.Gonzalo, J. y C. Granger. 1995. Estimation of common long-memory compo-nents in cointegrated systems, Journal of Business and Economic Statistics,13(1): 27-35.Guerrero, V.M. 2007. Time series smoothing by penalized least squares, Statisticsand Probability Letters, 77(12): 1225-1234.Guerrero, V.M. 2004.Nota sobre la estimaci´on del PIB mensual de M´exico,Estad´ıstica, 56(166): 35-60.Guerrero, V.M. y F.D.J. Corona. 2018. Actualizaci´on del sistema de indicadoresc´ıclicos de M´exico.Realidad, datos y espacio, Revista Internacional deEstad´ıstica y Geograf´ıa, 9(3): 29-64.Guerrero-de-Lizardi, C. 2020. PIB potencial y ciclo econ´omico en M´exico 1921-2019: una perspectiva desde la “medici´on en econom´ıa”, Revista Mexicanade Econom´ıa y Finanzas, 15(2): 185-204.Heath, J. 2012.Lo que Indican los Indicadores.C´omo Utilizar la Informa-ci´on Estad´ıstica para Entender la Realidad Econ´omica de M´exico, M´exico,Instituto Nacional de Estad´ıstica y Geograf´ıa.Hoffman, A. y H. Tapia. 2003. Potential Output in Latin America: A StandardApproach for the 1950-2002 Period, Santiago, Comisi´on Econ´omica paraAm´erica Latina y el Caribe.Islas, A. y V. Guerrero. 2019. Joint smoothing of GDP and unemployment witha bivariate HP filter, Estudios Econ´omicos, 34(1): 3-24.Johansen, S. 1991. Estimation and hypothesis testing of cointegration vectors inGaussian vector autoregressive models, Econometrica, 59(6): 1551-1580.Kichian, M. 1999. Measuring potential output within a state-space framework,Staff Working Paper No. 1999-9, Bank of Canada.Kuttner, K.N. 1994. Estimating potential output as a latent variable, Journalof Business and Economic Statistics, 12(3): 361-368.Laubach, T. y J.C. Williams.2003.Measuring the natural rate of interest,Review of Economics and Statistics, 85(4): 1063-1070.Lor´ıa, E. y L. de Jes´us. 2007. Los acervos de capital de M´exico: una estimaci´on,1980. I-2004.IV, El Trimestre Econ´omico, 74(2): 475-485.Lor´ıa, E., M.G. Ramos y L. De Jes´us. 2008. Producto potencial y ciclos econ´omi-cos en M´exico, 1980.1-2006.4, Estudios Econ´omicos, 23(1): 25-47.Mantescu, D. y D.T. Lazar. 2014. Estimation of potential GDP and output gap.Comparative Perspective, Amfiteatru Economic Journal, 16(37): 951-964.McElroy, T. 2008. Exact formulas for the Hodrick-Prescott filter, The Econo-metrics Journal, 11(1): 209-217.M´endez Sayago, J.A., J.M. M´endez Sayago y H.A. Hern´andez Escolar.2013.Productividad total de los factores, cambio t´ecnico, eficiencia t´ecnica y PIBpotencial en Latinoam´erica, Semestre Econ´omico, 16(34): 65-91.Newey, W.K. y K.D. West. 1987. Hypothesis testing with efficient method ofmoments estimation, International Economic Review, 28(3): 777-787.Noriega A. y C. Rodr´ıguez-P´erez. 2012. Estacionariedad, cambios estructuralesy crecimiento econ´omico en M´exico (1895-2008), El Trimestre Econ´omico,79(2): 333-378.EVALUACI ´ON DE DIFERENTEShttps://doi.org/10.24201/ee.v37i2.432313OCDE. 2012. OECD system of composite leading indicators, https://www.oecd.org/sdd/leading-indicators/41629509.pdf.Onalan, O. y H. Basegmez, H. 2018. Estimation of economic growth using theGrey Cobb-Douglas production function: Application for US economy, Jour-nal of Business, Economics and Finance, 7(2): 178-190.Ozbek, L. y U. Ozlale. 2005. Employing the extended Kalman filter in measuringthe output gap, Journal of Economics Dynamics and Control, 29(9): 1611-1622.Phillips, P.C. y Z. Shi. 2021. Boosting: Why you can use the HP filter, Interna-tional Economic Review, 62(2): 521-570.Santaella, J. 1998.Economic Growth in Mexico: Searching for Clues to itsSlowdown, Washington D.C., Inter-American Development Bank.Scacciavillani, F. y P. Swagel. 2002. Measures of potential output: An applica-tion to Israel, Applied Economics, 34(8): 945-957.Stock, J.H. y M.W. Watson. 1989. New indexes of coincident and leading eco-nomic indicators, en O.J. Blanchard y S. Fischer (eds.), NBER Macroeco-nomics Annual 4, Massachusetts, MIT Press.Theoduloz, T. 2010. Testing the gap: An application for Uruguay, Cuaderno deEconom´ıa, 4: 97-119.Ventosa-Santaul`aria, D., L.G. Hern´andez-Rom´an y A. Villag´omez-Amezcua. 2021.Recessions and potential GDP: The case of Mexico, Bulletin of EconomicResearch, 73(2): 179-195.View publication stats | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/366983963Leggett's plasma resonances and two-gap structures in the CVCs of MgB$_2$break junctions $-$ a direct evidence for a two-gap superconductivity inMgB$_2Preprint · January 2023DOI: 10.48550/arXiv.2301.02929CITATIONS0READS1916 authors, including:Some of the authors of this publication are also working on these related projects:Mobile and satellite communications View projectReconstruction of the laboratory building for the Ginzburg Center for High Temberature Superconductivity and Quantum Materials View projectYaroslav PonomarevLomonosov Moscow State University93 PUBLICATIONS 877 CITATIONS SEE PROFILESvetoslav A. KuzmichevLomonosov Moscow State University64 PUBLICATIONS 791 CITATIONS SEE PROFILEMatthias HeinTechnische Universität Ilmenau452 PUBLICATIONS 4,068 CITATIONS SEE PROFILEL.G. SevastyanovaLomonosov Moscow State University9 PUBLICATIONS 165 CITATIONS SEE PROFILEAll content following this page was uploaded by Svetoslav A. Kuzmichev on 13 January 2023.The user has requested enhancement of the downloaded file.Leggett's plasma resonances and two-gap structures in the CVCs of MgB2 break junctions – a direct evidence for a two-gap superconductivity in MgB2 Ya.G. Ponomarev1, S.A. Kuzmichev1, M.G. Mikheev1, M.V. Sudakova1, S.N. Tchesnokov1, N.Z. Timergaleev1, A.V. Yarigin1, M.A. Hein2, G. Müller2, H. Piel2, B.M. Bulychev3, K.P. Burdina3, V.K. Gentchel3, L.G. Sevastyanova3, S.I. Krasnosvobodtsev4, A.V. Varlashkin4 1 M.V. Lomonosov Moscow State University, Faculty of Physics, 119991 Moscow, Russia 2 Bergische Universität Wuppertal, Fachbereich Physik, D-42097 Wuppertal, Germany 3 M.V. Lomonosov Moscow State University, Faculty of Chemistry, 119991 Moscow, Russia 4 P.N. Lebedev Physical Institute, RAS, 119991 Moscow, Russia Originally published in the conference proceedings book: Bulletin of the V. Tarasov Center of Chemotronics of Glass No.2, 139 (2002) ISBN 5-7237-0388-9, Mendeleev University of Chemical Technology of Russia, Moscow PREFACE On the occasion of 20th anniversary of the first experimental observation of the Leggett collective plasma oscillation in MgB2 Before proceeding to the electron copy of the original publication from December 2002 [1] that was never available on-line, I want to sketch a general problem of Leggett collective plasma oscillation and make several remarks both about the nature of phenomena that was firstly observed by Prof. Yaroslav Georgievich Ponomarev in the spectra of tunneling contacts and published in the Mendeleev University bulletin [1] and on the history of our very first publications. Superconductivity in MgB2 was discovered occasionally in the end of 2000 [2]. Despite the strong boron isotope effect (observed in [3]) clearly points to the classical phonon nature of the pairing mechanism in MgB2, it becomes the first-ever-known two-gap superconductor (SC), which means that two types of the Cooper pairs having the distinct coupling energies (21 and 22) are developed in the SC state below Tc. These condensates are not totally independent: they weakly interact through interband coupling in the momentum space. This is somehow similar to the proximity effect between two SC in real space, but in the former case they induced their intrinsic properties to each other in every point of crystal in the momentum space (k-space). As a result, the SC order parameter of one of the condensates (j) being dependent both from his eigen properties (intraband coupling constant jj), as well as from the strength of the crossband coupling 12 or 21 (see works by Moskalenko [4] and Suhl et al. (so-called SMW-model) [5] done independently back in 1959). Even been initially in the weak-coupling limit of BCS theory (e.g. 11, 22 < 0.25 and having characteristic ratios 2eigen(0) / kBTc = 3.53), the variation of the coupling potential Vij in the intraband and crossband channels (Vintra Vinter) leads to the 21(0) / kBTc > 3.53 > 22(0) / kBTc deviation (when 1 is the large SC gap) in case of an extension of the BCS theory for two bands (so-called two-band BCS). Why this compound having the simple chemical formula MgB2 demonstrates the variety of sophisticated physics? One of the reasons is in its layered crystal structure and the complexity of the Fermi surfaces (at least a couple of 2D-hole barrels which are nearly orthogonal with two 3D electron and hole constructions [6–8]). The latter is in the contrast with the conventional SC materials (so-called BCS superconductors) having more or less isotropic crystal structure, conductivity and 3D-electron Fermi surfaces. In occasion of the SC state, Cooper pairs with the same properties (at least coupling energy 2) are developed at any and all conductive bands of the classical SC due to the crossband (interband) mixing of momenta. While this classical phenomenon being one of the important consequence of the Philip W. Anderson theorem, the magnesium diboride breaks its concept down in case of the 2D Fermi surfaces, crystal structure anisotropy and weak interband interaction. More detailed formulation one can find in [9]. Anthony J. Leggett (you can see several intriguing notes on his biography and “Reflections on the past, present and future of condensed matter physics” in [10]) in his pioneer work [11] predicted that collective oscillation of a superconducting plasma, which are caused by small fluctuations of the phase difference between two superconducting condensates, develop in two-gap (and at least two-band) superconductors. Charge carriers can flow from one band another band creating a crossband AC current in the k-space having some characteristic frequency L(k). In his Nobel lecture [12] Leggett calls this type of the collective excitations as “a sort of internal Josephson effect”, since they are intrinsic to the superconductor, and think that he was inspired by the “P.W. Anderson’s elegant formulation of the theory of superfluidity in 4He in terms of conjugate number and phase variables”. He also remembered that in 1966 his theory “sank more or less without trace, in part because by the time it was published it had already become clear that the experimental evidence for the existence of two-band superconductors in nature was dubious” [12]. The main result of Anthony Leggett’s theory [11] is that the square of the oscillation frequency L(k) is determined by two terms: a sound-like (gapless) in-phase phonon mode that depends on k, and a massless term that slightly depends on large k and gives finite frequency 0 in the k 0 limit (here k – is the wave-vector). For this reason, L(0) = 0 may be called the out-of-phase exciton-type mode. It was shown that 0 does not directly depend on the Coulomb interaction and could be obtained even for a system of neutral particles [11]. Finally, 2122101211221221 + 4(0)(0) (1) and valid in the limits of (a) T 0 strictly, (b) for small wave-vectors k, (c) in the low-energy limit (let say BCS weak-coupling constants ij < 0.25, while theoretical estimations [6] give 11 1, 22 0.3), (d) resulting Leggett plasma frequency 0 must correspond to the in-gap energies 0 22 < 21, so as not to be strongly damped by the quasiparticle continuum. This result was re-derived by Sharapov et al. [13] specially for MgB2. The numerical estimation made in [13] demonstrate the energy range 22(0) < 0 < 21(0) for the 0, contrary do the clause (d) limitations. Also note that a value of the exciton-type mode estimated for MgB2 in [13] is 0 1(0)+2(0). The latter raised a question on the possibility of the experimental observation of the Leggett mode, since it should be seriously damped. The limitation of 0 by the smallest SC gap value 22 (so-called softening of the Leggett mode or anticrossing with the gap edge) become a point of the theoretical discussions, for example, see Eq. 2 and Fig. 3 in the work by Karakozov et al. [14]; see also Figs. 3,4 from [15]. In the latter paper Klimin et al. argue that “The low frequency expansion thus becomes inapplicable when the Leggett mode frequency approaches the range close to the pair-breaking continuum edge.” Slightly above this sentence it has been written: “[it] does not capture the interplay of the Leggett collective mode with the pair-breaking continuum edge and hence crosses the value = 22 without any feature.” [15]. The same problem has been addressed in nice theoretical exercises of Arimitsu [16] (see Fig.1). What is the most important result of [11] for the experiment is the linear dependence of the 02 on the 1(0)2(0) product (in the low-energy limit, and T 0) given by Eq.(1). The same direct scaling was obtained in the number of theoretical studies [14,17–19]. The direct proportionality 02 ~ 1(0)2(0) can be checked by the gaps j variation with doping in (Mg,Al)B2 and Mg(B,C)2 systems. From the other side, this verification can help to distinguish between 12 Leggett’s and threshold cross-gap (1+2)2 dependencies of the 02. The story of the experimental discovery of Leggett collective mode started early in 2001, while the author of these notes began to work on his PhD thesis in Lomonosov MSU under the supervision of Prof. Ya.G. Ponomarev. His tunneling effects laboratory was developed in 1986 (as a consequence of the discovery of high-temperature SC cuprates). Ya.G. Ponomarev had extened classical “break-junction” tunneling setup of Moreland and Ekin [20] to be used with layered single crystals and realized the mechanically controlled planar “break-junction” (MCP-BJ) technique to produce S-c-S (S – bulk superconductor, c – constriction) contacts in ab-plane [21,22] and bulk natural arrays S-c-S-c-...-c-S. MCP-BJ technique should be used to study namely layered superconductors materials, see also our brief review [23]. Sadly, Prof. Ponomarev passed away in December 2015 after a severe and prolonged illness. Already in the January of 2001 we get samples of a newly discovered superconductor MgB2 made in the group of Prof. B.M. Bulychev from the Chemical faculty of Lomonosov Moscow State University (MSU). This occasion resulted in a change of the postgraduate work plan and its aim for the author. Our laboratory of tunneling effects has gone deep into the study of magnesium diboride electron properties. No one could guess at the time that we were dealing with the first two-gap (or two component) SC! Tunneling features of rather large amplitude at low bias region (caused by the small -gap) annoyingly entered the dI(V)/dV-curves, and for several months we tried to get rid of them, so that they did not “spoil” our spectra. In the beginning of the summer we have got new series of MgB2 samples from the Chemical faculty of MSU made in high-pressure chamber by Bulychev, Burdina and Sevastyanova with the different level of a structural disorder, as well as samples with the special made admixture of Mg-oxide (up to ~10%) produced by the magnesium vaporization method by Krasnosvobodtsev and Varlashkin (Lebedev Physical Institute, Russian Academy of Sciences). The latter samples had unprecedented properties: a width of the resistive transition to the SC state was as small as T 0.2 K (i.e. ~0.5% of Tc), and, in addition, they demonstrated unusually large -gap values and Tc's that reached 39–41 K (compare it with the standard maximum Tc 38.5 K)! But that is another interesting topic not related to the Leggett mode. During summer of 2002, in the dynamic conductance spectra of planar tunnel contacts (in both regimes, low-transparent and low-capacity Josephson SIS-contacts, and high-transparent semiballistic or diffusive Andreev SNS-contacts with incoherent transport and weak inelastic scattering [24–26]) based on MgB2, as well as MgB2 + MgO samples, Ya.G. Ponomarev discovered some reproducible additional fine structure corresponding to (a) the resonant excitation of some boson mode by the AC Josephson current in the range of energies that corresponds to small SC gap 2 for SIS-contacts (with the current deficiency), and (b) to the excess loss of energy (due to the multiple boson emission) by the so-called Andreev carriers (electrons involved in multiple Andreev reflections (MAR) [24–26]) in the bands with large gap 2 for the high-transparent SnS-contacts (with the excess current and so-called “foot” structure at low bias [24–26]). It is interesting to compare this phenomenon observed in MgB2 with the single or multiple spin-exciton resonant emission during MAR in SnS-contacts based on Fe-based SC of 1111 family (on the issue of the boson-mode) observed by us [27–29], see also the scheme of the emission process in Fig. 3 of [29]. Yaroslav Georgievich Ponomarev told us to search for the special looking fine structure at dI(V)/dV-spectra for observing its reproducibility in our tunneling break-junctions (note, only bulk properties or effects are reproducible in randomly shaped break-junctions!) and for discovering its temperature dependence. He was the first man, who compares the characteristic energy of both additional peculiarities (fine structures) in Josephson and Andreev transport regimes (SIS- vs. SNS-contacts data, as well as data for the corresponding SISIS and SNSNS arrays), ties it together and understands the same Leggett mode nature of the both effects. During 2002 the reproducibility of these effects was observed a dozen times and more or less verified for MgB2 having the largest Tc 35–40 K. For sure, the nature of the boson resonances found by Prof. Ponomarev required clarification, and we had to check it out. One of the most probable situations (both for transport and optical measurements) could be indirect tunneling of quasiparticles from the top of a valence band #1 to the bottom of a conductive band #2 (in this case they definitely change their band due to the inelastic process during tunneling). This should give something like threshold crossband excitation energy (+) ~ 8–11 meV value for MgB2 with a critical Tc ~ 34–38 K. Contrary to these expectations, we reproducibly observed the half of these values (4–5 meV) on the one hand, and on the other hand, the realization of the indirect (cross-gap) tunneling in our planar ScS contacts would definitely produce the (large amplitude) fundamental gap structure at eV = +, but not the additional fine-structure, as we have observed. Since that, we concluded in the Leggett mode nature of the resonances observed, and Prof. Ponomarev decided to make the experimental results public. The first results were published in the very end of 2002 in the "Bulletin of the Mendeleev University of Chemistry and Technology", see Ya.G. Ponomarev et al. [1]. The conference proceedings makes possible to quickly publish the results, which played a positive role, since the next paper on this issue was sent to an editorial office of a high-impact physical journal during the spring of 2003 [30], and its publication was continuously postponed by referees. Finally, in September 2003, the manuscript was readdressed to the editorial office of “Solid State Communications”, and was immediately accepted as a “hot topic publication”. It become available online in October 2003, but physically appears just in the 2nd issue in January 2004 [31], thus, formally, this led to the loss of two years, since the experimental discovery of the phenomenon predicted back in 1966. Note that the “arXiv:cond-mat” version [30] has color figures vs. black-and-white graphs in Solid State Communications [31]. It was very gratifying that Anthony J. Leggett referred to this work [31] in his Nobel lecture [12]. Aside from the pleasure to experimentally discover some new phenomena, we remained a little skeptical and curious, could this experimental resonant energy be driven something else than Leggett collective excitation? In result, we checked, how does this resonant energy L(0) vary with aluminum doping of Mg1−xAlxB2 and, correspondingly, its Tc? Since we can measure both bulk SC gaps (directly at T 0), thus, we checked the linear relationship between (momentum independent part of) the Leggett collective excitation energy squared and a product of the experimental values of the -gap and the -gap: 02 ~ (0)(0), according to the Leggett's equation (1) in the wide range of critical temperatures 10 K < Tc < 40.5 K. The result of this important experimental verification showed 02 (at T = 4.2 K << Tc) and was firstly published in Fig. 6 of Ya.G. Ponomarev, et al., "Leggett's mode in Mg1−xAlxB2" [32], and several years later in Fig. 2 of [14], as well as in paragraph 4.2 of [33]. Unfortunately, so far, we can not find any experimental work done by optical methods, in which this energy 0 would be measured by the Raman response on doped MgB2 together with SC gaps j and Tc variation. This issue is still waiting to be checked by optical spectroscopy. Subjecting self-criticism to the work of Ponomarev's laboratory, in which I was fortunate enough to participate, I need to mention such a shortcoming as the lack of discussion on the Andreev bound states (ABS) influence on our SNS-contact spectra. The development of ABSs in “long” SNS junctions [25, 34–38] has generally the same physical origin as the quantum size effect in (normal) metallic films as a result of the superposition of incident and reflected electron waves [39], as well as the Tomasch size effect in SIS/N tunneling structures [40], in which low energy carriers (|| ) participate in the (single) Andreev reflection at the S/N-interface of the structure, reversing, among other parameters, the sign of their charge. This leads to interference of the incident electron and reflected hole waves in ‘S’ at a distance of the order of mean-free-path, resulting in a series of peculiarities formation in the local electron density of states (DOS). In case of SNS contact low energy carriers are involved in MAR process, being the ground state of this tunneling system, yet at eV 0 bias voltage. They produce the single, several or even a comb of the nearly equidistant ABS inside the SC gap (i.e. in the “forbidden” range of energies), depending on a ratio of a metal layer width d to the SC coherence length 0 (see [41] for some details). Electrons and holes are prohibited from entering the energy range within the SC gap inside the SC, but not into the normal metal layer ‘N’. As the ratio d /0 is increase, more and more ABS appear inside the gap region, producing new maxima of local DOS, until a bunch of Andreev levels merge into a zone and the influence of the proximity of bulk ‘S’ to thin ‘N’ will stop. According to the theories of MAR effect [24–26] the most energetic is the first Andreev minima in dI(V)/dV-spectra of SNS contact (the so-called fundamental harmonic) that is biased at eV1 = 2. Consequently, any of the in-gap features, including ABS, have to appear at eV* < 2 [25]. Contrary to this, we have repeatedly observed extra features at energies large than 2, for example, see minima marked as “m=1”, “m=2”, “m=3” in Fig. 6 of [1,30], or the same in Fig. 5 of [31] (note that label “nL=1” points to 2 fundamental minima). These additional minima arise from the phenomena of the multiple boson emission by the Andreev carriers. The position (bias) of the m = 1,2,3… fine structure defined by (or from the point of view of the experimentalist, define) the trivial expression: eV1,m = 2 + m0 that demonstrates definitely overgap energies and leaving no chance to be originated from the comb of ABS. We thank P.I. Arseev and N.K. Fedorov for several short but fruitful discussions during 2010–2013, as well as A.V. Galaktionov for advices on the problem of ABS. S.A. Kuzmichev, December 2022 References 1. Ya.G. Ponomarev, S.A. Kuzmichev, M.G. Mikheev, et al., Bulletin of the V. Tarasov Center of Chemotronics of Glass No.2, 139 (2002). 2. J. Nagamatsu, N. Nakagawa, T. Muranaka, et al., Nature 410, 63 (2001). [DOI] 3. S.L. Bud’ko, G. Lapertot, C. Petrovic, et al., Phys. Rev. Lett. 86, 1877 (2001). [DOI] 4. V.A. Moskalenko, Fiz. Met. Metalloved. 4, 503 (1959); Sov. Phys. Usp. 17, 450 (1974). [link] 5. H. Suhl, B.T. Matthias, and L.R. Walker, Phys. Rev. Lett. 3, 552 (1959). [DOI] 6. A.Y. Liu, I.I. Mazin, J. Kortus, Phys. Rev. Lett. 87, 087005 (2001). [DOI] 7. Y. Kong, O. V. Dolgov, O. Jepsen, and O. K. Andersen, Phys. Rev. B 64, 020501 (2001). [DOI] 8. H. J. Choi, D. Roundy, H. Sun, et al., Nature 418, 758 (2002). [DOI] 9. I. I. Mazin, O. K. Andersen, O. Jepsen, et al., Phys. Rev. Lett. 89, 107002 (2002). [DOI] 10. Ping Ao, Science Bulletin 63, 1017 (2018) [DOI]; A.J. Leggett, Science Bulletin 63, 1019 (2018). [DOI] 11. A.J. Leggett, Prog. Theor. Phys. 36, 901 (1966). [DOI] 12. A.J. Leggett, Reviews of Modern Physics 76, 999 (2004). [DOI] 13. S.G. Sharapov, V.P. Gusynin, and H. Beck, Eur. Phys. J. B 30, 45 (2002). [DOI] 14. A.E. Karakozov, E.G. Maksimov, Ya.G. Ponomarev. JETP Letters 91, 24 (2010). [DOI] 15. S.N. Klimin, H. Kurkjian, J. Tempere, New J. Phys. 21, 113043 (2019). [DOI] 16. K. Arimitsu, arXiv: cond-mat/1812.05817 (2018). [link] 17. E.G. Maksimov, A.E. Karakozov, B.P. Gorshunov, et al. J. Exp. Theor. Phys. 115, 252 (2012). [DOI] 18. D.F. Agterberg, E. Demler, and B. Janko. Phys. Rev. B 66, 214507 (2002). [DOI] 19. A. Anishchanka, A.F. Volkov, K.B. Efetov. Phys. Rev. B 76, 104504 (2007). [DOI] 20. J. Moreland and J. W. Ekin, Appl. Phys. Lett. 47, 175 (1985). [DOI] 21. B.A. Aminov, A.I. Akimov, N.B. Brandt, et al., Sov. Low Temp. Phys. 15, 689 (1989). [link] 22. Ya.G. Ponomarev, N.B. Brandt, C.S. Khi, et al., Phys. Rev. B 52, 1352 (1995) [DOI]; Ya.G. Ponomarev, B.A. Aminov, M.A. Hein, et al., Physica C 243, 167 (1995). [DOI] 23. S.A. Kuzmichev, T.E. Kuzmicheva, Low Temp. Phys. 42, 1008 (2016). [review] 24. M. Octavio, M. Tinkham, G. E. Blonder, and T. M. Klapwijk, Phys. Rev. B 27, 6739 (1983). [DOI] 25. R. Kümmel, U. Gunsenheimer, and R. Nicolsky, Phys. Rev. B 42, 3992 (1990). [DOI] 26. U. Gunsenheimer, A.D. Zaikin, Phys. Rev. B 50, 6317 (1994). [DOI] 27. S.A. Kuzmichev, T.E. Kuzmicheva, N.D. Zhigadlo, Europhys. Lett. 119, 17007 (2017). [DOI] 28. T.E. Kuzmicheva, S.A. Kuzmichev, N.D. Zhigadlo, Phys. Rev. B 100, 144504 (2019). [DOI] 29. M.M. Korshunov, S.A. Kuzmichev, T.E. Kuzmicheva, Materials 15, 6120 (2022). [DOI] 30. Ya.G. Ponomarev, S.A. Kuzmichev, M.G. Mikheev, et al., arXiv: cond-mat/0303640 (2003). [link] 31. Ya.G. Ponomarev, S.A. Kuzmichev, M.G. Mikheev, et al., Solid State Commun. 129, 85 (2004). [DOI] 32. Ya.G. Ponomarev, S.A. Kuzmichev, M.G. Mikheev, et al. JETP Letters 85, 46 (2007). [DOI] 33. S.A. Kuzmichev, PhD thesis, Physical Faculty, M.V. Lomonosov Moscow State University, (2010). 34. I. O. Kulik, Sov. Phys. JETP 30, 944 (1970). [link] 35. C. W. J. Beenakker, Phys. Rev. Lett. 67, 3836 (1991) [DOI]; Erratum, Phys. Rev. Lett. 68, 1442 (1992). 36. Transport Phenomena in Mesoscopic Systems, edited by H. Fukuyama and T. Ando (Springer, Berlin, 1992). 37. A.V. Galaktionov and A.D. Zaikin, Phys. Rev B 65, 184507 (2002). [DOI] 38. S. Pinon, V. Kaladzhyan, and C. Bena, Phys. Rev B 101, 205136 (2020). [DOI] 39. I.M. Lifshits, A.M. Kosevich, Soviet Physics. Doklady 91, 795 (1953). 40. W.A. Tomasch, Tunneling Phenomena in Solids (Eds E. Burstein, S. Lundqvist) (New York: Plenum Press, 1969) p. 315. 41. In case T 0 and zero phase difference between SC banks of the contact, the 2nd ABS enters in-gap region when d /0 grows to 2; when d /0 > 15 one may approximate the distance between neighboring ABS as EABS 0 2 / ( + d /0). The electron copy of paper [1] is presented below. Selectable and updated REFERENCES of the original work 1. Ginzburg V.L. Usp. Fiz. Nauk 170 (2000) 619 [Phys. Usp. 43 (2000) 573] 2. Maksimov E.G. Usp. Fiz. Nauk 170 (2000) 1033 [Phys. Usp. 43 (2000) 965] 3. Kulić M.I. Phys. Rep. 338 (2000) 1 4. Izyumov Yu.A. Usp. Fiz. Nauk 169 (1999) 225 [Phys. Usp. 42 (1999) 215] 5. Abrikosov A.A. cond-mat/9912394; Int. J. Modern Phys. B 13 (1999) 3405; Physica C 341-348 (2000) 97 6. Alexandrov A.S. cond-mat/0104413; Physica C 363 (2001) 231; Alexandrov A.S., Sricheewin C. cond-mat/0102284; EPL 58 (2002) 576 7. Varelogiannis G. Physica C 317-318 (1999) 238 8. Pines D. Tr. J. Phys. 20 (1996) 535 9. Nagamatsu J. et al. Nature 410 (2001) 63 10. Bud’ko S.L. et al. Phys. Rev. Lett. 86 (2001) 1877 11. Liu A.Y., Mazin I.I., Kortus J. cond-mat/0103570; Phys. Rev. Lett 87 (2001) 087005; Mazin I.I. et al. cond-mat/0204013; Phys. Rev. Lett 89 (2002) 107002 12. Neaton J.B., Perali A. cond-mat/0104098 13. Choi H.J. et al. cond-mat/0111183; Nature 418 (2002) 758; Phys. Rev. B 66 (2002) 020513(R) 14. Leggett A.J. Progr. Theor. Phys. 36 (1966) 901 15. Pickett W. Nature 418 (2002) 733 16. Sharapov S.G., Gusynin V.P., Beck H. cond-mat/0205131; Eur. Phys. J. B 30 (2002) 45 17. Dulić D. et al. Phys. Rev. Lett. 86 (2001) 4144 18. Agterberg D.F., Demler E., Janko B. cond-mat/0201376; Phys. Rev. B 66 (2002) 214507 19. Muller C.J. et al. Physica C 191 (1992) 485 20. Ponomarev Ya.G. et al. Solid State Commun. 111 (1999) 513 21. Maksimov E.G., Arseyev P.I., Maslova N.S. Solid State Commun. 111 (1999) 391 Electronic layout of color FIGURES of the original work -25 -20 -15 -10 -50510 15 20 250.00.20.40.60.81.01.2L = 8 meV,S = 1.7 meVMgB2,T=4.2K4 S4 L1-a stack of 5 SIS contacts,2-a stack of 5 SIS contacts,3-a stack of 2 SIS contacts,4-a single SIS contact.4321dI/dV, arb. un.Vnorm, mV -30 -25 -20 -15 -10 -50510 15 20 25 30-5051015204S4LnL=221-a stack of 5 SnS contacts,2-a stack of 2 SnS contacts,3-a single SnS contact.321nS=13nL=1MgB2, BG series,T = 4.2 K,L=10 meV,S=1.9 meVdI/dV, arb. un.Vnorm, mV Fig. 1. A two-gap structure in normalized CVCs of SIS contacts based on MgB2 (T = 4.2 K). Fig. 2. Two sets of SGS with L and S in normalized CVCs of Andreev contacts. -20-15-10-505101520-4-2024nS=2nS=1nL=2nL=1L = 8 ± 0.5 meV,S = 1.9 meVMgB2,BBS series,T=4.2K21'11, 1' - tunneling single SIS contact,2 - Andreev single SnS contact.I, mA; dI/dV, arb. un.V, mV -10-8-6-4-20246810-15-10-505101520252'21'1MgB2, samp. BB2, T=4.2K, S = 1.8 meV,1, 1' - tunneling mode,2, 2' - Andreev mode-121I, mA; dI/dV, arb. un.I, mA; dI/dV, arb. un.V, mV Fig. 3. Two-gap structure in the CVCs of MB2 break junctions in the tunneling regime (1, 1’) and Andreev regime (2). Fig.4. S-structure in the CVCs of MB2 break junctions in the tunneling regime (1,1’) and Andreev regime (2,2’). -8-6-4-202468-20-10010203 2n=1E0 = 4 meVE0 - plasmonenergy,n - harmonicnumberVn=E0/2en,emission of Leggett's plasmonsby AC Josephson currentMgB2, samp. BB4,T = 4.2 KI, mA; dI/dV, arb. un.V, mV -30-20-100102030-10-8-6-4-20246810m=3m=2m=143 2nL = 1plasmon energyE0 = 4 ± 0.5 meV Vn,m = (2L + mE0)/enLMgB2,samp. KRW4,T = 4.2K,L = 7.5 meV,S = 1.3 meVI, mA; dI/dV, arb. un.V, mV Fig. 5. Structure in the CVC of a SIS MgB2 junction caused by generation of Leggett’s plasmons (T = 4.2 K, E0 = 4 meV). Fig. 6. Subharmonic gap structure modified by emission of plasmons. View publication stats | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/367298383PARADOSSO 01 FEBBRAIO 21Article · January 2023CITATIONS0READS81 author:Some of the authors of this publication are also working on these related projects:Paradosso Paper View projectEugenio Maria FalconeUniversidad de Extremadura3 PUBLICATIONS 0 CITATIONS SEE PROFILEAll content following this page was uploaded by Eugenio Maria Falcone on 20 January 2023.The user has requested enhancement of the downloaded file.14di vedere in prospettiva o in tre dimensioni, incapaci di vedere e capire una fotografia od un film. Uno stato ai primordi della comunicazione, dove il rapporto con l’immagine era di tipo empatico e l’occhio veniva usato come organo di tatto. Gli spazi o i rapporti euclidei, cioè quelli acquisiti durante l’esperienza sensoriale dell’infanzia, oggettivi e defini-bili, erano completamente sconosciuti. L’immagine veniva scandagliata dall’occhio così come noi oggi lo usiamo per leggere, rigo dopo rigo. Sempre l’esperienza di Wilson, con la sua ricerca nei confronti delle socie-tà non-letterate, dimostrò l’incapacità delle popolazioni non in possesso dell’alfabeto fonetico, nell’interpretazione dell’immagine filmata; su un filmato proiettato l’unico elemento riconosciuto dagli spettatori fu una gallina, elemento sfuggito allo stesso autore del film. Inoltre il rapporto degli spettatori con la storia narrata nel film era di tipo non passivo, cioè, se uno dei personaggi del film usciva fuori campo o cantava o soffriva, la platea intera spariva dietro il telone, o cantava o si alzava impaurita. La nostra società ‘letterata’ a questo punto, non dovrebbe subire tali coinvol-gimenti nei confronti delle ‘storie’ né un rapporto di consumatore ‘atti-vo’, ma ciò non accade, altrimenti come spiegare fenomeni quali le soap-opera, o programmi a quiz o a sfondo di forte coinvolgimento emotivo quali Real TV o Chi l’ha visto o programmi che organizzano incontri tra parenti d’oltreoceano dopo venti o trent’anni di separazione. E la gallina nascosta non spunta fuori!Una celebre canzonetta degli anni ottanta cantilenava - «...Vi-deo killed radio star...» - quando l’umana specie inventa nuove for-me di comunicazione o di rappresentazioni della realtà, deve necessa-riamente fare conseguire una morte, la fine dei sistemi precedenti, così l’avvento del cinema, avrebbe dovuto sopprimere il teatro, il cinema sonoro il cinema muto, il video, appunto, annientato il cinema. Ed oggi la televisione distrutto il cinema definitivamente, per non parlare di internet, il nuovo modo di comunicare. Poi ancora, la carta stampata sarebbe divenuta un obsoleto quanto trogloditico modo di esprimersi. Destinata a deperire su scaffali polverosi e divorata dalle tarme, la carta stampata, libri, giornali, riviste e tutto ciò che è in pasta di cellulosa do-vrebbe essere considerata una forma arcaica di comunicazione. La verità è invece che ognuna di queste forme di comunicazione, proseguendo su canali differenti, alcune volte paralleli, altre intersecandosi, altre an-cora in opposta direzione, continuano a vivere e ad avere una propria autonomia. Si mescolano, si contaminano ed alternano realtà a sogno, ideali a divagazioni, ideologie a mitizzazioni, ma sempre con specifiche caratteristiche. I guai arrivano quando i diversi linguaggi, non compresi a fondo o mescolati in maniera confusa, scambiano i parametri di riferi-mento ottenendo ‘prodotti’ ibridi. Abbiamo così: letteratura cinematografica, cinema televisivo, tele-visione cinematografica, pittura televisiva (installazioni), visione teatrale eccetera eccetera eccetera. In ogni caso, tali deformazioni concettuali de-rivano dalle rozze imprecisioni della comunicazione, dall’inadeguatezza delle parole che usiamo e che lontane da un rigore filologico o razionale, ingenerano imprecisioni ed errori. [...] “Il collegamento tra gli uomini av-viene per mezzo della favella, ma i nomi sono imposti alle cose secondo le comprensioni del volgo, e basta questa informe e inadeguata attribuzione di nomi a sconvolgere in modo straordinario l’intelletto. Né valgono cer-to, a ripristinare il naturale rapporto tra l’intelletto e le cose, tutte quelle definizioni e spiegazioni delle quali i dotti si servono sovente [...]”1. Ma è pur vero che il linguaggio, ovvero la parola, oggi non è più solamente composto da fonemi, ma anche da tutta una gamma di metodi di rappre-sentazione, facenti parte dei diversi vettori di comunicazione, appunto, la letteratura, la pittura, il teatro, il cinema, la televisione e infine il sistema multimediale, che è forse la summa di tutti. Il rischio che si corre però è sempre quello dell’appiattimento linguistico, cioè di quella fatidica scusa degli autori della comunicazione che addebitano al pubblico determinate preferenze linguistiche. Nascono così frasi del tipo -«è il pubblico che lo vuole...»-, come se già il pubblico conoscesse un tipo di pittura, un tipo di letteratura o un programma di quiz o un filmetto. In realtà, questa co-moda formula serve a quanti detengono il commercio della comunicazio-ne e del linguaggio. [...] “Perché le parole fanno gran violenza all’intelletto e turbano i ragionamenti, trascinando gli uomini a innumerevoli contro-versie e considerazioni vane”[...]2. E non dimentichiamo che le parole non sono più solo fonemi!3 Il linguaggio si ‘globalizza’ e lascia poco spazio alle diverse identità, le nostre campane ormai non fanno più din, don, dan, ma deng, deng, i nostri baci fanno smac i nostri pugni soc. E dire che nel secolo delle esplorazioni, delle grandi scoperte e dei navigatori spagno-li ed italiani, le lingue straniere nel mondo erano appunto lo spagnolo e l’italiano. Infatti la lingua del dominatore è sempre quella che s’impone ad un popolo colonizzato. Non esistono più linguaggi liberi, se non sono riconoscibili dalla massa. A questo punto vorrei menzionare alcuni studi fatti dall’inventore del villaggio globale Herbert Marshall McLuhan nel suo The Gutenberg galaxy (Armando Armando editore). In occasione di alcune ricerche fatte dal professor John Wilson dell’Istituto di studi afri-cani dell’Università di Londra, si è potuto affermare che le società non-letterate, cioè quelle società non in possesso dell’alfabeto fonetico, era-no incapaci di creare nuove forme di percezione. Infatti, erano incapaci 1. Francesco Bacone, Novum Organum,2. Ibidem,Come suonare i tam-tam con il computerdiEugenio Maria FalconeParadosso, Palermo anno I n°01, Febbraio 2021 ISSN2724-497015Anche perché la gallina non c’è, (la gallina a questo punto assume un valore; quello di valore aggiunto, nascosto, altro).Ci troviamo dunque davanti ad una strana situazione: da un lato le società non-letterate sono incapaci di crearsi nuove forme di percezione e hanno un atteggiamento ‘attivo’ nei confronti dell’opera d’arte o del-la rappresentazione; dall’altra le società letterate, creano continuamente nuove forme di percezione, hanno un atteggiamento di fruitore ‘passivo’ nei confronti dell’opera d’arte o della rappresentazione e malgrado inter-net e il fatidico villaggio globale, ormai divenuto un continuo scambio di dati in una unica lingua, con un unico alfabeto.Una situazione realmente ambigua, in quanto la qualità di ruolo ‘attivo’ dovrebbero essere posseduta dalle società non-letterate, e quella di ruolo ‘passivo’ solo da quelle letterate, la prima, evidenza di una inade-guata preparazione all’interpretazione della comunicazione, la seconda, evidenza della mancanza di un vero e proprio alfabeto, questa volta non fonetico, capace di interpretare la realtà e adeguato ai nuovi mezzi di co-municazione.Allora il villaggio globale, tanto auspicato non è altro che un si-stema unico, mal riuscito, dove un’unica lingua chiude orizzonti, invece che aprirne. La globalità non sta nel parlare tutti la stessa lingua, ma tutti capire la lingua di tutti. Si rischia di utilizzare un sistema avanzato di co-municazione in maniera inopportuna, come se per mezzo del computer si facessero suonare dei tam-tam per comunicare un messaggio.Il ‘valore aggiunto’, ovvero la fatidica gallina, ancora non salta fuo-ri, non viene riconosciuta, forse perché qualcuno non ce l’ha proprio mes-sa, per incapacità o semplicemente perché un popolo ignorante è sempre stato comodo ad un sistema autoritario. E come nelle scene del film Fah-renheit 451 di François Truffaut ci ritroveremo a ‘leggere’ libri composti solo da illustrazioni e quei pochi rimasti, stampati con caratteri tipografi-ci, pubblicamente banditi e dati al rogo, o nelle menti degli ‘uomini libro’.Si presenterà, allora ai nostri occhi un panorama desolato, denso di fumo rossastro, traboccante da alte ciminiere di impianti nucleari, unica fonte di energia rimasta. Tra cavi a fibre ottiche, fonte di luce variopinta e ratti abnormi, tra la spazzatura di non so quale tipo di commestibili, forse creati da strani incroci di ingegneria genetica, si muoverà agile e veloce uno strano essere, composto da parti splendidamente lucide, di una lega particolare. Al centro di esso, tra snodi ed ingranaggi argentati, un involu-cro di cristallo sfavillante. Una miriade di micro-scintille si sprigionano e corrono lungo sottilissimi fili, e immerso ad un liquido rosato, una massa grigia, un magnifico cervello umano, unico elemento di un’umanità total-mente sostituita da protesi.FotoLa villa come dicotomia del giardino diGiovanni Battista Maria Falcone ....................................................................Pg.1Rovine di Sicilia | Ruins of Sicilydi | byArturo Carlo Quintavalle ..................................................................................Pg.4Un Tempio a Palermo, tra Commemorazione e celebrazione, la Casa del Mutilato di Palermo diVincenza Maggiore ...........................................................................................Pg. 7Il “case-studio”: l’assassinio di Emanuele Notarbartolo di Annamaria Amitrano* ................................................................................... Pg.12Come suonare i tam-tam con il computer diEugenio Maria Falcone .................................................................................. Pg.14INDICEISSN 2724-49709 772724 497008View publication stats | RESEARCH GATE |
PSL is Dead. Long Live PSLKevin SmithRice UniversityHouston, U.S.A.kwsmith@rice.eduHai LinPalo Alto NetworksSanta Clara, U.S.A.halin@paloaltonetworks.comPraveen TiwariPalo Alto NetworksSanta Clara, U.S.A.prtiwari@paloaltonetworks.comMarjorie SayerPalo Alto NetworksSanta Clara, U.S.A.msayer@paloaltonetworks.comClaudionor CoelhoPalo Alto NetworksSanta Clara, U.S.A.ccoelho@paloaltonetworks.comAbstract—Property Specification Language (PSL) is a form oftemporal logic that has been mainly used in discrete domains(e.g. formal hardware verification). In this paper, we show thatby merging machine learning techniques with PSL monitors,we can extend PSL to work on continuous domains. We applythis technique in machine learning-based anomaly detection toanalyze scenarios of real-time streaming events from continuousvariables in order to detect abnormal behaviors of a system.By using machine learning with formal models, we leverage thestrengths of both machine learning methods and formal semanticsof time. On one hand, machine learning techniques can producedistributions on continuous variables, where abnormalities can becaptured as deviations from the distributions. On the other hand,formal methods can characterize discrete temporal behaviorsand relations that cannot be easily learned by machine learningtechniques. Interestingly, the anomalies detected by machinelearning and the underlying time representation used are discreteevents. We implemented a temporal monitoring package (TEF)that operates in conjunction with normal data science packagesfor anomaly detection machine learning systems, and we showthat TEF can be used to perform accurate interpretation oftemporal correlation between events.Index Terms—formal methods, PSL, anomaly detectionI. INTRODUCTIONProperty Specification Language (PSL) is a form of tem-poral logic that is designed to capture temporal relationsbetween discrete variables over discrete time. Due to thisnature, PSL has been mainly used in hardware design andverification since it was standardized by IEEE in 2004 [1],[2], [8]. There have been attempts to extend PSL to dealwith continuous variables over continuous time [6]. Due toits inherent limitation of expressibility, there have not beenmany successful applications.In recent years, anomaly detection has been widely usedin practice [3], [13]. There are many applications where real-time streaming events are monitored and analyzed in orderto detect abnormal behaviors. For example, if the amountof free memory of a computer is below a certain threshold,it can be considered as an anomaly. As another example,if there is an anomalous drop in purchase of a product inan online store, it is possible that the product is out ofstock, which needs attention. The state-of-the-art techniqueFig. 1. A New Framework for Anomaly Detectionfor anomaly detection is machine learning [4], [7], [9], [11],[12], [14]. Machine learning techniques learn distributionson continuous variables. Anomaly events can be captured asdeviations from established patterns (distributions). However,there are certain temporal behaviors and relations that cannotbe easily learned by machine learning techniques, but can beeasily characterized by formal languages such as PSL.In this paper, we propose a new framework called TEmporalFiltering (TEF) for anomaly detection (Fig. 1). The idea isto merge machine learning with PSL monitors. The machinelearning module takes as input a number of continuous vari-ables x1, x2, . . . . . . , xm, and outputs some discrete eventsy1, y2, . . . . . . , yn, which become the input of the PSL monitor.The PSL monitor encodes a user-defined temporal relation,which filters the output from the machine learning module. Inthis new framework, machine learning techniques extend thecapability of PSL by discretizing continuous time and events;the PSL monitor refines the results produced by the machinelearning module. This combination of machine learning andformal methods yields a whole that is greater than the sum ofits parts.The rest of this paper is organized as follows. In SectionII, we give a brief introduction to anomaly detection. SectionIII discusses the overall architecture of TEF. In Section IV,we describe how TEF is implemented. Section V illustrateshow TEF can be used to capture temporal relations. SectionVI summarizes the conclusions of the paper and future work.arXiv:2205.14136v1 [cs.LG] 27 May 2022II. ANOMALY DETECTIONAnomaly detection is the process of identifying events thatdeviate from established patterns. Anomaly detection is widelyused in many applications such as detecting cyber intrusions,credit card fraud, and health monitoring. In many applications,input variables have continuous values that vary with time,such as temperature. A time series anomaly detection modellearns baseline behavior from training data and predicts adiscrete set of anomalies. In time series anomaly detectionthe discretizing mechanism is simply that input timestampsare taken from discrete measurements, and there will alwaysbe a minimum nonzero granularity of inputs. These modelsprovide a rich set of examples where a continuous valuedproblem maps to a discrete space, where formal methods canbe applied.Anomaly detection modeling faces two major challenges.One, unbalanced data sets: anomalies are rare, and thus datasets will have few examples of anomalies which the modelcan learn. Two, characterization of anomalies: different typesof models detect different types of anomalies. Level basedmethods find metric outliers. Distribution based methods findanomalies in distributions. Features might not contain thesignal needed to detect anomalies. Being able to characterizeanomalies helps, but unseen anomalies cannot be character-ized. In such cases, model selection is difficult. The nextsections describe some common types of anomaly detectionmodels.A. Level-Based Anomaly DetectionIn these types of models, anomalies in continuous data aredefined as values beyond a specific threshold. The thresholdlevel is typically calibrated to the expected fraction of anoma-lies to be detected. A level that is too low results in falsenegatives, and one that is high results in false positives. Oftenan immediate limitation of level based models is they do notaccount for the frequency of threshold crossings. But thresholdcrossings and the timestamps they occur do make up a discreteevent space for study.B. Distribution-Based Anomaly DetectionDistribution based models learn the statistical distribution ofdata in a baseline or normal state. Anomalies are categorized asevents whose predicted probabilities are lower than a learnedthreshold. These models can learn more sophisticated behaviorthan level-based models - in particular, nonlinear decisionboundaries between anomalous and normal events can belearned. The challenge of tuning the model remains: both theparameters that affect the machine learning of the model, andthe final tuning of the anomaly probability thresholds. Thesechallenges are documented in [10].C. Forecasting Error MethodsA wide range of time series forecasting methods, such asARIMA, can be used to predict probable events from pastdata. Predictions from a recent past period can be comparedwith actual data values to determine if the actual values areanomalous. The comparison of prediction and truth can in turnbe level based or distribution based.D. Template-based Anomaly DetectionIf anomalies follow templates of behavior, established rulesof feature interaction and evolution through time, it’s reason-able to hope that a machine learning model will learn therules. The success of machine learning in many applicationshas led to extensive efforts in applying machine learningmodels to anomaly detection. Given enough features, enoughdata, enough compute power, the reasoning goes, a machinelearning model will learn all of the intricate influences thatdistinguish anomalous behavior from normal. First, we demon-strate that there are cases where enough data is theoreticallyimpossible. For instance, in a time stream of data, the eventconsisting of repeated events a followed by an event b are notpossible to learn from a finite dataset. Second, while it maybe possible to learn an underlying template for anomalies, theamount of data and compute resources required for sufficientlyaccurate results might be prohibitive.E. A Hybrid MethodBecause anomaly detection models create discrete sets ofevents they naturally can be described using PSL. Machinelearning can make PSL relevant in continuous applications.Because PSL can easily characterize infinite sets such as the”arbitrary stream of a followed by b” example as a simple PSLexpression: a[+]; b, machine learning anomaly detection canbe enhanced. We aim to show in this paper that time seriesanomaly detection together with TEF can improve overallmodel performance.III. TEF OVERVIEWTEF implements a subset of PSL. PSL is an extensionof linear temporal logic and adds a number of operators toexpress temporal constraints. In particular, PSL makes useof Sequential Extended Regular Expressions (SEREs), definedbelow. If a PSL formula is composed entirely of SEREs, it issaid to be written in SERE-style PSL. TEF implements mostof SERE-style PSL.Propositional formulas (i.e., boolean variables closed un-der conjunction, disjunction, and negation) are the atoms ofSEREs. SEREs are SERE atoms closed under the SEREoperators, which are analogous to the operators of regularexpressions and also include supplemental operators repre-senting useful syntactic sugar. Just as regular expressions areused to match strings, SEREs are used to match traces, thatis, sequences of truth assignments. A SERE atom matches atruth assignment when the truth assignment makes the atomtrue. The semantics of compound SEREs are shown in Fig. 3and Fig. 4.One of the advantages in using a temporal logic to specifyproperties is that logics are declarative. The result can be pre-cisely described without the use of control flow or statementsthat modify a program’s state. This allows those without abackground in software engineering to write useful properties.However, the operators of linear temporal logic, althoughconceptually simple, are not trivial to use in practice, and mostpeople are unfamiliar with them. An advantage of SERE-stylePSL is that the functionality of LTL operators is subsumedby the SERE operators and so familiarity with LTL is nota requirement to writing properties. The close resemblancebetween SEREs and standard regular expressions make theformer exceptionally easy to learn if one is familiar with thelatter. And regular expressions are common currency not justin software engineering, but in data analysis and related fieldsas well. This makes SERE-style PSL easily accessible to thosewith a wide variety of backgrounds.So SERE-style PSL is a natural choice for TEF, which aimsto provide a simple and accessible way to express and evaluatetemporal constraints. In fact, TEF extends SERE-style PSL inan intuitive and useful way by allowing the use of Booleanexpressions wherever Boolean variables may occur in SEREs.Another natural choice is the decision to package TEF as anextension to Pandas, which is also common currency in thePython world.TEF makes use of Python’s regex engine, which takes asinput a regex pattern (regular expression) and a string. Theregex engine reports all segments of the string that match theregex pattern. Python’s regex engine is highly optimized andefficient. Fig. 3 lists the main temporal operators, which TEFimplements, and compares them with Python’s regex patterns.In this table, r values are SEREs, s values are sequences ofrows in a DataFrame, r′ values are regex patterns, s′ values arestrings. Readers are referred to [15] for the formal semanticsof SERE-style expressions.At a high level, TEF takes advantage of the highly efficientregex engine and the similarity between SERE expressions andregex patterns (shown in Fig. 3). TEF reduces the problemof checking if a SERE expression matches rows of a dataframe to checking if a regex expression matches a string(Fig. 2). The reduction is based on the following observation.Let E be the set of all SERE expressions, R be the setof sequences of rows in a data frame, P be the set of allregex patterns, and S be the set of strings. There exist twofunctions f : E → P and g : R → S s.t. ∀e : E,r :R. matches(e, r) ↔ matches(f(e), g(r)). Intuitively, Fig. 3justifies this observation.Fig. 2. How TEF is implementedTEF also implements some additional operators as syntaticsugar (shown in Fig. 4). TEF also adds the “[]” operator for usein Boolean expressions, which allows further flexibility in theexpression of temporal relations. It functions in the followingway. Suppose we have a data frame with a column c. Thenc[−1] refers to the value at the previous row in column c.Also, c[1] refers to the value at the next row in column c.In general, if i is the index of the current row and j is aninteger, then c[j] refers to the value at row i + j of column c.As syntactic sugar, c may be used as an abbreviation for c[0].An illustrative example of a property making use of the “[]”operator is (c > c[−1])[∗5], which matches all segments of thetrace in which the value at c increases five times consecutively.Algorithm 1 An algorithm for converting a SERE expressione to a regex expression1: function PSL TO REGEX(e)2:if e is a Boolean expression b then3:return the disjunction of truth assignments thatmake b true, in string form.4:else if e is r1; r2 then5:s1 ← PSL TO REGEX(r1)6:s2 ← PSL TO REGEX(r2)7:return ‘(’ + s1 + s2 + ‘)’8:else if e is r1|r2 then9:s1 ← PSL TO REGEX(r1)10:s2 ← PSL TO REGEX(r2)11:return ‘(’ + s1+ ‘|’ +s2 + ‘)’12:else if e is r1&r2 then13:s1 ← PSL TO REGEX(r1)14:s2 ← PSL TO REGEX(r2)15:left = ’(’ + ’(?=’ + s1 + ’)’ + s2 + ’)’16:right = ’(’ + ’(?=’ + s2 + ’)’ + s1 + ’)’17:return ’(’ + left + ’|’ + right + ’)’18:else if e is r[∗] then19:s ← PSL TO REGEX(r)20:return ‘(’ + s + ‘*’ + ‘)’21:else if e is r[+] then22:s ← PSL TO REGEX(r)23:return ‘(’ + s + ‘+’ + ‘)’24:else if e is r[∗n] then25:s ← PSL TO REGEX(r)26:return ’((’ + s + ’)’ + ’{’ + n + ’}’ + ’)’27:else if e is r[∗n..m] then28:s ← PSL TO REGEX(r)29:return ’((’ + s + ’)’ + ’{’ + n + ‘,’ + m + ’}’ +’)’30:else if e is r[∗n..] then31:s ← PSL TO REGEX(r)32:return ’((’ + s + ’)’ + ’{’ + n + ‘,’ + ’}’ + ’)’33:else if e is r[∗..m] then34:s ← PSL TO REGEX(r)35:return ’((’ + s + ’)’ + ’{0’ + ‘,’ + m + ’}’ + ’)’36:end if37: end functionSERE SyntaxMeaningRegex SyntaxMeaningr1; r2matches s if s = s1s2, r1 matches s1 and r2 matches s2r′1r′2matches s′ if s′ = s′1s′2, r′1 matches s′1 and r′2 matches s′2r1|r2matches s if r1 matches s or r2 matches sr′1|r′2matches s′ if r′1 matches s′ or r′2 matches s′r1&r2matches s if r1 matches s and r2 matches s? = r′matches s′ if r′ matches s′, butitdoesnotconsumeanyr′r[∗]matches s if 0 or more concatenations of r matches sr′∗matches s′ if 0 or more concatenations of r′ matches s′r[+]matches s if 1 or more concatenations of r matches sr′+matches s′ if 1 or more concatenations of r′ matches s′r[∗n]matches s if n concatenations of r matches sr′{n}matches s′ if n concatenations of r′ matches s′r[∗n..m]matches s if between n and m concatenations of r matches sr′{n, m}matches s′ if between n and m concatenations of r′ matches s′r[∗n..]matches s if n or more concatenations of r matches sr′{n, }matches s′ if n or more concatenations of r′ matches s′r[∗..m]matches s if m or fewer concatenations of r matches sr{0, m}matches s′ if m or fewer concatenations of r′ matches s′Fig. 3. SERE expressions and Python’s regex patternsSyntaxIntuitive Meaning[+]True[+][∗]True[∗][∗n]True[∗n..][∗n..m]True[∗n..m][∗n..]True[∗n..][∗..m]True[∗..m]r[− >]!r[∗]; rr[− > n]!r[∗]; r[∗n]r[− > n..m]!r[∗]; r[∗n..m]r[= n](!r[∗]; r)[∗n]; !r[∗]r[= n..m](!r[∗]; r)[∗n..m]; !r[∗]r[= n..](!r[∗]; r)[∗n..]; !r[∗]r[= ..m](!r[∗]; r)[∗..m]; !r[∗]Fig. 4. Additional SERE-style operatorsIV. TEF IMPLEMENTATIONTEF shadows the Pandas DataFrame eval method. It takesas input a SERE-style PSL property in the form of a string,and it matches it against the DataFrame, which is interpretedas a trace. As we discussed in Section III, TEF reduces theproblem of checking if a SERE expression matches rows ofa data frame to checking if a Python regex pattern matches astring. In order to make the reduction work, we just need tobuild two functions: (1) f which converts a SERE expressionto a regex pattern. (2) g which converts a data frame to astring.To build the function f, we use a simple recursion based onhow the SERE expression is constructed. The simplest SEREexpression is a Boolean expression. To convert a Booleanexpression to a regex pattern, we rewrite the expression inthe form of the disjunction of its set of satisfying truth assign-ments, in string form. As a result, the length of the patternis exponential in the number of distinct boolean expressionsused in the property. In practice we have found that manyproblems need make use of only a few expressions. Often onlya single expression is used to express a useful property. Toconvert more complicated SERE expressions involving SEREoperators, the individual pieces are converted recursively, thenthe results are combined based on the SERE operator. Forexample, suppose that we want to convert r1; r2. We firstrecursively convert r1 and r2, get two regex strings, and thenconcatenate the two regex strings. The details of convertinga SERE expression into a Python regex pattern is shown inAlgorithm 1.To build the function g, TEF does two steps:• Step 1 (Booleanize a data frame) TEF identifies all theBoolean expressions used in the property, and evaluatesthem with respect to each row in the data frame. Itkeeps track of the results in a new data frame, in whichthe results of each Boolean expression are kept in newcolumns. The details of Booleanizing a data frame isshown in Algorithm 2.• Step 2 (convert to a string) The Booleanized DataFrameis converted to a string by using ‘,’ to seperate columnsand using ‘()’ to group characters within the same rows.After converting a SERE expression into a regex pattern andconverting a data frame into a string, Python’s regex engineis used to find matches. Results are returned in the form of alist of index pairs indicating where in the trace the propertyis satisfied.Algorithm 2 An algorithm for Booleanizing a data frame dfw.r.t. a compound Boolean expression e1: function BOOLEANIZE DATAFRAME(df, e)2:Initialize b df to be an empty data frame.3:for all row r in df do4:Initialize r′ to be an empty row.5:for all atom a in e do6:Evaluate a w.r.t. r, add the result to r′7:end for8:Add r′ to b df9:end for10:return b df11: end functionV. CASE STUDIESA. Analyzing Weather Patterns using TEFFigure 5 shows a data frame containing the weather in-formation in Amarillo, TX, in April 2021. The weather datais taken from [16]. We make the following queries to thedata frame. Pandas DataFrame eval method can handle queriesinvolving only one row. If the query refers to multiple rows,we need to use TEF eval method.1. Find all individual days, when the temperature is eithertoo hot (temp high ≥ 80) or too cold(temp low ≤ 40). Wecan use Pandas DataFrame eval method, which returns the setof individual rows, where the constraint is satisfied (Fig. 7).Fig. 5. Data Frame: dfFig. 6. Booleanized Data Frame: b dfFig. 7. Query using Pandas DataFrame eval method2. Find two consecutive days where a hot day (temp high≥ 80) is followed by a cold day (temp low ≤ 40). The PandasDataFrame eval method cannot handle this query, since itcannot reason about temporal relations involving multiple rowsof a data frame. We use TEF eval method instead. TEF doesthe following 3 steps to evaluate “temp high ≥ 80; temp low≤ 40” against the data frame df (Fig. 5).• Step 1: As discussed in Section IV, TEF identifies all theBoolean expressions used in the property, and evaluatesthem with respect to each row in df. This particular querycontains two Boolean expressions “high temperature ≥80” and “low temperature ≤ 40”. The Booleanized dataFig. 8. Query using TEF eval methodframe b df (Fig. 6) has two columns and has the samenumber of rows as df. b df is transformed into thefollowing string by using ‘,’ to seperate columns andusing ‘()’ to group characters within the same rows.(0,1)(0,0)(0,0)(0,0)(1,0)(1,0)(0,0)(0,0)(0,0)(0,1)(1,0)(0,1)(0,1)(0,0)(0,0)(0,1)(0,1)(0,1)(0,1)(0,1)(0,1)(0,1)(0,0)(0,0)(1,0)(1,0)(0,0)(0,0)(0,0)(0,0)• Step 2: The PSL property is compiled to a Python regexpattern: (\ (1, 0\) |\ (1, 1\))(\ (0, 1\) |\ (1, 1\)).• Step 3: Python’s regex engine is used to find matches.The result is shown in Fig. 8.3. Find two consecutive days when (1) temp high ≤ 80 andtemp low ≥ 40 (2) humidity ≥ 20 and humidity ≤ 70 (3)wind speed < 30. Again, this is a temporal relation involvingmultiple rows of the data frame. We should use TEF evalmethod. The result is shown in Fig. 9.Fig. 9. Query using TEF eval methodB. Analyzing Dow Jones Industrial Average (DJIA) indexusing TEFIn this section, we show how TEF can be used to analyzepatterns in Dow Jones Industrial Average (DJIA) index. Thisdataset contains data from 01/01/1980 to 12/31/2012, and istaken from [17]. All experiments are done on a Macbook, with2.6 GHz 6-Core Intel Core i7 and 16 GB 2667 MHz DDR4.1. [Rise after Drop] Find all periods where the index hasbeen dropping for 5 consecutive days and rises on the nextday. The result is shown in Fig. 10.Fig. 10. Rise after Drop2. [Fluctuation] Find all periods where the index decreaseson one day and increases on the next day, and this patternrepeats for at least 5 times. The result is shown in Fig. 11.Fig. 11. Flunctuation3. [Steady] Find all periods where the index stays between5000 and 6000 for at least 10 days. The result is shown inFig. 12.Fig. 12. Steady4. [Jump] Find all periods where the index increases by atleast 10% compared to the previous day. The result is shownin Fig. 13.C. Isolation Forest Anomaly Detection Models and TEFUsing examples from publicly available data sets, we showcases where false positives from a trained isolation forestmodel can be filtered by rules. The data is taken from [18]. TheFig. 13. Jumpentire dataset collection consists of 250 datasets from domainssuch as medical monitoring, motion sensors, and weather. Weidentified 130 examples where the model output could beimproved by a rule based filter.Fig. 14. Identify cluster of predictionsIn Fig. 14 the anomaly is difficult to detect as it hassimilar amplitude and frequency to the baseline pattern. Butthe detected anomalies are clustered in time and anomalies thatare farther apart could be ruled out by a simple rule filter. Wecan use TEF to specify that within a cluster, the total numberof anomalies needs to reach a certain threshold. For example,cluster2,5 = “anomaly[−2] + anomaly[−1] + anomaly +anomaly[1] + anomaly[2] ≥ 2” specifies that at least 2anomalies need to occur within a cluster of width 5. The widthof the cluster and the threshold can be adjusted for differentapplications. Fig. 15 is an example where the anomaly canbe detected by a gap that follows detected anomalies. Inorder to specify that a gap of width x needs to follow adetected anomaly, we can use: “[∗]; anomaly; !anomaly[∗x]”.The original data, shown in the upper graph, has an anomalythat is difficult to characterize.As a final example, Fig. 16 shows a case where a suitablefiltering rule would take into account both cluster and gapbehavior. “cluster2,5; !anomaly[∗5]; cluster2,5” specifies asequence of events: at least 2 anomalies occur within a clusterof width 5, followed by a gap of width 5, followed by at least2 anomalies occurring within a cluster of width 5. Again, thewidth of the cluster and the width of the gap can be adjustedfor different applications.Fig. 15. Identify by gapFig. 16. Identify by gapVI. CONCLUSIONS AND FUTURE WORKFor many years, the use of PSL has been limited to hardwaredesign and verification, since PSL is designed to deal withdiscrete variables over discrete time. In this paper, we proposethat PSL can be used in continuous domain for anomalydetection. The idea is to merge PSL with machine learningand use a hybrid framework for anomaly detection. Thishybrid framework consists of a machine learning moduleand a PSL monitor. The machine learning module outputsanomalies in the form of discrete time and events. The PSLmonitor further refines the output of the machine learningmodule by checking if some user-defined temporal relationis satisfied. We implemented a temporal monitoring package(TEF), and we show that TEF can be used to perform accurateinterpretation of temporal correlation between events.For future work we are experimenting with the followingcases: multivariate anomaly detection problems, and TEF inconjunction with ensemble models. With TEF, we have theopportunity to transform anomaly characterization rules intodata, and develop models from there.REFERENCES[1] Roderick Bloem, Stefan J. Galler, Barbara Jobstmann, Nir Piterman,Amir Pnueli, Martin Weiglhofer: “Specify, Compile, Run: Hardwarefrom PSL.” Electron. Notes Theor. Comput. Sci. 190(4): 3-16 (2007)[2] Marc Boule, Zeljko Zilic: “Efficient Automata-Based Assertion-CheckerSynthesis of SEREs for Hardware Emulation.” ASP-DAC 2007: 324-329[3] Varun Chandola, Arindam Banerjee, Vipin Kumar: “Anomaly detection:A survey.” ACM Comput. Surv. 41(3): 15:1-15:58. 2009.[4] Lei Cui, Youyang Qu, Gang Xie, Deze Zeng, Ruidong Li, ShigenShen, Shui Yu: “Security and Privacy-Enhanced Federated Learning forAnomaly Detection in IoT Infrastructures.” IEEE Trans. Ind. Informatics18(5): 3492-3500 (2022)[5] Cindy Eisner, Dana Fisman, “A Practical Introduction to PSL.” Serieson Integrated Circuits and Systems, Springer 2006, ISBN 978-0-387-35313-5, pp. 1-240[6] Ning Ge, Silvano Dal-Zilio, Hongyu Liu, Li Zhang, Lianyi Zhang, “RT-MOBS: A compositional observer semantics of time Petri net for real-time property specification language based on µ-calculus.” Sci. Comput.Program. 206: 102624 (2021)[7] Meenal Jain, Gagandeep Kaur, Vikas Saxena: “A K-Means clusteringand SVM based hybrid concept drift detection technique for networkanomaly detection.” Expert Syst. Appl. 193: 116510 (2022)[8] Fatemeh Negin Javaheri, Katell Morin-Allory, Dominique Borrione:Synthesis of Regular Expressions Revisited: From PSL SEREs toHardware. IEEE Trans. Comput. Aided Des. Integr. Circuits Syst. 36(5):869-882 (2017)[9]¨Omer Kasim: “An efficient and robust deep learning based networkanomaly detection against distributed denial of service attacks.” Comput.Networks 180: 107390 (2020)[10] Siwon Kim, Kukjin Choi, Hyun-Soo Choi, Byunghan Lee, SungrohYoon: “Towards a Rigorous Evaluation of Time-series Anomaly De-tection” https://arxiv.org/pdf/2109.05257.pdf[11] Jinbo Li, Hesam Izakian, Witold Pedrycz, Iqbal Jamal: “Clustering-based anomaly detection in multivariate time series data.” Appl. SoftComput. 100: 106919 (2021)[12] Hualin Ma, Liyan Zhang: “Attention-based framework for weakly su-pervised video anomaly detection.” J. Supercomput. 78(6): 8409-8429(2022)[13] Animesh Patcha, Jung-Min Park, “An overview of anomaly detectiontechniques: Existing solutions and latest technological trends.” Comput.Networks 51(12): 3448-3470. 2007.[14] Jiaming Pei, Kaiyang Zhong, Mian Ahmad Jan, Jinhai Li, “Personalizedfederated learning framework for network traffic anomaly detection.”Comput. Networks 209: 108906 (2022)[15] Ingo Pill, “Requirements Engineering and Efficient Verification of PSLProperties”, Doctoral Thesis, Institute for Software Technology, GrazUniversity of Technology, 2008.[16] https://www.wunderground.com/history/monthly/us/tx/amarillo/KAMA/date/2021-4[17] https://vincentarelbundock.github.io/Rdatasets/datasets.html[18] https://www.cs.ucr.edu/∼eamonn/time series data 2018/UCRTimeSeriesAnomalyDatasets2021.zip | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/367453366Maria y su lugar en la iglesia.Article · January 2017CITATIONS01 author:Some of the authors of this publication are also working on these related projects:Object holder connected to a steering wheel View projectLuis Alejandro Yarcho de la Puente38 PUBLICATIONS 0 CITATIONS SEE PROFILEAll content following this page was uploaded by Luis Alejandro Yarcho de la Puente on 26 January 2023.The user has requested enhancement of the downloaded file.María y su lugar en la iglesia María es nuestra hermana EN CRISTO, es bienaventurada entre todas las mujeres, y es amada por todo cristiano. Cristo por ser Dios, no es afectado al entrar en contacto con el pecado… ¿Acaso Cristo se va a contaminar tocando a muertos espirituales? Sea un inconverso, creyente o María (en el caso hipotético que esté afectada por el pecado original y de haber pecado como todo creyente)… Jesucristo tocó al leproso, al muerto y hasta a Judas; uno fue sanado, el segundo resucitó y con Judas, ya se sabe (al ser tocado por Jesús, por ejemplo en el beso que le dio Judas, no quedó sin pecado, ni tampoco cayó fulminado) y lo tocó, tocó al Verbo hecho carne. El cristiano es templo del Espíritu Santo y de Cristo NO PRACTICA PECADO, pero aún la carne sigue activa y puede pecar; por supuesto, uno pide perdón, se arrepiente y hace el propósito de no volver a hacerlo, permaneciendo en Su Palabra; asimismo María tuvo a Cristo en ella, pero eso no implica que nunca más (jamás) haya pecado y lo probaremos por las Escrituras en el primer punto. Todo cristiano ES TEMPLO DE DIOS (1Corintios 3:16 “¿No sabéis que sois templo de Dios”), DE CRISTO (Colosenses 1:27 “…que es Cristo en vosotros, la esperanza de gloria”),Y DEL ESPÍRITU SANTO (1Corintios 6:19 “¿O ignoráis que vuestro cuerpo es templo del Espíritu Santo”), que mora dentro de uno. Eso no significa que el cristiano no vaya a pasar por la muerte física; igual que todo redimido y salvo, envejecerá, aunque el Creador more en uno; tampoco eso significa que el cristiano, templo de Dios nunca más peque…, sino que Abogado tenemos con el Padre: “Hijitos míos, estas cosas os escribo para que no pequéis; y si alguno hubiere pecado, abogado tenemos para con el Padre, a Jesucristo el Justo.” (1Juan 2:1). Dios es INDEPENDIENTE AL HOMBRE, puso el ADN (ese X, del XY) en María con Su Sombra y nació COMO TODO SER HUMANO. Desde el principio hasta el fin se sujetó a las leyes por amor, incluidas las de la NATURALEZA. El que Dios mismo esté formándose en María como hombre (Dios manifestado en carne) no significa que María YA NUNCA PEQUE… Aunque haya tomado ADN de María lo limpia sin contaminarse Él como nos limpia a todo templo viviente actual donde mora. Ahondemos un poco más…, ahora somos, digo de los que son iglesia (el que no es iglesia no va entender), a los que somos templos de Dios. Y, ¿acaso esto es poca cosa?… ¿acaso Cristo es superior al Espíritu Santo o viceversa…?, no, los tres, Padre, Hijo y Espíritu Santo son UNO e IGUALES COMO DIOS. Bueno, siendo ahora el creyente TEMPLO DE DIOS y estar en contacto DIRECTO CON DIOS y tener RÍOS DE AGUA VIVA (el Espíritu Santo) FLUYENDO EN EL INTERIOR (que lo tienen sus hijos), y NO SOMOS FULMINADOS a pesar que podemos IR AL LUGAR SANTÍSIMO (sino leer Hebreos 9 y 10). Sin embargo, si el Señor no viene, encararemos ese TRÁNSITO donde este cuerpo ha de morir (la Palabra la llama ‘dormir’ para el creyente); si se quiere llamarle dormición, bien…; pero es por causa de la “paga del pecado”, por la herencia de Adán que esto sucede (Adán antes de pecar NO MORÍA). Claro está, que en Cristo seremos resucitados y tendremos cuerpos glorificados finalmente. Y podemos ser templos del Espíritu Santo, gracias justamente A CRISTO (“Cristo EN VOSOTROS, la esperanza de gloria” Colosenses 1:27); es Cristo que hace que no seamos fulminados, y por ese CORDERO que fue inmolado desde el principio del mundo (Apocalipsis 13:8) y lo mismo va para María. Cristo pudo estar en su vientre sin fulminarla (ni contaminarse… ya que es Dios, ¿cómo se iría a contaminar?; en tal caso no sería Dios… no sería el Cristo, el Hijo de Dios, Dios de los cristianos, sino un falso dios, otro cristo… que no es TODOPODEROSO), ya que está en Su Naturaleza… no puede pecar y no fulmina al pecador (a lo afectado por el pecado). Cristo no sólo debía estar en el vientre de María SINO EN SU CORAZÓN. La doctrina de impecabilidad, se ha logrado CAMBIANDO el sentido de unos pocos pasajes… se sostiene que fue llamada de una forma ÚNICA: “llena eres de gracia” o “plena de gracia” y que ese término griego por la que le llaman κεχαριτωμένη (kejaritoméne), y dicen equivocadamente que SÓLO es utilizado en ella… Esto amerita una explicación: Veremos que "gracia plena" no es la mejor traducción para kejaritoméne; utilizan la traducción latina de Jerónimo, pero finalmente la fuente inicial es el griego. Veremos que el mismo Jerónimo traduce la palabra de otra forma en otro de sus documentos, lo que rompre la burbuja. Kejaritoméne es un verbo, participio perfecto, pasivo, vocativo, femenino, singular de χαριτόω. Es FALSO que es un término “ÚNICO” sólo utilizado en María, en el sentido que le da en el catolicismo romano; se saca los textos de su uso natural encontrados en la Palabra. Para comenzar examinamos un texto que tiene el mismo verbo (aunque está en aoristo y no en perfecto). El mismo verbo con otra declinación se encuentra en Efesios 1:6 (ἐχαρίτωσεν - ejaritosen) verbo, indicativo AORISTO, activo, 3ra persona, singular, también de χαριτόω. El prefijo “ke”, y el sufijo “mene” de KEjaritoMÉNE son variantes que se añaden a χαριτόω = mostrar gracia.: “ke” es un prefijo del tiempo perfecto y “mene” es un sufijo por ser participio, femenino…, que se encuentra en Efesios 1:6 (ἐχαρίτωσεν - EjaritoSEN). Vemos el prefijo “e” y el sufijo “sen”. Donde “e” es el prefijo para aoristo y “sen” el sufijo para 3ra persona del singular. Se trata del mismo verbo: χαριτόω (jarito). En Efesios 1:6 dice: “para alabanza de la gloria de su gracia, con la cual nos HIZO ACEPTOS (ἐχαρίτωσεν - EjaritoSEN) en el Amado”; que en otras versiones es traducido: “que nos favoreció en el Amado” (Holman; Derby), “que nos fue dada GRATUITAMENTE” en la NIV; “ACEPTOS en el amado” en la Biblia del Peregrino (católica romana); “la cual nos DOTÓ en el Amado” en la Biblia de Serafín Ausejo (católica); y así sucesivamente. En general este verbo utilizado = gracia (χαριτόω - jarito), NO SÓLO ES APLICADA A MARÍA SINO A TODO CREYENTE (esto es válido para toda gracia, don); que es la raíz del verbo = gracia utilizado en ambos pasajes; y eso no significa que ella o algún creyente sea sin pecado. En ambos casos: Lucas 1:28, donde está el verbo en perfecto (tiempo presente; una acción completada en el momento de hablar) y en Efesios 1:6, donde está en aoristo (un sólo acto en tiempo pasado), se trata de TIEMPOS QUE IMPLICAN OBRAS COMPLETADAS. En sí, todo don dado a los creyentes ha sido PREPARADO Y HECHO CON ANTELACIÓN: “Porque a los que antes conoció, también los predestinó para que fuesen hechos conformes a la imagen de su Hijo.” (Romanos 8:29); “Porque POR GRACIA SOIS SALVOS por medio de la fe; y esto no de vosotros, pues es DON DE DIOS; no por obras, para que nadie se gloríe. Porque somos hechura suya, creados en Cristo Jesús para buenas obras, las CUALES DIOS PREPARÓ DE ANTEMANO para que anduviésemos en ellas.” (Efesios 2:8-10) Ahora veamos verbos distintos con la misma declinación a Lucas 1:28 (κεχαριτωμένη) en otros pasajes aceptados por el catolicismo romano, y examinemos como son traducidas; esto es: un verbo, presentado como un participio, en tiempo perfecto, voz pasiva, y caso vocativo. Sólo con un caso distinto queda descartada la traducción católico-romana. Se citan versiones católico/romanas: (Tobías 10:13) Edna dijo a Tobías: "Hermano querido (ἠγαπημένε = verbo, participio, perfecto, pasivo, vocativo), que el Señor del cielo te haga volver otra vez. No se va a armar todo un “tratado teológico” diciendo que ese “querer” era la “personificación” del querer, donde se da el “único” y “perfecto” querer, que implica que la persona que haya tenido ese “querer” sea sin pecado; por lo que “decretamos…”. (Daniel 13:52) Cuando estuvieron separados el uno del otro, llamó a uno de ellos y le dijo: "¡Viejo envejecido (πεπαλαιωμένε = verbo, participio, perfecto, pasivo, vocativo) en el mal, ahora caen sobre ti los pecados que cometiste antes…” - o - (Susana 1:52) Los apartaron, él llamó a uno y le dijo: ¡Envejecido (πεπαλαιωμένε = verbo, participio, perfecto, pasivo, vocativo) en años y en crímenes! Ahora vuelven tus pecados pasados… Tampoco se va a declarar que “nació envejecido”, de bebé y de niño fue “lleno de pecado”, que fue “sin gracia alguna” al nacer; sino que fue endureciéndose; y se cumplió en ese anciano “la completa sequedad”, y expresiones por el estilo sin sentido. (Hechos 23:3) “Entonces Pablo le dijo: ¡Dios te golpeará a ti, pared blanqueada (κεκονιαμένε = verbo, participio, perfecto, pasivo, vocativo)! ¿Estás tú sentado para juzgarme conforme a la ley, y quebrantando la ley me mandas golpear?” ¿Entonces el estado de Ananías fue de una pared blanqueada desde nacer (aún de niño)? ¿Y peor aún afirmar que lo fue completamente desde nacer y de niño? Eso sucede cuando ANTES DE ANALIZAR EL TEXTO SE TIENE UNA DOCTRINA POR JUSTIFICAR DE ANTEMANO… prueba la eiségesis en la interpretación de Lucas 1:28. Sigamos: Por ejemplo vemos en Romanos 15:14: “Pero estoy seguro de vosotros, hermanos míos, de que vosotros mismos estáis llenos de bondad, LLENOS DE TODO CONOCIMIENTO, de tal manera que podéis amonestaros los unos a los otros.”; “llenos de todo conocimiento”, (pepleromenoi) πεπληρωμένοι = llenos es TAMBIÉN un VERBO PARTICIPIO PERFECTO PASIVO y por eso no se va armar toda una teología de que “ya se sabe todo” y claro que es una OBRA CONSUMADA (“el CORDERO que fue inmolado desde el principio del mundo.” Apoc. 13:8), que se realiza y progresa en el presente en el creyente. Y fijarse en ese SUFIJO… no les parece conocido “menoi” que es un masculino, plural, y no en femenino singular como “menes” de Lucas 1:28, aplicado a María; pepleromenoi viene πληρόω. De la misma forma revisando en griego en la Palabra hay 1760 Verbos, utilizados como Participio, en tiempo Perfecto y voz Pasiva (desde ahora como: VPPP por la primera letra de cada palabra en la declinación del verbo), como en Lucas 1:28; si se revisa podemos sacar muchos ejemplos que contradicen su interpretación. Debido a la cantidad de pasajes, veremos algunos de Génesis (por ser los primeros de la lista tomando del A.T.) y luego iremos a pasajes de Lucas (por localizarse allí el pasaje que examinamos): • Génesis 6:12 “Y miró Dios la tierra, y he aquí que estaba CORROMPIDA (κατεφθαρμένη = VPPP)” – Pero sabemos que no siempre la tierra estuvo corrompida. Al principio no estuvo corrompida; con el diluvio tuvo una limpieza y al final de los tiempos será pasada por fuego y será limpia: No es un estado permanente. • En Génesis 39:22 se utiliza la palabra al cuidado de José de los presos (ἀπηγμένους = VPPP). Dejó de estar al cuidado. • En Génesis 40:6 cuando José se encuentra con el copero y el panadero que habían tenido sueños: “Vino a ellos José por la mañana, y los miró, y he aquí que estaban tristes (τεταραγμένοι = VPPP).” (uno de ellos luego fue restaurado) • O ver Génesis 41:36 Para la provisión guardada (πεφυλαγμένα = VPPP) previniendo los 7 años de hambre. (previnieron para un tiempo determinado limitado) • En Lucas 2:12 “Esto os servirá de señal: Hallaréis al niño envuelto (ἐσπαργανωμένον = VPPP) en pañales, acostado en un pesebre.” (ya no tiene pañales) • En Lucas 3:13 Cuando Jesús explica a los publicanos (recaudadores de impuestos) que no deberían de exigir màs de lo que la autoridad romana pedía. “Él les dijo: No exijáis más de lo que os está ordenado (διατεταγμένον = VPPP).” • El hombre paralítico llevado delante de Jesús en Lucas 5:18 y 5:24 no es un estado de plenitud permanente… παραλελυμένος (Paraleluménos = VPPP). • En Lucas 6:25 “¡Ay de vosotros, los que ahora estáis saciados (ἐμπεπλησμένοι = VPPP)! porque tendréis hambre.” El mismo Jesús aclara que ese verbo en participio, perfecto, pasivo es un estado con raíces no profundas y menos permanente… • “Habló a los invitados (VPPP) una parábola” en Lucas 14:7 y 8 (Ελεγεν δὲ πρὸς τοὺς κεκλημένους [VPPP] παραβολήν). La condición de invitados no es permanente… • Igual ver: Lucas 9:32; 11:25; 12:2; 16:20; 19:30; 24:38. Sigamos: Si examinamos este pasaje que se encuentra en las Biblias católico-romanas (Eclesiástico), sellaremos el asunto con creces, veamos: Eclesiástico o Sirac 18:17 “¿No ves que es mejor la palabra que el regalo? El hombre CARITATIVO sabe aunar las dos cosas.” ¡La palabra en griego traducida caricativo es: κεχαριτωμένῳ (KEjaritoMÉNO)! ¿Les parece conocida? Justo en Lucas 1:28 se utiliza: κεχαριτωμένη (KEjaritoMÉNE). Comparémoslas: KEjaritoMÉNO y KEjaritoMÉNE… ¿y aún siguen afirmando que es una palabra única que implica que es sin pecado y separada desde antes y demás yerbas que no se dicen para que no se pisen? κεχαριτωμένῳ (KEjaritoMÉNO) es un verbo, participio, perfecto, pasivo, dativo, masculino, singular y NADIE PODRÁ NEGAR LO QUE SIGNIFICA POR SU CONTEXTO… ¡Está hablando EN GENERAL del HOMBRE CARICATIVO! Está en perfecto porque Dios, como hemos señalado antes, ha preparado sus obras, que son un hecho consumado hoy… pero nada más fuera del contexto sería afirmar que esos hombres jamás pecan, que son sin pecado original y demás aberraciones que desdibujan lo que la Palabra habla de María creando OTRA María que no es la bíblica sino la pagana; y por consiguiente otro cristo que es dependiente de esta maría. Por eso los textos deben leerse dentro del contexto y SIN QUEBRAR LAS ENSEÑANZAS MÁS GENERALES: “Por tanto, como el pecado entró en el mundo por un hombre, y POR EL PECADO LA MUERTE, así LA MUERTE PASÓ A TODOS los hombres, por cuanto TODOS PECARON.” (Rom 5:12); “Porque NO HAY DIFERENCIA, por cuanto TODOS PECARON, y están destituidos de la gloria de Dios, siendo JUSTIFICADOS GRATUITAMENTE POR SU GRACIA, mediante la redención que es en Cristo Jesús” (Romanos 3:22-24) Y eso es lo que sucede con María y todo creyente…. “Porque así como en Adán TODOS MUEREN (María pasó por la muerte o dormición; consecuencia del pecado), también en Cristo todos serán vivificados.” (1Corintios 15:22). Por el contrario la expresión πλήρης χάριτος "pléres járitos" donde pléres = lleno y járitos = gracia, QUE SE APLICA A CRISTO (Juan 1:14) SI SIGNIFICA “LLENO DE GRACIA” = “Y aquel Verbo fue hecho carne, y habitó entre nosotros (y vimos su gloria, gloria como del unigénito del Padre), LLENO DE GRACIA y de verdad.” Esta fue también utilizada en Esteban: “Y Esteban, lleno de gracia y de poder, hacía grandes prodigios y señales entre el pueblo." (Hechos 6:8); y esto no significa que ahora Esteban sea sin pecado y demás… Algunos objetan que Eclesiástico (Sirac) y Lucas están escritos por diferentes autores y en tiempos distintos por lo que las misma palabra “kejaritoméne” es traducida en forma diferente; además que por el contexto una es tratada como adjetivo (Sirac) en Lucas como pronombre. Por último que el Latín refleja “mejor su real significado” = (“gracia plena” en Lucas)... A lo que debemos de decir que tanto el Sinaítico como el Vaticano son anteriores a la Vulgata; los 3 códices más antiguos Sinaítico, Vaticano y Alejandrino concuerdan y EL SEÑOR ESCOGIÓ POR ALGO EL GRIEGO y no el latín; por lo que eso de "reflejar mejor en Latín" es un mito. Pedro de Alejandría, que fallece en el 311 d.C (anterior a la Vulgata), cita Lucas 1:28 como “altamente favorecida” (Fragmentos de Escritos de Pedro V (9) De un Sermón). También Gregorio Tramaturgo que nació en Turquía (213 al 270 d.C.), y también anterior a Jerónimo (y la Vulgata 382 d.C – más de 100 años anterior) cita a Lucas 1:28 como “altamente favorecida”, en su Primera Homilía. Juan de Damasco (Damasceno) (675 – 749 d.C.) – en ‘Una exacta exposición de la fe ortodoxa’ Libro II cap. II., cita Lucas 1:28 como “altamente favorecida”. Lo curioso es aún Jerónimo (traductor de la Vulgata) también llega a citar Lucas 1:28 como “altamente favorecida”, en la Carta XXII a Eustoquio (38). ¿Qué más? Vayamos a versiones católico-romanas con su respectivo "imprimatur": La Biblia del Peregrino (católico romana): L. Alonso Schökel, traductor de la Biblia del Peregrino, profesor en el Pontifical Biblical Institute (Rome). Lucas 1:28 “Entró el ángel a donde estaba ella y le dijo: Alégrate, favorecida, el Señor está contigo.” Biblia Latinoamericana Edición XLI – Ediciones Paulinas, Verbo Divino (católico romana). “Entró el ángel a su casa y le dijo: 'alégrate tú, la Amada y Favorecida; el Señor está contigo.” Lucas 1:28 Y como se señaló antes Jerónimo llega a citar Lucas 1:28 en determinado momento: “altamente favorecida”. - Carta XXII a Eustoquio (38). Padres de la iglesia Gregorio Tramaturgo, Juan de Damasco = “altamente favorecida”. No hay evidencia de una “única” traducción que signifique impecabilidad, no estuvo en sus mentes; sino Jerónimo el mismo traductor de la Vulgata (que dicho sea de paso escribió un PRÓLOGO PROTECTOR y hoy no lo escuchan…) no hubiese traducido “kejaritoméne” como “altamente favorecida” en lugar de “llena eres de gracia” en otra oportunidad; por lo que vemos que para él las dos traducciones eran consideradas válidas. Tanto en Eclesiastés “kejaritoméne” como Lucas “kejaritoméno” se tiene LA MISMA DECLINACIÓN (participio, perfecto, pasivo), SÓLO VARÍAN EN EL GÉNERO. Lucas 1:28 es traducido tanto por los Padres de la Iglesia como por las versiones contemporáneas como PRONOMBRE (vocativo) y como ADJETIVO INDIFERENCIADAMENTE. Hemos visto que Jerónimo y algunos Padres tratan kejaritoméne también como adjetivo y no como pronombre. Sigamos examinando: Eclesiástico o Sirac 18:17 en la Nueva Vulgata traduce: “homine gratioso” = persona, hombre clemente, benevolente. Sin embargo, la Biblia Sacra juxta Vulgatam Clementinam 1598 traduce: “homine justificato” - “el hombre justificado” y la Biblia Sacra Iuxta Vulgatam Versionem (Vulgate Latin Bible), edited by R. Weber, B. Fischer, J. Gribomont, H.F.D. Sparks, and W. Thiele [at Beuron and Tuebingen] también “hombre justificado”. El contexto del verso nos refiere que “el regalo” y “las palabras” las tiene una persona al ser: Favorecido por Dios, justificado, lleno de la gracia de Dios, que es esencia lo que constituye una persona piadosa; ya que no es por sus méritos, sino por un don del Señor, dado antes de haber nacido. Vemos que de acuerdo al contexto la palabra tiene distintos significados pero no hay evidencia alguna en todos los casos para la peregrina idea de “impecabilidad”, lo cual es una grosera eiségesis. Las menciones de los Padres varían; y no se hacían problemas debido a que no existía la doctrina de la impecabilidad…; la prueba es que el mismo Jerónimo (y otros Padres) en determinado momento trata kejaritoméne como adjetivo, y lo traduce como “altamente favorecida”, demostrando que sí es una buena traducción; por lo tanto se trata de traducciones semejantes, y de ninguna manera es el “eslabón perdido” (bien perdido); esto se reafirma al ver el contexto de María en la Escrituras que puede revisarse en el presente documento. No es “necesario” que María sea sin pecado original o que luego de dar a luz nunca peque para asegurar la absoluta santidad de Cristo o declarar Su Divinidad. La Palabra de Dios al contrario nos presenta CLARAMENTE a otra María… una creyente, salva, redimida, sujeta a Cristo y sin protagonismo luego de empezar el ministerio de Jesús; y esto NO QUITA QUE SEA BIENAVENTURADA ENTRE TODAS LAS MUJERES. Mucha es la evidencia que se presenta; son MUCHAS las referencias y muy claras: 1. La Palabra muestra a una María AFECTADA por el pecado: Comencemos con Lucas 1:30 “Entonces el ángel le dijo: María, NO TEMAS, porque HAS HALLADO GRACIA delante de Dios.” (Lucas 1:30). Leemos que dice “NO TEMAS”… del griego: φοβέω = fobeo = imperativo, presente, voz media, 2da. P. singular = “no tengas en ti misma miedo”. Y es que María se había turbado. Lucas 1:29 y el ángel Gabriel explica que tipo de turbación tuvo: fobeo = temor. La PALABRA ES CLARA: “En el amor no hay temor (φόβος = fobos), sino que el perfecto amor echa fuera el temor; porque el temor lleva en sí castigo. De donde el que teme (φοβούμενος verbo, participio, presente, voz media, nominativo, masculino, singular de φοβέω), no ha sido perfeccionado en el amor.” (1Juan 4:18). María tuvo temor, al igual que José (Mateo 1:20), y es una prueba que la naturaleza de María estuvo afectada por el pecado, sino no hubiese temido. O cuando dice: “Cuando le vieron, se sorprendieron; y le dijo su madre: Hijo, ¿por qué nos has hecho así? He aquí, tu padre y yo te hemos buscado con ANGUSTIA. (Lucas 2:48) de ὀδυνώμενοι de ὀδυνάω (odunao), en realidad dice “siendo angustiados” (en agonía, ansiosos) ya que es un participio pasivo. Quiere decir que no tuvo paz. Todo ser humano, todo creyente necesita un SALVADOR DEL PECADO; y María dijo alto y fuerte: “Porque ha mirado la bajeza (ταπείνωσις = la humilde condición) de su sierva (esclava = doules);” (Lucas 1:48), “Y mi espíritu se regocija en Dios MI SALVADOR.” (Lucas 1:47). Salvador ¿de qué? ¡Si María mismo lo declara! ¿Cómo tendría María CONSCIENCIA de necesitar un SALVADOR si hubiese sido sin pecado? María por eso cuando envejeció fue enterrada; ¿por qué los cristianos de esa época la enterraron? Le llaman dormición y que luego fue resucitada y llevada al cielo… lo cual son TRADICIONES, pero igual… ya que TODO CRISTIANO antes de su 2da Venida DORMIRÁ y en su tiempo será resucitado; y lo que experimenta todo cristiano sin excepción puede llamársele “dormición”. Si María NO hubiese tenido EL VIRUS DEL PECADO NUNCA HUBIESE MUERTO O ENTRADO A UN ESTADO DE DORMICIÓN... ya que Adán y Eva antes de pecar NO MORÍAN, ERAN ETERNOS. Pero ella murió COMO TODOS y luego todos seremos resucitados y levantados hacia el Señor. La paga del pecado es la muerte... Ro.6:23. Pero hay mucho más…, si nos fijamos tanto ENOC COMO ELÍAS fueron llevados al cielo sin ver muerte, sin embargo ni judíos, ni cristianos les pedimos… “Caminó, pues, Enoc con Dios, y desapareció, porque LE LLEVÓ DIOS.” (Génesis 5:24) y “Por la fe Enoc fue traspuesto PARA NO VER MUERTE, y no fue hallado, porque lo traspuso Dios; y antes que fuese traspuesto, tuvo testimonio de haber agradado a Dios.” (Hebreos 11:5). “Y aconteció que yendo ellos y hablando, he aquí un carro de fuego con caballos de fuego apartó a los dos; y Elías subió al cielo en un torbellino.” (2Reyes 2:11)… sin embargo también dice: “Elías era hombre sujeto a pasiones semejantes a las nuestras, y oró fervientemente para que no lloviese, y no llovió sobre la tierra por tres años y seis meses.” (Santiago 5:17). Fueron personas sujetas al pecado, igual que todos, con excepción del postrer Adán (Cristo Jesús)… 2. María luego del parto esperó pase el tiempo DE SU PURIFICACIÓN y como todo pecador ofreció "Un par de tórtolas, o dos palominos" (que habían de ser SACRIFICADAS Y SU SANGRE DERRAMADA Lucas 2:22 y 24). “Y cuando se cumplieron los días DE LA PURIFICACIÓN DE ELLOS, conforme a la ley de Moisés, le trajeron a Jerusalén para presentarle al Señor (como está escrito en la ley del Señor: Todo varón que ABRIERE LA MATRIZ será llamado santo al Señor), Y para ofrecer conforme a lo que se dice en la ley del Señor: UN PAR DE TÓRTOLAS, O DOS PALOMINOS.” (Lucas 2:22-24). Entonces vemos que Lucas mismo menciona que JESÚS ABRE LA MATRIZ de María y es “llamado santo al Señor” (hemos citado Lucas 2:23), que María ofrece conforme a lo que se dice en la ley del Señor: UN PAR DE TÓRTOLAS, O DOS PALOMINOS.” (Lucas 2:24); esto es que derramaban sangre de las tórtolas, palomas o cordero (si tenían dinero) en lugar de la propia, se refiere a una justificación SUSTITUTORIA, de eso tratan los sacrificios en el contexto judío…, la sangre del cordero era derramada en lugar de la del pecador; que representa la JUSTIFICACIÓN por medio de Cristo Jesús. ES UN DERRAMAMIENTO DE SANGRE que fue hecho: "y sin derramamiento de sangre no se hace remisión de pecado". (Hebreos 9:22). Se están refiriendo a esta ley: “Habla a los hijos de Israel y diles: La mujer cuando conciba y dé a luz varón, será inmunda siete días; conforme a los días de su menstruación será inmunda. Y al octavo día se circuncidará al niño. Mas ella permanecerá TREINTA Y TRES DÍAS PURIFICÁNDOSE DE SU SANGRE; (Levítico 12:2-4). “Cuando los días de su purificación fueren cumplidos, por hijo o por hija, traerá un cordero de un año para holocausto, y un palomino o una tórtola para expiación, a la puerta del tabernáculo de reunión, al sacerdote; y él los ofrecerá delante del Señor, y HARÁ EXPIACIÓN POR ELLA, y SERÁ LIMPIA del flujo de su sangre. Ésta es la ley para la que diere a luz hijo o hija. Y si no tiene lo suficiente para un cordero, tomará entonces DOS TÓRTOLAS O DOS PALOMINOS, UNO PARA HOLOCAUSTO Y OTRO PARA EXPIACIÓN; y el sacerdote hará expiación por ella, y será limpia.” (Levítico 12:6-8). María OFRECIÓ LOS SACRIFICIOS DERRAMANDO SANGRE PARA HOLOCAUSTO Y PARA EXPIACIÓN; espero el tiempo indicado PARA PURIFICARSE. El propósito de que María haya sido VIRGEN en el momento de la CONCEPCIÓN por medio del ESPÍRITU SANTO, SIN INTERVENCIÓN DE JOSÉ, es porque Cristo es EL HIJO DE DIOS, es decir el VERBO HECHO CARNE = DIOS. Esto es que María era virgen como una garantía de QUE NO HAYA HABIDO INTERVENCIÓN HUMANA. Por eso también el apóstol Pablo nos enseña que Dios envió a su Hijo, “NACIDO DE MUJER Y NACIDO BAJO LA LEY” Gálatas.4:4 (aquí María aparece como parte del antiguo pacto; Jesucristo vino a redimir a aquellos que estaban bajo la ley). 3. Después de las Bodas de Caná, esto es al comenzar su ministerio, María es casi ignorada. Veremos cuál es el lugar en que es colocada: Cuando María pidió a Jesucristo que intervenga en las bodas de Caná, Jesús se dirige a ella diciéndole: “¿Qué tienes conmigo mujer?” (Comparar lo dicho por el endemoniado gadareno en Marcos 5:7 “Qué tienes conmigo, Jesús,...”); notemos además que no le llama madre sino mujer. Esta expresión en griego (τί ἐμοὶ καὶ σοί = ¿qué tienes conmigo…?) ha sido utilizada sólo en estos pasajes: Jueces 11:12, 1 Reyes 17:18; 2 Reyes 3:13; 2 Crónicas 35:21; Marcos 5:7; Lucas 8:28; Juan 2:4; en todas tiene un sentido de DISTANCIA o REPRENSIÓN. Igualmente si vemos la expresión “qué tienes con nosotros” en Mateo 8:29; Marcos 1:24 y Lucas 4:34 que profieren los demonios para expresar la distancia entre ellos y EL SEÑOR: “¿qué tienes con nosotros, Jesús Hijo de Dios? ¿Has venido acá para atormentarnos antes de tiempo?” o “¿Has venido para destruirnos?”. No se puede sacar la expresión del contexto en que es utilizada tanto en el Nuevo Testamento como en el Antiguo. Tertuliano anota, refiriéndose a este pasaje, en la Homilìa XXI (2 y 3), página 73 de Philip Schaff, D.D., LL.D. Nicene and Post Nicene Fathers of the Christian Church. - The Christian Literature Company 1890-1900. New York. Al final de 2: “Y por lo tanto, Él respondió así en este lugar, y de nuevo en otra parte, ‘¿Quién es mi madre, y quiénes son mis hermanos?’ (Mateo 12:48), porque todavía no pensaban correctamente de Él; y ella, porque lo había parido, afirmó, de acuerdo con la costumbre de otras madres, dirigirlo en todas las cosas, cuando debería haberlo reverenciado y adorado. Esta fue la razón por la que Él respondió como lo hizo en esa ocasión. Para considerar qué cosa era, que cuando todas las personas (alto y bajo) estaban de pie a su alrededor, cuando la multitud estaba tratando de escucharlo, y su doctrina había comenzado a ser presentada,... ...Porque si Él se preocupaba por los demás, y usaba todos los medios para implantar en ellos una opinión de sí mismo, mucho más lo haría en el caso de su madre... ...pero en todos los casos habría reclamado la superioridad como su madre, por lo tanto, respondió como lo hizo a los que le hablaron; de lo contrario, no podría haber guiado sus pensamientos (actuales de María sobre Jesús) de Su baja humildad actual a Su futura exaltación, si hubiera esperado que siempre fuera honrada por Él como si fuera un hijo, y no que Él debería venir como su Maestro. [3.] Fue por este motivo que Él dijo en este lugar: ‘Mujer, ¿qué tengo que ver contigo?’ y también por otra razón no menos apremiante. ¿Cuál fue esta razón? Fue, que Sus milagros podrían no ser puestos en prueba. LA SOLICITUD DEBERÍA HABER VENIDO DE AQUELLOS QUE LO NECESITABAN, NO DE SU MADRE.” Tertuliano afirma, que Jesús estaba por un lado guiando, preparando los pensamientos de María hacia Su futuro estado: con cuerpo GLORIFICADO; pero por otro lado buscaba enseñar que ella NO ES INTERMEDIARIA, que la petición debe ser DIRECTA a ÉL. María entendió lo dicho por Jesucristo y SE SUJETÓ diciendo: “HACED TODO LO QUE OS DIJERE” Juan 2:4-5; así también nosotros hacemos todo lo que Jesucristo nos dice DIRECTAMENTE y Él nos ordena EXPRESAMENTE que PIDAMOS EN SU NOMBRE; y no lo dice 1 sola vez, sino que lo repite una y otra vez, ver: Juan 16:24 “Hasta ahora nada habéis pedido mi nombre; pedid y recibiréis para que vuestro gozo sea cumplido”, o en Juan 16:26 “EN AQUEL DÍA PEDIRÉIS EN MI NOMBRE”. Y lo RECALCA MUCHAS VECES MÁS: Juan 14:13; Juan 14:14; Juan 15:16; Juan 16:23. • Juan 14:13 "y todo lo que pidiereis al Padre EN MI NOMBRE, lo haré, para que el Padre sea glorificado en el Hijo”. • Juan 14:14 “Si algo pidiereis EN MI NOMBRE, yo lo haré”. • Juan 15:16 “...para que todo lo que pidiereis al Padre EN MI NOMBRE, Él os lo dé”. • Juan 16:23 “De cierto, de cierto os digo, que todo cuanto pidiereis al Padre EN MI NOMBRE, os lo dará”. • Juan 16:24 “Hasta ahora nada habéis pedido EN MI NOMBRE, pedid y recibiréis, para que vuestro gozo sea cumplido”. • Juan 16:26 “en aquel día pediréis EN MI NOMBRE...”. Pero no sólo lo enseña Cristo, sino que el apóstol Pablo así lo entendió y lo enseñó a la iglesia y lo escribió de forma tan clara como para que todos lo entiendan sin dudas: “y TODO LO QUE HACÉIS; sea de palabra o de hecho, hacedlo TODO EN EL NOMBRE DEL SEÑOR JESÚS; dando gracias a Dios Padre por medio de Él” Colosenses 3:17. “y TODO (no sólo algunas cosas, TODO) LO QUE HACÉIS; sea de palabra o de hecho (es decir TODO), hacedlo TODO (se repite la palabra en el griego, πάντα = todo) EN EL NOMBRE DEL SEÑOR JESÚS; dando gracias a Dios Padre por medio de Él (aun las acciones de gracias, esto es TODO son dadas en el nombre de Jesús)” Colosenses 3:17. ¿Cuál es la parte que no se entiende? ¿No queda claro? No hay lugar en que se diga algo semejante sobre María… María en las Bodas de Caná cuando es confrontada por Cristo se rinde y SE SUJETA y dice: “HACED TODO LO QUE ÉL OS DIJERE”. Veamos lo que entendieron los Padres de la Iglesia sobre estos pasajes: Comencemos con Juan Crisóstomo, obispo de Constantinopla (347-407 d.C.), que al referirse a Colosenses 3:17 la llama LEY APOSTÓLICA; es decir que es UNA ENSEÑANZA FUNDAMENTAL OBLIGATORIA A SER GUARDADA: Crisóstomo. Homlilía LV (7) sobre Mateo 16:24. Schaff (pág. 328): "Bendito Dios". Por la LEY APOSTÓLICA que cumplen de inmediato, QUE ORDENA: "Todo lo que hagamos en palabra o en hecho, lo haremos en el nombre de nuestro Señor Jesucristo, dando gracias a Dios y al Padre por medio de Él". Sigamos con otra cita de Juan Crisóstomo (347-407 d.C.), que hace una detallada explicación sobre el texto: Las Homilías de Juan Crisóstomo sobre Colosenses. Homilía IX (Colosenses 3:16,17). Schaff. Pág.302, 303: “…y todo lo que hacéis; sea de palabra o de hecho, hacedlo todo en el nombre del Señor Jesús; dando gracias a Dios Padre por medio de Él”. “Porque si así lo hacemos, no habrá nada contaminado, nada sucio, dondequiera QUE CRISTO SEA LLAMADO. Si comes, si bebes, si te casas, si viajas, HAZLO TODO en el Nombre de Dios, es decir, INVOCÁNDOLO para que te ayude: en todo PRIMERO ORANDO, y así tome control de tus asuntos. ¿Vas a hablar un poco? Pon esto al frente (delante). Por esta causa también colocamos frente a (delante de) nuestras epístolas (cartas) el Nombre del Señor. Dondequiera que esté el Nombre de Dios, todo es auspicioso. Porque si los nombres de los cónsules aseguran los escritos, MUCHO MÁS LO SERÁ EL NOMBRE DE CRISTO. O Él quiere decir esto: Después de que Dios te diga algo y hagas todo, no introduzcas a los ángeles además. ¿Comes? Da gracias a Dios antes y después. ¿Duermes? Da gracias a Dios antes y después. ¿Te diriges al mercado? Haz lo mismo - nada mundano, nada de esta vida. HAZ TODO EN EL NOMBRE DEL SEÑOR, y todo será prosperado en ti. Dondequiera que se coloque EL NOMBRE, todas las cosas son auspiciosas. Si esto expulsa a los demonios, si elimina las enfermedades, mucho más hará que las actividades sean fáciles.” ¿Y qué es "hacer en palabra o en hecho"? Ya sea solicitando o realizando cualquier cosa… INVOCA AL HIJO, da gracias al Padre. PORQUE CUANDO SE INVOCA AL HIJO, SE INVOCA AL PADRE, y cuando se le agradece, se le ha agradecido al Hijo. (y sigue en este sentido) Veamos una cita de Atanasio (296-373): Atanasio contra los arrianos. Cuatro discursos contra los arrianos. Discurso III (12); [o también como Capítulo XXIII] . Schaff. Página 400. “Nadie, por ejemplo, oraría para recibir de Dios y los ángeles, o de cualquier otra criatura, ni nadie diría: "Que Dios y el ángel te den"; sino del Padre y del Hijo, debido a su unidad y la unidad de Su ofrenda. PORQUE POR MEDIO DEL HIJO SE DA LO QUE SE DA; y NO HAY NADA MÁS QUE EL PADRE QUE OPERA A TRAVÉS DEL HIJO; porque es GRACIA SEGURA al que lo recibe. Es clarísima la afirmación Orígenes; en Orígenes contra Celso. Libro V. Capítulos 4 y 5. Schaff. Pág. 544, dice: "TODA SÚPLICA, ORACIÓN, INTERCESIÓN Y ACCIÓN DE GRACIAS deben dirigirse al Dios Supremo, A TRAVÉS DEL SUMO SACERDOTE que está por encima de todos los ángeles, LA PALABRA VIVA Y DIOS… Capítulo 5. Porque invocar a los ángeles sin haber obtenido un conocimiento de su mayor naturaleza que la que poseen los hombres, sería contrario a la razón. Ambrosio (340 – 397 d.C.): Tratados Dogmáticos, Trabajos Èticos y Sermones. Exposición de la fe cristiana. Libro I Capítulo II (12). Schaff. 203. “Entonces, con espíritu fiel y mente devota, INVOQUEMOS A JESÚS nuestro Señor, creamos que Él es Dios, hasta el fin de que todo lo que pidamos al Padre, LO OBTENGAMOS EN SU NOMBRE. PORQUE LA VOLUNTAD DEL PADRE ES QUE ÉL SEA IMPLORADO A TRAVÉS DEL HIJO, el Hijo que el Padre sea implorado.” Cipriano, Obispo de Cartago (200 – 258 d.C.) en: Los Tratados de Cipriano. Tratato IV. Sobre la oración del Señor. (5) Schaff. Página 449, nos insta a orar de acuerdo al modelo de oración que Jesús nos enseñó y hacerlo en su Nombre: “Por lo tanto, hermanos, amados, oremos como Dios nuestro Maestro nos ha enseñado. Es una oración amorosa y amistosa para implorar a Dios con su propia palabra, para que llegue a sus oídos en la oración de Cristo. Deje que el Padre reconozca las palabras de Su Hijo cuando hacemos nuestra oración, y QUE EL QUE MORA EN NUESTRO PECHO, MORA EN NUESTRA VOZ. Y ya que LO TENEMOS COMO UN ABOGADO ANTE EL PADRE por nuestros pecados, permítanos, cuando como pecadores PEDIMOS POR NUESTROS PECADOS, expongamos las palabras de NUESTRO ABOGADO. Porque COMO ÉL DICE, “TODO LO QUE PIDAMOS AL PADRE EN SU NOMBRE, ÉL NOS DARÁ” ¡cuánto más efectivamente obtenemos lo que pedimos EN EL NOMBRE DE CRISTO, si lo pedimos en su propia oración! Y de aquí en adelante veremos que Cristo y los apóstoles enseñan a NO PONER A MARÍA EN UN LUGAR QUE NO LE CORRESPONDE… sigamos viendo. Ahora, no por ser Dios, Cristo dejó de sujetarse… fue obediente: “El cual, siendo en forma de Dios, no estimó el ser igual a Dios como cosa a que aferrarse, sino que se despojó a sí mismo, tomando forma de siervo, hecho semejante a los hombres; y estando en la condición de hombre, se humilló a sí mismo, HACIÉNDOSE OBEDIENTE HASTA LA MUERTE, Y MUERTE DE CRUZ. (Filipenses 2:6-8). Por eso HASTA los 30 años, que es la edad ceremonial para el servicio ministerial (Números 4:3, 47), Jesús no comienza su ministerio y estuvo sujeto a sus padres; E HIZO RECIÉN SU PRIMER MILAGRO ENTONCES (“este PRINCIPIO DE SEÑALES hizo Jesús en Caná de Galilea, y manifestó su gloria; y sus discípulos creyeron en Él.” Juan 2:11); por eso también tributó al César (Lucas 17:27,27). 4. Cristo enseñó varias veces el LUGAR de María, anticipándose a la tendencia natural que tiene el ser humano en exaltar a la madre: Veamos cómo en Marcos 3:21 se menciona que cuando se agolpó la gente en torno de Jesús “LOS SUYOS, VINIERON PARA PRENDERLE, PORQUE DECÍAN: ESTÁ FUERA DE SÍ”. En este capítulo, Marcos 3, vemos cómo la oposición contra Jesús aumenta drásticamente. Y es en ese contexto que los SUYOS fueron a PRENDERLO y luego unos versos más delante de ese mismo capítulo 3 de Marcos nuevamente menciona que su MADRE Y HERMANOS fueron a buscarlo, y por el contexto es para prenderle, para decirle que pare, pero CLARAMENTE Cristo los frena y no sale al llamado de su madre y dice su famosa expresión que su “MADRE Y HERMANOS son LOS QUE HACEN SU VOLUNTAD”. Nuevamente no podemos sacar los pasajes de su contexto: “Vienen después sus HERMANOS Y SU MADRE, y QUEDÁNDOSE AFUERA, ENVIARON A LLAMARLE. Y la gente que estaba sentada alrededor de él le dijo: Tu MADRE Y TUS HERMANOS ESTÁN AFUERA, y te buscan. Él les respondió diciendo: ¿QUIÉN ES MI MADRE Y MIS HERMANOS? Y mirando a los que estaban sentados alrededor de él, dijo: HE AQUÍ MI MADRE Y MIS HERMANOS. Porque todo aquel que hace la voluntad de Dios, ÉSE ES MI HERMANO, Y MI HERMANA, Y MI MADRE.” (Marcos 3:31-35). Por el contexto seguramente su madre y hermanos querían aconsejarle, prenderlo, advertirle o hacerlo desistir (Marcos 3:21 “LOS SUYOS, VINIERON PARA PRENDERLE, PORQUE DECÍAN: ESTÁ FUERA DE SÍ”.); pero Jesucristo aprovecha para enseñar a sus discípulos cuál es el orden; y no salió, ni con María y sus hermanos afuera. Pero Jesús sigue enseñando el lugar de María, veamos: 5. Es muy significativa la contestación de Jesús a las alabanzas hacia su madre (esa es la tendencia natural, ya que todos tenemos o hemos tenido madres): “Mientras Él (Cristo) decía estas cosas, una mujer de entre la multitud levantó la voz y le dijo: Bienaventurado el vientre que te trajo, y los senos que mamaste. Y Él dijo: Antes bienaventurados los que oyen la palabra de Dios, y la guardan.” Lucas 11:27-28 El énfasis es el OÍR SU PALABRA y PONERLA POR OBRA… y ya sabemos que Cristo ordenó pedir en Su Nombre. Pero veamos qué más nos enseña el apóstol Pablo: 6. El apóstol hace una comparación entre Melquisedec y Cristo y es indudable que habla de características comunes a AMBOS: “Porque este Melquisedec, ‘rey de Salem’, ‘sacerdote’ del Dios Altísimo, que salió a recibir a Abraham que volvía de la derrota de los reyes, y le bendijo, a quien asimismo dio Abraham los diezmos de todo; cuyo nombre significa primeramente ‘Rey de justicia’, y también ‘Rey de Salem’, esto es, ‘Rey de paz’; ‘sin padre, sin madre’, ‘sin genealogía’; que ‘ni tiene principio de días, ni fin de vida’, SINO HECHO SEMEJANTE AL HIJO DE DIOS, PERMANECE SACERDOTE PARA SIEMPRE.” Hebreos 7:1-3. Son características comunes entre Mequisedec y Cristo… ambos son reyes (de paz y justicia), sacerdote, ambos no tienen padre, NI MADRE, sin genealogía (descendencia), eterno… por eso es HECHO SEMEJANTE AL HIJO DE DIOS, esto es semejante a Cristo Jesús. Y una de las semejanzas es SIN MADRE; en la mente del apóstol no existe esa acepción de personas que hoy hacen con María… ¿cómo podría el apóstol hace semejante afirmación, si en ese tiempo fuesen conscientes de una veneración a María? Claro que la reconocían como madre de Jesús pero el apóstol define que como HIJO DE DIOS (que es un título Divino; ver al final del versículo) NO TIENE MADRE; es decir que como Dios no tiene madre. Cristo Jesús es al mismo tiempo perfecto hombre y perfecto Dios, y aunque es uno sólo como hombre tiene una madre = María, más como Dios NO TIENE MADRE. Y esto lo dice justamente el apóstol Pablo quien, como hemos visto, ordena a la iglesia el HACER TODO EN EL NOMBRE DE JESÚS (y dice: todo), aun las acciones de gracias, examine Colosenses 3:17. Entonces vemos cómo el apóstol escribe GUIADO POR EL ESPÍRITU SANTO: “sin padre (HUMANO = para el “Hijo de Dios” = Dios, es decir no de José, sino su padre es DIOS), SIN MADRE (humana = para el “Hijo de Dios” = Dios, es decir que su madre no es MARÍA, ya que EL HIJO DE DIOS = Dios no tiene madre, María es la madre de Jesús). Es por ESO que Jesús no llama madre a María sino MUJER al comenzar su ministerio para PROBAR A TODOS QUE VERDADERAMENTE ES EL HIJO DE DIOS, que es cuando comienza Su obrar Sobrenatural, ya que la Paternidad es de lo Alto; por eso SÍ DICE: “PADRE, EN TUS MANOS ENCOMIENDO MI ESPÍRITU”, porque es EL HIJO DE DIOS y tiene SU MISMA NATURALEZA. Por eso cuando confesaba que es el “Hijo de Dios” buscaban matarle y fue su acusación ante Pilato para crucificarle: • “Los judíos le respondieron: Nosotros tenemos una ley, y según nuestra ley debe morir, porque se hizo a sí mismo Hijo de Dios. Cuando Pilato oyó decir esto, tuvo más miedo. Y entró otra vez en el pretorio, y dijo a Jesús: ¿De dónde eres tú? (Juan 19:7-9). • También procuraron matarle cuando dijo que Dios era su “Padre”, (como consecuencia = “Hijo de Dios”): “Y Jesús les respondió: Mi Padre hasta ahora trabaja, y Yo trabajo. Por esto los judíos aún más procuraban matarle, porque no sólo quebrantaba el día de reposo, sino que también decía que Dios era su propio Padre, haciéndose igual a Dios. (esto es = “Hijo de Dios”) (Juan 5:17-18) • Cuando a partir de Juan 10:29 les dice Jesús varias veces que Su Padre era Dios vrs. 29, 30, 32; tratan de matarlo: “Entonces los judíos volvieron a tomar piedras para apedrearle. Jesús les respondió: Muchas buenas obras os he mostrado de mi Padre; ¿por cuál de ellas me apedreáis? Le respondieron los judíos, diciendo: Por buena obra no te apedreamos, sino por la blasfemia; porque tú, siendo hombre, te haces Dios. (Juan 10:31-33) Hay muchos, muchos pasajes en este sentido, en cuanto al título “Hijo de Dios” y su significado de Divinidad; ver el documento sobre Deidad de Cristo. 7. El llamar a María “esposa del Espíritu Santo” y “madre de Dios”, son excesos que llevan a sobredimensionarla y si se agrega la equivocación que hemos visto que es “sin pecado”…, se llega a un culto semejante a la desviación del mismo Israel en el Antiguo Testamento: "Los hijos recogen la leña, los padres encienden el fuego, y las mujeres amasan la masa, para hacer tortas A LA REINA DEL CIELO y para hacer ofrendas a dioses ajenos, para provocarme a ira." (Jeremías 7:18). Israel se oponía a Jeremías; vemos cuán perniciosa es esta práctica: "Respondieron a Jeremías, diciendo: La palabra que NOS HAS HABLADO EN NOMBRE DEL SEÑOR, NO LA OIREMOS DE TI; sino que ciertamente pondremos por obra toda palabra que ha salido de NUESTRA boca, para ofrecer incienso A LA REINA DEL CIELO, derramándole libaciones, COMO HEMOS HECHO NOSOTROS Y NUESTROS PADRES, NUESTROS REYES Y NUESTROS PRÍNCIPES (su tradición), en las ciudades de Judá y en las plazas de Jerusalén, y tuvimos abundancia de pan, y estuvimos alegres, y no vimos mal alguno (señales engañosas para los que se pierden). Mas desde que dejamos de ofrecer incienso a la reina del cielo y de derramarle libaciones, nos falta todo, y a espada y de hambre somos consumidos." (Jeremías 44:15-18) Esta práctica lamentablemente se ha “cristianizado”, e incorporado en María. Vemos que esos falsos cultos (a demonios) les daban su regalito (abundancia de pan, alegría); no les importaba la verdad sino el “estar bien”; ¿será toda coincidencia casual? Pero, sigamos: La IGLESIA (incluyendo a María) es la ESPOSA de Cristo (ver Efesios 5:23 al 32); todos los que nacen de nuevo son hijos de Dios (incluida María): "Mas a todos los que le recibieron, a los que creen en su nombre, les dio potestad de ser hechos hijos de Dios" (Juan 1:12). La declaración "madre de Dios" fue dada en su época (ya tardíamente por el año 431 d.C.) para resaltar la NATURALEZA DIVINA E INDIVISIBLE DE CRISTO y no de María; que es engendrado por el Espíritu Santo. En ese sentido los cristianos somos “hijos de Dios”, para resaltar Su naturaleza divina; es una declaración Cristocéntrica; nosotros somos pecadores redimidos. Por el hecho de confesar que somos hijos de Dios, no se va a crear toda una teología defendiendo el nacimiento sin pecado de los hombres, aunque ahora somos santos en Cristo; lo mismo se aplica a María. Es cierto que las naturalezas humana y divina de Cristo son indivisibles; Cristo es el Verbo hecho carne, sin embargo, eso no deifica a María. En Cristo no se puede aplicar ese RAZONAMIENTO HUMANO… “Como Sus Naturalezas son indivisibles… María no sólo es madre de Jesús sino de Dios mismo”. PUES NO; ya que Cristo fue engendrado por el Espíritu Santo; eso es un razonamiento Y FILOSOFÍA HUMANAS: “Mirad que nadie os engañe por medio de filosofías y huecas sutilezas, según las tradiciones de los hombres, CONFORME A LOS RUDIMENTOS DEL MUNDO, y NO SEGÚN CRISTO. Porque en Él habita corporalmente toda la plenitud de la Deidad, y VOSOTROS ESTÁIS COMPLETOS EN ÉL, QUE ES LA CABEZA de todo principado y potestad. (Colosenses 2:8-10) Es entre las distintas culturas paganas que aparece una madre de Dios, intercesora, sobre-exaltada; pero en la Palabra no es así… se exalta sólo a Cristo, que es el Hijo de Dios (Dios Padre), el Verbo encarnado, Dios Todopoderoso. La realidad es que ESTAMOS COMPLETOS YA EN CRISTO. María no es esposa del Espíritu Santo; María CONCIBIÓ DEL ESPÍRITU SANTO que es muy distinto; el Espíritu NO SE CASA en las Escrituras, sino EL HIJO… el Rey. Jesucristo, el Hijo de Dios es el que se casa con la iglesia; y en LAS BODAS DEL CORDERO están la iglesia y María. Cristo, no es un “bastardo”, porque tiene un Padre bien conocido, que es Dios mismo… Los espíritus no se casan; los ángeles no tienen sexo, y Dios es UNA TRINIDAD. Al hacerla esposa de Dios crean una (propongo el término…) “tetranidad”; y peor aún llegaría a ser (vamos a acuñar otro término…) “pentanidad” (con la adoración del pan). Entonces para ellos además de que Dios se hizo ser humano hay otro estado..., se convierte es ser inanimado, “pan y vino” (cuando el mismo doctor de su iglesia, san Agustín, dice que es la PRIMERA HEREJÍA, y la llama así, herejía… (leer, varias citas de él al respecto con su respectiva fuente, fue publicada en mi muro), y convierten a María en la práctica y por sus oficios en diosa, ya que si es “esposa del Espíritu Santo”, pasaría a ese nivel; POR ESO EN LA PRÁCTICA, terminan pidiendo a través de ella y le dan títulos que son sólo del Rey de Reyes. Por eso la iglesia es esposa de Cristo que se hizo hombre. A eso se debe tanto afán por decir que es sin pecado, es decir que nunca pecó, que fue virgen aún durante y después del parto, no tuvo hijos, que no pudo morir… que es reina de los cielos, la intercesora… Ese cromosoma X lo creó el Espíritu Santo, no lo tenía María. Por eso es HIJO DE DIOS; DIOS PADRE NO TIENE SEXO, ES ESPÍRITU; la naturaleza DIVINA ÚNICAMENTE proviene de Dios Padre; la PATERNIDAD, ASCENDENCIA DIVINA (de ese Cristo indivisible), es del Padre mismo. La Palabra de Dios NOS MUESTRA CÓMO INTERPRETARLA. En Efesios se hace claramente la comparación (ver: Efesios 5:23 al 32): Así como EL ESPOSO A LA ESPOSA ES CRISTO A LA IGLESIA... Igualmente en los evangelios LA ESPOSA ES LA IGLESIA... VER SINO quienes son las VÍRGENES que esperan la venida del ESPOSO...en Mateo 25. Y otra vez cuando dijo: "Mas vendrán días cuando el esposo les será quitado; entonces, en aquellos días ayunarán." (Lucas 5:35). ¿Acaso en Cantar de los Cantares está hablando de una MADRE que es esperada por un esposo?, o a una ESPOSA. Es una relación esposo – esposa… no madre – hijo. Tampoco es la relación de María con el Espíritu Santo…, YA QUE ese esposo es el rey Salomón, que es figura DEL REY DE REYES, CRISTO: "¡Oh, si él me besara con besos de su boca!" (Cantares 1:2) o "El rey me ha metido en sus cámaras;" (Cantares 1:4) - Vemos una relación esposo – esposa. Cantar de los Cantares es la relación del Rey con su ESPOSA, no con su mamá como algunos afirman. Y la Palabra y Cristo mismo son claros cuando llaman a la iglesia ESPOSA y VÍRGEN (10 vírgenes), que espera a su Esposo Cristo. María como parte de la iglesia, y SÓLO EN ESE SENTIDO, así como todo aquel que ha nacido de nuevo y es templo del Hijo de Dios y es SU ESPOSA, ES VÍRGEN, y TEMPLO... o ¿no se dan cuenta de que el creyente es TEMPLO DE CRISTO? ¿No es "Cristo en nosotros", la esperanza de Gloria? ("...que es Cristo en vosotros, la esperanza de gloria," Col 1:27). Es tremendo el hecho de que Dios, el TodoPoderoso MORE EN EL CREYENTE. A su verdadera iglesia se le llama ESPOSA Y VIRGEN (María, que es salva en ESE SENTIDO TAMBIÉN) Y María fue virgen sí o sí hasta el parto eso lo dice los evangelios y bien claro: “Y despertando José del sueño, hizo como el ángel del Señor le había mandado, y recibió a su mujer. Pero NO LA CONOCIÓ HASTA QUE dio a luz a su hijo; y le puso por nombre JESÚS. (Mateo 1:24-25) Y que creen que significa que José “recibió a su mujer”, pero que “NO LA CONOCIÓ HASTA QUE” dio a luz a su hijo; ¿qué podrá ser? ¿Es tan, tan difícil? ¡Por favor dejen ya las tradiciones y fariseísmos! Vamos a explicarlo: Lo que dice es que José luego de la explicación dada por el ángel ya “que quiso dejarla secretamente” (Mateo 1:19), recibió a María (su mujer), pero que “NO LA CONOCIÓ” es decir no tuvo SEXO con ella “HASTA QUE” dio a luz a su hijo. Ponemos el griego: καὶ (y) οὐκ (no) ἐγίνωσκεν (conoció) αὐτὴν (a ella) ἕως (hasta que) ἔτεκεν (dio a luz) υἱόν (al hijo) (Mateo 1:25); en griego están todas y cada una de las palabras y hablo de los códices Vaticano y Sinaítico del año 350 d.C.; entonces dice: “y no la CONOCIÓ HASTA QUE dio a luz al hijo”. La palabra ἐγίνωσκεν (egínosken de ginosko = γινώσκω) “conoció” se refiere al acto sexual, sino ver: “Conoció Adán a su mujer Eva, la cual concibió y dio a luz a Caín, y dijo: Por voluntad de YHVH he adquirido varón.” (Génesis 4:1), escrito en hebreo (yadá = conoció) pero traducido en la Septuaginta (LXX) como (egneo de ginosko = γινώσκω). Es evidente lo que significa cuando dice que Adán CONOCIÓ a su mujer la cual concibió y dio a luz a Caín, ¡pero vamos! Y así se repite la expresión en: Génesis 4:17 “y conoció Caín a su mujer la cual concibió”; Génesis 4:25 “Y conoció de nuevo Adán a su mujer la cual dio a luz un hijo, y llamó su nombre Set” (qué, ¿tenía alzhéimer?; ¿no?, entonces ya sabe); y así sigue…Génesis 38:26, Jueces 11:39, 1Reyes 1:4. Algunos argumentan el verbo “conocer” debería de haber estado en aoristo para indicar que fue un hecho puntual en el pasado y que luego María pasa a otro estado; pero el hecho de estar el verbo en imperfecto indica que la acción de “no la conoció” permanece constante, de tal forma que María fue siempre virgen. Pero se equivocan; es verdad que se puede utilizar el aoristo, pero esto no es exclusivo, ya que encontramos ejemplos en las Escrituras en las que se utiliza el IMPERFECTO y luego hay un cambio de estado (como en Mateo 1:25), veamos: "Así se quedó Jacob solo; y luchó con él un varón hasta que rayaba el alba." (Gen 32:24) - donde "luchó" = ἐπάλαιεν, está en imperfecto; y es EVIDENTE que luego DEJÓ DE LUCHAR. "Cuando venía Moisés delante de Jehová para hablar con él, se quitaba el velo hasta que salía; y saliendo, decía a los hijos de Israel lo que le era mandado." (Éxodo 34:34) - quitaba = περιῃρεῖτο = es imperfecto... y LUEGO SE VOLVÍA A PONER EL VELO. La figura que se ve en la Palabra y SE ENSEÑA UNA Y OTRA VEZ, explicando su significado, es la relación del esposo - esposa comparada con Cristo y su iglesia. Efesios capítulo 5 nos muestra que cómo relación NATURAL del ESPOSO con su ESPOSA se compara con la relación de CRISTO Y LA IGLESIA. Cantares justo muestra ESO MISMO... la relación de Salomón con su esposa; esto, tiene una primera interpretación del amor ENTRE ESPOSOS; y cantidad de pasajes hay que muestran el amor EROS.... (no se puede sacar el contexto histórico para otros fines). Pero hemos visto que esto no queda ALLÍ, también REPRESENTA LA RELACIÓN DE CRISTO CON SU IGLESIA... y eso es más que evidente para alguien que tenga sentido común espiritual: Salomón es UN REY. ¿Y QUIÉN MÁS ES EL REY SINO JESUCRISTO?, ¿No es claro y natural?, fluye directamente del texto mismo. Otra desviación se comete al poner a María como “el modelo” de la iglesia… Pero vemos, CÓMO CRISTO CORREGÍA EL QUE EXALTEN A MARÍA… lo que muestra que no es el modelo…, cuando vemos los discípulos Y APÓSTOLES SIGUIERON DIRECTAMENTE A JESÚS. El apóstol nos enseña: Sed imitadores de mí, así COMO YO DE CRISTO. (1Corintios 11:1) ¿Quién es entonces el modelo de la iglesia para el apóstol? Dice: “Puestos los OJOS EN JESÚS, el autor y consumador de la fe” (Hebreos 12:2). 8. Cristo es nuestro INTERCESOR EXCLUSIVO: Hebreos 7:25 dice: “por lo cual puede también salvar perpetuamente a los que por Él se acercan a Dios, VIVIENDO SIEMPRE PARA INTERCEDER POR ELLOS”. Romanos 8:34 “¿Quién es el que condenará? Cristo es el que murió; más aun, el que también resucitó, el que además está a la diestra de Dios, el que también INTERCEDE POR NOSOTROS”. Si tenemos libertad para entrar AL LUGAR SANTÍSIMO (Hebreos 10:19) y sentarnos en lugares celestiales con JESUCRISTO nuestro Sumo Sacerdote (Efesios2:6; 1:20); aún más, si somos templo de Dios (mora en nosotros), ¿para qué más intercesores (los santos y María) teniendo tanta excelencia…? Buscar otro intercesor es una afrenta a Dios. Además de ser TROPIEZO A MUSULMANES Y JUDÍOS (más de 1600 millones) que se dirigen directamente al Padre… Pero si nos damos cuenta el cristiano verdadero TAMBIÉN SE DIRIJE DIRECTAMENTE AL PADRE, ya que Cristo es Dios (hay otros documentos que explican esos puntos con más detalle y profundidad). 9. Otro pasaje que sacan del su contexto para decir que se refiere a María, es la mujer de Apocalipsis 12: Pero es FALSO que la mujer que aparece en Apocalipsis 12 sea María: “Apareció en el cielo una gran señal: una mujer vestida del sol, con la luna debajo de sus pies, y sobre su cabeza una corona de doce estrellas.” (Apocalipsis 12:1) Y lo explicaremos yendo simplemente a su contexto: • Las estrellas aluden a LAS 12 TRIBUS DE ISRAEL, sino ver el sueño de José (Génesis 37:9). La mujer es ISRAEL, que en ESA ÉPOCA (de la gran tribulación, que es EL CONTEXTO DEL TEXTO de Apocalipsis 12) SE CONVERTIRÁ RECIÉN, = parto = nacer de nuevo. Por eso huyen al desierto…v.6, que fue el consejo que dio Cristo a ISRAEL en Mateo 24:16 (nótese que Cristo daba instrucciones A LOS JUDÍOS). Todo esto acontece al final de los tiempos, luego de que el falso profeta y el anticristo hayan engañado a las naciones y hayan hecho esa PAZ MUNDIAL SUPERFICIAL, BASADA EN ALIANZAS HUMANAS (Daniel 2:43) y sabemos por Daniel 2 que ese último imperio basado en estas alianzas humanas será destruido por LA PIEDRA (Cristo), QUE LA DESMENUZARÁ…; Israel creerá (estará de PARTO) LUEGO que hayan construido ese 3er templo (y los judíos ya tienen todo preparado); LUEGO del pacto de 7 años, roto en la mitad del período (Apocalipsis 11:2,3; 12:6; 13:5); LUEGO que el anticristo se siente en ese TEMPLO y exija ADORACIÓN… (“tanto que se sienta en el templo de Dios como Dios, haciéndose pasar por Dios” 2Tesalonisenses 2:4), cosa que los judíos no aceptarán…. Es ENTONCES cuando entenderán y lo llorarán y creerán: “y mirarán a Mí, a quien traspasaron, y llorarán como se llora por hijo Unigénito, afligiéndose por Él como quien se aflige por el Primogénito. (Zacarías 12:10). Por eso Apocalipsis habla de esos 144,000 judíos (“Ciento cuarenta y cuatro mil sellados DE TODAS LAS TRIBUS DE LOS HIJOS DE ISRAEL.” Apocalipsis 7:4 y 14:3), que estará, en la época de la gran tribulación. La mujer de Apocalipsis, entonces no es la virgen María…; veamos un poco más de contexto bíblico: Sabemos por la Palabra y estuvo profetizado por Cristo que el templo sería destruido: “Porque vendrán días sobre ti, cuando tus enemigos te rodearán con vallado, y te sitiarán, y por todas partes te estrecharán, y te derribarán a tierra, y a tus hijos dentro de ti, y no dejarán en ti piedra sobre piedra, POR CUANTO NO CONOCISTE EL TIEMPO DE TU VISITACIÓN.” Lucas 19:. 43-44. ES DECIR QUE EL PUEBLO JUDÍO IBA TENER UN JUICIO… por eso fue dispersado. Por otro lado sabemos que la HIGUERA REPRESENTA A ISRAEL en la Palabra: “Así ha dicho YHVH Dios de Israel: Como a estos HIGOS buenos, así miraré a los transportados de Judá, a los cuales eché de este lugar a la tierra de los caldeos, para bien.” Jeremías 24:5 y “Como uvas en el desierto hallé a Israel; como la fruta temprana de la HIGUERA en su principio vi a vuestros padres.” Oseas 9:10 Por eso cuando Jesús maldijo la higuera (Marcos 11: 13-21) SE ESTABA REFIRIENDO A ISRAEL... Él ya sabía que los Judios lo rechazarían, y trajeron a esa maldición en práctica cuando entregaron a Jesús a Pilato: "Viendo Pilato que nada adelantaba, sino que se hacía más alboroto, tomó agua y se lavó las manos delante del pueblo, diciendo: Inocente soy yo de la sangre de este Justo; allá vosotros. Y respondiendo todo el pueblo, dijo: SU SANGRE SEA SOBRE NOSOTROS, Y SOBRE NUESTROS HIJOS.” Mateo 27:24-25. Pero Jesús también profetizó su restauración y este proceso YA HA COMENZADO (culminará cuando crean en Aquel que traspasaron y Ap.1:7; Zac. 12:10…): Marcos 13: 28-30 "Aprended la parábola de la higuera (Israel): Cuando ya su rama está tierna, y brotan las hojas, sabéis que el verano está cerca (podría ser 1948 o 1967..., el milagro ocurre el siglo XX… 1900 años después). Así también vosotros, cuando veáis que suceden estas cosas, conoced que está cerca, a las puertas. De cierto os digo, que no pasará esta generación (70-80, 100 ó 120 años o la edad), sin duda no pasará hasta que todas estas cosas acontezcan". Ahora sólo BROTAN las hojas luego pero DARÁ FRUTO cuando llegue la hora de SU PARTO. Eso es lo que se narra en el libro de Apocalipsis capítulo 12: Que es la mujer, es decir ISRAEL, DE PARTO; las estrellas aluden a LAS 12 TRIBUS DE ISRAEL. Esto queda claro al ver el sueño de José (Génesis 37:9), donde el sol, luna y estrellas se inclinan a él… esto es Jacob, Raquel y los hermanos, que sumando a José que tuvo el sueño; que son el pueblo de Israel. Es una regla de hermenéutica que Apocalipsis y las llaves de su simbolismo están en el Antiguo Testamento y resto del Nuevo, aquel que no respeta esto está: “sacando el texto del contexto para hacer un pretexto”. La mujer es Israel que en ESA ÉPOCA SE CONVERTIRÁ RECIÉN = parto = nacer de nuevo. Repito, por eso huyen al desierto que fue el consejo que dio Cristo a ISRAEL en Mateo 24:16 (Cristo dirije esas palabras a un público judío) y todo esto se cumple al final de los tiempos. Y ESO ES EXACTAMENTE LO QUE SE DICE en los capítulos 11 y 12 de Daniel, que narra ese comienzo de restauración, ese brote de hojas... veamos ahora: “Y dijo uno al varón vestido de lino, que estaba sobre las aguas del río: ¿CUÁNDO SERÁ EL FIN DE ESTAS MARAVILLAS? Y oí al varón vestido de lino, que estaba sobre las aguas del río, el cual alzó su diestra y su siniestra al cielo, y juró por el que vive por los siglos, que será por tiempo, tiempos, y la mitad de un tiempo. Y CUANDO SE ACABE LA DISPERSIÓN DEL PODER DEL PUEBLO SANTO, todas estas cosas serán cumplidas. Y yo oí, mas NO LAS ENTENDÍ. Y dije: Señor mío, ¿CUÁL SERÁ EL FIN DE ESTAS COSAS? Él respondió: Anda, Daniel, pues estas palabras están CERRADAS Y SELLADAS HASTA EL TIEMPO DEL FIN. Muchos serán limpios, y emblanquecidos y purificados; los impíos procederán impíamente, y ninguno de los impíos entenderá, pero los entendidos comprenderán.” Daniel 12:6-10. Por qué no entendió Daniel… ciertamente Daniel estaba en la dispersión… (600 a.C.; por eso no entendió), es clarísimo que se habla como SE DICE EXPRESAMENTE… del “TIEMPO DEL FIN”; y la señal en el tiempo del fin será: “CUANDO SE ACABE LA DISPERSIÓN DEL PODER DEL PUEBLO SANTO”. No tiene cumplimiento el año 520a.C. es evidente sino EN ESTOS TIEMPOS… “Pero tú, Daniel, cierra las palabras y sella el libro HASTA EL TIEMPO DEL FIN. Muchos correrán de aquí para allá, y la ciencia se aumentará.” (Daniel 12:4). Luego cuando el anticristo (que será aclamado y muy popular y vestido de ángel de luz y recibido hasta por las cabezas religiosas y demás) se siente en ese tercer templo que está por construirse y pida adoración… y sean perseguidos al rechazar el adorar al enemigo, creerán; eso es dar a luz, dar fruto. Por eso las 12 estrellas que representan las tribus de Israel; por eso huyen al desierto, porque será perseguida… al rehusar adorar al anticristo y porque es cuando recién entenderá. Los textos DENTRO DEL CONTEXTO… 10. La idea de la perpetua virginidada de Mar{ia Entre los Padres de la iglesia veremos a san Agustín; con citas revisadas y fuente que puede ser verificada; San Agustín (354-430 d.C.), doctor de la Iglesia, comentando el Salmo 35 (14), afirmó: "MARÍA, HIJA DE ADÁN, MURIÓ POR CAUSA DEL PECADO, y la carne del Señor nacida de María murió para borrar el pecado". Exposiciones al Libro de los Salmos XXXV (14) en algunas versiones se halla en el Salmo XXXIV (distinta numeración) Philip Schaff, D.D., LL.D. Nicene and Post Nicene Fathers of the Christian Church. - The Christian Literature Company 1890-1900. New York. BibleWorks version 9. – En otra fuente está en el Salmo XXXIV Sermón 2 (3). https://www.augustinus.it/spagnolo/esposizioni_salmi/esposizione_salmo_047_testo.htm - Traducido por Miguel Fuertes Lanero, OSA. “Por lo tanto, TODOS, SIN EXCEPCIÓN ALGUNA, estaban muertos en sus pecados, ya fueran pecados originales o voluntarios, pecados de ignorancia o pecados cometidos contra el conocimiento”. En San Agustín (Ciudad de Dios, Libro XX. cap. 6 ¿Cuál es la primera resurrección y cuál la segunda?) vol. ii. p. 425 - Philip Schaff, D.D., LL.D. Nicene and Post Nicene Fathers of the Christian Church. - The Christian Literature Company 1890-1900. New York. BibleWorks version 9. – Otra fuente: https://www.augustinus.it/spagnolo/cdd/index2.htm - Buscar Libro XX y luego capítulo 6 (1). Es clarísimo el pensamiento de San Agustín, doctor para la iglesia católica romana; no hay lugar a dudas: “María MURIÓ POR CAUSA DEL PECADO”, “Por lo tanto, TODOS, SIN EXCEPCIÓN ALGUNA, estaban muertos en sus pecados”. Otros Factores a considerar: • Dios se desagrada en que confiemos en hombres, sino únicamente en Dios, esto incluye a Cristo que es el “Hijo de Dios”: Job 15:15,16 “He aquí, en sus santos no confía, y ni aun los cielos son limpios delante de sus ojos; ¿Cuánto menos el hombre abominable y vil, que bebe la iniquidad como agua? - “Así ha dicho el Señor: Maldito el varón que confía en el hombre, y pone carne por su brazo, y su corazón se aparta de YHVH. (Jeremías 17:5). Vemos en estos pasajes que nadie es limpio sino Dios mismo, ni aún sus santos (en Cristo somos santos), por lo que nuestra confianza debe estar puesta exclusivamente en Dios, en nuestro caso únicamente en Cristo, el Hijo de Dios. Dios prohíbe que se pida a un muerto (aunque haya sido un santo) y lo considera entre las prácticas espiritistas. Deuteronomio18:11 "ni quien consulte a los muertos; porque es abominación para con YHVH cualquiera que haga estas cosas”. Los judíos no tienen tampoco intercesores a pesar que tienen muchos santos (Enoc, Moisés, Elías,... debemos considerar que tanto Enoc como Elías no vieron muerte, fueron llevados vivos a la presencia de Dios Hebreos 11:5 y 2Reyes 2:11); sin embargo se dirigen directamente a Dios como nosotros que oramos en el nombre de Cristo Jesús, QUE ES DIOS HECHO HOMBRE. • Cuando Juan apóstol vio en visión a Jesucristo sentado como el Cordero en el trono (Apocalipsis 5:6), no vio a ninguna “reina del cielo”. Debería de haber alguna mención, ya que Apocalipsis se escribe luego que María termina su tránsito en la tierra (más de 50 años después), en todo el libro NO HAY MENCIÓN de María como reina; sino a Cristo como Rey sentado en el Trono. • Mateo 1:25 enseña: “Y no la conoció hasta que dio a luz a su hijo”, aquí hay una (s) palabra (s) importante (s): hasta que (después sí la conoció); en Lucas 2:7 también menciona que Jesucristo es el hijo primogénito (el primero). Repetimos; en griego dice: οὐκ ἐγίνωσκεν αὐτὴν ἕως (no la conoció hasta que); es evidente que el texto nos habla de relaciones maritales, y nos señala que no hubieron HASTA QUE dio a luz a Jesús. Como confirma el texto de Lucas 2:7 Jesús fue el hijo PRIMOGÉNITO, UTILIZA LA PALABRA “πρωτότοκον” (protótokon), que significa PRIMER HIJO. En la Palabra no hay casualidades al utilizar los términos; muy bien hubiesen escrito: “Y María dio a luz a Jesús y José no la conoció jamás o nunca la conoció.”; pero por el contrario dice: “No la conoció hasta que…” dio a luz a su hijo; que implica que luego sí, y que Jesús fue el primero de varios hermanos. Una doctrina “tan importante”, hubiese merecido aclaraciones especiales… que no existen, en TODA LA PALABRA. El que María haya tenido hijos no afecta lo concerniente a la salvación, sino por el contrario ayuda a nos apoyemos SÓLO EN JESUCRISTO, la ROCA DE LOS SIGLOS. Por último repito, ya que sé que las memorias son frágiles; tenemos ÓRDENES EXPRESAS de pedir ÚNICAMENTE por medio de Jesús: El apóstol Pablo nos enseña a pedir sólo a Jesucristo: “y TODO lo que hacéis; sea de palabra o de hecho, hacedlo TODO en el NOMBRE DEL SEÑOR JESÚS; dando gracias a Dios Padre POR MEDIO DE ÉL” Colosenses 3:17. ¿No está claro?, dice TODO (lo que se hace de palabra y de hecho; es decir TODO), y vuelve a repetir HACEDLO TODO (en imperativo) y RECALCA TODO (no algunas cosas, sino TODO), en el NOMBRE DEL SEÑOR JESÚS… y termina especificando que AÚN LA ACCIONES DE GRACIAS se dan al Padre en SU NOMBRE. Está bien claro como para que nadie se confunda y en imperativo, de tal forma que no es una opción más. Queda FUERA entonces toda otra opción, aunque sea anunciada por un ángel o alguien que se abrogue autoridad… sabiendo de antemano que Dios no va a contradecir lo que ya está escrito. ¿Acaso no lo enseñó Jesucristo reiteradamente? ¿Acaso, sólo lo dijo UNA VEZ? ¿Por qué cree usted que repite Cristo Jesús tantas veces esta verdad? ¿No será porque COMO DIOS SABÍA QUE LA TENDENCIA SERÍA DIRIGIERSE A LOS SANTOS Y A MARÍA? 1. Juan 14:13 "y todo lo que pidiereis al Padre EN MI NOMBRE, lo haré, para que el Padre sea glorificado en el Hijo”. 2. Juan 14:14 “Si algo pidiereis EN MI NOMBRE, yo lo haré”. 3. Juan 15:16 “...para que todo lo que pidiereis al Padre EN MI NOMBRE, Él os lo dé”. 4. Juan 16:23 “De cierto, de cierto os digo, que todo cuanto pidiereis al Padre EN MI NOMBRE, os lo dará”. 5. Juan 16:24 “Hasta ahora nada habéis pedido EN MI NOMBRE, pedid y recibiréis, para que vuestro gozo sea cumplido”. 6. Juan 16:26 “en aquel día pediréis EN MI NOMBRE...”. Si Cristo es Dios, por lo que es omnisciente, omnipotente y misericordioso (murió por ti); puede contestar como Dios todas las oraciones (aunque todos oren al mismo tiempo), por lo que no necesitamos de más. Cristo satisface todos los requerimientos con una perfección infinita. ¿No está suficientemente claro que hay que pedir en SU NOMBRE? Finalmente, ¿No está CLARO EN LA PALABRA QUE CRISTO JESÚS ES NUESTRO INTERCESOR? Hebreos 7:25 dice: “por lo cual puede también salvar perpetuamente a los que por Él se acercan a Dios, VIVIENDO SIEMPRE PARA INTERCEDER POR ELLOS”. Romanos 8:34 “¿Quién es el que condenará? Cristo es el que murió; más aún, el que también resucitó, el que además está a la diestra de Dios, EL QUE TAMBIÉN INTERCEDE POR NOSOTROS”; otro pasaje que es útil para confirmar aún más lo afirmado es Efesios 3:11-12 que dice, “...en Cristo Jesús nuestro Señor, EN QUIEN TENEMOS SEGURIDAD Y ACCESO (a Dios) con confianza por medio de la fe en Él”. Bueno entonces NO HAY EXCUSA; quedan avisados que si piden dirigiéndose a María o a los “santos”, están en DESOBEDIENCIA a lo ESCRITO. Y no esperen que baje un ángel del cielo, un sueño o revelación que les confirme más… POR ALGO DIOS LO REGISTRA EN SU PALABRA. Por eso Jesús dijo: “Escudriñad las Escrituras” (Juan 5:39)… Si tratan de rezar, orar, comunicarse con el Padre por OTRO canal QUE NO ES SU HIJO, EL VERBO, EL HIJO DE DIOS, simplemente irán a otro dios y recibirán otro espíritu, ya que menospreciaron la verdad de su Palabra: “Mi pueblo a su ídolo de madera pregunta, y el leño responde; porque ESPÍRITU DE FORNICACIONES LO HIZO ERRAR,...”. Estos ídolos son precursores de la imagen final en los tiempos del fin que tendrá vida, hablará y matará a todo aquel que no la adore Apocalipsis 13:13-16 (Y se le permitió infundir aliento a la imagen de la bestia, para que la imagen hablase e hiciese matar a todo el que no la adorase); vemos que la bestia SI BUSCA QUE LA ADOREN EN FORMA DE IMAGEN. Hablando de esto hay que tener cuidado con representar a Dios por medio de imágenes como se aclara en Deuteronomio 4:12, 15, 16 que dice: “Guardad pues mucho vuestras almas; pues NINGUNA FIGURA VISTEIS el día que YHWH habló con vosotros de en medio del fuego; para que no os corrompáis y hagáis para vosotros escultura, imagen de figura alguna, efigie de varón u hembra,...”. Es algo terrible que hayan quitado el segundo mandamiento y dividido el último en 2 para que sigan habiendo 10 (y esto es algo en que los judíos hasta hoy se indignan puesto que ellos conocen). El enemigo puede y hace milagros: “y con todo engaño de iniquidad para los que se pierden, por cuanto no recibieron el amor de la verdad para ser salvos. Por esto Dios les envía un poder engañoso, para que crean la mentira, a fin de que sean condenados todos los que no creyeron a la verdad, sino que se complacieron en la injusticia.” 2Tesalonisences 2:10-12. Por eso la señal, milagro, profecía, visión, aparición…, sola NO ES PRUEBA DE SER DE DIOS, sino el que SE ENMARQUE EN SU PALABRA. El amar a Dios, que es el primer mandamiento, requiere del que dice que ama, el investigar, ESFUERZO, perseverancia. El que no tiene interés falta a la verdad. Esto no quita el amar al prójimo; sino lo complementa y da sentido y dirección. Los días en la tierra son pocos, e amor a Dios incluye: tiempo de COMUNIÓN CON ÉL en oración y alabanza; tiempo para INVESTIGAR Y ESCUDRIÑAR, sino nunca tendrá una certeza personal y perderá su alma. Bienvenidos comentarios o dudas sobre pasajes referentes a este tema. Aquel que llegó hasta aquí reciba un abrazo, y siga adelante. “Fieles son las heridas del que ama; Pero importunos los besos del que aborrece.” (Proverbios 27:6) Saludos View publication stats | RESEARCH GATE |
Volume 5 Nomor 2, 2022, 183 - 197 E-ISSN: 2655-0695 DOI: https://doi.org/10.32923/kjmp.v5i2.2546 183 Pelaksanaan Full Day School dan Pra Full Day School: Studi Kasus Di SDIT Albina Pangkalpinang dan SD Muhammadiyah Pangkalpinang Gustin1*, Ratna Dewi2, Satria Budiman3, Pung Purwadi4, Ria Budiarti5 Abstrak Tujuan dari penelitian ini untuk mengetahui pelaksanaan full day school di SDIT Albina Pangkalpinang dan pelaksanaan pra full day school di SD Muhammadiyah Pangkalpinang. Metode penelitian yang digunakan dalam penelitian ini adalah analisis deskriptif kuantitatif. Berdasarkan data yang dianalisis dengan menggunakan deskripsi data kuisioner ditemukan bahwa analisis pelaksanaan full day school di SDIT Albina Pangkalpinang dan pra full day school di SD Muhammadiyah pangkalpinang bahwa pada kategori sangat tinggi 34 siswa yaitu 55%, kategori tinggi sebanyak 19 siswa yaitu 31%, kategori rendah sebanyak 9 siswa yaitu 14%, dan kategori sangat rendah sebanyak 0 siswa yaitu 0%. Hal ini menunjukkan bahwa persentase terbesar perolehan skor tersebut berada pada kategori sangat tinggi, hal tersebut menunjukan kondisi yang baik. Output yang idealnya dari pelaksanaan program full day school dan pra full day school yaitu Mengoptimalkan satuan kurikulum pembelajaran dengan waktu belajar siswa, Menyediakan fasilitas olahraga dan mendukung kegiatan pengembangan diri, Peningkatan program ekstrakurikuler dan pemantapan dalam menerapkan budaya sekolah dengan landasan pendidikan karakter bagi peserta didik, Menentukan standart waktu untuk kegiatan ekstrakurikuler, Menyediakan kantin sekolah yang dikelola oleh sekolah, agar pemantauan gizi peserta didik teratur, Adanya kesepakatan dan komitmen sekolah dengan warga sekolah. Kata kunci: full day school; pra full day school; sekolah dasar; History: Received : 14 Juli 2022 Revised : 6 November 2022 Accepted : 16 November 2022 Published : 05 Desember 2022 12IAIN Syaikh Abdurrahman Siddik Bangka Belitung, Bangka, Indonesia 3SD Negeri 3 Pangkalpinang 4SD Negeri 6 Lubuk Besar 5SD Negeri 4 Toboali *Koresponden Penulis: gustin@iainsasbabel.ac.id Publisher: LPM IAIN Syaikh Abdurrahman Siddik Bangka Belitung, Indonesia Licensed: This work is licensed under a Lisensi Creative Commons Atribusi 4.0 Internasional. Pendahuluan Kementerian Pendidikan telah mengagendakan program penerapan full day school di sekolah secara umum. Pihak sekolah sejauh ini melaksanakan program full day school berdasarkan otonomi sekolah. Penerapan program full day school harus berdasarkan kesiapan dan perencanaan yang jelas bagi pihak sekolah. Umumnya, full day school ditetapkan dalam proses Pendidikan di Indonesia bertujuan untuk menyelesaikan masalah-masalah yang ada pada dunia Pendidikan di Indonsia. Berdasarkan telaah dari penelitian (Rosa et al., 2022) bahwa pengaruh sumber daya sekolah dan waktu belajar tambahan dalam sistem Pendidikan di Brasil mengalami banyak perubahan. Dimana prestasi dalam hasil ujian di sekolah oleh peserta didik mengalami perbaikan. Selain itu, dengan adanya bertambah waktu pembelajaran, tentunya akan menambah peluang peningkatan sarana dan prasarana oleh pemerintah dalam hal ini untuk penguatan pendidikan di sekolah tersebut dimana watu belajar yang awalnya hanya 4 jam sehari menjadi 8 jam sehari. Namun, berbeda dengan hasil penelitian (Neal et al., 2022) bahwa tidak ada bukti yang kuat bahwa dengan adanya pelaksanaan pembelajaran sehari penuh memiliki efek positif terhadap pembentukan sosial anak dan bahkan sebaliknya. Gustin, et all 184 Mutu Pendidikan di Indonesia perlu disikapi secara baik sehingga kualitas dan kuantitas mutu tentunya akan meningkat. Oleh karena itu perlu adanya kajian dimana kajian tersebut berfokus pada analisis lembaga-lembaga sekolah atau Pendidikan yang merata baik dari segi kebijakan maupun kualitas serta kuantitas sarana prasarannya. Hal ini tentunya membutuhkan restrukturisasi, reorientasi dan inovasi dalam Pendidikan kini sehingga membutuhkan percepatan atau akselerasi di berbagai bidang khususnya Pendidikan. Kemudian melalui telaah dari (Parra, 2022) bahwa terkait evaluasi proses pelaksanaan program pembelajaran dengan full day school di Kolombia mengungkapkan proses Pendidikan di negara tersebut sangat kompleks dilihat dari adanya istilah terdesentralisasi atau sangat tergantung pada otonom sekolah masing-masing sehingga berdampak pada hasil. Sehingga diperlukan resentralisasi secara nasional dalam pengambilan kebijakan yang seharusnya untuk mendukung program inovasi Pendidikan sehingga berorientasi pada pembangunan Pendidikan secara berkelanjutan dan menghasilkan sumber daya manusia yang utuh bagi pemerintah tersebut. Berdasarkan observasi tentang permasalahan yang berkaitan dengan program full day school di sekolah didapati bahwa hal-hal yang sering menjadi masalah diantaranya (1) Minimnya waktu orang tua di rumah karena tingginya tuntutan kerja, muncul anggapan anak akan melakukan kegiatan yang menuju kearah negatif; (2) pengawasan orang tua masih perlu ditingkatkan saat waktu belajar di sekolah berakhir; (3) waktu belajar dalam hal keagamaan masih minim (4) pola full day school belum dilandasi atas landasan yuridis yang kuat serta belum adanya program yang baku dan kompeten dalam program full day school tersebut; (5) Program full day school juga mengalami pro dan kontra dari pihak orang tua, guru, dan sekolah; (6) Siswa yang jauh dari pengawasan orang tua dan sekolah tentu akan mengundang tindakan yang bertolak belakang dengan etika; (7) Masih kurangnya kegiatan pengembangan ekstrakurikuler. Penelitian yang dilakukan oleh (Astuti, 2013) bahwa pelaksanaan pembelajaran dengan system full day school dilakukan lebih banyak waktu belajarnya dibandingkan program sebelumnya. Selain itu, dalam pembelajaran full day masih ditemukan permasalahan baik itu dari sarana prasarana pembelajaran, orang tua siswa dan lainnya dimana hal tersebut membutuhkan waktu sehingga menjadi terbiasa. Hal ini sejalan dengan penelitian yang dilakukan oleh (Hawi, 2015) mengungkapkan bahwa kebijakan full day dalam pembelajaran sekolah dasar yaitu suatu upaya yang dilakukan oleh pemerintah untuk menyiapkan potensi-potensi sumber daya manusia di masa yang akan datang sehingga siap secara komprehensi baik itu kognitif, afektif dan psikomotoriknya. Full day school merupakan pola pembelajaran di sekolah dimanan proses waktu pembelajaran dilaksanakan secara penuh mulai dari pukul 07.00 sampai menjelang 15.30 waktu Indonesia bagian Barat Berdasarkan permasalahan diatas maka perlu adanya penelitian mengenai pelaksanaan full day school dan pra full day school: studi kasus di SDIT Albina Pangkalpinangdan SD Muhammadiyah Pangkalpinang. Diharapkan dengan adanya penelitian ini dapat diketahui secara mendalam permasalahan dan solusi yang tepat terkait penerapan program full day schooldan pra full day school, sehingga dapat meningkatkan mutu pendidikan. Penelitian ini diharapkan memiliki manfaat untuk mewujudkan pelaksanaan program sekolah yang sesuai dan ideal bagi sekolah untuk menerapkan program yang dapat meningkatkan mutu pendidikan dan kualitas sekolah dalam pelaksanaan full day school di SDIT Albina Pangkalpinang dan pra full day school di SD Muhammadiyah Pangkalpinang pada ranah sarana dan prasarana, perkembangan sosial emosi, aktivitas fisik, ekstrakurikuler dan status gizi. Sehingga didapat pelaksanaan yang baik untuk sekolah yang menerapkan full day school maupun sekolah pra full day school. Pelaksanaan Full Day School dan Pra Full Day School: Studi Kasus di SDIT Albina Pangkalpinang dan SD Muhammadiyah Pangkalpinang 185 Metode Penelitian ini merupakan penelitian deskriptif kuantitatif. Metode kuantitatif disebut juga metode klasik atau sudah lama atau sering digunakan orang. Metode ini berupa data angka kuantitatif serta statistik digunakan untuk menganalisis temuan peneliti (Ismail Nurdin & Sri Hartati, 2019). Data kuantitatif didapatkan melalui dokumentasi, angket, dan wawancara secara mendalam kepada partisipan. Tempat pelaksanaan Penelitian ini di SDIT Albina Pangkalpiang dan SD Muhammadiyah Pangkalpinang. Teknik pengumpulan data dalam penelitian ini adalah kuisioner dan wawancara. Peneliti yang melaksanakan dan menganalisis data yang didapat dalam penelitian. Pada penelitian ini digunakan beberapa instrumen, yaitu: wawancara pelaksanaan full day school dan pra full day school, angket pelaksanaan full day school dan pra full day school, lembar observasi, dan dokumentasi. Pada instrumen soal yang digunakan, soal yang valid kami gunakan sebagai alat untuk memperoleh data terkait penelitian ini dan soal yang tidak valid tidak peneliti gunakan karena dinilai sulit dipahami. Jumlah item soal 55 item soal, jumlah soal yang valid berjumlah 49 item dan yang tidak valid atau yang tidak peneliti gunakan adalah 6 soal yang terdiri dari soal 29, 30, 32, 34, 36, dan 37. Analisis data penelitian ini yaitu dengan deskriptif kuantitatif. Data yang dikumpulkan kemudian dideskripsikan secara komprehensif, selanjutnya data yang ada tersebut diberikan pemaknaan sesuai dengan penelitian yang dilaksanakan secara kuantitatif (Sudaryono, 2016). Hasil dan Pembahasan Berdasarkan hasil penelitian dapat dideskripsikan terkait analisis pelaksanaan full day school dan pra full day school: Studi Kasus di SDIT Albina Pangkalpinang dan SD Muhammadiyah Pangkalpinang yang meliputi sarana dan prasarana, sosial emosi, aktivitas fisik, ekstrakurikuler, dan status gizi. Berdasarkan hasil penelitian yang dilakukan di SD Muhammadiyah Pangkalpinang terkait 5 aspek analisisfull day school dan prafull day school bahwa Sarana pembelajaran terdiri dari: infokus berjumlah lebih dari 4 dan yang masih bisa digunakan ada 2 dan memiliki laptop sekolah. Prasarana sekolah terdiri dari Perpustakaan, UKS (Usaha Kesehatan Sekolah), kantin, memiliki bangunan 2 lantai dan musala. Total ruangan kelas yang tersedia adalah 18 ruangan kelas. Prasarana sekolah terdiri dari Perpustakaan, UKS, kantin, Laboratorium olahraga, ruangan guru dan kepala sekolah. Kelas 5/B terletak di lantai 2 gedung sekolah, jumlah kelas dinilai cukup dengan jumlah siswa. Prasarana sekolah terdiri dari Perpustakaan, UKS, kantin, memiliki bangunan 2 lantai dan musala, lapangan putsal dan halaman upacara. Laboratorium IPA, Laboratorium Komputer, musala, kantor, lapangan Aula, ruang kepala sekolah, dan ruang belajar dan sering digunakan untuk ruangan ekstrakurikuler yang beragam seperti karate, tenis meja, jadi total ruangan 27 lokal, sarana di laboratorium komputer 20 unit komputer, laptop sekolah 5 unit, infocus 8 buah, fasilitas di dalam kelas menggunakan kipas angin, aula pertemuan dilengkapi AC (Air Conditioner), kantin dan koperasi sekolah tersedia, lapangan ekstrakurikuler memanah disesuaikan dengan ruangan yang tidak sedang digunakan untuk pembelajaran, UKS dan Dokter yang bertugas di UKS. Laboratorium IPA, Laboratorium Komputer, musala, kantor, lapangan Aula, ruang kepala sekolah, dan ruang belajar yang sering digunakan untuk ruangan ektrakurikuler yang beragam seperti karate, pimpong, jadi total ruangan 27 lokal, sarana di laboratorium komputer 20 unit komputer, laptop sekolah 5 unit, infocus 8 buah, fasilitas di dalam kelas menggunakan kipas angin, aula pertemuan dilengkapi AC (Air Conditioner), kantin dan koperasi sekolah tersedia, lapangan ekstrakurikuler memanah disesuaikan dengan ruangan yang tidak sedang digunakan untuk pembelajaran, UKS dan Dokter yang bertugas di UKS. Sesuai dengan penelitian yang dilakukan oleh (Parid & Alif, 2020) bahwa dalam pengelolaan Pendidikan yang baik dan benar perlu adanya sarana dan prasarana yang baik untuk mendukung hal tersebut mulai dari rancangan kebutuhan sampai dengan evaluasi keberadaan sarana prasarana tersebut sehingga tepat penggunaanya dan tepat sasarannya. Gustin, et all 186 Sarana dan prasarana di SDIT Albina memiliki Laboratorium IPA, Laboratorium Komputer, musala, kantor, lapangan Aula, ruang kepala sekolah, dan ruang belajar dan sering digunakan untuk ruangan ekstrakurikuler yang beragam seperti karate, jadi total ruangan 27 lokal, sarana yang dimiliki SDIT Albina di lab komputer 20 unit komputer, laptop sekolah 5 unit, infocus 8 buah, fasilitas di dalam kelas menggunakan kipas angin, aula pertemuan dilengkapi AC (Air Conditioner), kantin dan koperasi sekolah tersedia, lapangan ekstrakurikuler memanah disesuaikan dengan ruangan yang tidak sedang digunakan untuk pembelajaran, UKS tersedia sebagai prasarana dalam pemakaian ekstrakurikuler. Berdasarkan penelitian (Thompson & Sonnenschein, 2016) bahwa pembelajaran dengan pola sehari penuh dapat meningkatkan kemampuan akademik peserta didik khususnya bagi peserta didik yang memang mengalami kesulitan atau kendala dalam penguatan akademiknya. Selain berkaitan dengan sarana dan prasarana untuk meningkatkan kualitas suatu institusi sekolah diperlukan juga pengetahuan sosial emosi yang tinggi, yang mana pengembangan emosi siswa dalam menanamkan sikap kejujuran, kedisiplinan dikembangkan di dalam KBM. Untuk pengembangan sikap gotong royong dikembangkan melalui program kegiatan gotong royong bersama setiap sebulan sekali. Sesuai dengan penelitian (Dewi et al., 2020) bahwa peluang anak tersebut berkembang secara utuh baik pada masa sekolah dasar dimana kecerdasan berbahasa, kognitif, afektif dan psikomotoriknya. Setiap peserta didik memiliki perkembangan yang berbeda-beda dimana hal tersebut tentunya dipengaruhi oleh beberapa faktor khususnya secara primer yaitu keluarga dan secara sekunder yaitu lingkungan sekolah. Pengembangan sikap kejujuran, kedisiplinan dikembangkan di dalam KBM. Pengembangan sikap gotong royong dikembangkan melalui program kegiatan gotong royong bersama setiap sebulan sekali dan piket kelas yang berjalan setiap harinya mencerminkan budaya tanggung jawab dan menghargai lingkungan.Kegiatan yang dilakukan oleh siswa selalu terpantau oleh guru, pada jam istirahat Adab makan, disiplin dengan mengembangkan budaya antre, dan hukuman yang mendidik seperti hafalan ayat dan membaca istigfar sebanyak 100 kali serta nilai kejujuran dengan mengembangkan program market day pada hari jum’at. Mengembangkan budaya antre saat berwudhu,terlambat shalat akan dikenakan hukuman yang mendidik seperti hafalan ayat dan membaca istigfar sebanyak 100 kali. Berdasarkan penelitian (Soapatty & Suwanda, 2014) Menunjukkan bahwa sarana dan prasarana dalam pembelajaran sangat penting dalam mempengaruhi prestasi atau kinerja peserta didik. Kegiatan pembelajaran dimulai dari Pukul 07.15-07.30 WIB kegiatan apel pagi seperti hafalan ayat Al-Quran dan hafalan bacaan salat. Pukul 07.30-13.00 WIB mulai kegiatan belajar mengajar. Kegiatan belajar mengajar (KBM) guru menggunakan media pembelajaran dalam menyampaikan materi pelajaran.Sebelum memulai pembelajaran Pukul 07.15-07.30 WIB kegiatan apel pagi, dilanjutkan dengan pembacaan Surat Pendek dan hafalan, seperti hafalan do’a Sehari-hari dan hafalan bacaan salat Pukul 07.30-13.00 WIB mulai kegiatan belajar mengajar. dalam kegiatan belajar mengajar (KBM). Penggunaan media pembelajaran guru menyesuaikan dengan mata pelajaran dan kebutuhan pembelajaran. Setiap waktu yang ada merupakan bagian dari waktu belajar, dimana jika di waktu kelas berlangsung dari awal sampai akhir merupakan seperangkat pembelajaran yang harus dijadikan sebagai proses pembelajaran itu sendiri (Vashdev, 2016). Siswa belajar seperti pada umumnya dan peran guru dalam pembelajaran dinilai mendukung pemahaman siswa dalam pembelajaran. Menggunakan penilaian berdasarkan 3 Aspek, Apektif, kognitif, dan Psikomotorik anak. Pukul 07.30-13.00 WIB mulai kegiatan belajar mengajar. dalam kegiatan belajar mengajar guru menggunakan media pembelajaran dalam mengajar kendala dan permasalahan diselesaikan bersama seperti kekurangan fasilitas pembelajaran, ruangan yang ada di tingkat atas dan penilaian yang terpantau. Pelaksanaan Full Day School dan Pra Full Day School: Studi Kasus di SDIT Albina Pangkalpinang dan SD Muhammadiyah Pangkalpinang 187 Siswa belajar menggunakan metode umi yaitu kegiatannya membaca Al-Quran, hafalan doa-doa sehari-hari dan surat pendek. Setelah itu mulai pemberian materi ajar pada umumnya sampai waktu untuk istirahat. Sebelum masuk ke kelas siswa berwudu terlebih dahulu, salat duha berjamaah di kelas dan belajar menggunakan metode umi yaitu kegiatannya membaca Al-Quran, hafalan doa-doa sehari-hari dan surat pendek. Setelah itu mulai pemberian materi ajar pada umumnya sampai zuhur dan waktu untuk istirahat sekitar 1 jam. Untuk menjaga dan mengembangkan suatu misi atau tujuan pembelajaran yang islami maka suatu internalisasi nilai dan norma islam dapat menjadi pedoman dalam menentukan arah atau manfaat Pendidikan itu sendiri tentunya semua hal tersebut berpedoman pada Al-Quran (Mohammad Al Farabi, 2018). Kegiatan ekstrakurikuler di lakukan pada hari senin, rabu dan jumat. Ekstrakurikuler dibagi menjadi dua yaitu ekstrakurikuler pilihan seperti futsal, dokter cilik, mulok, nasyid, nari, TIK, science club, dan ekstrakurikurer wajib karate dan pramuka. Kegiatan ekstrakurikuler di lakukan pada hari senin, rabu dan jumat. Untuk hari senin kegiatan ekstrakurikuler pilihan yaitu futsal, dokter cilik, mulok, nasyid, nari, TIK, science club, untuk hari rabu dan jumat kegiatan ekstrakurikuler wajib karate dan pramuka. Proses pembelajaran atau kegiatan ekstrakurikuler keagamaan khususnya merupakan suatu bagian dasar yang penting bagi pribadi pebelajar. Dimana hal ini bertujuan untuk menemukan serta membangun minat siswa dalam bidang-bidang khusus atau tertentu (Mohamad Yudiyanto, 2021). Untuk status gizi siswa tergantung pendapatan atau perekonomian dan perhatian orang tua, karena itu berpengaruh terhadap asupan makanan yang didapatkan oleh siswa. Siswa membawa bekal uang dan bekal makanan. Sekolah tidak menyiapkan makanan untuk siswa.Olahraga yang teratur dijadwalkan bagi setiap kelas. Orang tua siswa dapat menyiapkan bekal untuk anaknya. Bagi sekolah yang menerapkan full daymereka menyiapkan makanan untuk siswa. Serta didukung dengan olahraga yang teratur dan sudah dijadwalkan bagi setiap kelas. Pada jam olahraga siswa dibimbing oleh guru PJOK. Pihak sekolah bekerjasama dengan koperasi dan pihak penyedia makanan untuk menyediakan santap siang siswa. Sayur sangat bermanfaat bagi kecerdasan anak dan diharapkan siswa sering mengkonsumsi sayur, dengan demikian siswa selalu disediakan lauk sayuran, pihak sekolah menggunakan program menanam sendiri sayuran di sekolah dan dikonsumsi sendiri untuk siswa sebagai sayur dan dimasak untuk makan siang siswa. Siswa dipantau status gizinya dengan bekerjasama melalui/UKS.Dokter Sekolah yang mengawasi dan membantu jika membutuhkan pertolongan, seperti sakit ringan, dan pemantauan makanan yang baik bagi siswa. Status gizi anak masa sekolah adalah suatu siklus dalam kehidupan yang tentunya akan berpengaruh satu dengan yang lainnya. Untuk menangani status gizi anak usia sekolah dapat diketahui dari pola pemberian zat-zat makanan yang dimanan hal tersebut tentunya sangat berguna bagi pertumbuhan dan perkembangan anak baik secara fisik mupun mentalnya (Lalu Juntra Utama & Yohanes Don Bosko Demu, 2021). Tabel 1 Distribusi Frekuensi analisis Pelaksanaan Full Day School dan pra Full Day School: Studi Kasus di SDIT Albina Pangkalpinang dan SD Muhammadiyah Pangkalpinang No Interval Frekuensi Frekuensi Komulatif 1 19-22 6 6 2 23-26 12 18 3 27-30 5 23 4 31-34 7 30 5 35-38 13 43 6 39-42 10 53 7 43-46 6 59 8 47-49 3 62 Total 62 Gustin, et all 188 Persentase skor penilaian pada variabel analisis Pelaksanaan Program full day school dan pra full day school: Studi Kasus di SDIT Albina Pangkalpinang dan SD Muhammadiyah Pangkalpinang yang menunjukan bahwa skor tertinggi adalah 36 yang mana 13 orang memilih skor antara 35 sampai 38. Hal ini dapat dideskripsikan bahwa skor 36 tersebut telah melewati skor atau nilai tengah atau median. Terkait skor terendah pada skor 19 sampai dengan skor 22 dengan frekuensi 6 responden dari 62 responden atau peserta didik. Kemudian Data tersebut Untuk lebih jelasnya terkait dengan penjelasan di atas dapat dilihat pada gambar di bawah ini: Gambar 1 Persentase Skor Penilaian analisisPelaksanaan Full Day School: dan pra Full Day School: Studi Kasus di SDIT Albina dan SD Muhammadiyah Pangkalpinang. Berdasarkan Perhitungan empat kategori kecendrungan normal bahwa: Tabel 2 Empat kategori kecendrungani analisis Pelaksanaan Full Day School dan pra Full Day School: Studi Kasus di SDIT Albina Pangkalpinang dan SD Muhammadiyah Pangkalpinang Hasil selengkapnya dapat dilihat pada tabel berikut: Tabel 3 Kategori Kecenderungan Analisis pelaksanaan program full day school dan pra Full Day School: Studi Kasus di SDIT Albina Pangkalpinang dan SD Muhammadiyah Pangkalpinang. Berdasarkan tabel tersebut, dapat diketahui variabel analisis Pelaksanaan Program full day school dan pra full day school: Studi Kasus di SDIT Albina Pangkalpinang dan SD Muhammadiyah Pangkalpinang pada kategori sangat tinggi sebanyak 34 siswa, kategori tinggi sebanyak 19 siswa, kategori rendah sebanyak 9 siswa, dan kategori sangat rendah sebanyak 0 siswa. Hal ini menunjukkan bahwa Pelaksanaan Program full day school dan pra full day school: Studi Kasus di Kategori Skor SangatTinggi X ≥ 32,7 Tinggi 32,7 X≥24,5 Rendah 24,5X ≥16,3 Sangat Rendah X <16,3 Skor Siswa Frekuensi Persentase Klasifikasi X ≥ 32,7 34 55% Sangat Tinggi 32,7 X≥24,5 19 31% Tinggi 24,5X ≥16,3 9 14% Rendah X <16,3 0 0% Sangat Rendah Jumlah 62 100% Pelaksanaan Full Day School dan Pra Full Day School: Studi Kasus di SDIT Albina Pangkalpinang dan SD Muhammadiyah Pangkalpinang 189 SDIT Albina Pangkalpinang dan SD Muhammadiyah Pangkalpinang yang meliputi sub variabel sarana dan prasarana, perkembangan sosial emosi, aktivitas fisik, ekstrakurikuler, dan status gizi dalam kategori sangat tinggi dan ditunjukkan oleh persentase terbesar perolehan skor tersebut, hal tersebut menunjukan bahwa pelaksanaan yang terbesar didapat dari penerapan program full day school di SDIT Albina Pangkalpinang. SD Muhammadiyah Pangkalpinang disimpulkan dalam penerapan pra full day school dinyatakan baik pada analisis Pelaksanaan pra fullday school: di SD Muhammadiyah Pangkalpinang. Kemudian, memperkuat hasil persentase di atas dapat dilihat dari hasil persentase keseluruhan responden siswa berkaitan dengan sub variabel tentang analisis Pelaksanaan Program full day school dan pra full day school: Studi Kasus di SDIT Albina Pangkalpinang dan SD Muhammadiyah Pangkalpinang yang meliputi sub variable sarana dan prasarana, perkembangan sosial emosi, aktivitas fisik, ekstrakurikuler, dan status gizi. Adapun Persentase keseluruhan per sub variabel responden siswa berkaitan dengan sub variable tersebut dapat dilihat pada tabel berikut: Tabel 4 Persentase Keseluruhan Responden siswa Berkaitan dengan Sub Variabel Sarana dan Prasarana Aspek Pencapaian Persentase (%) Sarana dan Prasarana Ya Tidak 44 56 Dari tabel di atas dapat dilihat bahwa pada aspek Sarana dan Prasarana berada pada persentase 44% ya atau berada pada kategori rendah. Hal ini mengartikan bahwa perlu adanya penguatansarana dan prasarana khususnya di SD Muhammadiyah Pangkalpinang. Kemudian persentase keseluruhan responden siswa berkaitan dengan sub variabel indikator atau aspek Sosial Emosi dapat dilihat pada tabel berikut: Tabel 5 Persentase Keseluruhan Responden siswa Berkaitan dengan Sub Variabel Sosial Emosi Aspek Pencapaian Persentase (%) Sosial Emosi Ya Tidak 82 18 Dari tabel di atas dapat dilihat bahwa pada aspek Sosial Emosi berada pada persentase 82% atau berada pada kategori sangat tinggi, walaupun ada persentase sangat rendah yaitu 18%, tentunya hal ini dipengaruhi oleh beberapa faktor. Untuk mengetahui lebih lanjut tentang hal ini dapat dilihat pada hasil paparan data kuantitatif sehingga hasil data indikator ini bersifat objektif. Selanjutnya, persentase keseluruhan responden siswa berkaitan dengan tiap-tiap sub variabel tentang pelaksanaan Program Full Day School dan pra Full Day School: Studi Kasus di SDIT Albina Pangkalpinang dan SD Muhammadiyah Pangkalpinang dapat dilihat pada tabel berikut ini: Tabel 6 Persentase Keseluruhan Responden Siswa Berkaitan Dengan Sub Variabel Aktivitas Fisik Aspek Pencapaian Persentase (%) Aktivitas Fisik Ya Tidak 73 27 Gustin, et all 190 Dari tabel di atas dapat dilihat bahwa pada sub variabel aktivitas fisik berada pada persentase 73% atau berada pada kategori tinggi, namun masih adapersentase aktivitas fisik yang masih rendah 27%. Kemudian, berkaitan dengan jumlah rata-rata persentase siswa sesuai kriteria yang dipilih pada sub variabel ekstrakurikuler dapat dilihat pada tabel berikuti ini: Tabel 7 Persentase Keseluruhan Responden siswa Berkaitan dengan Sub Variabel Ekstrakurikuler Aspek Pencapaian Persentase (%) Ekstrakurikuler Ya Tidak 73 27 Jumlah rata-rata persentase siswa sesuai kriteria yang dipilih pada sub variabel ekstrakurikuler persentasenya adalah 73% sedangkan tidak 27%. Data ini meunjukan bahwa persentase berada pada kategori tinggi. Kemudian, berkaitan dengan jumlah rata-rata persentase siswa sesuai kriteria yang dipilih pada sub variabel Status gizi dapat dilihat pada tabel berikuti ini: Tabel 8 Persentase Keseluruhan Responden siswa Berkaitan dengan Sub Variabel Status gizi Aspek Pencapaian Persentase (%) Status Gizi Ya Tidak 61 39 Jumlah rata-rata persentase siswa sesuai kriteria yang dipilih pada sub variabel status gizi persentasenya adalah 61% sedangkan tidak 39%. Data ini menunjukan bahwa persentase berada pada kategori tinggi. Kemudian terkait Persentase khususnya status gizi dilengkapi data IMT (Indeks massa tubuh) data IMT ini menjelaskan tentang persentase status gizi di SDIT Albina Pangkalpinang dan SD Muhammadiyah Pangkalpinang. Untuk persentase data IMT di SDIT Albina Pangkalpinang dapat dilihat tabel berikut: Tabel 9 Persentase Data IMT program full day school di SDIT Albina Pangkalpinang Presentase Status Gizi SDIT Albina Pangkalpinang 2017 Keterangan Presentase(%) Sangat kurus 0% Kurus 0% Normal 100% Gemuk 0% Obesitas 0% Persentase IMT di SD Muhammadiyah Pangkalpinang dapat dilihat pada tabel berikut: Tabel 10 Persentase Data IMT pra full day school di SD Muhammdiyah Pangkalpinang Presentase Status Gizi SD Muhammadiyah Pangkalpinang Tahun 2017 Keterangan Presentase(%) Sangat kurus 0% Kurus 0% Normal 94% Gemuk 6% Obesitas 0% Pelaksanaan Full Day School dan Pra Full Day School: Studi Kasus di SDIT Albina Pangkalpinang dan SD Muhammadiyah Pangkalpinang 191 Setelah diperoleh IMT dari rumus, maka didapatlah keterangan status gizi siswa, namun dalam hal ini menggunakan rumus dari tabel Antropometri Penilaian Status Gizi Anak Kementerian Kesehatan Republik Indonesia. Salah satu faktor penentu utama dari status gizi siswa adalah pola makan yang baik dan olahraga yang teratur. Jumlah rata-rata persentase siswa sesuai kriteria yang dipilih pada sub variabel status gizi persentasenya adalah 61% sedangkan tidak 39%. Keterangan umum dari masing-masing sekolah seperti: di SDIT Albina Pangkalpinang 100% normal, sedangkan di SD Muhammadiyah Pangkalpinang menunjukan 94% normal dan 6% gemuk. Berdasarkan hasil di atas mata terkait sarana dan prasarana tentunya tiap sekolah memiliki fasilitas dan kelengkapan sekolah yang beragam, namun dalam hal ini sarana dan prasarana yang digunakan di Sekolah Dasar pada umumnya adalah ruang teori/kelas, ruang kantor, ruang guru, perpustakaan, laboratorium IPA dan laboratorium computer. Sarana dan sarana pelengkap seperti kamar mandi, kantin, gedung pertemuan, dan gedung olahraga tentunya diharapkan menjadi prioritas bagi setiap instansi pendidikan/sekolah. Terkait perkembangan social emosi peserta didik tentunya dihubungkan dengan pelaksanaan budaya sekolah dan serangkaian kegiatan yang selalu dilakukan di lingkungan sekolah secara berkelanjutan. Sehingga secara langsung ataupun continue akan terus bertahan dan menjadi kebiasaan yang membentuk perkembangan social emosi peserta didik. Terkait dengan aktivitas fisik belajar di dalam kelas pada umumnya jam pelajaran di tingkat Sekolah Dasar umumnya sama, namun pada kesempatan ini aktivitas fisik tentunya menggunakan media, metode, dan strategi dalam pembelajaran sehingga pembelajaran menjadi lebih bermakna. Terkait dengan kegiatan pengembangan bakat dan minat siswa dalam hal ini kegiatan ekstrakurikuler tentunya menggunakan sarana dan prasarana sehingga dalam hal ini dinyatakan bahwa bagi setiap sub variabel tentunya saling berkaitan dengan demikian maka pengembangan bakat dan minat peserta didik tentunya akan berdampak lebih baik. Mengingat lagi program dari full day maupun pra full day tentunya akan mengenalkan cara untuk mengembangkan bakat dan minat peserta didik. Terkait dengan aspek status gizi peserta didik tentunya mampu dipahami dari pendapatan dan gaya hidup peserta didik, karena mengingat lagi meskipun siswa yang sekolah di sekolah full day maupun pra full day tentunya masih memiliki perbedaan yang mencolok, dan factor yang sangat mempengaruhi ialah latar belakang dan pemenuhan gizi peserta didik di lingkungan keluarga. Gambar 2. Skema analisis Pelaksanaan Program Full Day School Gustin, et all 192 Sangat Tinggi55%Tinggi31%Rendah14%Sangat Rendah0%Berdasarkan data yang di analisis dengan menggunakan deskripsi data kuisioner ditemukan bahwa analisis pelaksanaan full day school dan pra full day school di SDIT Albina Pangkalpinang dan SD Muhammadiyah Pangkalpinang bahwa Berdasarkan analisis statistik menyatakan tahapan analisis Pelaksanaan Full Day School dan pra full day school: Studi Kasus SD Muhammadiyah Pangkalpinang dan SDIT Albina Pangkalpinang menunjukan skor rerata sebesar 33,40; median sebesar 35.00; modus 36; standar deviasi sebesar 8,141; skor minimum sebesar 19; dan skor maksimum sebesar 49. Untuk menunjukan persentase skor penilaian pada variabel analisis Pelaksanaan Full Day School dan praFull Day School: Studi Kasus di SDIT Albina Pangkalpinang dan SD Muhammadiyah Pangkalpinang tersebut dapat dilihat pada gambar berikut: Gambar 3 Persentase Skor Penilaian Pada Variabel analisis Pelaksanaan Program Full Day School dan pra full day school: Studi Kasus di SDIT Albina Pangkalpinang dan SD Muhammadiyah Pangkalpinang. Gambar di atas menunjukan bahwa skor tertinggi adalah 36 yang mana 13 orang memilih skor antara 35 sampai 38. Hal ini dapat dideskripsikan bahwa skor 36 tersebut telah melewati skor atau nilai tengah atau median. Berdasarkan data kecendrungan analisis Pelaksanaan Program Full Day School dan pra Full Day School: Studi Kasus di SDIT Albina Pangkalpinang dan SD Muhammadiyah Pangkalpinang. Menunjukan bahwa persentase skor perolehan analisis full day schooldan pra full day schooldapat dilihat pada gambar berikut: Gambar 4 Persentase Skor Perolehan Analisis Full Day School dan pra Full Day School Berdasarkan gambar tersebut di atas, dapat diketahui variabel analisis Pelaksanaan Program Full Day School dan pra Full Day School: Studi Kasus di SDIT Albina Pangkalpinang dan SD Muhammadiyah Pangkalpinang pada kategori sangat tinggi 34 siswa yaitu 55%, kategori tinggi sebanyak 19 siswa yaitu 31%, kategori rendah sebanyak 9 siswa yaitu 14%, dan kategori sangat Pelaksanaan Full Day School dan Pra Full Day School: Studi Kasus di SDIT Albina Pangkalpinang dan SD Muhammadiyah Pangkalpinang 193 62%38%sarana dan prasaranayatidak82%18%perkembangan sosial emosiyatidak70%30%aktivitas fisik yatidakrendah sebanyak 0 siswa yaitu 0%. Hal ini menunjukkan bahwa kategori sangat tinggi yang ditunjukkan oleh persentase terbesar perolehan skor tersebut, hal tersebut menunjukan kondisi yang baik. Kemudian, memperkuat hasil persentase di atas dapat dilihat dari hasil persentase keseluruhan responden siswa berkaitan dengan tiap-tiap indikator atau aspek tentang Analisis full day school yang meliputi 5 sub variabel. Persentase keseluruhan responden siswa berkaitan dengan sub variabel sarana dan prasarana di SDIT AlbinaPangkalpinang dapat dilihat pada gambar berikut ini. Gambar 5 Diagram Sarana dan Prasarana Berdasarkan gambar tersebut menunjukan bahwa 62% sarana dan prasarana sudah terpenuhi, namun 38% sarana prasarana masih belum lengkap atau belum memenuhi. Sehingga perlu adanya peningkatan sarana dan prasarana di sekolah tersebut. Gambar 6 Diagram Perkembangan Sosial Emosi Berdasarkan gambar tersebut menunjukan bahwa 82% perkembangan sosial emosinya sangat baik, hanya 18 % perkembangan sosial emosi siswa yang kurang baik, namun secara keseluruhan perkembangan sosial emosi siswa di SDIT Albina Pangkalpinang sudah sangat baik. Gambar 7 Diagram Aktivitas Fisik Berdasarkan gambar tersebut menunjukan bahwa 70% aktivitas fisik siswa terpenuhi, namun 30% aktivitas fisik siswa masih belumterpenuhi atau belum dapat diterapkan. Sehingga perlu adanya peningkatan aktivitas fisik siswa sekolah tersebut. Gustin, et all 194 95%5%ekstrakurikulerya38%62%sarana dan prasaranayatidak Gambar 8 Diagram Ekstrakurikuler Berdasarkan gambar tersebut menunjukan bahwa 95% siswa ekstrakurikuler terpenuhi, namun 5% masih belum ada. Sehingga perlu adanya peningkatan program ekstrakurikuler di sekolah tersebut. Gambar 9 Diagram Status Gizi Berdasarkan gambar tersebut menunjukan bahwa 76% siswa terpenuhi status gizinya, namun 24% masih belum terpenuhi. Sehingga perlu adanya peningkatan program status gizi di sekolah tersebut. Presentase grafik SD Muhammadiyah Pangkalpinang dalam setiap sub variabel sebagai berikut: Gambar 10 Diagram sarana dan prasarana Berdasarkan gambar tersebut menunjukan bahwa 38% siswa mengungkapkan bahwa sarana dan prasarana sudah terpenuhi, namun 62% sarana prasarana masih belum lengkap atau belum memenuhi. Sehingga sangat perlu adanya peningkatan sarana dan prasarana di sekolah tersebut. Gambar 11 Diagram perkembangan sosial emosi 76%24%status giziyatidak76%24%perkembangan sosial emosiYATIDAKPelaksanaan Full Day School dan Pra Full Day School: Studi Kasus di SDIT Albina Pangkalpinang dan SD Muhammadiyah Pangkalpinang 195 76%24%aktivitas fisikYaTidakBerdasarkan gambar tersebut menunjukan bahwa 76% perkembangan sosial emosinya sangat baik, hanya 24 % perkembangan sosial emosi siswa yang masih kurang, namun secara keseluruhan perkembangan sosial emosi siswa di SD Muhammadiyah Pangkalpinang sudah baik. Gambar 12 Diagram aktivitas fisik Berdasarkan gambar tersebut menunjukan bahwa 76% aktivitas fisik siswa terpenuhi, namun 24% aktivitas fisik siswa masih belum terpenuhi atau belum dapat diterapkan. Sehingga perlu adanya peningkatan aktivitas fisik siswa sekolah tersebut. Gambar 13 Diagram ekstrakurikuler Berdasarkan gambar tersebut menunjukan bahwa 24% siswa menjawab ekstrakurikuler terpenuhi, namun 76% masih belum ada. Sehingga sangat perlu adanya peningkatan program ekstrakurikuler di sekolah tersebut. Gambar 14 Diagram status gizi Berdasarkan gambar di atas menunjukan bahwa 47% siswa terpenuhi status gizinya, namun 53% masih belum terpenuhi. Sehingga sangat perlu adanya peningkatan program status gizi di sekolah tersebut agar status gizi anak lebih baik. Tabel 22 Persentase Sub variabel SDIT Albina Pangkalpinang pada pelaksanaan full day school SDIT Albina pangkalpinang Sarana dan Prasarana Aktifitas Fisik Sosial Emosi Ekstrakulikurer Status Gizi Ya 62% 70% 82% 95% 76% Tidak 38% 30% 18% 5% 24% 24%76%ekstrakurikuler yatidak47%53%status giziyatidakGustin, et all 196 Tabel 23 Persentase Sub variabel SD Muhammadiyah Pangkalpinang pada pelaksanaan pra full day school SD Muhammadiyah Pangkalpinang Sarana dan Prasarana Aktifitas Fisik Sosial Emosi Ekstrakulikurer Status Gizi Ya 38% 76% 76% 24% 47% Tidak 62% 24% 24% 76% 53% Kesimpulan Data yang dianalisis dengan menggunakan deskripsi data kuisioner ditemukan bahwa analisis pelaksanaan full day school di SDIT Albina Pangkalpinang dan pra full day school di SD Muhammadiyah pangkalpinang bahwa pada kategori sangat tinggi 34 siswa yaitu 55%, kategori tinggi sebanyak 19 siswa yaitu 31%, kategori rendah sebanyak 9 siswa yaitu 14%, dan kategori sangat rendah sebanyak 0siswa yaitu 0%. Hal ini menunjukkan bahwa analisis Pelaksanaan Program full day school dan pra full day school: Studi Kasus di SDIT Albina Pangkalpinang dan SD Muhammadiyah Pangkalpinang pada kategori sangat tinggi yang ditunjukkan oleh persentase terbesar perolehan skor tersebut menunjukan kondisi yang baik. Berdasarkan analisis tersebut maka Perlunya output yang idealnya dari pelaksanaan program full day school dan pra full day school yakni mengoptimalkan satuan kurikulum pembelajaran dengan waktu belajar, menyediakan fasilitas sarana dan prasarana, peningkatan program ekstrakurikuler dan pemantapan dalam menerapkan budaya sekolah dengan landasan pendidikan karakter bagi peserta didik, menentukan standart waktu untuk kegiatan ekstrakurikuler, menyediakan kantin sekolah dan klinik di sekolah yang dikelola oleh sekolah, agar pemantauan gizi peserta didik teratur, adanya kesepakatan dan komitmen sekolah dengan warga sekolah. Ucapan terima kasih Ucapan terima kasih peneliti kepada semua pihak, khususnya SDIT Albina Pangkalpinang dan SD Muhammadiyah pangkalpinang karena telah membantu menyelesaikan penelitian ini yaitu analisis full day school dan pra full day school. Diharapkan dengan adanya penelitian ini maka di masa yang akan datang dapat menjadi upaya kebijakan sehingga pengembangan program yang mendukung kegiatan sekolah sehingga seluruh aspek lingkungan sekolah lebih memadai. Selain itu, diharapkan juga guru dapat memperhatikan siswa terkait penerapan sistem di sekolah agar siswa dapat meningkatkan kemampuan akademik dan non akademik secara menyeluruh. Referensi Astuti, M. (2013). Implementasi Program Fullday School Sebagai Usaha Mendorong Perkembangan Sosial Peserta Didik TK Unggulan Al-Ya’lu Kota Malang. Jurnal Kebijakan Dan Pengembangan Pendidikan, 1(2), Article 2. https://doi.org/10.22219/jkpp.v1i2.1561 Cendekia, M. S., Dr. Drs. Ismail Nurdin, M. S., Dra. Sri Hartati, M. S., & 228/JTI/2019, A. I. (2019). METODOLOGI PENELITIAN SOSIAL. MEDIA SAHABAT CENDEKIA. https://books.google.co.id/books?id=tretDwAAQBAJ Pelaksanaan Full Day School dan Pra Full Day School: Studi Kasus di SDIT Albina Pangkalpinang dan SD Muhammadiyah Pangkalpinang 197 Dewi, M. P., S, N., & Irdamurni, I. (2020). PERKEMBANGAN BAHASA, EMOSI, DAN SOSIAL ANAK USIA SEKOLAH DASAR. Jurnal Ilmiah Pendidikan Dasar, 7(1), 1–11. https://doi.org/10.30659/pendas.7.1.1-11 Hawi, H. A. (2015). Sistem Full-Day School di Sekolah Dasar Islam Terpadu (SDIT) Studi Kasus di Izzuddin Palembang. Jurnal Istinbath, 15(2). Lalu Juntra Utama, L. J. U., SST, & Yohanes Don Bosko Demu, Y. D. B. D., S. KM. (2021). Dasar-Dasar Penanganan Gizi Anak Sekolah. Media Sains Indonesia. Mohamad Yudiyanto, M. Y., S. kom. (2021). Revitalisasi Peran Ekstrakurikuler Keagamaan di Sekolah. Rinda Fauzian. Mohammad Al Farabi, D. M. A. F. (2018). Pendidikan Orang Dewasa Dalam Al-Qur’an. Kencana. Neal, J. W., Brutzman, B., & Durbin, C. E. (2022). The role of full- and half-day preschool attendance in the formation of children’s social networks. Early Childhood Research Quarterly, 60, 394–402. https://doi.org/10.1016/j.ecresq.2022.04.003 Parid, M., & Alif, A. L. S. (2020). Pengelolaan Sarana dan Prasarana Pendidikan. Tafhim Al-’Ilmi, 11(2), 266–275. https://doi.org/10.37459/tafhim.v11i2.3755 Parra, J. D. (2022). Decentralisation and school-based management in Colombia: An exploration (using systems thinking) of the Full‐Day Schooling programme. International Journal of Educational Development, 91, 102579. https://doi.org/10.1016/j.ijedudev.2022.102579 Rosa, L., Bettinger, E., Carnoy, M., & Dantas, P. (2022). The effects of public high school subsidies on student test scores: The case of a full-day high school in Pernambuco, Brazil. Economics of Education Review, 87, 102201. https://doi.org/10.1016/j.econedurev.2021.102201 Soapatty, L., & Suwanda, T. (2014). Pengaruh sistem sekolah sehari penuh (full day school) terhadap prestasi akademik siswa SMP Jati Agung Sidoarjo. Kajian Moral Dan Kewarganegaraan, 2(2), 717–733. Sudaryono. (2016). Metode Penelitian Pendidikan. Prenada Media. Thompson, J. A., & Sonnenschein, S. (2016). Full-day kindergarten and children’s later reading: The role of early word reading. Journal of Applied Developmental Psychology, 42, 58–70. https://doi.org/10.1016/j.appdev.2015.11.005 Vashdev, G. (2016). Setiap Waktu adalah Waktu Belajar—Happiness Inside (Snackbook). Noura e-Lite. | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/236248028Libsharp - spherical harmonic transforms revisitedArticle in Astronomy and Astrophysics · March 2013DOI: 10.1051/0004-6361/201321494 · Source: arXivCITATIONS60READS2282 authors, including:Dag Sverre SeljebotnUniversity of Oslo8 PUBLICATIONS 1,090 CITATIONS SEE PROFILEAll content following this page was uploaded by Dag Sverre Seljebotn on 08 August 2014.The user has requested enhancement of the downloaded file.arXiv:1303.4945v2 [physics.comp-ph] 22 Apr 2013Astronomy & Astrophysics manuscript no. paperc⃝ ESO 2013April 23, 2013Libsharp – spherical harmonic transforms revisitedM. Reinecke1 and D. S. Seljebotn21 Max-Planck-Institut f¨ur Astrophysik, Karl-Schwarzschild-Str. 1, 85741 Garching, Germanye-mail: martin@mpa-garching.mpg.de2 Institute of Theoretical Astrophysics, University of Oslo, P.O. Box 1029 Blindern, N-0315 Oslo, Norwaye-mail: d.s.seljebotn@astro.uio.noReceived 18 March 2013 / Accepted 14 April 2013ABSTRACTWe present libsharp, a code library for spherical harmonic transforms (SHTs), which evolved from the libpsht library and ad-dresses several of its shortcomings, such as adding MPI support for distributed memory systems and SHTs of fields with arbitraryspin, but also supporting new developments in CPU instruction sets like the Advanced Vector Extensions (AVX) or fused multiply-accumulate (FMA) instructions. The library is implemented in portable C99 and provides an interface that can be easily accessed fromother programming languages such as C++, Fortran, Python, etc. Generally, libsharp’s performance is at least on par with that of itspredecessor; however, significant improvements were made to the algorithms for scalar SHTs, which are roughly twice as fast whenusing the same CPU capabilities. The library is available at http://sourceforge.net/projects/libsharp/ under the terms ofthe GNU General Public License.Key words. methods: numerical – cosmic background radiation – large-scale structure of the Universe1. MotivationWhile the original libpsht library presented by Reinecke(2011) fulfilled most requirements on an implementation ofspherical harmonic transforms (SHTs) in the astrophysical con-text at the time, it still left several points unaddressed. Some ofthose were already mentioned in the original publication: sup-port for SHTs of arbitrary spins and parallelisation on computerswith distributed memory.Both of these features have been added to libpsht in themeantime, but other, more technical, shortcomings of the libraryhave become obvious since its publication, which could not befixed within the libpsht framework.One of these complications is that the library design didnot anticipate the rapid evolution of microprocessors during thepast few years. While the code supports both traditional scalararithmetic as well as SSE2 instructions, adding support for thenewly released Advanced Vector Extensions (AVX) and fusedmultiply-accumulate instructions (FMA3/FMA4) would requireadding significant amounts of new code to the library, which isinconvenient and very likely to become a maintenance burden inthe long run. Using proper abstraction techniques, adding a newset of CPU instructions could be achieved by only very smallchanges to the code, but the need for this was unfortunately notanticipated when libpsht was written.Also, several new, highly efficient SHT implementationshave been published in the meantime; most notably Wavemoth(Seljebotn 2012) and shtns (Schaeffer 2013). These codesdemonstrate that libpsht’s computational core did not makethe best possible use of the available CPU resources. Note thatWavemoth is currently an experimental research code not meantfor general use.To address both of these concerns, the library was redesignedfrom scratch. The internal changes also led to a small loss offunctionality; the new code no longer supports multiple simul-taneous SHTs of different type (i.e. having different directionsor different spins). Simultaneous transforms of identical type arestill available, however.As a fortunate consequence of this slight reduction in func-tionality, the user interface could be simplified dramatically,which is especially helpful when interfacing the library withother programming languages.Since backward compatibility is lost, the new namelibsharp was chosen for the resulting code, as a shorthand for“Spherical HARmonic Package”.The decision to develop libsharp instead of simply usingshtns was taken for various reasons: shtns does not supportspin SHTs or allow MPI parallelisation, it requires more mainmemory than libsharp, which can be problematic for high-resolution runs, and it relies on the presence of the FFTW li-brary. Also, shtns uses a syntax for expressing SIMD opera-tions, which is currently only supported by the gcc and clangcompilers, thereby limiting its portability at least for the vec-torised version. Finally, libsharp has support for partial spher-ical coverage and a wide variety of spherical grids, includingGauss-Legendre, ECP, and HEALPix.2. Problem definitionThis section contains a quick recapitulation of equations pre-sented in Reinecke (2011), for easier reference.We assume a spherical grid with Nϑ iso-latitude rings (in-dexed by y). Each of these in turn consists of Nϕ,y pixels (indexedby x), which are equally spaced in ϕ, the azimuth of the first ringpixel being ϕ0,y.A continuous spin-s function defined on the sphere with aspectral band limit of lmax can be represented either as a set ofspherical harmonic coefficients salm, or a set of pixels pxy. Thesetwo representations are connected by spherical harmonic synthe-1M. Reinecke and D. S. Seljebotn: Libsharp – spherical harmonic transforms revisitedsis (or backward SHT):pxy =lmax�m=−lmaxlmax�l=|m|salm sλlm(ϑy) exp�imϕ0,y + 2πimxNϕ,y�(1)and spherical harmonic analysis (or forward SHT):sˆalm =Nϑ−1�y=0Nϕ,y−1�x=0pxy wy sλlm(ϑy) exp�−imϕ0,y − 2πimxNϕ,y�,(2)where sλlm(ϑ) := sYlm(ϑ, 0) and wy are quadrature weights.Both transforms can be subdivided into two stages:pxy =lmax�m=−lmaxFm,y exp�imϕ0,y + 2πimxNϕ,y�with(3)Fm,y :=lmax�l=|m|salm sλlm(ϑy),and(4)sˆalm =Nϑ−1�y=0Gm,y sλlm(ϑy)with(5)Gm,y := wyNϕ,y−1�x=0pxy exp�−imϕ0,y − 2πimxNϕ,y�.(6)Eqs. 3 and 6 can be computed using fast Fourier transforms(FFTs), while Eqs. 4 and 5, which represent the bulk of the com-putational load, are the main target for optimised implementa-tion in libsharp.3. Technical improvements3.1. General remarksAs the implementation language for the new library, ISO C99was chosen. This version of the C language standard is moreflexible than the C89 one adopted for libpsht and has gainedubiquitous compiler support by now. Most notably, C99 allowsdefinition of new variables anywhere in the code, which im-proves readability and eliminates a common source of program-ming mistakes. It also provides native data types for complexnumbers, which allows for a more concise notation in manyplaces. However, special care must be taken not to make use ofthese data types in the library’s public interface, since this wouldprevent interoperability with C++ codes (because C++ has a dif-ferent approach to complex number support). Fortunately, thisdrawback can be worked around fairly easily.A new approach was required for dealing with the growingvariety of instruction sets for arithmetic operations, such as tradi-tional scalar instructions, SSE2 and AVX. Rewriting the librarycore for each of these alternatives would be cumbersome anderror-prone.Instead we introduced the concept of a generic “vec-tor” type containing a number of double-precision IEEE val-ues and defined a set of abstract operations (basic arithmetics,negation, absolute value, multiply-accumulate, min/max, com-parison, masking and blending) for this type. Depending on theconcrete instruction set used when compiling the code, these op-erations are then expressed by means of the appropriate opera-tors and intrinsic function calls. The only constraint on the num-ber of values in the vector type is that it has to be a multiple ofthe underlying instruction set’s native vector length (1 for scalararithmetic, 2 for SSE2, 4 for AVX).for b = <all submaps or "blocks">for m = [0;mmax]// OpenMPfor l = [m;lmax]precompute recursion coefficientsend lfor y = <all rings in submap b>// SSE/AVXfor l = [m;lmax]compute s_lambda_lm(theta_y)for j = <all jobs>accumulate F(m,theta_y,j)end jend lend yend mfor y = <all rings in submap b>// OpenMPfor j = <all jobs>compute map(x,y,j) using FFTend jend yend bFig. 1. Schematic loop structure of libsharp’s shared-memorysynthesis code.for b = <all submaps or "blocks">for y = <all rings in submap b>// OpenMPfor j = <all jobs>compute G(m,theta_y,j) using FFTend jend yfor m = [0;mmax]// OpenMPfor l = [m;lmax]precompute recursion coefficientsend lfor y = <all rings in submap b>// SSE/AVXfor l = [m;lmax]compute s_lambda_lm(theta_y)for j = <all jobs>compute contribution to a_lm(j)end jend lend yend mend bFig. 2. Schematic loop structure of libsharp’s shared-memoryanalysis code.Using this technique, adding support for new vector instruc-tion sets is straightforward and carries little risk of breaking ex-isting code. As a concrete example, support for the FMA4 in-structions present in AMD’s Bulldozer CPUs was added andsuccessfully tested in less than an hour.3.2. Improved loop structureAfter publication of SHT implementations, which perform sig-nificantly better than libpsht, especially for s = 0 transforms(Seljebotn 2012; Schaeffer 2013), it became obvious that some2M. Reinecke and D. S. Seljebotn: Libsharp – spherical harmonic transforms revisitedbottleneck must be present in libpsht’s implementation. Thiswas identified with libpsht’s approach of first computing awhole l-vector of sλlm(ϑ) in one go, storing it to main mem-ory, and afterwards re-reading it sequentially whenever needed.While the l-vectors are small enough to fit into the CPU’s Level-1 cache, the store and load operations nevertheless caused somelatency. For s = 0 transforms with their comparatively low arith-metic operation count (compared to the amount of memory ac-cesses), this latency could not be hidden behind floating pointoperations and so resulted in a slow-down. This is the most likelyexplanation for the observation that libpsht’s s = 0 SHTs havea much lower FLOP rate compared to those with s � 0.It is possible to avoid the store/load overhead for the sλlm(ϑ)by applying each value immediately after it has been computed,and discarding it as soon as it is not needed any more. This ap-proach is reflected in the loop structure shown in Figs. 1 and2, which differs from the one in Reinecke (2011) mainly by thefusion of the central loops over l.In this context another question must be addressed: the loopsmarked as “SSE/AVX” in both figures are meant to be executedin “blocks”, i.e. by processing several y indices simultaneously.The block size is equivalent to the size of the generic vectortype described in Sect. 3.1. The best value for this parameterdepends on hardware characteristics of the underlying computerand therefore cannot be determined a priori. Libsharp alwaysuses a multiple of the system’s natural vector length and esti-mates the best value by running quick measurements whenevera specific SHT is invoked for the first time. This auto-tuning stepapproximately takes a tenth of a wall-clock second.Due to the changed central loop of the SHT implementation,it is no longer straightforward to support multiple simultaneoustransforms with differing spins and/or directions, as libpshtdid – this would lead to a combinatorial explosion of loop bodiesthat have to be implemented. Consequently, libsharp, whilestill supporting simultaneous SHTs, restricts them to have thesame spin and direction. Fortunately, this is a very common casein many application scenarios.3.3. Polar optimisationAs previously mentioned in Reinecke (2011), much CPU timecan be saved by simply not computing terms in Eqs. (4) and(5) for which sλlm(ϑ) is so small that their contribution to theresult lies well below the numerical accuracy. Since this situationoccurs for rings lying close to the poles and high values of m,Schaeffer (2013) referred to it as “polar optimisation”.To determine which terms can be omitted, libsharp usesthe approach described in Pr´ezeau & Reinecke (2010). In short,all terms for which√m2 + s2 − 2ms cos ϑ − lmax sin ϑ > T(7)are skipped. The parameter T is tunable and determines the over-all accuracy of the result. Libsharp models it asT = max(100, 0.01lmax),(8)which has been verified to produce results equivalent to those ofSHTs without polar optimisation.4. Added functionality4.1. SHTs with arbitrary spinWhile the most widely used SHTs in cosmology are performedon quantities of spins 0 and 2 (i.e. sky maps of Stokes I andQ/U parameters), there is also a need for transforms at otherspins. Lensing computations require SHTs of spin-1 and spin-3quantities (see, e.g., Lewis 2005). The most important motiva-tion for high-spin SHTs, however, are all-sky convolution codes(e.g. Wandelt & G´orski 2001; Pr´ezeau & Reinecke 2010) anddeconvolution map-makers (e.g. Keih¨anen & Reinecke 2012).The computational cost of these algorithms is dominated by cal-culating expressions of the formRmk(ϑ) =lmax�l=max(|m|,|k|)almb∗lkdlmk(ϑ),(9)where a and b denote two sets of spherical harmonic coeffi-cients (typically of the sky and a beam pattern) and d are theWigner d matrix elements. These expressions can be interpretedand solved efficiently as a set of (slightly modified) SHTs withspins ranging from 0 to kmax ≤ lmax, which in today’s applica-tions can take on values much higher than 2.As was discussed in Reinecke (2011), the algorithms usedby libpsht for spin-1 and spin-2 SHTs become inefficientand inaccurate for higher spins. To support such transforms inlibsharp, another approach was therefore implemented, whichis based on the recursion for Wigner d matrix elements presentedin Pr´ezeau & Reinecke (2010).Generally, the spin-weighted spherical harmonics are relatedto the Wigner d matrix elements viasλlm(ϑ) = (−1)m�2l + 14π dl−ms(ϑ)(10)(Goldberg et al. 1967). It is possible to compute the dlmm′(ϑ) us-ing a three-term recursion in l very similar to that for the scalarYlm(ϑ):�cos ϑ −mm′l(l + 1)�dlmm′(ϑ)=�(l2 − m2)(l2 − m′2)l(2l + 1)dl−1mm′(ϑ)+�[(l + 1)2 − m2][(l + 1)2 − m′2](l + 1)(2l + 1)dl+1mm′(ϑ)(11)(Kostelec & Rockmore 2008). The terms depending only on l,m, and m′ can be re-used for different colatitudes, so that thereal cost of a recursion step is two additions and three multipli-cations.In contrast to the statements made in McEwen & Wiaux(2011), this recursion is numerically stable when performed inthe direction of increasing l; see, e.g., Sect. 5.1.2 for a practicalconfirmation. It is necessary, however, to use a digital floating-point representation with enhanced exponent range to avoid un-derflow during the recursion, as is discussed in some detail inPr´ezeau & Reinecke (2010).4.2. Distributed memory parallelisationWhen considering that, in current research, the required bandlimit for SHTs practically never exceeds lmax = 104, it seemsat first glance unnecessary to provide an implementation sup-porting multiple nodes. Such SHTs fit easily into the memoryof a single typical compute node and are carried out within afew seconds of wall clock time. The need for additional paral-lelisation becomes apparent, however, as soon as the SHT is nolonger considered in isolation, but as a (potentially small) part3M. Reinecke and D. S. Seljebotn: Libsharp – spherical harmonic transforms revisitedfor m = <all local m>// OpenMPfor l = [m;lmax]precompute recursion coefficientsend lfor y = <all rings in the map>// SSE/AVXfor l = [m;lmax]compute s_lambda_lm(theta_y)for j = <all jobs>accumulate F(m,theta_y,j)end jend lend yend mrearrange F(m,theta_y,j) among tasks // MPIfor y = <all local rings>// OpenMPfor j = <all jobs>compute map(x,y,j) using FFTend jend yFig. 3. Schematic loop structure of libsharp’s distributed-memory synthesis code.for y = <all local rings>// OpenMPfor j = <all jobs>compute G(m,theta_y,j) using FFTend jend yrearrange G(m,theta_y,j) among tasks // MPIfor m = <all local m>// OpenMPfor l = [m;lmax]precompute recursion coefficientsend lfor y = <all rings in the map>// SSE/AVXfor l = [m;lmax]compute s_lambda_lm(theta_y)for j = <all jobs>compute contribution to a_lm(j)end jend lend yend mFig. 4. Schematic loop structure of libsharp’s distributed-memory analysis code.of another algorithm, which is libsharp’s main usage scenario.In such a situation, large amounts of memory may be occupiedby data sets unrelated to the SHT, therefore requiring distribu-tion over multiple nodes. Moreover, there is sometimes the needfor very many SHTs in sequence, e.g. if they are part of a sam-pling process or an iterative solver. Here, the parallelisation toa very large number of CPUs may be the only way of reduc-ing the time-to-solution to acceptable levels. Illustrative exam-ples for this are the Commander code (Eriksen et al. 2008) andthe artDeco deconvolution mapmaker (Keih¨anen & Reinecke2012); for the processing of high-resolution Planck data, the lat-ter is expected to require over 100GB of memory and severalhundred CPU cores.Libsharp provides an interface that allows collective execu-tion of SHTs on multiple machines with distributed memory. Itmakes use of the MPI1 interface to perform the necessary inter-process communication.In contrast to the standard, shared-memory algorithms, itis the responsibility of the library user to distribute map dataand alm over the individual computers in a way that ensuresproper load balancing. A very straightforward and reliable wayto achieve this is a “round robin” strategy: assuming N comput-ing nodes, the map is distributed such that node i hosts the maprings i, i+N, i+2N, etc. (and their counterparts on the other hemi-sphere). Similarly, for the spherical harmonic coefficients, nodei would hold all alm for m = i, i + N, i + 2N, etc. Other efficientdistribution strategies do of course exist and may be advanta-geous, depending on the circumstances under which libsharpis called; the library makes no restrictions in this respect.The SHT algorithm for distributed memory architectures isanalogous to the one used in the S2HAT package2 and firstpublished in Szydlarski et al. (2013); its structure is sketchedin Figs. 3 and 4. In addition to the S2HAT implementation,the SHT will be broken down into smaller chunks if the av-erage number of map rings per MPI task exceeds a certainthreshold. This is analogous to the use of chunks in the scalarand OpenMP-parallel implementations and reduces the memoryoverhead caused by temporary variables.It should be noted that libsharp supports hybrid MPI andOpenMP parallelisation, which allows, e.g., running an SHT oneight nodes with four CPU cores each, by specifying eight MPItasks, each of them consisting of four OpenMP threads. In gen-eral, OpenMP should be preferred over MPI whenever sharedmemory is available (i.e. at the computing node level), since theOpenMP algorithms contain dynamic load balancing and have asmaller communication overhead.4.3. Map synthesis of first derivativesGenerating maps of first derivatives from a set of alm is closelyrelated to performing an SHT of spin 1. A specialised SHT modewas added to libsharp for this purpose; it takes as input aset of spin-0 alm and produces two maps containing ∂ f/∂ϑ and∂ f/(∂ϕ sin ϑ), respectively.4.4. Support for additional spherical gridsDirect support for certain classes of spherical grids has been ex-tended in comparison to libpsht; these additions are listed be-low in detail. It must be stressed, however, that libsharp can –very much as libpsht does – perform SHTs on any iso-latitudegrid with equidistant pixels on each ring. This very general classof pixelisations includes, e.g., certain types of partial sphericalcoverage. For these general grids, however, the user is responsi-ble for providing correct quadrature weights when performing aspherical harmonic analysis.4.4.1. Extended support for ECP gridsLibpsht provides explicit support for HEALPix grids, Gauss-Legendre grids, and a subset of equidistant cylindrical projection1 http://en.wikipedia.org/wiki/Message_Passing_Interface2 http://code.google.com/p/s2hat-library/4M. Reinecke and D. S. Seljebotn: Libsharp – spherical harmonic transforms revisited(ECP) grids. The latter are limited to an even number of rings atthe colatitudesϑn = (n + 0.5)πN,n ∈ [0; N − 1].(12)The associated quadrature weights are given by Fej´er’s first rule(Fej´er 1933; Gautschi 1967).Libsharp extends ECP grid support to allow even and oddnumbers of rings, as well as the colatitude distributionsϑn = nπN ,n ∈ [1; N − 1](13)(corresponding to Fej´er’s second rule), andϑn = nπN ,n ∈ [0; N](14)(corresponding to Clenshaw-Curtis quadrature). This last pixeli-sation is identical to the one adopted in Huffenberger & Wandelt(2010).Accurate computation of the quadrature weights for thesepixelisations is nontrivial; libsharp adopts the FFT-based ap-proach described in Waldvogel (2006) for this purpose.4.4.2. Reduced Gauss-Legendre gridThe polar optimisation described in Sect. 3.3 implies that it ispossible to reduce the number of pixels per ring below the the-oretically required value of 2lmax + 1 close to the poles. Eq. (7)can be solved for m (at a given s, lmax and ϑ), and using 2m + 1equidistant pixels in the corresponding map ring results in a pix-elisation that can represent a band-limited function up to the de-sired precision, although it is no longer exact in a mathemati-cal sense. If this number is further increased to the next numbercomposed entirely of small prime factors (2, 3, and 5 are used inlibsharp’s case), this has the additional advantage of allowingvery efficient FFTs.Libsharp supports this pixel reduction technique in theform of a thinned-out Gauss-Legendre grid. At moderate to highresolutions (lmax > 1000), more than 30% of pixels can be saved,which can be significant in various applications.It should be noted that working with reduced Gauss-Legendre grids, while saving considerable amounts of mem-ory, does not change SHT execution times significantly; all po-tential savings are already taken into account, for all grids, bylibsharp’s implementation of polar optimisation.4.5. Adjoint and real SHTsSince Eq. (1) is a linear transform, we can introduce the notationp = Ya(15)for a vector a of spherical harmonic coefficients and correspond-ing vector p of pixels. Similarly, one can write Eq. (2) asa = YTWp,(16)where W is a diagonal matrix of quadrature weights. When in-cluding SHTs as operators in linear systems, one will often needthe adjoint spherical harmonic synthesis, YT, and the adjointspherical harmonic analysis, WY. For instance, if a is a ran-dom vector with covariance matrix C in the spherical harmonicdomain, then its pixel representation Ya has the covariance ma-trix YCYT. Multiplication by this matrix requires the use of theadjoint synthesis, which corresponds to analysis with a differ-ent choice of weights. Libsharp includes routines for adjointSHTs, which is more user-friendly than having to compensatefor the wrong choice of weights in user code, and also avoids anextra pass over the data.For linear algebra computations, the vector a must also in-clude alm with m < 0, even if libsharp will only computethe coefficients for m ≥ 0. The use of the real spherical har-monics convention is a convenient way to include negative mwithout increasing the computational workload by duplicatingall coefficients. For the definition we refer to the appendix ofde Oliveira-Costa et al. (2004). The convention is supported di-rectly in libsharp, although with a restriction in the storagescheme: The coefficients for m < 0 must be stored in the samelocations as the corresponding imaginary parts of the complexcoefficients, so that the pattern in memory is [al,m, al,−m].5. EvaluationMost tests were performed on the SuperMUC Petascale Systemlocated at the Leibniz-Rechenzentrum Garching. This systemconsists of nodes containing 32GB of main memory and 16 IntelXeon E5-2680 cores running at 2.7GHz. The exception is thecomparison with Wavemoth, which was performed on the Abelcluster at the University of Oslo on very similar hardware; XeonE5-2670 at 2.6 GHz.The code was compiled with gcc version 4.7.2. The Intelcompiler (version 12.1.6) was also tested, but produced slightlyinferior code.Except where noted otherwise, test calculations were per-formed using the reduced Gauss-Legendre grid (see Sect. 4.4.2)to represent spherical map data. This was done for the pragmaticpurpose of minimising the tests’ memory usage, which allowedgoing to higher band limits in some cases, as well as to demon-strate the viability of this pixelisation.The band limits adopted for the tests all obey lmax = 2n − 1with n ∈ N (except for those presented in Sect. 5.2.2). This isdone in analogy to most other papers on the subject, but leads tosome unfamiliar numbers especially at very high lmax.The number of cores used for any particular run always is apower of 2.5.1. Accuracy tests5.1.1. Comparison with other implementationsThe numerical equivalence of libsharp’s SHTs to existing im-plementations was verified by running spherical harmonic syn-thesis transforms on a Gauss-Legendre grid at lmax = 50 andspins 0, 1, and 2 with both libsharp and libpsht, and com-paring the results. The differences of the results lay well withinthe expected levels of numerical errors caused by the finite preci-sion of IEEE numbers. Spherical harmonic analysis is implicitlytested by the experiments in the following sections.5.1.2. Evaluation of SHT pairsTo test the accuracy of libsharp’s transforms, sets of spin=0alm coefficients were generated by setting their real and imagi-nary parts to numbers drawn from a uniform random distribu-tion in the range [-1;1[ (with exception of the imaginary partsfor m = 0, which have to be zero for symmetry reasons). Thisdata set was transformed onto a reduced Gauss-Legendre gridand back to spherical harmonic space again, resulting in ˆalm.5M. Reinecke and D. S. Seljebotn: Libsharp – spherical harmonic transforms revisited10-1610-1510-1410-1310-1210-1110-1010-910-8100101102103104105SHT errorlmaxlmaxlmax3/2εrmsεmaxFig. 5. Maximum and rms errors for inverse/forward spin=0SHT pairs at different lmax.The rms and maximum errors of this inverse/forward trans-form pair can be written asεrms :=��lm |salm − sˆalm|2�lm |salm|2and(17)εmax := maxlm |salm − sˆalm| .(18)Fig. 5 shows the measured errors for a wide range of bandlimits. As expected, the numbers are close to the accuracy limitof double precision IEEE numbers for low lmax; rms errors in-crease roughly linearly with the band limit, while the maximumerror seems to exhibit an l3/2max scaling. Even at lmax = 262143(which is extremely high compared to values typically requiredin cosmology), the errors are still negligible compared with theuncertainties in the input data in today’s experiments.Analogous experiments were performed for spins 2 and 37,with very similar results (not shown).5.2. Performance testsDetermining reliable execution times for SHTs is nontrivial atlow band limits, since intermittent operating system activity cansignificantly distort the measurement of short time scales. Alllibsharp timings shown in the following sections were ob-tained using the following procedure: the SHT pair in questionis executed repeatedly until the accumulated wall-clock time ex-ceeds 2 seconds. Then the shortest measured wall-clock time forsynthesis and analysis is selected from the available set.5.2.1. Strong scaling testTo assess strong scaling behaviour (i.e. run time scaling for agiven problem with fixed total workload), a spin=2 SHT withlmax=16383 was carried out with differing degrees of parallelisa-tion. The accumulated wall-clock time of these transforms (syn-thesis + analysis) is shown in Fig. 6. It is evident that the scalingis nearly ideal up to 16 cores, which implies that parallelisationoverhead is negligible in this range. Beyond 16 cores, MPI com-munication has to be used for inter-node communication, andthis most likely accounts for the sudden jump in accumulatedtime. At even higher core counts, linear scaling is again reached,although with a poorer proportionality factor than in the intra-node case. Finally, for 1024 and more cores, the communicationtime dominates the actual computation, and scalability is lost. 0 0.5 1 1.5 2 2.5 3 1 10 100 1000Twall Ncore [h]NcoreFig. 6. Strong scaling scenario: accumulated wall-clock time fora spin=2 SHT pair with lmax=16383 run on various numbers ofcores. 0 5 10 15 20 25 30 1 10 100 1000Twall [s]NcoreFig. 7. Weak scaling scenario: wall-clock time for a spin=0 SHTpair run on various numbers of cores, with constant amount ofwork per core (lmax(N) = 4096 · N1/3 − 1).5.2.2. Weak scaling testWeak scaling behaviour of the algorithm is investigated bychoosing problem sizes that keep the total work per core con-stant, in contrast to a fixed total workload. Assuming an SHTcomplexity of O(l3max), the band limits were derived from theemployed number of cores N via lmax(N) = 4096 · N1/3 − 1. Theresults are shown in Fig. 7. Ideal scaling corresponds to a hori-zontal line. Again, the transition from one to several computingnodes degrades performance, whereas scaling on a single node,as well as in the multi-node range, is very good. By keeping theamount of work per core constant, the breakdown of scalabil-ity is shifted to 4096 cores, compared with 1024 in the strongscaling test.It is interesting to note that the scaling within a single nodeis actually slightly superlinear; this is most likely because in thissetup, the amount of memory per core decreases with increasingproblem size, which in turn can improve the amount of cachere-use and reduce memory bandwidth per core.5.2.3. General scaling and efficiencyThe preceding two sections did not cover cases with small SHTs.This scenario is interesting, however, since in the limit of smalllmax those components of the SHT implementation with com-plexities lower than O(l3max) (like the FFT steps of Eqs. 3 and6M. Reinecke and D. S. Seljebotn: Libsharp – spherical harmonic transforms revisited10-610-410-2100102104106108100101102103104105Twall Ncore [s]lmaxlmax3s=0s=2Ncore=1Ncore=16Ncore=64Ncore=512Ncore=4096Fig. 8. Accumulated wall-clock time for spin=0 and spin=2 SHTpairs at a wide range of different band limits. For every run thenumber of cores was chosen sufficiently small to keep paralleli-sation overhead low. 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7100101102103104105fraction of theoretical peaklmaxs=0s=2Ncore=1Ncore=16Ncore=64Ncore=512Ncore=4096Fig. 9. Fraction of theoretical peak performance reached by var-ious SHT pairs. For every run the number of cores was chosensufficiently small to keep parallelisation overhead low.6) may begin to dominate execution time. Fig. 8 shows the totalwall-clock time for SHT pairs over a very wide range of bandlimits; to minimise the impact of communication, the degree ofparallelisation was kept as low as possible for the runs in ques-tion. As expected, the l3max scaling is a very good model forthe execution times at lmax ≥ 511. Below this limit, the FFTs,precomputations for the spherical harmonic recursion, memorycopy operations and other parts of the code begin to dominate.In analogy to one of the tests described in Reinecke (2011),we computed a lower limit for the number of executed floating-point operations per second in libsharp’s SHTs and comparedthe result with the theoretical peak performance achievable onthe given hardware, which is eight operations per clock cycle(four additions and four multiplications) or 21.6 GFlops/s percore. Fig. 9 shows the results. In contrast to libpsht, whichreached approximately 22% for s = 0 and 43% for s = 2, bothscalar and tensor harmonic transforms exhibit very similar per-formance levels and almost reach 70% of theoretical peak in themost favourable regime, thanks to the changed structure of theinner loops. For the lmax range that is typically required in cos-mological applications, performance exceeds 50% (even whenMPI is used), which is very high for a practically useful algo-rithm on this kind of computer architecture. 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1 2 3 4 5 6 7 8 9 10speedup factornshts=0s=2Fig. 10. Relative speed-up when performing several SHTs simul-taneously, compared with sequential execution. The SHT had aband limit of lmax = 8191. 0 5 10 15 20 25 30 35 40 45104105memory overhead [%]lmaxs=0s=2Fig. 11. Relative memory overhead, i.e. the fraction of totalmemory that is not occupied by input and output data of the SHT.For low lmax this is dominated by the program binary, for highlmax by temporary arrays.5.2.4. Multiple simultaneous SHTsThe computation of the sλlm(ϑ) coefficients accounts for roughlyhalf the arithmetic operations in an SHT. If several SHTs withidentical grid geometry and band limit are computed simulta-neously, it is possible to re-use these coefficients, thereby re-ducing the overall operation count. Fig. 10 shows the speed-upscompared to sequential execution for various scenarios, whichincrease with the number of transforms and reach saturationaround a factor of 1.6. This value is lower than the na¨ıvely ex-pected asymptotic factor of 2 (corresponding to avoiding half ofthe arithmetic operations), since the changed algorithm requiresmore memory transfers between Level-1 cache and CPU regis-ters and therefore operates at a lower percentage of theoreticalpeak performance. Nevertheless, running SHTs simultaneouslyis evidently beneficial and should be used whenever possible.5.2.5. Memory overheadEspecially at high band limits, it is important that the SHT li-brary does not consume a large amount of main memory, toavoid memory exhaustion and subsequent swapping or codecrashes. Libsharp is designed with the goal to keep the size ofits auxiliary data structures much lower than the combined sizeof any SHT’s input and output arrays. A measurement is shown7M. Reinecke and D. S. Seljebotn: Libsharp – spherical harmonic transforms revisited 0 0.2 0.4 0.6 0.8 1 1.2 1.4101102103104time ratio shtns/libsharplmaxscalarNcore=4Ncore=8Ncore=16Fig. 12. Performance comparison between libsharpand shtnsfor varying lmax and number of OpenMP threads. Note that re-duced autotuning was used for shtns at lmax=16383 (see text).in Fig. 11. Below the lowest shown band limit of 2047, memoryoverhead quickly climbs to almost 100%, since in this regimememory consumption is dominated by the executable and theconstant overhead of the communication libraries, which on thetesting machine amounts to approximately 50MB. In the impor-tant range (lmax ≥ 2047), memory overhead lies below 45%.5.2.6. Comparison with existing implementationsTable 1 shows a performance comparison of synthesis/analysisSHT pairs between libsharp and various other SHT implemen-tations. In addition to the already mentioned shtns, Wavemoth,S2HAT and libpsht codes, we also included spinsfast(Huffenberger & Wandelt2010),SSHT(McEwen & Wiaux2011) and Glesp (Doroshkevich et al. 2005) in the comparison.All computations shared a common band limit of 2047 and wereexecuted on a single core, since the corresponding SHTs are sup-ported by all libraries and are very likely carried out with a com-paratively high efficiency by all of them. The large overall num-ber of possible parameters (lmax, spin, number of simultaneoustransforms, degree and kind of parallelisation, choice of grid,etc.) prevented a truly comprehensive study.Overall, libsharp’s performance is very satisfactory andexhibits speed-ups of more than an order of magnitude in sev-eral cases. The table also demonstrates libsharp’s flexibility,since it supports all of the other codes’ “native” grid geometries,which is required for direct comparisons.The three last columns list time ratios measured under dif-ferent assumptions: RAVX reflects values that can be expectedon modern (2012 and later) AMD/Intel CPUs supporting AVX,RSSE2 applies to older (2001 and later) CPUs with the SSE2 in-struction set. Rscalar should be used for CPUs from other vendorslike IBM or ARM, since libsharp does not yet support vectori-sation for these architectures.Fig. 12 shows the relative performance of identical SHTpairs on a full Gauss-Legendre grid with s=0 for libsharp andshtns. For these measurements the benchmarking code deliv-ered with shtns was adjusted to measure SHT times in a similarfashion as was described above. The plotted quantity is shtnswall-clock time divided by libsharp wall-clock time for vary-ing lmax and number of OpenMP threads. It is evident that shtnshas a significant advantage for small band limits (almost an or-der of magnitude) and maintains a slight edge up to lmax=8191.It must be noted, however, that the measured times do not in-clude the overhead for auto-tuning and necessary precalcula-tions, which in the case of shtns are about an order of mag-nitude more expensive than the SHTs themselves. As a conse-quence, its performance advantage only pays off if many identi-cal SHT operations are performed within one run. The origin ofshtns’s performance advantage has not been studied in depth;however, a quick analysis shows that the measured time differ-ences scale roughly like l2max, so the following explanations arelikely candidates:– libsharp performs all of its precomputations as part of thetime-measured SHT– libsharp’s flexibility with regard to pixelisation and stor-age arrangement of input and output data requires some ad-ditional copy operations– at low band limits the inferior performance of libsharp’sFFT implementation has a noticeable impact on overall runtimes.The relative performance of both libraries is remarkably in-sensitive to the number of OpenMP threads; this indicates thatthe performance differences are located in parallel code regionsas opposed to sequential ones.For lmax = 16383, the time required by the default shtnsautotuner becomes very long (on the order of wall-clock hours),so that we decided to invoke it with an option for reduced tuning.It is likely a consequence of this missed optimisation that, at thisband limit, libsharp is the better-performing code.6. ConclusionsJudging from the benchmarks presented in the preceding sec-tion, the goals that were set for the libsharp library have beenreached: it exceeds libpsht in terms of performance, supportsrecent developments in microprocessor technology, allows usingdistributed memory systems for a wider range of applications,and is slightly easier to use. On the developer side, the modulardesign of the code makes it much more straightforward to addsupport for new instruction sets and other functionality.In some specific scenarios, especially for SHTs with com-paratively low band limits, libsharp does not provide the bestperformance of all available implementations, but given its ex-treme flexibility concerning grid types and the memory layoutof its input/output data, as well as its compactness (≈ 8000 linesof portable and easily maintainable source code without externaldependencies), this compromise certainly seems acceptable.The library has been successfully integrated into version 3.1of the HEALPix C++ and Fortran packages. There also existsan experimental version of the SSHT3 package with libsharpreplacing the library’s original SHT engine. Libsharp is alsoused as SHT engine in an upcoming version of the Python pack-age NIFTy4 for signal inference (Selig et al. 2013). Recently,the total convolution code conviqt (Pr´ezeau & Reinecke 2010),which is a central component of the Planck simulation pipeline(Reinecke et al. 2006), has been updated and is now basedon libsharp SHTs. There are plans for a similar update ofthe artDeco deconvolution map maker (Keih¨anen & Reinecke2012).A potential future field of work is porting libsharp toIntel’s “many integrated cores” architecture5, once sufficientcompiler support for this platform has been established. The3 http://www.mrao.cam.ac.uk/˜jdm57/ssht/index.html4 http://www.mpa-garching.mpg.de/ift/nifty/5 http://en.wikipedia.org/wiki/Intel_MIC8M. Reinecke and D. S. Seljebotn: Libsharp – spherical harmonic transforms revisitedTable 1. Performance comparison with other implementations at lmax = 2047, ncore = 1.CodeversiongridspinnSHTRAVXRSSE2Rscalarshtns2.31Gauss-Legendre010.840.880.91Wavemoth (brute-force)Nov 2011HEALPix (Nside=1024)011.630.98—”””051.591.09—Wavemoth (butterfly)Nov 2011HEALPix (Nside=1024)050.960.66—libpshtJan 2011Gauss-Legendre014.062.302.30”””052.661.751.62”””212.501.481.20”””252.151.441.08spinsfastr104ECP (Clenshaw-Curtis)0157.0432.1215.31”””0528.3918.729.38”””2116.9910.204.73”””258.605.662.56SSHT1.0b1MW sampling theorem0120.9115.609.46”””2113.409.294.99S2HAT2.55betaHEALPix (Nside=1024)0112.337.333.60Glesp2Gauss-Legendre0155.3231.2614.95Notes. All tests had a band limit of lmax = 2047 and were carried out on a single core. The grids used by libsharp and the respective comparisoncode were identical in each run. RAVX denotes the quotient of wall-clock times for the respective code and libsharp in the presence of the AVXinstruction set, RSSE2 is the quotient when SSE2 (but not AVX) is supported, and Rscalar was measured with both SSE2 and AVX disabled. Thelibsharp support for the MW sampling theorem used for the SSHT comparisons is experimental. For Wavemoth, butterfly matrix compressioncan optionally be enabled. In the benchmark given we requested an accuracy of 10−4, which led to an extra requirement of 4 GB of precomputeddata in memory. Note that when running on a single core, Wavemoth is at an advantage compared to the normal situation where the memory busis shared between multiple cores.hardware appears to be very well suited for running SHTs, andthe porting by itself would provide a welcome test for the adapt-ability of the library’s code design.Acknowledgements. We thank our referee Nathana¨el Schaeffer for his construc-tive remarks and especially for pointing out a missed optimisation opportu-nity in our shtns installation, which had a significant effect on some bench-mark results. MR is supported by the German Aeronautics Center and SpaceAgency (DLR), under program 50-OP-0901, funded by the Federal Ministryof Economics and Technology. DSS is supported by the European ResearchCouncil, grant StG2010-257080. The presented benchmarks were performed asproject pr89yi at the Leibniz Computing Center Garching.Referencesde Oliveira-Costa, A., Tegmark, M., Zaldarriaga, M., & Hamilton, A. 2004,Phys. Rev. D, 69, 063516, arXiv:astro-ph/0307282Doroshkevich, A. G., Naselsky, P. D., Verkhodanov, O. V., et al. 2005,International Journal of Modern Physics D, 14, 275, arXiv:astro-ph/0305537Eriksen, H. K., Jewell, J. B., Dickinson, C., et al. 2008, ApJ, 676, 10,arXiv:0709.1058Fej´er, L. 1933, Mathematische Zeitschrift, 37, 287Gautschi, W. 1967, SIAM Journal on Numerical Analysis, 4, 357Goldberg, J. N., Macfarlane, A. J., Newman, E. T., Rohrlich, F., & Sudarshan,E. C. G. 1967, Journal of Mathematical Physics, 8, 2155Huffenberger, K. M. & Wandelt, B. D. 2010, ApJS, 189, 255, arXiv:1007.3514Keih¨anen, E. & Reinecke, M. 2012, A&A, 548, A110, arXiv:1208.1399Kostelec, P. & Rockmore, D. 2008, Journal of Fourier Analysis and Applications,14, 145, 10.1007/s00041-008-9013-5Lewis, A. 2005, Phys. Rev. D, 71, 083008, arXiv:astro-ph/0502469McEwen, J. D. & Wiaux, Y. 2011, IEEE Transactions on Signal Processing, 59,5876, arXiv:1110.6298Pr´ezeau, G. & Reinecke, M. 2010, ApJS, 190, 267, arXiv:1002.1050Reinecke, M. 2011, A&A, 526, A108, arXiv:1010.2084Reinecke, M., Dolag, K., Hell, R., Bartelmann, M., & Enßlin, T. A. 2006, A&A,445, 373, arXiv:astro-ph/0508522Schaeffer, N. 2013, Geochem. Geophys. Geosyst., 14, arXiv:1202.6522Selig, M., Bell,M. R., Junklewitz,H., et al. 2013, ArXiv e-prints,arXiv:1301.4499Seljebotn, D. S. 2012, ApJS, 199, 5, arXiv:1110.4874Szydlarski, M., Esterie, P., Falcou, J., Grigori, L., & Stompor, R. 2013,Concurrency and Computation: Practice and Experience, arXiv:1106.0159Waldvogel, J. 2006, BIT Numerical Mathematics, 46, 195Wandelt, B. D. & G´orski, K. M. 2001, Phys. Rev. D, 63, 123002,arXiv:astro-ph/00082279View publication stats | RESEARCH GATE |
Microorganisms 2021, 9, 430. https://doi.org/10.3390/microorganisms9020430 www.mdpi.com/journal/microorganisms Communication Deep Ultraviolet Light‐Emitting Diode Light Therapy for Fusobacterium nucleatum Soichiro Fukuda 1,†, Shunsuke Ito 2,†, Jun Nishikawa 1,*, Tatsuya Takagi 1, Naoto Kubota 1, Ken‐ichiro Otsuyama 1, Hidehiro Tsuneoka 1, Junzo Nojima 1, Koji Harada 3, Katsuaki Mishima 3, Yutaka Suehiro 4, Takahiro Yamasaki 4 and Isao Sakaida 2 1 Department of Laboratory Science, Graduate School of Medicine, Yamaguchi University, Ube 7558505, Japan; a006upu@yamaguchi‐u.ac.jp (S.F.); tkgcrzccg@docomo.ne.jp (T.T.); a003upu@yamaguchi‐u.ac.jp (N.K.); otsuyama@yamaguchi‐u.ac.jp (K.‐i.O.); htsune@yamaguchi‐u.ac.jp (H.T.); nojima‐j@yamaguchi‐u.ac.jp (J.N.) 2 Department of Gastroenterology and Hepatology, Graduate School of Medicine, Yamaguchi University, Ube 7558505, Japan; g029ub@yamaguchi‐u.ac.jp (S.I.); sakaida@yamaguchi‐u.ac.jp (I.S.) ³ Department of Oral and Maxillofacial Surgery, Graduate School of Medicine, Yamaguchi University, Ube 7558505, Japan; harako@yamaguchi‐u.ac.jp (K.H.); kmishima@yamaguchi‐u.ac.jp (K.M.) 4 Department of Oncology and Laboratory Medicine, Graduate School of Medicine, Yamaguchi University, Ube 7558505, Japan; ysuehiro@yamaguchi‐u.ac.jp (Y.S.); t.yama@yamaguchi‐u.ac.jp (T.Y.) * Correspondence: junnis@yamaguchi‐u.ac.jp; Tel.: +81‐836‐22‐2835 † These authors equally contributed to this work. Abstract: Background: Fusobacterium nucleatum, which is associated with periodontitis and gingivi‐tis, has been detected in colorectal cancer (CRC). Methods: We evaluated the bactericidal effect of deep ultraviolet (DUV) light‐emitting diode (LED) light therapy on F. nucleatum both qualitatively and quantitatively. Two DUV‐LEDs with peak wavelengths of 265 and 280‐nm were used. DNA damage to F. nucleatum was evaluated by the production of cyclobutane pyrimidine dimers (CPD) and pyrimidine (6–4) pyrimidone photoproducts (6–4PP). Results: DUV‐LEDs showed a bacteri‐cidal effect on F. nucleatum. No colony growth was observed after 3 min of either 265 nm or 280 nm DUV‐LED irradiation. The survival rates of F. nucleatum under 265 nm DUV‐LED light irradiation dropped to 0.0014% for 10 s and to 0% for 20 s irradiation. Similarly, the survival rate of F. nucleatum under 280 nm DUV‐LED light irradiation dropped to 0.00044% for 10 s and 0% for 20 s irradiation. The irradiance at the distance of 35 mm from the DUV‐LED was 0.265 mW/cm2 for the 265 nm LED and 0.415 mW/cm2 for the 280 nm LED. Thus, the radiant energy for lethality was 5.3 mJ/cm2 for the 265 nm LED and 8.3 mJ/cm2 for the 280 nm LED. Amounts of CPD and 6–4PP in F. nucleatum irra‐diated with 265 nm DUV‐LED light were 6.548 ng/μg and 1.333 ng/μg, respectively. Conclusions: DUV‐LED light exerted a bactericidal effect on F. nucleatum by causing the formation of pyrimidine dimers indicative of DNA damage. Thus, DUV‐LED light therapy may have the potential to prevent CRC. Keywords: colorectal cancer; deep ultraviolet; DNA damage; Fusobacterium nucleatum; light‐emit‐ting diode 1. Introduction The development of gastrointestinal cancer involves lifestyle factors such as alcohol use, smoking [1], obesity [2], and eating habits [3] as well as certain infectious diseases [4]. The International Agency for Research Cancer (IARC) identified Helicobacter pylori as a definite oncogenic factor for gastric cancer from epidemiological studies in 1994 [5], and eradication therapy for H. pylori has been shown to suppress the development of meta‐chronous gastric cancers [6,7]. Unlike gastric cancer, the mortality rate from colorectal cancer (CRC) has increased significantly in Japan, and thus effective methods to prevent CRC are required. Citation: Fukuda, S.; Ito, S.; Nishikawa, J.; Takagi, T.; Kubota, N.; Otsuyama, K.‐i.; Tsuneoka, H.; Nojima, J.; Harada, K.; Mishima, K.; et al. Deep Ultraviolet Light‐Emitting Diode Light Therapy for Fusobacterium nucleatum. Microorganisms 2021, 9, 430. https://doi.org/10.3390/ microorganisms9020430 Academic Editor: Curtis A. Machida Received: 23 January 2021 Accepted: 15 February 2021 Published: 19 February 2021 Publisher’s Note: MDPI stays neu‐tral with regard to jurisdictional claims in published maps and insti‐tutional affiliations. Copyright: © 2021 by the authors. Li‐censee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and con‐ditions of the Creative Commons At‐tribution (CC BY) license (http://crea‐tivecommons.org/licenses/by/4.0/). Microorganisms 2021, 9, 430 2 of 10 Gut dysbiosis has been associated with the development of CRC [8]. Recent studies have identified Fusobacterium nucleatum (F. nucleatum), Streptococcus bovis, enterotoxigenic Bacteroides fragilis, Enterococcus faecalis, Escherichia coli, and Peptostreptococcus anaerobius as CRC candidate pathogens [9]. F. nucleatum has been reported to be detected in CRC [10]. Fusobacterium is an anaerobic Gram‐negative bacillus present in the oral cavity and diges‐tive tract of healthy individuals [11], and F. nucleatum is associated with oral inflammatory diseases such as periodontitis and gingivitis [12–14]. In 2012, Kostic et al. characterized the composition of the microbiota in colorectal carcinoma using whole genome sequenc‐ing and showed that Fusobacterium sequences were enriched in carcinomas [15]. Castel‐larin et al. also found that only F. nucleatum was significantly increased in CRC tumors relative to control specimens by using RNA sequencing [16]. Mima et al. reported that the amount of F. nucleatum DNA in CRC tissue is associated with shorter survival [17], and Yamaoka et al. found that it may potentially serve as a prognostic biomarker [18]. Thus, accumulating evidence suggests that F. nucleatum is associated with the development and progression of CRC in humans. Several studies have revealed a mechanism by which F. nucleatum contributes to CRC development. Rubinstein et al. showed that FadA adhesin on F. nucleatum binds to E‐cad‐herin and activates β‐catenin signaling in CRC cells, and then, CRC growth is induced by transcription of c‐Myc and cyclin‐D1 [19]. Yang et al. showed that CRC cell lines infected with F. nucleatum formed larger tumors more rapidly in nude mice than did uninfected cells. Infection of cells with F. nucleatum increased the expression of miR21 by activating TLR4 signaling to MYD88, leading to the activation of nuclear factor NFκB [20]. Yu et al. found that F. nucleatum was abundant in CRC tissues of patients with recurrence after chemotherapy. F. nucleatum targeted TLR4 and MYD88 innate immune signaling and spe‐cific microRNAs to activate the autophagy pathway and alter CRC chemotherapeutic re‐sponse [21]. Gur et al. found that the Fap2 protein of F. nucleatum directly interacted with T‐cell immunoglobulin and the ITIM domain (TIGIT), leading to the inhibition of NK cell cytotoxicity [22]. F. nucleatum adheres to and invades CRC cells and then induces onco‐genic and inflammatory responses to stimulate their growth. We have considered deep ultraviolet (DUV) light‐emitting diode (LED) light therapy for the treatment of F. nucleatum. Because F. nucleatum forms a biofilm in the periodontal pocket, injection of an antibacterial agent does not provide a sufficient bactericidal effect, and infection often recurs [23]. LEDs are semiconductor devices that emit light and are ideal for downsizing equipment [24]. Recently, LEDs that can emit ultraviolet light have been developed [25], and we have confirmed that DUV‐LEDs have various bactericidal effects against bacteria and fungi in vitro [26]. In the present study, the bactericidal effect of DUV‐LEDs on F. nucleatum was examined. 2. Materials and Methods 2.1. Fusobacterium nucleatum Strain and Growth Conditions F. nucleatum (ATCC 25586) and Escherichia coli (NBRC3972) were purchased from NITE Biological Resource Center (NBRC). The frozen bacterial stock of F. nucleatum was cultivated in Gifu Anaerobic Medium (GAM) agar in an anaeropack at 37 °C for 72 h. E. coli were cultured on heart infusion agar (Eiken Kagaku Co., Ltd.,Tokyo, Japan) at 37 °C for 16 h. Then, cell suspensions were prepared using phosphate‐buffered saline (PBS). 2.2. Irradiation by DUV‐LEDs Two DUV‐LEDs with peak wavelengths of 265 and 280‐nm (VPS131 (265‐nm LED) and VPS161 (280‐nm LED), NIKKISO CO., LTD., Tokyo, Japan) were used (Figure 1). Total radiant flux of the DUV‐LEDs was 9.4 mW for the 265 nm and 17.0 mW for the 280 nm wavelength. As the half‐power angle of the LEDs was 130 degrees, we set the distance between the LED and the plate at 35 mm, which enabled uniform DUV‐LED illumination (Figure 2). The rated current for the 265 nm LEDs set by the company is 350 mA. We Microorganisms 2021, 9, 430 3 of 10 used a direct current‐regulated power supply in constant current mode (PW36‐1.5AD, TEXIO TECHNOLOGY CORPORATION, Yokohama, Japan). The current was set to 350 mA or 50 mA to irradiate LED light. Irradiance was measured with an MCPD‐9800 array spectrometer (Otsuka Electronics, Osaka, Japan). Figure 1. Spectral radiant flux of the DUV‐LEDs. Figure 2. The apparatus for DUV‐LED irradiation. The size of the DUV‐LEDs was 3.5 mm × 3.5 mm. The size of the substrate of the LEDs was 35 mm × 35 mm × 5 mm. The length of the fins of the heat sink was 32 mm. 2.3. Qualitative Test of Bactericidal Effect on Agar Medium Surface by DUV‐LED Light The suspension of F. nucleatum was adjusted to 1 McFarland turbidity standard with PBS and then spread on GAM agar plates (diameter, 35 mm) with a cotton swab. We con‐ducted time series tests with irradiation times of 0, 10, 20, 30, 60, and 180 s. Colonies of F. nucleatum were observed after anaerobic cultivation for 6 days. As the non‐irradiated con‐trol, F. nucleatum was left in the air for the same time periods as the DUV‐LED irradiation. 2.4. Quantitative Test of Bactericidal Effect of DUV‐LED Light Bactericidal effects were determined using a colony‐forming assay. The cell suspen‐sion of F. nucleatum was adjusted to 1.0 × 106 colony forming units (CFU)/mL. One millili‐ter of cell suspension was dispensed onto a 35 mm dish and irradiated by DUV‐LED light. We conducted time series tests with irradiation times of 0, 10, 20, and 30 s. Then, 10 μL of 10‐fold serial dilutions of the cell suspension was plated onto GAM agar in quadruplicate, and the plates were incubated at 37 °C for 6 days. Thereafter, the number of colonies was counted, and survival rate was expressed as a percentage of the non‐irradiated control. The experiment was performed three times, and the average percentage of surviving F. nucleatum was evaluated [27,28]. Microorganisms 2021, 9, 430 4 of 10 2.5. Cytotoxicity of Human Keratinocytes by DUV‐LED HaCaT cells, a spontaneously immortalized human keratinocyte line, were cultured in Dulbecco’s Modified Eagle Medium—high glucose (4.5 g/L) with L‐glutamine and with sodium pyruvate (Capricorn) supplemented with 10% fetal bovine serum. HaCaT cells (4.0 × 105/dish) were seeded in 35 mm dishes and cultured at 37 °C in 5% CO2 for 24 h. The cells were then washed twice with PBS, after which we immediately irradiated the cells with DUV‐LED light. Two milliliters of fresh medium was added, and the mixture was cultured for 24 h. Then, the cells were harvested, and cell viability was evaluated by trypan blue dye exclusion assay. The number of viable cells was counted, and the survival rate was expressed as a percentage of the non‐irradiated control. The experiment was per‐formed three times, and the average percentage of surviving HaCaT cells was evaluated. 2.6. CPD and 6–4PP Quantification by ELISA The DNA damage in F. nucleatum was evaluated based on the production of cyclo‐butane pyrimidine dimers (CPD) and pyrimidine (6–4) pyrimidone photoproducts (6–4PP) with an OxiSelect™ UV‐Induced DNA Damage ELISA Kit (CELL BIOLABS Inc., San Diego, CA, USA). The bacterial suspension was adjusted to a 2 McFarland turbidity stand‐ard. Then, 2 mL of the suspension was dispensed onto a 35 mm dish, followed by DUV‐LED light irradiation for 30 s. DNA was extracted from 1.8 mL of the bacterial solutions with a QIAamp DNA FFPE Tissue Kit (QIAGEN, Venlo, The Netherlands). The ELISA procedure was performed according to the manufacturer’s protocol. We used DNA sam‐ples from E. coli irradiated by DUV from a UV lamp as a positive control and DNA sam‐ples from F. nucleatum cultured without DUV irradiation as a negative control. The UV lamp (GL‐15, Toshiba, Japan) was installed in a safety cabinet (Panasonic Healthcare, To‐kyo, Japan). 2.7. Statistics Differences in survival rates between the 265 nm DUV‐LED and 280 nm DUV‐LED were analyzed by t‐test using StatFlex Ver.6.0 (Artech Co., Ltd., Osaka, Japan). 3. Results 3.1. Qualitative Test of Bactericidal Effect on Agar Medium Surface by DUV‐LED Light DUV‐LEDs with peak wavelengths of 265 and 280‐nm showed similar bactericidal effect on F. nucleatum cultured on an agar medium surface in a time‐dependent manner. No colony growth was observed after 180 s of either 265 nm or 280 nm DUV‐LED irradi‐ation (Figure 3). In this experiment, even if the sterilized dish was returned to the culture, no bacterial growth was observed. We confirmed that an increase in the ambient temper‐ature was not observed when we supplied a current of 350 mA to the DUV‐LEDs. Microorganisms 2021, 9, 430 5 of 10 Figure 3. Qualitative test of the bactericidal effect of DUV‐LED light irradiation on Fusobacterium nucleatum. (a) The growth of colonies of F. nucleatum was observed after irradiation with 265 nm LED and 280 nm LED at each time point. A non‐irradiated control was used. (b) An enlarged im‐age of the agar plates. 3.2. Quantitative Test of Bactericidal Effect by DUV‐LED Light When we supplied a current of 350 mA to the DUV‐LED, the survival rate of F. nu‐cleatum under 265 nm DUV‐LED light irradiation dropped to 0.0014% for 10 s and to 0% (below limit of detection) for 20 s irradiation. Similarly, the survival rate of F. nucleatum under 280 nm DUV‐LED light irradiation dropped to 0.00044% for 10 s and 0% (below limit of detection) for 20 s irradiation (Figure 4). When 50 mA was supplied to the DUV‐LED, the survival rates of F. nucleatum under 265 nm DUV‐LED light irradiation were 45.34% for 10 s and 16.13% for 20 s irradiation. Whereas those of F. nucleatum under 280 nm DUV‐LED light irradiation were 54.72% for 10 s and 21.34% for 20 s irradiation (Figure 4). There was no significant difference in the survival rates of F. nucleatum between the 265 nm DUV‐LED and 280 nm DUV‐LED irradiations. To confirm the presence of any surviving bacteria, 0.5 mL of these residual suspensions was cultured in liquid medium, but no F. nucleatum growth was observed. When we supplied a current of 350 mA to the DUV‐LED, the irradiance at the dis‐tance of 35 mm from the DUV‐LED was 0.265 mW/cm2 for the 265 nm LED and 0.415 mW/cm2 for the 280 nm LED. The irradiation time under 265 nm and 280 nm DUV‐LED light for which the survival rates were 0% was 20 s. Thus, the radiant energy for lethality was 5.3 mJ/cm2 for the 265 nm LED and 8.3 mJ/cm2 for the 280 nm LED. Figure 4. Quantitative test of the bactericidal effect of DUV‐LED light irradiation on Fusobacterium nucleatum. The survival rates of F. nucleatum after irradiation with 265 nm LED (black bars) and 280 nm LED (grey bars) are shown for the application of 350 mA or 50 mA of current. Error bars indicate standard error. Microorganisms 2021, 9, 430 6 of 10 3.3. Evaluation of HaCaT Cell Damage by DUV‐LED Light When we supplied a current of 350 mA to the DUV‐LED, the survival rates of HaCaT cells under 265 nm DUV‐LED light irradiation were 4.38% for 10 s and 2.08% for 20 s irradiation. Similarly, the survival rates of HaCaT cells under 280 nm DUV‐LED light ir‐radiation dropped to 3.31% for 10 s and 2.26% for 20 s irradiation (Figure 5). When 50 mA was supplied to the DUV‐LED, the survival rates of HaCaT cells under 265 nm DUV‐LED light irradiation were 90.14% for 10 s and 48.29% for 20 s irradiation, whereas those of HaCaT cells under 280 nm DUV‐LED light irradiation were 90.06% for 10 s and 47.76% for 20 s irradiation (Figure 5). There was no significant difference in the survival rates of HaCaT cells between the 265 nm and 280 nm DUV‐LED irradiations. Figure 5. Evaluation of HaCaT cell damage by DUV‐LED light. The survival rates of HaCaT cells after irradiation with 265 nm LED (black bars) and 280 nm LED (grey bars) are shown for the ap‐plication of 350 mA or 50 mA of current. Error bars indicate standard error. 3.4. Detection of DUV‐LED Light‐Induced DNA Damage by Formation of CPD and 6–4PP CPD and 6–4PP were detected by ELISA in the DNA samples from F. nucleatum irra‐diated with DUV‐LED light. The respective amounts of CPD and 6–4PP in F. nucleatum irradiated with 265 nm DUV‐LED were 6.548 ng/μg and 1.333 ng/μg, whereas those in F. nucleatum irradiated with 280 nm DUV‐LED were 7.963 ng/μg and 1.593 ng/μg. As a con‐trol, we prepared E. coli irradiated with the UV lamp. The respective amounts of CPD and 6–4PP in the E. coli were 2.622 ng/μg and 0.276 ng/μg (Figure 6). Microorganisms 2021, 9, 430 7 of 10 Figure 6. Detection of DUV‐LED light‐induced DNA damage by quantification of CPD and 6–4PP with ELISA. DNA samples extracted from F. nucleatum after 30 s irradiation with 265 nm LED light (black bars) or 280 nm LED light (grey bars) were measured for amounts of CPD and 6–4PP per microgram of DNA. E. coli irradiated with a UV lamp (hatched bar) were used as a positive control. 4. Discussion Light irradiation from DUV‐LEDs exerted a bactericidal effect on F. nucleatum in a time‐dependent manner. We previously reported that the Gram‐negative bacteria Pseudo‐monas aeruginosa and E. coli were highly sensitive to DUV‐LED light [28]. F. nucleatum is an anaerobic Gram‐negative bacterium with a similar sensitivity to DUV‐LED as that of other Gram‐negative bacteria because their cell wall is thinner than that of Gram‐positive bacteria [29]. That we did not conduct experiments to form biofilms in vitro is a limitation of the present study. DUV light at a wavelength of around 260 nm is absorbed by DNA [30]. When cells are irradiated with UV light, covalent bonds of the pyrimidine base form with the adjacent thymine or cytosine base that result in the creation of pyrimidine dimers. These dimers interfere with DNA replication and transcription that can lead to cell death, mutation, and chromosomal instability [31]. The mechanism of the bactericidal effect exerted by DUV‐LEDs is thought to be due to DNA damage because DNA extracted from DUV‐LED‐irra‐diated F. nucleatum contains pyrimidine dimers. While the light wavelength around 280 nm is absorbed by proteins, the effect on proteins of cell membranes is one of the issues being pursued. Although fusobacteria are found in CRC [10,15–18], the mechanisms by which they hone in on and localize to CRC have been underexplored. Abed et al. suggested that fuso‐bacterial Fap2 and host Gal‐GalNAc are involved in fusobacterial CRC localization and enrichment. They showed that intravenously injected F. nucleatum localizes to mouse tu‐mor tissues in a Fap2‐dependent manner, suggesting that fusobacteria use a hematog‐enous route to reach colon adenocarcinomas [32]. Komiya et al. reported that 75% (6/8) of CRC patients exhibited identical strains of F. nucleatum in their CRC and saliva specimens [33]. In addition, transient bacteremia is common during periodontal disease with bacte‐rial loads reaching 104 bacteria/mL blood 15 min after tooth brushing in humans [34]. These findings suggest that F. nucleatum in CRC originates in the oral cavity. We think that sterilizing F. nucleatum in the oral cavity by DUV‐LED light may reduce the number of F. nucleatum in the intestine and could potentially lead to the prevention of CRC. In addition, we previously showed that stage IV CRC patients with a high F. nucleatum copy Microorganisms 2021, 9, 430 8 of 10 number had a significantly shorter overall survival time than those with a low copy num‐ber [18], indicating that DUV‐LED therapy may also improve survival in patients with advanced CRC. DNA damage by DUV is elicited not only in microorganisms but also in human cells, and chronic exposure to DUV light has been established as a human health hazard [35]. In a mouse model, significant hyperplasia and intercellular edema were induced in the epidermis after chronic irradiation of 254 nm DUV at 4500 mJ/cm2 [36]. The radiant energy required for lethality of F. nucleatum in the present study was much less, 5.3 mJ/cm2 for the 265 nm LED and 8.3 mJ/cm2 for the 280 nm LED. By reducing the current supplied to the LEDs, the survival rate of HaCaT cells increased. We must first determine appropriate doses of DUV that can selectively inactivate F. nucleatum in an animal model. DUV‐LED light therapy is not a specific treatment targeting one pathogen such as vaccines do, so it may disturb the oral bacterial flora. Treatments that affect all oral bacte‐ria can lead to nitric oxide pathways and adversely affect blood pressure regulation [37]. Thus, we must be aware of these adverse events. In the case of general oral care, however, the number of bacteria is reduced by 60 to 70%, but this is effective in preventing aspira‐tion pneumonia [38]. Consequently, it is considered clinically useful to reduce multiple types of bacteria as a whole rather than suppressing just one type of bacteria. We believe that this also needs to be verified by animal experiments. LEDs have unique characteristics of compactness, durability, and low heat produc‐tion, and thus, DUV‐LEDs can be applied in narrow spaces such the oral cavity. A mouth‐piece‐type phototherapy device should be considered as an effective application to irra‐diate DUV‐LED light to a localized infection. It might be necessary to use a thin fiber to irradiate DUV into the periodontal pockets. 5. Conclusions Irradiation with DUV‐LED light exerted a bactericidal effect on F. nucleatum by caus‐ing the formation of pyrimidine dimers indicative of DNA damage. DUV‐LED devices can be useful tools for the inactivation of F. nucleatum. Thus, DUV‐LED light therapy may have the potential to prevent CRC. Author Contributions: J.N. (Jun Nishikawa) and T.Y. conceived and designed the study. S.F., S.I., T.T., and N.K. acquired the data. K.‐i.O., H.T., J.N. (Junzo Nojima), and K.H. analyzed the data. K.M., Y.S., and I.S. revised the manuscript. All authors have read and agreed to the published ver‐sion of the manuscript. Funding: This research was supported by AMED under Grant Number JP19lm0203008. Data Availability Statement: All data generated or analyzed during the study are included in this published article. Conflicts of Interest: The authors declare that they have no competing interests. References 1. Steevens, J.; Schouten, L.J.; Goldbohm, R.A.; van den Brandt, P.A. Alcohol consumption, cigarette smoking and risk of subtypes of oesophageal and gastric cancer: A prospective cohort study. Gut 2010, 59, 39–48, doi:10.1136/gut.2009.191080. 2. Aune, D.; Greenwood, D.C.; Chan, D.S.; Vieira, R.; Vieira, A.R.; Navarro Rosenblatt, D.A.; Cade, J.E.; Burley, V.J.; Norat, T. Body mass index, abdominal fatness and pancreatic cancer risk: A systematic review and non‐linear dose‐response meta‐analysis of prospective studies. Ann. Oncol. 2012, 23, 843–852, doi:10.1093/annonc/mdr398. 3. Magalhães, B.; Peleteiro, B.; Lunet, N. Dietary patterns and colorectal cancer: Systematic review and meta‐analysis. Eur, J. Cancer Prev. 2012, 21, 15–23, doi:10.1097/CEJ.0b013e3283472241. 4. Selgrad, M.; Malfertheiner, P.; Fini, L.; Goel, A.; Boland, C.R.; Ricciardiello, L. The role of viral and bacterial pathogens in gas‐trointestinal cancer. J. Cell Physiol. 2008, 216, 378–388, doi:10.1002/jcp.21427. 5. Infection with Helicobacter pylori. IARC Monger Eval. Carcinog Risks Hum. 1994, 61, 177–240. 6. Uemura, N.; Mukai, T.; Okamoto, S.; Yamaguchi, S.; Mashiba, H.; Taniyama, K.; Sasaki, N.; Haruma, K.; Sumii, K.; Kajiyama, G. Effect of Helicobacter pylori eradication on subsequent development of cancer after endoscopic resection of early gastric can‐cer. Cancer Epidemiol Biomarkers Prev. 1997, 6, 639–642. Microorganisms 2021, 9, 430 9 of 10 7. Fukase, K.; Kato, M.; Kikuchi, S.; Inoue, K.; Uemura, N.; Okamoto, S.; Terao, S.; Amagai, K.; Hayashi, S.; Asaka, M. Japan Gast Study Group. Effect of eradication of Helicobacter pylori on incidence of metachronous gastric carcinoma after endoscopic resec‐tion of early gastric cancer: An open‐label, randomised controlled trial. Lancet 2008, 372, 392–397, doi:10.1016/S0140‐673661159‐9. 8. Chattopadhyay, I.; Dhar, R.; Pethusamy, K.; Seethy, A.; Srivastava, T.; Sah, R.; Sharma, J.; Karmakar, S. Exploring the Role of Gut Microbiome in Colon Cancer. Appl. Biochem. Biotechnol. 2021, 25, doi:10.1007/s12010‐021‐03498‐9. 9. Cheng, Y.; Ling, Z.; Li, L. The Intestinal Microbiota and Colorectal Cancer. Front. Immunol. 2020, 11, 615056, doi:10.3389/fimmu.2020.615056. 10. Liu, Y.; Baba, Y.; Ishimoto, T.; Iwatsuki, M.; Hiyoshi, Y.; Miyamoto, Y.; Yoshida, N.; Wu, R.; Baba, H. Progress in characterizing the linkage between Fusobacterium nucleatum and gastrointestinal cancer. J. Gastroenterol. 2019, 54, 33–41, doi:10.1007/s00535‐018–1512–9. 11. Signat, B.; Roques, C.; Poulet, P.; Duffaut, D. Fusobacterium nucleatum in periodontal health and disease. Curr. Issues Mol. Biol. 2011, 13, 25–36. 12. Yang, N.Y.; Zhang, Q.; Li, J.L.; Yang, S.H.; Shi, Q. Progression of periodontal inflammation in adolescents is associated with increased number of Porphyromonas gingivalis, Prevotella intermedia, Tannerella forsythensis, and Fusobacterium nucleatum. Int. J. Paediatr Dent. 2014, 24, 226–233, doi:10.1111/ipd.12065. 13. Kistler, J.O.; Booth, V.; Bradshaw, D.J.; Wade, W.G. Bacterial community development in experimental gingivitis. PLoS ONE 2013, 8, e71227, doi:10.1371/journal.pone.0071227. 14. Fujii, R.; Saito, Y.; Tokura, Y.; Nakagawa, K.I.; Okuda, K.; Ishihara, K. Characterization of bacterial flora in persistent apical periodontitis lesions. Oral Microbiol. Immunol. 2009, 24, 502–505, doi:10.1111/j.1399‐302X.2009.00534.x. 15. Kostic, A.D.; Gevers, D.; Pedamallu, C.S.; Michaud, M.; Duke, F.; Earl, A.M.; Ojesina, A.I.; Jung, J.; Bass, A.J.; Tabernero, J.; et al. Genomic analysis identifies association of Fusobacterium with colorectal carcinoma. Genome Res. 2012, 22, 292–298, doi:10.1101/gr.126573.111. 16. Castellarin, M.; Warren, R.L.; Freeman, J.D.; Dreolini, L.; Krzywinski, M.; Strauss, J.; Barnes, R.; Watson, P.; Allen‐Vercoe, E.; Moore, R.A.; et al. Fusobacterium nucleatum infection is prevalent in human colorectal carcinoma. Genome Res. 2012, 22, 299–306, doi:10.1101/gr.126516.111. 17. Mima, K.; Nishihara, R.; Qian, Z.R.; Cao, Y.; Sukawa, Y.; Nowak, J.A.; Yang, J.; Dou, R.; Masugi, Y.; Song, M.; et al. Fusobacterium nucleatum in colorectal carcinoma tissue and patient prognosis. Gut 2016, 65, 1973–1980, doi:10.1136/gutjnl‐2015‐310101. 18. Yamaoka, Y.; Suehiro, Y.; Hashimoto, S.; Hoshida, T.; Fujimoto, M.; Watanabe, M.; Imanaga, D.; Sakai, K.; Matsumoto, T.; Nish‐ioka, M.; et al. Fusobacterium nucleatum as a prognostic marker of colorectal cancer in a Japanese population. J. Gastroenterol. 2018, 53, 517–524, doi:10.1007/s00535‐017‐1382‐6. 19. Rubinstein, M.R.; Wang, X.; Liu, W.; Hao, Y.; Cai, G.; Han, Y.W. Fusobacterium nucleatum promotes colorectal carcinogenesis by modulating E‐cadherin/beta‐catenin signaling via its FadA adhesin. Cell Host Microbe 2013, 14, 195–206, doi:10.1016/j.chom.2013.07.012. 20. Yang, Y.; Weng, W.; Peng, J.; Hong, L.; Yang, L.; Toiyama, Y.; Gao, R.; Liu, M.; Yin, M.; Pan, C.; et al. Fusobacterium nucleatum increases proliferation of colorectal cancer cells and tumor development in mice by activating Toll‐like receptor 4 signaling to nuclear factor‐κB, and up‐regulating expression of MicroRNA‐21. Gastroenterology 2017, 152, 851–866, doi:10.1053/j.gas‐tro.2016.11.018. 21. Yu, T.; Guo, F.; Yu, Y.; Sun, T.; Ma, D.; Han, J.; Qian, Y.; Kryczek, I.; Sun, D.; Nagarsheth, N.; et al. Fusobacterium nucleatum promotes chemoresistance to colorectal cancer by modulating autophagy. Cell 2017, 170, 548–563, doi:10.1016/j.cell.2017.07.008. 22. Gur, C.; Ibrahim, Y.; Isaacson, B.; Yamin, R.; Abed, J.; Gamliel, M.; Enk, J.; Bar‐On, Y.; Stanietsky‐Kaynan, N.; Coppenhagen‐Glazer, S.; et al. Binding of the Fap2 protein of Fusobacterium nucleatum to human inhibitory receptor TIGIT protects tumors from immune cell attack immunity. Immunity 2015, 42, 344–355, doi:10.1016/j.immuni.2015.01.010. 23. Kanwar, I.; Sah, A.K.; Suresh, P.K. Biofilm‐mediated antibiotic‐resistant oral bacterial infections: Mechanism and combat strat‐egies. Curr. Pharm Des. 2017, 23, 2084–2095, doi:10.2174/1381612822666161124154549. 24. Wu, C.Y.; Wu, T.F.; Tsai, J.R.; Chen, Y.M.; Chen, C.C. Multistring LED backlight driving system for LCD panels with color sequential display and area control. IEEE Trans. Ind. Electron. 2010, 55, 3791–3800. 25. Hirayama, H.; Noguchi, N.; Yatabe, T.; Kamata, N. 227 nm AlGaN light‐emitting diode with 0.15 mW output power realized using a thin quantum well and AlN buffer with reduced threading dislocation density. Appl. Phys. Express 2008, 1, 051101‐3. 26. Takagi, T.; Nishikawa, J.; Yanagihara, M.; Fukuda, S.; Kubota, N.; Kobayashi, Y.; Otsuyama, K.I.; Nojima, J.; Tsuneoka, H.; Sakai, K.; et al. Microbicidal effect of deep ultraviolet light‐emitting diode irradiation. Lasers Med. Sci. 2020, 9, doi:10.1007/s10103‐020‐03143‐7. 27. Thomas, P.; Sekhar, A.C.; Mujawar, M.M. Nonrecovery of varying proportions of viable bacteria during spread plating gov‐erned by the extent of spreader usage and proposal for an alternate spotting‐spreading approach to maximize the CFU. J. Appl Microbiol. 2012, 113, 339–350, doi:10.1111/j.1365‐2672.2012.05327.x. 28. Okamoto, T.; Nishikawa, J.; Yanai, H.; Nakamura, H.; Takeuchi, H.; Kurai, S.; Akada, J.K.; Sakaida, I. In vitro bactericidal effects of near‐ultraviolet light from light‐emitting diodes on Helicobacter pylori. Scan. J. Gastroenterol. 2013, 48, 1484–1486, doi:10.3109/00365521.2013.837953. 29. Kim, D.K.; Kim, S.J.; Kang, D.H. Bactericidal effect of 266 to 279nm wavelength UVC‐LEDs for inactivation of Gram positive and Gram negative foodborne pathogenic bacteria and yeasts. Food Res. Int. 2017, 97, 280–287, doi:10.1016/j.foodres.2017.04.009. Microorganisms 2021, 9, 430 10 of 10 30. Tataurov, A.V.; You, Y.; Owczarzy, R. Predicting ultraviolet spectrum of single stranded and double stranded deoxyribonucleic acids. Biophys. Chem. 2008, 133, 66–70, doi:10.1016/j.bpc.2007.12.004. 31. Sinha, R.P.; Häder, D.P. UV‐induced DNA damage and repair: A review. Photochem Photobiol Sci. 2002, 1, 225–236, doi:10.1039/b201230h. 32. Abed, J.; Emgård, J.E.; Zamir, G.; Faroja, M.; Almogy, G.; Grenov, A.; Sol, A.; Naor, R.; Pikarsky, E.; Atlan, K.A.; et al. Fap2 mediates Fusobacterium nucleatum colorectal adenocarcinoma enrichment by binding to tumor‐expressed Gal‐GalNAc. Cell Host Microbe 2016, 20, 215–225, doi:10.1016/j.chom.2016.07.006. 33. Komiya, Y.; Shimomura, Y.; Higurashi, T.; Sugi, Y.; Arimoto, J.; Umezawa, S.; Uchiyama, S.; Matsumoto, M.; Nakajima, A. Patients with colorectal cancer have identical strains of Fusobacterium nucleatum in their colorectal cancer and oral cavity. Gut 2019, 68, 1335–1337, doi:10.1136/gutjnl‐2018‐316661. 34. Ashare, A.; Stanford, C.; Hancock, P.; Stark, D.; Lilli, K.; Birrer, E.; Nymon, A.; Doerschug, K.C.; Hunninghake, G.W. Chronic liver disease impairs bacterial clearance in a human model of induced bacteremia. Clin. Transl. Sci. 2009, 2, 199–205, doi:10.1111/j.1752‐8062.2009.00122.x. 35. American Conference of Governmental Industrial Hygienists (ACGIH®). 2017 TLVs and BEIs Book; American Conference of Governmental Industrial Hygienists (ACGIH®): Cincinnati, OH, USA, 2017; ISBN 978‐1‐607260‐90‐5. 36. Narita, K.; Asano, K.; Morimoto, Y.; Igarashi, T.; Nakane, A. Chronic irradiation with 222‐nm UVC light induces neither DNA damage nor epidermal lesions in mouse skin, even at high doses. PLoS ONE 2018, 13, e0201259, doi:10.1371/jour‐nal.pone.0201259. 37. Barbadoro, P.; Ponzio, E.; Coccia, E.; Prospero, E.; Santarelli, A.; Rappelli, G.G.L.; DʹErrico, M.M. Association between hyper‐tension, oral microbiome and salivary nitric oxide: A case‐control study. Nitric Oxide. 2021, 106, 66–71, doi:10.1016/j.niox.2020.11.002. 38. Yoneyama T, Yoshida M, Matsui T, Sasaki, H. Oral care and pneumonia. Oral Care Working Group. Lancet 1999, 354, 515, doi:10.1016/s0140‐673675550‐1. | RESEARCH GATE |
Corporate Ownership & Control / Volume 14, Issue 2, Winter 2017, Continued - 1 211 STABILITY IN ISLAMIC, CONVENTIONAL, AND SOCIALLY RESPONSIBLE BANKS: EVIDENCE FROM MENA COUNTRIES Majed Alharthi* *Assistant Professor in Finance, College of Business, Department of Finance, King Abdul Aziz University, Saudi Arabia, Rabigh, majed.alharthi2@gmail.com Abstract This study empirically estimates financial stability and its determinants in 40 Islamic banks, 168 conventional banks, and 8 socially responsible banks (SRBs) in MENA region during the period 2005-2012. The dependent variables in this study are capital ratio (equity to total assets) and z-score. The statistical approaches to find the relationship between financial stability indicators and their determinants are ordinary least square (OLS) and fixed effects model FEM). The results suggest that the SRBs are the most stable banks while, Islamic banks are highly risky. Moreover, conventional banks score the minimum capitalisation. The stability in Islamic banks is positively affected by ROA and age. Furthermore, the main determinants of capitalisation in Islamic banks are operating leverage, GDP, and market capitalisation. In conventional banking, size and profitability are important to stability. The capitals have effective associations with lending, ROA, and market development. In SRBs, banks achieve better stability in countries with higher inflation. This study could help bankers, policy makers and economists who focus on MENA region. The coverage of period 2005-2012 could be a limitation and the availability of data for the Islamic and socially responsible banks in MENA area could be another limitation as well. Keywords: Islamic Banks, Conventional Banks, Socially Responsible Banks, Z-Score, Capital Ratio JEL Classification: G21 DOI: 10.22495/cocv14i2c1p7 1. INTRODUCTION After the global financial crisis (GFC) in 2007, focusing on insolvency risk analysis became more important as the majority of banks achieved a huge amount of loss or bankruptcy. The stability indicators are z-score, as employed by Ghosh (2014), and capital (equity to total assets) ratio (Horvàth et al., 2014). In this study, both measures are used: z-score and capital ratio during the period 2005-2012. The z-score was developed by Boyd et al. (1993) and statistically concerns the probability of bankruptcy. A higher z-score means the bank is more stable and less likely to go bankrupt. Regarding the capitalisation ratio, better values indicate that the firm is well capitalised and stable. After measuring the stability and comparing values between Islamic, commercial, and socially responsible banks (SRB), finding the determinants of stability is the main topic. Financial stability can be affected by many factors, for example internal variables such as size of bank (log of total assets) and external factors like inflation. However, these variables can affect the stability positively or negatively. In fact, the relationship between stability and its determinants can help with the decision about which variable to use more or less - or even neglect altogether – in order to avoid insolvency risks and to raise financial stability in the banking sector. In accordance with macroeconomics, identifying the factors that affect banks leads to allocating more recovery plans by the banks’ policy makers. However, this study identifies the stability (z-score and capitalisation which can be the explained variable) and its determinants (explanatory variables). To understand more about the characteristics of socially responsible banks, we have to consider the main aims of this type of bank to be as follows (Kansal et al., 2014): 1. SRBs are concerned more with social issues (e.g. sponsoring community events, local scholarships, training courses, summer training for students) rather than achieving profits. 2. SRBs care about environmental issues, such as energy saving, green funds, and organic agriculture (Kansal et al., 2014). The most recent study on SRB was made by Gutiérrez-Nieto et al. (2016). This study has focused on the credit score system for socially responsible lending. This study concludes that SRB is a financial institution that only fund target groups or causes, generally social and/or environmentally orientated. When considering Islamic and conventional banks, the main difference between them is that conventional banks deal with interest (Riba), whereas Islamic banks operate interest-free. During the time of the global crisis, the banking sector in the Middle East and North Africa faced many hurdles due to a reduction in deposits and loans as internal factors. In addition, the external Corporate Ownership & Control / Volume 14, Issue 2, Winter 2017, Continued - 1 212 factor of economic recession allowed an increment in cash flows and GDPs. Most MENA countries’ incomes depend on the export of oil. As a result, oil prices have affected MENA’s economies as well as their banking sector. So, this study aims to compare the financial stability in MENA region between Islamic, conventional, and socially responsible banks. Further, this study investigates the reasons behind the instability. This study aims to answer several questions: Q1: Are Islamic, conventional, and socially responsible banks financially stable over the period of the study? Which type of bank is the most stable? Q2: What are the determinants of financial stability in Islamic, conventional, and socially responsible banks? Are the determinants different for these bank types? Q3: How do the internal and external factors affect stability? Are the influences positive or negative? In fact, this study makes several contributions to the current literature. Firstly, it is the first study that concerns the financial stability of the socially responsible banking system. Secondly, comparing the financial stability of Islamic, conventional, and socially responsible banks is a contribution to the literature. Thirdly, discovering what impact the financial crisis had on SRBs will add to financial knowledge. The study is organised as follows. Section 2 reviews the previous literature. Section 3 presents the data and methodology. Section 4 shows and discusses the empirical results. Finally, Section 5 concludes the study. 2. LITERATURE REVIEW AND HYPOTHESES 2.1. Literature Review As previously mentioned, stability studies have become more important since the global financial crisis in 2007. Many studies concentrate on conventional banks (Cubillas & González, 2014; Fu et al., 2014; William, 2014), but there is a limited number of studies analysing Islamic banks or comparing Islamic and commercial banking stability (Beck et al., 2013; Ghosh, 2014). In terms of cooperative banks, only one study was conducted by Mirzaei et al. (2013) that compared Islamic, real estate, savings and cooperative banks in the Middle East and Eastern Europe using z-score as a stability indicator (dependent variable). However, stability indicators can be non-performing loans, credit risk, net interest margin and z-score. Most studies focusing on z-score indicate the percentage of bankruptcy. Many studies found the determinants of stability through statistical regressions such as OLS, which is the most common model (Chalermchatvichien et al., 2014; Jeon & Lim, 2013; Lee & Chih, 2013; Srairi, 2013). Rajhi and Hassairi (2013) discuss the Islamic banking stability for MENA and Southeast Asian regions for the period 2000-2008. This study explains the causes of stability as size of banks, loans services, liquidity and GDP. In contrast, efficiency ratio and inflation led to instability. These results allow managers to attract more clients to borrow; one way could be by minimising the lending interest. Additionally, bankers could consider reducing their costs, as efficiency ratio has a negative and significant sign. With regards to the MENA region, Srairi (2013) compared the determinants of risk using 10 countries over the period 2005-2009. This study evaluated 175 Islamic and conventional banks in MENA. There are three types of explanatory variables, namely ownership, bank-specific variables and financial indicators. Srairi (2013) adopted OLS regression as a statistical approach to examine the determinants of z-score. This study yielded three main results: family banks tend to be more stable than company and state-owned banks; concentration (equity % participation by the largest shareholder of the bank), size, loan growth, operating leverage, diversification, banking sector development, and shareholders’ rights and bank concentration (assets of 3 largest banks to total assets of all banks in the country) were found to have a significant and positive z-score which leads to making the banks less risky; and efficiency ratio (cost to income) was found to be decreasing the z-score which raises the insolvency risk. Overall, during the period, conventional banks have more mean z-score (21.7) than Islamic banks (20.8) which makes the conventional banks more stable and resistant against crises in MENA countries. A significant comparison has been conducted between Middle Eastern banks and Eastern European banks over the period 1999-2008, examining 1929 banks by Mirzaei et al. (2013). The empirical results explain that for banks in the Middle East, market share, interest rate, capital ratio and overheads to total assets ratio have a significant and negative relationship with z-score, while inflation and bank size were found to be decreasing the z-score (increasing the risk). Regarding the Eastern European banks, the findings indicate that z-score (stability) was influenced significantly and positively by market share, interest rate spread, capital ratio, off-balance sheet to total assets, bank age, inflation, and GDP. In contrast, overheads to total assets ratio is negative and significant with z-score at the 5% level. Overall, the most stable banks were found to be the foreign banks (Middle Eastern and Eastern European) through the period. Concerning the GCC banking market, Ghosh (2014) tested the relation between risk and capital for 57 conventional banks and 46 Islamic banks in the GCC region for the period 1996-2011. The main finding (after employing the 2SLS model) shows that banks generally increase capital in response to an increase in risk. However, the determinants of risk identified by Ghosh (2014) such as funding (short-term funding over total assets) and listed banks explain the z-score significantly and positively. In contrast, income diversification was found to have a significant and negative relationship with z-score (the risk indicator) which made banks more risky through the period. Regarding to capitalisation, the relationship between the capital ratio and size was found to be significant and negative, while ROA were improving the capital significantly over the period. In general, Ghosh (2014) proposed that the GFC does not impact the stability of banks in GCC. Corporate Ownership & Control / Volume 14, Issue 2, Winter 2017, Continued - 1 213 2.2. Hypotheses’ formulations Based on the literature review on stability, this study examines the determinants of stability in MENA region using the most effective internal variables, namely z-score, capital ratio, bank size, loan intensity, credit risk, ROA, operating leverage, age of bank, and foreign, domestic and public ownerships. On the other side, GDP, inflation, market capitalisation and global financial crisis can be examined as external variables. 2.2.1. Internal variables 1. Z-score: Horvàth et al. (2014) consider the commercial banking sector in their study examining the determinants of capitalisation including z-score in the Czech Republic for the period 2000-2010. The association between capitalisation and z-score was positive and significant. This demonstrates that more capitalisations led to financial stability and less default risk. H1. There is a significant relationship between z-score and stability. 2. Capital ratio: Capitalisation is one of the most important bank-specific variables to describe stability in the banking sector. There are many examples showing the importance of capitalisation such as the studies of Ghosh (2015), Köhler (2015) and Tabak et al. (2015), who claim that higher equity leads to raising constancy in the banking industry. Consequently, more capitalisation lets banks face any threat of failure. On the contrary, Tabak et al. (2013) argue that capitalisation decreased the stability in the Latin American banking sector for the period 2001-2008. In order to avoid any bankruptcy risks, banks need to reduce their capitalisation (which enhances the stability). H2. There is a significant relationship between capital ratio and stability. 3. Bank size: The size of banks (total assets) plays a very important role on stability in the recent studies. There are various points of view considered regarding banking size. Some studies confirm that size keeps banks stable with less default risk and others go against this orientation. An example of a study that considered the relationship between stability and bank size as positive is the most recent study of Tabak et al. (2015), who examined the stability (z-score) determinants of 76 Brazilian commercial banks for the period 2001-2011. The result of this study supports having more total assets in the Brazilian banking industry. The same result was found by Cubillas and González (2014) and Agoraki et al. (2011). Conversely, Köhler (2015) proves a negative and significant correlation between size and stability (z-score). H3. There is a significant relationship between bank size and stability. 4. Loan intensity: Köhler (2015) found that banks with a larger loan portfolio have significantly higher z-scores. This encourages banks to raise their lending activities due to being further from insolvency risk, as noted for European commercial banks in the period 2002-2011. This contradicts the results of Rumler and Waschiczek (2014). Berger et al. (2009) analysed both stability indicators (z-score and capital ratio) for 23 countries for the period 1999-2005. This study discouraged banks from giving more loans as lending reduced the capital and increased the default (bankruptcy) risk which logically made banks unstable. H4. There is a significant relationship between loan intensity and stability. 5. Credit risk: Soedarmono et al. (2011) conducted a study which concentrated on examining the financial stability using a sample of commercial banks from 12 Asian countries between 2001 and 2007. They document that the relationship between liquidity and stability ratios (z-score and capitalisation) are positive and significant (the same conclusion as Nguyen & Nghiem, 2015; Dima et al., 2014; Jeon & Lim, 2013). This means that banks could enhance their stability through providing fewer loans to cover the withdrawals of clients. On the other hand, Lee and Chih (2013) find a negative and significant correlation between z-score and loans to deposits ratio. As a result, increasing the loans to deposits ratio let banks take less risk over the period 2004-2011. The results of Dong et al. (2014) revealed that a loan to deposits ratio was irrelevant to stability in the Chinese commercial banking sector through the period 2003-2011. H5. There is a significant relationship between credit risk and stability. 6. Return on assets (ROA): The statistical results of Anginer et al. (2014) claim that the profitability ratio (ROA) improved the steadiness of banks over the period 2004-2009. This result also concludes that earnings of banks are very important in terms of profits and can save banks from default risk. This is consistent with Ghosh (2014) in terms of capital ratio. The rest of the studies in the literature could not provide any further evidence that profits influence stability and risk in banking systems (Ghosh, 2014; Srairi, 2013; Tabak et al., 2015Tan & Floros, 2013; Zhang et al., 2015). H6. There is a significant relationship between profitability and stability. 7. Operating leverage: A few studies have examined the determination of fixed assets intensity on stability. Srairi (2013) concentrated on the stability indicators of 10 MENA countries including 175 Islamic and conventional banks for the period 2005-2009. Based on the results, the Islamic and conventional banks in MENA countries were recommended to purchase more fixed assets as the z-score and operating leverage were found to be significant and positive, which is consistent with Williams’s (2014) findings. Berger et al. (2009) had an opposite result which suggested that fixed assets (negative correlation with z-score) made the financial stability worse and raised the risk of failure over the period 1999-2005 in their sample of 23 countries (1091 Asian commercial banks). This result allows policy makers to sell more fixed assets; depreciation could be due to the high cost of the fixed assets. In contrast, this study approved that Corporate Ownership & Control / Volume 14, Issue 2, Winter 2017, Continued - 1 214 fixed assets intensity increased the capitalisation significantly over the period. H7. There is a significant relationship between operating leverage and stability. 8. Age: Lee and Chih (2013) compared the stability (z-score) of small (185) and large (57) banks in China for the period 2004-2011. Lee and Chih (2013) argue that the experience in the Chinese banking sector affected the stability of large banks. Higher experience time leads to steadier banks with fewer insolvency risks. In this study, age is unimportant for small banks (similar to Dedu & Chitan, 2013). Schaeck and Cihàk (2014) included z-score and capital ratio as dependent variables to find if the age impact the stability or not examining ten European Countries for the period 1995-2005. As a result of this study, time trend was found to be highly important to both z-score and capitalisation. Another study also focused on age; Mirzaei et al. (2013) compared the stability (z-score) between emerging economies and advanced economies in the Middle East and Eastern Europe through the period 1999-2008. This study concludes that older banks in advanced economies were financially more settled and less risky. On the other side, age was found to be an insignificant variable to stability (z-score). H8. There is a significant relationship between age and stability. 9. Foreign ownership: Berger et al. (2009) argue that increment in levels of foreign banks increased the probabilities of failure for 23 countries over the period 1999-2005, as the relationship between z-score and foreign ownership was significant and negative (similar to Kasman & Kasman, 2015). This finding contradicts the conclusion of Mirzaei et al. (2013) who approve that foreign banks allow the banking sector to be more stable and less risky. On the other side, Berger et al. (2009) claim that foreign banks enhance capital ratios significantly. H9. There is a significant relationship between foreign ownership and stability. 10. Domestic ownership: Tabak et al. (2013) examined the determinants of z-score inefficiency for 17 Latin American countries through the period 2001-2008. The results suggest that private banks significantly increased financial instability. García-Kuhnert et al. (2015) found an insignificant correlation between z-score (financial stability) and private banks. H10. There is a significant relationship between domestic ownership and stability. 11. Public ownership: Finally, Rumler and Waschiczek (2014) examined the impact of public banking roles on financial stability for commercial Austrian banks during the period 1995-2010. Their findings conclude that government involvement in the banking sector was important to improve their financial effectiveness. This result is consistent with the study of Agoraki et al. (2011) but contradicts to the result of ElBannan (2015) and Barakat and Hussainey (2013). H11. There is a significant relationship between public ownership and stability. 2.2.2. External variables 1. Gross Domestic Product (GDP): The GDP growth is a macroeconomic indicator that has been examined by most studies in stability of the banking industry. The majority of studies prove that there is a positive relationship between GDP development and stability (z-score) in the banking industry (e.g. Köhler, 2015 and Anginer et al., 2014). In contrast, a minority of studies proposed that GDP decreases stability and increases the probability of bankruptcy. Examples for this case can be seen in the studies of Cubillas and González (2014) and Dong et al. (2014), who confirm that banks in better GDP growth found hurdles in growing (investing) their capital (negative association between capital ratio and GDP growth). According to capitalisation, Nguyen and Nghiem (2015) also confirmed that GDP affected the stability negatively. However, Chalermchatvichien et al. (2014) could not estimate any correlation between GDP and z-score (stability). H12. There is a significant relationship between GDP and stability. 2. Inflation: Rumler and Waschiczek (2014) investigated the factors that determine the bank-taking risk, focusing on the Austrian banking industry for the period 1995-2010. In fact, they found that inflation reduced the bank risk-taking. As a result, the constancy of Austrian commercial banks was enhanced. Many studies support Rumler and Waschiczek’s (2014) result (for example, Barakat & Hussainey; 2013; Bertay et al., 2013; Bourkhis & Nabi, 2013; Tan & Floros, 2013). In the private banking sector, Nguyen and Nghiem (2015) and Horvàth et al. (2014) found the same result in terms of capitalisation. This result encourages banks to expand their activities in high inflation rates but some studies discourage banks from operating more due to a negative and significant relationship between z-score and inflation rates (see Cubillas & González, 2014; Delis et al., 2012; Houston et al., 2010; Köhler, 2015; Mirzaei et al., 2013). Nguyen and Nghiem (2015) arrived at the same conclusion in accordance with the public banking sector. In fact, the inflation rates are not always an influential variable to stability (Delis et al., 2012; Srairi, 2013). H13. There is a significant relationship between inflation and stability. 3. Market capitalisation: Nguyen et al. (2012) indicate that financial development in the Asian economies such as in the stock market is very important to the banking industry, as they investigated the financial stability (z-score) determinants of 151 Asian commercial banks including Bangladesh, India, Pakistan and Sri Lanka over the period 1998-2008. Results from Dima et al. (2014) show the same association between stock market growth and stability (z-score) using a sample of commercial banks in 63 developed and developing countries through the period 1997-2010. In addition, Lee and Hsieh (2014) also found a positive and significant relationship between capital ratio and stock market earnings. Anginer et al. (2014) and Tan and Floros (2013) argue that stock Corporate Ownership & Control / Volume 14, Issue 2, Winter 2017, Continued - 1 215 market development did not influence stability in the banking industry. H13. There is a significant relationship between market capitalisation and stability. 4. Global financial crisis (GFC): For GFC, there is no study finding a positive relationship with stability. Williams’s (2014) findings suggest that Asian commercial banks faced a risk of bankruptcy over the GFC period. In addition, Anginer et al. (2014) conclude that bank stability was affected by GFC. Some studies found no influence of GFC in banks, such as Ghosh (2014), Bourkhis and Nabi (2013) and Nguyen et al. (2012). H14. There is a significant relationship between market capitalisation and stability. 3. METHODOLOGY 3.1. Data of the study The data in this study was extracted from two main sources: Bankscope (Bankscope, 2016) and World Bank databases (World Bank, 2016). For Bankscope, the data was extracted from balance sheets and income statements of 216 banks, and of those being 40 Islamic banks, 168 conventional banks, and 8 socially responsible banks (more details in Table 1 below) during the period 2005-2012. The data has been gathered from Middle Eastern and North African (MENA) regions including Islamic, conventional and socially responsible banks. Regarding the banks, data was collected from 20 countries, namely Algeria, Egypt, Iran, Iraq, Lebanon, Libya, Malta, Morocco, Israel, Jordan, Palestine, Syria, Tunisia and Yemen, as well as the Gulf Cooperation Council (GCC) countries, which are considered to be oil exporter countries in the Middle Eastern region, namely Bahrain, Kuwait, Oman, Qatar, the Kingdom of Saudi Arabia, and United Arab Emirates (Bankscope, 2016). The main purposes of choosing MENA countries are: 1. Most MENA countries have the same culture and language (Arabic). 2. MENA countries contain Islamic, conventional, and socially responsible banks (globally, the highest number of Islamic banks can be found in MENA region). 3. Availability of data for MENA region. 4. The first international Islamic bank was located in the Middle East in 1975 in Jeddah, Saudi Arabia, known as Islamic Development Bank (Islamic Development Bank, 2016); whereas, the first domestic Islamic bank was established in Dubai, UAE in 1975, known as Dubai Islamic Bank (Dubai Islamic Bank, 2016). 5. Some MENA countries lead in the global export of oil, especially GCC countries whose GDP is based on the oil sector. However, only 8 MENA countries out of 12 members are in the Organisation of Petroleum Exporting Countries (OPEC Organisation, 2016). 6. MENA countries considered to have emerging economies. So, studying their stability would enhance their economics. 3.2. Independent variables The bank-specific variables in this study are z-score, capital ratio, size of banks, loan intensity, credit risk, ROA, operating leverage, age of banks, z-score, and domestic, foreign and public ownerships. On the other side, four main country indicators are examined as GDP, inflation, market capitalisation and global financial crisis. In Table 2, we conclude the descriptive statistics for the independent variables for Islamic, conventional and socially responsible banks for the period 2005-2012. 3.3. Dependent variables Based on the literature, the dependent variables would be the natural logarithm of z-score (e.g. Chalermchatvichien et al., 2014) and capital ratio (Nguyen & Nghiem, 2015). Table 2 illustrates the data description of z-score and capital ratio for MENA countries over the period 2005-2012. Table 2 explains that based on both stability’s indicators (z-score and capital ratio), the socially responsible banks were found to be the most financially settled type of banking by far (scoring averages 4.877 and 0.696 for z-score and capital ratio, respectively). This occurred due to SRBs having high capitals over the period 2005-2012, which allowed them to face insolvency risks. These results encourage all banks in MENA region to provide more social services to be more fixed and less risky. In contrast, Islamic banks were unstable and highly risky (average z-score = 2.715). Finally, conventional banks scored the least capitalisation (0.15). The reasons behind the stability’s indicators can be revealed in results sections through finding the determinants of financial stability. This study employs z-score as an explained variable to describe stability. Boyd et al. (1993) proposed the z-score formula: Z-score = (ROA+E/TA)/(S.D.ROA) (1) where, ROA: return on assets E/TA: equity to total assets (or capital) ratio S.D. ROA: standard deviation of return on assets According to capitalisation ratio, Horváth et al. (2014) employed the capital ratio as follows: Capital ratio = Equity / Total Assets (2) Corporate Ownership & Control / Volume 14, Issue 2, Winter 2017, Continued - 1 216 Table 1. Number of banks in each country N Countries GDP (million US$) in 2015 World Islamic Conventional SRBs Total Rank Banks Bank 1 Saudi Arabia 777,870 20 3 9 0 12 2 UAE 402,340 28 6 17 0 23 3 Iran 367,098 31 7 0 1 8 4 Israel 290,643 36 0 8 0 8 5 Egypt 271,427 39 2 21 0 23 6 Iraq 229,327 45 0 2 0 2 7 Algeria 212.453 48 1 9 3 13 8 Qatar 202,450 49 3 6 0 9 9 Kuwait 175,787 55 2 6 0 8 10 Morocco 103,824 60 0 8 1 9 11 Oman 77,116 63 0 7 0 7 12 Syria 71,998 65 0 5 0 5 13 Libya 65,516 69 0 5 0 5 14 Tunisia 46,995 82 1 8 2 11 15 Lebanon 45,019 85 0 28 0 28 16 Yemen 40,415 89 4 1 0 5 17 Jordan 33,858 90 1 7 0 8 18 Bahrain 32,791 92 9 15 0 24 19 Malta 9,545 135 0 4 1 5 20 Palestine (Gaza) 6,641 148 1 2 0 3 Total 40 168 8 216 Source: International Monetary Fund (2016) Table 2. Variable definitions and summary statistics Variables Islamic Banks Conventional Banks Socially Responsible Banks All Banks Definition Obs Mean S.D. Obs Mean S.D. Obs Mean S.D. Obs Mean S.D. Dependent variables Z-score log(Z-score), where Z-score = (ROA + capital ratio)/S.D. (ROA) 291 2.715 0.883 1277 3.079 1.050 64 4.877 0.949 1653 3.073 1.103 Capital ratio Capital/total assets 291 0.226 0.215 1277 0.150 0.202 64 0.696 0.255 1653 0.191 0.253 Independent variables Bank-specific variables Size Log (total assets) 291 2.715 0.883 1277 8.052 1.754 64 4.877 0.949 1653 8.032 1.759 Loan intensity Loans/total assets 291 0.226 0.215 1277 0.470 0.388 64 0.696 0.255 1653 0.458 0.366 Credit risk Loans/deposits 291 7.907 1.755 1277 1.104 14.677 64 8.406 0.963 1653 1.372 16.534 ROA Return on assets = net income/total assets 291 0.485 0.222 1277 0.015 0.039 64 0.134 0.132 1653 0.012 0.050 Operating leverage Fixed assets/total assets 291 2.722 24.643 1277 0.016 0.016 64 0.809 0.221 1653 0.017 0.019 Age Log (years since establishment) 291 0.007 0.068 1277 3.665 0.593 64 0.005 0.013 1653 3.552 0.644 Foreign ownership Dummy = 1 if a bank owned by foreign, else zero 291 0.022 0.019 1277 0.434 0.496 64 0.019 0.014 1653 0.415 0.493 Domestic ownership Dummy = 1 if a bank owned by local, else zero 291 3.109 0.612 1277 0.379 0.485 64 3.460 0.595 1653 0.378 0.485 Government ownership Dummy = 1 if a bank owned by government, else zero 291 0.285 0.452 1277 0.193 0.395 64 0.500 0.504 1653 0.226 0.419 Country-specific variables GDP Log ( GDP) 291 25.202 1.218 1277 25.053 1.079 64 25.014 1.056 1653 25.113 1.145 Inflation Inflation rates 291 0.104 0.100 1277 0.065 0.072 64 0.055 0.053 1653 0.071 0.079 Market capitalisation Market capitalisation to GDP 291 0.568 0.501 1277 0.567 0.492 64 0.228 0.243 1653 0.558 0.492 Global Financial Crisis Dummy = 1 for the period 2007-2009, otherwise zero 291 0.395 0.490 1277 0.388 0.488 64 0.375 0.488 1653 0.390 0.488 Corporate Ownership & Control / Volume 14, Issue 2, Winter 2017, Continued - 1 217 3.4. Main models The main models of the study that obtained from OLS and fixed effects models (through STATA 14) can be as follows: Z-scoit = α + β1 EQTAit + β2 LTAit + β3 LOANSTAit + β4 LOANSDEPOit + β5 ROAit + β6 FATAit + β7 LAGEt + β8 FOREi + β9 DOMi + β10 GOVi + β11 LGDPt + β12 INFLATIONt + β13 MCAPt + β14 GFCt + ɛit (3) i = 1….n; t = 1….n Capit = α + β1 LOGZit + β2 LTAit + β3 LOANSTAit + β4 LOANSDEPOit + β5 ROAit + β6 FATAit + β7 LAGEt + β8 FOREi + β9 DOMi + β10 GOVi + β11 LGDPt + β12 INFLATIONt + β13 MCAPt + β14 GFCt + ɛit (4) i = 1….n; t = 1….n As mentioned, Z-scoit denotes the dependent variable which is the z-score and Capit represents capital ratio, i is the observations, t is time, α is the constant, β denotes the coefficient of variables and ɛit is the error term. On the other side, the independent variables are size (LTA), loans intensity (LOANSTA), credit risk (LOANSDEPO), return on assets (ROA), operating leverage (FATA), age (LAGE), foreign ownership (FORE), domestic ownership (DOM), public ownership (GOV), gross domestic production (GDP), inflation rates (INFLATION), market capitalisation (MCAP), and global financial crisis (GFC). However, before examining the relationship between the dependent and independent variables, we need to conduct a correlation matrix to insure that there is no multicollinearity. As a result, Table 3 indicates that the maximum amount is 36.16% (the correlation between size and age) which is less than 80% (Studenmund, 2005). This means that no potential multicollinearity problem exists. 4. DATA ANALYSIS AND RESULTS The determinants of financial stability of Islamic, conventional and socially responsible banks using OLS regression are displayed in Table 4. The findings suggest that the hypotheses which support z-score in Islamic banks are H4, H6, H8 and H14, while the determinants of capital ratio are H3, H4, H7, H10, H11, H12 and H14. According to conventional banks, H3-H12 and H14 affected z-score significantly, whereas H3-H6, H8 and H14 were found to be significant with capitalisation. Regarding the socially responsible banks, H3, H7, H8 and H12 confirmed a significant relationship with z-score and H5-H7, H10 and H12 impacted capital ratio effectively. H3. Bank size: OLS findings suggest that larger conventional and socially responsible banks were more stable and less risky than smaller banks. Many studies confirm that higher total assets enhance stability, such as the studies of Tabak et al. (2015) and Cubillas and González (2014). The reason for this result could be due to the fact that larger banks are more likely to gain profits from economies of scale than smaller banks, which may have a higher degree of production differentiation and loan diversification. In Islamic and conventional banks, the relationship between capital ratio and size of banks is strongly negative at the 0.1% level. This indicates that smaller sized banks are better capitalised than larger sized banks (consistent with Ghosh, 2014). H4. Loan intensity: The results confirm that Islamic banks that provide more loans tend to be financially constant (in line with Köhler, 2015), but at the same time, lower their capitals significantly (similar to ElBannan, 2015). Regarding the conventional banks, the opposite situation was demonstrated in that loans allowed conventional banks to be riskier (Bourkhis & Nabi, 2013) but supported the capitalisation significantly. H5. Credit risk: The findings show that conventional banks suffering from increment in credit risk are affected in their financial stability and capitalisation. Consequently, conventional banks could increase deposits and minimise loans. This can be achieved through encouraging clients to make more deposits with higher deposit interests. Furthermore, banks could discourage customers from applying for more loans by raising the lending interest. This finding is in line with Nguyen and Nghiem (2015) concerning Indian commercial banks. On the other side, socially responsible banks’ capitalisation decreased the credit risk significantly. H6. ROA: The profitability was found to be highly important to Islamic and conventional banks’ financial stability. However, there is a negative and significant correlation between capitalisation and ROA in socially responsible banks. This relationship shows that the costs of banks was found to be greater than income. Banks could find strategies to cut costs by achieving more stability and avoiding any default risks. H7. Operating leverage: Concentrating on operating leverage ratio, there is a positive and significant association between operating leverage and stability in conventional and socially responsible banks. In other words, greater concentration of fixed assets against total assets leads to an increase in stability. The assumption is correlated with Wang et al. (2015). This underlines strategies with fixed assets such as purchasing more fixed assets. According to capitalisation, Islamic banks have a strong and significant relationship (at 0.1%) between capital and fixed assets intensity. This contradicts with the correlation between capital ratio and operating leverage in SRBs. This confirms that SRBs could reduce fixed assets (selling or depreciation) in order to enhance their capitals. In this case, the costs of fixed assets significantly increased over the period. Corporate Ownership & Control / Volume 14, Issue 2, Winter 2017, Continued - 1 218 Table 3. Correlation matrix for variables N Independent Variables 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 1 Z-score 1 2 Capital ratio 0.086 1 3 Size 0.141 -0.201 1 4 Loan intensity -0.035 0.237 0.204 1 5 Credit risk 0.011 0.129 -0.036 -0.011 1 6 ROA 0.120 0.071 0.058 0.190 -0.034 1 7 Operating ratio -0.021 0.350 -0.277 0.235 0.017 -0.136 1 8 Age 0.242 -0.201 0.362 0.076 -0.065 0.023 -0.121 1 9 Foreign banks -0.135 0.099 -0.235 -0.133 -0.003 -0.035 0.069 -0.155 1 10 Domestic banks 0.072 -0.184 -0.012 0.012 -0.031 -0.010 -0.080 0.081 -0.658 1 11 Government banks 0.092 0.162 0.308 0.117 0.037 0.043 0.003 0.106 -0.387 -0.395 1 12 GDP -0.128 0.045 0.333 0.236 0.000 0.074 0.015 -0.008 -0.136 -0.089 0.263 1 13 Inflation -0.133 0.026 -0.092 -0.082 0.077 -0.102 0.079 -0.127 0.017 -0.090 0.066 -0.086 1 14 Market capitalisation -0.145 0.084 0.064 0.104 0.029 0.046 -0.093 0.010 0.030 0.089 -0.142 -0.122 0.019 1 15 Financial crisis -0.009 0.001 -0.013 -0.015 -0.019 -0.039 0.012 -0.010 0.017 0.001 -0.023 -0.019 0.057 0.051 1 Table 4. OLS results Stability Islamic banks Conventional banks Socially responsible banks Z-score Capital Ratio Z-score Capital Ratio Z-score Capital Ratio Bank-specific variables (H1) Z-score -0.00199 -0.00285 0.0139 (-0.16) (-0.72) -0.78 (H2) Capital ratio -0.0491 -0.143 0.883 (-0.16) (-0.72) -0.78 (H3) Size -0.0121 -0.053*** 0.0715*** -0.0274*** 0.293* 0.0128 (-0.27) (-6.32) -3.51 (-9.83) -2.42 -0.8 (H4) Loan intensity 0.594* -0.279*** -0.289** 0.131*** -1.066 0.139 -2.24 (-5.47) (-2.98) -9.89 (-1.38) -1.44 (H5) Credit risk -0.000051 0.00106** 0.00523** 0.0017*** 0.194 -0.166*** (-0.03) -2.97 -2.74 -6.44 -0.52 (-4.05) (H6) ROA 3.471*** 0.237 5.597*** 2.552*** 10.16 -2.250*** -4.93 -1.61 -5.7 -21.13 -1.98 (-3.83) (H7) Operating leverage 1.652 2.800*** 4.092* 0.343 17.33* -2.876** -0.65 -5.79 -2.03 -1.2 -2.45 (-3.39) (H8) Age 0.569*** -0.0585** 0.155** -0.0142* -0.569*** 0.0329 -6.38 (-3.10) -3.09 (-2.00) (-3.86) -1.6 (H9) Foreign ownership -1.858*** -0.0187 -0.567* 0.0252 (-5.18) (-0.37) (-2.11) -0.72 (H10) Domestic ownership 0.103 -0.0693** -1.505*** -0.0529 -0.832 -0.721*** -0.95 (-3.21) (-4.26) (-1.05) (-0.89) (-12.20) (H11) Public ownership 0.105 -0.00642 -1.752*** 0.0248 -0.427 0.0329 -0.7 (-0.21) (-5.00) -0.5 (-1.04) -0.63 Macroeconomic variables (H12) GDP -0.0372 0.054*** -0.178*** -0.00467 -0.885*** 0.0518* (-0.61) -4.6 (-6.12) (-1.12) (-6.41) -2.32 (H13) Inflation 0.2 0.0983 -1.750*** 0.0446 9.784*** -0.0872 -0.41 -1.01 (-4.47) -0.8 -8.22 (-0.38) (H14) Market capitalisation -0.331** 0.138*** -0.212*** 0.0286*** -0.93 0.12 (-3.16) -6.99 (-3.55) -3.38 (-1.86) -1.92 (H15) Global financial crisis -0.0142 0.0233 0.0315 -0.00241 -0.0759 0.0224 (-0.16) -1.35 -0.56 (-0.30) (-0.75) -1.82 Sigma 1.738 -0.543* 8.328*** 0.447*** 25.86*** -0.683 _cons -1.27 (-1.98) -10.33 -3.79 -8.73 (-1.17) R2 0.3818 0.5775 0.1445 0.5384 0.9011 0.9085 Number of banks 40 40 168 168 8 8 Obs 291 291 1277 1277 64 64 Corporate Ownership & Control / Volume 14, Issue 2, Winter 2017, Continued - 1 219 H8. Age: According to age of banks, older Islamic and conventional banks scored better z-scores than new banks, which means that older banks are less risky than new banks due to their having more experience in the banking sector and dealing with failure risks. Studies by Lee and Chich (2013) and Mirzaei et al. (2013) demonstrate the same result. This goes against the relationship between age and stability in SRBs as the most recent established banks tend to be steadier than older banks. Depending on capitalisation, the age of bank inversely and significantly influences capitals in conventional banks. H9. Foreign ownership: With regard to ownership, the OLS model confirms that an increment in levels of foreign banks raised the risk of insolvency in conventional and socially responsible banks. Kasman and Kasman (2015) claim the same finding in Turkey. As a result, the international banks are discouraged to invest in the banking sector in MENA region. H10. Domestic ownership: Based on the OLS coefficient, the conventional banks in MENA were threatened with bankruptcy over the period 2005-2012. In addition, the concentration of local Islamic and socially responsible banks decreased the capitalisation negatively and significantly at the level of 0.1%. H11. Public ownership: Based on conventional banks’ empirical results, the involvement of government in banking operations results in instability and higher default risks. Barakat and Hussainey (2013) propose the same association. In Islamic banks, public ownership allowed for worse capitalisation. Based on the results of H9, H10 and H11 above, the banking sector in MENA region did suffer from the economic recession during the period 2005-2012. This could be due to Arab Spring (republic revelations) that occurred in 2011 in some MENA countries. Ghosh (2015) confirms that Arab Spring badly and negatively affected the banking sector in 12 MENA countries. H12. Gross domestic production: The findings state that GDP (economic) growth significantly decreases the stability of conventional and socially responsible banks. In other words, these banks cannot exploit the growth of the economy. This result is consistent with the findings of Cubillas and González (2014). In contrast, Nguyen et al. (2012) confirm that GDP development leads to more stable banks. Depending on capital ratio, the Islamic and socially responsible banks in countries with higher GDP were found to be well capitalised compared to countries with low GDP rates. H13. Inflation: the inflation rates strongly impact the stability of conventional banks in MENA region (similar to Köhler, 2015). In contrast, the inflation supported the stability of SRBs and allowed them to be further from bank-risk taking (in line with Nguyen & Nghiem, 2015). H14. Market capitalisation: Finally, market capitalisation was found to have a negative and significant relationship to stability in both the Islamic and conventional banking sectors. This contrasts with Nguyen et al. (2012), who note that development in the stock market makes banks steadier against losses. On the contrary, the capitals of Islamic and conventional banks were strongly and effectively increased in countries with greater stock market indices. To include a robust test, we can test the data through fixed effects model (FEM) as in Table 5 below. The main differences between the findings of OLS and FEM are: The relationship between the financial stability (z-score) and the capitalisation became strongly significant and positive (in line with Tabak et al., 2015) for Islamic, conventional and socially responsible banks compared to OLS coefficients which have insignificant correlation. The FEM approved that higher capitalised Islamic, conventional and socially responsible banks are more stable than lower capitalised banks. Horvàth et al. (2014) claims the same result in Czech Republic case study. For the macroeconomic factors, FEM confirms that during the global financial crisis (2007-2009), Islamic banks could increase their capitals efficiently. According to Hartmann et al. (2005), central bank tries to prevent systematic risk as higher risk results to have weaker financial products and services which leads to contraction in economy from the macroeconomics point of view. However, concentrating on microeconomics concept, banks strive to be stable through having better capitalisation and having less probability of bankruptcy risk. Well regulations imposed from central banks to banks yield to more stability in banking systems. Corporate Ownership & Control / Volume 14, Issue 2, Winter 2017, Continued - 1 220 Table 5. FEM results Islamic banks Conventional banks Socially responsible banks Stability Z-score Capital Ratio Z-score Capital Ratio Z-score Capital Ratio Bank-specific variables (H1) Z-score 0.163*** 0.0839*** 0.144*** (12.82) (8.21) (4.98) (H2) Capital ratio 2.492*** 0.691*** 2.429*** (12.82) (8.21) (4.98) (H3) Size -0.0980* -0.0216* -0.29*** -0.048*** 0.0897 0.0217 (-2.50) (-2.15) (-10.02) (-4.63) (0.85) (0.84) (H4) Loan intensity -0.129 -0.0744** -0.0828 0.260*** 0.494 0.00705 (-1.19) (-2.70) (-1.53) (15.16) (1.39) (0.08) (H5) Credit risk 0.00012 0.000192 0.00019 -0.00003 0.292 -0.174*** (0.23) (1.45) (0.32) (-0.19) (1.71) (-5.04) (H6) ROA 1.551*** -0.132* 2.258*** 1.822*** 3.610 -1.913*** (7.45) (-2.26) (6.72) (17.19) (1.65) (-4.07) (H7) Operating leverage 0.703 0.0131 3.000** -0.560 11.99*** -2.648*** (0.75) (0.05) (3.24) (-1.73) (4.04) (-3.55) (H8) Age (H9) Foreign ownership 0.0996 -0.171 0.00150 (0.74) (-1.68) (0.04) (H10) Domestic ownership 0.0556 0.0210 -0.109 -0.00740 (0.46) (0.95) (-1.43) (-0.28) (H11) Public ownership 0.0488 (1.43) Macroeconomic variables (H12) GDP 0.128 -0.0180 0.420*** 0.0306 0.137 0.00755 (1.86) (-1.02) (8.80) (1.78) (0.90) (0.20) (H13) Inflation -0.131 -0.00404 -0.0436 -0.0525 0.182 0.158 (-0.91) (-0.11) (-0.33) (-1.15) (0.24) (0.85) (H14) Market capitalisation 0.0570 -0.0284* 0.150*** -0.0332* 0.0377 0.0510 (1.09) (-2.14) (3.91) (-2.48) (0.17) (0.96) (H15) Global financial crisis -0.0216 0.0140* -0.00774 0.00546 -0.0102 0.00755 (-0.96) (2.44) (-0.46) (0.94) (0.76) Sigma -0.329 0.433 -5.20*** -0.599 -0.204 _cons (-0.20) (1.05) (-4.83) (-1.58) (-0.26) R2 0.6290 0.5693 0.3665 0.6109 0.6109 0.7633 Number of banks 40 40 168 168 8 8 Obs 291 291 1277 1277 64 64 * p < 0.05, ** p < 0.01, *** p < 0.001, t statistics in parentheses. 5. CONCLUSION This study aimed to find the determinants of financial stability in Islamic, conventional, and socially responsible banks in MENA region covering the period 2005-2012, using z-score and capital ratio. The socially responsible banks were found to be the most stable. Furthermore, Islamic banks were more risky and unstable. In addition, conventional banks had the minimum capital ratios. According to the determinants of stability in Islamic banks, the results conclude that z-score was strongly and positively affected by both ROA and age. According to capital ratio, the main determinants of capitalisation are operating leverage, GDP and market capitalisation. These variables highly support capitalisation in Islamic banks. In the conventional banking sector, the size of banks and profitability were found to be very important to their financial stability. Focusing on capitals of conventional banks, capital ratio had effective associations with lending, ROA and financial market development. In SRBs, banks achieved better financial stability in countries with higher inflation rates. However, the global financial crisis had insignificant relationships with z-scores and capitalisation in Islamic, conventional, and socially responsible banks. The greatest limitation to this study was the availability of data, which forced the researcher to reduce the sample of banks, particularly in Islamic and socially responsible banks. Furthermore, it is often difficult to contact banks with a view to collecting more data. An area for future research could be to cover more recent periods, and include the Arab Spring period which could potentially add more to the literature review. REFERENCES 1. Agoraki, M.-E.K, Delis, M.D., Pasiouras, F., 2011. Regulations, competition and bank risk-taking in transition countries. Journal of Financial Stability 7(1), 38-48. 2. Anginer, D., Demirguc-Kunt, A., Zhu, M., 2014. How does deposit insurance affect bank risk? Evidence from the recent crisis. Journal of Banking and Finance 48, 312-321. 3. Bankscope Database, 2016. Bureau Van DijK. Available at: https://bankscope.bvdinfo.com 4. Barakat, A., Hussainey, K., 2013. Bank governance, regulation, supervision, and risk reporting: Evidence from operational risk disclosures in European banks. International Review of Financial Analysis 30, 254-273. 5. Beck, T., Demirguc-Kunt, A., Merrouche, O., 2013. Islamic vs. conventional banking: Business model, efficiency and stability. Journal of Banking and Finance 37, 433-447. Corporate Ownership & Control / Volume 14, Issue 2, Winter 2017, Continued - 1 221 6. Berger, A.N., Klapper, L.F., Turk-Ariss, R., 2009. Bank Competition and Financial Stability. Journal of Financial Services Research 35, 99-118. 7. Bertay, A.C., Demirguc-Kunt, A., Huizinga, H., 2013. Do we need big banks? Evidence on performance, strategy and market discipline. Journal of Financial Intermediation 22 (4), 532-558. 8. Bourkhis, K., Nabi, M.S., 2013. Islamic and conventional banks' soundness during the 2007-2008 financial crisis. Review of Financial Economics 22, 68-77. 9. Boyd, J., Graham, S., Hewitt, R., 1993. Bank holding company mergers with nonbank financial firms: Effects on the risk of failure . Journal of Banking and Finance 17, 43-63. 10. Chalermchatvichien, P., Jumreornvong, S., Jiraporn, P., 2014. Basel III, capital stability, risk-taking, ownership: Evidence from Asia. Journal of Multinational Financial Management 28, 28-46. 11. Cubillas, E., Gonzalez, F., 2014. Financial liberalization and bank risk-taking: International evidence, Journal of Financial Stability 11, 32-8. 12. Dedu, V., Chitan, G., 2013. The influence of internal corporate governance on bank performance - an empirical analysis for Romania. Procedia - Social and Behavioral Sciences 99, 1114-1123. 13. Dima, B., Dinca, M.S., Spulbăr, C., 2014. Financial nexus: Efficiency and soundness in banking and capital markets. Journal of International Money and Finance 47, 100-124. 14. Dong, Y., Meng, C., Firth, M., Hou, W., 2014. Ownership structure and risk-taking: Comparative evidence from private and state-controlled banks in China. International Review of Financial Analysis 36, 120-130. 15. Dubai Islamic Bank, 2016. Dubai, UAE. Available at: http://www.dib.ae 16. ElBannan, M.A., 2015. Do consolidation and foreign ownership affect bank risk taking in an emerging economy? An empirical investigation. Managerial Finance 41(9), 874-907. 17. Hartmann, P., Straetmans, S., Vries. C.D., 2005. Banking System Stability. Working Paper Series No. 527. European Central Bank. 18. Fu, X., Lin, Y., Molyneux, P., 2014. Bank competition and financial stability in Asia Pacific. Journal of Banking and Finance 38, 64-77. 19. García-Kuhnert, Y., Marchica, M.-T., Mura, R., 2015. Shareholder diversification and bank risk-taking. Journal of Financial Intermediation, 1-34. 20. Ghosh, S., 2014. Risk, capital and financial crisis: Evidence for GCC banks. Borsa Istanbul Review 14(3), 145-157. 21. Ghosh, S., 2015. Political transition and bank performance: How important was the Arab Spring? Journal of Comparative Economics. 22. Gutiérrez-Nieto, B., Serrano-Cinca, C., Camón-Cala, C., 2016. A Credit Score System for Socially Responsible Lending. Journal of Business Ethics 133(4), 691-701. 23. Horvàth, R., Seidler, J., Weill, L., 2014. Bank Capital and Liquidity Creation: Granger-Causality Evidence. Journal of Financial Services Research 45(3), 341-361. 24. International Monetary Fund, 2016. Available at: http://www.imf.org 25. Islamic Development Bank, 2016. Jeddah, Saudi Arabia. Available at: http://www.isdb.org 26. Jeon, J.Q., Lim, K.K., 2013. Bank competition and financial stability: A comparison of commercial banks and mutual savings banks in Korea. Pacific-Basin Finance Journal 25, 253-272. 27. Kansal, M., Joshi, M., & Batra, G.S. (2014). Determinants of corporate social responsibility disclosures: Evidence from India. Advances in Accounting, 30, 217-229. 28. Kasman, S., Kasman, A., 2015. Bank competition, concentration and financial stability in the Turkish banking industry. Economic Systems 39, 502-517. 29. Köhler, M., 2015. Which banks are more risky? The impact of business models on bank stability. Journal of Financial Stability 16, 195-212. 30. Lee, T., Chih, S., 2013. Does financial regulation affect the profit efficiency and risk of banks? Evidence from China’s commercial banks. North American Journal of Economics and Finance 26, 705-724. 31. Mirzaei, A., Moore, T., Liu, G., 2013. Does market structure matter on banks’ profitability and stability? Emerging vs. advanced economies. Journal of Banking and Finance 37, 2920-2937. 32. Nguyen, T.P., Nghiem, S.H., 2015. The interrelationships among default risk, capital ratio and efficiency: Evidence from Indian banks. Managerial Finance 41(5), 507-525. 33. Organization of the Petroleum Exporting Countries, 2016. Vienna, Austria. Available at: http://www.opec.org 34. Rajhi, W., Hassairi, S.A., 2013. Islamic banks and financial stability: a comparative empirical analysis between MENA and Southeast Asia countries. Region et Development 37, 149-177. 35. Rumler, F., Waschiczek, W., 2014. Have changes in the financial structure affected bank profitability? Evidence for Austria. The European Journal of Finance 1-22. 36. Schaeck, K., Cihàk, M., 2014. Competition, Efficiency, and Stability in Banking. Financial Management 43 (1), 215-241. 37. Soedarmono, W., Machrouh, F., Tarazi, A., 2011. Bank market power, economic growth and financial stability: Evidence from Asian banks. Journal of Asian Economics, 22(6), 460-470. 38. Srairi, S., 2013. Ownership structure and risk-taking behaviour in conventional and Islamic banks: Evidence for MENA countries. Borsa Istanbul Review 13, 115-127. 39. Tabak, B.M., Fazio, D.M., Cajueiro, D.O., 2013. Systemically important banks and financial stability: The case of Latin America. Journal of Banking & Finance 37, 3855-3866. 40. Tabak, B.M., Gomes, G.M.R., Medeiros, M.D.S., 2015. The impact of market power at bank level in risk-taking: The Brazilian case. International Review of Financial Analysis 40, 154-165. 41. Tan, Y., Floros, C., 2013. Risk, capital and efficiency in Chinese banking. Journal of International Financial Markets, Institutions and Money 26, 378-393. 42. Wang, Z., Chen, J., Wan, Y., Jin, Y., Mazzanti, J.A., 2015. Information Disclosure and Bank Risk-Taking under a Partially Implicit Deposit Insurance System: Evidence from China. The Australian Economic Review 48(2), 163-176. 43. Williams, B., 2014. Bank risk and national governance in Asia. Journal of Banking and Finance 49, 10-26. 44. World Bank, 2016. Washington, DC. Available at: http://www.worldbank.org 45. Zhang, Z., Xie, L., Lu, X., Zhang, Z., 2015. Determinants of Financial Distress in Large Financial Institutions: Evidence from U.S. Bank Holding Companies. Contemporary Economic Policy, 1-18. | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/357964499From Act to Action: (Meta)pragmatic instruction with young EFL learners inNorwayThesis · December 2021CITATIONS0READS1091 author:Some of the authors of this publication are also working on these related projects:Pragmatic development in Norwegian primary school EFL learners View projectFrom Act to Action: (Meta)pragmatic instruction with young EFL learners in Norway View projectAnders MyrsetUniversity of Stavanger (UiS)11 PUBLICATIONS 57 CITATIONS SEE PROFILEAll content following this page was uploaded by Anders Myrset on 28 January 2022.The user has requested enhancement of the downloaded file. From Act to Action: (Meta)pragmatic instruction with young EFL learners in Norway by Anders Otterbech Jølbo Myrset Thesis submitted in fulfilment of the requirements for the degree of PHILOSOPHIAE DOCTOR (PhD) Faculty of Arts and Education Department of Education and Sports Science 2021 University of Stavanger NO-4036 Stavanger NORWAY www.uis.no ©2021 Anders Myrset ISBN: 978-82-8439-051-2 ISSN: 1890-1387 PhD: Thesis UiS No. 619 iii Acknowledgements They say it takes a village to raise a child. Well, if there was ever a time to say that this is true, I am convinced this is it – writing a PhD. The only thing is that my village is huge, going across borders. Thanking each and everyone would result in a long piece of prose, which I do not have enough space for. Thus, below is a list of people, but it is in no way complete. If you are not mentioned here and feel like you should have been I am really sorry. First and foremost, I have to thank my supervisors and friends, Dr Milica Savić and Professor Emeritus Ion Drew. I am afraid words will never fully describe just how grateful I am for everything you have done for me. Without the two of you I would have never even considered embarking on this journey. You planted the first seeds that turned into the idea of a) doing a PhD and b) exploring pragmatics. Furthermore, you have been there with your guidance and support through all the highs and the lows. It has been three long years, but the two of you have made it a fun adventure and for that I am eternally grateful. In addition, you have kept supporting my development, from a mere toddler taking his first steps into the runner I am today. I would not be where I am without you and this thesis is a testament to that: all the clear thoughts presented on these pages are in here because of you, any ‘leftover’ is my own doing – my zone of proximal development could not be stretched any further at this point (but how far it has been stretched!). For all the support, engaging discussions, countless coffees, numerous walks, the collaboration (written works and presentations) – the list is endless, but I appreciate everything and look forward to more of the same in the future. I also want to express my sincere thanks to everyone at the school where I conducted my research. I wish I could mention each and every one of you by name, but alas, it is not possible due to ethics. I trust you know who you are. To the learners, thank you for enthusiastically participating in the research, and for sharing all your wonderful and interesting reflections. I can only hope that you have learnt as much as I have from this project. To the teachers, thank you for being so understanding and engaged in my research, and for allowing me to come in and steal from your precious time with your learners. I am especially grateful to the English teacher, who went above and beyond during my fieldwork. I am also grateful to Dr Annamaria Pinter and Professor Maria Economidou-Kogetsidis for their invaluable feedback during my 50% seminar (half-viva). Thank you both for generously sharing your knowledge about research with children and pragmatics. The feedback you gave me then has been invaluable during the continued work on my thesis iv and has helped me present my study more clearly. I am also thankful for the work we have done together since my 50%. My sincere gratitude to Dr Troy McConachy, who facilitated my stay abroad at the University of Warwick. I appreciate all the engaging conversations we had during my stay, all the feedback, and that you let me come in to share my research with your students. Furthermore, I would like to thank you for the collaboration and support following my visit. I look forward to our continued work in the future. I also want to thank Professor Emeritus Andrew Cohen for sharing his insights about pragmatics research and providing feedback on the design of my study at an early stage of my PhD; Dr Miodrag Đorđević for helping me with the statistics; Dušan Cvetković for drawing the cartoons I used during the instruction; and Mia Eriksen for translating my data. To the PhD community in educational sciences at the University of Stavanger: thank you. I especially want to thank my friends, James, Cecilie, and Anastasia. You were already experienced PhD students when I started, so thank you for sharing your insights in abundance. I am also grateful to you for reading and giving me feedback on various drafts of my manuscripts. I am also grateful to my family, especially my parents for their never-ending support and for always letting me find my own path; my brothers, Ola and Eirik, for always being understanding and for always being in my corner; and my cousin, Ingri, for giving me advice when I applied for the PhD and for helping me out with my written Norwegian. I would like to thank all my friends for being so understanding whilst I have lived in this bubble. I appreciate that you are still around now that I am ready to come back out into the world. I especially want to thank Marius for being there since our very first steps into the education system in primary school; Jarle for being so enthusiastic and supportive about my work; Sam for always reminding me of the beauties of writing; and Kjersti and Shalini for being part of bringing the sunshine that has helped me stay (somewhat) sane through all of this. Finally, to the pioneer who first attempted to roast, grind, and boil the beans forever known as coffee, I thank you and salute you. Stavanger, 2021 v Summary This doctoral research project is a case study of the impact of teaching English pragmatics to Norwegian primary school learners in 7th grade (aged 12-13). The importance and the impact of teaching second/foreign language (L2) pragmatics have been much discussed in both empirical and theoretical work, shifting the focus from whether pragmatics is teachable to the affordances of various teaching approaches. However, the evidence is largely based on (young) adult learners, with young language learners (YLLs) comprising an underexplored group. Similarly, YLLs’ development of pragmatic ability, i.e. ability to produce and interpret language in context, and metapragmatic awareness, i.e. reflections about language use, remain largely uncharted waters. Hence, the discussions about how L2 pragmatics can be taught and researched are largely informed by research with older language learners. This forms the backdrop for the present doctoral study, which specifically investigates the impact of teaching L2 requests to the target group. The impact of instruction is explored through the learners’ request production, their use of scientific concepts to express metapragmatic understandings, and their engagement with the project. Informed by sociocultural theory (SCT), the instruction adopted a concept-based approach to teaching L2 pragmatics to two intact classes in a Norwegian primary school. The overarching aim of the instruction was to foster agency, that is to promote the learners’ ability to make informed choices in communication. In addition, the study was influenced by the growing body of literature on research with children, which aims to enable them to express their views and be listened to, that is, to give them a voice. Informed by this view, the current study included a focus on giving learners a voice through the use of innovative data elicitation techniques. This thesis is the synopsis of an article-based Ph.D. which comprises four articles (I-IV). Article I presents a systematic review, investigating the data elicitation techniques used in prior research exploring YLLs’ metapragmatic awareness, i.e. their verbalised reflections about language use, contextual considerations, and/or their interplay. The review revealed that previous research was sparse and that the elicitation techniques employed largely mirrored those used with adults. In light of these findings and informed by literature on research with children, the article presents three elicitation techniques, developed and used by the authors in research projects with learners aged 9-13, with aims to scrutinise their affordances. Article II investigates the impact of the instruction on the learners’ request production strategies. The data was collected through a video-prompted oral discourse completion vi task (VODCT), which was administered in a pre-post-delayed design, enabling the researcher to investigate both short- and long-term changes in strategy use following the instruction. These changes were measured through statistical tests. The study revealed significant longer-term retention of some request strategies, e.g. internal modification through modal verbs, whilst others revealed no significant changes. Article III explores the learners’ use of scientific concepts to express their metapragmatic understandings. The analysis was conducted through a framework aiming to identify metapragmatic episodes and subsequently three excerpts were analysed in-depth to explore how the learners used in discussions the scientific concepts introduced during the instruction. The study revealed that, although used relatively infrequently in the dataset as a whole, scientific concepts were used to discuss the importance of linguistic variation, the communicative value of hints, and to compare strategies in the first language (L1) and the L2. Thus, the study reveals a potential for teaching pragmatics through concept-based approaches. Finally, Article IV investigates how the learners appraised various components of the project, including the different data elicitation techniques, and how they explained their appraisals. The study revealed that the target of instruction (requests) presented a novel topic, which the learners found engaging and relevant. In addition, the learners were positive to their perceived learning outcomes and to the focus on choices related to requests of which they became aware. The study provides valuable insights into YLLs’ engagement in pragmatics research and the importance of giving them a voice in projects of this kind. First and foremost, the thesis contributes to our limited understanding of whether and how pragmatics can be taught with YLLs, both generally and within SCT-informed instructional pragmatics research. From the perspective of SCT-informed instruction, the instructional approach employed presents a novel focus: whilst prior research has employed concept-based approaches for teaching L2 pragmatics with adults, the present study is, to the best of the author’s knowledge, the only one of its kind to investigate the affordances of such approaches with YLLs. The study shows that an explicit focus on pragmatics is indeed feasible with YLLs and that the focus of instruction and the teaching approaches resonated with the learners (Articles II, III, and IV). In addition, since YLLs’ voices have largely been under-communicated within the field of instructional pragmatics, this thesis contributes to addressing this gap (Articles I and IV). The thesis contributes to our understanding of the affordances of explicit instruction with YLLs through concept-based approaches, both from the perspective of teaching practice and research, and adds to the knowledge about participant-friendly vii methodologies aiming to promote, and ultimately act upon, children’s perspectives in pragmatics research. viii Sammendrag Dette doktorgradsprosjektet er en kasusstudie som undersøker påvirkningen av undervisning i engelsk pragmatikk til norske grunnskoleelever i syvende trinn (12-13 år). Betydningen og effekten av å undervise i andre-/fremmedspråkspragmatikk (S2-pragmatikk) har vært mye diskutert, både i empiri og teori, og gått fra et fokus på om pragmatikk kan læres til å fokusere på mulighetene og utfordringene ved ulike undervisningsmetoder. Empirien som denne forskningen tar utgangspunkt i, er imidlertid stort sett basert på studier av studenter, altså unge voksne. Språkinnlærere i grunnskolealder representerer dermed en gruppe det er gjort lite forskning på i denne sammenhengen. Derfor vet vi lite om unge språkelevers utvikling av pragmatiske evner (hvordan de uttrykker og tolker språk i kontekst), og deres metapragmatiske bevissthet (deres refleksjoner om språkbruk). Som en følge av dette, tar diskusjonene knyttet til undervisning og forskning på S2-pragmatikk utgangspunkt i forskning på eldre språkelever. Med bakgrunn i dette er formålet med denne studien å undersøke hvordan målrettet undervisning av engelske anmodninger påvirker språkinnlærerenes språkbruk. Undervisningens påvirkning utforskes gjennom å undersøke elevenes produksjon av anmodninger, deres bruk av vitenskapelige begreper for å uttrykke metapragmatisk bevissthet, og deres engasjement i prosjektet. Undervisningen tok utgangspunkt i et sosiokulturelt læringssyn (SCT), og tok i bruk en begrepsbasert tilnærming til undervisning av S2-pragmatikk til to klasser i en norsk barneskole. Undervisningens overordnede mål var å fremme handlingsfrihet (agency), det vil si å fremme elevenes evne til å ta informerte valg i kommunikasjon. I tillegg tok studien utgangspunkt i forskningslitteratur som tematiserer hvordan en kan gjøre barn i stand til å uttrykke sine synspunkter og bli lyttet til – altså hjelpe dem til å utvikle en stemme (voice). Med utgangspunkt i dette synet har fokuset i denne studien vært å gi elevene verktøy til å uttrykke sin egen stemme gjennom bruken av innovative datainnsamlingsteknikker. Denne avhandlingen er en artikkelbasert ph.d. som består av fire artikler (I-IV). Artikkel I presenterer en systematisk review, som undersøker datainnsamlingsteknikkene brukt i tidligere forskning på unge språkelevers metapragmatiske bevissthet, det vil si deres verbaliserte refleksjoner om språkbruk, kontekstuelle hensyn og/eller deres samspill. Reviewen viste at det var lite tidligere forskning og at innsamlingsteknikkene i stor grad gjenspeilet de som har blitt brukt med voksne. I lys av disse funnene og basert på litteratur om forskning med barn, presenterer artikkelen tre datainnsamlingsteknikker. Teknikkene er utviklet og brukt av ix artikkelforfatterne i to forskningsprosjekter med elever i alderen 9-13 år, og tar sikte på å belyse mulighetene og utfordringene ved disse teknikkene. Artikkel II undersøker undervisningens påvirkning på elevenes anmodningsstrategier. Dataene ble samlet inn gjennom en test der en undersøker elevenes anmodningsstrategier, på engelsk kalt video-prompted oral discourse completion task (VODCT), som ble gjennomført i en pre-, post-, og forsinket post-test, slik at forskeren kunne undersøke både kortsiktige og langsiktige endringer i strategibruk etter undervisningen. Disse endringene ble målt gjennom statistiske tester. Studien avdekket signifikante langsiktige endringer i bruk av noen anmodningsstrategier, for eksempel intern nedtoning gjennom modale verb, mens andre strategier ikke avdekket noen signifikante endringer. Artikkel III utforsker elevenes bruk av vitenskapelige begreper for å uttrykke sine metapragmatiske forståelser. Analysen ble utført gjennom et rammeverk som tok sikte på å identifisere metapragmatiske hendelser. Deretter ble tre utdrag analysert i dybden for å utforske hvordan elevene brukte de vitenskapelige begrepene som ble introdusert i løpet av undervisningen, i diskusjoner. Studien viste at selv om vitenskapelige begreper ble brukt relativt sjelden i datasettet som helhet, ble vitenskapelige begreper brukt til å diskutere betydningen av språklig variasjon, den kommunikative verdien av hint, og for å sammenligne strategier på førstespråket (S1) og S2. Studien avdekker dermed et potensial for å undervise i pragmatikk gjennom begrepsbaserte tilnærminger. Artikkel IV undersøker hvordan elevene evaluerte ulike deler i prosjektet, inkludert datainnsamlingsteknikkene, samt hvordan de utdypet disse evalueringene i diskusjon. Studien viste at undervisningsmålet (opplæring i anmodninger) var et nytt tema som elevene fant engasjerende og relevant. I tillegg var elevene positive til det de vurderte som eget læringsutbytte og til fokuset på bevisstgjøringen av ulike valg knyttet til anmodninger. Studien gir verdifull innsikt i unge språkelevers deltakelse i pragmatikforskning, og viktigheten av å gi dem en stemme i slike prosjekter. Først og fremst bidrar avhandlingen til vår begrensede forståelse av hvordan pragmatikk kan undervises til unge språkelever, både generelt og innenfor forkning på SCT-informert pragmatikkundervisning. Fra et SCT-basert undervisningsperspektiv presenterer tilnærmingen i dette prosjektet et nytt fokus: Mens tidligere forskning har benyttet begrepsbaserte tilnærminger for å undersøke undervisning av S2-pragmatikk for voksne, undersøker den nåværende studien mulighetene og utfordringene med slike tilnærminger hos unge språkelever. Studien viser at et eksplisitt fokus på pragmatikk faktisk er gjennomførbart med unge språkelever, og at undervisningsfokuset og undervisningsmetodene resonerte med elevene (artikkel II, III og IV). I tillegg, siden unge språkelevers stemmer i stor grad har vært underkommunisert innen forskning på x pragmatikkundervisning, bidrar denne avhandlingen til å adressere dette gapet (artikkel I og IV). Avhandlingen bidrar til vår forståelse av mulighetene og utfordringene med eksplisitt undervisning med unge språkelever gjennom begrepsbaserte tilnærminger, både fra et undervisnings- og et forskningsperspektiv, og bidrar til kunnskapen om deltakervennlige metoder som tar sikte på å fremme, og til slutt handle ut fra, barns perspektiver i pragmatikkforskning. xi Table of Contents Acknowledgements .......................................................................................... iii Summary ........................................................................................................... v Sammendrag .................................................................................................. viii 1 Introduction ............................................................................................... 1 1.1 The present study ...................................................................................... 2 1.1.1 A focus on requests ................................................................................ 2 1.1.2 A sociocultural approach to teaching pragmatics ................................... 3 1.1.3 Focus on young language learners (YLLs) ............................................. 4 1.1.4 The curriculum and English teaching in Norway ................................... 4 1.1.5 An overview of the study ....................................................................... 6 1.2 Contributions of the study ......................................................................... 8 1.3 Structure of the synopsis ........................................................................... 9 2 Theory and previous research ................................................................. 11 2.1 Requests .................................................................................................. 11 2.1.1 Request strategies ................................................................................. 12 2.2 Sociocultural theory ................................................................................ 13 2.2.1 Learner development ............................................................................ 15 2.2.2 Spontaneous and scientific concepts .................................................... 16 2.2.3 Teaching pragmatics through concepts ................................................ 19 2.2.4 Research on concept-based pragmatics instruction .............................. 23 2.3 Previous research with young language learners .................................... 27 2.3.1 Research on young language learners’ metapragmatic awareness........ 27 2.3.2 Pragmatics instruction with young learners .......................................... 37 2.4 Concluding remarks ................................................................................ 42 3 Methodology ........................................................................................... 45 3.1 Overview of research design ................................................................... 45 3.1.1 Case study ............................................................................................ 46 3.1.2 Mixed methods approach ..................................................................... 49 3.1.3 Project coherence ................................................................................. 49 3.1.4 Sample .................................................................................................. 51 3.2 Instruction ............................................................................................... 52 3.2.1 Pedagogical approach and classroom procedures ................................. 53 xii 3.2.2 The pragmalinguistic dimension .......................................................... 54 3.2.3 The sociopragmatic dimension ............................................................. 59 3.3 Data collection ........................................................................................ 64 3.3.1 Request production data ....................................................................... 64 3.3.2 Readers Theatre .................................................................................... 71 3.3.3 Interview data ....................................................................................... 75 3.3.4 Overview of the data collection and analysis ....................................... 82 3.4 Scientific quality: Quality criteria ........................................................... 83 3.5 Reflexivity .............................................................................................. 89 3.6 Ethical considerations ............................................................................. 91 4 Summary of articles ................................................................................ 95 4.1 Article I – “If an astronaut were on the moon…”: Eliciting metapragmatic data from young L2 learners ........................................... 95 4.2 Article II – 'You could win Masterchef with this soup. Can I get some more?' Request production and the impact of instruction on young EFL learners. ................................................................................................... 99 4.3 Article III – Scientific concepts as meaning-making resources for young EFL learners in the learning of pragmatics ........................................... 101 4.4 Article IV – Giving young language learners a voice: learner feedback on pragmatics instruction ........................................................................... 103 5 Discussion and conclusion .................................................................... 107 5.1 Overall findings .................................................................................... 107 5.2 Limitations ............................................................................................ 112 5.3 Implications for teaching L2 pragmatics ............................................... 114 5.4 Contributions......................................................................................... 116 5.5 Conclusion ............................................................................................ 117 6 References ............................................................................................. 121 Appendices ................................................................................................... 141 Appendix 1 – Worksheet directness ................................................................. 141 Appendix 2 – Homework (H1) ......................................................................... 143 Appendix 3 – Worksheet, supportive moves .................................................... 144 Appendix 4 – Requests produced by learners .................................................. 145 Appendix 5 – Request perception journey ....................................................... 148 xiii Appendix 6 – Match request and interlocutor .................................................. 150 Appendix 7 – Dice game .................................................................................. 151 Appendix 8 – Homework (H2) ......................................................................... 152 Appendix 9 – Labels Dice game ...................................................................... 153 Appendix 10 – Script, worksheet ..................................................................... 154 Appendix 11 – Example of a pre-written script (RT Cycle 1) ......................... 156 Appendix 12 – An example of a folder for script-writing (RT Cycle 2) .......... 159 Appendix 13 – Questions for the VODCT ....................................................... 163 Appendix 14 – Interview guide ........................................................................ 166 Appendix 15 – Consent form and NSD approval ............................................. 169 Articles .......................................................................................................... 174 Article I - “If an astronaut were on the moon…”: Eliciting metapragmatic data from young L2 learners ......................................................................... 175 Article II - 'You could win Masterchef with this soup. Can I get some more?' Request production and the impact of instruction on young EFL learners .............................................................................................................. 211 Article III - Scientific concepts as meaning-making resources for young EFL learners in the learning of pragmatics ................................................... 241 Article IV - Giving young language learners a voice: learner feedback on pragmatics instruction ........................................................................... 265 List of Figures Figure 1.1: Overview of the fieldwork with the techniques used and the data they elicited. The “A” followed by a roman numeral (e.g. A-I) refers to the article in which the data was presented. ................ 6 Figure 2.1: A visual representation of a hierarchy of scientific concepts relating to directness, with sub-concepts for directness levels, as well as examples of pragmalinguistic resources within each sub-concept (also presented in Article III (Myrset, 2021))...... 18 Figure 2.2 Systematic review process (also presented in Article I) ................ 29 Figure 3.1: Overview of the fieldwork with the techniques used and the data they elicited. The “A” followed by a roman numeral (e.g. A-I) refers to the article in which the data was presented. .............. 48 xiv Figure 3.2: Coherence of project .................................................................... 50 Figure 3.3: Slides for introducing the context ................................................ 59 Figure 3.4: A screenshot from one of the videos used in the project as a visual prompt (https://youtu.be/P5Vi4j1F4sE). ................................. 66 Figure 3.5: A request appraisal sheet (also presented in Article III (Myrset, 2021)) ....................................................................................... 77 Figure 3.6: Appraisal sheet for the project components (also presented in Article IV (Myrset, In preparation)) ........................................ 77 Figure 3.7 Categories occurring in the discussions about the instruction ....... 81 List of Tables Table 1.1 The main research question of the study, and the titles, aims or research questions of the four articles........................................ 7 Table 2.1: Pragmatics instruction using concept-based approaches ............... 24 Table 2.2: Previous research on YLLs’ metapragmatic awareness ................ 30 Table 2.3: Pragmatics instruction studies with young language learners ....... 39 Table 3.1: Scientific concepts employed during the instruction (also presented in Article III (Myrset, 2021)) ................................................... 54 Table 3.2: Overview of the sessions in the first two weeks focusing on the pragmalinguistic dimension. .................................................... 56 Table 3.3: Overview of the sessions in the last two weeks focusing on the sociopragmatic dimension. ...................................................... 61 Table 3.4: Videos used in the VODCT (videos retrieved: 16.08.19). ............. 67 Table 3.5: Coding manual for requests ........................................................... 70 Table 3.6 Coding framework, adapted from Fortune and Thorp (2001) (adapted from Article III (Myrset, 2021)). .............................. 79 Table 3.7: The research questions, data sources and data analysis approaches in the four articles .................................................................... 82 Table 3.8: Quality criteria for the present study ............................................. 84 Table 4.1: Articles of the doctoral research project ........................................ 95 xv List of articles Article I: Myrset, A., & Savić, M. (2021). “If an astronaut were on the moon…”: Eliciting metapragmatic data from young L2 learners. Applied Pragmatics, 3(2), 163-196. Article II: Myrset, A. (Pending revisions). 'You could win Masterchef with this soup. Can I get some more?' Request production and the impact of instruction on young EFL learners. Journal of Pragmatics. Article III: Myrset, A. (2021). Scientific concepts as meaning-making resources for young EFL learners in the learning of pragmatics. Intercultural Communication Education, 4(2), 191–212. https://doi.org/10.29140/ice.v4n2.485 Article IV: Myrset, A. (In preparation). Giving young language learners a voice: learner feedback on pragmatics instruction. [To be submitted to TESL-EJ] List of abbreviations DCT – Discourse completion task EFL – English as a foreign language L1 – First language L2 – Second/foreign/additional language LK06 – Norwegian national curriculum 2006-2020 LK20 – Norwegian national curriculum 2020- RT – Readers Theatre SCT – Sociocultural theory VODCT – Video-prompted oral discourse completion task YLL – Young Language Learner xvi Introduction 1 1 Introduction This doctoral research project is a case study exploring the impact of teaching English pragmatics on Norwegian primary school learners in 7th grade (aged 12-13). More specifically, informed by sociocultural theory (SCT), the instruction focused on the teaching of requests in an English as a foreign language (EFL) context1. The study aimed to explore the learners’ language production and understandings of language use in connection with requests following the instruction, as well as their engagement with the project. The data the study is based on is presented in four articles attached at the end of this synopsis. In essence, pragmatics is "the study of language from the perspective of users, especially of the choices they make, the constraints they encounter in using language in social interaction and the effects their use of language has on other participants in the act of communication" (Crystal, 1997, p. 301). In other words, pragmatics deals with how language is performed and interpreted within the context in which it occurs. It is thus of vital importance in communication, particularly when interacting in a foreign, second, or additional language2 (L2). For this reason, and in the wake of early calls for more research on the teachability of L2 pragmatics (Kasper, 1997), there has been much theoretical discussion and empirical research aiming to answer two main questions: Firstly, can pragmatics be taught successfully to L2 learners? Secondly, is there a need for L2 pragmatics instruction? Today, the consensus is yes on both counts, and the field of L2 pragmatics has since moved to exploring the affordances of different teaching approaches. These have largely been informed by three broader language ideologies, namely the interlanguage pragmatics (often referred to as ILP) paradigm, 1 Whereas Norwegian learners have a generally high proficiency in English (Education First, 2020), the role of English as a foreign versus a second language has been debated in Norway (Rindal, 2014; Vattøy, 2017). For instance, Rindal (2014) argues that whilst English has traditionally been treated as a foreign language in Norwegian education, English plays a significant role in work and higher education. In addition, Norwegians experience substantial exposure to English through media (audio and visual) and travel (Rindal, 2014). This is also evident in policy, where Norwegian (first language, L1) and English (L2) are the only two languages which are taught as compulsory subjects from 1st grade, and with their own curricula (Udir 2020b), with English being referred to as the first foreign language in policy reports (Norwegian Ministry of Education and Research, 2003). However, the role of English as a second or foreign language remains somewhat opaque (Rindal, 2014; Vattøy, 2017), which is further emphasised in the distinction between English and foreign languages in the curriculum. However, English does not have a role as an official language. Thus, the learners in this thesis are viewed as learners of English as a foreign language (EFL). 2 Given the increasingly multilingual nature of language classrooms, English as an additional language (EAL) has been used by some authors (e.g. Lorenz et al., 2021; Portolés & Martí, 2017) rather than the labels ‘foreign’ or ‘second language’. Introduction 2 sociocultural theory, and intercultural pragmatics (McConachy & Spencer-Oatey, 2020). Of these, ILP, which is grounded in cognitive theories of L2 acquisition (Ohta, 2005) is the most widely researched – and by extension influential – perspective, whilst the latter two have received increased traction in recent years. Nevertheless, despite a growing interest in the teaching of L2 pragmatics from various language perspectives, young language learners (YLLs), here defined as those aged approximately 5-13 (Drew & Hasselgreen, 2008), remains a largely overlooked group. It is uncertain whether this has to do with a view that pragmatics is considered too advanced for these learners, that pragmatic ability is deemed less important for them, or simply that access to adult participants is more easily attainable (e.g. students in university settings). What is clear, however, is that despite calls for introducing pragmatics at an early stage of language teaching and learning (e.g. Bardovi–Harlig & Mahan-Taylor, 2003; Ishihara, 2013), the majority of studies explore the affordances of pragmatics instruction as they relate to adult learners (Plonsky & Zhuang, 2019; Taguchi, 2015). Thus, there is a gap of knowledge when it comes to pragmatics instruction with YLLs, both within the Norwegian and an international context, which the present study aims to address. 1.1 The present study 1.1.1 A focus on requests The present study focuses on the teaching of requests. Requests are “attempts by the speaker to get the hearer to do something. They may be very modest attempts as when I ask you to do it, or they may be very fierce attempts as when I insist that you do it” (Searle, 1979, p. 13). Originating from theoretical discussions in language philosophy and Speech Act Theory as ‘directives’ (e.g. Austin, 1962; Searle, 1965; Searle, 1979), requests, and other speech acts, have since become empirically described (Cohen, 1996). Within empirical research, one of the seminal works is Blum-Kulka et al.’s (1989) Cross-Cultural Pragmatics: Requests and Apologies, which presented a large-scale study comparing requests and apologies in seven countries, focusing on both L1 and L2 speakers. Importantly, in order to explore cross-cultural and intralinguistic variation, Blum-Kulka et al. (1989) provided a coding manual, which has since been extensively used in research. This coding manual served as a framework in the instruction and analysis in the present study. From the perspective of the L2, request production has been widely researched, with (young) adults (e.g. Awedyk, 2003; Infantidou & Tzanne, 2012; Krulatz, 2016) and YLLs (e.g. Achiba, 2003; Ellis, 1992; Portolés & Safont, 2018; Savić, 2015; Savić et Introduction 3 al., 2021)3. However, less evidence is provided in relation to teaching requests to YLLs, despite the fact that requests are frequently used in communication (Stavans & Shafran, 2018), and the development of request production and comprehension begins at an early age (Cekaite, 2013; Portolés, 2015). Thus, considering the sparse research on teaching requests to YLLs, and their frequent use and early development, they were considered an appropriate focus (pragmatic target) of the instruction. 1.1.2 A sociocultural approach to teaching pragmatics The instruction was informed by sociocultural theory and adopted a concept-based approach (e.g. van Compernolle, 2014). Within this approach, pragmatics is seen as mediated action. What this means is that rather than successful pragmatic performance being viewed as “adherence to social conventions” (van Compernolle, 2014, p. 42), the focus is on making informed choices in communication. These choices are informed by two dimensions: pragmalinguistics, that is, the link between pragmatics and grammar, and sociopragmatics, the link between pragmatics and culture (Leech, 1983; Thomas, 1983; van Compernolle, 2014). Thus, mediated action involves taking into consideration, for instance, the context and interpersonal aspects (sociopragmatics) in order to make informed pragmalinguistic choices, which results in accomplishing a goal in communication (e.g. requesting, apologising) (van Compernolle, 2014). Following this view, language teaching and learning aims to foster agency, i.e. “the socio-culturally mediated capacity to act and to assign meaning to one’s actions” (van Compernolle, 2014, p. 21), rather than teaching and learning rules of thumb, e.g. generalised prescriptions about language norms, politeness, and appropriateness in given contexts (Liddicoat & McConachy, 2019; van Compernolle, 2014). In order to foster agency, concept-based approaches aim to introduce scientific concepts, with a view that these foster a deeper, conceptual understanding of language use (Nicholas, 2015; Vygotsky, 1934/2012). Furthermore, in addition to conceptual development, metapragmatic awareness, that is, learners’ own understandings and reflections about pragmatic phenomena, such as politeness, is viewed as serving a vital mediating role for agency (e.g. Morollón Martí, Forthcoming). In the present study, metapragmatic awareness is viewed as being displayed through verbalised reflections about language use, contextual considerations, or their interplay, to varying degrees of sophistication. With the SCT perspective on pragmatics in mind, teaching requests involves raising awareness of the multitude of pragmalinguistic strategies. Drawing on Blum-Kulka et al. (1989), there is a range of request strategies. However, apart from the minimal unit 3 Portolés and Safont (2018) explored requests in three languages, that is, Spanish (majority language), Catalan (minority language), and English (foreign language). In their study, English is referred to as the third language (L3). Introduction 4 that is necessary to realise the request itself (referred to as the head act), all these strategies are non-essential. In other words, they can be included to modify the head act, but are not necessary. In the present study, increasing the learners’ pragmalinguistic repertoire was considered a prerequisite for fostering agency. This repertoire, mediated by the learners’ conceptual understandings and metapragmatic awareness, would provide tools for pragmalinguistic variation in communication, where meaning is dynamically negotiated (Spencer-Oatey, 2008). Therefore, teaching pragmatics with a view to developing agency was considered a useful approach for this study, thus exploring the affordances of such approaches with a previously uncharted group, namely YLLs. 1.1.3 Focus on young language learners (YLLs) Because of its focus on YLLs, this study was largely informed by literature on research with children (e.g. Christensen & James, 2017; Eckhoff, 2019; Pinter & Kuchah, 2021). Within this interdisciplinary paradigm of research, often referred to as the ‘new sociology of childhood’, an emphasis is placed on the role of children in research, moving from being mere objects of study to taking active part in the research (e.g. Fielding, 2001). From the perspective of children as experts of their own worlds, the aim is to provide them with a voice, which is manifested in the methodological considerations of a research project, for instance, in the choice of the data elicitation techniques, the analyses, and in the research reports. This focus on children being active agents in the research and being given a voice is often attributed to the introduction of the United Nation’s Convention on the Rights of the Child (UNCRC, 1989), which states that children have a right to share their views about matters concerning them. Thus, it could be argued that the introduction of the UNCRC promoted a view of children as active agents. A paradox, however, is that whilst children are amongst the most institutionally governed citizens, they also carry the least influence in decision-making (Kellett, 2010). This includes democratic engagement in educational settings (Kuchah & Milligan, 2021). Nevertheless, governments and state institutions – and thus, by extension, schools and researchers – are required to provide spaces where children can voice their opinions and democratically engage in matters concerning them, which is apparent in the Norwegian curriculum (Udir, 2020a). Providing children with a voice was therefore an important consideration in the present study. 1.1.4 The curriculum and English teaching in Norway The current project was conducted during the introduction of a new curriculum (LK20) in Norway. Thus, the project occurred in a transitioning period between the national curriculum of 2006 (LK06) and the new curriculum of 2020. Consequently, the learners Introduction 5 participating in this study were still taught in accordance with the LK06. However, the study is even more relevant in the light of the new curriculum. The national curriculum consists of three parts: 1) the core curriculum, which addresses the overarching purposes, obligations, and values of education, as decreed by law; 2) the overarching principles of the subject, i.e. “Purpose” (Udir, 2006a) or “About the subject” (Udir, 2020b), which includes the relevance and central values of the subject, across all grades (primary and secondary levels); and 3) the subject-specific competence aims and assessment, in primary school after 2nd, 4th, and 7th grades respectively. With the UNCRC (1989) in mind, the core curriculum both in the LK06 (Udir, 2006b) and the LK20 (Udir, 2020a) is highly relevant, as it states that the school should promote democratic values and facilitate active participation4. With regard to the purposes of the English subject, both curricula emphasise a focus on world Englishes and the ability to communicate across cultural backgrounds5. Finally, concerning the subject-specific competence aims specifically related to pragmatics, aims with similar foci can be found in both curricula, with a progression from 2nd to 7th grade. For instance, at the time of the instruction in the present study, the curriculum stated that learners were expected to be able to “use expressions of politeness and appropriate expressions for the situation” after 7th grade (Udir, 2006a)6. In light of the competence aims and the curriculum, it is relevant to draw attention to the English proficiency of Norwegian learners, which may help shed light on the teaching context. English has been a compulsory subject from 1st grade in Norway since 1997, and is in fact the only additional language taught as a compulsory subject, while others, e.g. German or Spanish, being elective subjects from 8th grade. Norwegian learners of English are relatively proficient in English and are currently ranked fifth on the English proficiency index (Education First, 2020). With regard to the primary level, the expected levels according to the CEFR are approaching A1 (2nd grade), A1-A2 (4th grade), and A2-B1 (7th grade) (Hasselgreen, 2005). The present study took place in the 7th grade. 4 The LK20 is more specific in this respect and states that “[t]he pupils must experience that they are heard in the day-to-day affairs in school, that they have genuine influence and that they can have impact on matters that concern them” (Udir, 2020a). This mirrors Article 12 in the UNCRC (1989), i.e. “States Parties shall assure to the child who is capable of forming his or her own views the right to express those views freely in all matters affecting the child, the views of the child being given due weight in accordance with the age and maturity of the child”. 5 However, there is a shift from an explicit focus on L1 English contexts in the LK06 to a focus on intercultural communication, regardless of L1 backgrounds, in the LK20. 6 Albeit broader in scope, an aim focusing on the ability to use ‘polite expressions’ can also be found in the LK20: “express himself or herself in an understandable way with a varied vocabulary and polite expressions adapted to the receiver and situation” (Udir, 2020b). Introduction 6 To sum up, at the time of the instruction, the LK06 included specific learning aims that emphasised a focus on pragmatics. In addition, the purpose of the English subject was to foster the ability to communicate across cultures and backgrounds. Thus, the present study aimed to teach requests with the view to developing agency. Furthermore, in line with the core curriculum, the present study aimed to provide the learners with a voice in the project. However, despite the study being grounded in the LK06, it has become increasingly relevant in light of the LK20, where the link to the UNCRC (1989) is even clearer, and the focus on intercultural communication is emphasised. 1.1.5 An overview of the study With the aforementioned sections as a backdrop, I turn to the present study, which aimed to explore the teaching of English requests, using a concept-based approach, with two intact 7th-grade classes in a primary school. The instruction lasted for four weeks (four hours total), with the researcher teaching the material. Data was collected prior to and following the instruction, resulting in approximately three months of fieldwork. The study was conducted in the spring of 2019 and aimed to explore the impact of a concept-based approach to teaching requests on the learners’ request production and awareness, and their engagement with pragmatics. Figure 1.1 provides a chronological overview of the fieldwork, which lasted approximately three months. Figure 1.1: Overview of the fieldwork with the techniques used and the data they elicited. The “A” followed by a roman numeral (e.g. A-I) refers to the article in which the data was presented. Introduction 7 In the study, the researcher adapted data elicitation methods, e.g. the discourse completion task (DCT) and group interviews, informed by the literature on research with children within 'the new sociology of childhood'. Considerations taken during the research included building trust with the participants, using participant-friendly techniques, and combining tasks to facilitate responses in different modalities. The research is presented in four articles. In Article I, learner-produced scripts from Readers Theatre (see section 3.3.2) were used to prompt metapragmatic discussions. Article II employed the Video-prompted Oral Discourse Completion Task (VODCT, see section 3.3.1) to elicit production data in a pre-post-delayed design. Articles III and IV employed task-based interviews to enable both verbal and non-verbal responses (see section 3.3.3). Table 1.1 presents the main research question of the study and an overview of the four articles, stating the aims and research questions for each article (Article I-IV). Table 1.1 The main research question of the study, and the titles, aims or research questions of the four articles. Main research question How does a concept-based approach to teaching requests impact young language learners’ request production and awareness, and their engagement with pragmatics? Articles I Myrset, A. & Savić, M. (2021). “If an astronaut were on the moon…”: Eliciting metapragmatic data from young L2 learners. Applied Pragmatics, 3(2), 163-196. - To provide an overview of the methods used to elicit metapragmatic data in research with young language learners - To present three data collection techniques designed and used in two research projects conducted by the authors II Myrset, A. (Pending revisions). 'You could win Masterchef with this soup. Can I get some more?' Request production and the impact of instruction on young EFL learners. Journal of Pragmatics. - To what extent does concept-based instruction of EFL requests with young learners influence o the learners’ linguistic repertoire of head acts, and internal and external modification strategies? o the learners’ linguistic variation depending on familiarity and age of the interlocutor? III Myrset, A. (2021). Scientific concepts as meaning-making resources for young EFL learners in the learning of pragmatics. Intercultural Communication Education, 4(2), 191-212. - Do young language learners employ scientific concepts to express metapragmatic understandings following a period of concept-based instruction? If so, how? Introduction 8 IV Myrset, A. (In preparation). Giving young language learners a voice: learner feedback on pragmatics instruction. To be submitted to TESL-EJ. - How did young language learners appraise various components in a project related to pragmatics instruction? - What were the learners’ explanations behind their appraisals? Four articles are included in this synopsis. Of these, three articles (II-IV) aimed to address the three important aspects (i.e. production, awareness, and engagement) in the overarching research question, whereas Article I, an investigation of previous elicitation techniques used in research exploring YLLs’ metapragmatic awareness through a systematic review, provided the background and rationale for the selection of techniques. Furthermore, this article presents some elicitation techniques employed in the present study. Article II explores the requests produced by the learners in a pre-, post-, and delayed post-test, through the VODCT. Article III draws on discussions emerging in group interviews and investigates the learners’ use of scientific concepts to express their metapragmatic understandings about requests. Finally, Article IV discusses the learners’ perceptions about the project, aiming to give them a voice in research. 1.2 Contributions of the study Considering the gap in research with YLLs both in Norway and more broadly within the field, this study adds to knowledge both locally and globally. For the Norwegian context, despite the curriculum providing pragmatics-related learning aims already at the primary school level, empirical research investigating learners' pragmatic development remains sparse. Most studies of request production have focused on older English L2 speakers, such as teachers (Krulatz, 2016), learners in upper secondary school (Brubæk, 2012), and university students (Awedyk, 2003). Some developmental studies have investigated request production (Savić, 2015; Savić et al., 2021) and metapragmatic awareness (Savić, 2021; Savić & Myrset, Forthcoming-a, Forthcoming-b) of young English language learners in primary school. However, despite calls for instruction studies in L2 pragmatics (Brubæk, 2012; Savić, 2015), none have been conducted in the Norwegian context to the best of the author's knowledge. The global context paints a similar picture, that is, some research has explored YLLs’ L2 (meta)pragmatic development in English (e.g. Achiba, 2003; Lee, 2010; Portolés, 2015). However, few studies have explored L2 pragmatics instruction with YLLs (e.g. Ishihara, 2013); thus, the evidence pertaining to YLLs remains sparse compared to that with adults, which is a general trend in applied linguistics (Pinter, 2014). With this in mind, the paucity of empirical evidence results in a lack of knowledge in relation to Introduction 9 what can be taught within pragmatics, and ultimately how pragmatics teaching can be approached. Thus, the current study adds to previous knowledge by providing evidence regarding the teachability of pragmatics and learners’ engagement with this process. Consequently, the study may serve as a support for researchers, teacher educators, teacher students and teachers, both in Norway and globally. Another contribution of the study lies in the teaching approach it has adopted. Informed by SCT, the study adopted a concept-based approach for teaching pragmatics (Morollón Martí, Forthcoming; van Compernolle, 2014). Whereas this approach has gained traction within the field, its affordances have been explored exclusively with (young) adult learners. Thus, by tailoring a concept-based approach specifically for YLLs, the present study provides empirical evidence of YLLs’ pragmatic gains and affective responses to this kind of instruction, broadening the scope of instructional pragmatics studies by focusing on an uncharted group of learners. Finally, in terms of methodology, the study was largely informed by literature on conducting research with children. This involved adapting elicitation techniques aiming to provide the learners with a voice, whilst at the same time ensuring that the techniques generated relevant data. Since prior pragmatics research with YLLs has largely been based on research methods mirroring those used with adults (Culpeper et al., 2018), and thus not taking into account the potential differences between children and adults (Pinter, 2014; Punch, 2002b), the current research study provides a novel approach to data collection within the field. The use of innovative data elicitation techniques (e.g. Readers Theatre) and an emphasis on child voices, opening up for a discussion about their involvement in research, present a major contribution to the field of pragmatics. 1.3 Structure of the synopsis This synopsis provides insight into the project as a whole and how the four articles are linked together by offering a more detailed overview of its theoretical and methodological underpinnings, as well as the most important findings. Chapter 2 presents the theoretical concepts and empirical studies relevant to the current project. Chapter 3 sheds light on the methodological considerations regarding the design of the study, the instruction, and the data collection and analysis procedures. Chapter 4 presents summaries of the four articles, which are attached at the end of this synopsis. Chapter 5 provides a discussion and conclusion based on the findings from the project as a whole, as well as its limitations and suggestions for future directions of research within the context of YLLs and pragmatics instruction. Introduction 10 Theory and previous research 11 2 Theory and previous research This chapter presents the theoretical underpinnings of the study presented herein. In addition, it provides reviews of the relevant previous research. Section 2.1 explores the nature of requests, the pragmatic target of the instruction in the present study, and how they develop in language learning. In addition, the request strategies relevant for the current study are presented in this section. Section 2.2 presents sociocultural theory (SCT), which provided the guiding principles for the instruction, focusing on some of the core aspects related specifically to instructional pragmatics and pragmatic development. These include the zone of proximal development, spontaneous and scientific concepts, and learner agency (e.g. van Compernolle, 2014; Vygotsky, 1934/2012, 1978). In addition, this section provides a review of previous studies that have adopted SCT as a pedagogical foundation for teaching pragmatics through concept-based approaches. Section 2.3 presents reviews of research investigating pragmatics with YLLs. More specifically, this section provides an overview of the previous research that has explored YLLs’ metapragmatic awareness, grounded in a systematic review7, and explores previous instructional pragmatics research with YLLs. Finally, Section 2.4 concludes the chapter with a brief summary. 2.1 Requests Requests are attempts at moving the hearer to perform an action (Searle, 1979), most often to the benefit of the speaker (Ishihara & Cohen, 2014; Pérez-Hernández, 2021). The way in which requests are performed may vary as they occur in a “relationship between form, meaning, and pragmatic prerequisites involved” with potentially “high social stakes involved for both interlocutors in choice of linguistic options” . (Blum-Kulka et al., 1989, p. 11). Thus, the act of requesting involves balancing two sets of considerations: pragmalinguistic ones, i.e. the link between pragmatics and grammar, and sociopragmatic ones, i.e. the link between language and culture (e.g. Kasper, 2001; Leech, 1983; Thomas, 1983; van Compernolle, 2014). With this in mind, requests are multifaceted and contextually situated (Ishihara & Cohen, 2014; Ogiermann, 2009; Pérez-Hernández, 2021; Spencer-Oatey, 2008), which means that producing requests involves choosing between a range of pragmalinguistic strategies as well as taking into consideration the context in which they occur. Requests, which have been vastly researched (Ishihara & Cohen, 2014; Pérez-Hernández, 2021), occur from an early stage of L1 development (Bernicot, 1994; 7 The systematic review was conducted by the author and his supervisor and provided the backdrop for Article I (Myrset & Savić, 2021) of this thesis. Theory and previous research 12 Cekaite, 2013; Dorcheh & Baharlooie, 2016; Zufferey, 2014). Indeed, requests can be produced non-verbally, such as through pointing, and verbally (accompanied by gestures) by toddlers, and by the age of two and a half include a “wide repertoire […] that gradually become[s] more sophisticated and refined” (Cekaite, 2013, p. 2). These early stages of development are characterised by learners engaging with their social world, such as interacting with adults and peers (Bernicot, 1994; Cekaite, 2013). Furthermore, requests are commonly used in communication (Pérez-Hernández, 2021; Stavans & Shafran, 2018). Thus, it is clear that requests are an important part of communication, reflected by their early appearance, with production and reception becoming increasingly sophisticated with age and frequent use in everyday life, which is why they were chosen as the pragmatic target in the current study. 2.1.1 Request strategies Following Blum-Kulka et al. (1989, p. 275)8, requesting involves a range of strategies9 starting from the head act, i.e. “the minimal unit which can realize a request”. These head acts can be manifested in various ways, depending on their directness. Directness is the “degree to which the speaker's illocutionary intent is apparent from the locution” (Blum-Kulka et al., 1989, p. 278). In other words, the directness is dependent on the level of transparency, leaving more or less responsibility for interpretation on the hearer (Blum-Kulka, 1987). The directness of the head act, or request proper, can thus be viewed as on a continuum from transparent to opaque, and can be divided into three overarching categories: direct, conventionally indirect, and non-conventionally indirect (henceforth referred to as ‘hints’). On this continuum, direct requests are the most transparent, being “realized by requests syntactically marked as such, such as imperatives, or by other verbal means that name the act as a request, such as performatives” (Blum-Kulka & Olshtain, 1984, p. 201), e.g. ‘Close the window!’ (imperative) and ‘I am asking you to close the window.’ (performative). Situated in the middle of the continuum, conventionally indirect requests are realised through contextual preconditions, and are conventionalised within a language (Blum-Kulka & Olshtain, 1984), e.g. ‘Could you close the window?’. Finally, the head act characterised by various degrees of opacity are hints. These are 8 Since the publication of Blum-Kulka et al. (1989), a number of taxonomies related to request strategies have been proposed (e.g. Alcón Soler et al. 2005; Woodfield & Economidou-Kogetsidis, 2010). However, the framework proposed by Blum-Kulka et al. remains the most widely cited (Kádár & Haugh, 2014; Spencer-Oatey, 2008; Spencer-Oatey & Kádár, 2021) and has provided the foundation for more recent taxonomies. Thus, their framework was used in the current study. 9 The strategies presented are those relevant for the current study. Theory and previous research 13 realised “by either partial reference to object or element needed for the implementation of the act ('Why is the window open'), or by reliance on contextual clues ('It's cold in here')” (Blum-Kulka & Olshtain, 1984, p. 201). The head act can be modified internally and externally through modification strategies. These strategies enable the speaker to soften or increase the force of the requests. For example, requests can be internally softened through lexical downgraders, i.e. optional lexical devices that soften the force of the request, such as possibly/perhaps and the marker please10; or aggravated through lexical uptoners, which add negative connotations to the request, e.g. “[c]lean up that mess!” (Blum-Kulka et al., 1989, p. 286). Another internal modification strategy is syntactic modification, which is achieved through optional syntactic devices that mitigate the requestive force, e.g. negation (‘you couldn’t lend me some money, could you?’), as well as modal verbs, which frequently occur in requests, e.g. ‘can/may/could I have a glass of water?’. When it comes to external modification, it can be realised through alerters, i.e. elements to get the hearer’s attention, such as titles/roles (e.g. teacher or Mrs), attention getters (e.g. excuse me), and supportive moves. Similar to internal modification, supportive moves, which precede or follow the head act, can be mitigating or aggravating. Examples of mitigating supportive moves are: preparators, i.e. announcements of an upcoming request through enquiring about the hearer’s availability or by asking for permission to make a request; grounders, i.e. providing a reason, explanation, or justification that accompanies the request; promises of reward, i.e. announcing that fulfilling the request will be rewarded; and sweeteners, i.e. showing appreciation through, for instance, compliments (Blum-Kulka et al., 1989; Blum-Kulka & Olshtain, 1984). On the opposite end, aggravating moves can be realised through, for example, insults and threats. In sum, starting from the head act, requesting can involve a range of strategies that modify the request internally and externally. These strategies play various functions in requesting as they can either mitigate or aggravate the force of the request and represent the linguistic options that are available to speakers in communication. 2.2 Sociocultural theory The core of sociocultural theory (SCT) is rooted in a belief that development occurs in a unity between biological conditions and the social world (Lantolf & Poehner, 2014; Lantolf et al., 2018), with its origins in Vygotsky’s work (1934/2012, 1978). In other 10 ‘Please’ is often referred to as a ‘politeness marker’ (e.g. Blum-Kulka et al., 1989; Pérez-Hernández, 2021). Theory and previous research 14 words, a child’s cognitive development is influenced by social interaction, where, for instance, values, beliefs, and strategies for problem-solving are acquired in collaboration with more knowledgeable people from their social world. This development occurs from infancy, but for Vygotsky, formal learning was “the natural initial stage of development of scientific knowledge” (Gal'perin, 1992, p. 69; Negueruela, 2003). In other words, the educational context was considered a platform in which development could move from simply learning from the social environment to directing focus towards theoretical knowledge. Vygotsky’s theories became internationally recognised following the translated publication of Mind in Society (Vygotsky, 1978), both as a theoretical lens for investigating development and as an influential foundation for (language) teaching (Gredler, 2012; Kinginger, 2002; Lantolf et al., 2018)11. Theoretical constructs from Vygotsky’s work (1934/2012, 1978), e.g. the zone of proximal development (ZPD), have since inspired new concepts, such as 'scaffolding'. The ZPD and scaffolding have since become prominent for understanding how learning occurs in education and how development can be mediated, i.e. through involvement of, for instance, a person or concepts as a third factor in interaction (Kozulin, 2018; Lantolf & Poehner, 2014; Negueruela, 2003; van Compernolle, 2014; Vygotsky, 1978). The appearance of SCT also sparked new ways of viewing second and foreign language acquisition, which Vygotsky also theorised in his own work, such as conceptual knowledge in a foreign language in Thought and language (Vygotsky, 1934/2012). Consequently, the SCT paradigm has become “a theory that L2 scholars draw heavily upon” (Lantolf et al., 2018, p. 5), with aims to understand the process of language learning and its relation to the socio-cultural context. Following the surge of SCT-informed research, explorations of pragmatics instruction have also been approached through this theoretical lens, namely through dialogic (Ishihara, 2013; Ishihara & Chiba, 2014) and concept-based approaches (e.g. Morollón Martí, Forthcoming; Nicholas, 2015; van Compernolle, 2012, 2014; van Compernolle et al., 2016), the latter drawing heavily on works by Gal'perin (1979, 1989, 1992) and Negueruela (2003). Consequently, some constructs grounded in SCT that have been fundamental for informing pragmatics instruction specifically will be further explored here. These are scientific and spontaneous concepts, and the zone of proximal development. 11 See Lantolf et al. (2018) for a detailed discussion about the history and influence of SCT on language learning. Theory and previous research 15 2.2.1 Learner development In essence, the zone of proximal development (ZPD) is “the distance between the actual developmental level as determined by independent problem solving and the level of potential development as determined through problem solving under adult guidance or in collaboration with more capable peers” (Vygotsky, 1978, p. 86). In other words, the ZPD captures the essence of SCT by describing a relationship between the individual and their social world, namely how children’s development occurs through interaction with adults and peers. The ZPD has thus become widely used in research and teaching, the latter with a view that “learning is a socially creative activity, and […] an effective way of people to develop as learners" (Holzman, 2018, p. 51). Development through the ZPD occurs from early childhood, through schooling and beyond, and is linked to what Vygotsky (1998) referred to as ‘periods of crisis’ (i.e. turning-points). These periods of crisis occur at different ages, respectively newborn, 1, 3, 7, 13, and 17 (Vygotsky, 1998), and there is thus a gradual shift, in concert with children’s social and emotional growth (McKay, 2006), moving from a reliance on parents/guardians as the capable others to peers serving such roles. One such turning point, according to Vygotsky, occurs at the age of 13, when learners are in the last year of primary school in Norway, which is the age in focus in the present study. At this stage, children move from attention to what is obvious, to understanding and deduction, and ultimately to a higher form of intellectual activity. As a result, these learners will have a larger cognitive capacity for abstract thought and a focus on scientific concepts (discussed below). The focus on learner reflections and scientific concepts was thus considered appropriate for the instruction in the present study. With the ZPD in mind, development does not happen solely by maturation and interaction with the environment. Rather, it happens through help from more capable others: first, through adults interpreting the world to children and, later, through external mediators, with “specifically constructed activities, formal educational activities being the most prominent of them” (Kozulin, 2018, p. 38). These external mediators facilitate development towards each period of crisis, and, in the case of language, shape children's ability to communicate in and with their surroundings. Furthermore, formal learning, i.e. education, aims to foster development of self-regulation (agency), in which reflection and metacognition play a vital role (Fox & Riconscente, 2008; Kozulin, 2018). Considering that learners develop through interaction with their surroundings, two concepts proposed by Vygotsky (1934/2012) Theory and previous research 16 become highly relevant: spontaneous and scientific12 concepts, both characterised and differentiated by how they are acquired. It is important to note, however, that although they follow different trajectories for acquisition, both play an important role in development (Karpov, 2018; Vygotsky, 1934/2012), and the one does not exclude the other. There is a strong link between the ZPD and conceptual knowledge, namely spontaneous concepts indicate the current stage of development, whilst scientific concepts indicate directions and potential for future development (Vygotsky, 1934/2012). 2.2.2 Spontaneous and scientific concepts Spontaneous concepts are empirical and characterised by developing without conscious attention (van Compernolle, 2014), in a “spontaneous manner in the course of engagement in social activities” (Negueruela, 2003, p. 63). In other words, spontaneous concepts are acquired from lived experiences, and their development “knows no systematicity and goes from the phenomena upward toward generalizations” (Vygotsky, 1934/2012, p. 157). For example, through interaction with their environment, children learn how to use the L1, e.g. the verb ‘give’, before they develop knowledge and more in-depth understanding about the overarching functions and meanings of verbs. As Vygotsky (1934/2012, p. 205) points out, “[t]he inception of a spontaneous concept can usually be traced to a face-to-face meeting with a concrete situation”, and a spontaneous concept is formed on the basis of generalisations from lived experiences in “the absence of systematic instruction” (Karpov, 2018, p. 102). Spontaneous concepts are thus by nature unsystematic13, and their limitation lies in a learner’s “inability to use these concepts freely and to form abstractions” (Vygotsky, 1934/2012, pp. 157-158). Consequently, spontaneous concepts lack transferability to other contexts different from where they were encountered. At the same time, spontaneous concepts are also vital as they provide the foundation for the development of scientific concepts (Infante, 2018; Karpov, 2018; Vygotsky, 1934/2012). Whereas spontaneous concepts are unsystematic and empirically acquired through lived experiences, scientific concepts are characterised by their systematic, hierarchical, and abstract nature (Karpov, 2018; Vygotsky, 1934/2012). Their 12 In the literature, these are also referred to as ‘everyday’, ‘empirical’ or ‘experiential’ concepts (spontaneous); and ‘academic’ or ‘theoretical’ concepts (scientific) (e.g. Morollón Martí, Forthcoming; Neguerela, 2003; van Compernolle, 2014; Vygotsky, 1934/2012). 13 Karpov (2018) uses the example of birds to describe the unsystematic nature of spontaneous concepts. At an early stage of a child’s development birds are characterised by their ability to fly; thus, pre-schoolers may not define a penguin as a bird, but will do so with a bat. Theory and previous research 17 acquisition is grounded in systematic instruction14, and whilst the development of scientific concepts can be found in other forms of learning, e.g. apprenticeship (Negueruela, 2003), they are particularly prevalent in school. As Vygotsky (1934/2012, pp. 166-167) holds, “[i]nstruction is one of the principal sources of the schoolchild’s concepts and is also a powerful force in directing their evolution”. Thus, to Vygotsky, education was more than simply acquiring new knowledge. Rather, education could foster development through scientific concepts which are unavailable in everyday life (Lantolf & Zhang, 2017). Furthermore, as opposed to the limited transferability of spontaneous concepts, scientific concepts are independent of specific contexts, and can thus be recontextualised to other (diverse) situations (Infante, 2018; Negueruela, 2003; Swain et al., 2015; Vygotsky, 1934/2012), thereby providing knowledge that goes beyond what can be spontaneously acquired. An instructional focus on scientific concepts therefore provides an orienting basis for action in diverse situations (Gal'perin, 1989; Morollón Martí, Forthcoming). This systematic focus on scientific concepts enables learners to generalise and provides a foundation for awareness and reflection (Karpov, 2018; Negueruela, 2003; van Compernolle, 2014; Vygotsky, 1934/2012; Zuckerman, 2004), with thinking becoming “independent of their personal experiences” (Karpov, 2018, p. 103). As Vygotsky (1934/2012, p. 181) argues, “[r]eflective consciousness comes to the child through the portals of scientific concepts”, where the formal learning mediates development in the ZPD (Zuckerman, 2004). With this in mind, scientific concepts provide a set of interrelated features that can guide action in diverse contexts (Gal'perin, 1989; Negueruela, 2003; Vygotsky, 1934/2012), and provide a foundation for reflection about the object in question. It is important to note, however, that Vygotsky argued that both forms of conceptual knowledge (spontaneous and scientific) are vital to development, closely connected, and develop over time, and that the “introduction of new concepts does not preclude spontaneous development, but rather charts new paths for it” (Vygotsky, 1934/2012, p. 161). This conceptual development was also theorised by Vygotsky in relation to foreign language learning and its contingency on the L1. However, it is worth pointing out that Vygotsky’s context was vastly different from the one that 21st-century learners reside in, where the borders between learning a first language and other languages are substantially more blurred, e.g. exposure to languages through different forms of media, leisurely travel, and migration. Nevertheless, the distinction between spontaneous and scientific concepts is still relevant due to their conditions for acquisition. 14 Drawing again on Karpov’s (2018) example of birds, scientific concepts provide other characteristics that provide a system, such as vertebrate and animals that lay eggs. Theory and previous research 18 2.2.2.1 Sociocultural theory and L2 pragmatic development When it comes to languages and L2 acquisition, Vygotsky argued that the systems of meaning in the L1 are transferred to the process of learning the L2, whilst “a foreign language facilitates mastering the higher forms of the native language” (Vygotsky, 1934/2012, p. 207). Whilst the L2 serves as support for mastery in the native language, in Vygotsky’s view, the L1 also plays a mediating role in the L2 through the meanings already established in the L1. Thus, the L1 and lived experiences can serve as scaffolding in the process of L2 learning (Chavarría & Bonany, 2006; McConachy, 2018). Indeed, Chavarría and Bonany (2006, p. 136) argue that “the L1 may be strategically used as a means of communication in the classroom”, with one positive aspect being as a “stepping stone into potentially difficult contents (e.g. textual or cultural aspects)”. From the perspective of pragmatics, specifically in relation to requests, one could thus argue that the use of strategies and their linguistic manifestations are mainly developed spontaneously through everyday interaction in the L1. Considering that requests occur early in language development (Cekaite, 2013), and are used frequently, even before conscious attention, this knowledge about requesting can be brought into the L2 and used as a scaffold for production and interpretation. Unless pragmatic phenomena, such as requests, are systematically addressed through scientific concepts, learners are potentially deprived of tools for reflection and making informed decisions about their meaning and use. Figure 2.1 provides an example of scientific concepts related to a pragmalinguistic aspect of requesting: directness of head acts. Such concepts provide abstract knowledge focusing on the (intended) meaning potential of strategies, e.g. hints, rather than focusing only on specific pragmalinguistic resources, e.g. “Do you have a pencil?”. Figure 2.1: A visual representation of a hierarchy of scientific concepts relating to directness, with sub-concepts for directness levels, as well as examples of pragmalinguistic resources within each sub-concept (also presented in Article III (Myrset, 2021)) Theory and previous research 19 2.2.3 Teaching pragmatics through concepts Within SCT, mediated action is central for instructional pragmatics (van Compernolle, 2014; van Compernolle et al., 2016). This means that rather than aiming to teach adherence to social conventions, the instruction aims to provide learners with conscious control over their choices. This control includes an “ability to break with pragmalinguistic and/or sociopragmatic conventions in order to achieve a desired effect in light of present circumstances, constraints, and potential conflicts and/or points of tension” (van Compernolle, 2014, p. 42). Thus, SCT-informed concept-based approaches focus on teaching concepts related to the pragmalinguistic and sociopragmatic dimensions (Al Jumah, 2021; Nicholas, 2015), adhering to the view that social action is mediated by pragmalinguistics, which is again mediated by sociopragmatics (van Compernolle, 2014). In other words, concepts related to pragmalinguistic and sociopragmatic dimensions serve as an orienting basis for making choices in communication. 2.2.3.1 Agency and rules of thumb Within SCT, agency is defined as “the socioculturally mediated capacity to act and to assign meaning to one’s actions” (van Compernolle, 2014, p. 21). In other words, it is the ability to make informed choices and act on them, mediated by sociocultural, contextual, and interpersonal conditions, which allows learners to create and expand meaning (Levi & Poehner, 2018; Martin, 2004; Mercer, 2011). What is central in SCT-informed instruction is that it aims to foster learner agency rather than teaching pragmatic ‘rules of thumb’ (Morollón Martí, Forthcoming; Nicholas, 2015; van Compernolle, 2014). Such rules of thumb are prescriptive generalisations of linguistic forms as inherently im/polite, in/formal or in/appropriate in certain communicative contexts, or mapping specific pragmalinguistic forms onto specific sociopragmatic features, e.g. familiarity with and age of the interlocutor (Liddicoat & McConachy, 2019; McConachy & Liddicoat, 2016; McConachy & Spencer-Oatey, 2020; Nicholas, 2015; van Compernolle, 2014, 2018). In this respect, teaching pragmatic rules of thumb to some extent aligns with a traditional perspective of politeness (e.g. Brown & Levinson, 1987; Leech, 1983), in which contextual variables were reduced to three static variables, i.e. the relative power (P) and distance (D) between the interlocutors, and the imposition (R) of the speech act. Following van Compernolle (2014, 2018), when aiming to foster agency and the ability to negotiate social meaning, rules of thumb are problematic for three main reasons. Theory and previous research 20 Firstly, whilst rules of thumb may provide practical guidelines, such prescriptive rules are inconsistent in communication. Secondly, since teaching rules of thumb assign specific forms to particular contexts, they inhibit agency. Finally, rules of thumb focus on what to say to whom, rather than taking into account the meaning potential of various choices. One could also argue that teaching prescriptive norms as rules, e.g. “you should say X to Y”, is counterproductive as it would be impossible to teach learners about all available situations and contexts of language use. Rules of thumb may therefore “have potential pernicious effects on L2 development since they direct L2 learners to form hypotheses and understandings of language and communication in a simplified, incomplete, and unsystematic fashion” (Negueruela, 2003, p. 85). Teaching pragmatics by presenting a “set of doctrinal, norm-referenced rules of thumb” (van Compernolle, 2014, p. 5) raises a question about the role of the native speaker, where instructional pragmatics has often relied on native speaker performance as a benchmark (e.g. Eslami-Rasekh et al., 2004; Félix-Brasdefer, 2006; Hosseini & Safari, 2018)15. However, the native speaker construct is in itself ambiguous (Davies, 2004), with norms being fluid (Spencer-Oatey & Kádár, 2021), and languages being characterised by intralinguistic variation, which has, for instance, been identified in native speakers of different varieties of English (Barron, 2008, 2021). This is not to argue that it is not useful to acquire knowledge about L2 pragmatic norms (McConachy, 2013), but rather that one should avoid viewing native speaker norms as prescriptive rules of thumb, since such rules may result in overgeneralisations (van Compernolle, 2014). Furthermore, with English serving as a lingua franca, where meaning is negotiated against a backdrop of different linguistic and cultural backgrounds, coming across as a native speaker is not necessarily the aim of acquiring the language (House, 2010; Taguchi, 2011). As House (2010, p. 365) holds: Localized, regionalized or otherwise appropriated varieties – whose linguistic surface is English, but whose speakers creatively conduct pragmatic shifts in their use of this auxiliary language – are taking over the linguistic landscape. Non-native speakers of English anywhere in the world are developing their own discourse strategies, speech act modifications, genres and communicative styles in their use of ELF. The focus on fostering agency in teaching and learning L2 pragmatics has thus become increasingly emphasised from the perspective of the intercultural learner (e.g. Liddicoat & McConachy, 2019; McConachy, 2013, 2018; McConachy & Liddicoat, 2016; 15 This bears resemblance to the traditional perspective of politeness, namely the Model Person, i.e. a speaker or hearer that is “a wilful fluent speaker of a natural language” (Brown & Levinson, 1987, p. 58). Theory and previous research 21 Morollón Martí, Forthcoming). Furthermore, the focus on agency aligns with discursive perspectives of politeness, in which politeness starts from a lay conceptualisation of the term (Eelen, 2001; Watts, 2003), rather than a “a superordinate, universal term that can then be applied universally to any socio-cultural group at any point in time” (Watts, 2003, p. 9). As McConachy (2018, p. 26) argues, “it is difficult for language learners to develop a true sense of agency in their use of L2 pragmatics if they are socialized into a view of language as a highly constrained system”. Rather, there is a need for learners to reflect and develop awareness about pragmatic variation in language use in which learners construct “a more dynamic perspective on language as a whole” (McConachy, 2018, p. 26). Such a view is thus closely linked to operationalisations of metapragmatic awareness within SCT-informed approaches to pragmatics instruction. 2.2.3.2 Metapragmatic awareness within concept-based approaches to pragmatics instruction When it comes to teaching pragmatics, there is consensus that learners benefit more from explicit input, i.e. the teacher providing direct metapragmatic explanations, as opposed to implicit instruction16 (Kasper, 2001; Plonsky & Zhuang, 2019; Taguchi, 2015). This distinction, however, is not as dichotomous as often presented, and input should be viewed as on a continuum from implicit to explicit (Taguchi, 2015). Furthermore, what presents itself as a challenge within research on L2 pragmatics teaching is that the term metapragmatic awareness is used inconsistently or not always clearly defined (McConachy, 2018; Nikula, 2002) and explicitly operationalised. It is thus difficult to compare the impact of various instructional studies since the kinds of metapragmatic information with which lerarners were provided may have varied considerably. Broadly speaking, metapragmatic awareness is “a crucial force behind the meaning-generating capacity of language in use” (Verschueren, 2000, p. 439), and refers to reflexivity about language usage. This reflexivity concerns language users’ awareness about how language is used in communication with others (Culpeper & Haugh, 2014)17. In L2 pragmatics research, metapragmatic awareness has been operationalised in markedly different ways. For example, in some studies, metapragmatic awareness has 16 Implicit pragmatics instruction does not provide any metapragmatic explanations, but rather aims to develop learners’ implicit understandings in which they deduce their own rules through, for instance, input flood and consciousness raising activities (Kasper, 2001; Plonsky & Zhuang, 2019; Taguchi, 2015). 17 Such reflexivity may involve interpersonal evaluations, such as those pertaining to politeness (Spencer-Oatey & Kádár, 2021), for instance valency, i.e. evaluative categorisations of language use through scales ranging from e.g. good-bad, polite-impolite, or appropriate-inappropriate (Kádár & Haugh, 2013). Theory and previous research 22 been operationalised as “knowledge of what is considered (in)appropriate language use in a given context rather than why” (McConachy & Liddicoat, 2016, p. 16), or as mappings between form, function, and context (McConachy & Spencer-Oatey, 2020), and metapragmatic explanations have thus been presented as rules of thumb (van Compernolle, 2014). In the case of requests, such mappings could, for instance, be linking specific strategies to specific relational categories, such as ‘friend’, ‘boss’, or ‘colleague’. Other studies have conceptualised metapragmatic awareness more holistically and focused on learners’ (co-)construction of understandings about different pragmatic phenomena, such as politeness or self-representation (e.g. Liddicoat & McConachy, 2019; McConachy, 2013, 2018; McConachy & Liddicoat, 2016; Morollón Martí, Forthcoming; van Compernolle, 2014). The different views of metapragmatic awareness could partly be explained by different fundamental paradigms of L2 acquisition and use (McConachy & Spencer-Oatey, 2020), and also how politeness and pragmatic behaviours are viewed. For example, Tajeddin et al. (2012) assessed EFL learners’ metapragmatic awareness before and after instruction through tightly controlled questionnaires, focusing on imposition, power (referred to as status), and distance (referred to as intimacy), and compared their responses to those of native speakers. Thus, the study by Tajeddin et al. (2012) aligns with traditional theories of politeness (e.g. Brown & Levinson, 1987; Leech, 1983), by relying on static contextual variables and a model person. McConachy (2013), on the other hand, investigated learners’ metapragmatic awareness through their interpretations, e.g. emotional dimensions, in discussions between the instructor and the learners in which the learners (co-)constructed their understandings in dialogues. This aligns with discursive views of politeness (e.g. Eelen, 2001; Spencer-Oatey, 2008; Spencer-Oatey & Kádár, 2021; Watts, 2003), in which interpretations of politeness originate from the individual’s own understandings negotiated through interaction. In other words, one study treated metapragmatic awareness as knowledge about (and conformity with) target language norms, whereas the other viewed it as an interpretative tool constituting “a bridge from the learners’ culture to the L2” (McConachy, 2013, p. 102), the latter being more in line with views about the role of metapragmatic awareness in SCT. Following the central tenets of SCT, in which agency is an important part of a learner’s language development, metapragmatic awareness becomes an important aspect of the language learning process18. As language is essentially about making choices (Culpeper & Haugh, 2014), metapragmatic awareness enables learners to make informed choices in communication, which ultimately leads to agentive language use (Morollón Martí, 18 Indeed, Verschueren (2000) makes a direct link between sociocultural theory and the dynamic processes that occur in interpretation and production of language. Theory and previous research 23 Forthcoming; van Compernolle, 2014). Metapragmatic awareness thus plays a vital mediating role (Morollón Martí, Forthcoming), in which both the pragmalinguistic and sociopragmatic dimensions are taken into account to perform social actions. Thus, in the present study metapragmatic awareness is defined as verbalised reflections about language use, contextual considerations, or their interplay, which can vary in their degree of sophistication. 2.2.4 Research on concept-based pragmatics instruction Whereas some studies have approached pragmatics instruction with YLLs through SCT (e.g. Ishihara 2013; Ishihara & Chiba, 2014), to the best of the author’s knowledge, no studies have explored the impact of introducing scientific concepts (see Section 2.2.2), i.e. concept-based instruction, with these age groups. However, recent years have seen a growing interest in concept-based approaches for teaching pragmatics with older learners (e.g. Morollón Martí, Forthcoming; van Compernolle, 2014; van Compernolle et al., 2016). A common denominator in these studies are references to the work by Negueruela (2003), who first introduced concept-based approaches for L2 instruction19. Consequently, the literature review was conducted through forward searches or citation searching (Booth et al., 2016). This forward search aimed to retrieve primary studies that had cited Negueruela (2003), focusing specifically on research exploring pragmatics. The forward search provided a first indication of studies using concept-based approaches for teaching pragmatics. An additional string search was conducted in Google Scholar. This search included the terms: 1) “concept based” AND 2) (“pragmatics instruction” OR “teaching pragmatics”) Table 2.1 presents the findings from the searches20. 19 Negueruela’s (2003, p. 230) study focused on teaching grammatical concepts, i.e. “indicative/subjunctive, conditionals, relative pronouns, articles, aspect, and verbal tense”, in L2 Spanish to university students. 20 Six of the retrieved citations are not included, of which five were not empirical research (e.g. van Compernolle, 2018) and one could not be accessed (van Compernolle & Henery, 2016). Theory and previous research 24 Table 2.1: Pragmatics instruction using concept-based approaches Author Target language Pragmatic target Age Al Jumah (2021) English Requestive behaviour 19-24 Henery (2015) French CBPIa 19-22 Kim (2013) English Sarcasm 29-37 Kuepper and Feryok (2020) German CBPI 17-20 (University students) Morollón Martí (Forthcoming) Spanish (Im)politeness Adult learnersb Nicholas (2015) English Speech acts 19-21 van Compernolle (2011) French CBPI Adult learnersb van Compernolle (2012) French CBPI Adult learners van Compernolle (2013a) French CBPI Adult learners van Compernolle (2013b) French CBPI Adult learners van Compernolle (2014) French CBPI Adult learners van Compernolle (2015) French CBPI Adult learners van Compernolle et al. (2016) Spanish CBPI Adult learners van Compernolle and Henery (2014) French CBPI Adult learners van Compernolle and Henery (2015) French CBPI Teacher van Compernolle (2016) French CBPI Adult learners van Compernolle and Kinginger (2013) French CBPI Adult learners van Compernolle and Williams (2012) French CBPI Adult learners a CBPI refers to Concept-Based Pragmatics Instruction. These studies are based on the same dataset or have departed from the teaching material developed by van Compernolle (2012, 2014). These focus on teaching sociopragmatic concepts related to self-presentation, social distance, and power hierarchies. b These studies do not refer to ages, per se, but rather refer to the participants as a “university learner of French” (van Compernolle, 2011, p. 3267) or Spanish learners “at a northeastern U.S. university” (van Compernolle et al., 2016, p. 341). These are thus referred to as adult learners. As displayed in Table 2.1, some trends can be found in previous research using concept-based approaches for teaching pragmatics. First and foremost, all the studies focus on adult learners. In addition, the majority of studies focus on teaching L2 French. Finally, the vast majority of the studies use material developed by van Compernolle (2012, 2014) as mediational tools for instruction, in both French (e.g. Henery 2015) and Spanish (van Compernolle et al. 2016). Thus, some of these will be presented below. Theory and previous research 25 In addition, some studies have also explored the teaching of EFL speech acts (Al-Jumah, 2021; Nicholas, 2015). Focusing on the sociopragmatic dimension, van Compernolle (2012; 2014) introduced concepts related to self-presentation, social distance, and power hierarchies. Simply put, the underlying premise is that sociopragmatic knowledge informs pragmalinguistic choices, resulting in social action. For instance, through the sociopragmatic concepts of “presenting oneself as tee-shirt-and-jeans or as suit-and-tie”, learners were provided with generalisable knowledge to describe pragmatic phenomena, such as formality and social distance (van Compernolle, 2014, p. 77). These concepts served as an orienting basis for making pragmalinguistic choices, e.g. the second person pronouns tu and vous in French (e.g. Henery, 2015) and tú vs. usted in Spanish (van Compernolle et al., 2016). Studies drawing on van Compernolle’s (2012, 2014) have revealed that the learners used the sociopragmatic concepts when providing their reasonings in dialogic verbalised reflection with the instructor (van Compernolle, 2014). For instance, one learner in van Compernolle’s (2014) study displayed a shift in her understandings of social distance when making pragmalinguistic choices (tu vs. vous). During these reflections, the instructor served as a mediator by co-constructing the reflections with the learner, focusing the learners’ attention towards examples of interlocutors (a professor and a child) and pointing to the scientific concept (distance) to ground the reflection. Henery (2015) used transcripts from TV interviews, in which learners were asked to provide their interpretations of the language used. Similar to van Compernolle (2014), Henery found that the instruction enabled the learners to draw attention to sociopragmatic concepts (self-representation, social distance, and power) when interpreting pragmalinguistic choices, e.g. “and then:: they’re both equally- they’re both tu” (Henery, 2015, p. 327, original highlights). In addition to a focus on individual learners’ verbal responses, van Compernolle et al. (2016), who focused on Spanish, investigated sociopragmatic development through written responses, to which they assigned scores (Awareness of concepts, Type of awareness, and Agency) for the learners (n= 19). They found that following the instruction, all but one learner had an increased score. This suggests that the concept-based instruction had improved the conceptual understandings and awareness of sociopragmatics, which ultimately led to agentive language use. Consequently, the instruction provided the learners with tools for verbalisations and problem-solving. Al Jumah (2021) investigated the teaching of EFL requests, focusing on both pragmalinguistic and sociopragmatic dimensions with Iraqi learners. The sociopragmatic concepts introduced were: power, distance, and imposition (referred to Theory and previous research 26 as ‘size of the request’), as well as self-presentation. In addition, the students were told that “age, gender, frequency of the request, etc., may also play influencing roles depending on the situation” (Al Jumah, 2021, p. 61). The pragmalinguistic dimension focused on the “main stages of request making” (Al Jumah, 2021, p. 62), i.e. pre-request, main request, post-request, and response to request. Data was collected through open-ended interviews, appropriateness judgement questionnaires, and strategic interaction scenarios (planning and producing requests). Thus, the study allowed for the exploration of both the learners’ understandings and production related to requesting. With regard to production, there was a shift from direct to conventionally indirect requests as a result of the instruction (seven sessions, 4-5 weeks). The learners also employed a wider range of head act strategies. Similarly, internal modification devices increased following the instruction, albeit to a lesser extent than head act strategies, that is, the learners used a wider range of available resources following the instruction, whilst the use of ‘please’ decreased. Al Jumah (2021) argues that the limited use of ‘please’ could be related to the request situations or L1 influence. With regard to the sociopragmatic dimension, distance was mentioned in a pre-intervention interview. However, following the instruction, the learners also showed an awareness of power and imposition when requesting. Furthermore, scientific concepts related to power, distance and imposition served as an orienting basis in their verbalisations, that is, the learners could employ these to ground their reasonings during the interviews. Through the scientific concepts, the learners were able to provide more nuanced explanations, displaying their awareness about social factors (power, distance, imposition), cultural differences between Arabic and English, and “awareness of the American cultural values influencing requests and how they are different from Arabic” (Al Jumah, 2021, p. 131) 21. Finally, Nicholas (2015) focused on two overarching concepts, interactional competence and requesting, in a study with Japanese learners. The focus on interactional competence aimed to introduce the learners to sub-concepts related to speech-as-action, adjacency pairs, conversation sequencing, and preferred and 21 It is worth noting that whereas Al Jumah (2021, p. 97) highlights “cultural differences in requests including politeness level, directness, clarity, briefness, (non)-compliance to the request, request organization, etc.” as “American cultural values”, it is not clear how these were presented or addressed during the instruction. Theory and previous research 27 dispreferred turns22; whilst the focus on requesting introduced request stages (pre-stage, main request, and post-stage). In addition, Nicholas introduced both verbal and non-verbal behaviours related to request responses. The instruction included stages of orientation (introducing a concept), execution (interaction tasks), and control (reflection) to ensure that the learners internalised conceptual understandings. The study revealed that the learners’ conceptual understandings developed during the instruction, that is, they became more sophisticated and nuanced over the course of the instruction. Furthermore, in informal post-instruction interviews, the majority of learners reported that their L2 proficiency related to requests had improved. Nicholas (2015, p. 392) argues that “[b]y promoting the development of a more nuanced, conceptual understanding of English language pragmatic norms among EFL learners, they will be equipped to apply their knowledge to a variety of contexts”. At the same time, Nicholas points out that such approaches need to consider the age and educational background of the learners. What this review highlights is that YLLs represent a group that has been overlooked with regard to teaching pragmatics through concept-based approaches. Although YLLs are largely overlooked more generally in pragmatics instruction (Plonsky & Zhuang, 2019), this could also suggest an assumption that concept-based approaches are too challenging for these age groups. Considering the focus of the present study, prior research with YLLs’ focusing on their metapragmatic awareness and affordances of instruction may, thus, help shed light on findings that informed the present study. 2.3 Previous research with young language learners 2.3.1 Research on young language learners’ metapragmatic awareness As part of the research, the present study set out to explore the learners’ metapragmatic awareness, which plays a vital mediating role in L2 learning (Morollón Martí, Forthcoming; van Compernolle, 2014), through the instruction itself and in the interviews following the instruction. Consequently, as a point of departure for this doctoral research project, a systematic review was conducted aiming to explore the 22 These concepts were operationalised in the study as follows: Speech-as-action, i.e. “the view that speech is doing something, such as apologizing” (Nicholas, 2015, p. 384, original highlights); adjacency pairs, i.e. the isolated speech act and its response; conversational sequencing, i.e. situating speech acts in “sequences of talk-in-interaction” (p. 387); and preferred and dispreffered turns, i.e. “utterances by speakers have a response that is socially comfortable (preferred), and responses that are not (dispreferred)” (p. 387). Theory and previous research 28 previous research on YLLs’ metapragmatic awareness. The result of this review is presented in Article I (Myrset & Savić, 2021), which focuses on the research methods employed in previous research. A systematic review is in essence "a methodology using a systematic, transparent process for gathering, synthesizing, and appraising the findings of studies on a particular topic or question" (Sweet & Moynihan, 2007, p. 5). In addition, a systematic review is reproducible (Booth et al., 2016), and employs strings of search terms to gather the (potentially) relevant studies in various databases. In the present study the review process, which is presented in Figure 2.2, included string searches in four databases, namely Academic Search Premier (ASP), ERIC, Web of Science (WoS), and SCOPUS23. A string search involves creating a set of search terms, for instance, related to metapragmatic awareness24, using the Boolean operator ‘OR’ (string 1 in the figure)25. Each string is first searched for separately, followed by a search in which the strings are combined using the Boolean operator ‘AND’. The search was limited to peer-reviewed studies published in English between the years 2000 and 2019. Limiting the scope to empirical studies published after the year 2000 is in line with Rose (2000, p. 56), who argued that “there is clearly a need to refine and develop this form of data elicitation via metapragmatic assessment – not yet done with child L2 learners”. The search resulted in 594 references (864 before excluding duplications). These were reviewed and references were excluded based on the following criteria: - Publications in languages other than English - Publications that were not empirical - Publications with participants younger than 5 and older than 13 - Publications not focusing on metapragmatic awareness 23 During the initial stage of the process, the researchers and a university librarian piloted a search individually before agreeing on a satisfactory string of search terms and sets (Booth et al., 2016). 24 As displayed in Figure 2.2, the string search for metapragmatic awareness included a wide range of search terms. The reasoning behind the wide range of terms was the inconsistent use of the term ‘metapragmatic awareness’ itself (e.g. McConachy, 2018). In other words, rather than solely retrieving studies explicitly using the term, the search aimed to retrieve research falling within the definition of the study, i.e. “the ability to verbalize reflections on linguistic forms, contextual features and/or their interplay” (Myrset & Savić, 2021, p. 165). 25 The search in the present study included a combination of proximity operators and truncation. Proximity operators specify that terms can be near each other rather than right next to each other (e.g. “young W1 learner” where W1 indicates that any one word can be between ‘young’ and ‘learner’). Truncation includes terms with various word endings (e.g. “child*” would include publications using words such as child and children). In addition, the searches were adapted for each database by including their thesaurus, or controlled vocabulary (Booth et al., 2016). Theory and previous research 29 Figure 2.2 Systematic review process (also presented in Article I) Considering that the review was conducted in 2019, more research has since added to the original knowledge. Thus, an additional stage aimed to retrieve relevant research published after the systematic review had been conducted. In this post-review stage, Google Scholar was used to retrieve relevant research. Furthermore, for the purposes of Article I, the systematic review focused specifically on the data elicitation techniques used in previous studies, and included both L1 and L2 research “due to the limited number of studies in L2 contexts” (Myrset & Savić, 2021, p. 166). However, studies conducted in L1 contexts may also indicate what learners are capable of understanding and provide reasonings for. More specifically, learners’ L1 and lived experiences may serve as a scaffold for L2 development (Chavarría & Bonany, 2006; McConachy, 2013, 2018; Savić & Myrset, Forthcoming-a). In other words, how learners come to understand pragmatics phenomena in their L1 may serve as a mediational tool when Theory and previous research 30 learning and communicating in the L2. Thus, both L1 and L2 research helped inform the instructional approach in the present study. For this reason, L1 studies are also included in Table 2.2, which presents the empirical research retrieved in the systematic review and in the post stage (marked in grey). Table 2.2: Previous research on YLLs’ metapragmatic awareness Author Context Age Focus Participants Adams et al. (2018) L1 English 6–11 Metapragmatic ability 82 Ben-Shlomo and Sela (2021) L1 Hebrew 4-10 Conversational violations 75 Bernicot et al. (2007) L1 French 6–10 Nonliteral forms (hints, idioms and implicatures) 60 Bosco et al. (2006) L1 Italian 3–8 Recognition and repair of communicative failures 80 Buson and Billiez (2013) L1 French 9–11 Stylistic variation 196 Caillies et al. (2012) L1 French 7–11 Irony comprehension 20 Chang (2016) L1 Mandarin Chinese 9–19 Apologies 240 Chang (2018) L1 Mandarin Chinese 9–19 Apologies 240 Collins et al. (2014) L1 English 6–11 Metapragmatic ability 88 Hsieh and Hsu (2010) L1 Mandarin Chinese 6, 9, adults Idiom comprehension 32 Ishihara (2013) Japanese EFL learners 9 Pragmatic awareness and metapragmatic judgments of formality and politeness 3 Ishihara and Chiba (2014) Japanese EFL learners 7-12 Pragmatic awareness and metapragmatic judgments of formality and politeness 5 Lacroix et al. (2010) L1 French 6–17 Idiom comprehension 57 Laval (2003) L1 French 6, 9, adults Idiom comprehension 48 Lee (2010) Cantonese EFL learners 7–12 Comprehension of direct and indirect speech acts 176a Lockton et al. (2016) L1 English 6–10 Conversational pragmatic ability and metapragmatic awareness 39 Savić (2021) Norwegian EFL learners 9-13 Co-construction of metapragmatic understandings in relation to requests 79 Theory and previous research 31 Savić and Myrset (Forthcoming-a) Norwegian EFL learners 9-13 Learners’ evaluations of request behaviours 79 Savić and Myrset (Forthcoming-b) Norwegian EFL learners 9-13 Metapragmatic appraisals of requests 79 Zhang and Yan (2012) Mandarin EFL learners 6 Sociopragmatic awareness regarding request and reply strategies 128 a In Lee’s (2010) study, a selection of the learners (n= 60) participated in a think-aloud and verbal protocol, which aimed to explore the learners’ processing strategies, i.e. their reasonings, during a comprehension exercise. 2.3.1.1 L1 metapragmatic awareness As displayed in Table 2.2, the majority of studies explore children’s L1 metapragmatic awareness26, with a selection of these exploring children with developmental disabilities (Adams et al., 2018; Caillies et al., 2012; Collins et al., 2014; Lacroix et al., 2010; Lockton et al., 2016)27. These studies have explored YLLs’ metapragmatic awareness related to non-literal forms, such as hints, idioms, and irony (e.g. Bernicot et al., 2007; Caillies et al., 2012), pragmatic violations (e.g. Adams et al., 2018; Ben-Shlomo & Sela, 2021), communicative failures (Bosco et al., 2006), and speech acts (Chang, 2016, 2018)28. The number of participants ranged from 20 (Caillies et al., 2012) to 240 (Chang, 2016, 2018). The majority of these studies employed some form 26 It is worth noting that the studies found in the review did not always use the label ‘metapragmatic awareness’. This lends support to scholars (e.g. McConachy, 2018; Nikula, 2002) who argue that the term is used inconsistently. Furthermore, Buson and Billiez (2013,p. 328) point to this inconsistency in definitions as a factor making it difficult to pin down exactly when children’s “metapragmatic competence” starts developing: “[i]f certain authors posit that metapragmatic activity exists from the age of 2 or 3, whereas others situate it at 8-10, this is perhaps because of varying definitions of what metapragmatic competence actually is”. 27 Due to the participants in focus, these will not be explored further. 28 Some of the L1 studies are cross-sectional studies which look at larger samples of learners and report on the statistics. Whereas the data elicitation included explanations in addition to the tasks, thus providing methodological insights for the systematic review, there is limited evidence that provides insights into knowledge and experiences that could inform understandings in the L2. For instance, Chang’s (2016, 2018) studies were conducted through a questionnaire in which the learners made judgements of the severity of an offence and their emotional response (whether they would feel embarrassed) through a Likert-scale response (Chang, 2016), followed by a suggested appropriate response in form of an apology. However, due to the vast number of participants, the study does not go into details about the learners’ responses. For this reason, these will not be elaborated on. Theory and previous research 32 of audio-visual stimulus followed by verbal explanations, that is, the researchers invited the learners to explain their reasonings for their choices. Studies focusing on YLLs’ L1 metapragmatic awareness show that learners are indeed capable of comprehending and explaining pragmatic phenomena from the age of six (e.g. Laval, 2003), with evidence of development with age (Ben-Shlomo & Sela, 2021; Bernicot et al., 2007; Hsieh & Hsu, 2010). For instance, learners aged 9-11 are capable of drawing on contextual cues when spontaneously commenting on stylistic variation, i.e. formality and informality, in utterances (Buson & Billiez, 2013). Indeed, Buson and Billiez found eight broad categories of interpretations. These categories ranged from focusing solely on the content of the utterance (Category 1) to referring to the “characteristics of the interaction […] [and] other criteria” (Categories 5-8) (Buson & Billiez, 2013, p. 335). In addition, the learners used evaluative statements, e.g. “speaks kind of badly”, “it doesn’t sound nice the way he talks”, and “rude words” (2013, pp. 333-335), as a frame when providing their explanations. In a study with L1 French learners, Bernicot et al. (2007) focused on indirect requests, i.e. hints, as one of the nonliteral forms. This study focused on whether learners could comprehend nonliteral language forms, and subsequently provide an explanation of the form. The study revealed that whilst most of the learners (aged 6, 8, and 10) could comprehend hints, the number of elaborate explanations was low (less than 7%). At the same time, there were very few irrelevant responses. Indeed, the majority resorted to explanations focusing on the “context or the utterance itself”, suggesting that the learners understood hints, but were not able to provide explanations for them. In addition, replication studies, such as Ben-Shlomo and Sela (2021) replicating the study by Collins et al. (2014), and Hsieh and Hsu (2010) replicating Laval (2003), have made it possible to observe similarities across L1s (English, French, Hebrew, Mandarin Chinese). In the latter replication study, Hsieh and Hsu (2010) explored Mandarin Chinese learners’ idiom comprehension through a word-card task and a picture selection task. The picture selection task involved the use of stories and pictures leading up to the use of an idiomatic utterance. The learners were then asked to choose between two pictures: one reflecting the idiomatic meaning, the other the literal meaning. The study revealed that a reliance on literal interpretations remained dominant, but that the youngest learners in the study (aged 6) had started to produce idiomatic answers when provided with a context. This reliance on the context was also clear when the learners provided their explanations. Adults focused more on the linguistic conventions in their explanations, whereas the younger learners (aged 6 and 9) relied mainly on the contextual aspects for providing their reasonings. These findings mirror to some extent Laval’s (2003) findings, where the youngest learners (aged 6) relied on the context to Theory and previous research 33 provide their explanations, whilst older learners (aged 9) started incorporating their linguistic knowledge about idiomatic meanings. Thus, there seems to be some correspondence in the results between the two studies in different contexts. Furthermore, Hsieh and Hsu (2010, p. 520) found that learners also used their “knowledge of the world” to make inferences about language use. Ben-Shlomo and Sela’s (2021) replication study focused on awareness of conversation rules and violations. More specifically, they explored the developmental trajectories of L1 Hebrew learners (aged 4-10) explaining pragmatic violations in videos. Four questions, including one in which the learners were asked to provide their reasoning, were asked to elicit metapragmatic understandings. The study revealed that the learners highlighted violations of non-verbal features, e.g. proxemics, and communicative collaboration, i.e. staying on topic. Furthermore, the researcher found a correlation between age and performance, suggesting a linear developmental trajectory with regard to L1 metapragmatic awareness. Furthermore, the study largely reflected the findings from Collins et al. (2014)29, thus providing “a first step in generalizing our understanding regarding metapragmatic development across languages and cultures” (Ben-Shlomo & Sela, 2021, p. 58). In other words, the ways in which YLLs come to understand pragmatic phenomena, and the resources that they employ, continue to develop from when the learners are roughly 5 to 13 years of age. 2.3.1.2 L2 metapragmatic awareness Moving to L2 contexts, few studies have explored YLLs’ metapragmatic awareness (see Table 2.2), with the cross-sectional studies conducted with EFL learners providing a mixed picture with regard to developmental patterns: Lee (2010) did not find any clear trajectories, whereas Savić and Myrset (Forthcoming-b) found some evidence of development. The following section first explores studies from other contexts before moving to the findings with Norwegian learners. In one study, Lee (2010) explored Cantonese EFL learners’ comprehension of direct and indirect speech acts in a cross-sectional study with learners aged 7, 9, and 12. The overarching study focused on responses to a multiple-choice comprehension exercise. However, 60 learners completed the exercise with a verbal protocol, in which the learners explained their choices during the exercise. This enabled Lee to explore their processing strategies. The study revealed that the majority of the learners attended to the semantics of utterances, e.g. creating a relationship between cause and result. 29 The study by Collins et al. (2014, p. 31) compared learners with a communication impairment and children with “typical language development”. The latter was used as data for comparison in the study by Ben-Shlomo and Sela (2021). Theory and previous research 34 However, to various extents, the learners also attended to sociopragmatic features, such as speakers’ intentions and feelings, and the context in which the utterance was made. Furthermore, the learners compared the L1 and L2, and employed their world knowledge to provide their reasonings. Interestingly, there were no clear developmental trajectories in the processing strategies, which Lee (2010, p. 363) assigns to “unknown socio-cultural factors such as school instruction”. However, the study revealed that YLLs draw on various frames of reference to make sense of pragmatic phenomena. In anoher study, Zhang and Yan (2012) explored the impact of immersion programs on YLLs’, aged 6, sociopragmatic competence. The study was based around request and response scenarios and sought to explore the learners’ understandings of these. The study included three questions for each request scenario. The first question aimed to explore the learners’ comprehension of the speaker’s intentions, e.g. “What does Xiaogang expect his peer to do by saying, ‘Could you pick up the teddy bear on the floor for me?’” (2012, p. 38). The second question focused on the learners’ ability to compare the appropriateness of two requests with the same intentions. Finally, drawing on the comparison in the second question, the learners were asked to explain their reasoning for their choice, i.e. “Why do you think the former (latter) is a better choice for Xiao Gang (Xiao Hong)?” (2012, p. 45). The study revealed that the learners in immersion programs had higher levels of sociopragmatic awareness in English as opposed to the non-immersion group. These findings were significant in relation to requests, but not responses. The authors ascribe the findings to L1 transfer, that is that the norms in the two languages are similar in responses as opposed to requests. Consequently, the significant difference in relation to requests, the authors argue, was a result of the communicative approach in immersion programs, in which the learners used the language in a variety of authentic contexts. It is important to note, however, that this study was tightly controlled, i.e. focusing on “distinguishing the polite request forms from the impolite ones” (Zhang & Yan, 2012, p. 42), with the learners providing explanations for their choices. While the data presented through statistics revealed that a larger number of learners in the immersion program provided appropriate explanations for their choices, there is little detail about the content of the reflections themselves. Thus, it is not possible to deduce how the learners provided their reasonings or the frames of reference they invoked to provide these. Rather, the study reveals that the learners in the immersion program, as opposed to those in the non-immersion program, employed a wider range of aspects, i.e. intentions, appropriateness, and reasoning, to make sense of requests. Theory and previous research 35 In a third study outside of Norway, Ishihara (2013)30 focused on three YLLs, aged 9, in an instructional study (see also Section 2.3.2). The study revealed that the learners spontaneously identified non-verbal cues, such as the lowering of a hat, and made judgements about the situational formality using a formality judgement scale. Furthermore, comments such as “[i]s it rude language” (Ishihara, 2013, p. 142) show that the learners resorted to valency to question the appropriateness of utterances. The learners’ L1 was also used, for instance to translate English requests, and to scaffold understandings about the interplay between pragmalinguistics and sociopragmatics. In the Norwegian context, one project has investigated EFL learners in 3rd, 5th, and 7th grade, aged roughly 9, 11, and 13 respectively (Savić, 2021; Savić & Myrset, Forthcoming-a, Forthcoming-b). This project aimed to explore learners’ (n= 79) metapragmatic awareness related to EFL requests through verbalised learner reflections. Studies from this project have revealed that the learners drew on various frames of reference, co-constructed ideas, and drew on pragmalinguistic and sociopragmatic features to provide their understandings of requests and pragmatic behaviours. In Savić and Myrset (Forthcoming-b), the learners provided reasonings for their appraisals of requests as appropriate or inappropriate. In this study, conventionally indirect requests were appraised more positively than direct requests and hints, especially by younger learners. Indeed, when asked how those requests appraised less positively could be improved, a common tendency was to change the requests into conventionally indirect requests or adding the marker ‘please’, which was also often highlighted as a reason for positive evaluations. Overall, the study revealed that the learners drew on a range of pragmalinguistic features, such as word choice and using supportive moves, and sociopragmatic considerations, such as interlocutor age and familiarity, or their interplay. An interesting finding from this study was that although learners in all grades produced hints, e.g. “I like the little kite” (3rd grade), they displayed uncertainties when appraising such requests. This was reflected in both the appraisals and their reasonings in the subsequent discussions, which Savić and Myrset argue could be an indication of the learners having difficulties determining the communicative intent, with similar findings among learners in L1 contexts (e.g. Bernicot et al., 2007). Savić and Myrset (Forthcoming-a) explored the interpretative frames that the learners drew on, and their evaluative stances, when discussing pragmatic practices. The study revealed a range of interpretative frames that the learners employed to make sense of 30 Ishihara’s study and Ishihara and Chiba’s (2014) were instructional studies. These are thus also presented in Section 2.3.2 Theory and previous research 36 pragmatic phenomena. These frames included the perceived feelings of the hearer, stereotypes about L1 and L2 and their speakers, lived experiences, and historical and cultural knowledge. For instance, one group of learners compared and contrasted L1 and L2 politeness, arguing that the English are more polite than Norwegians – a commonly held stereotype. Savić and Myrset (Forthcoming-a) argue that the learners’ various frames of reference and positive evaluative stances towards the L2 and its speakers highlight a potential for challenging stereotypes and nuancing the view of certain linguistic resources as inherently polite in language teaching, suggesting that the primary level is a useful time to start focusing on teaching pragmatics. Finally, Savić (2021) investigated the learners’ co-construction of metapragmatic understandings, drawing specifically on data generated from the discussions in 3rd and 5th grade. The study revealed that learners used a variety of discursive devices to co-construct metapragmatic understandings. These included expanding on each other’s utterances, grounding their positionings through personal and hypothetical stories, invoking perceived thoughts and utterances by other people (i.e. hypothetical quotes), and taking on various perspectives to deduce an argument (i.e. heterogeneous positionings). Savić (2021) argues that the dynamics of collaborative dialogue in the research setting, in which the learners formed ‘thinking societies’ for the creation of ideas, reveal a potential for using peer discussion for teaching pragmatics, where the teacher has a crucial role for the creation and scaffolding of metapragmatic discussion. 2.3.1.3 A summary of young language learners’ metapragmatic awareness In sum, research has identified some developmental patterns with regard to YLLs’ metapragmatic awareness, both in L1 (e.g. Ben-Shlomo & Sela, 2021; Hsieh & Hsu, 2010) and L2 contexts (Savić & Myrset, Forthcoming-b). Furthermore, learners have been found to draw on a range of frames of reference to facilitate the expression of their metapragmatic understandings. These include comparisons between L1 and L2, lived experiences, evaluations (e.g. valency), and world knowledge (e.g. Ishihara, 2013; Savić & Myrset, Forthcoming-a; Savić & Myrset, Forthcoming-b). Furthermore, the studies reveal that the learners attend to both pragmalinguistic features, such as word choice, semantics, and content of utterances (e.g. Buson & Billiez, 2013; Lee, 2010; Savić & Myrset, Forthcoming-b), and sociopragmatics, such as perceived feelings of the interlocutor, situation, and familiarity (Lee, 2010; Savić & Myrset, Forthcoming-a, Forthcoming-b). Research from the Norwegian context reveals a rather complex picture related to YLLs’ metapragmatic awareness, which helped inform the present study and the choices related to the instruction. First and foremost, learners aged 9, 11 and 13 were found to Theory and previous research 37 pay attention to both pragmalinguistic and sociopragmatic aspects when discussing the (in)appropriateness of requests (Savić & Myrset, Forthcoming-b); thus, the instruction in the present study aimed to focus attention on both the pragmalinguistic and sociopragmatic dimension. Furthermore, in relation to making sense of pragmatic phenomena, the learners use various frames of reference, such as lived experience, hypothetical stories, cultural and historical knowledge, and stereotypes (Savić, 2021; Savić & Myrset, Forthcoming-a). As a result, the instruction aimed to probe reflections in which the learners could employ their previous experiences and knowledge as a scaffold. Finally, learners actively co-construct metapragmatic understandings through collaborative dialogue (Savić, 2021), which was incorporated in the instruction through pair, group and whole-class activities and discussions, scaffolded in various ways. Thus, the resources YLLs bring to the learning situation were utilised in the current study as a springboard for teaching L2 pragmatics and fostering agency, which is considered of vital importance within sociocultural theory (SCT). Furthermore the way in which learners co-construct metapragmatic understandings in groups suggests that collaborative dialogue and instruction informed by SCT (Vygotsky, 1934/2012, 1978) may be appropriate for teaching pragmatics with YLLs. Consequently, an SCT-informed approach to instruction was selected for the present study, which in addition to broadening the learners’ pragmalinguistic repertoire, aimed to develop their metapragmatic awareness through reflections. 2.3.2 Pragmatics instruction with young learners When it comes to pragmatics instruction with young learners (aged 5-13), there is still a limited pool of research. For instance, recent meta-analyses (Plonsky & Zhuang, 2019; Taguchi, 2015) have found that explicit instruction is more favourable, but this claim is largely based on findings from studies with (young) adults (e.g. Alcón Soler, 2005; Halenko & Jones, 2011). The following review departed from the findings in Plonsky and Zhuang’s (2019, p. 291) meta-analysis, which provided a rigorous and exhaustive literature search, comprising “four library-housed databases (ERIC, LLBA, PsycINFO, ProQuest Dissertations and Theses) and two non-library databases (Google and Google Scholar)”. In these databases, Plonsky and Zhuang used various string searches to explore both broad terms, such as ‘pragmalinguistics’ and ‘teaching’, and narrow terms, such as ‘request’ and ‘refusal’. In addition, they used ancestry searches through investigating references from previous reviews, and forward searches, through exploring articles that had cited “seminal review papers and existing meta-analyses on pragmatics instruction” (2019, p. 291). Importantly, in open science attempts to ensure transparency, the authors provided the references retrieved and the coding schemes Theory and previous research 38 available for download31. Consequently, this provided a rigorous point of departure for reviewing relevant literature for the current doctoral research project. Following this review, the author conducted searches to further explore relevant literature published after 2016, the year Plonsky and Zhuang’s review was conducted. This process employed string searches combining Boolean operators, e.g. “AND” or “OR”, truncation, e.g. “learner*”, and categories of search terms. The overarching categories of the search terms were: 1) young language learners, 2) pragmatics instruction, and 3) target language (English). Combined, these categories generated the following search: 1) child* OR "young learner*" OR "young language learner*" AND 2) "pragmatic* instruction" OR (pragmalinguistic* AND instruction) OR (sociopragmatic* AND instruction) AND 3) English OR ESL OR EFL OR EAL OR "English as a second language" OR "English as a foreign language” OR “English as an additional language” Whereas the search confirmed that instructional pragmatic studies focusing on young learners are extremely sparse, some studies investigating the teachability of pragmatics with these age groups have been identified, thus providing insights into both the approaches and targets of instruction (see Table 2.3 for an overview). These studies are discussed in this section32. 31 Available at https://www.iris-database.org/ (Accessed, 19.04.2021) 32 Whilst the current study focused on requests, due to the paucity of research with young learners, studies focusing on other pragmatic targets are included. Furthermore, some studies focused on groups of participants where only some were within the age group defined herein as young language learners. These have also been included. The review is limited to those that focused on L2 English as target. Consequently, studies such as Lyster (1994), which explored French, have been excluded. Theory and previous research 39 Table 2.3: Pragmatics instruction studies with young language learners Methods Pictorial DCTa Pre-post FJT, DCT, SVDCTb FJT, DCT, SVDCT DCT Pre-post DCT Think-alouds Pre-post-delayed a Discourse Completion Task (DCT) b Formality Judgment Task (FJT), Student-generated Visual DCT (SVDCT). Instruction Short Dialogic (constructivism) Very short Dialogic (SCT) Short Dialogic (SCT) Half-semester (5 weeks) Explicit Very short Explicit Collaborative vs individual + control Participants N= 38 3-6 years N= 3 (all male), 9 years N= 5 7-12 years n=24 ( all male), 12-18 n=74 (all female), 13-14 years Target Requests, Thanking Formality: Requests, greetings Formality: Requests, greetings Refusals Requests Context Iran, EFL Japan, EFL China, Hong Kong, EFL (Japanese learners) Iran, EFL South Korea, EFL Author Alemi and Haeri (2020) Ishihara (2013) Ishihara and Chiba (2014) Sa'd and Gholami (2017) Taguchi and Kim (2016) As displayed in Table 2.3, very few studies have investigated the teachability of pragmatics with young learners. The majority of the studies included requests as a pragmatic target, with Sa’d and Gholami (2017) being the only study focusing on a different target (refusals). With regard to the instructional approach, following Bardovi-Harlig (2015), the length of the studies ranged from very short (Ishihara, 2013; Taguchi & Kim, 2016), short (Alemi & Haeri, 2020; Ishihara & Chiba, 2014), to half-semester long (Sa'd & Gholami, 2017). There are variations in materials used in the various studies, and the studies seem to align with two paradigms of pragmatics research Theory and previous research 40 (McConachy & Spencer-Oatey, 2020): interlanguage pragmatics (ILP) (Sa’d & Gholami, 2017; Taguchi & Kim, 2016) of explicit and implicit instruction (in line with Kasper's (2001) definitions), and dialogic approaches rooted in SCT (Ishihara, 2013; Ishihara & Chiba, 2014)33, in which the learners co-constructed knowledge through dialogue with the teacher as mediator. All the studies report on the benefits of pragmatics instruction, albeit to a lesser or greater extent. Although the studies focused on different age groups and pragmatic targets for instruction, some trends in the testing approaches can be found within the different studies. All the studies used a written discourse completion task (DCT) as a technique for testing the learners' language production. In addition, most studies employed a form of oral response, through think-alouds (Taguchi & Kim, 2016), and classroom interaction (Ishihara, 2013; Ishihara & Chiba, 2014). Furthermore, a formality judgment task was employed in two studies (Ishihara, 2013; Ishihara & Chiba, 2014). Although all the studies report on language production through a DCT, they vary considerably in both how the data were analysed and their results. In pre-post designs, Taguchi and Kim (2016) and Sa'd and Gholami (2017) analysed the responses based on the frequencies of use of linguistic resources. However, only Taguchi and Kim (2016) tested the long-term effect of instruction, concluding that the retention was only apparent in some linguistic categories. Their study revealed a short-term effect on head act scores, with both treatment groups (individual and collaborative) outperforming a control group in the post-test. In addition, the frequencies of internal (amplifiers and hedges) and external (grounders and preparators) modifications revealed "a mixed picture" (p. 429). Amplifiers and hedges were used to a limited extent, though still revealing an instructional effect with both treatment groups. However, this effect was not long-lasting and disappeared in the delayed post-test. For external modification, the test revealed a high use of grounders in all groups, suggesting that the learners were 33 Alemi and Haeri (2020) focus on the learners’ ability to produce utterances relevant for a specific situation, e.g. asking to go to the bathroom, and through linking an action with an utterance, e.g. waving for goodbye. The authors attribute their approach to discovery learning and learning through experience. They explored development through Robot-Assisted Language Learning (RALL), in which a humanoid robot served as a teaching assistant with half the group of learners. In the RALL group, the robot engaged in conversations and performed actions together with the teacher, and the learners practised with the robot. The learners in the non-RALL group received the same treatment, without the robot. Both groups were tested with a pictorial DCT, i.e. provided with a pictured and asked to produce the correct speech act. Not being able to produce the correct speech act was viewed as the child having acquired “neither comprehension nor production”. The learners in the RALL-group outperformed the learners in the non-RALL group. Considering, the differences in age from the present study, and the way in which the learners were tested, this study will not be presented beyond this. Theory and previous research 41 already familiar with this strategy. Preparators, on the other hand, were used more frequently following the instruction, and were the only modification that showed signs of longer-term retention, with both groups outperforming the control in the delayed post-test. Consequently, the study revealed some positive effects of instruction with no significant differences between individual and collaborative learning in terms of production, but the length of the instruction most likely did not provide "enough time to boost their robust learning" (Taguchi & Kim, 2016, p. 434). Focusing specifically on SCT-informed pragmatics instruction, Ishihara (2013) and Ishihara and Chiba (2014), whose samples comprised fewer learners, three and five participants respectively (see Table 2.3), report on the learners’ production through the DCTs and observations. In these two studies, there seem to be conflicting results. In Ishihara’s (2013) study the learners were able to produce pragmatic formulas, although they relied on external scaffolding from the teacher to do so. Ishihara argues that this could indicate that the learners had not yet internalised the formulas, which were thus not readily available in production. In Ishihara and Chiba (2014, p. 97), on the other hand, the oldest learners were able to produce and vary between the pragmatic targets (“Can you pass the X, please?; Could you pass the X, please?; Can I have the X, please?; and May I have the X, please?”). Considering that these studies were similar in the teaching approaches, the reported differences could be related to the learners’ age (9 vs 12 years), duration of instruction (120 vs 180 minutes), or individual learner differences34. However, the small samples render it impossible to reach firm conclusions or generalisations. When it comes to sociopragmatic aspects, Ishihara (2013) and Ishihara and Chiba (2014) used visual aids as support. For instance, to facilitate the learners making evaluations about politeness, the terms ‘polite’ and ‘impolite’ were presented on a continuum on which learners could indicate their perceptions through pointing or marking. Thus, this could be interpreted in such a way that the instruction facilitated evaluations through valency (e.g. Kádár & Haugh, 2013; Spencer-Oatey & Kádár, 2021). In the study, politeness was linked to the pragmatic target, e.g. “the levels of politeness and formality of the target expressions introduced in the instruction were often closely intertwined” (Ishihara & Chiba, 2014, p. 91). However, in the two studies, less emphasis was placed on theoretical constructs related to politeness, suggesting that the instruction focused on learners relying on lived experiences, i.e. their own perceptions, and (co-)constructing understandings with their peers and the teacher through dialogue. In Ishihara’s study, the learners became increasingly attuned to sociopragmatic dimensions, such as making evaluative judgments about 34 Ishihara (2013) argues that the learners’ difficulties in producing target formulas could be related to limited exposure to the language. Theory and previous research 42 appropriateness and situational formality, and visual cues. Comments related to the latter appeared both spontaneously and directed by the teacher. Thus, the development occurred through collaboration between the teacher and the learners. Ishihara and Chiba (2014, p. 15) also observed such “interactive and collaborative meaning-making”, in which peers co-constructed their ideas. Furthermore, non-verbal responses through drawings in the SVDCT revealed understandings about sociopragmatic features, such as age, gender, and physical distance. Thus, the two studies revealed both pragmalinguistic and sociopragmatic development to various extents. Furthermore, these studies seems to align with instruction favouring the development of agency (e.g. Morollón Martí, Forthcoming), in which learners developed their own understandings, as opposed to teaching rules of thumb. As revealed through the review, pragmatics instruction with young learners is an under-explored area of research. Previous research suggests that there is potential for teaching pragmatics with YLLs. However, due to the discrepancies in instructional approaches, pragmatic targets, length, and design, more studies are necessary. Consequently, the field of pragmatics still needs to “determine what type of instruction may be more compatible with how they generally learn and what aspects of pragmatics may be beneficial to teach them” (Ishihara, 2010, p. 946). Studies investigating young learners’ pragmatic development (Achiba, 2003; Ellis, 1992; Rose, 2000; Savić, 2015; Savić et al., 2021) and the pragmatics-related content in EFL language textbooks (Jakupčević & Portolan Ćavar, 2021; Limberg, 2016; Schauer, 2019) give reason to believe that young language learners receive some form of pragmatics input, implicitly or explicitly. At the same time, pragmatic phenomena are still largely overlooked or disregarded in the YLL language classroom (Glaser, 2018). There are thus vast knowledge gaps concerning effective ways of teaching pragmatics, as well as the pragmatic targets that should be taught with these age groups. 2.4 Concluding remarks This chapter has aimed to conceptualise the theoretical underpinnings of the present research project and demonstrate how they have informed the present study. Requests have been chosen as the pragmatic target due to their early occurrence in language development, their frequency of occurrence in speech, and a range of strategies and pragmalinguistic resources through which they can be realized. They were addressed in the classroom through an SCT-informed concept-based approach, which aimed to foster agency mediated through scientific concepts and the learners’ metapragmatic awareness. Previous research using concept-based approaches was presented through a literature review. This review revealed that whereas such approaches have gained traction in instructional pragmatics, prior research has focused on adult learners, mainly Theory and previous research 43 in university settings. In addition, the majority of these studies have focused on other languages and pragmatic targets than those explored in the present study. Thus, whilst previous studies using concept-based approaches helped inform the instruction in the present project, they did not provide insights in relation to YLLs. Consequently, two more reviews were conducted, both of which confirmed that YLLs remain largely under-explored within pragmatics. The first review, focusing on YLLs’ metapragmatic awareness, aimed to provide insights into the frames of reference and strategies that learners employ when expressing their understandings, and the pragmatic topics that they can discuss. In line with an SCT perspective, this review focused on both L1 and L2 studies, as L1 lived experiences may serve as a scaffold in SCT. This review revealed that learners draw on, for instance, their world knowledge, L1 and L2 differences, and speaker intentions and feelings to make sense of various pragmatic phenomena (e.g. nonliteral forms, speech acts, and formality). Thus, the research suggests a potential for (meta)pragmatic learning through fostering reflection. The second review related to YLLs focused on previous instructional pragmatics research, specifically that focusing on English as a target language. This review revealed a paucity of research, with the majority of studies focusing on requests as a pragmatic target. However, these instructional studies vary considerably in the teaching approaches. The previous research on pragmatics instruction with YLLs reveal that pragmatics is indeed teachable with this group of learners, but questions remain unanswered about the affordances of various approaches. However, two studies informed by SCT reveal a potential for grounding pragmatics instruction in SCT to facilitate pragmatic development. Thus, the successful implementation of dialogic approaches in previous research and the findings on YLLs’ metapragmatic awareness suggest that concept-based approaches aiming to foster agency may have a potential with YLLs, which was the focus of the present study. Against this backdrop, I turn to the present study, which aimed to explore the impact of concept-based approaches for teaching pragmatics with YLLs in relation to their pragmalinguistic development, their metapragmatic awareness, and their engagement with pragmatics. Theory and previous research 44 Methodology 45 3 Methodology This chapter provides an overview of the methodological approaches, considerations, and choices for data collection, in addition to the detailed description of the instruction carried out in this project. The case study adopted a mixed methods design. Section 3.1. provides an overview of the research design and its coherence. Since the aim of the study was to explore the impact of concept-based approaches for teaching pragmatics to young language learners (YLLs) of English, Section 3.2 provides an overview of the instruction the learners received. In the current study, data was collected prior to and following the instruction, and included an adapted video-prompted oral DCT, group interviews, and Readers Theatre. Section 3.3 provides a detailed account of these techniques and the data analyses. Section 3.4 presents the quality assessment of the study, that is, the degree to which it is scientifically sound. Section 3.5 presents topics related to researcher reflexivity. Finally, the ethical considerations related to the study are addressed in Section 3.6. The current project aimed to investigate the impact of EFL pragmatics instruction in a Norwegian primary school, with learners in 7th grade (aged 12-13). Thus, the methodological considerations in the present study were largely informed by literature on research with children (see for instance Christensen & James, 2017; Eckhoff, 2019; Pinter & Kuchah, 2021), or the ‘New Sociology of Childhood’. This debate about YLLs participation and involvement in research has been heavily influenced by the United Nation’s Convention on the Rights of the Child (UNCRC, 1989), which has led to discussions about participatory research with children (e.g. Shamrova & Cummings, 2017), the varying degrees to which they are included in the research (e.g. Fielding, 2001; Hart, 1992; Lansdown, 2005), and their influence in research (e.g. Kellett, 2010; Lundy, 2007). However, within applied linguistics research, children have traditionally been “objects of research” and Pinter (2019, p. 421) calls for broadening the scope “to include children's own views, perspectives and experiences”. Consequently, this chapter draws on relevant literature beyond the field of applied linguistics when relevant. 3.1 Overview of research design The present study aimed to answer the following overarching research question: How does a concept-based approach to teaching requests impact young language learners’ request production and awareness, and their engagement with pragmatics? The overarching research question has since been divided into aims and research questions Methodology 46 to explore this question from various perspectives through four articles (see Chapter 4 for summaries). 3.1.1 Case study Considering that the present study concerns an in-depth exploration of the impact of instruction, the project is in essence a case study. Although there are some ambiguities with regard to the definition of the case study (Nunan, 1992; Simons, 2009; Starman, 2013; Swanborn, 2010; Verschuren, 2003), there seems to be a consensus that it entails the study of “an individual, a family, a program, a nation, or another structure or entity” (Duff, 2020, p. 145). Furthermore, it usually entails investigating a phenomenon over time (Lew et al., 2018). Thus, the case study offers explorations of a phenomenon from various perspectives, and includes generating vivid and detailed insights into the object of research (Lew et al., 2018; Swanborn, 2010). It is perhaps for this reason that case studies are frequently used in education and applied linguistics (Duff, 2020; Lew et al., 2018; Miles, 2015; Starman, 2013; Verschuren, 2003), which strive for in-depth explorations into complex phenomena such as language learning and teaching. However, a challenge with the case study, which is also evident in the various definitions (see Simons, 2009; Verschuren, 2003 for comprehensive presentations of definitions), is the differing views of whether it is a research method or whether it serves as a frame for the design (Miles, 2015; Starman, 2013). The view of the case study as a method could perhaps help explain the criticism towards its scientific rigour. Due to the in-depth focus on what can be a single participant (Lew et al., 2018), or “the ‘N of 1’ problem” (Stoecker, 1991, p. 91), the case study has been accused “of being at the nonrigorous and nonscientific end of the research spectrum” (Mukhija, 2010, p. 418). However, the criticism is generally grounded in the quantitative paradigm of research, in which scientific rigour is commonly tested against criteria of external/internal validity and reliability (Dörnyei, 2007; Miles, 2015; Mukhija, 2010; Stoecker, 1991; Swanborn, 2010), for instance, the generalisability of the research which “is often categorically dismissed with case study” (Duff, 2020, p. 150). Although some argue for the possibility of generalisations within case study research (see Starman, 2013 for a discussion of various views), generalisations to larger populations may not be the aim of the research. Rather, the aim is to gain in-depth insights into processes occurring within the case. Another concern raised about case studies relates to researcher biases (Mukhija, 2010; Starman, 2013; Stoecker, 1991), that is, the researcher imposing their own views (subjectivity) on the data. This is a particularly relevant critique in “the traditional case study” which is “conducted most Methodology 47 often by a single researcher” (Duff, 2020, p. 145). With this in mind, the common consensus is that case study research requires rigid analyses, heightened self-awareness, and detailed descriptions to maintain objectivity (Simons, 2009; Starman, 2013; Swanborn, 2010). Given the general aim of case study research and the role of the researcher within it, some of the critique about the validity and reliability of case studies may be viewed as somewhat unfounded. Thus, rather than assessing the scientific rigour through a quantitative lens, other forms of quality assessment may be more appropriate. However, the criticism highlights a concern for case studies, and should be taken into account through critically assessing the quality of research (see Section 3.4). Considering the ambiguities regarding the definition and the purposes of the case study, the present study aligns with Simons (2009, p. 21), who holds that a [c]ase study is an in-depth exploration from multiple perspectives of the complexity and uniqueness of a particular project, policy, institution, programme or system in a ‘real life’ context. It is research-based, inclusive of different methods and is evidence-led. The primary purpose is to generate in-depth understanding of a specific topic (as in a thesis), programme, policy, institution or system to generate knowledge and/or inform policy development, professional practice and civil or community action. Within this view, the case study is viewed as a form of inquiry, that is, a “selection of what will be explored” (Starman, 2013, p. 32), rather than being limited to a research method (usually qualitative) (see also Miles, 2015). In the context of the present study, for instance, the researcher aimed to explore the impact of instruction, which forms the case, through prolonged engagement. The case study, thus, provided a framework for investigating the instruction from various perspectives to gain in-depth insights into the learners’ request production and awareness, and their engagement with pragmatics. Furthermore, as several authors posit (Duff, 2020; Simons, 2009; Starman, 2013; Swanborn, 2010), a case study may include the use of a range of methods, such as surveys and interviews. Following this, the fieldwork of the present study lasted for approximately three months, and included various forms of data collection prior to and following the instruction. Figure 3.1 provides an overview of the data collection. Methodology 48 Figure 3.1: Overview of the fieldwork with the techniques used and the data they elicited. The “A” followed by a roman numeral (e.g. A-I) refers to the article in which the data was presented. As displayed in Figure 3.1, the data was collected at different points of time in relation to the instruction and with the purpose of eliciting different types of data. The data generated was used in the four articles comprising this thesis (Articles I-IV). The video-prompted oral discourse completion task (VODCT, see Section 3.3.1), used in Article II, aimed to elicit requests in a pre-, post-, and delayed post-test. With aims to explore the learners’ metapragmatic awareness, the scripts and requests produced in the second cycle of Readers Theatre (Section 3.3.2) were used as a discussion prompt in the group interview, presented in Article I35. This group interview also employed appraisal tasks which aimed to prompt metapragmatic discussions presented in Article III, and perceptions about the project in Article IV (Section 3.3.3). The approach of eliciting data through various techniques is in line with the case study approach, where researchers use “a combination of data generation processes, such as tests, questionnaires, [and] interviews” (Duff, 2020, p. 147). Thus, the case study lends itself 35 The project as a whole generated a large pool of data. This is a common feature in case studies and qualitative research (e.g. Duff, 2020; Lew et al. 2018). However, considering research quality, it is also important to provide rich descriptions of the data (see also Section 3.4). Thus, researchers should be reflexive about the intersection between data collection and analysis (Tracy, 2010). In the present study, the articles present rich descriptions of the data generated and the analysis. Consequently, the data generated in the second interview was not used in the articles presented in this thesis, but will rather be used for future publications. With this in mind, the second interview will not be elaborated on in this thesis. Methodology 49 to the use of mixed methods (Simons, 2009; Swanborn, 2010), which was also the approach employed in the present study. 3.1.2 Mixed methods approach In the current project, the data was elicited concurrently. The analyses in the project – statistical analyses of requests (Article II) and qualitative analyses of interviews (Articles I, III, and IV) – entailed a mixing of methods with both a qualitative and a quantitative strand. This study is thus positioned on a continuum between the two strands, thus drawing on the strengths and reducing the weaknesses from both strands (Gobo, 2008; Johnson & Onwuegbuzie, 2004). Mixing methods was considered the optimal approach as its interactive nature between methods helps to explain phenomena, which in effect may enhance the integrity and credibility of the findings (Biesta, 2012; Creswell & Plano Clark, 2011; Dörnyei, 2007). Opting for mixed methods corresponds well with trends in applied linguistics and educational research, where combining the two strands has become increasingly popular and influential (Johnson & Onwuegbuzie, 2004; Phakiti et al., 2018). The mixing in this project was a concurrent triangulation, or QUAL+quan (Dörnyei, 2007), in which the data was collected at the same time (see Figure 3.1), but with the weighting placed on the qualitative strand (Jones et al., 2006; Mackey & Bryfonski, 2018). 3.1.3 Project coherence Considering the use of mixed methods and the case study framing, the present study is positioned within pragmatism. Case studies are largely driven by research questions and action (Duff, 2020; Starman, 2013), in which researchers find appropriate techniques to answer their research questions (Simons, 2009), evidenced in the range of methods employed (e.g. Lew et al., 2018; Swanborn, 2010). This also aligns with Gobo (2008), who calls for a less categorical distinction between quantitative and qualitative strands by rather foregrounding components such as cognitive modes, response types, and research techniques. The critical selection of various techniques is also relevant for research with children (e.g. Griffin, 2019; Punch, 2002b). For instance, drawing on Gobo (2008), selecting techniques that engage different response modes, may be of particular value in research with children. This may include taking into account children’s cognitive growth (McKay, 2006), through incorporating mixed methods, and enabling different response types, e.g. oral, written, or kinaesthetic, aimed at increasing participation (O'Kane, 2017; Pinter, 2014). Furthermore, combining tasks with interviews may enable children Methodology50to feel comfortable in the research setting, ensure multimodal responses, and give learners control over their responses (Punch, 2002b). Such considerations were also incorporated in the present study (see Section 3.3), for instance, combining tasks and interviews and using videos (Yamada-Rice, 2017). At the core of such considerations, however, is the focus of inquiry, namely, the techniques were selected on the basis of what was being explored, that is, techniques driven by the research question. Figure 3.2 provides an overview of the project’s coherence. Figure 3.2: Coherence of projectAs displayed in Figure 3.2, pragmatism as an ontological foundation was considered a useful positioning as “it offers a practical and outcome-oriented method of inquiry that is based on action […] and it offers a method for selecting methodological mixes that can help researchers better answer many of their research questions” (Johnson & Onwuegbuzie, 2004, p. 17). As such, pragmatism can be seen as a view in which the inquiries and research questions dictate the choices of data elicitation techniques and analysis. The epistemological positioning in this thesis is social constructivism(Hacking, 1999), and views language as constructed by social actors through social interaction, which is reflected both in the socio-cultural approach to the instruction and in the data elicitation techniques relying heavily on interaction and co-construction of meaning among peers. For instance, the current study aimed to explore both the learners’ language production and their reflections, requiring different forms of data elicitation techniques (e.g. language production tasks and interviews) and analyses (statistics and discourse analysis). Consequently, this called for using techniques and analyses belonging to different methodological strands. Analysis• Descriptive statistics• Inferential statistics• Content analysis• Discourse analysisTechniques• VODCT• Group interviews• Appraisal tasks• Script writingMethodology• Mixed methodsEpistemology• Social constructivism• Socio-cultural theoryOntology• PragmatismMethodology 51 3.1.4 Sample Conducting research in educational settings, such as a primary school, can be challenging. Dörnyei (2007, p. 188) points out that teachers can often be very busy, making it difficult to “bring them on board, and it is a real challenge in almost every case to keep up their commitment”. In this study, the teachers would play a significant role in facilitating the project by organising the learners into groups and allowing the researcher to conduct research during regular classes. Thus, the project employed two sampling strategies: convenience sampling and homogenous sampling. In other words, in order to find teachers that would be willing to participate for the duration of the project, the researcher used his network of schools to come into contact with teachers. Once the teachers were recruited, the participants were selected from a specific entity, namely 7th graders, which Dörnyei (2007, p. 127) refers to as homogenous sampling, allowing the researcher to “conduct an in-depth analysis to identify common patterns in a group with similar characteristics”. This sampling strategy resulted in participants from two intact classes (Bardovi–Harlig, 2015) of 51 learners, comprising one class of 25 learners (14 girls and 11 boys) and one of 26 learners (12 girls and 14 boys). Of these, all but one agreed to participate in the study. This learner was involved in the instruction, but did not participate in the data collection. Choosing not to participate did not put the learner at a disadvantage (Mayo, 2021)36. Instead, the researcher made the decision not to audio record the learners during the lessons to ensure that they could all participate equally. All the learners were born in Norway and were, thus, fluent in Norwegian. As regards their English proficiency, the learners’ expected level, according to the CEFR, was in the range of A2-B1 (Hasselgreen, 2005). For the data collection, the learners were divided into groups of 4-5, so-called ‘friendship groups’ (Pinter, 2014), in attempts to reduce the researcher-child power imbalance (Griffin, 2019; Gu et al., 2005). The criterion for organising them into groups was that the learners should be able to work well together rather than being at similar levels of proficiency. As a result, the groups were mixed ability groups. These groups remained permanent for the duration of the data collection. Organising the learners into groups resulted in a total of 12 groups (6 per class). However, one learner had lived abroad in an English-speaking country, and in order to avoid this learner’s background influencing the data generated, this group was not included in the data analysis. Instead, this learner’s group (four learners) served as a pilot group for the duration of data 36 Indeed, this was explicitly mentioned in the consent form (Appendix 15) and relates to ethical considerations of not doing harm (see also Section 3.6). Methodology 52 collection to ensure that the tasks and questions were appropriate37. As a result of one learner opting out and four learners serving as a pilot group, the data presented in the articles were generated from 11 groups (n= 46 learners). 3.2 Instruction The instruction introduced requests as a pragmatic target with a focus on promoting agency, that is, making informed choices in language use. The researcher taught the material38. The primary language during the instruction was English, but Norwegian (L1) served as a scaffold for meaning-making, for instance, when prompting learner reflections (e.g. Chavarría & Bonany, 2006; McConachy, 2018; Zuckerman, 2004). In order to promote agency, the instruction had the following aims: 1) introduce the pragmalinguistic dimension, i.e. head act directness levels, and internal and external modification strategies, through scientific concepts; 2) raise awareness of how the sociopragmatic dimension may influence request choices; 3) raise awareness of individual differences in perceptions of appropriateness. The instruction was distributed over four weeks (four hours in total), and used aims from the national English subject curriculum, LK06, as a point of departure. The LK06 (Udir, 2006a) stated that by the end of 7th grade learners should be able to: - use expressions of politeness and appropriate expressions for the situation - express oneself to obtain help in understanding and being understood in different situations - converse about the way people live and socialise in different cultures in English-speaking countries and in Norway, including the Sami culture Considering that these aims are very broad, and thus do not reflect a daily focus of language instruction, they were broken down by using aims from the Common European Framework of Reference for Languages (CEFR, 2018), and finally aims for the individual sessions were developed (see Tables 3.2 and 3.3). In addition, while the use of the term 'expressions of politeness' in the LK06 implied that certain language resources are inherently polite, the instruction itself aimed to problematise this view of language. The term ‘polite’ was therefore used as a starting point for raising the 37 In addition to piloting the data elicitation techniques with this group, all the data elicitation techniques had been used in previous research with learners of similar ages (Myrset, 2014; Myrset & Drew, 2016; Savić & Myrset, Forthcoming-b), but had been adapted for the present study. 38 In preliminary meetings, the teachers voiced concerns about teaching the material. Together they decided that the researcher would serve as a teacher. Whilst the initial plan had been for the teacher to teach the material, the change was both an ethical and methodological consideration in attempts to not take more time than necessary from the teachers’ schedule (Dörnyei, 2007). Methodology 53 learners' awareness of the contextually situated and sometimes idiosyncratic interpretations of politeness and appropriateness (e.g. Eelen, 2001; Spencer-Oatey, 2008; Watts, 2003). 3.2.1 Pedagogical approach and classroom procedures Sociocultural theory (SCT) was the pedagogical foundation for instruction. More specifically, a concept-based approach to teaching pragmatics was employed (see Section 2.2). Previous studies employing concept-based approaches (e.g. Morollón Martí, Forthcoming; Nicholas, 2015; van Compernolle, 2014) have tended to focus their instruction on sociopragmatic concepts, assuming that the learners were already familiar with the pragmalinguistic resources. The current project focused on YLLs and aligns with previous research employing concept-based approaches in the overarching principles: explicit instruction of scientific concepts, reflection as an important tool for development, metapragmatic awareness as a vital mediating tool for making choices, avoiding the teaching of rules of thumb, and fostering agency. However, considering that a broad pragmalinguistic repertoire is a prerequisite for agency, the project introduced scientific concepts related to pragmalinguistics as the point of departure. Grounding the instruction in SCT was also considered appropriate as it entails a dialogic approach of collaboration common in Norwegian schools, taking into account the instructional context (Bardovi–Harlig, 2015). Thus, the learners would develop knowledge through meaning-making with peers and the teacher as a mediator (Kozulin, 2018). Furthermore, as learners bring with them understandings from their L1 when learning the L2, the L1 was used as a basis, or scaffolding, when making sense of pragmatic phenomena (McConachy, 2013, 2018). Consequently, the instruction viewed the L1 and lived experiences as resources for meaning-making rather than as obstacles for L2 development (e.g. Chavarría & Bonany, 2006; McConachy, 2018). The instruction had two overarching themes: the pragmalinguistic and the sociopragmatic dimension of requesting. However, as Kasper (2001) points out, in pragmatics, language and contexts are interrelated and one cannot be taught without the other. As a consequence, these were addressed simultaneously when necessary. The rationale for dividing the two was that the instruction on pragmalinguistics would provide the learners with a pragmalinguistic repertoire that they could employ and vary when being introduced to sociopragmatic features of communication. The instruction lasted for four weeks and included three sessions per week – one 30-minute and two 15-minute sessions. The first six sessions focused on the pragmalinguistic dimension, while the last six focused on the sociopragmatic one. The Methodology 54 following sections (3.2.2 and 3.2.3) present each of the two overarching themes in detail. 3.2.2 The pragmalinguistic dimension In order to present and discuss the pragmalinguistics related to requesting, the researcher introduced scientific concepts during the first two weeks of the instruction period. The scientific concepts, e.g. ‘in-between’, were adapted to make them suitable for the target group, which is in line with van Compernolle (2014, p. 45), who argues that concepts need to “be simplified for pedagogical use […] without compromising [their] coherence and systematicity”. The framework of request strategies developed by Blum-Kulka et al. (1989) was thus chosen as it provides systematicity, but the terms were adapted for the purposes of the instruction. Table 3.1 presents the scientific concepts introduced during the instruction. Table 3.1: Scientific concepts employed during the instruction (also presented in Article III (Myrset, 2021)) Terminology Blum-Kulka et al. (1989) Adapted Directness levels Direct Direct a Conventionally indirect In-between Non-conventionally indirect/hints Hint a Internal modification Alerter Attention getters Address term Address term Lexical downgraders Polite wordsb External modification Grounder Reason Sweetener Compliment Promise of reward Promise a The two terms 'direct' and 'hint' are very similar to their Norwegian equivalents (direkte and hint) and were thus employed during the instruction b The term ‘polite’ was used for three reasons: 1) Considering the age of the learners the term itself was one that the learners were familiar with and could attach meaning to. 2) It was grounded in the learning aim from the national curriculum. 3) The term functioned as a starting point for raising the learners’ awareness about the contextually situated and sometimes idiosyncratic interpretations of the term (e.g. Watts, 2003). To foster the internalisation of the scientific concepts (e.g. van Compernolle, 2014; Vygotsky, 1934/2012), after being introduced, the concepts were used consistently for the duration of instruction. For instance, in the sessions following the introduction of directness levels, the terminology, i.e. ‘direct’, ‘in-between’, and ‘hint’, would be used Methodology 55 when they were relevant in the subsequent discussions and activities. Each session opened with a short repetition of the concept from the previous session before a new concept was introduced by the researcher, complemented with scaffolded discussions. Following each introduction, the learners were given activities adapted from those used in previous research and pragmatics literature (see Table 3.2) in which they could practise the pragmalinguistic forms. During these activities, which were either written or oral, the learners were reminded of previously introduced concepts. For instance, during the activity focusing on external modification, e.g. ‘reason’, the learners were asked to choose between being ‘direct’, ‘in-between’, or to ‘hint’ when making a request containing a ‘reason’. Each session ended with a short discussion to prompt reflection. Table 3.2. provides an overview of each session (1-6) focusing on the pragmalinguistic dimension. Methodology 56 Table 3.2: Overview of the sessions in the first two weeks focusing on the pragmalinguistic dimension. Session Concept Aim In focus Activities 1 Directness Learners will be able to perform and respond to requests. Directness strategies, i.e. direct, conventionally indirect, and hints. May, could, would Class discussion: Using video clips from Dumb and Dumberer and The Big Bang Theory in which the characters request. What did they ask for in the clips, and what did they say? a The requests from the clips are then presented on a PowerPoint slide, and subsequently placed on a continuum from ‘direct’ to ‘hint’. Following this, more pragmalinguistic resources are added to the continuum. Activity (adapted from Eslami and McLeod (2010) and Rinnert and Iwai (2010)): Three requests for each directness level are given on a worksheet. In pairs, the learners revise them into a different directness level, e.g. a direct request to a hint. See Appendix 1 for worksheet. 2 Downgraders Learners will be able to utilise downgraders to modify requests. Please, perhaps, possibly Class discussion: Can you think of any words we use in Norwegian that make requests sound less ‘harsh’? Is there a way to make them 'softer'? What do you think I mean by 'softer'? Provide the learners with examples of words. Activity (adapted from Rinnert and Iwai (2010)): In pairs, the learners make requests and add downgraders. Methodology 57 3 Alerters Learners will be able to utilise alerters to get someone's attention when modifying requests. Excuse me, pardon me, sorry Class discussion: Using a learner as prop. S/he stands with his/her back towards the class: What can we say to get his/her attention, in Norwegian or English? Show the learners some examples of alerters: Where do you think it makes sense to place these in a request? Activity: Onion circle. The learners stand in two concentric circles facing each other. The learner in the outer circle produces a request with an alerter (and downgrader) to a learner in the inner circle who responds. The learners in the outer circle take a step to the right so that they are facing a new learners, and repeat the process. Switch to inner circle producing request and outer circle responding. 4 Address Terms Learners will be able to utilise address terms to modify requests. Mr, Mrs, miss, sir, madam. Used as contrasting examples: dude, mate, love Class discussion: Using two learners as props. They bow, one deeper than the other: What is going on? Why do you think s/he is bowing deeper than him/her? What if I tell you it has to do with something called 'status' and ‘respect’? Who do you think is trying to show respect to the other? Can you think of ways we could do this by using language? Activity: The learners are mobile in the classroom. They approach each other, and produce requests with an address term. H1 Homework week 2 Internal modification strategies. Worksheet: Two pairs of requests are provided on a worksheet. Learners ask a parent/guardian to choose the request in the pair that they think is “nicer”. Afterwards the parent/guardian explains their choiceb. The learners’ findings are used as prompts for a class discussion, aiming to explore differences in perceptions. Did the parents/guardians in the class have different preferences? Why do the learners think that is? See Appendix 2 for worksheet. Methodology 58 5 Supportive moves Learners will be able to utilise supportive moves to modify requests. Grounder, sweetener, promise of reward Class discussion: Show three head acts on the board, each with arrows before and after. Sometimes we add something before or after the request itself. Add a grounder, a sweetener and a promise of reward, e.g. “You’re such a great driver. Could you give me a lift?” What do these do to the request? Pair activity: A worksheet where the learners make requests with a supportive move. See Appendix 3 for worksheet. Group activity: As a group (4-5), the learners make a request with an aggravating move (referred to as 'insults' and 'threat'). After deciding on a request, they write it down on a piece of paper and make a tableau (three-dimensional image) where they act as the speaker, hearer(s), and bystanders. Discussion: How did they come up with the request, and their choices when making the tableau. 6 Summary Learners will be able to utilise internal and external modification strategies to modify requests. Class discussion: What have we talked about so far? What have we added to the requests? Activity: Individual work. Using a computer, the learners write as many request formulations as they can think of, e.g. "Can I have a pencil?", "Can I have a pencil, please?", in 10 minutes. See Appendix 4 for examples of requests produced. a Underlined sentences are questions used to prompt the learners. The questions were asked in English, but often followed up in Norwegian. b Whereas this activity focused on decontextualised utterances, that is, making a pragmalinguistic choice without considering the sociopragmatic dimension, the activity aimed to focus specifically on the pragmalinguistics of requests to raise awareness of individual differences in perceptions. This was grounded in a previous study (Savić & Myrset, Forthcoming-b) in which the learners often viewed specific words and phrases, e.g. ‘please’ and ‘can I/you’, as inherently polite, regardless of the context. This was a view that he wanted the learners to question in the present study. Thus, the activity aimed to raise awareness that such individual preferences can be found even in decontextualised utterances. The activity was then used to stimulate a discussion about contextual situatedness of request appropriateness as well as about variability in individual speaker preferences. Methodology 59 3.2.3 The sociopragmatic dimension Whereas the first two weeks (pragmalinguistic dimension) were based on introducing a concept accompanied by terminology, the focus on the sociopragmatic dimension (the final two weeks) had a more consciousness-raising nature, in which the conceptual knowledge related to pragmalinguistics could be employed. As a result, this component had a stronger emphasis on prompting verbalised reflections (McConachy, 2013, 2018; van Compernolle, 2014, 2018; Zuckerman, 2004), and encouraged the use of L1 and lived experiences as part of meaning-making (e.g. Chavarría & Bonany, 2006; Eun & Lim, 2009; McConachy, 2013, 2018). The topics in focus for the sociopragmatic dimension were interlocutor characteristics (familiarity and age), place of interaction, and situation. The sessions started by engaging the learners’ prior knowledge and lived experiences related to the topic (e.g. What types of roles do you have (sibling, son/daughter, friend, etc.)? Do you use different language depending on who you are talking to?) followed by activities and reflection. The discussions and activities (see Table 3.3) aimed to prompt the learners to consider the context and utilise the pragmalinguistic resources that they had acquired during the first two weeks. Furthermore, the use of scientific concepts related to pragmalinguistic strategies was encouraged to facilitate internalisation and to serve as an orienting basis (Gal'perin, 1992; Morollón Martí, Forthcoming; van Compernolle, 2014; Vygotsky, 1934/2012). The contextual features were introduced through the use of images. First, requests were presented without the context, before using different images of situations to show how the requests, and their appropriateness, related to contextual factors. Figure 3.3 provides examples of how the context was introduced to the learners. Figure 3.3: Slides for introducing the context Through the first slide in Figure 3.3, the learners were presented with a direct request and told that it was written in capitalised letters to signal that someone was yelling ‘Give me a hose!’39. The class was then asked what the statement meant to ensure that 39 The scientific concepts, in this particular case ‘direct’, were also used in these discussions. Methodology 60 everyone understood its meaning before proceeding. The researcher then showed the request in relation to the picture from the DIY shop and explained to the learners that the customer was yelling this remark to the clerk. The learners were subsequently asked what they thought about the request being performed in the shop in this manner. Finally, the researcher showed the slide from the fire and asked the learners what they thought about the request in this situation. The aim of the discussion was to make the learners aware of how the context affects interpretation and meaning, i.e. the relationship between pragmalinguistics and sociopragmatics. In addition to the slides presented above, three more requests were used to help draw the learners' attention to the context. This was achieved through visual aids and discussions. By using class discussions, the learners were able to (co-)construct metapragmatic understandings as a group (Swain, 1997). Furthermore, this approach enabled them to use their own experiences as frames of reference and the L1 as a scaffold (e.g. Chavarría & Bonany, 2006; Eun & Lim, 2009). Table 3.3. provides an overview of each session (7-12) focusing on the sociopragmatic dimension. Methodology 61 Table 3.3: Overview of the sessions in the last two weeks focusing on the sociopragmatic dimension. Session Topic Aim In focus Activities 7 Introducing the context Learners will develop an awareness of how the context can influence the meaning of requests. Introducing the impact context has on the force of a request and on interpretation. Group and class discussions: Discussing various requests and deciding in which contexts they would be appropriate. What do you think about using this request in this situation? a Activity: Request perception journey (adapted from Hancock (1995)): The learners are provided with a worksheet. A pair of requests is shown on the board, and the learners choose the one they prefer by going right or left depending on their choiceb. Three such choices result in reaching one of eight cities. By raise of hand the learners respond to which city they ended up in. The ensuing discussion focused on how they perceived appropriateness and their thoughts about the learners ending up at different 'destinations'. See Appendix 5 for worksheet and request pairs. 8 Employing requests with different interlocutors: Familiarity and age Learners will develop an awareness of how their language may change depending on age and familiarity, and practise using their linguistic resources with this in mind. Family members: Young, old Friends: Young, old Strangers: Young, old Group and class discussions: What types of roles do you have (sibling, son/daughter, friend, etc.)? Do you use different language depending on who you are talking to? Worksheet in pairs (adapted from Yates (n.d)): The learners are expected to produce request for the same thing in two different situations. Individual worksheet (adapted from Yates and Springall (2010) and using Hill et al. (1986)): Matching request and interlocutor by drawing a line between the two. Each selection task has 6-8 requests and 11 options for interlocutors, of which two are blank and one states that the learner would not use this form of request. See Appendix 6 for worksheets. Methodology 62 9 Employing requests in various places Learners will develop an awareness that requests are performed in different places, and practise using their linguistic resources with this in mind. Place Prompts to help the learners: restaurant vs fast food restaurant; at home vs at a friend's house Class discussion: What types of places can you think of where you would use requests? Dice game (adapted from Kontra (n.d.)): In pairs, the learners use a sheet with a list of interlocutors (e.g. sibling or old stranger) and a list of what to request (e.g. water or money), both lists are numbered 1-6. The learners roll a dice twice to select an interlocutor and what to request, and then produce a request. See Appendix 7 for worksheet. H2 Homework week 4 Pragmalinguistics and sociopragmatics. Worksheet: The learners ask their parents/guardians what they think are important considerations when requesting. Of the things they list, the parents/guardians choose the three they think are the most important. The learners then write why they think their parent/guardian chose those three. See Appendix 8 for worksheet. 10 Employing requests in various situations Learners will develop an awareness that requests are performed in different places and situations, and practise using their linguistic resources with this in mind. Place and situation Prompts to help the learners: restaurant vs fast food; school vs home; wedding vs birthday Group and class discussions: Follow-up from previous session. Did you vary your requests depending on what and who you were asking when you rolled the dice? Dice game 2.0 (adapted from the previous dice game and Kontra (n.d.)): The class is divided into two groups (A and B). Group A is mobile and learners in Group B are seated by their desk. Learner A (L-A) receives a dice and a note with alternatives for what to request, numbered 1-6. Learner B (L-B) receives a post-it note with an assigned role and a sheet to keep tally of the types of request strategies employed. The learner attaches the post-it to their chest. L-A approaches L-B, rolls the dice, reads the assigned role of L-B, and performs the request. L-As and L-Bs switch roles when half the time of the activity has passed. See Appendix 9 for sheet used to count linguistic resources. Methodology 63 11 Employing requests with different interlocutors and in different places Learners will develop an awareness of the interplay between pragmalinguistics and sociopragmatics, and practise using their linguistic resources with this in mind. Interlocutor, place, situation. Prompt: remind the learners of the previous topics that were brought up. Class discussion: What do you think is important to think about when you are making requests? What have we talked about in the past couple of weeks? Pair activity: Story completion task (inspired by the DCT (e.g. Blum-Kulka et al., 1989)). Learners receive a script where parts of the story are pre-written. The learners fill in the gaps to finish the dialogue and story. Optional: A second page with blanks to continue the story. See Appendix 10 for script. 12 Summary Learners will be able to utilise internal and external modification strategies to requests, and be aware of some contextual considerations when requesting. Group and class discussions: Discussing the content of the four weeks. a Underlined sentences are questions used to prompt the learners. The questions were asked in English, but often followed up in Norwegian. b Whereas this activity deals with decontextualised utterances, the aim of the activity was to raise awareness of individual differences in preferences even when drawing solely on the pragmalinguistics. In a previous study (Savić & Myrset, Forthcoming-b), the researcher found that learners often viewed specific words and phrases, e.g. ‘please’ and ‘can I/you’, as inherently polite, regardless of the context. This was a view that he wanted the learners to question in the present study. For this reason, the activity first separated the utterances from the context. The learners were thus asked to choose a direction based solely on the utterance. The raising of hands for each destination following the activity then provided a visual representation of these individual preferences, even within a group of peers. Finally, the ensuing discussion aimed to stimulate reflections about why the learners thought their peers had reached a different destination, followed by drawing in the contextual situatedness of request appropriateness as well as about variability in individual speaker preferences. Methodology 64 3.3 Data collection This section presents the elicitation techniques employed and the data analysis procedure for each dataset in this project. The techniques included a video-prompted oral discourse completion task (VODCT), Readers Theatre (RT), group interviews, and appraisal tasks. All the learner data comprised of verbal responses and were thus audio-recorded. The audio-recordings were later transcribed verbatim by the researcher. The techniques generated data for the four articles presented in this thesis, which contributed to answering the overarching research question, that is, How does a concept-based approach to teaching requests impact young language learners’ request production and awareness, and their engagement with pragmatics?. 3.3.1 Request production data 3.3.1.1 Elicitation technique: video-based oral discourse completion tasks (VODCT) In order to explore the impact of instruction on the learners’ request production, this study employed a video-prompted oral discourse completion task (VODCT) to elicit requests. The VODCT was conducted in a pre-post-delayed design, that is, two weeks prior to the instruction (pre-test), in the week following the instruction (post-test), and six weeks after the instruction (delayed post-test). This technique was adapted from a study by Savić and Myrset (Forthcoming-b), focusing on the development of young Norwegian EFL learners’ (meta)pragmatic awareness in which the VODCT served as a pre-task to elicit requests40. Thus, the previous study served as a pilot. Furthermore, the researcher’s prior engagement with the VODCT ensured that he was familiar with its procedures and the data it generated (see also Section 3.4 for quality criteria). The VODCT was conducted in English. The nature of a discourse completion task (DCT) is that the participants are provided with the social situation leading up to a speech act and asked to state what they would say in the given situation (Economidou-Kogetsidis, 2013). DCTs have become standard for eliciting data in pragmatics research (Bardovi–Harlig, 2018). For instance, Félix-Brasdefer (2010) found that approximately half of the studies on refusals and rejections conducted in cross-cultural, single-moment, and ILP research, employed written DCTs. Similarly, with regard to requests, the written DCT has been commonly used (e.g. Brubæk, 2012; Cenoz & Valencia, 1996; Krulatz, 2016; Ogiermann & Bella, 40 The VODCT has since been employed in a study by Savić et al. (2021), which aimed to explore the pragmalinguistic development of Cypriot Greek and Norwegian learners of English, aged roughly 9, 11, and 13. Methodology 65 2020; Pinto & Raschio, 2007; Svanes, 1989). A possible reason for this is access to large amounts of data while controlling the social variables within the scenarios (Blum-Kulka et al., 1989; Economidou-Kogetsidis, 2013). Thus, the written DCT generates comparable data. However, in spite of its advantages, the written DCT has also been subject to criticism, for instance, because it does not reflect the modality it seeks to investigate (Economidou-Kogetsidis, 2013). In other words, it represents a so-called 'written-for-oral' test (Bardovi–Harlig, 2018), in which the written test aims to explore language that usually occurs in speech. Furthermore, the written DCT provides scenarios with hypothetical interlocutors, resulting in a potential gap between utterances elicited through the test scenarios with those that occur in face-to-face interaction (Woodfield, 2008). This latter point could also be extended to the situations themselves, namely, the DCT may invite the participant to respond to scenarios where both the interlocutor and situations are hypothetical, i.e. the participants may not have had any real-life experience with such situations. With this in mind, whilst the written DCT enables the researcher to elicit large quantities of comparable data, the results do not mirror authentic speech. Considering that the setting for data collection was an EFL classroom, and that the researcher employed a pre-post-delayed design, a form of DCT was selected to ensure comparability between the tests. However, the researcher aimed to remedy some of the criticism by eliciting the requests verbally, that is by conducting oral-for-oral testing (Bardovi–Harlig, 2018). In addition, employing oral DCTs was considered appropriate, as they can be administered regardless of the participants’ proficiency level (Bardovi–Harlig, 2018). Furthermore, since the study focused on YLLs, the test was conducted in friendship groups (Gu et al., 2005; Pinter & Zandian, 2014), which aimed to empower the learners and make them feel comfortable in the research setting. In addition, drawing on literature on research with children (Johnston, 2008; Punch, 2002a; Yamada-Rice, 2017), the test incorporated visual elicitation techniques, that is, language learning videos to prompt request production and the printed pictures of each request situation as an additional stimulus. This approach aimed to engage the learners by mirroring activities with which they were familiar in their everyday lives. In addition, the visuals also aimed to draw attention away from the researcher. These considerations were particularly important for the pre-test, considering that it represented the learners’ first experience as participants in the present study. Given that requesting is contextual in nature (e.g. Kasper, 2001; Spencer-Oatey, 2008), the researcher selected videos that provided a range of situations and interlocutors to investigate whether the learners used different strategies depending on the situation, familiarity, and age of the interlocutor. Furthermore, the videos included situations that the learners were familiar with from their everyday lives, thus aiming to reduce the Methodology 66 hypotheticality of the test (Woodfield, 2008). Thus, the VODCT included eight language learning videos in which child characters produced requests to interlocutors of different familiarity and age, for example a friend (familiar, same age), a sales assistant (unfamiliar adult), and a teacher (familiar adult). The same videos were used in all three tests (pre-post-delayed). Whilst this may be viewed as a potential limitation (see also Section 5.2), intervals of 5-6 weeks between the tests were found to be enough time to avoid retest biases (Brown et al., 2008; Randall et al., 2016). Table 3.4 provides an overview of the videos, which comprised a total of 23 request scenarios. Furthermore, in addition to the videos, the learners were provided with a supplementary printed screenshot of the situation in question to serve as a visual stimulus (Figure 3.4). Figure 3.4: A screenshot from one of the videos used in the project as a visual prompt (https://youtu.be/P5Vi4j1F4sE). Methodology 67 Table 3.4: Videos used in the VODCT (videos retrieved: 16.08.19). Video Test Situation Interlocutor Familiarity Request Library https://youtu.be/f14VXicnOvs Pre, Post Library Library Playground Friend Friend Friend Familiar Familiar Familiar Play outside Play outside in snow Build a snowman Museum https://youtu.be/IfT4XhfUFkk Pre, Post Outside school Street Friend Stranger Familiar Unfamiliar Go to museum Directions Phone call https://youtu.be/gWOqA3pUaTk Pre, Post On the phone On the phone Friend's parent Friend Unfamiliar Familiar Speak to friend Go outside and play Fast food https://youtu.be/49QFHWIky-k Pre, Post Counter Counter Counter Sales assistant Sales assistant Sales assistant Unfamiliar Unfamiliar Unfamiliar Burger, fries, coke Two cheeseburgers Two cokes Shopping https://youtu.be/P5Vi4j1F4sE Pre, Post, Del Market Market Store Store Sales assistant Sales assistant Parent Parent Unfamiliar Unfamiliar Familiar Familiar Doll Car Kite Hat Classroom https://youtu.be/UFy02dqJCUs Pre, Post, Del Classroom Classroom Classroom Classroom Friend Teacher Teacher Friend Familiar Familiar Familiar Familiar Crayon Green paper Orange paper Yellow pencil Dinner https://youtu.be/CtsG2klnuGk Pre, Post, Del At the table At the table At the table Friend's parent Friend's parent Friend's parent Unfamiliar Unfamiliar Unfamiliar Fork More soup Ask to come back Restaurant https://youtu.be/IWMMkp35d6Y Pre, Post, Del Restaurant Restaurant Waiter Parent Unfamiliar Familiar Green salad Glass of water Methodology 68 3.3.1.2 Procedures of the VODCT The VODCT took place in the school. The researcher was provided with a group room opposite the learners’ classroom and was allowed to take the groups of learners out of the regular classes. This ensured that the VODCT, which lasted approximately 30 minutes per group, could be conducted within three days. The pre-post-delayed design was conducted with the pre-test two weeks before instruction, the post-test during the week following the instruction, and the delayed post-test approximately six weeks after the instruction was completed (see also Figure 3.1). Furthermore, considering the school context, the researcher attempted to avoid unintended stress for the learners and to create a relaxed atmosphere in which the learners felt comfortable speaking, thus reducing an inherent researcher-child power imbalance (e.g. Griffin, 2019; Gu et al., 2005), by emphasising that there were no incorrect answers (Beauchamp et al., 2019; McKay, 2006; Mukherji & Albon, 2015; Pinter & Zandian, 2014; Punch, 2002b). Each video was introduced by the researcher (e.g. This video is about a girl and a boy who go to the market to look at a doll and a toy car. Afterwards, a girl goes to the store with her parents. She looks at a kite and a hat.) and played up to the point leading to a request to provide the context. The video was paused before a child character uttered a request. During this pause, the researcher presented the visual stimulus (Figure 3.4) and asked the group to suggest what they thought the child would say in the video, e.g. "The girl really likes the doll. What does she say to the sales assistant?” (see Appendix 13 for the questions for each video). The first two cycles (pre- and post-test) were organised in the same manner, by playing the eight videos in random order and supplying the visual prompts for each request scenario. However, during the second cycle, the researcher sensed that the learners were growing impatient with the task, most likely due to its repetitive nature and, as a result, decided to change the final cycle of the VODCT. According to Dörnyei (2007), making such a change is quite common in educational research, due to its emergent and fluid nature, and allows the researcher to be flexible. Thus, in the delayed post-test, four of the eight videos were selected for the test. These were selected to include the three main interlocutor characteristics in the videos (see Table 3.4), namely friend (same age), familiar adults (e.g. parents), and unfamiliar adults (e.g. waiter). Furthermore, in addition to only using four of the videos (i.e. Shopping, Classroom, Dinner, Restaurant), rather than playing the videos, the researcher reminded the learners orally about the context, showed them the visual stimuli, and asked what they thought the Methodology 69 children would say41. The questions prompting the request production were phrased in the same manner in all three tests. Considering the ‘oral-for-oral’ testing (Bardovi–Harlig, 2018), the response type in the VODCTs was verbal. The groups were audio-recorded during the three tests and their responses were transcribed. The VODCTs in the pre-, post-, and delayed post-test resulted in a total of 2180 requests. 3.3.1.3 Analysis of request production data The requests produced in the VODCTs were analysed quantitatively. More specifically, each request was analysed and coded into different categories (see Table 3.5) in the statistics software SPSS (IBM, 2019), using a coding manual adapted from Blum-Kulka et al. (1989). The purpose of this coding was to quantify the learners’ language use in the VODCT to "a construct that cannot be directly seen or observed" (Roever & Phakiti, 2018). In other words, the coding provided the researcher with the frequencies of language use, for instance, the number of times the learners used sweeteners (e.g. 'You could win Masterchef with this soup. Can I get some more?') in each test and in which scenarios. Drawing on Ellis (1992), the requests labelled 'Requests for information', e.g. 'How much is it?', were omitted prior to the analysis, leaving 2015 (pre: 699, post: 872, delayed: 444) requests for analysis. Although the requests were elicited in groups, each request was analysed in full, including instances in which the learners within a group produced similar or even identical requests potentially by drawing on their peers (Holzman, 2018; Vygotsky, 1934/2012). Whilst this may be seen as a limitation, the decision to include all the requests produced was influenced by the instructional approach of the study as well as learning processes occurring in the classroom, in which learners co-construct ideas in collaborative dialogues (Swain, 1997). All the learners in the groups produced requests in all three tests. 41 In addition, the researcher had planned to have the learners act out some of the situations in the form of a role play. However, this was attempted with the pilot group, but turned out to confuse the learners more than making the test more engaging. Consequently, the role plays were not conducted with the 11 groups generating the data (see also Section 3.4 for quality assessment). Methodology 70 Table 3.5: Coding manual for requests Variable Value Test 1: Pre-test 2: Post-test 3: Delayed post-test Learner 1-46 (number assigned to each learner)a Group 1-11 (number assigned to each group)a Directness 1: Direct 2: CI (Conventionally indirect) 3: Hint 4: Request for information* Modals 1: Can 2: May 3: Could 4: Would 5: Other Attention getter 1: Excuse me 2: Pardon me 3: Sorry 4: Hey 5: Hello 6: Hi 7: Yo Address term 1: First name 2: Address term (e.g. professional title) 3: Mum/Dad Downgrader 1: Maybe 2: Perhaps 3: Possibly 4: Please 5: Please+ (Please + additional downgrader) Supportive move 1: Grounder 2: Sweetener 3: Preparator Situation 1-23 (number assigned to each situation) a Requests for information, e.g. 'How much is it?' were not included in the final analysis The coding resulted in raw frequencies of instances of language use, or nominal data42. It has become increasingly common in applied linguistic research to investigate "rates of occurrence of a particular linguistic feature" (Egbert & LaFlair, 2018, p. 525). Nominal data thus assigns numerical values within a group, e.g. modal verbs, for ease of analysis (Christmann, 2012; Cox, 2017). This was the type of data used to analyse 42 This type of data has also been referred to as "[c]ategorical data […] frequency data, nominal data, or data with nonquantitative outcomes" (Egbert and LaFlair, 2018, p. 525). Methodology 71 the learners' request production and it allowed for an investigation of frequencies of use of specific pragmalinguistic resources. The quantified language use in the learners’ requests enabled the researcher to run statistical tests, thus distinguishing the research from "an anecdotal account of events" (Turner, 2014, p. 12). Two tests were conducted in cooperation with a statistician: a chi-square test of independence and a z-test for difference of proportions. The chi square test is a versatile tool as it only requires frequencies (Roever & Phakiti, 2018), and is common for measuring differences in nominal data (Egbert & LaFlair, 2018). Thus, it was considered appropriate for the current study. In essence, the chi-square test is an inferential statistic test that investigates discrepancies between observed and expected frequencies (Christmann, 2012; Roever & Phakiti, 2018). In this project specifically, the chi-square test measured the significance of the differences in the frequency of use of specific request strategies and pragmalinguistic resources. The z-test, on the other hand, measures the increase or decrease of use (Abbott, 2017; Bruce, 2015). By using the frequencies from the pre-test as baseline, the z-test aimed to highlight the increase or decrease of use of linguistic resources and whether these changes were statistically significant or not. In other words, it enabled the researcher to explore the increase or decrease in frequencies between the tests, namely, from pre- to post-test, from post- to delayed post-test, and from pre- to delayed post-test. The results from the statistical analyses are presented in Article II (Myrset, Pending revisions). 3.3.2 Readers Theatre As part of the project, two cycles of Readers Theatre (RT), a group reading aloud of a text (Drew, 2018), were conducted. The aim is to read the text aloud to an audience, normally peers in a classroom context. In educational contexts, three types of scripts can be used: pre-written scripts, adapted scripts, or learner-produced scripts. The first cycle43 was conducted in the week before the instruction and employed pre-written scripts produced by the researcher. This cycle aimed to introduce the learners to the topic (requests), RT as an activity, and the researcher. The second cycle was conducted in the week following the instruction, and the learners produced scripts based on a story outline. The aim of the second cycle was to collect scripts, including requests, produced by the learners, which could later be used to prompt metapragmatic discussions (presented in Article I). Both cycles were completed with the learners performing in front of their peers. In a previous study (Myrset, 2014; see also Myrset & Drew, 2016), 43 A cycle refers to a full sequence of RT. For instance, when using pre-written scripts in the present study (Cycle 1), the cycle opened with handing out the scripts. The learners then read through the text as a group and assigned roles. After the learners had been assigned roles they rehearsed the script through repeated reading. Finally, they performed their script in front of their peers (see also Young and Rasinski, 2009, for a detailed plan of a full cycle). Methodology 72 the researcher investigated the cognitive and affective effects of RT among Norwegian learners of English in 6th grade. Myrset (2014) also conducted two cycles of RT: first, by providing the learners with pre-written scripts and second, by learners producing their own Christmas story. The findings from the previous study, with learners of similar characteristics (age and educational context), suggested that RT was indeed a feasible method to include with the learners in the current study44. 3.3.2.1 The first RT cycle: purpose and procedures In the present study, the first cycle, which employed pre-written scripts and was conducted prior to the instruction (see Figure 3.1 for overview), served three purposes: 1) to introduce the researcher to the learners and facilitate rapport-building, 2) to introduce requests as a topic, and 3) to familiarise the learners with RT scripts before they wrote their own scripts in the second cycle (see Section 3.3.2.2). When it comes to introducing the researcher into the teaching context, previous research has found that RT has affective benefits such as increasing learners’ confidence and motivation (e.g. Bridges, 2008; Casey, 2006; Drew & Pedersen, 2010, 2012; Young & Rasinski, 2009, 2018). Such benefits were also evident in the researcher’s prior study with 6th graders (Myrset, 2014; Myrset & Drew, 2016). In the present study, it was important for the researcher to build rapport with the learners so that they would feel comfortable with him (Kuchah & Milligan, 2021; Kuchah & Pinter, 2012; Punch, 2002b; Urbach & Banerjee, 2019; Zandian, 2021). Thus, RT was considered an engaging activity that could serve as a rapport-building activity to familiarise the learners with the researcher. Additionally, considering its affective benefits, RT was considered an engaging way to introduce – and potentially spark the learners’ curiosity about (Ness, 2019) – requests. Thus, the first cycle aimed to introduce the topic of the project through different request scenarios. As preparation for the first cycle, the researcher produced six scripts (see Appendix 11 for an example). These scripts were based on outlines that the researcher had produced in advance. The six outlines involved different situations in which the main characters made requests, e.g. asking for help, or asking to get something. The main characters were all children, whereas the interlocutors were of different familiarity and age. When preparing the scripts, the criterion was to include a variety of linguistic resources for requesting, namely directness levels, and internal and external modification (Blum-Kulka et al., 1989). After the researcher had written the scripts, he sent them to a native speaker of English to ensure their quality. 44 It is important to note that data was not collected during the cycles. Thus, the learner who had not given consent to take part in the study could participate in the cycles on equal terms with their peers. Methodology 73 The scripts were written as a developed rather than traditional model of RT (Shepard, 2004), with learners reading the roles of either narrators or characters45. In other words, the narrators would tell the story and provide the audience with the context (e.g. Arthur has been saving his allowance to buy a PlayStation 4, and there’s a sale on the console at the local store. Unfortunately, he is still 30 pounds short, so he decides to ask his dad to lend him some money.), and the characters would provide the dialogues (e.g. Arthur: Dad, they have a sale on PlayStation 4, but I still haven’t saved up enough money to buy it. Could I borrow 30 pounds?). The reason for using the developed model was twofold: firstly, in a previous study (Myrset, 2014), the 6th grade learners reported that they could more easily comprehend the performances of a developed as opposed to a traditional model and also opted for a developed model when they produced and performed their scripts in the second RT cycle. Secondly, the narration and dialogues enabled the learners to contextualise the dialogues, and ultimately the requests, within the performances they were watching; namely, narrators provided the background story and context, such as where the characters were or what they did, while the characters performed the dialogues. During the first RT cycle, each friendship group was handed a script46. They were given two English lessons (45min*2) to rehearse. As preparation, the learners read through the script individually or together as a group. Following the first reading of the text, each group assigned the different roles from the scripts. Before rehearsals started, some of the learners chose to highlight the lines that they would be reading. The rehearsals involved very little guidance by the English teacher and the researcher, who spent their time during rehearsals moving between the groups, answering questions about pronunciation of words, or providing tips about how to make the performance more effective. Since the scripts involved more characters than group members, some learners chose to use simple props or costumes in order to shift between scenes or characters. The cycle was completed with the learners performing in front of their peers, where 90 minutes had been allotted to the performances. 3.3.2.2 The second RT cycle: purpose and procedures As a result of the first cycle, the learners were familiar with the nature of RT and the structure of scripts. This provided a foundation for the second RT cycle, in which the learners produced their own scripts. In this cycle, the learners were provided with outlines based on the scripts from the first RT cycle, which provided the learners with a foundation for the story, as well as including scenarios in which they had to produce 45 The traditional model does not divide the reading roles according to whether they involve narration or reading the role of a character, as is the case in the developed model. In the traditional model, the readers usually stand or sit in a row and alternate their reading in a fixed order. 46 These procedures were identical for both classes. Methodology 74 requests. The researcher had taken great care when writing the outlines to ensure that they did not influence the requests produced by the learners. The outlines were included in a four-page folder (see Appendix 12 for an example), consisting of writing guidelines, the outline, cartoons of the characters created by a cartoonist, and an example text of a script. The scripts produced by the learners were later used as a discussion prompt in the group interviews. In other words, the researcher would bring the script produced by the group to prompt discussions about the learners’ choices, specifically those relating to deciding on request formulations during the writing process (see Article I). In the second RT cycle, each group in the class was provided with a different outline, which ensured that each group performed a unique story for their peers47. As in the researcher's previous study (Myrset, 2014), the learners were provided with more time in this cycle. The production and rehearsals of the script were conducted over a week (approximately 5 hours total). At the school, each learner had access to their personal Chromebook, and the groups were thus able to co-write the script in the same document. During the introduction of the cycle, the English teacher and the researcher instructed the learners on how to structure the writing and urged them to use dialogues as much as possible. Similar to the first cycle, the learners did most of the work without guidance from the English teacher and the researcher, allowing the latter to move between the groups and answer questions. However, to ensure that the teacher and the researcher did not influence the learners’ request production, for questions specifically related to the request situations, the learners were prompted with questions such as "Think about what you have learnt in the past four weeks". This ensured that the learners produced the requests themselves, which was important for the discussions during the interviews. Following the script production, the cycle followed a similar pattern to the first RT cycle: repeated readings of their text (rehearsals), and planning the performance. In the week following the script production and rehearsals, the learners performed for the class. The script production resulted in 11 scripts, ranging from 308 to 1032 words. The scripts themselves were not analysed for the requests produced, but were used in group interviews as a stimulus to facilitate discussion about the considerations related to formulating the requests the groups had made during the writing process. Illustrative examples of discussions prompted by the script-production are presented in Article I 47 The groups were provided with an outline that did not match the script they had been given in the first cycle. Methodology 75 (Myrset & Savić, 2021) in relation to the affordances of RT to elicit metapragmatic data. 3.3.3 Interview data 3.3.3.1 Elicitation techniques: Group interviews Articles III and IV aimed to investigate the learners’ (meta)pragmatic awareness and perceptions about the project. In order to elicit relevant data, the learners were interviewed in groups. The interviews were semi-structured with an unstructured response mode (Cohen et al., 2000). This meant that the questions followed a guide (see Appendix 14 for interview guide), but were raised in a manner that allowed the respondents to answer freely (Rolland et al., 2020). The semi-structured interview is the most commonly used in research (Polkinghorne, 2005), also in educational research, as it provides both objectivity and depth and is particularly suitable for YLLs (Griffin, 2019; Holmes, 2019; Johnston, 2008; Kingdon, 2019). In addition, if necessary, the researcher would prompt, or ask follow-up questions, enabling the learners to extend or elaborate on a topic. These prompts aimed to generate the “richness, depth of response, comprehensiveness, and honesty that are some of the hallmarks of successful interviewing” (Cohen et al., 2000, p. 278). Using this approach allowed the researcher to compare the answers provided in the various groups, while making attempts to keep the participants relaxed, thus avoiding a stressful or overwhelming environment (Mukherji & Albon, 2015; Pinter, 2014). Furthermore, to allow the learners to speak freely, and thus provide them with a voice (e.g. James, 2007; Lundy, 2007; McTavish et al., 2012), the learners were invited to use both Norwegian and English (Pinter & Zandian, 2014). Since the topics in the interviews aimed to prompt reflections, and verbalising reflections can be difficult for YLLs (Zuckerman, 2004), the default language was Norwegian to reduce cognitive load (McKay, 2006; Pinter & Zandian, 2014). This resulted in the learners mainly using Norwegian when they explained their choices. However, the learners would resort to English when, for instance, they provided specific examples or used the scientific concepts introduced during the instruction. In addition, the interviews were task-based, which aimed to facilitate the learners’ expression of ideas (e.g. Andrews, 2021; Holmes, 2019; Lyndon, 2019; Mayo, 2021; Punch, 2002a, 2002b). Using tasks in research with children may serve as a resource for the learners to express their understandings through different modalities and make communication easier (O'Kane, 2008). Facilitating communication was particularly important in the present study: the interviews aimed to generate reflections which may be difficult to verbalise (Zuckerman, 2004), thus, in addition to the use of Norwegian, the tasks included as additional response modes (Gobo, 2008), such as non-verbal. Methodology 76 Consequently, the task functioned as a resource facilitating shared understandings between the researcher and the learners (McTavish et al., 2012; Rollins, 2018), and ensured that learners could participate equally regardless of language proficiency (Pinter, 2014). The tasks also gave the learners additional time to think during their responses (Punch, 2002b) and provided a visual stimulus in the discussions (Holmes, 2019; Lyndon, 2019). Finally, task-based interviews were also considered important for building rapport, namely by decreasing the adult-child power imbalance and making the learners comfortable to speak freely (Beauchamp et al., 2019; Griffin, 2019; Gu et al., 2005; Johnston, 2008; Kuchah & Pinter, 2012; McTavish et al., 2012; Mukherji & Albon, 2015; Punch, 2002b). This latter point of enabling learners to speak freely was considered vital in the present study, as it provided the foundation for giving the learners a voice (Lundy, 2007; McTavish et al., 2012; Schiller & Einarsdóttir, 2009) . 3.3.3.2 Interview procedures Similar to the VODCT and the RT cycles, the interviews were conducted in the friendship groups (Kuchah & Pinter, 2021; Pinter & Zandian, 2014) to ensure that the learners felt comfortable in the research setting and alleviate the power imbalance (e.g. Punch, 2002b). In addition, considering that the research was conducted in the school setting, the learners were reminded that there were no correct or incorrect answers (McKay, 2006; Mukherji & Albon, 2015; Pinter & Zandian, 2014). In this way, the researcher aimed to create an environment in which the learners could express themselves freely, negotiate and co-construct meaning. In addition, the task-based component included two appraisal tasks, adapted from a previous study by Savić and Myrset (Forthcoming-b). The scripts from RT cycle 2 (see Section 3.3.2.2) served as an additional prompt. The group interviews aimed to elicit data on the learner reflections about the writing process, specifically their considerations when producing requests (presented in Article I); metapragmatic understandings (presented in Article III); and their perceptions of the project, e.g. elicitation techniques and instruction (presented in Article IV). In addition to the questions, two versions of the appraisal task were employed during the first interview. The first variant included the appraisal of two requests from the scripts from the first RT cycle (Figure 3.5). These requests were taken from the same story on which the learners worked in RT cycle 2. Thus, each group within a class appraised a different set of requests, matching the outline they had developed in RT cycle 2. As a result, the learners could appraise, the request itself and give reasons for their appraisals, but also compare the requests with the ones from their own script. Unless asked, the researcher did not inform the learners that he had written the scripts Methodology 77 for the first cycle to avoid influencing the learners' appraisals and to ensure that the learners felt comfortable and could speak freely. Figure 3.5: A request appraisal sheet (also presented in Article III (Myrset, 2021)) During the request appraisal task, the learners were first asked if they thought that the request was a “nice” (������������������������), a “so-so” (������������������������), or a “not so nice” (☹) way to ask (see Appendix 14 for interview guide), and to leave a mark on the emoticon using a marker. Each marker in the group was a different colour, which enabled the learners to see (visual stimulus) where they had left a mark during the discussion. In addition, it enabled the researcher to analyse each individual response. Following the appraisal, the learners were asked to provide their reasoning (e.g. Why do you think it was blue (������������������������)? What would we have to do to get it up to green (������������������������)?). This task aimed to prompt metapragmatic discussions, some of which were included in Article III (Myrset, 2021). The second appraisal task aimed to facilitate the discussion about the learners’ perceptions of the project (Figure 3.6). The appraisal task included four components from the project: Performances (RT cycle 1), Class activities (the instruction), Writing scripts (RT cycle 2), and Videos (VODCT). In addition, the sheet had two empty slots, in which the learners could decide for themselves what they wanted to appraise to provide them with agency in the task (Pinter & Zandian, 2014). Figure 3.6: Appraisal sheet for the project components (also presented in Article IV (Myrset, In preparation)) Methodology 78 For each component the learners were first asked if they could remember it (e.g. Do you remember the activity when I gave you a script?). Then they were asked to brainstorm aspects of the component that they could remember. Following the reminder and the brainstorming, they were asked to appraise the component. Similar to the request appraisals, the learners used markers to make their appraisals, which served as a visual stimulus during the discussions and enabled subsequent analyses. The questions were structured in a similar way as with the first emoticon task, both for the appraisal (e.g. Do you think it was fun, so-so, or not so fun?) and to prompt discussions (e.g. Why do you think it was blue? What would we have to do to get it up to green?) (see Appendix 14 for interview guide). In addition, the learners were invited to select topics for appraisals in the two open slots. In these, the learners mainly chose specific activities (e.g. the request perception journey) or approaches during the instruction (e.g. working in groups or the use of PowerPoint slides). These were subsequently discussed using questions similar to those for the pre-decided components. 3.3.3.3 Analysis of interview data The transcribed interviews were analysed using NVivo 12 (QSR, 2016), a computer-assisted qualitative data analysis software. The analyses for the interview data were conducted through content analysis, "a highly flexible, pragmatic, and systematic method used for investigation of a wide range of topics" (Selvi, 2020, p. 450). This approach to analysis, in which the data can be understood through an inductive (data-driven) or deductive (theory-driven) approach, has become more widely used by researchers in applied linguistics in the last couple of decades (Elo & Kyngäs, 2008; Selvi, 2020). Whilst the labels ‘inductive’ and ‘deductive’ provide insights into how the data was approached, that is, in a data-driven or a theory-driven manner, the present study adopts Selvi’s (2020, p. 442) less categorical view between deductive and inductive content analysis, in which there are “possible convergences” between the two. With this in mind, in the present study, the data was approached deductively for Article III and inductively for Article IV. For both Articles III and IV, the preparation stage involved acquiring an overview of the data, which involved reading the transcripts for content through open coding (Dörnyei, 2007; Elo & Kyngäs, 2008). Following Saldaña (2016), coding was practised as a cyclical, heuristic process, in which the codes were constantly refined. This process allowed the researcher to arrange the data in a systematic manner in a process where the data was “divided, grouped, reorganized, and linked in order to consolidate meaning and develop explanation” (Saldaña, 2016, p. 9). In addition, the interviews were transcribed by the researcher, which is "an important first step in data analysis" (Bailey, 2008, p. 129; Dörnyei, 2007). However, following the first stage of coding, the analyses for Articles III and IV followed different patterns. Methodology 79 Article III aimed to explore the learners’ metapragmatic understandings in the interviews, with a specific focus on their use of scientific concepts. The open coding revealed that the learners collaboratively engaged with each other’s ideas in order to (co-)construct meaning (Swain, 1997). This co-construction occurred over multiple turns, or within “identifiable units of a collaborative activity” (Fortune & Thorp, 2001, p. 146). Fortune and Thorp (2001) refer to these units as ‘episodes’ and present four stages of analysis for such episodes. These stages are to 1) identify the episodes, 2) organise the data into categories, 3) organise the data into subcategories, and 4) quantify the data. The present study followed a similar pattern. Firstly, following the open coding, the researcher coded the transcriptions to identify the episodes that occurred within each group, more specifically those in which the learners expressed metapragmatic understandings. Secondly, the episodes were organised into overarching categories. Thirdly, the researcher adapted the framework developed by Fortune and Thorp (2001), which originally aimed to highlight language related episodes (i.e. the learners’ use of metalanguage in relation to grammar) to focus specifically on metapragmatic awareness and scientific concepts. The original codes by Fortune and Thorp (2001, p. 150) were: M (metalanguage alone, e.g. “Shall we put ‘a’ or ‘the’ before ‘zoo’?”), M+G (metalanguage and grammatical terminology, e.g. “‘A zoo’? Or is it the definite article?”), M+R (metalanguage and rule or generalisation, e.g. “We use ‘the’ when something’s been mentioned before.”), and M+T (metalanguage and text knowledge to inform the decision, e.g. “No, ‘the zoo’. We know which zoo. The one the keeper worked in.”). The adapted codes are presented in Table 3.6. Table 3.6 Coding framework, adapted from Fortune and Thorp (2001) (adapted from Article III (Myrset, 2021)). Code Meaning Explanation P Scientific concepts for pragmatics Instances in which learners were able to identify linguistic resources by using scientific concepts. Metapragmatics M+R Metapragmatics and rule Episodes in which the learners took a firm stance, or resorted to evaluations, such as valency, in order to provide a rule. M+L Metapragmatics and pragmalinguistics Episodes in which the learners reflected on language use. M+C Metapragmatics and sociopragmatics Episodes in which the learners reflected on the context. M+EX Metapragmatics and example Episodes in which the learners used an example of a specific linguistic resource, e.g., excuse me, or provided a request. M+P Metapragmatics and scientific concepts for pragmatics Episodes in which the learners used scientific concepts in their reflections. Methodology 80 Comparing the original and the adapted framework, two codes could be re-employed with minor changes: M+R, in which the learners provided a rule, and M+G, in which the learners used terminology. However, in the present study, the latter code was renamed and used exclusively when the learners used scientific concepts as part of their reflections (M+P). The codes M and M+T could not be transferred to the present study. However, two codes were added to reflect the definition of metapragmatic awareness used in the study as well as the instructional targets, that is, learners’ verbalised reflections about language use (M+L) and contextual considerations (M+C). In addition, drawing on previous research (Savić & Myrset, Forthcoming-b), as well as the data from the open coding (Elo & Kyngäs, 2008; Saldaña, 2016), the code M+EX was added, identified when the learners resorted to specific pragmalinguistic resources as examples in their reflections. Finally, an additional code (P) was added to highlight instances where the learners used scientific concepts, that is, to indicate the extent to which they had started to internalise the concepts (e.g. van Compernolle, 2014), but not to express metapragmatic understandings. The adapted framework provided the researcher with frequencies of occurrences in the various categories, thus representing the final stage of Fortune and Thorp’s (2001) analysis – quantification. In addition, the coding helped identify the episodes in which the learners used scientific concepts to scaffold their understandings. Three such episodes were selected for an in-depth analysis. These were episodes where “a) the learners collaboratively engaged in the discussion and b) the scientific concepts served different purposes for the discussion, that is, concluding remarks, a springboard for the discussion, and as prompts introduced by the researcher” (Myrset, 2021, p. 200). Drawing on previous literature (Bloome et al., 2008; Marková et al., 2007), the excerpts were analysed using discourse analysis. The analysis emphasised how the learners “jointly co-construct[ed] messages and meanings, and [how] they change[d] their positions” (Marková et al., 2007, p. 202) and what roles scientific concepts served in facilitating metapragmatic understandings. Furthermore, the discourse analysis was selected as it enabled the researcher to focus on both the content and the discursive practices taking place in the episodes (Bloome et al., 2008; Marková et al., 2007). The interview data presented in Article IV aimed to explore the learners’ perceptions about the project and was analysed inductively. In other words, the analysis was data-driven and focused on the topics emerging in the interviews (Selvi, 2020). Following Elo and Kyngäs (2008), considering the limited previous research on the topic, an inductive approach was deemed more appropriate. Due to the focus of Article IV, the dataset comprised the discussions concerning the appraisal of the project components Methodology81(see Figure 3.6). Subsequently, the analysis was conducted in three stages, which aimed to identify 1) the component of discussion, e.g. the instruction; 2) the content discussed for each component, e.g. worksheets; and 3) the learners’ evaluative positionings, through lexical items or comments. Following the coding, the codes were reorganised into categories (Elo & Kyngäs, 2008). For instance, the codes ‘dice game’, ‘cities’ (request perception task), and ‘onion circle’ were all assigned to the overarching category ‘Specific activities’. Thus, the analysis provided an overview of ‘component – evaluation – category’, e.g. Instruction – Positive – Specific activities. Figure 3.7exemplifies the result of the inductive coding for the instruction component (Class activities), also presented in Article IV.Figure 3.7 Categories occurring in the discussions about the instructionAs displayed in Figure 3.7, the analysis provided the researcher with an overview of the component discussed (i.e. Instruction), the learners’ evaluative positioning (i.e. positive-negative), and the topics raised to provide their reasonings (e.g. specific activities). In addition to providing the overview of categories, the learners’ appraisals were also analysed and quantified. Thus, the coding followed a similar pattern to the stages presented by Fortune and Thorp (2001): first, the discussions about each component were identified; second, the content and evaluative positionings for each component were organised into codes; third, the codes were reorganised into categories (Elo & Kyngäs, 2008); fourth, the appraisals of the components were quantified. The percentages of the appraisals, which were presented in the final report, along with the categories identified in the interviews, aimed to ensure a broad description of the data(Elo & Kyngäs, 2008; Krippendorf, 1989), both during the analysis and in the final report.The data-driven approach in Article IV aimed to avoid a priori interpretations of the data. However, such approaches to analysis are perhaps particularly prone to researcher biases (Elo & Kyngäs, 2008; Selvi, 2020). Thus, it is important to assess the quality of the study (see Section 3.4). This is particularly important in research with children, such as the present study, which aims to give children a voice (e.g. Lundy, 2007). As Punch (2002b, p. 326) argues, in research with children there is a “danger of imposing adult views” in the interpretations. Thus, considering the inductive approach taken in Article IV, in addition to providing the categories identified through coding and the frequencies Methodology 82 of the learners’ appraisals, the report prioritised incorporating direct quotations from the data. 3.3.4 Overview of the data collection and analysis In the present study, the overarching RQ was addressed through four articles, each of which draw on different datasets and data analyses. The overarching research question of the study is: How does a concept-based approach to teaching requests impact young language learners’ request production and awareness, and their engagement with pragmatics? Table 3.7 provides an overview of how the techniques and the analyses presented in this synopsis informed the articles. Table 3.7: The research questions, data sources and data analysis approaches in the four articles Article Instrument Focus Data analysis Software Approach I Aim: 1) To provide an overview of the methods used to elicit metapragmatic data in research with young language learners and 2) to present three data collection techniques designed and used in two research projects conducted by the authors Systematic review RT Data elicitation techniques Systematic Review Excel II RQ: To what extent does concept-based instruction of EFL requests with young learners influence - the learners’ linguistic repertoire of head acts, and internal and external modification strategies? - the learners’ linguistic variation depending on familiarity and age of the interlocutor? VODCT Request production Descriptive and inferential statistics SPSS QUAN+qual III RQ: -Do young language learners employ scientific concepts to express metapragmatic understandings following a period of concept-based instruction? If so, how? Group interviews Appraisal task Metapragmatic awareness Scientific concepts Framework for metapragmatic episodes Quantification Discourse analysis NVivo QUAL+quan IV RQ: How did young language learners appraise various components in a project related to pragmatics instruction? - What were the learners’ explanations behind their appraisals? Group interviews Appraisal task Perceptions of the project. Inductive content analysis Quantification NVivo QUAL+quan Methodology 83 3.4 Scientific quality: Quality criteria When it comes to the quality of research, Marsden (2020, p. 16) argues that methodological transparency "runs in tandem with the life cycle of a research project". In other words, ensuring the quality of the research is an ongoing process, from planning the design to presenting the data. These stages are often measured in light of internal and external validity, that is, simply put, the extent to which the outcomes are true (internal validity) and whether the findings are generalisable to larger populations (external validity) (Dörnyei, 2007; Mackey & Gass, 2005; Rogers & Révész, 2020). However, this understanding of scientific quality is rooted within the quantitative paradigm (Dörnyei, 2007; Tracy, 2010). Thus, these criteria do not transfer easily to the qualitative strand, which has resulted in proposals of new frameworks for quality criteria for qualitative research (Lincoln & Guba, 1985; Tracy, 2010), which has different aims (Lew et al., 2018). A challenge with the present case study is that mixed methods research is positioned between the quantitative and qualitative paradigms. This has resulted in various views about the role of quality frameworks within the mixed methods research paradigm, i.e. whether new frameworks should be developed or be adopted from the quantitative and/or the qualitative strand (Dörnyei, 2007; Fàbregues & Molina-Azorín, 2017). For instance, Fàbregues and Molina-Azorín (2017) conducted a systematic review to investigate the discourses regarding quality assessment in mixed methods research literature. While the study revealed an increase in literature addressing quality within mixed methods research, there was still inconsistent use of terminology and a lack of agreement on core quality criteria in the 64 publications included in the review48. Indeed, the review found that a large selection of the studies used terminology from quantitative or qualitative research. Furthermore, Fàbregues and Molina-Azorín (2017, p. 2859) found three positions concerning agreement on quality criteria: 1) agreement results in clarity, 2) agreement is “neither desirable nor feasible, given that quality is heavily context-dependent”, and 3) an intermediate position, that is, a minimum agreement on quality is necessary for clarity. Thus, Fàbregues and Molina-Azorín (2017) call for greater consistency in the terminology used, as well as agreement on the core quality criteria. Considering the current state of quality assessment within mixed methods research, the present study aligns with a view of adopting quality criteria and terminology from the qualitative or quantitative strand. This study is in essence a case study, which included – from a quantitative perspective – a small sample. Furthermore, the data collection techniques and analyses were predominantly qualitative, i.e. with the weighting in most 48 In their study, 4028 publications were screened. Methodology 84 of the articles on the qualitative strand (QUAL+quan). The study thereby favours a qualitative view on quality, presented through Lincoln and Guba’s (1985) criteria for trustworthiness. These criteria were selected as they “have the advantage of parsimony and they are frequently referred to in the literature” (Bryman et al., 2008, p. 266). Their criteria comprise credibility, transferability, dependability, and confirmability49 (Lincoln & Guba, 1985). Credibility refers to the truth value, that is, the trustworthiness of the research. Transferability relates to the extent to which the research is applicable to other contexts. Dependability refers to the stability or consistency of the findings and whether these could be reproduced. Finally, confirmability relates to the findings deriving from the data and whether they could be confirmed by others. Drawing on a number of sources (Dörnyei, 2007; Roy et al., 2019; Tracy, 2010) to supplement Lincoln and Guba (1985), Table 3.8 presents an overview of the quality assessment (criteria and strategies) for the present study, which will be discussed below. However, as displayed in the table, there is a good deal of overlap, with one strategy addressing more than one quality criterion. Consequently, rather than being discussed separately, the criteria and various strategies will be highlighted when relevant. Table 3.8: Quality criteria for the present study Criterion Strategy employed Credibility Prolonged engagement, or time in the field Mixed methods (Triangulation) Sample Data collection and piloting Data analysis and peer checking Transferability Contextualisation and thick descriptions Sample (intact classes) Elicitation techniques Data analysis Teaching material Dependability Audit trail Examining outliers or negative cases Confirmability Audit trail Thick descriptions and direct quotations When it comes to credibility, the present study, firstly, involved prolonged engagement in the research setting. Whereas the time spent is dependent on the research context (Tracy, 2010), the quantity of engagement increases the study’s credibility (Dörnyei, 2007; Roy et al., 2019). The present study involved approximately three months of 49 Dörnyei (2007) refers to these as the qualitative counterparts of the quantitative quality criteria: credibility (internal validity), transferability (external validity), dependability (reliability), and confirmability (objectivity). Methodology 85 fieldwork and data collection before and after the instruction50. This ensured that the researcher had time to gain in-depth insights into the research context. In addition, the prolonged engagement enabled the researcher to build rapport and trust with the teachers and learners (Roy et al., 2019), which is crucial for reducing the power imbalance in research with children (e.g. Kuchah & Pinter, 2012; Pinter, 2014; Punch, 2002b). The emerging rapport was particularly evident in the data presented in Article IV, where the learners gave their honest, sometimes negative, responses about the project (Holmes, 2019). Similarly, Article II reports on a learner who questioned the study design by saying that the test would not necessarily show the full extent of the learning outcomes. The inclusion of such responses in the report can be viewed as an attempt by the researcher to be sincere and honest through transparency about negative cases (Dörnyei, 2007; Duff, 2020; Tracy, 2010), thus addressing dependability. This transparency also relates to Articles II and III. In Article II some requests strategies did not reveal significant changes. However, the researcher decided to include the frequencies and test results as an appendix to provide future researchers with the additional insights. Similarly, in Article III the frequencies of all the metapragmatic episodes were included, thus showing that the use of scientific concepts occurred in a comparatively small number of metapragmatic episodes. The mixed methods research design of the study also adds to the credibility of the research (Creswell & Plano Clark, 2011; Dörnyei, 2007). More specifically, the mixing of methods allowed the researcher to explore the impact of instruction from various perspectives (Dörnyei, 2007). It enabled the researcher to explore the instructional impact on the learners’ request production (Article II), their use of scientific concepts (Article III), and their engagement with the project (Article, IV). The weighting on the quantitative strand in Article II enabled the researcher to observe changes through descriptive and inferential statistics rather than providing "an anecdotal account of events" (Turner, 2014, p. 12), with the qualitative strand providing additional insights, such as requests produced and learner comments. In Articles III and IV, the weighting on the qualitative strand provided in-depth insights about the discussions that emerged in the interviews, whilst the quantitative strand, in the form of frequencies, helped identify relevant episodes for qualitative analysis and revealed the extent to which such episodes were represented in the data. Thus, the use of both strands strengthened the research. In addition, the use of an interview guide ensured that all the groups were asked the same questions, which provided the researcher with a rich dataset to explore 50 Considering the prolonged engagement, it was important to account for participant attrition (Dörnyei, 2007), that is, the lack of physical or attentive presence. For instance, if learners missed parts of the study, this should be accounted for by, for example, excluding them from the analysis. With this in mind, the study included a cut-off for the instruction at 75% attendance (Bardovi-Harlig, 2015), which all the learners exceeded. Methodology 86 the research phenomenon from various perspectives (Dörnyei, 2007; Mackey & Bryfonski, 2018), so that the data "provide[d] for and substantiate[d] meaningful and significant claims" (Tracy, 2010, p. 841). Furthermore, the sampling resulted in two intact classes, with 11 groups (46 learners) generating the data in the VODCT (Article II) and the group interviews (Articles I, III, and IV). Therefore, in addition to using a range of methods, the combination of multiple groups in the sample and data collected at various stages of the fieldwork contributes to the credibility by providing rich data, while the use of intact classes yielded data that may be transferable to other contexts (Lincoln & Guba, 1985; Tracy, 2010). Mackey and Gass (2005) argue that reflection about the design of the study is important. This is particularly important in studies with prolonged engagement, in which the researcher reflexively plans a thorough design which is both systematic and flexible at the same time (Watt, 2007)51. This ensures that the elicitation techniques generate data relevant for what is being investigated (Creswell & Plano Clark, 2011), which is further ensured through piloting the instruments (Dörnyei, 2007). Furthermore, two of the techniques, namely the Emoticon task, which was adapted in the current study, and Readers Theatre (RT), were scrutinised in Article I, both in relation to the data generated and their instrumentation (Punch, 2002b). In terms of the transferability of elicitation techniques, Articles I and IV in particular provide thick descriptions of the techniques and their procedures (Tracy, 2010), allowing for replication or adaptation in future studies and other contexts. To ensure credibility of the study, the techniques were piloted. For Article I the researchers and a university librarian separately conducted a trial run for the systematic review (Booth et al., 2016). This piloting of the search terms helped refine the string searches that were used in the final search. In addition, this article presents data elicited through RT, which the researcher had used in a previous study with Norwegian EFL learners in 6th grade (Myrset, 2014). He was thus familiar with the structure and the potential challenges for the learners of a similar age when producing and performing 51 Watt (2007) suggests writing about the design in journal entries in order to reflect on the data collection techniques. Drawing on this, the researcher created mind maps and timelines that aimed to provide visual overviews of the fieldwork, that is, the stages of data collection, the types of data collected, etc. This enabled the researcher to critically reflect on the design and discuss it with peers (peer-checking) by drawing on the mind map. The reflexive engagement with the techniques and the timeline can be traced to methodological decisions in the study, for instance, in the attempts of not imposing the researcher’s views: firstly, the open-ended questions aimed to reduce the risk of asking leading questions during the data collection and enabling the learners to speak freely. Secondly, the questions leading up to the requests in the VODCT avoided words that could influence the request production (e.g. ‘want’). Finally, the instructions and discussions regarding the appraisal task were based on the emoticons and their colours, to avoid the researcher imposing words or evaluations on the learners. Methodology 87 the texts. For Article II, the researcher employed the VODCT, which also had been used in a previous study (Savić & Myrset, Forthcoming-b) with a similar age group. In the current project, he employed some of the videos from the previous study and structured the questions in the same way. As a result, he was familiar with the instrumentation of the technique and the types of data it would generate. Moreover, his previous experience with the instrument ensured that the tests were conducted in the same way with all the groups. For Articles III and IV the instruments included appraisal tasks, which were adapted from the Emoticon task used in Savić and Myrset (Forthcoming-b) with 7th-grade Norwegian learners. In addition to the researcher’s aforementioned familiarity with the techniques, the group that was not included in the data analysis, due to one learner’s experience of living in an English-speaking country, served as a pilot group for the duration of the study. Indeed, their role as a pilot group resulted in a change of the delayed post-test (VODCT). Furthermore, the pilot group ensured that the learners could understand and respond to the open-ended questions in the interviews and that the tasks were engaging. The employed data-eliciting techniques generated a large pool of data that was subsequently analysed quantitatively and qualitatively, which by itself contributes to the study’s credibility (Creswell & Plano Clark, 2011; Dörnyei, 2007). During the analysis the researcher used peer checking at various stages (Dörnyei, 2007). In Article I, in addition to cooperating with a university librarian, the two researchers cooperated in the inclusion/exclusion process. A challenge in this selection process was the inconsistent use of the term 'metapragmatic awareness' (McConachy, 2018; Nikula, 2002), or not using the label at all. With publications where the researchers were uncertain, they would discuss the study in relation to the relevant criteria and agree whether the publication should be included or not. Thus, in addition to the rigid criteria for inclusion, the collaboration between the researchers strengthened the credibility of the final choice of the studies included in the review. This is evident in the final result of the review, which includes articles not using the term 'metapragmatic awareness' but still investigating the phenomena falling within the definition of metapragmatic awareness employed in the systematic review (e.g. Bosco et al., 2006; Lee, 2010). Furthermore, a step-by-step guide of the review process in the final report ensured that the researchers provided an audit trail, which strengthens the dependability and confirmability of the study. In Article II, two stages of the analysis relate to peer checking. Firstly, during the coding, which followed a coding manual (Blum-Kulka et al., 1989), the author would discuss requests in the data with a peer who was familiar with the coding manual and had previously used it in their research to ensure that the request strategies were assigned the appropriate codes. Secondly, following the coding of the data, the Methodology 88 researcher collaborated with a statistician to ensure that the tests were appropriate for the data. This latter point addresses a potential limitation within mixed methods research, namely that researchers are often trained within one paradigm (Mackey & Bryfonski, 2018). Thus, acknowledging this limitation by collaborating with a statistician strengthens the credibility of the analysis. In addition, the provision of the raw frequencies and the use of Blum-Kulka et al.’s (1989) coding manual ensured the dependability and confirmability of the research through an audit trail. Article III involved multiple rounds of coding (Elo & Kyngäs, 2008; Saldaña, 2016) before the researcher employed an adapted framework (Fortune & Thorp, 2001). This framework was peer checked to ensure that it was a) appropriate for the material, and b) transparent in the final report. The adaptation of Fortune and Thorp’s (2001) framework and the detailed description of the framework can also be viewed as ensuring transferability, namely that the framework may be used in future research and contribute to “methodological craft skills” (Tracy, 2010, p. 846). Furthermore, in this study the transcripts were translated by both the researcher and a professional translator separately (Thompson & Dooley, 2020), which strengthens their credibility and reduces the possibility of misinterpretation of the data. Finally, the excerpts that were analysed in-depth were presented through thick descriptions and direct quotations in the final reports. This strengthens their quality through an honest account of the researcher’s interpretations (Tracy, 2010), ensuring the dependability and confirmability of the results by leaving the excerpts as an audit trail. Finally, Article IV follows a similar pattern to Article III. During the coding stage, the researcher conducted multiple rounds of coding. Following these rounds, the analysis was peer checked to ensure that the researcher did not misinterpret the learners’ responses before abstracting the data to overarching categories (Elo & Kyngäs, 2008). Furthermore, the transcripts were translated separately by the researcher and a peer to ensure their credibility. Direct quotations from these transcripts were included in the article to ensure confirmability. The frequencies of the learners’ appraisals also provide contextualisation of the findings. In addition to the data presented, this article also provided thick descriptions of the techniques and activities that the learners appraised, providing further contextualisation. These thick descriptions may thus also result in transferability of practical use (Tracy, 2010), in which the materials and procedures may be used in the language classroom or in future intervention studies. Methodology 89 3.5 Reflexivity Given the complex nature of the study and the weighting largely placed on the qualitative strand, it was important for the researcher to consider his own biases, an essential component in educational research (Musgrave, 2019). Reflexivity is the “careful consideration of the phenomenon under study, as well the ways a researcher’s own assumptions and behavior may be impacting the inquiry” (Watt, 2007, p. 82). The view that “the researcher is an instrument” (Lew et al., 2018, p. 83) involves reflecting on one’s own personal background, motivations, and the impact of the research (Musgrave, 2019). In the present study, the researcher’s positioning has been discussed in relation to a number of methodological choices, such as the design and researcher’s positionality (see Section 3.1); the data collection techniques informed by research with children, for instance giving the learners a voice (see Section 3.3); the quality assessment of the techniques and analysis, such as peer checking to ensure that the data was not misinterpreted (see Section 3.4); and the ethical considerations made during the fieldwork (see Section 3.6). Since these issues are highlighted in sections throughout this chapter, the current section will discuss some issues related to reflexivity that are not addressed elsewhere, particularly related to the researcher’s background, motivations, and role. When it comes to the personal background and motivations, the present study was conducted in a primary school, which was a familiar context for the researcher as a former primary school teacher. In addition to having worked as a primary school teacher, the researcher also had prior experience in tertiary education as a teacher educator and as a co-author of English language textbooks for primary schools in Norway. During this time the researcher was also involved in a cross-sectional study exploring YLLs’ metapragmatic awareness, which included revisiting schools to share the findings with the teachers. These prior experiences resulted in many discussions about English language teaching with a particular focus on pragmatics. The discussions revealed limited (explicit) focus on pragmatics. Whilst these findings were anecdotal at best, this limited focus was also apparent in the research. Thus, instructional pragmatics, with YLLs in particular, provided an interesting topic that was considered useful for both teaching practice and research. This served as the motivation for this study. The researcher’s former role as a teacher was also an important factor. In case study research and studies with prolonged engagement, the researcher can be viewed as being on a continuum between ‘insider’ and ‘outsider’. Following Hellawell (2006, pp. 484-485), the insider is someone “who possesses a priori intimate knowledge of the community and its members” and the outsider is someone who “is not a priori familiar with the setting and people s/he is researching“. Considering the researcher’s prior Methodology 90 experience as a teacher, he was an insider to the national teaching context, but an outsider in the local teaching context (school, class). This was advantageous from the perspective of building rapport with the learners and teachers (Hellawell, 2006), whilst at the same time having a personal distance from the participants. The insider-outsider role, however, is not static when conducting research with prolonged engagement. For instance, in the present research, which involved the researcher as a teacher and his conscious attempts of building rapport with the learners to make them comfortable in the data collection setting, there was a gradual shift on the continuum towards becoming an insider within the local context. Thus, the pre-planned data collection techniques and questions, and the peer checking of the analysed material were important to counter researcher biases which may have developed in concert with a movement along the outsider-insider continuum. The researcher spent almost full days in the school for several weeks. Considering his heavy presence, and the fact that he taught all the sessions for the duration of the instruction, it was important to avoid becoming too involved in the role as educator during the fieldwork, while at the same time attempting to reduce an inherent adult-child power gap (e.g. Griffin, 2019; Gu et al., 2005; Kuchah & Pinter, 2012). Following Kuchah and Pinter (2012, p. 286), rapport-building activities, such as participating in activities during recess or chatting with the participants when opportunities arise52 may reduce this power gap, thus "establishing confidence and a favourable interview atmosphere". However, drawing on personal experiences and prior research, the researcher was aware that such settings may also involve conflicts between peers. Thus, as an outsider in the local context (Hellawell, 2006), the researcher decided prior to the fieldwork that he would only be present with another member of staff who could intervene during an incident that required problem-solving. This was both an ethical and methodological consideration emerging from reflexivity prior to data collection: children may be sensitive to criticism, especially from unfamiliar adults, thus school staff would be better equipped to intervene (see, for instance, Pálmadóttir & Einarsdóttir, 2016, for a discussion about not interfering in pedagogical work). In addition, the researcher intervening in peer conflicts could potentially be attached to him, thereby affecting rapport and making the learners less comfortable with him in the data collection setting and thus influencing the data generated. Consequently, during the rapport-building activities outside of the classroom, the researcher ensured that he 52 One example of such involvement is that as part of their homework, the learners conducted interviews about requestive behaviours with an adult at home (Appendix 8). James (pseudonym) did not have someone he could interview that day and asked the researcher to be the interviewee during recess, which the researcher agreed to. Methodology 91 was not “directing their [the learners’] activities and intervening in disputes” (Corsaro & Molinary, 2017, p. 16). 3.6 Ethical considerations Making ethical considerations is "paramount for the continuing success of any research field" (Sterling & De Costa, 2018, p. 163). There is an overarching principle in research of respecting persons, doing no harm, and justice (De Costa et al., 2020). Pimple (2002, p. 192) proposes that “the ethics of any particular research product or project can be divided into three categories: (A) Is it true? (B) Is it fair? (C) Is it wise?”. The first relates to truthfulness in presenting data. The second relates to the relationship between the researcher and others, e.g. fellow researchers, and participants. The last refers to the research agenda and the world, e.g. is it morally acceptable? However, De Costa et al. (2020) point out that the ethical considerations have by and large been focusing on institutional guidelines, e.g. ethical review boards, thus not taking into account the emergent nature of ethics in the research process, so-called ‘situated ethics’ (Ebrahim, 2010). Sterling and De Costa (2018, p. 163) also point out that “research ethics takes on a different role when the data being collected and analyzed comes from human beings”. Consequently, the process of conducting research is often more ‘messy’ than a generic ethical one-size-fits-all. The ethical considerations thus pertain to both institutional guidelines and those within the research context, e.g. research field and place of fieldwork, thereby falling into two categories (Kuchah & Pinter, 2021). The first category is the legal frameworks and guidelines researchers are required to uphold, i.e. macro-ethics (De Costa et al., 2020), which will be referred to as ‘formal ethics’. The second category is the ethical considerations taken by the researcher, i.e. micro-ethics (De Costa et al., 2020), and will be referred to as ‘informal ethics’. Both formal and informal ethics are equally important in research. However, whereas formal ethics are requirements that every researcher must adhere to, informal ethics will change depending on the focus of the research. Formal ethics ensure appropriate conduct by all researchers in line with Pimple's (2002) categories, e.g. honest reports of data, co-authorship, and fair treatment of the participants, regardless of the research field. Such requirements can be found locally at the institution, e.g. university guidelines53; nationally, e.g. the Norwegian legal framework forskningsetikkloven54, the National Committee for Research Ethics in the 53 https://www.uis.no/nb/forskning/forskningsetikk-ved-uis (02.09.2021) 54 https://lovdata.no/dokument/NL/lov/2017-04-28-23 (Accessed: 02.09.2021) Methodology 92 Social Sciences and the Humanities (NESH)55 or the Norwegian Centre for Research Data (NSD)56; or globally, e.g. The Vancouver Recommendations (ICMJE, 2019) or The General Data Protection Regulations (GDPR)57. Regarding this project, all relevant considerations were made to ensure that the conduct of formal ethics was followed through an approval of the project design by NSD; equal participation for co-authorship, following The Vancouver Recommendations; storage and anonymisation of data in line with the GDPR; and, when reporting on the data, not presenting falsified or tampered data. This project includes one co-written article (Article I, Myrset & Savić, 2021). The Vancouver Recommendations (ICMJE, n.d.) state that authorship has “important academic, social, and financial implications” and “implies responsibility and accountability for published work”. This is arguably also an ethical consideration relating to Pimple's (2002, p. 192) second category (is it fair?) as it includes “issues such as relationships among researchers (authorship and plagiarism)”. The recommendations state that authorship entails substantial contributions in the design, drafting or revising the work, final approval, and accountability for all aspects of the work. In Myrset and Savić (2021), both authors were involved in all aspects of the project, from design to the final product, thus ensuring that the recommendations were followed. Following Politou et al. (2018), the relatively new regulations in the GDPR state that participants must have the right to withdraw themselves and be forgotten from the study. This has implications not only for how data should be treated if a participant chooses to revoke their consent, but also for how the consent forms are phrased. To ensure that the consent form followed the new GDPR regulations, it was approved by the NSD (See Appendix 15 for consent form and NSD approval). The NSD also approved the project with regard to how the data was being stored and anonymised during analysis, and in the report. In this project, the identifiable data, e.g. names of the participants, were kept separate from the other data and each participant was assigned a number during transcription. The raw data material was stored on an encrypted device that required a pin code to access the files. Later, in publications, the participants were given pseudonyms, thus ensuring anonymity (Pálmadóttir & Einarsdóttir, 2016; Pimple, 2002; Truscott et al., 2019). 55https://www.forskningsetikk.no/en/guidelines/social-sciences-humanities-law-and-theology/guidelines-for-research-ethics-in-the-social-sciences-humanities-law-and-theology/ (Accessed: 02.09.2021) 56 https://www.nsd.no/ (Accessed: 02.09.2021) 57 https://gdpr-info.eu/ (Accessed: 02.09.2021) Methodology 93 Moving to informal ethics, concerning the consent forms, the children participating in the study were underage, and consequently informed parental consent was required and obtained for all the participants. However, it is questionable in such instances whether the children have actually given their consent (Ebrahim, 2010; Truscott et al., 2019). Children’s consent can be given both verbally or non-verbally (Ebrahim, 2010; Pálmadóttir & Einarsdóttir, 2016). In addition, given that the parental consent form provides information about the nature of the study, the children should also be allowed to make an informed choice about participating (Truscott et al., 2019). In the current project, the researcher started the pre-test by introducing himself and asking the learners whether they knew why he was there. Their explanations would thus ensure that they had understood his role and the nature of the research project. If learners said that they did not know the reasons for his presence, the researcher explained this and answered questions before proceeding with the data collection. This process was revisited throughout the data collection to ensure the continued consent of the learners. The research setting and the instruction itself also called for ethical considerations. For instance, continuing in the vein of consent, one learner did not consent to participation. Thus, the researcher decided not to audio record during the instruction so that the learner could participate equally with their peers and not experience any negative consequences (Kuchah & Pinter, 2021; Mayo, 2021). Furthermore, considering that the teacher taught English to both the classes included in the study, it was considered unethical to assign one class as a control group, that is, to only give the instruction to one group (Mayo, 2021). Indeed, drawing on Pimple’s (2002, p.192) notion of fairness, dividing the learners into control and treatment groups was considered unethical both in relation to the learners and the teacher: for the learners, it would potentially deprive one group of the potential benefits of the instruction, and for the teacher, it would require them to prepare two different sets of lessons for the duration of the fieldwork. In addition to the English teacher, it was considered imperative to involve all the 7th grade teachers to gain their trust and acceptance (Corsaro & Molinary, 2017), and to ensure that they – like the child participants – were treated fairly (Pimple, 2002). Prior to the study, the researcher had preliminary meetings with them where he presented a detailed plan of what the project would entail to ensure that they were aware of the time required before they agreed to participate. Furthermore, during the fieldwork, the researcher held informal meetings where he updated the teachers on the progress of the study. Finally, after the project, the researcher returned to the school and presented the findings to the teachers who had been involved and to the school administration (Shamrova & Cummings, 2017), which also gave the learners a voice through providing an audience and influence (Lundy, 2007). In addition, all the teaching material developed was shared with the English teacher so that they could benefit from the project long-term. Methodology 94 Regardless of the specific research focus, all research methods employed with children need to be designed and implemented in such a way as to ensure that they are adjusted to the cognitive, social and emotional development of young research participants. Consequently, when conducting research with children, generating 'good data' is not the only concern, but also ensuring that the elicitation techniques are “non-invasive, non-confrontational, and participatory […] diminishing the ethical problems of imbalanced power relationships between researcher and researched at the point of data collection and interpretation” (Morrow & Richards, 1996, p. 100). Truscott et al. (2019, p. 21) argue that although techniques developed for research with children have been “ethically and epistemologically motivated to facilitate children’s participation in research, diffuse inherent power dynamics between children and adults, and assist researchers to ‘tune in’ and ‘listen’ to children’s voices”, this does not ensure that they are ethically fool-proof. In other words, close attention to the development of the techniques employed does not ensure that the research is conducted ethically. Consequently, Ebrahim (2010, p. 290) argues that ethics is contextually situated and requires a “reflexive stance to how ethics is mediated by situational factors”. An example of such situated ethics in this project was the role of the researcher in the school context as a former teacher and as an adult. In the current project the researcher engaged with the learners during recess (Kuchah & Pinter, 2012), while at the same time ensuring that he would not have to direct or intervene (Corsaro & Molinary, 2017). In sum, ethical considerations are important for any research project (Pimple, 2002; Sterling & De Costa, 2018). In the present study, both formal and informal ethics were followed and revisited on multiple occasions during the various stages. This was considered crucial given the context of the research, that is, prolonged engagement in an educational setting with young learners. Summary of articles 95 4 Summary of articles The current doctoral research project comprises four articles. The study aimed to explore the impact of concept-based approaches for teaching pragmatics. More specifically, it examined young language learners’ request production and awareness, and their engagement with pragmatics. Considering the paucity of research focusing on YLLs in pragmatics and in applied linguistics more generally (Pinter, 2014), the articles (Table 4.1) aimed to contribute to this knowledge gap. This section presents a summary of each article. Table 4.1: Articles of the doctoral research project Article Aim/research question I Aim: 1 “[T]o provide an overview of the methods used to elicit metapragmatic data in research with young language learners through a systematic review” […] 2 To provide “a thorough description of three techniques the authors have employed to collect metapragmatic data” II RQ: To what extent does concept-based instruction of EFL requests with young learners influence 1 the learners’ linguistic repertoire of head acts, and internal and external modification strategies? 2 the learners’ linguistic variation depending on familiarity and age of the interlocutor? III RQ: Do young language learners employ scientific concepts to express metapragmatic understandings following a period of concept-based instruction? If so, how? IV RQ: How did young language learners appraise various components in a project related to pragmatics instruction? What were the learners’ explanations behind these appraisals? 4.1 Article I – “If an astronaut were on the moon…”: Eliciting metapragmatic data from young L2 learners This article aimed to explore elicitation techniques employed in previous empirical research on YLLs’ metapragmatic awareness, defined as “the ability to verbalize reflections on linguistic forms, contextual features and/or their interplay” (p. 2). Following the presentation of techniques in prior research, three techniques developed by the authors were scrutinised by discussing their affordances in the light of literature on the considerations specifically related to conducting research with children (e.g. O’Kane, 2008; Punch 2002a, 2002b), and by providing examples of data that the Summary of articles 96 techniques generated. These techniques were the ‘Emoticon task’, the ‘Ranking circle’, and ‘Readers Theatre’58. In order to explore the prior research, the authors conducted a systematic review aimed to retrieve research conducted with YLLs, defined as those aged 5-13 (Drew & Hasselgreen, 2008), specifically focusing on the elicitation techniques employed in the studies. The systematic review, which included research published between 2000-2019 in journals indexed in Academic Search Premier, ERIC, Scopus, and Web of Science, revealed sparse research with these age groups. Furthermore, the majority of the studies focused on L1 learners (n=12) in various contexts (e.g. French, English, Mandarin Chinese), whilst only 3 studies focused on L2 EFL learners (Japanese, Cantonese, and Mandarin). In addition to the limited number of studies retrieved, the systematic review revealed that the techniques employed in previous studies largely reflected those used with (young) adult learners (Culpeper et al., 2018), e.g. one-on-one interviews with learners, pre-decided topics, and the learners being expected to respond through questions-and-answers or questionnaires. With this review as a backdrop, the Emoticon task, the Ranking circle, and Readers Theatre were presented, both in relation to the methodological considerations made in the design and implementation of the studies, by drawing on literature on conducting research with children (e.g. Gu et al., 2005; Punch, 2002a, 2002b), and in relation to the data that they generated. As opposed to the majority of the studies retrieved through the systematic review, the three techniques presented were employed in groups (3-5 learners), thus empowering the learners through peer activities (Gu et al., 2015; McKay, 2006). The Emoticon task was a request appraisal task in which the learners appraised the requests they themselves had produced through a VODCT, followed by a discussion in which they were invited to explain their choices. Three to four requests were selected for appraisals with each group, where learners were asked to place stickers of different colours on an A3 sheet with three emoticons, based on whether they thought the request was ‘nice’ (������������������������), ‘so-so’ (������������������������), or ‘not so nice’ (☹). These appraisals enabled individual, non-verbal responses. In addition, the different colours of the stickers ensured that the appraisals could be traced to the individual learner and served as a visual representation and reminder of the choices they had made during the ensuing discussion. The data revealed interesting findings both in relation to the methodology and metapragmatic awareness. Methodologically, the data showed that although the 58 The techniques presented were used in two research projects: a cross-sectional study with 9-13-year-olds (see also Savić, Forthcoming; Savić & Myrset, Forthcoming-a, Forthcoming-b), and the current doctoral research project. Summary of articles 97 requests were produced by the learners themselves in the VODCT, they ascribed the requests to the characters from the videos. Furthermore, the emoticons and their colours (green (������������������������), blue (������������������������), and red (☹)) were used as part of the meaning-making process, for instance, by using the colours to make evaluations about requests (e.g. “it [the request] was better than red [☹]”). Thus, the visuals enabled the learners to explain their choices without access to metalanguage. Indeed, both the researcher who facilitated the discussions and the learners used the colours as a point of reference in the discussion. This enabled the researcher to use language relatable to the learners and avoid researcher influence through terms such as im/polite and in/appropriate. The metapragmatic data generated through the Emoticon task showed its potential for eliciting relevant data on various aspects of requesting. For instance, the exchanges presented revealed how the learners viewed certain features as polite, e.g. asking (‘can’) rather than demanding (‘I will have’), and that the learners attended to the content of supportive moves (compliment), discussing how to reformulate them in an attempt to improve the request. The Ranking circle was a group discussion and ranking task. The activity aimed to explore the learners’ awareness of linguistic and contextual features related to EFL requests and employed two sub-tasks. First, the learners engaged in a discussion prompted by the question ‘What is important to think about when we ask for something in English?’, thus ensuring that the topics were learner-generated (Pinter & Zandian, 2014). Second, the topics raised within each group were ranked by the learners on a sheet with three concentric circles. Three circles represented the perceived importance of the topics: the inner circle represented the most important issues, whereas the outer circle represented the least important ones. The placement of each topic required the learners to agree on its importance. Thus, the task involved two rounds of discussions, i.e. brainstorming topics and agreeing on their importance, providing additional opportunities for reflection and contribution (Punch, 2002a). The exchanges presented revealed that the task created opportunities to co-construct metapragmatic understandings, focusing for instance on the interplay between politeness and context. In addition, both the L1 and L2 were used as frames of reference in the discussion. Through the task, the learners were also provided with opportunities to compare Norwegian and English requestive behaviours, e.g. discussing the use of address terms, based on the learners’ perceptions. In the case of address terms, the learners’ perceptions about different L1/L2 uses of address terms were reflected in the Summary of articles 98 different rankings, with address terms being perceived as very important in English and not important in Norwegian59. The third and final elicitation technique presented in the article is Readers Theatre (RT) (see also Section 3.3.2), which aimed to explore the learners’ reasonings about request production during script writing. The data presented was elicited through interviews where the learners were provided with hard copies of the scripts they had produced in the second RT cycle. The exchanges presented in the article revealed how the learners displayed understandings of request variations depending on the interlocutor, i.e. an interplay between linguistic and contextual features. For instance, familiarity and relationships (friend-friend and parent-child) were highlighted in order to discuss the use of request strategies. Furthermore, the L1 was used as a frame of reference with the learners considering the way in which they interacted with their own parents, both when writing scripts and providing their reasonings for specific request formulations in the interview. Although not all the groups generated elaborate discussion related to the writing process, the examples provided in the article revealed RT’s potential for eliciting metapragmatic data. The three techniques were organised in ways that aimed to enable participation and engagement (Gu et al., 2005; Punch, 2002b). For instance, the learners were organised in groups and were seated in a semi-circle, in which the tasks were in focus. Furthermore, the learners were constantly reminded that the researcher(s) were not looking for a correct answer, but rather that the learners should provide their opinions. The authors argue that such considerations seemed to have created a non-threatening atmosphere. Furthermore, to ensure reliability (Punch 2002b), the learners were asked to comment on how they perceived the tasks themselves, which they seemed largely positive towards. The article revealed how there is a need within the research field to explore YLLs’ metapragmatic awareness. Furthermore, the previous studies focusing on YLLs were largely informed by techniques used with adults, thus not taking into account considerations that may be highly relevant with YLLs. With this in mind, the article contributes to the research field by a) proposing novel, participant-friendly techniques for eliciting metapragmatic data with young L2 learners, and b) bringing the discussion 59 Thorough analyses of exchanges prompted by the Ranking circle have been presented in Savić (Forthcoming) and Savić & Myrset (Forthcoming-a). Summary of articles 99 about considerations relevant for conducting research with children into the limelight of pragmatics research. 4.2 Article II – 'You could win Masterchef with this soup. Can I get some more?' Request production and the impact of instruction on young EFL learners. This article explores the impact of pragmatics instruction, specifically focusing on the pragmalinguistic dimension and whether the instruction had an impact on the learners’ request production by broadening their linguistic repertoire. Pragmatics instruction informed by SCT places emphasis on agency (Henery, 2015; Morollón Martí, Forthcoming; van Compernolle, 2014) i.e. making informed choices in interaction. In this article it is argued that a prerequisite for agentive language use is having a broad pragmalinguistic repertoire. With this in mind, the study employed a VODCT, adapted from Article I (Myrset & Savić, 2021), to prompt request production in a pre-, post-, and delayed post-test in order to explore changes in request production through statistical analyses: first, by investigating the overall changes in the use of head acts and internal and external modification strategies, and subsequently by exploring differences depending on the context (familiarity with and age of the interlocutor). Additionally, some illustrative examples of requests produced and comments by the learners provided further insights in the discussion of the paper. Thus, the article incorporates the mixing of methods, although the weighting is largely placed on the quantitative strand. The analysis of changes in the use of pragmalinguistic resources paints a mixed picture, with changes in some categories, whilst others remained static. For instance, when it comes to directness, there were significant changes in the use of direct and conventionally indirect requests between the pre- and post-test. However, these changes disappeared in the delayed post-test, suggesting no longer-term retention. Modals, on the other hand, revealed significant longer-term changes, with a decreased use of can and an increased use of may and could, the changes in the latter being the most significant (p < .001). Similarly, the use of supportive moves almost doubled from the pre- to delayed post-test, and the increased use of sweeteners longer-term could be observed in the z-test (Z= -2.59, p = .01). As for attention getters, address terms, and lexical downgraders, few significant changes were identified, and in the case of lexical downgraders, please was the preferred strategy. In order to explore changes in the use of pragmalinguistic resources depending on familiarity and age of the interlocutor, the request situations in the videos in which the learners produced the requests were divided into three categories of familiarity and Summary of articles 100 age60: unfamiliar adult, familiar adult, and friend. The analysis revealed significant differences in directness, particularly in requests produced to a friend, namely the use of hints increased from the pre- to delayed post-test (Z = -2.97, p = .003). With regard to modals, which had shown significant changes overall, the most notable differences were observed with unfamiliar adults. There were no statistically significant changes among lexical downgraders, with please being dominant with all interlocutors. However, downgraders were used more often with familiar (26.3%) and unfamiliar adults (28.8%) compared to with friends (9.9%). When it comes to supportive moves, they were used more often with unfamiliar adults, and sweeteners were only employed in situations with these interlocutors. In the case of familiar adults, all but one instance – a preparator – were grounders. Thus, the chi-square test revealed statistical significance with unfamiliar adults (X2 = 6.3879, p = .041) and friends (X2 = 8.5849, p = .014), but not with familiar adults. Overall, the results reveal that the majority of requests produced were conventionally indirect, which reflects the findings from previous research with young Norwegian EFL learners (Savić, 2015). Furthermore, some learners chose to comment on their use of hints during the VODCT, which suggests that the learners had internalised a conceptual understanding of their communicative function and willingly externalised these during the test. This comprehension of hints is in stark contrast to findings from previous research in which learners produced hints, but had difficulties comprehending them (Savić & Myrset, Forthcoming-b). The increased variation in request strategies following the instruction suggests that the learners had acquired a wider repertoire, which influences their ability for agentive language use. The lack of variation in attention getters and address terms, where learners mainly resorted to strategies with which they were already familiar, could also be an indication of agency. For instance, although the address terms Mr and Mrs, which are near extinct in the Norwegian context (Fretheim, 2005), were introduced, the learners did not seem to employ these, suggesting that these were at odds with their culturally situated preferences. In the case of external modification, the study revealed an overreliance on please. At the same time the learners started to use other forms (perhaps and possibly), which suggests a potential for more focused attention with this strategy. Finally, with supportive moves, the results revealed an increase, both in frequency and in variation. Compared to previous research with learners in this age group, this suggests an expansion of the learners’ linguistic repertoire. Furthermore, the ways in which the learners produced requests suggests that they had become familiar with the function of supportive moves, enabling them to vary the content and orientation (object/hearer) of the sweetener. 60 Alerters and internal modification strategies were not analysed due to the lack of significant changes in the overall tests. Summary of articles 101 The article shows how teaching requests through a focus on scientific concepts may have an impact on the learner’s linguistic repertoire, thus providing the learners with a foundation for agency. The study adds empirical evidence to the sparse research on pragmatics instruction with YLLs by focusing on Norwegian learners. Furthermore, it shows that explicit instruction also has a potential with 12/13-year-old learners. In the case of concept-based approaches to teaching pragmatics, this study expands the scope of SCT instructional studies, which have so far solely focused on adults (e.g. Nicholas, 2015; van Compernolle, 2014; van Compernolle et al., 2016; van Compernolle & Henery, 2014), and provides evidence of the potential of focusing on concepts with YLLs. 4.3 Article III – Scientific concepts as meaning-making resources for young EFL learners in the learning of pragmatics The instruction in this research project focused on introducing scientific concepts relating to requests. Following this vein, Article III investigates the learners’ use of scientific concepts to articulate metapragmatic understandings following the instruction. The theoretical framework is thus rooted in SCT (e.g. Vygotsky, 1934/2012), specifically related to pragmatics and concept-based instruction (e.g. Morollón Martí, Forthcoming; van Compernolle, 2014, see also Section 2.2). Furthermore, this article aligns with holistic perspectives of metapragmatic awareness (e.g. McConachy, 2018; McConachy & Liddicoat, 2016; Morollón Martí, Forthcoming; van Compernolle, 2014), which was defined as “being displayed through verbalised reflections about language use, contextual considerations, or their interplay, to varying degrees of sophistication (McConachy & Liddicoat, 2016; Myrset & Savić, Forthcoming)” (Myrset, 2021, p. 192). The article draws on data generated in group interviews following the instruction. As part of these interviews, the Emoticon task (Article I) was adapted in order to prompt discussions. The analysis entailed a mixed methods approach, resulting in both frequencies of occurrence of metapragmatic episodes and in-depth analyses of some episodes, with weighting on the latter. The raw data was first analysed through a coding framework, adapted from Fortune and Thorp (2001), which aimed to identify metapragmatic episodes, i.e. “identifiable units of collaborative dialogue in which learners display metapragmatic awareness, with or without the researcher as a mediator” (Myrset, 2021, p. 192). This coding provided the frequencies of the various types of discussions that surfaced in the interviews: 1) metapragmatics with a rule, e.g. through the use of valency; 2) metapragmatics grounded in pragmalinguistics; 3) metapragmatics grounded in sociopragmatics; 4) metapragmatics grounded in a Summary of articles 102 linguistic example, e.g. by providing a request; or 5) metapragmatics grounded in the use of scientific concepts (see Table 2 in the article). With regard to answering the question of whether the learners used scientific concepts as part of their discussions, the coding revealed that learners did indeed employ scientific concepts related to requests in their discussions. However, episodes in which scientific concepts were used occurred much less frequently (n= 20) than other categories, e.g. metapragmatics grounded in pragmalinguistics (n=176) (see Table 3 in the article for frequencies). Indeed, the episodes containing scientific concepts comprised only 3.5% of the total number of episodes and were the only category that did not appear in all groups. This supports claims by Vygotsky (1934/2012) that the process of internalising concepts is long and complex. The coding also enabled the researcher to identify the episodes in which the learners used scientific concepts to express metapragmatic understanding. Three such episodes were selected as they presented instances of learners collaboratively engaging in discussions and where the scientific concepts served different purposes for expressing understandings. The episodes were analysed focusing on the content and the discursive practices in the discussions (Bloome et al., 2008; Marková et al., 2007). The in-depth analysis of the episodes revealed that the collaborative dialogue enabled the learners to co-construct meaning and offer support to each other. Furthermore, a range of topics surfaced in the discussion. The first discussion focused on the choices, i.e. agency, related to requesting, in which the learners incorporated sociopragmatic features by contrasting a friend with a distant interlocutor, and valency was used as an evaluative frame (Kádár & Haugh, 2013). Importantly, towards the end of the episode, as a concluding remark, a learner resorted to scientific concepts to ground the discussion and demonstrate an awareness of the interplay between pragmalinguistics and sociopragmatics. More specifically, the concepts were used to highlight agentive language use by emphasising the importance of knowing the difference between levels of directness in order to make informed choices in communication. In the second episode, the discussion was initiated through the use of a scientific concept to discuss the communicative value of hints. This focus on the communicative value (appropriateness) of hints was a noteworthy finding in the light of previous research by Savić and Myrset (Forthcoming-b) in which the learners produced hints, but seemed insecure when appraising them. In Article III, through a scientific concept and their understanding of hints as requests, the learners could discuss the request in light of evaluative frames, the hearer’s perspective, and L1 behaviours. Summary of articles 103 The third and final episode presented a discussion in which the learners compared request strategies in the L1 and the L2. This discussion developed from an impromptu question by the researcher and showed how the learners had gained deeper insights into their own L1 by using the scientific concepts introduced in L2 instruction to produce request strategies in the L1. Thus, the episode supports claims by Vygotsky (1934/2012, p. 207) that “a foreign language facilitates mastering the higher forms of the native language”, where the L2 instruction facilitated deeper insights into the L1. What the findings in this article suggest is that a focus on scientific concepts in instruction provides tools for reflection about language use. Furthermore, it provides evidence that explicit instruction may also be beneficial with learners in these age groups. This is particularly important considering the sparse research on YLLs, since claims about the potential for pragmatics instruction have largely been based on findings with adults or on general YLL characteristics (e.g. Ishihara, 2010; Plonsky & Zhuang, 2019). Thus, the findings suggest that language teachers should aim to develop YLLs’ metapragmatic awareness, and explicit input, in this case through scientific concepts, may enable such a development. 4.4 Article IV – Giving young language learners a voice: learner feedback on pragmatics instruction The final article of this thesis investigates the learners’ perceptions about the project. Such perceptions about instruction and research projects seem to be largely overlooked among both child and adult participants. However, within the overarching discourse of conducting research with children, providing the participants with a voice in research has been increasingly emphasised, especially following the introduction of the United Nations Convention on the Rights of the Child (UNCRC). Furthermore, children’s right to be heard is stated in the Norwegian core curriculum. Thus, this paper aimed to address this gap by inviting the learners to appraise and comment on various components in the project. The project and the learners’ feedback are presented in this article through Lundy’s (2007) four facets of voice, i.e. Space, Voice, Audience, and Influence. The aim was to provide insights from the learners who participated in the project. In order to elicit these insights, the researcher used an adapted version of the Emoticon task in Article I (Myrset & Savić, 2021), in which the learners appraised four pre-decided components of the project (Readers Theatre (RT) cycles 1 and 2, the instruction itself, and the VODCT), as well as two aspects/topics of their choice. The topics raised by the learners were, for instance, specific activities, their own RT performances, group work, and the researcher. Thus, the open slots covered a wide range of topics and Summary of articles 104 increased the learners’ participation by giving them agency in choosing the topics they found relevant. The article draws on data from the group interviews conducted in the week following the second cycle of RT, and presents both descriptive statistics, i.e. percentages of the learner appraisals, as well as recurring categories and quotations from the interviews. Thus, the article incorporates a mixing of methods, with the weighting on the qualitative strand. The interview data was coded inductively to avoid a priori interpretations, and aimed to identify 1) the project component being discussed, e.g. RT cycle 1; 2) the categories occurring in the discussions of each component; and 3) the lexical items signalling the learners’ evaluative positionings, i.e. positive or negative stances. When it comes to the appraisals, the descriptive statistics revealed that the learners were generally positive towards both the instruction and RT. The VODCT, on the other hand, was appraised more negatively. The recurring categories when the learners discussed the VODCT were the monotony of the task and the videos used being childish. It is likely that the monotony of the VODCT was a result of the repetitive nature of the task, with eight videos being played in a consecutive order and paused before each request scenario. Such insights are valuable in the research field as the DCT has been has been extensively used in the field, but only scrutinized in relation to the authenticity of the data rather than in relation to learner perceptions (e.g. Ishihara & Chiba, 2014; Taguchi & Kim, 2016). Furthermore, the negative appraisals provide credibility to the findings of the study as a whole as they indicate that the learners offered honest accounts of their views. This suggests that the learners were comfortable in the research context and that the researcher had managed to reduce an inherent adult-child power gap (Kuchah & Pinter, 2012; Punch, 2002b). Moving to the components that were appraised positively, both cycles of RT were appraised positively by the learners. The recurring categories in both cycles were: the autonomy the learners experienced, collaboration with peers, and the novelty of RT. In addition, the atmosphere of RT, which gave the learners confidence to speak aloud in front of their peers, was highlighted. When it comes to autonomy, the learners pointed to the decision-making and use of imagination involved when producing scripts. Furthermore, the learners also highlighted the collaborative aspect, where some learners found working in groups a fruitful enterprise, whilst other groups found it challenging to collaborate. The majority of the learners seemed to be positive towards RT and the topics occurring seemed to largely mirror previous research on the positive effects of RT as a method for developing literacy and oral skills (Drew, 2018). Considering that RT presents a novel approach within pragmatics, it is argued that the findings presented Summary of articles 105 in the article show potential for further explorations of RT in pragmatics research and teaching. The instruction was also appraised positively, the common categories for positive appraisals being: learning outcomes, the novelty of the instructional focus, and specific activities. The learners seemed to perceive the instruction as having provided them with more linguistic resources. In addition, the learners highlighted choices related to requesting, indicating that the aim of fostering agency and moving away from teaching rules of thumb (e.g. Liddicoat & McConachy, 2019; Nicholas, 2015; van Compernolle, 2014) had made an impact on their language use. With regard to the novelty of instruction, some groups mentioned movement as a positive aspect. Furthermore, the focus itself (requesting) and the instruction materials were highlighted as novel aspects of the instruction. In addition, the article argues that the researcher teaching the material, instead of the learners’ English teacher, may have added to the novelty of the instruction. Indeed, some groups chose to appraise the researcher in the open slots. The open slots also provided the learners with agency in the data collection and many groups seemed to use this opportunity to provide nuances to what they had discussed in the pre-decided components, for instance by appraising specific activities from the instruction. This article contributes to the field of pragmatics research by emphasising the importance of involving learners in the research process. Through giving learners a voice, this article provides insights into their perceptions about the relevance of requesting as an instructional target, the general teaching approach and specific activities both in terms of perceived learning outcomes and learner engagement. Such insights are crucial for advancing our understandings of how pragmatics can be taught and how it can be researched with these age groups, as well as how it can be made relevant to YLLs’ lives. Summary of articles 106 Discussion and conclusion 107 5 Discussion and conclusion This final chapter discusses the results obtained in the four articles included in the doctoral research project. Following the discussion of the results, the limitations of the study are stated, as well as the implications for teaching and researching pragmatics with YLLs. Finally, the study’s contribution to the field is discussed before the chapter is concluded. 5.1 Overall findings The articles included in this doctoral research project aimed to answer the following overarching research question: How does a concept-based approach to teaching requests impact young language learners’ request production and awareness, and their engagement with pragmatics? When it comes to the impact of the instruction, the findings in Articles II and III reveal that pragmatics instruction influenced the young learners’ appropriation of linguistic resources and scientific concepts related to requesting. From the perspective of agency, this is a valuable insight, namely agency requires a broad pragmalinguistic repertoire in order to make informed choices and to act on these to create meaning (e.g. Levi & Poehner, 2018; Martin, 2004; Mercer, 2011; van Compernolle, 2014). Furthermore, Article IV revealed that not only did the instruction provide a foundation for agentive language use, but the learners were also aware of their own development and the opportunities their new knowledge offered in communication. Thus, the instruction had provided a foundation from which the learners could produce a variety of request strategies (Article II), engage with scientific concepts in metapragmatic reflection (Article III), and reflect on their own development and engagement in learning (Article IV), thus aligning with Vygotsky’s (e.g. 1978, 1934/2012) views on the role of education (formal learning), that is, to scaffold children’s development into self-regulated learners (Kozulin, 2018). On this path to becoming self-regulated learners, reflection and metacognition are key dimensions (Kozulin, 2018), and scientific concepts play an integral role in this development as these facilitate abstraction (Fox & Riconscente, 2008; Vygotsky, 1934/2012). Thus, the focus on teaching pragmatics through scientific concepts, which has previously been shown to facilitate development with (young) adult learners (e.g. Morollón Martí, Forthcoming; Nicholas, 2015), has also yielded positive results in the present study. The results presented in Articles II, III, and IV thereby show the affordances of this instructional approach for fostering agency even with YLLs. Discussion and conclusion 108 During the instruction, reflection was viewed as an important tool for developing agency (Kozulin, 2018; McConachy, 2013, 2018; van Compernolle, 2014, 2018; Vygotsky, 1934/2012, 1978). Three perspectives in relation to reflections were studied: how the learners used scientific concepts in their reflections (Article III), reflections related to their perceptions about the project (Article IV), and designing data elicitation techniques to prompt reflections (Article I). Following Zuckerman (2004, p. 10), highly developed reflection constitutes three main abilities, namely “(a) to consider the goals, motives, methods, and means of one’s own and other people’s actions and thoughts […] (b) to take other people’s point of view […] and (c) to understand oneself; study one’s own strong points and limitations in order to find the ways to excel or to accept one’s shortcomings”. Whereas Zuckerman (2004) argues that these are achieved in adulthood, the emergence of all these abilities was identified in Articles III and IV, namely the learners’ ability to consider the goals and motives of actions and take on the perspective of others (Article III), as well seeing their own language development (Article IV). This provides another indication of the affordances of the instructional approach adopted in the present study: the instructional focus on reflection had made an impact on the learners’ zone of proximal development (Holzman, 2018; Kozulin, 2018; Vygotsky, 1934/2012, 1978; Zuckerman, 2004). In other words, the repeated engagement with the material through reflections with their peers and the researcher, provided support for the learners to reorganise their knowledge (Vygotsky, 1934/2012, 1978). Continuing in the vein of reflection and agency, the concept-based approach provided learners with knowledge that directly influenced their ability to articulate their understandings about pragmatics phenomena (Article III). The concepts served as an orienting basis when expressing the learners’ understandings in verbalised reflections (e.g. Gal'perin, 1979; Negueruela, 2003; van Compernolle, 2014) Thus, in addition to having access to a range of pragmalinguistic resources (Article II), agency also involves making choices about using these in concert with the sociopragmatic dimension (e.g. Al Jumah, 2021; Nicholas, 2015; van Compernolle, 2014, 2018). Action is the result of pragmalinguistic and sociopragmatic choices (van Compernolle, 2014). What Article III revealed is that the learners started externalising conceptual knowledge related to requesting when articulating their understandings about requesting. Their discussions included a focus on both dimensions. An interesting finding from this study is that not only did the concepts enable deeper insights with regard to English requests, but they also provided a framework for the learners to gain insights into their L1 (Vygotsky, 1934/2012). The learners’ emergent use of scientific concepts suggests that the concept-based approach provided the learners with a framework in which they could act and assign meaning rather than resorting to rules of thumb (e.g. Liddicoat & McConachy, 2019; McConachy & Liddicoat, 2016; van Compernolle, 2014). However, these Discussion and conclusion 109 concepts were used much less frequently in the metapragmatic reflections than other categories. Thus, it was only a relatively small number of learners who could readily use them in discussions, suggesting that the internalisation of scientific concepts is a process that takes time (Vygotsky, 1934/2012). Considering that the instruction was relatively short (four hours), the findings reveal a potential for teaching pragmatics through concept-based approaches with 12/13-year-old learners. Although internalisation requires time for the learners to readily externalise conceptual knowledge, they already showed vast progress in this respect. The relationship between scientific concepts and reflection highlights another important dimension, namely learners’ metapragmatic awareness (Article III). Pragmatics instruction informed by SCT adopts a holistic perspective on metapragmatic awareness, which is closely linked to fostering agency as opposed to teaching rules of thumb (e.g.McConachy, 2018; Nicholas, 2015; van Compernolle, 2014). The findings from Article III revealed that the learners started externalising the scientific concepts to express their understandings. This can be viewed such that the learners’ metapragmatic awareness became more sophisticated (McConachy & Liddicoat, 2016), as the scientific concepts enabled abstract thinking and generalisations. Thus, it adds a new dimension to the sparse research identified in Article I: the instruction enabled the learners to frame their understandings through conceptual knowledge. What the findings suggest is that the instruction had started taking the learners on a path towards internalisation, whereby they had gained in-depth knowledge about requesting and could externalise this knowledge as part of their reflections. With this in mind, the instruction provided the learners with a broader pragmalinguistic repertoire that they could choose from (Article II), and the internalised concepts enabled them to reflect about language use on an abstract and generalised plain. Building on the previous research findings indicating that YLLs draw heavily on their L1 to mediate pragmatic understandings (e.g. Ishihara, 2013; Lee, 2010; Savić & Myrset, Forthcoming-a; Savić & Myrset, Forthcoming-b), the use of both the L1 and the L2 was an integral part of the instruction. Whereas the use of the L1 is debated in language teaching (Ellis, 2012), scholars within pragmatics argue that the L1 serves as a scaffold in L2 pragmatics (e.g. Chavarría & Bonany, 2006; Eun & Lim, 2009; McConachy, 2018), which has been confirmed in empirical studies exploring YLLs’ metapragmatic awareness (Lee, 2010; Savić, 2021; Savić & Myrset, Forthcoming-a). Similar to previous research, this study has further enforced this view, where learners were invited to use the L1 as part of their reflections, both during the instruction and in the interviews. Discussion and conclusion 110 Strategic use of the L1 (and lived experiences) may provide an entry point for approaching pragmatics in the YLL classroom, especially given the scarcity of pragmatics-related learning outcomes and materials in beginner-level language teaching, most likely driven by an assumption that YLLs’ mastery of the L2 is insufficient and that pragmatics is “simply an area to be fine-tuned once the learners’ proficiency has reached an intermediate or advanced level” (Ishihara, 2013, p. 136). It is worth pointing out, however, that the researcher and the participants in this study shared an L1, which made it possible for the researcher to scaffold learning through the L1, as shown in Article III. Thus, the findings about the affordances of mediating L2 development through the L1 are perhaps particularly context dependent. Around the world, learners are becoming increasingly multilingual (Lorenz et al., 2021; Portolés & Martí, 2017), with multilingualism becoming “the norm rather than the exception” (Portolés, 2015, p. 13). Thus, the language classroom is becoming increasingly multilingual as well. This offers a new set of opportunities for building on the multitude of language resources in the classroom and raising awareness of inter- and intra-language variation in pragmatic behaviours, with learners coming to understand pragmatics from diverse linguistic and cultural perspectives (see, for instance, chapters in McConachy & Liddicoat, Forthcoming). At the same time, this requires a different set of teacher competences (Lorenz et al., 2021). The path towards acquiring such competences arguably begins in teacher education (Krulatz & Dahl, 2016; Portolés & Martí, 2017). Whereas the multilingual perspective goes beyond the scope of the present study, the findings indicate that shared languages may be successfully utilised in pragmatics instruction, thus providing insights for future teaching and teacher education. Future research in the Norwegian context could pursue this avenue further, especially since the acknowledgment of multilingualism as a resource is also reflected in the new English subject curriculum in Norway, where learners should be able to “explore and talk about some linguistic similarities between English and other languages that the pupil is familiar with and use this in their language learning” after 7th grade (Udir, 2020b). Continuing with the curriculum, the present study was grounded in aims from the English subject curriculum, the LK06 (Udir, 2006a), but is even more relevant in the light of the new LK20 curriculum (Udir, 2020b)61. The instructional approach in the current study sheds light on how some learning aims could be further nuanced, i.e. that the learners are indeed capable of more in-depth understandings than viewing certain expressions as inherently polite, and how more complex goals from the core curriculum and interdisciplinary topics can be addressed in concert. On the one hand, the LK20 61 Considering the recent renewal of the national curriculum, which was introduced during the writing of this thesis (August, 2020), the findings of the study will be discussed in relation to the LK20. Discussion and conclusion 111 states that a learner should “express himself or herself in an understandable way with a varied vocabulary and polite expressions adapted to the receiver and situation” after 7th grade (Udir, 2020b, the author's highlights), with corresponding competence aims after 2nd and 4th grade. Meanwhile, the central values of the English subject offer a view that “English shall help the pupils to develop an intercultural understanding of different ways of living, ways of thinking and communication patterns” (Udir, 2020b, the author's highlights). This is further emphasised in the interdisciplinary topic Democracy and citizenship62, which states that English should help “the pupils to develop their understanding of the fact that the way they view the world is culture dependent” (Udir, 2020b, the author’s highlights). Whereas the subject-specific aims invoking ‘polite expressions’ seem to be rooted in the traditional view of politeness (e.g. Brown & Levinson, 1987; Leech, 1983), and teaching rules of thumb, in which specific language forms are viewed as inherently polite and can be mapped to specific contexts, the overarching principles recognise that understandings about language use are culturally dependent. Thus, the latter perspective can be argued to be broadly grounded in discursive views of politeness (e.g. Spencer-Oatey, 2008; Spencer-Oatey & Kádár, 2021; Watts, 2003), and supports a focus on developing metapragmatic awareness as a tool for fostering agency. What the instructional focus and subsequent findings of the present study (Articles II, III, and IV) thus suggest is that fostering agency through reflection about pragmatics is not beyond the reach of learners in primary school, but rather an aim that can and should be emphasised in primary teaching. This also relates to the choice of using the L1 during the instruction, namely that the L1 enabled the learners to articulate and challenge understandings they could not readily achieve in the L2. Finally, with regard to the methodology, the design of the data elicitation techniques in this study was heavily informed by literature on research with children (e.g. Brown & Perkins, 2019; Christensen & James, 2017; Eckhoff, 2019). Whereas this focus is most explicitly emphasised in Articles I and IV, all the articles were influenced by this paradigm, namely the VODCT (Article II), adapted versions of the Emoticon task (Article III and IV), and the learner-produced RT scripts (Article I), were all informed by this literature. Hence, the data elicitation techniques were designed to facilitate YLL participation and expression of thoughts and thus incorporated the use of visual stimuli and pictures (e.g. Johnston, 2008; Punch, 2002b), videos and technology (Punch, 2002a; Yamada-Rice, 2017), and data elicitation in groups (e.g. Pinter, 2014). Importantly, the study aimed to provide the learners with a voice (Kellett, 2010; Lundy, 62 The renewed curriculum introduced three interdisciplinary topics, which aim to focus on societal issues from various perspectives across the subjects. These are: Health and life skills, Democracy and citizenship, and Sustainable development. The former two are to be included in the English subject. Discussion and conclusion 112 2007). Learner voices (Article IV) were included and provide insights into both the elicitation techniques and the instruction, further supporting the benefits of innovative elicitation techniques. Firstly, the learners seemed to be positive towards RT, for instance, highlighting confidence and collaboration, which has also been found in prior research (e.g. Drew & Pedersen, 2010, 2012; Myrset & Drew, 2016). Considering that RT is a ‘whole language’ approach to teaching, with both cognitive and affective benefits, this emphasises the potential of RT for data collection and instruction within pragmatics (Articles I and IV). Secondly, the VODCT, was viewed less positively by the learners, which provides new insights into the field of pragmatics. The written DCT, a commonly used technique, has been criticised for the authenticity of the data it generates (Bardovi–Harlig, 2018; Economidou-Kogetsidis, 2013; Woodfield, 2008), but less so with regard to how learners perceive such tests. In the present study, videos and oral responses were used to address some of this criticism. Considering that the negative feedback from the learners focused mainly on the content of the videos (childish) and the number of videos (monotony), this suggests that the technique itself has potential for future use. Finally, the learners were mainly positive to the target of instruction (requests), which they viewed as novel and relevant. Furthermore, the focus on agency through concepts resonated with them, thereby suggesting that pragmatics instruction is both beneficial and perceived as relevant for YLLs. 5.2 Limitations This study and the findings presented should be viewed in the light of some limitations, the two major limitations of this project relating to the sampling and the design of the instruction. The sample comprised two intact classes of EFL learners (n = 51), of which 46 learners generated the data presented in the articles. Thus, the study is relatively small-scale. This provided an in-depth study of the participants and enabled a mixed methods approach with data collection before and after the instruction. In addition, the sampling strategy, accessing participants from a specific group (homogenous sampling) through the researcher’s network (convenience sampling), may have resulted in the research being carried out in a setting in which the teachers – and possibly by extension the learners – were positive to participation, which may potentially affect the credibility of the study (Dörnyei, 2007). At the same time, the study involved time constraints, a lack of resources (the study being carried out in its entirety by the researcher), and considerable time taken from regular teaching. Furthermore, the sample being positive towards participation may result in a willingness to contribute and create a rich dataset (Dörnyei, 2007). However, since the mixed methods approach aimed to provide a detailed focus on various aspects of the chosen case, the sampling strategy and the focus on a relatively small group of learners were considered optimal for the research design. Discussion and conclusion 113 Another limitation concerning the sample is that the study does not include a control group, which is commonly used in instructional pragmatics research. The inclusion of a control group would have enabled the researcher to make comparisons with groups of learners who did not receive the treatment and would have ultimately strengthened the claims of the findings. At the same time, as pointed out in Articles I, III and IV, the considerable time spent with the classes during the instruction enabled the researcher to develop rapport with the learners. Thus, a multisite design, in which data was elicited in a different classroom context, with learners unfamiliar with the researcher or the data collected by proxy, would also have created challenges for generating comparable data, where one group would have been more comfortable with the researcher. Since the school in which the project was conducted comprised only two 7th-grade classes, the lack of a control group could have been remedied by offering the treatment to only one class, using the other as control. However, this would have limited the number of participants receiving the treatment, thus reducing the richness of the data (Dörnyei, 2007). In addition, with the same English teacher teaching both classes, it was considered unethical to conduct the instruction with one class (Mayo, 2021), both for the sake of the learners and the teacher. Finally, and most importantly, the focus of the instruction on scientific concepts related to requesting rendered the inclusion of the control group unfeasible; namely, the control group could not have been expected to use these concepts to scaffold metapragmatic understandings without having had any exposure to them. Another limitation related to the data collection may be the use of friendship groups throughout the project. Whereas this was a conscious choice by the researcher to make the learners comfortable in the research setting and reduce the power imbalance between the researcher and the children (e.g. Beauchamp et al., 2019; Gu et al., 2005; Kuchah & Pinter, 2012), the data generated in groups did not enable the researcher to explore individual responses and progress specifically related to request production. At the same time, this kind of data reflects learners’ collaborative co-construction of knowledge in classroom settings (Swain, 1997; Vygotsky, 1934/2012). More importantly, however, as revealed in Article IV, some learners appraised group work negatively, but did not want to elaborate on their negative appraisals. This may indicate that the learners were shy or did not feel comfortable sharing their experiences with the other group members (Kuchah & Pinter, 2021), which was important to keep in mind during the interpretation and presentation of the data. Such an interplay between the potential advantages and disadvantages of employing friendship groups further points to the complexity of methodological choices in research with children. As for the limitations of the instruction itself, two aspects are particularly important to consider: the researcher teaching the material and the duration of the instruction. The Discussion and conclusion 114 instruction was carried out by the researcher, rather than by the regular teacher, and as discussed in Article IV, this may have influenced the learners’ perceptions and positive attitudes towards the instruction as it added to its novelty. Furthermore, an expert mediator, i.e. someone with in-depth knowledge about pragmatics, may have influenced the results of the instruction. Thus, while the present study reveals that the approach adopted in this study is possible with YLLs, it does not provide insights into such approaches in authentic teaching contexts. The learners and the researcher also shared the same L1, which made it an accessible scaffold that could be actively used during the instruction, which may not reflect all teaching contexts. However, this suggests that the approaches and findings in this study are highly relevant for teacher education and multilingual pedagogies. When it comes to the duration of the instruction, it was relatively short (4 hours total). Thus, the findings are limited to what was achieved over the course of a month and do not reflect an authentic teaching context in which language related phenomena are introduced and revisited over time. With these limitations in mind, it is not possible to make generalisations to larger populations. Similarly, with the researcher teaching the material the results may not be directly transferable to other teaching contexts. Thus, the researcher aimed to be transparent in the articles about how the data was collected and analysed, and how the results were presented. For instance, the rich data and detailed descriptions (particularly in articles I, III, and IV) may have resonance with and be transferrable to other contexts (Tracy, 2010). Similarly, in order to account for the lack of a control group, the researcher has been careful with the ways in which the data is presented in the articles, e.g. referring to the impact rather than the effects of instruction, as the latter suggests the use of a control group. 5.3 Implications for teaching L2 pragmatics The current research project offers some implications for teaching pragmatics, both within the Norwegian context and globally. For instance, the findings from the current study suggest that providing explicit input through concepts is plausible with YLLs, and that such an approach can indeed serve to develop learner agency. Consequently, concept-based approaches may be useful in L2 pragmatics instruction with YLLs. Explicit input has more generally been favoured with adult learners (Plonsky & Zhuang, 2019). However, the sparse research on YLLs has led to uncertainties about the affordances of such approaches in YLL classrooms (Ishihara, 2010, 2013). While the present findings suggest that explicit input, with emphasis on scientific concepts, may indeed foster YLL’s pragmatic development, it is important to note that the current study explores the upper ages of YLLs, i.e. 12-13 years. Thus, some discretion is Discussion and conclusion 115 advised when considering the extent to which explicit input is appropriate with learners younger than those included in this study. This instruction adopted a view that the L1 can serve as scaffolding for L2 pragmatic development. This enabled the learners to draw on both languages (knowledge and experiences) when articulating their understandings (e.g. Chavarría & Bonany, 2006). Provided that teachers and learners share one or more additional languages, this approach may be useful to adopt when focusing on pragmatic phenomena. Thus, rather than viewing the L1 as an obstacle for L2 development, the conscious use of the L1 should be viewed as a resource that serves as a springboard in the classroom for making sense of pragmatic phenomena, where a body of individuals with different language abilities and experiences may potentially expand each other’s perspectives emerging through collaborative dialogues. This is particularly useful from the perspective of teaching pragmatics with YLLs and in the growing multilingual classrooms (Lorenz et al., 2021; Portolés & Martí, 2017). The study placed emphasis on providing learners with a voice. When asked to appraise and comment on the instruction, the learners were indeed capable of providing well thought-through and nuanced feedback. This shows that YLLs should indeed be involved and have an impact in the decision-making process (Lundy, 2007). In fact, facilitating children’s democratic engagement in matters concerning them in the school context, or learner-centred education (Kuchah & Milligan, 2021), is a right stated in both the Norwegian curriculum (Udir, 2020a) and the UNCRC (1989). However, YLLs’ right to express their views is often overlooked (Kuchah & Milligan, 2021; Kuchah & Pinter, 2021), leaving adults solely in charge of the decision-making. The findings from this study show that providing learners with a voice may offer highly useful input for teaching. Thus, rather than treating children as objects, teachers should strive to provide spaces in which YLLs become active agents in their education. However, it is worth noting that expectations about child-adult interactions and child agency are culturally engrained, and in some cultural contexts children may be perceived “as recipients, not generators of knowledge (Kuchah & Pinter, 2012)” about matters concerning them in education (Kuchah & Milligan, 2021, p. 169). Furthermore, the learners provided feedback on the instructional design (Article IV) and were generally positive to the incorporation of movement and collaboration, as well as RT. Thus, RT, which is a ‘whole language’ approach to teaching, can serve to train different aspects of language development at the same time, making it a low-threshold approach to teaching pragmatics in the classroom. Furthermore, the learners’ feedback suggests that pragmatics instruction should incorporate movement and collaborative tasks, thus acknowledging YLLs’ physical and social growth (McKay, 2006). Through Discussion and conclusion 116 movement and collaborative tasks, the learners are provided with hands-on experience in producing language as well as opportunities for reflecting on language use. Such experience enables the learners to practise the immediacy of production in communication in classroom settings, where the stakes are lower. A classroom will perhaps always be scrutinised for its authenticity, but collaborative tasks present a huge step in the right direction. 5.4 Contributions This study has contributed both to the Norwegian and global context of pragmatics research. Although some research has explored pragmatics in the Norwegian context, both in L1 Norwegian (e.g. Fretheim, 2005; Svanes, 1989; Urbanik & Svennevig, 2019) and L2 English (e.g. Awedyk, 2003; Brubæk, 2012; Krulatz, 2016), few production studies have explored pragmatics with YLLs (Savić, 2015; Savić et al., 2021). Thus, the current study contributes to investigating this under-represented group of learners. Furthermore, to the best of my knowledge, pragmatics instruction presents uncharted waters in the Norwegian context. Since the national curriculum (LK20) includes aims addressing topics related to pragmatics already in primary school, the current study contributes both empirically and pedagogically through presenting evidence of learners’ production, awareness, and engagement. Furthermore, the study provides teaching materials used with these learners and RT scripts, which may directly contribute to addressing learning aims in the Norwegian curriculum, especially within the newly introduced interdisciplinary topic Democracy and citizenship, as it relates to the English subject. In addition, from a global perspective, materials addressing pragmatics, for instance in language learning textbooks (Jakupčević & Portolan Ćavar, 2021; Limberg, 2016; Schauer, 2019), remain limited or tend to present oversimplified rules of thumb. Thus, the materials developed by the researcher for this study provide teachers and researchers with activities that aim to explore pragmatics in more sensitive and nuanced ways. In a similar vein, the overarching approach to teaching is a contribution to instruction. The instruction in this study adopted a concept-based approach by tailoring it specifically for learners aged 12-13. With this in mind, the current study contributes to: 1) empirical research using concept-based approaches by focusing on YLLs, and 2) the more general discussion about pragmatics instruction with young learners. This discussion has so far largely derived from evidence found in research with adults and/or YLL characteristics (Ishihara, 2010; Plonsky & Zhuang, 2019). By providing in-depth explorations of two intact classes of YLLs, with findings from learners’ request Discussion and conclusion 117 production, their reflections, and their engagement, this study provides empirical evidence showing the affordances of explicit input in instructional pragmatics. One such affordance was the learners’ use of scientific concepts to express metapragmatic understandings, which can arguably be attributed to the instruction itself. These scientific concepts served as a tool during the discussion, facilitating abstract thinking and generalising. Thus, the current study adds a new dimension to prior research on YLL’s metapragmatic awareness, both more generally (e.g. Lee, 2010; Savić, 2021; Savić & Myrset, Forthcoming-a, Forthcoming-b) and in instructional settings (Ishihara, 2013; Ishihara & Chiba, 2014). Finally, this study (Article IV) systematically investigated learner perceptions about the project. Thus, the study contributes to our understandings of how learners perceive instruction. Such learner feedback seems to be largely overlooked in pragmatics research: with YLLs and adults alike, the focus has mainly been on instructional outcomes (production and awareness) rather than the learners’ engagement with the instruction, resulting in perceptions about instruction being either excluded (e.g. Taguchi & Kim, 2016) or added as an addendum in the report (e.g. Nicholas, 2015). In this study, the systematic exploration of the learners’ perceptions of the project through feedback was informed by literature on research with children (e.g. O'Kane, 2008; Pinter, 2014; Pinter & Zandian, 2014, 2015; Punch, 2002a, 2002b). This focus on methodological considerations (Articles I and IV) highlights another salient contribution to the field, namely, designing and using innovative methods to elicit (meta)pragmatic data and facilitating learner voices in research, which has been largely overlooked. More importantly, the articles also invite for a discussion about methodological considerations when conducting research with YLLs within pragmatics. 5.5 Conclusion The case study presented herein was designed and conducted in order to explore the impact of EFL pragmatics instruction with YLLs. More specifically, informed by SCT and concept-based approaches for teaching pragmatics, the current study focused on the teaching of English requests. The aim was to investigate the impact of instruction on the learners’ request production and awareness, as well as their perceptions about the project as a whole. The research was conducted in two intact 7th-grade classes, with one group of four learners serving as a pilot group for the duration of the project. The fieldwork lasted approximately three months, with data collection prior to and following one month (4 hours) of instruction. Discussion and conclusion 118 The study comprises four research articles. Article I focused on elicitation techniques in research on YLLs’ metapragmatic awareness and served as a background for the design of the study. The remaining three articles focused on reporting the findings from the study: firstly, Article II investigated the learners’ request production; secondly, Article III explored the learners’ use of scientific concepts to express metapragmatic understandings; and finally, Article IV focused on the learners’ perceptions about the project. Overall, despite focusing on a limited number of learners, the articles have revealed a huge potential for teaching pragmatics to YLLs. The in-depth focus on various aspects of pragmatics shows that the 12-13-year-old learners were more than ready for learning pragmatics. Fostering agency through language instruction will ultimately serve as a foundation for interaction with people from diverse L1 and cultural backgrounds, where the increased linguistic repertoire enables the learners to vary between strategies and their metapragmatic awareness serves a mediating role for making informed communicative choices. Despite early calls urging researchers to explore whether, how, and when pragmatics should be taught (Kasper, 1997), the field of pragmatics still needs more evidence with YLLs. Thus, similar calls can be reiterated today with emphasis on YLLs. For instance, as research has revealed that pragmatics content is presented in YLL textbooks (e.g. Schauer, 2019) and that pragmatic development is evidenced in YLLs (e.g. Savić, 2015), a potential research avenue could be to explore the current state of pragmatics instruction in classrooms through observations and teacher interviews. In other words, research could explore whether teachers of YLLs a) teach pragmatics, b) are aware of pragmatics, as well as c) their beliefs about pragmatics and pragmatics instruction. These would serve as a useful point of departure for teacher education and pragmatics research. Specifically related to the approach used in this study, more studies employing concept-based approaches with YLLs would provide further insights into their applicability with these age groups. Drawing on the findings and materials from the current study, future research could investigate requests and request responses, or conduct multiple cycles focusing on the pragmalinguistic and sociopragmatic dimensions. In other words, future research using such approaches could revisit both dimensions to explore whether they foster internalisation of concepts and generate more sophisticated reflections about language use. Another possible research avenue includes investigating a broader range of instructional targets. In this study requests were selected as the pragmatic target of instruction. These were selected due to their early appearance in children’s speech and their frequent use in communication. Furthermore, requests have been extensively Discussion and conclusion 119 researched, and even with the limited focus on YLLs in pragmatics research, requests have been a common focus with this group. However, other pragmatic foci, e.g. other speech acts or idioms, should also be explored in future studies. Such studies would inform the field about the pragmatic targets that are attainable with YLLs. Furthermore, studies exploring learners’ (meta)pragmatic development would also provide evidence that may inform instructional targets. In addition, future studies could explore the longitudinal impact of instruction, i.e. teaching pragmatics over the course of a semester or more. Based on the design and findings from this study, with short periods of instruction (15-30 minutes per session), there is reason to believe that in longitudinal studies pragmatics could be more fully incorporated along with other foci. It is possible that such a design would make it easier to gain access to schools and learners. Such longitudinal studies would be a useful addition to the field since research focusing on YLLs, including the current doctoral research study, is largely based on relatively short periods of instruction. In most of these studies, the instruction itself has lasted four hours or less, which limits the current state of knowledge with regard to gains through longer-term input. An additional avenue in longitudinal studies could be to explore how pragmatics can be integrated more fully into English language teaching. Specifically related to the Norwegian context, to the best of my knowledge, this is the only study that has explored pragmatics instruction with learners in primary school. The study has revealed that pragmatics can be addressed systematically in the classroom and that learners largely respond positively. However, the study focused on learners in 7th grade. Given that the curriculum states that topics related to pragmatics should be covered all through primary school, e.g. learners should be able to “ask and answer simple questions, follow simple instructions and use some polite expressions” after 2nd grade and “use a number of common small words, polite expressions and simple phrases and sentences to obtain help to understand and be understood” in 4th grade (Udir, 2020b), future studies should aim to explore how such aims can be addressed in early language teaching. Such studies would both be highly important for the Norwegian context and contribute to the field of pragmatics internationally. Regardless of the pragmatic target, length of the instruction, or teaching context, an aspect that pragmatics research should take into account is the learners’ perceptions about the research in which they are participating, which has become increasingly emphasised in research with children (e.g. Eckhoff, 2019; Fielding, 2001). However, as argued in Article IV, this should also be highlighted within the field of pragmatics, as feedback on, for instance, instruction may provide insights into the affordances of various pragmatics teaching approaches to engage and motivate learners. At the very Discussion and conclusion 120 least, researchers should reflect more critically on children’s role as participants, e.g. their agency in the research, their ability to voice their opinion, and how they are presented in the data. References 121 6 References Abbott, M. L. (2017). Using Statistics in the Social and Health Sciences with SPSS and Excel. John Wiley & Sons, Inc. Achiba, M. (2003). Learning to request in a second language: A study of child interlanguage pragmatics. Multilingual Matters. Adams, C., Lockton, E., & Collins, A. (2018). Metapragmatic explicitation and social attribution in social communication disorder and developmental language disorder: A comparative study. Journal of Speech, Language, and Hearing Research, 61, 604-618. https://doi.org/10.1044/2017_jslhr-l-17-0026 Al Jumah, K. A. J. (2021). Utilising concept-based instruction in teaching pragmatics: Exploring the development of requesting behaviour of Iraqi Arabic-speaking EFL learners [Unpublished doctoral thesis, University of Otago]. Dunedin, New Zealand. Alcón Soler, E. (2005). Does instruction work for learning pragmatics in the EFL context? System, 33(3), 417-435. https://doi.org/10.1016/j.system.2005.06.005 Alcón Soler, E., Safont Jordà, P., & Martínez-Flor, A. (2005). Towards a typology of modifiers for the speech act of requesting: A socio-pragmatic approach. RAEL: revista electrónica de lingüística aplicada, 4, 1-35. Alemi, M., & Haeri, N. S. (2020). Robot-assisted instruction of L2 pragmatics: Effects on young EFL learners’ speech act performance. Language Learning & Technology, 24(2), 86-103. https://doi.org/10125/44727 Andrews, J. (2021). "I don’t want to talk any more": Reflecting on research into young children's perspectives on their multilingual lives. In A. Pinter & K. Kuchah (Eds.), Ethical and Methodological Issues in Researching Young Language Learners in School Contexts (pp. 147-164). Multilingual Matters. Austin, J. L. (1962). How to Do Things with Words. Oxford University Press. Awedyk, W. (2003). Request strategies in Norwegian and English. Folia Scandinavica, 7, 287-300. Bailey, J. (2008). First steps in qualitative analysis: transcribing. Family Practice, 25, 127-131. https://doi.org/10.1093/fampra/cmn003 Bardovi–Harlig, K. (2015). Designing instructional effect studies for L2 pragmatics: A guide for teachers and researchers. In S. Gesuato, F. Bianchi, & W. Cheng (Eds.), Teaching, Learning and Investigating Pragmatics (pp. 135-164). Cambridge Scholars Publishing. Bardovi–Harlig, K. (2018). Matching modality in L2 pragmatics research design. System, 75, 13–22. https://doi.org/10.1016/j.system.2018.03.007 Bardovi–Harlig, K., & Mahan-Taylor, R. (2003). Introduction to teaching pragmatics. English Teaching Forum, 37-39. Barron, A. (2008). Contrasting requests in inner circle Englishes: A study in variational pragmatics. In M. Pütz & J. Neff-van Aertselaer (Eds.), Developing contrastive pragmatics: Interlanguage and cross-cultural perspectives (pp. 355-402). Mouton de Gruyter. References 122 Barron, A. (2021). Pragmatic competence in EIL. In Z. Tajeddin & M. Alemi (Eds.), Pragmatics Pedagogy in English as an International Language (pp. 19-34). Routledge. Beauchamp, G., Haughton, C., Ellis, C., Sarwar, S., Tyrie, J., Adams, D., & Dumitrescu, S. (2019). Using video to research outdoors with young children. In Z. Brown & H. Perkins (Eds.), Using Innovative Methods in Early Years Research: Beyond the Conventional (pp. 124-137). Routledge. Ben-Shlomo, O. R., & Sela, T. (2021). Conversational categories and metapragmatic awareness in typically developing children. Journal of Pragmatics, 172, 46-62. https://doi.org/10.1016/j.pragma.2020.11.010 Bernicot, J. (1994). Speech acts in young children: Vygotsky’s contribution. European Journal of Psychology of Education, 9(4), 311-319. https://doi.org/10.1007/BF03172904 Bernicot, J., Laval, V., & Chaminaud, S. (2007). Nonliteral language forms in children: In what order are they acquired in pragmatics and metapragmatics? Journal of Pragmatics, 39, 2115-2132. https://doi.org/10.1016/j.pragma.2007.05.009 Biesta, G. (2012). Combining methodologies: Mixed methods. In J. Arthur, M. Waring, R. Coe, & L. V. Hedges (Eds.), Research methods and methodologies in education. (pp. 147-152). Bloome, D., Carter, S. P., Christian, B. M., Madrid, S., Otto, S., Shuart-Faris, N., & Smith, M. (2008). Discourse analysis in classrooms: Approaches to language and literacy research. Teachers College Press. Blum-Kulka, S. (1987). Indirectness and politeness in requests: Same or different? Journal of Pragmatics, 11(2), 131-146. https://doi.org/10.1016/0378-2166(87)90192-5 Blum-Kulka, S., House, J., & Kasper, G. (1989). Cross-Cultural Pragmatics: Requests and Apologies. Ablex Publishing Company. Blum-Kulka, S., & Olshtain, E. (1984). Requests and apologies: A cross-cultural study of speech act realization patterns (CCSARP). Applied Linguistics, 5(3), 196-213. https://doi.org/10.1093/applin/5.3.196 Booth, A., Sutton, A., & Papaioannou, D. (2016). Systematic Approaches to a Successful Literature Review (2nd ed.). Sage Publications Ltd. Bosco, F. M., Bucciarelli, M., & Bara, B. G. (2006). Recognition and repair of communicative failures: A developmental perspective. Journal of Pragmatics, 38, 1398-1429. https://doi.org/10.1016/j.pragma.2005.06.011 Bridges, C. L. (2008). Effects of readers’ theatre on English language learners: A strategy for oral language and reading improvement TNTESOL Journal, 1, 20-29. Brown, G., Irving, E., & Keegan, P. (2008). An Introduction to Educational Assessment, Measurement and Evaluation (2nd ed.). Pearson Education. Brown, P., & Levinson, S. C. (1987). Politeness: Some universals in language use. Cambridge University Press. Brown, Z., & Perkins, H. (Eds.). (2019). Using Innovative Methods in Early Years Research: Beyond the Conventional. Routledge. References 123 Brubæk, S. (2012). Pragmatic competence in English at the VG1 level: To what extent are Norwegian EFL students able to adapt to contextual demands when making requests in English? Acta Didactica Norge, 6(1), Art. 20. https://doi.org/10.5617/adno.1089 Bruce, P. C. (2015). Introductory Statistics and Analytics: A Resampling Perspective. John Wiley & Sons, Incorporated. Bryman, A., Becker, S., & Sempik, J. (2008). Quality criteria for quantitative, qualitative and mixed methods research: A view from social policy. International Journal of Social Research Methodology, 11(4), 261-276. https://doi.org/10.1080/13645570701401644 Buson, L., & Billiez, J. (2013). Representations of stylistic variation in 9- to 11-year-olds: Cognitive processes and salience. Linguistics, 51(2), 325-354. https://doi.org/10.1515/ling-2013-0013 Caillies, S., Hody, A., & Calmus, A. (2012). Theory of mind and irony comprehension in children with cerebral palsy. Research in Developmental Disabilities, 33, 1380-1388. https://doi.org/10.1016/j.ridd.2012.03.012 Casey, S. C., Rachel. (2006). Bringing reading alive through readers’ theater. Illinois Reading Council Journal, 34(4), 17-25. CEFR. (2018). Collated Representative Samples of Descriptors of Language Competences Developed for Young Learners (Vol. 2). Council of Europe. Cekaite, A. (2013). Child pragmatic development. In C. A. Chapelle (Ed.), The Encyclopedia of Applied Linguistics (pp. 1-7). Blackwell Publishing, Ltd. Cenoz, J., & Valencia, J. F. (1996). Cross-cultural communication and interlanguage pragmatics: American vs. European requests. In L. F. Bouton (Ed.), Pragmatics and language learning (Vol. 7) (pp. 47-53). University of Illinois, Division of English as an International Language. Chang, Y.-F. (2016). Apologizing in Mandarin Chinese: A study on developmental patterns. Concentric: Studies in Linguistics, 42(1), 73-101. https://doi.org/10.6241/concentric.ling.42.1.04 Chang, Y.-F. (2018). The effect of an interlocutor’s social status on the use of apology strategies: A cross-sectional study. Concentric: Studies in Linguistics, 44(1). https://doi.org/10.6241/concentric.ling.44.1.04 Chavarría, M. I., & Bonany, E. B. (2006). Raising awareness of pragmatics in the EFL classroom: A proposal. Cultura, Lenguaje y Representación/Culture, Language and Representation, 3(3), 133-144. Christensen, P., & James, A. (Eds.). (2017). Research with Children: Practices and Perspectives (3rd ed.). Routledge. Christmann, E. P. (2012). Beyond the Numbers: Making Sense of Statistics. NSTA Press. Cohen, A. D. (1996). Speech acts. In S. L. McKay & N. H. Hornberger (Eds.), Sociolinguistics and Language Teaching (pp. 383-420). Cambridge University Press. Cohen, L., Manion, L., & Morrison, K. (2000). Research Methods in Education (5th ed.). RoutledgeFalmer. Collins, A., Lockton, E., & Adams, C. (2014). Metapragmatic explicitation ability in children with typical language development: Development and validation of References 124 a novel clinical assessment. Journal of Communication Disorders, 52, 31-43. https://doi.org/10.1016/j.jcomdis.2014.07.001 Corsaro, W. A., & Molinary, L. (2017). Entering and observing in children’s worlds: A reflection on a longitudinal ethnography of early education in Italy. In P. Christensen & A. James (Eds.), Research with Children: Perspectives and Practices (3rd ed., pp. 11-30). Routledge. Cox, V. (2017). Translating Statistics to Make Decisions: A Guide for the Non-Statistician. Apress. Creswell, J. W., & Plano Clark, V. L. (2011). Desigining and Conducting Mixed Methods Research (2nd ed.). Sage. Crystal, D. (1997). The Cambridge Encyclopedia of Language (2nd ed.). Cambridge University Press. Culpeper, J., & Haugh, M. (2014). Pragmatics and the English Language. Palgrave Macmillan. Culpeper, J., Mackey, A., & Taguchi, N. (2018). Second Language Pragmatics. From Theory to Research. Routledge. Davies, A. (2004). The Native Speaker in Applied Linguistics. In A. Davies & C. Elder (Eds.), The Handbook of Applied Linguistics (pp. 431-450). Blackwell Publishing Ltd. De Costa, P., Lee, J., Rawal, H., & Li, W. (2020). Ethics in applied linguistics research. In J. McKinley & H. Rose (Eds.), The Routledge Handbook of Research Methods in Applied Linguistics (pp. 122-130). Routledge. Dorcheh, H. H., & Baharlooie, R. (2016). Development of pragmatic competence. Journal of Language Teaching and Research, 7(1), 152-157. https://doi.org/10.17507/jltr.0701.17 Dörnyei, Z. (2007). Research Methods in Applied Linguistics: Quantitative, Qualitative, and Mixed Methodologies. Oxford University Press. Drew, I. (2018). Reading in the second language classroom: Consideration of first language approaches in second language contexts. Acta Didactica Norge, 12(2), Art. 10. https://doi.org/10.5617/adno.5570 Drew, I., & Hasselgreen, A. (2008). Young language learner (YLL) research: An overview of some international and national approaches. Acta Didactica Norge, 2(1), Art. 5. https://doi.org/10.5617/adno.1024 Drew, I., & Pedersen, R. R. (2010). Readers theatre: A different approach to English for struggling readers. Acta Didactica Norge, 4(1), Art. 7. https://doi.org/10.5617/adno.1051 Drew, I., & Pedersen, R. R. (2012). Readers theatre: A group reading approach to texts in mainstream classes. In A. Hasselgreen, I. Drew, & B. Sørheim (Eds.), The Young Language Learner: Research-Based Insights into Teaching and Learning (pp. 71-84). Fagbokforlaget. Duff, P. A. (2020). Case study research: Making language learning complexities visible. In J. McKinley & H. Rose (Eds.), The Routledge Handbook of Research Methods in Applied Linguistics (pp. 144-153). Routledge. Ebrahim, H. B. (2010). Situated ethics: possibilities for young children as research participants in the South African context. Early Child Development and Care, 180(3), 289-298. https://doi.org/10.1080/03004430701822958 References 125 Eckhoff, A. (Ed.). (2019). Participatory Research with Young Children. Springer. Economidou-Kogetsidis, M. (2013). Strategies, modification and perspective in native speakers’ requests: A comparison of WDCT and naturally occurring requests. Journal of Pragmatics, 53, 21-38. https://doi.org/10.1016/j.pragma.2013.03.014 Education First. (2020). English Proficiency Index. https://www.ef.no/epi/ Eelen, G. (2001). A Critique of Politeness Theories. St Jerome Publishing. Egbert, J., & LaFlair, G. T. (2018). Statistics for categorical, nonparametric, and distribution-free data. In A. Phakiti, P. De Costa, L. Plonsky, & S. Starfield (Eds.), The Palgrave Handbook of Applied Linguistics Research Methodology (pp. 523-540). Palgrave Macmillan. Ellis, R. (1992). Learning to communicate in the classroom: A study of two language learners' requests. Studies in Second Language Acquisition, 14(1), 1-23. Ellis, R. R. (2012). Language Teaching Research & Language Pedagogy. John Wiley & Sons, Inc. Elo, S., & Kyngäs, H. (2008). The qualitative content analysis process. Journal of Advanced Nursing, 62(1), 107-115. https://doi.org/10.1111/j.1365-2648.2007.04569.x Eslami-Rasekh, Z., Eslami-Rasekh, A., & Fatahi, A. (2004). The effect of explicit metapragmatic instruction on the speech act awareness of advanced EFL students. TESL-EJ, 8(2), A-2. Eslami, Z., & McLeod, K. (2010). Pragmatics: Teaching speech acts. In D. H. Tatsuki & N. R. Houck (Eds.), Pragmatics: Teaching Speech Acts (pp. 19-28). Teachers of English to Speakers of Other Languages, Inc. (TESOL). Eun, B., & Lim, H.-S. (2009). A sociocultural view of language learning: The importance of meaning-based instruction. TESL Canada Journal, 27(1), 13-26. https://doi.org/10.18806/tesl.v27i1.1031 Fàbregues, S., & Molina-Azorín, J. F. (2017). Addressing quality in mixed methods research: a review and recommendations for a future agenda. Quality & Quantity, 51, 2847-2863. https://doi.org/10.1007/s11135-016-0449-4 Félix-Brasdefer, J. C. (2006). Pedagogical intervention and the development of pragmatic competence in learning Spanish as a foreign language. Issues in Applied Linguistics, 16(1), 59-84. https://doi.org/10.5070/L4161005093 Félix-Brasdefer, J. C. (2010). Data collection methods in speech act performance: DCTs, role plays, and verbal reports. In A. Martínez-Flor & E. Usó-Juan (Eds.), Speech Act Performance: Theoretical, empirical and methodological issues (pp. 41-56). John Benjamins. Fielding, M. (2001). Students as radical agents of change. Journal of educational change, 2, 123-141. https://doi.org/10.1023/A:1017949213447 Fortune, A., & Thorp, D. (2001). Knotted and entangled: New light on the identification, classification and value of language related episodes in collaborative output tasks. Language Awareness, 10(2), 143-160. https://doi.org/10.1080/09658410108667031 Fox, E., & Riconscente, M. (2008). Metacognition and self-regulation in James, Piaget, and Vygotsky. Educational Psychology Review, 20, 373–389. https://doi.org/10.1007/s10648-008-9079-2 References 126 Fretheim, T. (2005). Politeness in Norway; How can you be polite and sincere. In L. Hickey & M. Stewart (Eds.), Politeness in Europe (pp. 145-158). Multilingual Matters, Ltd. Gal'perin, P. Y. (1979). The role of orientation in thought. Soviet Psychology, 18(2), 19-45. https://doi.org/10.2753/RPO1061-0405180284 Gal'perin, P. Y. (1989). Organization of mental activity and the effectiveness of learning. Soviet Psychology, 27, 65-82. https://doi.org/10.2753/RPO1061-0405270365 Gal'perin, P. Y. (1992). Stage-by-stage formation as a method of psychological investigation. Journal of Russian and East European Psychology, 30, 60-80. https://doi.org/10.2753/RPO1061-0405300460 Glaser, K. (2018). Enhancing the role of pragmatics in primary English teacher training. Glottodidactica. An International Journal of Applied Linguistics, 45(2), 119-131. https://doi.org/10.14746/gl.2018.45.2.06 Gobo, G. (2008). Doing Ethnography. Sage Publishing. Gredler, M. E. (2012). Understanding Vygotsky for the classroom: is it too late? Educational Psychology Review, 24(1), 113-131. https://doi.org/10.1007/s10648-011-9183-6 Griffin, K. M. (2019). Participatory research interviewing practices with children. In A. Eckhoff (Ed.), Participatory Research with Young Children (pp. 55-71). Springer. Gu, P. Y., Hu, G., & Zhang, L. J. (2005). Investigating language learner strategies among lower primary school pupils in Singapore. Language and Education, 19(4), 281-303. https://doi.org/10.1080/09500780508668682 Hacking, I. (1999). The Social Construction of What? Harvard University Press. Halenko, N., & Jones, C. (2011). Teaching pragmatic awareness of spoken requests to Chinese EAP learners in the UK: Is explicit instruction effective? System, 39, 240-250. https://doi.org/10.1016/j.system.2011.05.003 Hancock, M. (1995). Pronunciation Games. Cambridge University Press. Hart, R. A. (1992). Children's Participation: From Tokenism to Citizenship. UNICEF International Child Development Centre. Hasselgreen, A. (2005). The new læreplan proposal for English: reading between the lines. Språk og språkundervisning, 2, 7-10. Hellawell, D. (2006). Inside–out: analysis of the insider–outsider concept as a heuristic device to develop reflexivity in students doing qualitative research, . Teaching in Higher Education, 11(4), 483-494. https://doi.org/10.1080/13562510600874292 Henery, A. (2015). On the development of metapragmatic awareness abroad: two case studies exploring the role of expert-mediation. Language Awareness, 24(4), 316-331. https://doi.org/10.1080/09658416.2015.1113982 Hill, B., Ide, S., Ikuta, S., Kawasaki, A., & Ogino, T. (1986). Universals of Linguistic Politeness: Quantitative Evidence from Japanese and American English. Journal of Pragmatics, 10, 347-371. https://doi.org/10.1016/0378-2166(86)90006-8 References 127 Holmes, S. (2019). Play-based interview techniques with young children. In Z. Brown & H. Perkins (Eds.), Using Innovative Methods in Early Years Research: Beyond the Conventional (pp. 92-108). Routledge. Holzman, L. (2018). Zones of proximal development: Mundane and magical. In J. P. Lantolf, M. E. Poehner, & M. Swain (Eds.), The Routledge Handbook of Sociocultural Theory and Second Language Development (pp. 42-55). Routledge. Hosseini, S. M. B., & Safari, M. (2018). Exploring the potential of explicit/implicit teaching through plays for EFL learners’ pragmatic development. The Journal of Asia TEFL, 15(4), 1065-1082. https://doi.org/10.18823/asiatefl.2018.15.4.12.1065 House, J. (2010). The pragmatics of English as a lingua franca. In A. Trosborg (Ed.), Pragmatics Across Languages and Cultures (pp. 363-387). De Gruyter Mouton. Hsieh, S. C.-Y., & Hsu, C.-C. N. (2010). Idiom comprehension in mandarin-speaking children. Journal of Psycholinguistic Research, 39, 505-522. https://doi.org/10.1007/s10936-009-9145-z IBM. (2019). IBM SPSS Statistics for Windows In (Version 26.0) IBM Corp. ICMJE, I. C. o. M. J. E. (n.d.). Defining the Role of Authors and Contributors. http://www.icmje.org/recommendations/browse/roles-and-responsibilities/defining-the-role-of-authors-and-contributors.html Infante, P. (2018). Mediated development: Promoting learner internalization of L2 concepts through cognitive-process focused activities. In J. P. Lantolf, M. E. Poehner, & M. Swain (Eds.), The Routledge Handbook of Sociocultural Theory and Second Language Development (pp. 229-246). Routledge. Infantidou, E., & Tzanne, A. (2012). Levels of pragmatic competence in an EFL academic context: A tool for assessment. Intercultural Pragmatics, 9(1), 47-70. https://doi.org/10.1515/ip-2012-0003 Ishihara, N. (2010). Instructional pragmatics: Bridging teaching, research, and teacher education. Language and Linguistics Compass, 4(10), 938-953. https://doi.org/10.1111/j.1749-818X.2010.00242.x Ishihara, N. (2013). Is it rude language? Children learning pragmatics through visual narratives. TESL Canada Journal, 30(7), 135-149. https://doi.org/10.18806/tesl.v30i7.1157 Ishihara, N., & Chiba, A. (2014). Teacher-based or interactional?: Exploring assessment for children’s pragmatic development. Iranian Journal of Language Testing, 4(1), 84-112. Ishihara, N., & Cohen, A. D. (2014). Teaching and Learning Pragmatics: Where Language and Culture Meet. Routledge. Jakupčević, E., & Portolan Ćavar, M. (2021). An analysis of pragmatic content in EFL textbooks for young learners in Croatia. Language Teaching Research, 1–24. https://doi.org/10.1177/1362168820986936 James, A. (2007). Giving voice to children's voices: Practices and problems, pitfalls and potentials. American Anthropologist, 109(2), 261-272. https://doi.org/10.1525/aa.2007.109.2.261 References 128 Johnson, R. B., & Onwuegbuzie, A. J. (2004). Mixed methods research: A research paradigm whose time has come. Educational Researcher, 33(7), 14-26. https://doi.org/10.3102/0013189X033007014 Johnston, J. (2008). Methods, Tools and Instruments for Use with Children. University of Oxford. Jones, S., R, Torres, V., & Arminio, J. (2006). Negotiating the Complexities of Qualitative Research in Higher Education: Fundamental Elements and Issues. Routledge. Kádár, D. Z., & Haugh, M. (2013). Understanding Politeness. Cambridge University Press. Karpov, Y. (2018). Acquisition of scientific concepts as the content of school instruction. In J. P. Lantolf, M. E. Poehner, & M. Swain (Eds.), The Routledge handbook of sociocultural theory and second language development (pp. 102-116). Routledge. Kasper, G. (1997). Can pragmatic competence be taught? . http://www.nflrc.hawaii.edu/NetWorks/NW06/ Kasper, G. (2001). Classroom research on interlanguage pragmatics. In K. R. Rose & G. Kasper (Eds.), Pragmatics in Language Teaching (pp. 33-62). Cambridge University Press. Kellett, M. (2010). Small shoes, big steps! Empowering children as active researchers. American Journal of Community Psychology, 46, 195-203. https://doi.org/10.1007/s10464-010-9324-y Kim, J. (2013). Developing conceptual understanding of sarcasm in a second language through concept-based instruction [Unpublished doctoral thesis, The Pennsylvania State University]. Pennsylvania. Kingdon, Z. (2019). Using the mosaic approach as an ethnographic methodology. In Z. Brown & H. Perkins (Eds.), Using Innovative Methods in Early Years Research: Beyond the Conventional (pp. 109-123). Routledge. Kinginger, C. (2002). Defining the zone of proximal development in US foreign language education. Applied Linguistics, 23(2), 240-261. https://doi.org/10.1093/applin/23.2.240 Kontra, E. H. (n.d.). Introducing Pragmatic Awareness at Low Levels. https://americanenglish.state.gov/files/ae/resource_files/awareness-kontra.pdf Kozulin, A. (2018). Mediation and internalization: Conceptual analysis and practical applications. In J. P. Lantolf, M. E. Poehner, & M. Swain (Eds.), The Routledge Handbook of Sociocultural Theory and Second Language Development (pp. 23-41). Routledge. Krippendorf, K. (1989). Content analysis. In E. Barnouw, G. Gerbner, W. Schramm, T. L. Worth, & L. Gross (Eds.), International Encyclopedia of Communication (Vol. 1, pp. 403-407). Oxford University Press. Krulatz, A. (2016). Competent non-native users of English? Requestive behavior of Norwegian EFL teachers. Nordic Journal of English Studies, 15(4), 24-44. https://doi.org/10.35360/njes.383 Krulatz, A., & Dahl, A. (2016). Baseline assessment of Norwegian EFL teacher preparedness. Journal of Linguistics and Language Teaching, 7(2), 199-218. References 129 Kuchah, K., & Milligan, L. O. (2021). Navigating cultural and methodological complexities in research with children in a Sub-Saharan African Context. In A. Pinter & K. Kuchah (Eds.), Ethical and Methodological Issues in Researching Young Language Learners in School Contexts (pp. 165-182). Multilingual Matters. Kuchah, K., & Pinter, A. (2012). ‘Was this an interview?’ Breaking the power barrier in adult-child interviews in an African context. Issues in Educational Research, 22(3), 283-297. Kuchah, K., & Pinter, A. (2021). Researching young language learners in school context: Setting the scene. In A. Pinter & K. Kuchah (Eds.), Ethical and Methodological Issues in Researching Young Language Learners in School Contexts (pp. 1-24). Multilingual Matters. Kuepper, M.-C., & Feryok, A. (2020). Concept-based pragmatics instruction: Teaching German address pronouns to New Zealand tertiary students. Language and Sociocultural Theory, 6(3), 158–183. https://doi.org/10.1558/lst.37359 Lacroix, A., Aguert, M., Dardier, V., Stojanovik, V., & Laval, V. (2010). Idiom comprehension in French-speaking children and adolescents with Williams' syndrome. Research in Developmental Disabilities, 31(2), 608-616. https://doi.org/10.1016/j.ridd.2009.12.011 Lansdown, G. (2005). Can You Hear Me? The Right of Young Children to Participate in Decisions Affecting Them. Bernard van Leer Foundation. Lantolf, J. P., & Poehner, M. E. (2014). Sociocultural Theory and the Pedagogical Imperative in L2 Education: Vygotskian Praxis and the Research/Practice Divide. Routledge. Lantolf, J. P., Poehner, M. E., & Swain, M. (2018). Introduction. In J. P. Lantolf, M. E. Poehner, & M. Swain (Eds.), The Routledge Handbook of Sociocultural Theory and Second Language Development (pp. 1-20). Routledge. Lantolf, J. P., & Zhang, X. (2017). Concept-based language instruction. In S. Loewen & M. Sato (Eds.), The Routledge Handbook of Instructed Second Language Acquisition (pp. 146-165). Routledge. Laval, V. (2003). Idiom comprehension and metapragmatic knowledge in French children. Journal of Pragmatics, 35(5), 723-739. https://doi.org/10.1016/S0378-2166(02)00137-6 Lee, C. (2010). An Exploratory Study of the Interlanguage Pragmatic Comprehension of Young Learners of English. Pragmatics, 20(3), 343-373. https://doi.org/10.1075/prag.20.3.03lee Leech, G. (1983). The Principles of Pragmatics. Longman. Levi, T., & Poehner, M. E. (2018). Employing dynamic assessment to enhance agency among L2 learners. In J. P. Lantolf, M. E. Poehner, & M. Swain (Eds.), The Routledge Handbook of Sociocultural Theory and Second Language Development (pp. 295-309). Routledge. Lew, S., Yang, A. H., & Harklau, L. (2018). Qualitative methodology. In A. Phakiti, P. De Costa, L. Plonsky, & S. Starfield (Eds.), The Palgrave Handbook of Applied Linguistics Research Methodology (pp. 79-101). Palgrave Macmillan. References 130 Liddicoat, A. J., & McConachy, T. (2019). Meta-pragmatic awareness and agency in language learners’ constructions of politeness. In T. Szende & G. Alao (Eds.), Pragmatic and Cross-Cultural Competences: Focus on Politeness (pp. 11-25). Peter Lang. Limberg, H. (2016). "Always remember to say please and thank you”: Teaching politeness with German EFL textbooks. In K. Bardovi–Harlig & J. C. Félix-Brasdefer (Eds.), Pragmatics and Language Learning (Vol. 14, pp. 265–291). National Foreign Language Resource Center, University of Hawai‘i. Lincoln, Y. S., & Guba, E. G. (1985). Naturalistic Enquiry. Sage Publications. Lockton, E., Adams, C., & Collins, A. (2016). Do children with social communication disorder have explicit knowledge of pragmatic rules they break? A comparison of conversational pragmatic ability and metapragmatic awareness. International Journal of Language & Communication Disorders, 51(5), 508-517. https://doi.org/10.1111/1460-6984.12227 Lorenz, E., Krulatz, A., & Torgersen, E. N. (2021). Embracing linguistic and cultural diversity in multilingual EAL classrooms: The impact of professional development on teacher beliefs. Teaching and Teacher Education, 105, 103428. https://doi.org/10.1016/j.tate.2021.103428 Lundy, L. (2007). ‘Voice’ is not enough: conceptualising Article 12 of the United Nations Convention on the Rights of the Child. British Educational Research Journal, 33(6), 927-942. https://doi.org/10.1080/01411920701657033 Lyndon, H. (2019). The use of drawing methods with young children in research. In H. Perkins & Z. Brown (Eds.), Using Innovative Methods in Early Years Research Beyond the Conventional (pp. 63-76). Routledge. Lyster, R. (1994). The effect of functional-analytic teaching on aspects of french immersion students' sociolinguistic competence. Applied Linguistics, 15(3), 263-287. https://doi.org/10.1093/applin/15.3.263 Mackey, A., & Bryfonski, L. (2018). Mixed methodology. In A. Phakiti, L. Plonsky, P. De Costa, & S. Starfield (Eds.), The Palgrave Handbook of Applied Linguistics Research Methodology (pp. 103-122). Palgrave Macmillan. Mackey, A., & Gass, S. M. (2005). Second Language Research: Methodology and Design (1st ed.). Routledge. Marková, I., Linell, P., Grossen, M., & Orvig, A. S. (2007). Dialogue in Focus Groups: Exploring Socially Shared Knowledge. Equinox. Marsden, E. (2020). Methodological transparency and its consequences for the quality and scope of research. In J. McKinley & H. Rose (Eds.), The Routledge Handbook of Research Methods in Applied Linguistics (pp. 15-28). Routledge. Martin, J. (2004). Self-regulated learning, social cognitive theory, and agency. Educational Psychologist, 39(2), 135-145. https://doi.org/10.1207/s15326985ep3902_4 Mayo, M. d. P. G. (2021). ‘Are you coming back? It was fun’: Turning ethical and methodological challenges into opportunities in task-based research with children. In A. Pinter & K. Kuchah (Eds.), Ethical and Methodological Issues in Researching Young Language Learners in School Contexts (pp. 68-84). Multilingual Matters. References 131 McConachy, T. (2013). Exploring the meta-pragmatic realm in English language teaching. Language Awareness, 22(2), 100-110. https://doi.org/10.1080/09658416.2011.644795 McConachy, T. (2018). Developing Intercultural Perspectives on Language Use: Exploring Pragmatics and Culture in Foreign Language Learning. Multilingual Matters. McConachy, T., & Liddicoat, A. J. (2016). Meta-pragmatic awareness and intercultural competence: The role of reflection and interpretation in intercultural mediation. In F. Dervin & Z. Gross (Eds.), Intercultural Competence in Education: Alternative Approaches for Different Times (pp. 13-30). Palgrave Macmillan. McConachy, T., & Liddicoat, A. J. (Eds.). (Forthcoming). Teaching and Learning Second Language Pragmatics for Intercultural Understanding. Routledge. McConachy, T., & Spencer-Oatey, H. (2020). Developing pragmatic awareness. In K. P. Schneider & E. Ifantidou (Eds.), Developmental and Clinical Pragmatics (pp. 393-427). Mouton de Gruyter. McKay, P. (2006). Assessing Young Language Learners. Cambridge University Press. McTavish, M., Streelasky, J., & Coles, L. (2012). Listening to children's voices: Children as participants in research. International Journal of Early Childhood, 44(3), 249-267. https://doi.org/10.1007/s13158-012-0068-8 Mercer, S. (2011). Understanding learner agency as a complex dynamic system. System, 39, 427-436. https://doi.org/10.1016/j.system.2011.08.001 Miles, R. (2015). Complexity, representation and practice: Case study as method and methodology. Issues in Educational Research, 25(3), 309-318. Morollón Martí, N. (Forthcoming). Concept-based instruction for teaching and learning L2 (im)politeness In T. McConachy & A. J. Liddicoat (Eds.), Teaching and Learning Second Language Pragmatics for Intercultural Understanding. Routledge. Morrow, V., & Richards, M. (1996). The ethics of social research with children: An overview. Children & Society, 10, 90–105. https://doi.org/10.1111/j.1099-0860.1996.tb00461.x Mukherji, P., & Albon, D. (2015). Research Methods in Early Childhood: An Introductory Guide (2 ed.). Sage. (2010) Mukhija, V. (2010). N of one plus some: An alternative strategy for conducting single case research. Journal of Planning Education and Research, 29(4), 416–426. https://doi.org/10.1177/0739456X10362770 Musgrave, J. (2019). Reflexivity in educational research. In Z. Brown & H. Perkins (Eds.), Using Innovative Methods in Early Years Research: Beyond the Conventional (pp. 7-18). Routledge. Myrset, A. (2014). A case study of the benefits and challenges of readers theatre in a 6th grade Norwegian EFL class [Unpublished master's thesis, University of Stavanger]. Stavanger. Myrset, A. (2021). Scientific concepts as meaning-making resources for young EFL learners in the learning of pragmatics. Intercultural Communication Education, 4(2), 191–212. https://doi.org/10.29140/ice.v4n2.485 References 132 Myrset, A. (In preparation). Giving young language learners a voice: learner feedback on pragmatics instruction. Myrset, A. (Pending revisions). 'You could win Masterchef with this soup. Can I get some more?' Request production and the impact of instruction on young EFL learners. Journal of Pragmatics. Myrset, A., & Drew, I. (2016). A case study of readers theatre in a primary Norwegian EFL class. Nordic Journal of Modern Language Methodology, 4(1), 49-66. https://doi.org/10.46364/njmlm.v4i1.292 Myrset, A., & Savić, M. (2021). “If an astronaut were on the moon…”: Eliciting metapragmatic data from young L2 learners. Applied Pragmatics, 3(2), 163-196. Negueruela, E. (2003). A sociocultural approach to teaching and researching second language: systemic theoretical instruction and second language development [Unpublished doctoral thesis, Pennsylvania State University]. Ness, M. (2019). When students generate questions: Participatory-based reading instruction in elementary classrooms. In A. Eckhoff (Ed.), Participatory Research with Young Children (pp. 73-87). Springer. Nicholas, A. (2015). A concept-based approach to teaching speech acts in the EFL classroom. ELT journal, 69(4), 383-394. https://doi.org/10.1093/elt/ccv034 Nikula, T. (2002). Teacher talk reflecting pragmatic awareness: A look at EFL and content-based classroom settings. Pragmatics, 12(4), 447-467. https://doi.org/10.1075/prag.12.4.03nik Norwegian Ministry of Education and Research. (2003). Country Report: Norway. Strasbourg: Council of Europe Retrieved from https://rm.coe.int/language-education-policy-profile-norway-country-report/16807b3b42 Nunan, D. (1992). Research Methods in Language Learning. Cambridge University Press. O'Kane, C. (2008). The development of participatory techniques. Facilitating children’s views about decisions which affect them. In P. Christensen & A. James (Eds.), Research with children. Perspectives and practices (2nd ed., pp. 125-155). Routledge. O'Kane, C. (2017). Participatory research on kinship care in East Africa. In P. Christensen & A. James (Eds.), Research with Children: Perspectives and Practices (3rd ed., pp. 180-202). Routledge. Ogiermann, E. (2009). Politeness and in-directness across cultures: A comparison of English, German, Polish and Russian requests. Journal of Politeness Research, 5, 189-216. https://doi.org/10.1515/JPLR.2009.011 Ogiermann, E., & Bella, S. (2020). An interlanguage study of request perspective: Evidence from German, Greek, Polish and Russian learners of English. Contrastive Pragmatics, 1(2), 180–209. https://doi.org/10.1163/26660393-BJA10003 Ohta, A. S. (2005). Interlanguage pragmatics in the zone of proximal development. System, 33, 503-517. https://doi.org/10.1016/j.system.2005.06.001 Pálmadóttir, H., & Einarsdóttir, J. (2016). Video observations of children's perspectives on their lived experiences: Challenges in the relations between References 133 the researcher and children. European Early Childhood Education Research Journal, 24(5), 721-733. https://doi.org/10.1080/1350293X.2015.1062662 Pérez-Hernández, L. (2021). Speech Acts in English: From Research to Instruction and Textbook Development. Cambridge University Press. Phakiti, A., De Costa, P., Plonsky, L., & Starfield, S. (2018). Applied linguistics research: Current issues, methods, and trends. In A. Phakiti, P. De Costa, L. Plonsky, & S. Starfield (Eds.), The Palgrave Handbook of Applied Linguistics Research Methodology (pp. 5-30). Palgrave Macmillan. Pimple, K. D. (2002). Six domains of research ethics: A heuristic framework for the responsible conduct of research. Science and Engineering Ethics, 8(2), 191-205. https://doi.org/10.1007/s11948-002-0018-1 Pinter, A. (2014). Child participant roles in applied linguistics research. Applied Linguistics, 35(2), 168-183. https://doi.org/10.1093/applin/amt008 Pinter, A. (2019). Research issues with young learners. In S. Garton & F. Copland (Eds.), The Routledge Handbook of Teaching English to Young Learners (pp. 411-424). Routledge. Pinter, A., & Kuchah, K. (Eds.). (2021). Ethical and Methodological Issues in Researching Young Language Learners in School Contexts. Multilingual Matters. Pinter, A., & Zandian, S. (2014). ‘I don’t ever want to leave this room’: benefits of researching ‘with’ children. ELT journal, 68(1), 64-74. https://doi.org/10.1093/elt/cct057 Pinter, A., & Zandian, S. (2015). ‘I thought it would be tiny little one phrase that we said, in a huge big pile of papers’: children’s reflections on their involvement in participatory research. Qualitative Research, 15(2), 235-250. https://doi.org/10.1177/1468794112465637 Pinto, D., & Raschio, R. (2007). A comparative study of requests in heritage speaker Spanish, L1 Spanish, and L1 English. International Journal of Bilingualism, 11(2), 135-155. https://doi.org/10.1177/13670069070110020101 Plonsky, L., & Zhuang, J. (2019). A meta-analysis of L2 pragmatics instruction. In N. Taguchi (Ed.), The Routledge Handbook of Second Language Acquisition and Pragmatics (pp. 287-307). Routledge. Politou, E., Alepis, E., & Patsakis, C. (2018). Forgetting personal data and revoking consent under the GDPR: Challenges and proposed solutions. Journal of Cybersecurity, 0(0), 1-20. https://doi.org/10.1093/cybsec/tyy001 Polkinghorne, D. E. (2005). Language and meaning: Data collection in qualitative research. Journal of Counseling Psychology, 52(2), 137-145. https://doi.org/10.1037/0022-0167.52.2.137 Portolés, L. (2015). Multilingualism and Very Young Learners. De Gruyter Mouton. Portolés, L., & Martí, O. (2017). Translanguaging as a teaching resource in early language learning of English as a an additional language (EAL). Bellaterra Journal of Teaching & Learning Language & Literature, 10(1), 61-77. https://doi.org/10.5565/rev/jtl3.698 Portolés, L., & Safont, P. (2018). Examining authentic and elicited data from a multilingual perspective. The real picture of child requestive behaviour in the References 134 L3 classroom. System, 75, 81-92. https://doi.org/10.1016/j.system.2018.03.012 Punch, S. (2002a). Interviewing strategies with young people. Children and Society, 16, 45-56. https://doi.org/10.1002/chi.685 Punch, S. (2002b). Research with children: The same or different from research with adults? Childhood, 9(3), 321-341. https://doi.org/10.1177/0907568202009003005 QSR. (2016). NVivo. In (Version 12) QSR International Pty Ltd. https://www.qsrinternational.com/nvivo-qualitative-data-analysis-software/home Randall, J. G., Villado, A. J., & Zimmer, C. U. (2016). Is retest bias biased? Journal of Personnel Psychology, 15(2), 45-54. https://doi.org/10.1027/1866-5888/a000149 Rindal, U. (2014). What is English? Acta Didactica Norge, 8(2), Art. 14. https://doi.org/10.5617/adno.1137 Rinnert, C., & Iwai, C. (2010). I want you to help me: Learning to soften English requests. In D. H. Tatsuki & N. R. Houck (Eds.), Pragmatics: Teaching Speech Acts (pp. 29-46). Teachers of English to Speakers of Other Languages, Inc. (TESOL). Roever, C., & Phakiti, A. (2018). Quantitative Methods for Second Language Research: A Problem-Solving Approach. Routledge. Rogers, J., & Révész, A. (2020). Experimental and quasi-experimental designs. In J. McKinley & H. Rose (Eds.), The Routledge Handbook of Research Methods in Applied Linguistics (pp. 133-143). Rolland, L., Dewaele, J.-M., & Costa, B. (2020). Planning and conducting ethical interviews: Power language and emotions. In J. McKinley & H. Rose (Eds.), The Routledge Handbook of Research Methods in Applied Linguistics (pp. 279-289). Routledge. Rollins, J. (2018). Being participatory through play. In I. Coyne & B. Carter (Eds.), Being Participatory: Researching with Children and Young People Co-constructing Knowledge Using Creative Techniques (pp. 79-102). Springer. Rose, K. R. (2000). An exploratory cross-sectional study of interlanguage pragmatic development. Studies in Second Language Acquisition, 22, 27-67. https://doi.org/10.1017/S0272263100001029 Roy, K., Hart, J. R., & Golojuch, L. (2019). Authenticity in qualitative data analysis: Notes on racial and gender diversity in team ethnography of young men of color. In Á. E. Humble & M. E. Radina (Eds.), How Qualitative Data Analysis Happens: Moving Beyond "Themes Emerged" (pp. 97-111). Routledge. Sa'd, S. H. T., & Gholami, J. (2017). Teaching Iranian elementary EFL learners to say ‘no’ politely: An interlanguage pragmatic study. The Electronic Journal for English as a Second Language, 21(1), 1-23. Saldaña, J. (2016). The Coding Manual for Qualitative Researchers (3rd ed.). Sage. Savić, M. (2015). “Can I very please borrow it?”: Request development in young Norwegian EFL learners. Intercultural Pragmatics, 12(4), 443-480. https://doi.org/10.1515/ip-2015-0023 References 135 Savić, M. (2021). Co-constructing metapragmatic understandings: How young EFL learners talk about making requests. The European Journal of Applied Linguistics and TEFL, 10(2). Savić, M., Economidou-Kogetsidis, M., & Myrset, A. (2021). Young Greek Cypriot and Norwegian EFL learners: Pragmalinguistic development in request production. Journal of Pragmatics, 180, 15-34. https://doi.org/10.1016/j.pragma.2021.04.006 Savić, M., & Myrset, A. (Forthcoming-a). “But in England they’re certainly very polite, so you mustn’t forget that” Young EFL learners making sense of pragmatic practices. In T. McConachy & A. J. Liddicoat (Eds.), Teaching and Learning Second Language Pragmatics for Intercultural Understanding. Routledge. Savić, M., & Myrset, A. (Forthcoming-b). “Hey, you, can I loan your yellow pencil?” Young Norwegian EFL learners’ metapragmatic appraisal of requests. In N. Halenko & J. Wang (Eds.), Pragmatics in English language learning. Cambridge University Press. Schauer, G. A. (2019). Teaching and Learning English in the Primary School: Interlanguage Pragmatics in the EFL Context (Vol. 18). Springer. Schiller, W., & Einarsdóttir, J. (2009). Special Issue: Listening to young children’s voices in research – changing perspectives/changing relationships. Early Child Development and Care, 179(2), 125-130. https://doi.org/10.1080/03004430802666932 Searle, J. R. (1965). What is a speech act? In M. Black (Ed.), Philosophy in America (pp. 221-239). Allen & Unwin. Searle, J. R. (1979). Expression and Meaning: Studies in the Theory of Speech Acts. Cambridge University Press. Selvi, A. F. (2020). Qualitative content analysis. In J. McKinley & H. Rose (Eds.), The Routledge Handbook of Research Methods in Applied Linguistics (pp. 440-452). Routledge. Shamrova, D. P., & Cummings, C. E. (2017). Participatory action research (PAR) with children and youth: An integrative review of methodology and PAR outcomes for participants, organizations, and communities. Children and Youth Services Review, 81, 400-412. https://doi.org/10.1016/j.childyouth.2017.08.022 Shepard, A. (2004). Readers on stage: Resources for Reader's Theater (or Readers Rheatre), with Tips, Play Scripts, and Worksheets. Shepard Publications. Simons, H. (2009). Case Study Research in Practice. Sage Publications Ltd. Spencer-Oatey, H. (2008). Face, (im)politeness and rapport. In H. Spencer-Oatey (Ed.), Culturally speaking: Culture, communication and politeness theory (pp. 11-47). Continuum. Spencer-Oatey, H., & Kádár, D. Z. (2021). Intercultural Politeness: Managing Relations across Cultures. Cambridge University Press. Starman, A. B. (2013). The case study as a type of qualitative research. Journal of Contemporary Educational Studies/Sodobna Pedagogika, 64(1), 28-43. References 136 Stavans, A., & Shafran, R. W. (2018). The pragmatics of requests and refusals in multilingual settings. International Journal of Multilingualism, 15(2), 149-168. https://doi.org/10.1080/14790718.2017.1338708 Sterling, S., & De Costa, P. (2018). Ethical applied linguistics research. In A. Phakiti, P. De Costa, L. Plonsky, & S. Starfield (Eds.), The Palgrave Handbook of Applied Linguistics Research Methodology (pp. 163-182). Palgrave Macmillan. Stoecker, R. (1991). Evaluating and rethinking the case study. The Sociological Review, 39(1), 88-112. https://doi.org/10.1111/j.1467-954X.1991.tb02970.x Svanes, B. (1989). En undersøkelse av realisasjonsmønsteret for språkhandlingen “å be noen om å gjøre noe”. Maal og minne, 1(2), 89-107. Swain, M. (1997). Collaborative dialogue: Its contribution to second language learning. Revista Canaria de Estudios Ingleses, 34, 115-132. Swain, M., Kinnear, P., & Steinman, L. (2015). Sociocultural Theory in Second Language Education : An Introduction through Narratives. Multilingual Matters. Swanborn, P. (2010). Case Study Research: what, why and how? Sweet, M., & Moynihan, R. (2007). Improving Population Health: The Uses of Systematic Review. https://www.milbank.org/wp-content/uploads/2016/06/0712ImprovingPopulationHealthFinal.pdf Taguchi, N. (2011). Teaching pragmatics: Trends and issues. Annual Review of Applied Linguistics, 31, 289-310. https://doi.org/10.1017/S0267190511000018 Taguchi, N. (2015). Instructed pragmatics at a glance: Where instructional studies were, are, and should be going. Language Teaching, 48(1), 1-50. https://doi.org/10.1017/S0261444814000263 Taguchi, N., & Kim, Y. (2016). Collaborative dialogue in learning pragmatics: Pragmatic-related episodes as an opportunity for learning request-making. Applied Linguistics, 37(3), 416-437. https://doi.org/10.1093/applin/amu039 Tajeddin, Z., Keshavarz, M. H., & Zand-Moghadam, A. (2012). The effect of task-based language teaching on EFL learners’ pragmatic production, metapragmatic awareness, and pragmatic self-assessment. Iranian Journal of Applied Linguistics, 15(2), 139-166. Thomas, J. (1983). Cross-cultural pragmatic failure. Applied Linguistics, 4(2), 91-112. https://doi.org/10.1093/applin/4.2.91 Thompson, G., & Dooley, K. (2020). Ensuring translation fidelity in multilingual research. In J. McKinley & H. Rose (Eds.), The Routledge Handbook of Research Methods in Applied Linguistics (pp. 63-75). Routledge. Tracy, S. J. (2010). Qualitative quality: Eight "big-tent" criteria for excellent qualitative research. Qualitative Inquiry 16(10), 837-851. https://doi.org/10.1177/1077800410383121 Truscott, J., Graham, A., & Powell, M. A. (2019). Ethical considerations in participatory research with young children. In A. Eckhoff (Ed.), Participatory Research with Young Children (pp. 21-38). Springer. Turner, J. L. (2014). Using Statistics in Small-Scale Language Education Research. Routledge. References 137 Udir. (2006a). English subject curriculum (ENG1-03). https://www.udir.no/kl06/ENG1-03?lplang=eng Udir. (2006b). Generell del av læreplanen [in Norwegian]. https://www.udir.no/laring-og-trivsel/lareplanverket/utgatt/generell-del-av-lareplanen-utgatt/ Udir. (2020a). Core curriculum – values and principles for primary and secondary education. https://www.udir.no/lk20/overordnet-del/?lang=eng Udir. (2020b). Curriculum in English (ENG01‑04). https://www.udir.no/lk20/eng01-04?lang=eng UNCRC. (1989). The United Nations Convention on the Rights of the Child. https://www.unicef.org/child-rights-convention/convention-text Urbach, J., & Banerjee, R. (2019). Participatory research with young children from special oopulations: Issues and recommendations. In A. Eckhoff (Ed.), Participatory Research with Young Children (pp. 127-144). Springer. Urbanik, P., & Svennevig, J. (2019). Managing contingencies in requests: The role of negation in Norwegian interrogative directives. Journal of Pragmatics, 139, 109-125. https://doi.org/10.1016/J.PRAGMA.2018.10.014 van Compernolle, R. A. (2011). Developing second language sociopragmatic knowledge through concept-based instruction: A microgenetic case study. Journal of Pragmatics, 43, 3267–3283. https://doi.org/10.1016/j.pragma.2011.06.009 van Compernolle, R. A. (2012). Developing sociopragmatic capacity in a second language through concept-based instruction [Unpublished doctoral thesis, Pennsylvania State University]. van Compernolle, R. A. (2013a). Concept appropriation and the emergence of L2 sociostylistic variation. Language Teaching Research, 17(3), 343-362. https://doi.org/10.1177/1362168813482937 van Compernolle, R. A. (2013b). From verbal protocols to cooperative dialogue in the assessment of second language pragmatic competence. Intercultural Pragmatics, , 10(1), 71-100. https://doi.org/10.1515/ip-2013-0003 van Compernolle, R. A. (2014). Sociocultural Theory and L2 Instructional Pragmatics. Multilingual Matters. van Compernolle, R. A. (2015). The emergence of sociolinguistic competence in L2 classroom interaction. In N. Markee (Ed.), The Handbook of Classroom Discourse and Interaction (pp. 265-280). John Wiley & Sons, Inc. van Compernolle, R. A. (2016). Evaluating L2 pragmatic appropriateness and authenticity in synchronous computer-mediated strategic interaction scenarios. In R. A. van Compernolle & J. McGregor (Eds.), Authenticity, Language and Interaction in Second Language Contexts (pp. 10-34). Multilingual Matters. van Compernolle, R. A. (2018). Concept-based pragmatics instruction: Principles and applications. In J. P. Lantolf, M. E. Poehner, & M. Swain (Eds.), The Routledge Handbook of Sociocultural Theory and Second Language Development (pp. 211-228). Routledge. van Compernolle, R. A., Gomez-Laich, M. P., & Weber, A. (2016). Teaching L2 Spanish sociopragmatics through concepts: A classroom-based study. The References 138 Modern Language Journal, 100(1), 341-361. https://doi.org/10.1111/modl.12318 van Compernolle, R. A., & Henery, A. (2014). Instructed concept appropriation and L2 pragmatic development in the classroom. Language Learning, 64(3), 549-578. https://doi.org/10.1111/lang.12054 van Compernolle, R. A., & Henery, A. (2015). Learning to do concept-based pragmatics instruction: Teacher development and L2 pedagogical content knowledge. Language Teaching Research, 19(1), 351-372. https://doi.org/10.1177/1362168814541719 van Compernolle, R. A., & Henery, A. (2016). Evaluating L2 pragmatic appropriateness and authenticity in synchronous computer-mediated strategic interaction scenarios. In R. A. van Compernolle & J. McGregor (Eds.), Authenticity, Language and Interaction in Second Language Contexts. Multilingual Matters, 2016. van Compernolle, R. A., & Kinginger, C. (2013). Promoting metapragmatic development through assessment in the zone of proximal development. Language Teaching Research, 17(3), 282-302. https://doi.org/10.1177/1362168813482917 van Compernolle, R. A., & Williams, L. (2012). Promoting sociolinguistic competence in the classroom zone of proximal development. Language Teaching Research, 16(1), 39–60. https://doi.org/10.1177/1362168811423340 Vattøy, K.-D. (2017). Learning English in Norway. Language Issues: The ESOL Journal, 28(2), 51-53. https://doi.org/10.13140/RG.2.2.10696.78081 Verschueren, J. (2000). Notes on the role of metapragmatic awareness in language use. Pragmatics, 10(4), 439-456. https://doi.org/10.1075/prag.10.4.02ver Verschuren, P. J. M. (2003). Case study as a research strategy: Some ambiguities and opportunities. International Journal of Social Research Methodology, 6(2), 121-139. https://doi.org/10.1080/13645570110106154 Vygotsky, L. S. (1978). Mind in Society: The Development of Higher Psychological Processes (M. Cole, V. John-Steiner, S. Scribner, & E. Souberman, Eds.). Harvard University Press. Vygotsky, L. S. (1998). The Collected Works of L. S. Vygotsky. Vol. 5: Child psychology (R. W. Rieber, Ed.). Plenum. Vygotsky, L. S. (2012). Thought and language. The MIT Press. (Original work published 1934) Watt, D. (2007). On becoming a qualitative researcher: The value of reflexivity. The Qualitative Report, 12(1), 82-101. https://doi.org/10.46743/2160-3715/2007.1645 Watts, R. J. (2003). Politeness. Cambridge University Press. Woodfield, H. (2008). Problematising discourse completion tasks: Voices from verbal report. Evaluation & Research in Education, 21(1), 43-69. https://doi.org/10.2167/eri413.0 Woodfield, H., & Economidou-Kogetsidis, M. (2010). “I just need more time”: a study of native and non-native students' requests to faculty for late submission. Multilingua J. Cross-cult. Interlang. Commun., 29(1), 77-118. References 139 Yamada-Rice, D. (2017). Using visual and digital research methods with young children. In P. Christensen & A. James (Eds.), Research with Children: Perspectives and Practices (pp. 71-86). Routledge. Yates, L. (n.d). Softening short requests. https://americanenglish.state.gov/files/ae/resource_files/short-yates.pdf Yates, L., & Springall, J. (2010). Soften up! Successful requests in the workplace. In D. H. Tatsuki & N. R. Houck (Eds.), Pragmatics: Teaching Speech Acts (pp. 67-86). Teachers of English to Speakers of Other Languages, Inc. (TESOL). Young, C., & Rasinski, T. (2009). Implementing readers theatre as an approach to classroom fluency instruction. The Reading Teacher, 63(1), 4-13. https://doi.org/10.1598/RT.63.1.1 Young, C., & Rasinski, T. (2018). Readers Theatre: effects on word recognition automaticity and reading prosody. Journal of Research in Reading, 41(3), 475-485. https://doi.org/10.1111/1467-9817.12120 Zandian, S. (2021). Constructing joint understandings of research with children. In A. Pinter & K. Kuchah (Eds.), Ethical and Methodological Issues in Researching Young Language Learners in School Contexts (pp. 48-67). Multilingual Matters. Zhang, L., & Yan, R. (2012). Impact of immersion teaching on English sociopragmatic awareness of Chinese kindergarten children: A polite study. International Education, 41(2), 33-45. Zuckerman, G. (2004). Development of reflection through learning activity. European Journal of Psychology of Education, 19(1), 9-18. https://doi.org/10.1007/BF03173234 Zufferey, S. (2014). Acquiring Pragmatics: Social and Cognitive Perspectives. Routledge. References 140 Appendices 141 Appendices Appendix 1 – Worksheet directness Appendices 142 Appendices143Appendix 2 – Homework (H1)Appendices144Appendix 3 – Worksheet, supportive movesAppendices 145 Appendix 4 – Requests produced by learners Example, Learner 1 Appendices 146 Example, Learner 2 Appendices 147 Example, Learner 3 Appendices 148 Appendix 5 – Request perception journey Requests from the slides Appendices 149 Worksheet for the learners. Appendices150Appendix 6 – Match request and interlocutorExample of taskAppendices 151 Appendix 7 – Dice game Appendices152Appendix 8 – Homework (H2)Appendices 153 Appendix 9 – Labels Dice game Appendices 154 Appendix 10 – Script, worksheet Fill in the blanks Appendices 155 Optional: Continue the story Appendices 156 Appendix 11 – Example of a pre-written script (RT Cycle 1) Appendices 157 Appendices 158 Appendices 159 Appendix 12 – An example of a folder for script-writing (RT Cycle 2) Appendices 160 Appendices 161 Appendices 162 Appendices 163 Appendix 13 – Questions for the VODCT Intro Fast food: This video is about two children who go to McDonald’s and order food. Questions: - The boy is hungry. This is what he orders. What do you think he says? - The girl is hungry. This is what she orders. What do you think she says? - She also orders this. What do you think she says? Intro classroom: This video is about four learners who are drawing in class. Questions: - The girl doesn’t have a crayon. What does she ask her friend? - The girl doesn’t have green paper. What does she ask the teacher? - The girl doesn’t have orange paper. What does she ask the teacher? - The girl doesn’t have a yellow pencil. What does she ask her friend? Intro At the shop: This video is about a girl and a boy who go to the market to look at a doll and a toy car. Afterwards, a girl goes to the store with her parents. She looks at a kite and a hat. Questions: - The girl really likes the doll. What does she say to the sales assistant? - The boy really likes the car. What does she say to the sales assistant? - The girl really likes the kite. What does she say to her parents? - The girl really likes the hat. What does she say to her parents? Intro At the table: This video is about two children who are visiting their friend and her mother for dinner. Appendices 164 Questions: - The girl cannot use chopsticks and asks the mother for a fork. What do you think she says? - The boy is still hungry and asks the mother for more. What do you think he says? - The two friends enjoyed their stay and ask to come back. What do you think they say? Intro Snowman This video is about three children sitting inside at a library when it starts snowing outside. Questions: - The boy asks the girls to go outside and play. What do you think he says? - The boy looks out the window. It has started to snow. He asks them again. What do you think he says? - The boy would like to make a snowman with the girls. What do you think he says? Intro may I talk to Kate This video is about a boy who calls his friend to arrange a play date. Questions: - The boy asks the mother to speak to Kate. What do you think he says? - The boy asks Kate to meet in the park to play badminton. What do you think he says? Intro restaurant This video is about a boy and his mum at a restaurant. Questions: - The boy is hungry and orders this. What do you think he says? Appendices 165 - The boy burnt his tongue because the food was hot. He asks his mum for water. What do you think he says? Intro museum This video is about a group of friends who go to the museum. Questions: - The boy suggests going to the museum. What do you think he says? - They ask the lady for directions to the museum. What do you think they say? Appendices 166 Appendix 14 – Interview guide (translated from Norwegian) Questions based on scripts: Why did you choose this particular request? - What directness level is it? - Did you all agree at once? - Do you remember any alternatives you discussed? Would you change the request if we changed the person they are asking? - What if s/he’s talking to a ___________ (add characteristics?) Appraisal task. - Do you think this request was a nice, a so-so, or a not so nice way to ask? - If green (������������������������): Why do you think it was green? What made it green? - If blue (������������������������): Why do you think it was blue? What would we have to do to get it up to green? - If red (☹): Why do you think it was red? What would we have to do to get it up to blue or green? Potential prompts as follow-up: What’s important to think about when making requests? - Prompt: Situation? - Prompt: Who we talk to? What does it mean to be polite? Questions about project: Appraisal task: Performance: What did you think about the performance? - Do you think it was fun, so-so, or not so fun? - If green (������������������������): Why do you think it was green? What made it green? - If blue (������������������������): Why do you think it was blue? What would we have to do to get it up to green? - If red (☹): Why do you think it was red? What would we have to do to get it up to blue or green? Follow-up: Was there anything you particularly liked / didn’t like about the performance? Appendices 167 Class activities: What did you think about our classes in general? - Do you think it was fun, so-so, or not so fun? - If green (������������������������): Why do you think it was green? What made it green? - If blue (������������������������): Why do you think it was blue? What would we have to do to get it up to green? - If red (☹): Why do you think it was red? What would we have to do to get it up to blue or green? Follow-up: Was there anything you particularly liked / didn’t like about the classes? Writing scripts: What did you think about writing scripts? - Do you think it was fun, so-so, or not so fun? - If green (������������������������): Why do you think it was green? What made it green? - If blue (������������������������): Why do you think it was blue? What would we have to do to get it up to green? - If red (☹): Why do you think it was red? What would we have to do to get it up to blue or green? Follow-up: Was there anything you particularly liked / didn’t like about writing scripts? Videos: What did you think about the videos? - Do you think the activity was fun, so-so, or not so fun? - If green (������������������������): Why do you think it was green? What made it green? - If blue (������������������������): Why do you think it was blue? What would we have to do to get it up to green? - If red (☹): Why do you think it was red? What would we have to do to get it up to blue or green? Follow-up: Was there anything you particularly liked / didn’t like about the videos? Open slots: Is there anything else you want to appraise? - Do you think it was fun, so-so, or not so fun? - If green (������������������������): Why do you think it was green? What made it green? - If blue (������������������������): Why do you think it was blue? What would we have to do to get it up to green? - If red (☹): Why do you think it was red? What would we have to do to get it up to blue or green? Appendices 168 Potential prompts as follow-up: What have you learnt during this project? - What was particularly difficult / interesting / easy / fun? - Did you feel that you were able to use what you have learned in the videos and the scripts? Do you think what you have learned will be useful for you in the future? If so, when? Appendices169Appendix 15 – Consent form and NSD approvalConsent form.Appendices 170 Appendices 171 Appendices 172 NSD Approval Appendices 173 Articles 174 Articles Article I: Myrset, A., & Savić, M. (2021). “If an astronaut were on the moon…”: Eliciting metapragmatic data from young L2 learners. Applied Pragmatics, 3(2), 163-196. Article II: Myrset, A. (Pending revisions). 'You could win Masterchef with this soup. Can I get some more?' Request production and the impact of instruction on young EFL learners. Journal of Pragmatics. Article III: Myrset, A. (2021). Scientific concepts as meaning-making resources for young EFL learners in the learning of pragmatics. Intercultural Communication Education, 4(2), 191–212. https://doi.org/10.29140/ice.v4n2.485 Article IV: Myrset, A. (In preparation). Giving young language learners a voice: learner feedback on pragmatics instruction. [To be submitted to TESL-EJ] Articles 175 Article I - “If an astronaut were on the moon…”: Eliciting metapragmatic data from young L2 learners Due to copyright restrictions, this article is not available here. See: https://benjamins.com/catalog/ap.19027.myr. Access may be restricted.Articles 211 Article II - 'You could win Masterchef with this soup. Can I get some more?' Request production and the impact of instruction on young EFL learners Due to copyright restrictions, this article is not available here.Articles 240 Articles 241 Article III - Scientific concepts as meaning-making resources for young EFL learners in the learning of pragmatics Articles 242 Articles243Articles244Articles 245 Articles246Articles 247 Articles248Articles 249 Articles250Articles 251 Articles252Articles253Articles254Articles255Articles256Articles257Articles258Articles259Articles260Articles261Articles262Articles263Articles 264 Articles 265 Article IV - Giving young language learners a voice: learner feedback on pragmatics instruction Due to copyright restrictions, this article is not available here.Articles291View publication stats | RESEARCH GATE |
Research ArticleQualitative Health Research2021, Vol. 0(0) 1–13© The Author(s) 2021Article reuse guidelines:sagepub.com/journals-permissionsDOI: 10.1177/10497323211058164journals.sagepub.com/home/qhrHow do Positive Deviants OvercomeHealth-Related Stigma? An Exploration ofDevelopment of Positive Deviance AmongPeople With Stigmatized HealthConditions in IndonesiaSarju Sing Rai1,2, Elena V. Syurina1, Ruth M. H. Peters1, Annisa Ika Putri1,Irwanto Irwanto3, and Marjolein B. M. Zweekhorst1AbstractA constructivist grounded theory approach was used to understand how some people living with stigmatized healthconditions develop positive deviance to overcome stigma. We examined interviews from 13 identified positive deviantsliving with four different stigmatized health conditions (HIV, leprosy, schizophrenia, and diabetes) in Indonesia. Positivedeviance develops in the form of psychological empowerment through improvement of self-belief and perception(intrapersonal component), development of understanding and skill to exert control in life (interactional component),and self-discovery of successful behaviors and strategies to avert stigma (behavioral component). Positive deviants, afterbeing empowered, start empowering others affected by sharing their knowledge and fostering social awareness andacceptance. The findings revealed the presence of problem-solving ability and agency within the community of stig-matized individuals in Indonesia and warrant researchers to partner with the community to expedite the diffusion oftransferable positive deviant strategies within and outside the communities.Keywordsstigma; chronic illness and disease; power empowerment; adaptation; coping; enduring; southeast Asia; Asian people;grounded theory research strategies; IndonesiaIntroductionPositive deviance is an approach that seeks to solvecomplex societal problems by uncovering solutions thatalready exist in the society (Bradley et al., 2009; Marshet al., 2004). The approach entails identifying individualscalled “positive deviants” within the society who havemanaged to overcome a commonly shared problem bydevising their own unique techniques and mechanismsusing readily available resources within the communityand then learning and replicating those successful strat-egies to help others in the community (Marsh et al., 2004;Pascale & Monique, 2010). As these unique traits andstrategies are derived within the community using resourcesthat are available to all, this approach ensures easy diffusion,uptake, and sustenance of solutions in the community(Marsh et al., 2004; Positive Deviance Initiative, 2010).Studies have highlighted this very scope of positivedeviance as an approach to discovering solutions to avertexisting health-related adversities in the community(Bradley et al., 2009; Marsh et al., 2004). To date, positivedeviance has proven as an effective approach to tackling someof society’s complex health-related problems from childmalnutrition (Marsh et al., 2004), smoking cessation (Awofesoet al., 2008), weight loss/control (Stuckey et al., 2011),1Athena Institute, Faculty of Science, Vrije Universiteit Amsterdam,Amsterdam, The Netherlands2Barcelona Institute for Global Health (ISGlobal), University ofBarcelona, Barcelona, Spain3Faculty of Psychology, Atma Jaya Catholic University, Jakarta, IndonesiaCorresponding Author:Sarju Sing Rai, VU Amsterdam, De Boelelaan 1085, Amsterdam 1081 HV,The Netherlands.Email: s.s.rai@vu.nlretention in medical treatment (Assefa et al., 2014), to post-traumatic stress disorder from sexual abuse (Herlof Andersen,2008).The positive deviance approach may hold promise inaddressing the widely prevalent and complex issue ofhealth-related stigma in the society. Weiss et al. (2006)define health-related stigma as “a social process that in-stigates devaluation, exclusion and rejection of peoplebased on their health condition.” The experience of health-related stigma among people living with health conditionsis broadly categorized into two—enacted and felt stigma.Enacted stigma entails the experience of negative socialjudgment and discrimination, while felt stigma entailsinternalized shame associated with having a health con-dition and fear of experiencing discriminatory acts (Lekaset al., 2011; Scambler, 2004). Health-related stigma is apersistent problem in many communities around the world(Kemp et al., 2019; Sercu & Bracke, 2017; Stangl et al.,2019; Woodgate et al., 2020) and is known to negativelyaffect the quality of life of people living with stigmatizedhealth conditions (Holzemer et al., 2009; Mosanya et al.,2013; Tsutsumi et al., 2007) and hamper the treatment andmanagement of their disease (Sercu & Bracke, 2017; Stanglet al., 2019). Studies have shown that health-related stigma,particularly felt stigma, can be improved through variousstigma reduction strategies and interventions targeted to-wards people living with stigmatized conditions (Dadunet al., 2019; Lusli et al., 2016; Peters et al., 2016; Rao et al.,2019). For example, Dadun et al. (2019) and Lusli et al.(2016) reported significant reduction in felt stigma amongpeople with leprosy in Indonesia through socio-economicand rights-based counseling interventions respectively.However, existing interventions as such are often cost-intensive and non-sustainable in the long run (Kemp et al.,2019). Many of such stigma reduction interventions in-volve solutions that reside outside the community, whichinclude resources and practices that require heavy invest-ments and adopter buy-ins, and are often managed byexternal agents with minimal to non-existent involvementof the community (Kemp et al., 2019; Singhal et al., 2010).Studies have shown that community-based interventionsthat focus on locally derived and -led solutions can beeffective (Bogart & Uyeda, 2009), have greater uptake(Seyfang & Smith, 2007), and likely to be sustainable in thelong run (McLeroy et al., 2003).The concept of positive deviance has prospects inuncovering sustainable and effective solutions to stigmareduction that are self-derived and already exist within thecommunity, ensuring maximal community uptake andownership. People living with stigmatized health condi-tions have their own personal agency, and some succeed toovercome stigma in their lives through such self-derivedinternal mechanisms and strategies (Peters et al., 2014;Shih, 2004). Insights from the life experiences of thesepositive deviants can help disentangle the unique mecha-nisms and strategies used by them to mitigate their negativeexperiences of stigma and help deconstruct their successfulapproaches to identify those that can be transferred,practiced, and embedded within the community to helpothers. Understanding positive deviance can not only helpuncover and improve understanding on the mechanism ofself-agency, but also help in identification of effective,innovative, and sustainable solutions for stigma reductionthat may already exist in the community. However, this has,to our knowledge, not been done yet, and still remains anunderdeveloped and underexplored area of research.Therefore, we aimed to explore and understand theprocess of development of positive deviance to overcomestigma among people living with four distinct healthconditions—HIV, leprosy, schizophrenia, and diabetes—in Indonesia. The context of Indonesia as a developingupper-middle income country with high burden of bothcommunicable and non-communicable diseases providesan ideal prospect for this study (IHME, 2010; Kemp et al.,2019; Stangl et al., 2019). Studies have found that In-donesia not only has high burden of infectious diseases,particularly HIV (UNAIDS, 2017) and leprosy (WHO,2017), and non-communicable mental illness like schizo-phrenia (HRW, 2016) or a chronic metabolic disorder likediabetes (IDF, 2017), but also rampant stigma and dis-crimination associated with these four health conditions(Lusli et al., 2016; Pujilestari et al., 2014; Rai et al., 2020;Subu, 2015; Sianturi et al., 2019). Further, studies haveindicated that while the cause/origin and severity of stigmamay differ across different health conditions, the experi-ence of stigma and their consequences remain largelysimilar (Rai et al., 2020; Rao et al., 2019; Van Brakel et al.,2019). Taking into consideration such similarities that existin stigma experiences across different health conditions, wesought to explore the pathway in which people living withone of these four stigmatized diseases in Indonesia developpositive deviance to overcome stigma.Positive deviance approach for social change“Positive deviance (PD) is an approach to social change thatenables communities to discover the wisdom they alreadyhave, and then to act on it” (Pascale & Monique, 2010).The positive deviance approach strives for sustainablebehavioral and social change through identification ofsolutions that are community-derived, diffused and owned(Positive Deviance Initiative, 2010; Van Dick & Scheffel,2015). There are many methodologies for the positivedeviance approach (Marsh et al., 2004; Positive DevianceInitiative, 2010; Pascale & Monique, 2010). Nonetheless,the characteristics of the different methodologies are verysimilar. All methodologies acknowledge the importanceof identifying the needs of the community, the problem,2Qualitative Health Research 0(0)and the positive deviants. Furthermore, all methods try todiscover successful behaviors and/or practices of thepositive deviants; this is often called the positive devianceinquiry (Positive Deviance Initiative, 2010). The onlydifference exists in the final stage of the process, where thesteps can lead to either monitoring and evaluation (Marshet al., 2004; Positive Deviance Initiative, 2010) or con-tinuous research on the positive deviants within a com-munity and extend the positive deviance research to othercommunities (Fowles, 2007).One of the widely used methodologies is the oneproposed by the Positive Deviance Initiative (PositiveDeviance Initiative, 2010). This methodology consistsof five basic steps, namely, (i) defining the problem; (ii)determining the presence of positive deviants (usually fourto six people, as they are rare); (ii) discovering successfulbehavior that offers solution to the problem; (iv) designingactivities using the discovered solution; and (v) monitoringand evaluation. The first four steps are also known as thefour Ds (Positive Deviance Initiative, 2010). Within thescope of this study, only the four Ds of the positive de-viance methodology were utilized, with focus on assimi-lating the findings and recommending subsequent actionsfor the fourth D instead of designing activities.MethodsStudy Settings and ProcedureThis study was a part of an exploratory community-basedstudy on health-related stigma conducted in Jakarta andwest Java in Indonesia between March and June 2018among people with/affected by either of the four stigma-tized health conditions—HIV, leprosy, schizophrenia, anddiabetes. This study comprised of an exploratory qualita-tive design to understand the process of development ofpositive deviance. These four specific health conditionswere chosen because of their higher prevalence (MOH,2013; IDF, 2017; UNAIDS, 2017; WHO, 2017) in com-parison to other south-east Asian nations and stigmatizationin the Indonesian society (Lusli et al., 2016; Peters et al.,2013; Pujilestari et al., 2014; Subu, 2015; Sianturi et al.,2019). Participants were recruited purposively from thecommunity through referrals from community-based or-ganizations and peer-support groups. Those who were overthe age of 16 (age of consent) and who were willing andconsented to participate were included in the study. Thisstudy was based on in-depth interviews.Selection of Positive DeviantsA multi-stakeholder consultation with local researchers,community leaders, peer leaders, and representatives ofcommunity-basedorganizationswasundertakentoidentify positive deviants across the four conditions. Therole of the stakeholders in this study was limited to theidentification, selection, and recruitment of the partici-pants. As per the consultation, a four criteria definition forpositive deviants was derived. Positive deviants werethose who (i) successfully managed to overcome stigma intheir lives, (ii) were not ashamed of their disease, (iii)innovated unique mechanisms and ways to avert stigma,and (iv) managed to reintegrate into the society. A two-step identification and recruitment of positive deviantswas then conducted. In the first step, 21 individuals (HIV:6; leprosy: 6; schizophrenia: 5; diabetes: 4) were identifiedthrough referrals from local resource persons. This tookplace in different ways for each condition: possiblepositive deviants with leprosy were identified in Cirebonthrough recommendations from peers and resource per-sons from local leprosy-related organization, those withHIV and schizophrenia were identified through peers andcommunity-based organizations in Jakarta, and those withdiabetes were recruited through peer-recommendation andpatient support groups in Jakarta. In the second step, theidentified positive deviants were further confirmed throughpreliminary in-depth interviews using the aforementionedcriteria. Based on the interviews, 13 positive deviants werefinally selected (HIV: 4; leprosy: 3; schizophrenia: 4; di-abetes: 2). A final consultation was conducted among theresearch team and community stakeholders for consensuson the selection of positive deviants.Data CollectionQualitative semi-structured in-depth interviews wereconducted with identified positive deviants. Informedconsent was sought prior to participation in the study.Those who agreed were either interviewed in their homesor at the NGO offices in privacy. Prior to interview, so-ciodemographic information on age, gender, ethnicity,religion, income, etc. was collected from the participants.The interviews started in an exploratory manner andprogressed toward more in-depth questions to uncover theparticipants’ self-reported trajectory/process of positivedeviance. On average, the interviews lasted an hour.During the interviews, a guide was used (SupplementaryFile 1). The following topics were addressed by the in-terview guide to uncover the development of positivedeviance: general information, life history, experience ofliving with the health condition, experience of stigma anddiscrimination, turning point in life, strategies used toavert stigma, perceived changes after overcoming stigma,and current works and activities. Focused probing wasconducted to discover successful strategies and behavioremployed by participants to mitigate stigma in their lives. Theinterviews included exhaustive probing along with membercheck/respondent validation to ensure trustworthiness of theRai et al.3collected data. The interviews were recorded, transcribed,and translated into English. Data management and analysiswere performed using Atlast.ti software.Data AnalysisData analysis was independently carried out by the twoco-authors—one from the field of global health (SarjuSing Rai) and other from psychology (Elena Syurina). Theobtained codes were discussed among the authors: commoncodes were combined, while the different and discrepantcodes were further deliberated on and noted. All such codes/categories were discussed with and agreed upon by all co-authors in each stage of the analysis to make final decisionson the codes/categories. The constructivist grounded theoryapproach outlined by Charmaz (Charmaz, 2014) was usedfor data analysis. Using Charmaz’s approach, data analysisentailed an extensive process of coding and memo writing.The analysis was conducted as follows:Initial CodingThe transcripts were read line-by-line and incident-by-incident coding was conducted within each interview toobtain the initial codes.Focused CodingAfter obtaining the initial codes, they were compared acrossthe interviews from all positive deviants. Codes were firstclustered into categories based on the two preliminaryevolving themes that emerged from analysis—developmentof positive deviance and impact of positive deviance. As thecoding process proceeded, it became clear that some of thecategories corresponded to most positive deviants, whichbecame the main theoretical categories. There were somecategories that were less mentioned and formed sub-categories. The codes were then categorized and prelimi-nary categories were obtained.Theoretical CodingIn the theoretical coding process, the categories obtainedthrough focused coding from all 13 interviews were furtherstudied, compared, and organized. The process entailedinterlinking and synthesizing of those categories that led tointegration of interrelated categories into themes. Threemain themes emerged from theoretical coding, viz. “triggers”that led to positive deviance, “empowerment” as an actualprocess of positive deviance, and “impact” of positive de-viance. Under each theme were interrelated categories andsub-categories. The process of theoretical coding led to apreliminary theoretical framework obtained through assimila-tion and thematic conceptualization of interrelated categories.Application of Extant Theoretical ModelsAs per the constructivist grounded theory approach, theresultant theoretical concepts were compared with ex-isting relevant literature in the later stage of data analysis.The theoretical categories pertaining to the theme “em-powerment” was found to bear strong similarity to theexisting theoretical models of psychological empowerment.Upon further review and comparison, the authors deemed fitto integrate the theoretical concept of “empowerment” intothe existing theoretical model of psychological empower-ment by Zimmerman (Zimmerman, 1995).Zimmerman’s psychological empowerment model(Zimmerman, 1995) was used to describe the process ofdevelopment of positive deviance. The model posits thatpsychological empowerment entails three components,viz. intrapersonal, interactional, and behavioral compo-nent, which all work together in unison to empower anindividual. The first component—intrapersonal empow-erment—refers to how individuals think about themselvesand perceive control, self-efficacy, and motivation in theirlives. This component is about the individual’s perceptionof self-esteem, self-view, and locus of control. The secondcomponent—interactional empowerment—refers to thecritical awareness and understanding people have abouttheir condition, community, and society (norms, structures,etc.), and the ability and skillset individuals have to make achange. This component posits that individuals need to learnabout their options and scenario in a given context in order tobe able to exert control in their environment. The finalcomponent—behavioral empowerment—occurs when in-dividuals have both intrapersonal and interactional em-powerment to drive a positive behavior change. Behavioralempowerment refers to actual actions/behaviors/strategiesundertaken to make a change and directly influence out-comes in one’s life (Zimmerman, 1995).Memo WritingMemo writing was a crucial part of the entire data analysisprocess as outlined by Charmaz (Charmaz, 2014). Theauthors wrote down memos on the evolving codes andresultant categories and themes in every step of theanalysis. The memos helped in thorough reflection,comparison, and elucidation of the emerging codes andcategories and guided the whole analytical process.Final Theoretical FrameworkA final theoretical framework of the process of developmentof positive deviance was obtained after integration of the mainthemes with the extant theoretical model. The final theoreticalmodel includes four thematic stages that underline the processof development of positive deviance among the participants.4Qualitative Health Research 0(0)Rigor of StudyThe analytical process was peer debriefed and discussedin each stage to ensure the quality and trustworthiness ofcodes, categories, and themes. The authors made sure allcodes, categories, and themes were obtained until the pointof saturation. The results of the analysis were carefullyreviewed and refined in all stages. Rigorous vetting anddeliberation were carried out among the authors in deter-mining the causal and consequential relationships betweenthe categories and developing the preliminary theoreticalframework from thematic coding, and then integrating thetheoretical concepts into existing theoretical models afterliterature review. The final framework was thus obtainedafter consensus from all authors on the final themes and thelogical connection and pathway they describe. An outlineof the analytical process and resultant outcomes in eachstage of analysis is presented in Supplementary File 2.Ethical ApprovalThis study was performed in line with the principles of theDeclaration of Helsinki. Approval was granted by the EthicsCommittee of Atma Jaya Catholic University, Indonesia(Approval ID: FR-UAJ-26-13/R0). All participants pro-vided verbal and written consent to participate in the study.ResultsDemographic Information About ParticipantsA total of 13 individuals were identified as positive de-viants. Out of them, four were living with HIV, three wereaffected by leprosy (two with visible physical impair-ments), four with schizophrenia, and two with diabetes(with no diabetic complications and resulting impairments).There were six male and female respondents, while oneidentified as transgender. The majority of the respondentsbelonged to Jawa and Sunda ethnicity (67.6%), and fol-lowed Islam (69.2%). The median age of the positivedeviants was 41.31 years (SD 14.22) with an age rangebetween 20 and 75. The mean duration of living with/having had the condition was 12.31 years (SD 5.99) with arange between 1 and 25 years.The Process of Development of Positive DevianceThe final theoretical framework obtained from the con-structivist grounded theory analysis described four stagesof the process of development of positive deviance amongthe participants: (i) initiation, where specific triggers helpedinitiate the process of psychological empowerment amongpositive deviants, (ii) self-strengthening, where the processof psychological empowerment repeated in a loop whilefurther strengthening its components, (iii) impact, whereafter being empowered, positive deviants started helpingothers and contributing to the society, and (iv) reinforcement,where the process of impact and psychological empower-ment repeated in a reiterative cycle, enriching each other inevery iteration. The framework illustrating the four stages ofdevelopment of positive deviance is outlined in Figure 1.InitiationTheoretical coding showed that participants developedpositive deviance through psychological empowerment.The process of psychological empowerment adapted fromZimmerman’s model (1995) constituted three components:a.Intrapersonal component: Positive deviants reportedof intrapersonal component of psychological em-powerment that helped them improve their self-beliefand perception and led them to take charge of theirlives. The categories of intrapersonal componentincluded: (1) “self-acceptance,” wherein positivedeviants talked about how accepting themselves andtheir health condition was the first step to feelingbetter about themselves and improving self-esteem;(2) “self-confidence,” which played an important rolefor them to be self-aware of their ability to perceivecontrol of their life; (3) “positive outlook,” whereinpositive deviants reported of having positive insightsand outlook to life, which ignited a sense of hope-fulness and normalcy in their lives; and (4) “moti-vation,” which provided them the incentive to keepgoing on and striving for better.Motivation was further categorized into three sub-categories based on its sources: (i) “empathy towardsother stigmatized individuals,” where motivation camefrom being able to relate to others living with the samehealth condition as them and having empathy towardsthem; (ii) “experience of stigma/discrimination,” whereinmotivation came from the experience of stigma in theireveryday life from family, friends, and healthcare pro-fessionals; and (iii) “concern for loved ones,” wheremotivation came from the concern that positive deviantshad for their loved ones—especially close family mem-bers (children, spouse, father, mother, etc.).b.Interactional component: Positive deviants re-ported of interactional components that helpedthem develop critical awareness and understandingof their disease and the environment, and gainskills to manage stigma. This included (1) “ex-periential knowledge,” wherein positive deviantsreported of how their personal experiences ofliving with the health condition provided a betterRai et al.5understanding of not only their disease but also thecontext of stigma surrounding them; (2) “infor-mation,” wherein they talked about how gainingrelevant knowledge and information, either fromother experts or through information resources,helped them understand their disease and think ofways to manage them; (3) “situational assessmentand awareness,” in which they talked about theimportance of having a critical awareness of theenvironment and the social norms in one’s dailylife, and understanding measures to fit-in as a wayto avert stigmatization; (4) “disclosure consider-ations,” wherein they highlighted the importanceof understanding the people, their beliefs andspecific contexts before disclosing one’s healthstatus to others; (5) “support from others,” wherethey talked about how perception of support fromothers, especially family and friends, provide pos-itive reinforcement in their lives; and (6) “resourcemobilization skills,” wherein positive deviantsdiscussed how critical awareness of available re-sources or ways to accessing different resources isparamount to effective management of their life.c.Behavioral component: Behavioral component inthis study entails successful behavior changes ex-hibited by the positive deviants to overcome stigma.Positive deviants reported of two different conceptsof behavior change—“personal changes” that helpedthem in their everyday life to overcome stigma, and‘transformational leadership’ in which they startedhelping others affected to fight against stigma.Personal changes included behaviors like (1) “self-care,”wherein positive deviants highlighted the importance ofcaring for themselves for not only effective management andtreatment of their disease but also for overall well-being.Self-care entailed both physical/medical care (taking med-ications regularly, following through with health check-ups,etc.) and spiritual care (meditation, prayers, etc.); (2) “in-difference to others’ stigmatizing views/behavior,” whereinthey talked about how they ignored and did not care aboutwhat others thought of them or talked about them. By thispoint, they perceived enough control of their life and en-vironment that the public views did not affect them; (3)“disclosure of one’s status,” wherein disclosure of theirhealth status was a very important action for positive de-viants, which they said not only helped them feel liberatedfrom shame and stigma associated with their disease, but alsohelped garner support and understanding from others; (4)and “situational adjustments to assert control of the envi-ronment,” where they talked about how adjusting accordingto different contexts, cultures, and situations help them avertstigma and judgment in their daily lives. They highlightedthe importance of critical awareness of the environment andnorms in helping decide the required adjustments needed.Transformational leadership included behaviors like(1) “leading by example (being a role model),” whereinhaving understanding and empathy towards others livingwith their health conditions, positive deviants discussedhow they started helping others by showcasing their ownlife experiences of living with the condition and mentoringand guiding them; (2) “helping and supporting peers,”where positive deviants actively helped others affectedthrough actions like home visits to track their treatment,emotional support and counseling, connecting/referringthem to relevant people within their network, lendingmoney when in need, etc.; (3) “activism,” wherein theytalked about their passion to fight for their rights and thoseof other people affected. They showed aversion to in-equality and injustice and actively spoke against those,while demanding equality and justice; and (4) “advocacy,”wherein some positive deviants reported of going a stepfurther ahead from activism towards advocacy where theysystematically worked with different stakeholders on af-fecting policy changes to improve the condition of peopleliving with the health condition as them.A summary of the categories of the three aforemen-tioned components, their description, and illustrativequotes are provided in Supplementary File 3.Role of Triggers in InitiatingPsychological EmpowermentPositive deviants were found to already have differentlevels of intrapersonal and interactional component, as theyhad already gone through various experiences and expo-sures in their lives to provide for these two components ofempowerment. However, to initiate the process of psy-chological empowerment, the two components needed tolead to behavioral component (i.e. to initiate behavioralFigure 1. The process of development of positive deviance.6Qualitative Health Research 0(0)changes to avert stigma). For this to happen, a “trigger” wasnecessary, which the participants called “the turning pointin their life.” Triggers were found to act like sparks thatstarted the process of psychological empowerment by ei-ther driving interpersonal or interactional component thatultimately led to behavior change.For positive deviants who had adequate interactionalcomponent but lacked sufficient intrapersonal component,a trigger in the form of added motivation helped in ini-tiating the empowerment process. The motivationaltrigger came in different forms. For some positive devi-ants, motivation came after remission of their disease orimprovement of their health after treatment. For others,motivation took shape in the form of indignation andrevolt against discrimination. One positive deviant livingwith HIV gave an account of how at first she was com-placent and tolerated discrimination from others, but whenit was directed to her loved ones—she revolted: “[…] aslong as the discrimination was only directed to me... […] Ijust accepted it. But…when I saw people close to megetting discriminated against, I got angry. […] my hus-band needed to get operated and we got denied fromhospital […]. I ended up getting really angry. I reported itto ministry of health and I took it to the media. I ended upmeeting the president of the hospital because I also didn’twant it to get to the media. (Female, HIV)”In case of positive deviants with adequate intrapersonalcomponent but not enough interactional component, atrigger in the form of information that built their knowledgeand skill helped strengthen interactional empowerment todrive behavior change and initiate the process of positivedeviance. Information received through internet, socialmedia, friends, family, healthcare professionals, and otherindividuals living with the condition (peers) helped ignitetheir passion to make positive change(s) in their lives. Onepositive deviant living with HIV who was in prison whentested positive for HIV recounted his experience of howinformation helped trigger positive deviance in his life:“The information that I got inside (the prison) … aboutHIV was limited. […] So, I called my friends and told themabout my status, wishing they could look for books [onHIV]. But it turns out there were none, it was not that easy.[…] I could only look at it from google, but Informationthat I got were limited. I tried to look for more in-formation…luckily, there was a doctor from America whowas conducting research inside the prison, so he/she told us[…] as long as you have HIV, you could get medication totreat HIV. Knowing that, I fought for my rights and startedarguing why we needed to wait until our condition gotworse, if we were ready! (Male, HIV)”Triggers initiated psychological empowerment in the livesof the respondents in different stages of their lives. Somepositive deviants encountered triggers at the earliest pointwhen they came to know of their disease status, while forsome it took a longer time—until a specific event or expe-rience drove them to make changes in their lives. Further, thestrength of triggers also differed as per individuals. For some,triggers had to be strong and cathartic to catalyze positivechange like a traumatic experience of discrimination, while forsome, they were very mild and gentle, much like an epiphany.Self-StrengtheningOnce initiated, the process of psychological empower-ment was found to strengthen further by functioning in areiterative loop, with each empowerment componentleading to strengthening of either of the other two com-ponents, and so forth (see Figure1). One positive deviantwith schizophrenia described this process which startedwith him recovering and feeling better and confident(intrapersonal empowerment), which led him to activelysearch for information and resources regarding his con-dition and learning ways to improve it (interactional em-powerment), finally leading to making important changesin his life to improve his condition (behavioral empow-erment): “I recovered (from schizophrenia) and I startedlooking more into what was I sick with. I looked it up (ingoogle) and I learned that the symptoms matched with mine.But I did not stop there and started looking for more. Icreated my Facebook account […] and I found KPSI (NGOon Schizophrenia) and contacted them. After meeting otherslike me I was then sure that the pills really do help with ourrecovery. So, I regularly took my pill from 2010 until now.Thank God I haven’t relapsed since then. (Male, Schizo-phrenia).” Positive deviants, once entered into the processof psychological empowerment, were found to undergo theprocess reiteratively further building up on each of the threecomponents that led to self-strengthening with time.ImpactWith initiation, development and strengthening of psy-chological empowerment, the positive deviants ultimatelysought to bring about two distinctive impacts—directedtoward others living with their health condition and to-ward the outside world. The first type was to strengthenand empower other stigmatized individuals from theirgroup through “knowledge diffusion.” Positive deviantsreported of how they were able to help others understandtheir disease and ways to manage it and avert stigma andother adversities in life. They talked about sharing withothers the behaviors and strategies they used in their ownlives, which helped them. One positive deviant with HIVtalked about the advice he gives to others living with HIV:“I feel happy that I have been able to help others with HIV.I always tell them that accepting yourself first is the mostimportant thing. If you can’t accept yourself, how willothers accept you. If you are able to accept yourself,Rai et al.7others will follow. Even if others can’t accept you, at leastyou can accept yourself. (Male, HIV).”The second impact was to “foster wider social awarenessand acceptance” of people living with the condition. Positivedeviants talked about their yearning for acceptance andinclusion in the society, for which they made sure to portraya positive picture of how people with health conditions asthem are just like any other person and are a productivemember of the society. One positive deviant with HIVexplained: “[…] I always try to show people that despite ofhaving HIV, I am a productive person who contributes to thesociety. People will set aside the fact that you have HIV….even if you’re gay and LGBT ….if they see you have skillsand are productive. I taught in some government agenciesand they don’t care that I have HIV, they just appreciate whatcan I give to the company. So that’s what I always emphasize“what can you give to other people?”. If you do good things,people will respect and appreciate you. (Male, HIV).”ReinforcementSuccessfully achieving impact was found to further en-courage and reinforce the positive deviants’ psychologicalempowerment, which in turn encouraged them to strive formore impact—thus creating a reiterative positive loop be-tween the process of psychological empowerment and theresultant impact (as illustrated in Figure 1). Positive deviantstalked about how they felt more driven and motivated andstrived to work harder to help others after seeing the positiveimpact they have been able to bring in the community and topeople’s lives. One positive deviant with leprosy said inregards to helping others affected by leprosy: “When I sawhow I was able to motivate other people with leprosy, Ibecame even more motivated. I started learning new skillsand working with other villagers to improve our villageinfrastructure. I showed everybody that even though I haveleprosy, I can positively contribute to the society. Seeing mywork, my friends with leprosy are even more motivated. …(Male, Leprosy).” Another positive deviant with HIV talkedabout how he managed to show the public that “people withHIV can equally contribute in the society” by teachingaerobics to the police, and reported receiving a lot of supportfrom the public, which further fueled his desire to do more toimprove their understanding and perception towards peopleliving with HIV in the society.DiscussionThrough this study, we attempted to explore and under-stand the process of development of positive devianceamong people living with stigmatized health conditions inIndonesia as a way to avert stigma. Positive deviantsexhibited psychological empowerment triggered by theirunique life events and experiences. Once initiated, theprocess of psychological empowerment was found to self-strengthen by functioning in a reiterative loop—with eachempowerment component leading to strengthening ofeither of the other two components and so forth. Withinitiation and self-strengthening of psychological em-powerment, the ultimate impact that the positive deviantswere trying to bring about were empowerment of otherstigmatized individuals from their group through diffusionof successful strategies in the community and striving forsocial change by fostering wider social awareness andacceptance of people living with the condition. Theprocesses of impact and psychological empowermentwere found to repeat in a reiterative cycle, enriching andbuilding upon each other in every iteration.Positive deviance developed in the form of psycho-logical empowerment which helped individuals overcomethe experiences of stigma and related adversities in ev-eryday life. Depending on the field of research and thesetting, positive deviance can take different forms (HerlofAndersen, 2008; Marsh et al., 2004). Among Vietnamesefarmers facing the issue of child malnutrition, positivedeviance came in the form of going against the com-munity tradition and diversifying the children’s diet toinclude other macronutrients other than just carbohydrates(Marsh et al., 2004). For men dealing with childhoodexperience of sexual abuse, positive deviance took shapein the form of focusing and building on a positive identitythat identifies less with the experience of abuse (HerlofAndersen, 2008). In this study, psychological empower-ment was indicative of positive deviants’ ability to per-ceive control in their life (interpersonal empowerment)and develop understanding of the social contexts andsystems and ways to maneuver them (interactional em-powerment), which enabled them to exert required actionsto bring about positive changes in their lives that helpedthem overcome stigma and related adversities (behavioralempowerment) (Zimmerman, 1995).Further, once the process was initiated, the three com-ponents of psychological empowerment reiteratively self-strengthened each other while also leading the positivedeviants to make an impact by empowering others andfostering public awareness. This in turn further strengthenedthe process of psychological empowerment among positivedeviants that again led to more impact—thus creating apositive reiterative loop. This is consistent with Fre-drickson’s broaden-and-build theory (Fredrickson, 2001)which posits that positive psychological processes oftenenhance and grow in each thought-action repertoire creatingan upward spiral. Studies on positive psychology havedemonstrated this particular phenomenon where positiveprocesses like coping with adversities, psychological resil-ience, and emotional well-being keep building and growingin a cyclical manner with each positive thought-actionrepertoire (Aspinwall, 1998; Fredrickson & Joiner, 2002).8Qualitative Health Research 0(0)The findings indicate that psychological empowermentcan be an effective way to overcome stigma among peopleliving with stigmatized health conditions. This corrobo-rates with the works of Oyserman and Swim (2002) andShih (2004) which posited that empowerment can helpstigmatized individuals overcome adversities caused bystigma, build self-efficacy, and reject negative labels as-sociated with their stigmatizing trait. This may be apossible mechanism that can lead to reduction of feltstigma. Hence, interventions focused on psychologicalempowerment may have prospects in stigma reductionamong those living with different health conditions. How-ever, there may be two different approaches to designingsuch interventions. The first approach is the conventionaletic empowerment approach led by external agents (re-searchers, program managers, etc.) whereby specific inter-ventions are designed and introduced targeting the differentcomponents of psychological empowerment (Chopra &Ford, 2005; Sadan, 1997). Studies have indicated thatsuch interventions can be effective (Eisman et al., 2016;Peterson & Reid, 2003). Eisman et al. found that targetingthe three empowerment components not only improved theoverall psychological empowerment, but also increased thelikelihood of positive behaviors among youth (Eisman et al.,2016). Further there have been stigma reduction interven-tions focused on empowering people living with stigmatizedhealth conditions, which have shown promising results inreducing felt-stigma (Dadun et al., 2019; Ebenso et al., 2007;Peters et al., 2016). However, studies have found that suchinterventions are often resource-intensive, have poor uptakeby target beneficiaries, and not sustainable on the long run(Singhal et al., 2010; Singhal & Svenkerud, 2019; Stanglet al., 2019). This may not only have happened because ofthe obvious reason that the etic approach is often designedand led by external agents, but also because such inter-ventions extensively focus on successful strategies whichmay work only in context of certain individuals and settings,and not the wider community.It is important to understand that for individuals likepositive deviants, psychological empowerment is a self-derived mechanism, while for others, it might not be asobviously feasible. It can therefore be beneficial to followa second approach—the positive deviance approach(Marsh et al., 2004; Positive Deviance Initiative, 2010;Pascale & Monique, 2010) that focuses in identifying themost usable and transferable behaviors or strategiesemployed by positive deviants that can be easily im-plemented, practiced, and shared by others in the com-munity, which may help lead those who are not positivedeviants into the path of psychological empowerment. Forexample, strategies such as selective disclosure of one’shealth condition can be emulated by people living withstigmatized health conditions. As opposed to full dis-closure, selective disclosure entails disclosing of one’shealth condition to a selected group of people who thestigmatized individual trusts and/or expects understandingand support from (Corrigan & Rao, 2012). While thisstrategy may not help individuals fully mitigate stigma asfeelings of shame and apprehension may still exist, it isknown to provide a needed reprieve to stigmatized in-dividuals by gaining supporters/confidants and getting asense of control of how, when and who they disclose theirhealth condition to (Black & Miles, 2002; Corrigan &Rao, 2012). This strategy has already proven effectiveamong people living with HIV (Black & Miles, 2002;Kumar et al., 2015) and mental illness (Bos et al., 2009;Corrigan & Rao, 2012) in coping with stigma. Hence,such positive deviant strategies have prospects for broaderacceptance, ownership, and practice in the community asthey are derived from people from the community withsimilar experiences and contexts (Pascale & Monique,2010). The positive deviance approach may thus have theanswer to designing future stigma interventions that arenot only effective but also sustainable.Positive deviants living with stigmatized conditionshad passion and drive to help others in the community andyearned to make an impact among both other stigmatizedpeers and the society. Besides helping other stigmatizedindividuals, becoming role models, and advocating fortheir rights, they shared personal stories and specificstrategies in their community. This may have already ledto a diffusion of innovation (transference and transfor-mation of tacit knowledge to implicit knowledge) withintheir community. Studies on psychological empowermenthave shown that people who are empowered tend to em-power others (Eisman et al., 2016; Sadan, 1997; Zimmerman,1995). Eisman et al. (Eisman et al., 2016) found thatpsychological empowerment among youths in turn helpedincrease their prosocial involvement and contribution to thecommunity. Positive deviants were further involved inadvocacy to help other oppressed stigmatized individuals.This is consistent with the findings from Sadan (Sadan,1997) who noticed that empowered individuals focused onadvocacy to ensure even the weakest people can accessempowerment process by creating the minimal environ-mental conditions.Further, they strived for social change and were known toput themselves out in the public eye to foster understandingand awareness of their condition. They aimed to elicit publicunderstanding and empathy toward themselves and otherpeople living with their condition—paving the way to socialacceptance, reintegration, and possible inclusion. Sadan(Sadan, 1997) noted similar behavior among empoweredpeople where empowerment extended beyond personalcontext like individual achievements—toward a path ofsocial transformation. On discovering one’s right and abilityto control one’s destiny, empowered individuals were knownto systematically change their lives and environment, andRai et al.9consequentlystriveforsocialandpoliticalchange(Wallerstein & Bernstein, 1994; Sadan, 1997).The findings from this study show prospects of thepositive deviance approach to catalyze social change fromwithin the community and provide comprehensive in-formation on the “know-how” for those affected to em-power themselves and overcome stigma and its negativeeffects. However, it is also important for the externalagents and environment to also facilitate the process ofsocial change. As per the social change theory it is im-portant to influence both the normative and socialstructures in order to bring about social transformation(De La Sablonnière, 2017). The positive deviance ap-proach is mostly focused in the normative level (Pascale &Monique, 2010; Singhal et al., 2010). While the approachmay extend to exert influence on the social structural level(Marsh et al., 2004), the attempt may still fall short. Inorder to deal with a deeply rooted and complex issue ofhealth-related stigma, it is important for allies from out-side of the community to help in efforts to effect change inthe social structural level (Singhal et al., 2010; De LaSablonnière, 2017; Singhal & Svenkerud, 2019).In this context, researchers, community leaders, andhealth program managers have an important role withinthe positive deviance approach that first and foremostincludes understanding and acknowledging the existenceof problem-solving ability and innovation within thecommunity of stigmatized individuals. This is followedby partnering with the community members in identifyingpositive deviants and their successful strategies that can beused with ease within the community, and wider diffusionof those strategies. Researchers can help expedite thisprocess of diffusion of innovation within the community.Further, researchers can play a very important role in thecross-diffusion of such successful strategies from onecommunity to another, and subsequent participatorymonitoring and evaluation of the process of social change.It is important for researchers and program managers tobuild an equitable alliance with the community membersand support positive deviants as role models and leaders inthe community. Further they can also play an importantpart in linking the community with other relevant stake-holders and advocacy partners to establish a united front toaddress both the normative and social structural levelchanges needed for stigma reduction in the society.This study also has certain limitations. While we un-covered how the process of self-strengthening and impactreiteratively reinforce each other, we could not determine theextent of the reiterations, specifically if such processes keepon building up throughout the whole life of the positivedeviants or if they plateau at some point in life. Future studiesshould explore this further. Further, this study does not takeinto account the differences in the origin and severity ofstigma across different health conditions, but rather focuseson the overall process of overcoming stigma. It is thereforeimportant to take these unaccounted factors into consider-ation while formulating stigma reduction strategies based onthe findings from this study. Despite the limitations, thisstudy has several notable strengths. This study was able toidentify 13 positive deviants through rigorous vetting. Marshet al. (Marsh et al., 2004) have recommended a sample sizeof four to six participants in research involving the positivedeviance approach owing to the rarity of positive deviants inthe community. However, by partnering with the communitystakeholders, we were able to identify and verify a muchlarger number of positive deviants than expected. To ourknowledge, this was the first study to use the positive de-viance approach to explore and understand how some in-dividuals manage to overcome health-related stigma. Futurestudies should further expand on this approach by part-nering with the community as an ally to design activities orinterventions that can help other stigmatized individualsovercome stigma. Further, it is recommended that thepositive deviant behavior and strategies identified in thestudy be utilized in stigma reduction interventions to assessits transferability, translationality, and effectiveness in non-positive deviants living with stigmatized health conditions.ConclusionThe positive deviance approach employed in this study tounderstand how positive deviants overcome stigma dem-onstrates the presence of problem-solving ability and agencywithin the community of stigmatized individuals in In-donesia. Positive deviance took shape in the form of psy-chological empowerment. The process of how positivedeviance develops among people living with stigmatizedhealth conditions and the successful behavior they employ toavert stigma in their lives may hold promise in helping othersaffected by stigmatized diseases in Indonesia in overcomingstigma and fostering social inclusion through the process ofpsychological empowerment. Positive deviants have beencrucial in not only coming up with these unique solutionsand strategies to avert stigma, but also diffusion of thosethrough knowledge sharing and support to other stigmatizedindividuals. Further, they strive for social change by fos-tering awareness and relatedness, and changing the per-ception of the societies’ way of viewing people living withhealth conditions. Researchers can play an important role inthis process by partnering with the community and helpingexpedite the process of diffusion of positive deviant strat-egies within and outside the communities.Declaration of Conflicting InterestsThe author(s) declared no potential conflicts of interest withrespect to the research, authorship, and/or publication of thisarticle.10Qualitative Health Research 0(0)FundingThe author(s) disclosed receipt of the following financial supportfor the research, authorship, and/or publication of this article:This study was supported by Leprosy Research Initiative (LRI),706.18.20.LRI; Erasmus Mundus Joint Doctorate (EMJD)Fellowship-Trans Global Health Consortium, 2013-0039.ORCID iDsSarju Sing Rai https://orcid.org/0000-0001-5914-1471Annisa Ika Putri https://orcid.org/0000-0002-4398-3567Supplemental MaterialSupplemental material for this article is available online.ReferencesAspinwall, L. G. (1998). Rethinking the role of positive affect inself-regulation. Motivation and Emotion, 22(1), 1–32.https://doi.org/10.1023/A:1023080224401Assefa, Y., Lynen, L., Wouters, E., Rasschaert, F., Peeters, K., &Van Damme, W. (2014). How to improve patient retentionin an antiretroviral treatment program in Ethiopia: a mixed-methods study. BMC Health Services Research, 14(1), 45.https://doi.org/10.1186/1472-6963-14-45Awofeso, N., Irwin, T., & Forrest, G. (2008). Using positivedeviance techniques to improve smoking cessation out-comes in New South Wales prison settings. Health Pro-motion Journal of Australia: Official Journal of AustralianAssociation of Health Promotion Professionals, 19(1), 72.https://doi.org/10.3316/informit.037570737860686Black, B. P., & Miles, M. S. (2002). Calculating the risks andbenefits of disclosure in African American women who haveHIV. Journal of Obstetric, Gynecologic, & Neonatal Nursing,31(6), 688–697. https://doi.org/10.1177/0884217502239211Bogart, L. M., & Uyeda, K. (2009). Community-based partic-ipatory research: Partnering with communities for effectiveand sustainable behavioral health interventions. Health Psy-chology, 28(4), 391–393. https://doi.org/10.1037/a0016387Bos, A. E., Kanner, D., Muris, P., Janssen, B., & Mayer, B.(2009). Mental illness stigma and disclosure: Conse-quences of coming out of the closet. Issues in MentalHealth Nursing, 30(8), 509–513. https://doi.org/10.1080/01612840802601382Bradley, E. H., Curry, L. A., Ramanadhan, S., Rowe, L.,Nembhard, I. M., & Krumholz, H. M. (2009). Research inaction: Using positive deviance to improve quality of healthcare. Implementation Science, 4(1), 25. https://doi.org/10.1186/1748-5908-4-25Charmaz, K. (2014). Constructing grounded theory. SAGE.Chopra, M., & Ford, N. (2005). Scaling up health promotioninterventions in the era of HIV/AIDS: Challenges for arights-based approach. Health Promotion International,20(4), 383–390. https://doi.org/10.1093/heapro/dai018Corrigan, P. W., & Rao, D. (2012). On the self-stigma of mentalillness: Stages, disclosure, and strategies for change. TheCanadian Journal of Psychiatry, 57(8), 464–469. https://doi.org/10.1177%2F070674371205700804Dadun, D., Peters, R. M., van Brakel, W. H., Bunders, J. G.,Irwanto, I., & Regeer, B. J. (2019). Assessing the impact ofthe twin track socio-economic intervention on reducingleprosy-related stigma in Cirebon district, Indonesia. In-ternational Journal of Environmental Research and PublicHealth, 16(3), 349. https://doi.org/10.3390/ijerph16030349De La Sablonnière, R. (2017). Toward a psychology of socialchange: A typology of social change. Frontiers in Psy-chology, 8, 397. https://doi.org/10.3389/fpsyg.2017.00397Ebenso,B., Fashona,A., Ayuba, M., Idah, M., Adeyemi,G., & S-Fada,S. (2007). Impact of socio-economic rehabilitation on leprosystigmainNorthernNigeria:Findingsof aretrospective study.AsiaPacific Disability Rehabilitation Journal, 18(2), 98–119.Eisman, A. B., Zimmerman, M. A., Kruger, D., Reischl, T. M.,Miller, A. L., Franzen, S. P., & Morrel-Samuels, S. (2016).Psychological empowerment among urban youth: Mea-surement model and associations with youth outcomes.American Journal of Community Psychology, 58(3–4),410–421. https://doi.org/10.1002/ajcp.12094Fowles, E. R. (2007). Collaborative methodologies for ad-vancing the health of underserved Women. Family &Community Health, 30(1), S53–S63.Fredrickson, B. L. (2001). The role of positive emotions inpositive psychology: The broaden-and-build theory ofpositive emotions. American Psychologist, 56(3), 218.https://doi.org/10.1037/0003-066X.56.3.218Fredrickson, B. L., & Joiner, T. (2002). Positive emotions triggerupward spirals toward emotional well-being. PsychologicalScience, 13(2), 172–175. https://doi.org/10.1111/1467-9280.00431Herlof Andersen, T. (2008). Men dealing with memories ofchildhood sexual abuse: conditions and possibilities of‘positive deviance. Journal of Social Work Practice, 22(1),51–65. https://doi.org/10.1080/02650530701872355Holzemer, W. L., Human, S., Arudo, J., Rosa, M. E., Hamilton,M. J., Corless, I., Robinson, L., Nicholas, P. K., Wantland,D. J., Moezzi, S., Willard, S., Kirksey, K., Portillo, C., Sefcik,E., Rivero-M´endez, M., & Maryland, M. (2009). ExploringHIV stigma and quality of life for persons living with HIVinfection. Journal of the Association of Nurses in AIDS Care,20(3), 161–168. https://doi.org/10.1016/j.jana.2009.02.002HRW. (2016). Living in Hell: Abuses against People with psycho-social Disabilities in Indonesia. https: //www.hrw.org/report/2016/03/20/living-hell/abuses-against-people-psychosocial-disabilities-indonesia (Accessed 20 August 2019).IDF. (2017). Prevalence of Diabetes in Indonesia. https://idf.org/our-network/regions-members/western-pacific/members/104-indonesia.html (Accessed 15 November 2019).IHME. (2010). GBD Profile: Indonesia. http://www.healthdata.org/sites/default/files/files/country_profiles/GBD/ihme_gbd_country_report_indonesia.pdf (Accessed 20 August 2019).Rai et al.11Kemp, C. G., Jarrett, B. A., Kwon, C. S., Song, L., Jette, N.,Sapag, J. C., Bass, J., Murray, L., Rao, D., & Baral, S.(2019). Implementation science and stigma reduction in-terventions in low- and middle-income countries: A sys-tematic review. BMC Medicine, 17(1), 6. https://doi.org/10.1186/s12916-018-1237-xKumar, S., Mohanraj, R., Rao, D., Murray, K. R., & Manhart,L. E. (2015). Positive coping strategies and HIV-relatedstigma in south India. AIDS Patient Care and STDs, 29(3),157–163. https://doi.org/10.1089/apc.2014.0182Lekas, H.-M., Siegel, K., & Leider, J. (2011). Felt and enactedstigma among HIV/HCV-coinfected adults: The impact ofstigma layering. Qualitative Health Research, 21(9),1205–1219. https://doi.org/10.1177/1049732311405684Lusli, M., Peters, R., van Brakel, W., Zweekhorst, M., Iancu, S.,Bunders, J., & Regeer, B. (2016). The impact of a rights-basedcounselling intervention to reduce stigma in people affected byleprosy in Indonesia. PLoS Neglected Tropical Diseases, 10(12),e0005088. https://doi.org/10.1371/journal.pntd.0005088Marsh, D. R., Schroeder, D. G., Dearden, K. A., Sternin, J., &Sternin, M. (2004). The power of positive deviance. BMJ:British Medical Journal, 329(7475), 1177. https://doi.org/10.1136/bmj.329.7475.1177Mcleroy, K. R., Norton, B. L., Kegler, M. C., Burdine, J. N., &Sumaya, C. V. (2003). Community-based interventions.American Journal of Public Health, 93(4), 529–533.https://doi.org/10.2105/AJPH.93.4.529MOH – Ministry of Health Indonesia. (2013). Riset KesehatanDasar. https://www.depkes.go.id/resources/download/general/HasilRiskesdas2013.pdf (Accessed 20 August 2019).Mosanya, T. J., Adelufosi, A. O., Adebowale, O. T., Ogunwale,A., & Adebayo, O. K. (2013). Self-stigma, quality of lifeand schizophrenia: An outpatient clinic survey in Nigeria.International Journal of Social Psychiatry, 60(4), 377–386.https://doi.org/10.1177/0020764013491738Oyserman, D., & Swim, J. (2002). Stigma: An insider’s view.Journal of Social Issues, 57(1), 1–14. https://doi.org/10.1111/0022-4537.00198Pascale, R., & Monique, S. J. S. (2010). The power of positivedeviance. Harvard Business Press.Peters, R. M., DadunLusli, M., Miranda-Galarza, B., VanBrakel, W. H., Zweekhorst, M. B., Damayanti, R., Seda,F. S., Bunders, J. F., & Irwanto (2013). The meaning ofleprosy and everyday experiences: An exploration in cir-ebon, indonesia. Journal of Tropical Medicine, 2013,507034. https://doi.org/10.1155/2013/507034Peters, R. M., Hofker, M. E., Van Brakel, W. H., Marjolein,Z. B. M., Seda, F. S. S. E., Irwanto, I., & Bunders, J. F. G.(2014). Narratives around concealment and agency for stigma-reduction: A study of Women affected by Leprosy in CirebonDistrict, Indonesia. Disability, CBR & Inclusive Development,25(4), 5–21. http://doi.org/10.5463/dcid.v25i4.389Peterson, N. A., & Reid, R. J. (2003). Paths to psychologicalempowermentinanurbancommunity:Senseofcommunity and citizen participation in substance abuseprevention activities. Journal of Community Psychology,31(1), 25–38. https://doi.org/10.1002/jcop.10034Peters, R. M., Zweekhorst, M. B., Van Brakel, W. H., Bunders,J. F., & Irwanto (2016). ‘People like me don’t make thingslike that’: Participatory video as a method for reducingleprosy-related stigma. Glob Public Health, 11(5–6),666–682. https://doi.org/10.1080/17441692.2016.1153122Positive Deviance Initiative. (2010). Basic field guide to thepositive deviance approach. Tufts University.Pujilestari, C. U., Ng, N., Hakimi, M., & Eriksson, M. (2014). Itis not possible for me to have diabetes”-community per-ceptions on diabetes and its risk factors in Rural PurworejoDistrict, Central Java, Indonesia. Global Journal of HealthScience, 6(5), 204–218. https://doi.org/10.5539/gjhs.v6n5p204Rai, S. S., Irwanto, I., Peters, R. M., Syurina, E. V., Putri, A. I.,Mikhakhanova, A., Naniche, D., & Zweekhorst, M. B.(2020). Qualitative exploration of experiences and conse-quences of health-related stigma among Indonesians withHIV, Leprosy, Schizophrenia and Diabetes. Kesmas: Na-tional Public Health Journal, 15(1), 7–16. http://dx.doi.org/10.21109/kesmas.v15i1.3306Rao, D., Elshafei, A., Nguyen, M., Hatzenbuehler, M. L., Frey,S., & Go, V. F. (2019). A systematic review of multi-levelstigma interventions: state of the science and future di-rections. BMC Medicine, 17(1), 41. https://doi.org/10.1186/s12916-018-1244-ySadan, E. (1997). Empowerment and community planning:Theory and practice of people-focused social solutions.Hakibbutz Hameuchad Publishers.Scambler, G. (2004). Re-framing stigma: Felt and enactedstigma and challenges to the sociology of chronic anddisabling conditions. Social Theory & Health, 2(1), 29–46.https://doi.org/10.1057/palgrave.sth.8700012Sercu, C., & Bracke, P. (2017). Stigma, social Structure, and thebiomedical framework: Exploring the stigma experiencesof inpatient service users in two Belgian psychiatric hos-pitals. Qualitative Health Research, 27(8), 1249–1261.https://doi.org/10.1177/1049732316648112Seyfang, G., & Smith, A. (2007). Grassroots innovations forsustainable development: Towards a new research andpolicy agenda. Environmental Politics, 16(4), 584–603.https://doi.org/10.1080/09644010701419121Shih, M. (2004). Positive stigma: Examining resilience andempowerment in overcoming stigma. The ANNALS of theAmerican Academy of Political and Social Science, 591(1),175–185. https://doi.org/10.1177/0002716203260099Sianturi, E. I., Perwitasari, D. A., Islam, M. A., & Taxis, K.(2019). The association between ethnicity, stigma, beliefsabout medicines and adherence in people living with HIVina rural area in Indonesia. BMC Public Health, 19(1), 55.https://doi.org/10.1186/s12889-019-6392-2Singhal, A., Shirley, S., & Marston, E. H. (2010). Turning diffusionof innovation paradigm on its head: The positive deviance12Qualitative Health Research 0(0)approach to social change. In A. Vishwanath & G. Barnett(Eds.), Advances in the study of the diffusion of innovations:Theory, methods, and application. Peter Lang Publishers.Singhal, A., & Svenkerud, P.J. (2019) Flipping the diffusion ofinnovations paradigm: Embracing the positive devianceapproach to social change. Asia Pacific Media Educator.https://doi.org/10.1177/1326365X19857010Stangl, A. L., Earnshaw, V. A., Logie, C. H., Van Brakel, W.,Simbayi, L. C, Barre, I., & Dovidio, J.F. (2019). The healthstigma and discrimination framework: A global, crosscuttingframework to inform research, intervention development,and policy on health-related stigmas. BMC Medicine, 17(1),31. https://doi.org/10.1186/s12916-019-1271-3Stuckey, H. L., Boan, J., Kraschnewski, J. L., Miller-Day, M.,Lehman, E. B., & Sciamanna, C. N. (2011). Using positivedeviance for determining successful weight-control prac-tices. Qualitative Health Research, 21(4), 563–579. https://doi.org/10.1177/1049732310386623Subu, M. A. (2015). Persistent Taboo understanding mentalillness and stigma among Indonesian adults throughgrounded theory. Universit´e d’Ottawa/University of Ottawa.Tsutsumi, A., Izutsu, T., Islam, A. M., Maksuda, A., Kato, H., &Wakai, S. (2007). The quality of life, mental health, andperceived stigma of leprosy patients in Bangladesh. SocialScience & Medicine, 64(12), 2443–2453. https://doi.org/10.1016/j.socscimed.2007.02.014UNAIDS(2017).Indonesia.http://www.unaids.org/en/regionscountries/countries/indonesia (Accessed 6 Febru-ary 2019).Van Brakel, W. H., Cataldo, J., Grover, S., Kohrt, B. A., Ny-blade, L., Stockton, M., Wouters, E., & Yang, L. H. (2019).Out of the silos: identifying cross-cutting features of health-related stigma to advance measurement and intervention.BMC Medicine, 17(1), 13. https://doi.org/10.1186/s12916-018-1245-xVan Dick, G., & Scheffel, R. (2015). Positive deviance. A lit-erature review about the relevance for health promotion.https://hdl.handle.net/1956/9282 (Accessed 10 February2019).Wallerstein, N., & Bernstein, E. (1994). Introduction to communityempowerment, participatory education, and health. SAGE.Weiss, M. G., Ramakrishna, J., & Somma, D. (2006). Health-related stigma: rethinking concepts and interventions.Psychology, Health & Medicine, 11(3), 277–287. https://doi.org/10.1080/13548500600595053WHO (2017). Global leprosy update, 2017: Reducing the diseaseburden due to leprosy. https://apps.who.int/iris/bitstream/handle/10665/274289/WER9335.pdf?ua=1 (Accessed 6February 2019).Woodgate, R. L., Comaskey, B., Tennent, P., Wener, P., &Altman, G. (2020). The wicked problem of stigma for youthliving with anxiety. Qualitative Health Research, 30(10),1491–1502. https://doi.org/10.1177/1049732320916460Zimmerman, M. A. (1995). Psychological empowerment: Issuesand illustrations. American Journal of Community Psychology,23(5), 581–599. https://doi.org/10.1007/BF02506983Author BiographicsSarju Sing Rai is a Global Health Professional who isworking as a Post-doctoral Researcher at the AthenaInstitute, Vrije Universiteit Amsterdam, Amsterdam, theNetherlands.Elena V. Syurina is an Assistant Professor at the AthenaInstitute, Vrije Universiteit Amsterdam, Amsterdam, theNetherlands.Ruth M.H. Peters is an Assistant Professor at the AthenaInstitute, Vrije Universiteit Amsterdam, Amsterdam, theNetherlands.Annisa Ika Putri is a Researcher at the Athena Institute,Vrije Universiteit Amsterdam, Amsterdam, the Netherlands.Irwanto Irwanto is a Professor at the Faculty of Psy-chology, Atma Jaya Catholic University, Jakarta, Indonesia.Marjolein B.M. Zweekhorst is a Professor at the AthenaInstitute, Vrije Universiteit Amsterdam, Amsterdam, theNetherlands.Rai et al.13 | RESEARCH GATE |
Citation: Li, G.; Li, L. An AnalysisMethod of Symplectic Dual Systemfor Decagonal Quasicrystal PlaneElasticity and Application. Crystals2022, 12, 636. https://doi.org/10.3390/cryst12050636Academic Editor: Shujun ZhangReceived: 4 April 2022Accepted: 25 April 2022Published: 28 April 2022Publisher’s Note: MDPI stays neutralwith regard to jurisdictional claims inpublished maps and institutional affil-iations.Copyright:© 2022 by the authors.Licensee MDPI, Basel, Switzerland.This article is an open access articledistributedunderthetermsandconditions of the Creative CommonsAttribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).crystalsArticleAn Analysis Method of Symplectic Dual System for DecagonalQuasicrystal Plane Elasticity and ApplicationGuangfang Li 1,2 and Lianhe Li 2,3,*1College of Science, Inner Mongolia Agricultural University, Hohhot 010018, China; liguangfang@126.com2Center for Applied Mathematics Inner Mongolia, Hohhot 010022, China3College of Mathematics Science, Inner Mongolia Normal University, Hohhot 010022, China*Correspondence: nmglilianhe@163.comAbstract: The symplectic solution system of decagonal quasicrystal elastic mechanics is considered.Hamiltonian dual equations together with the boundary conditions are investigated by utilizingthe principle of minimum potential energy. Then the symplectic eigenvectors are given on the basisof the variable separation method. As application, analytical solution for decagonal quasicrystalcantilever beam with concentrated load is discussed. The analytical expressions of the stresses anddisplacements of the phonon field and phason field are obtained. The present method allows for theexploration of new analytic solutions of quasicrystal elasticity that are difficult to obtain by otheranalytic methodsKeywords: quasicrystal; Hamiltonian system; eigenvector; analytical solution1. IntroductionIn 1984, Shechtman et al. initially discovered quasicrystals (QCs) [1]. QCs are non-periodic but ordered structural forms between crystals and glass. The elastic behaviorof QCs varies from that of ordinary crystals [2]. Experimental and theoretical studies onmechanical and physical properties of QCs have been encouraged [3]. The dislocation of astraight and moving screw in one-dimensional hexagonal QCs was studied by Fan et al. [4].The fracture mechanics problem of cubic QCs with a crack or an elliptical hole was solvedby Gao et al. [5]. The Stroh formalism has been applied successfully to the study of two-dimensional deformation of quasicrystal materials [6,7]. One dimensional hexagonal QCswith planar cracks are mentioned in reference [8]. The Stroh-like formalism for the bendingtheory of decagonal quasicrystal plates was developed by Li et al. [9]. Fundamental quanti-ties for the generalized elasticity and dislocation theory of QCs were provided by Lazar andAgiasofitou [10]. Ding et al. discussed two kinds of contact problems in three-dimensionalicosahedral QCs by a complex variable function method [11]. Lazar and Agiasofitou de-rived material balance laws for quasicrystalline materials with dislocations [12]. The elasticfield near the tip of an anticrack in a homogeneous decagonal quasicrystalline materialwas investigated by Wang et al. [13]. Guo et al. studied a mathematical model for nonlo-cal vibration and buckling of embedded two-dimensional decagonal quasicrystal layerednanoplates [14].The Hamiltonian system exists extensively and is universally applicable. The symplec-tic approach is the variable separation method which is actually based on the Hamiltoniansystem [15]. This method does not assume the trial function in advance, but introduces theproblem into the Hamiltonian system, and uses the variable separation method to solve thedifferential eigenvalue problem. Analytical solutions can then be obtained by the expan-sion of eigenfunctions. Many researchers use the symplectic method to study problemsin mechanics and engineering science since it is helpful for finding analytical solutions ofsome basic elasticity problems. The symplectic approach was first used in computationalsolid mechanics by Feng [16]. Zhong introduced the symplectic approach in analyticalCrystals 2022, 12, 636. https://doi.org/10.3390/cryst12050636https://www.mdpi.com/journal/crystalsCrystals 2022, 12, 6362 of 9solid mechanics [17]. In 2002, Zhong’s group originated the symplectic elasticity approachand developed it to form a systematic methodology [18]. The symplectic approach wasthen used successfully to research elasticity [19,20], piezoelectricity [21,22], functionallygraded effects [23] and differential equations [24–26], etc.However, the symplectic approach for quasicrystal elasticity has not been developedin a systematic way due to the complexity of QCs’ structure. In the present study, the basicequations of decagonal quasicrystals are first transferred to the Hamiltonian dual equationswith the help of the variation principle. By introducing dual variables, the dual systemis established directly, and a complete eigen-solution space is obtained. The solutions ofthe problem can be reduced to the zero-eigenvectors of the corresponding Hamiltonianoperator matrix and all their Jordan form eigenvectors. Then the analytical solution for theproblem is given by linear combination of these eigenvectors.2. Theoretical Formulation2.1. Basic EquationsSuppose that along the z direction, the decagonal quasicrystal is periodic, and quasi-periodic in the x-y plane.Based on the quasicrystal elasticity theory [2], the strain-displacement relations areεij = 12(∂ui∂xj+ ∂uj∂xi), wij = ∂wi∂xj.(1)If ignoring the body forces, the equilibrium equations can be written as∂σxx∂x + ∂σxy∂y = 0, ∂σyx∂x + ∂σyy∂y = 0,∂Hxx∂x + ∂Hxy∂y= 0, ∂Hyx∂x + ∂Hyy∂y= 0.(2)and the constitutive equations are as followsσxx = C11εxx + C12εyy + R(wxx + wyy),σyy = C12εxx + C11εyy − R(wxx + wyy),σxy = σyx = 2C66εxy + R(wyx − wxy),Hxx = K1wxx + K2wyy + R(εxx − εyy),Hyy = K1wyy + K2wxx + R(εxx − εyy),Hxy = K1wxy − K2wyx − 2Rεxy,Hyx = K1wyx − K2wxy + 2Rεxy,(3)in which C66 = (C11 − C12)/2. Here, σij(σij = σji) , ui and εij(εij = εji), respectively, repre-sent the stress, displacement and strain of the phonon field. Hij , wi and wij, respectively,represent the stress, displacement and strain of the phason field. Cij and Ki are the elasticconstants of the phonon field and the phason field, respectively. R represents the couplingelastic constant of the phonon-phason field.2.2. Variational Principle and Hamiltonian Dual EquationConsider a decagonal quasicrystal rectangular domain with length l and width 2hin the x-y coordinates (0 ≤ x ≤ l and −h ≤ y ≤ h). We use a dot to represent thedifferentiation with respect to x, i.e., (.) = ∂/∂x.Crystals 2022, 12, 6363 of 9The strain energy density U(ux, uy, wx, wy,.ux,.uy,.wx,.wy) can be expressed asU�ux, uy, wx, wy,.ux,.uy,.wx,.wy�= 12�σxxεxx + σyyεyy + 2σxyεxy + Hxxwxx + Hyywyy + Hxywxy + Hyxwyx�= 12�C11�.u2x +� ∂uy∂y�2�+ C66� ∂ux∂y +.uy�2+ K1�.w2x +� ∂wx∂y�2+.w2y +� ∂wy∂y�2��+C12.ux∂uy∂y + R��.ux − ∂uy∂y��.wx + ∂wy∂y�+� ∂ux∂y +.uy��.wy − ∂wx∂y��+K2�.wx∂wy∂y − ∂wx∂y.wy�.(4)The principle of minimum potential energy of the problem can be represented asδΠ = δ� l0� h−h L(ux, uy, wx, wy,.ux,.uy,.wx,.wy)dydx = 0(5)where L(ux, uy, wx, wy,.ux,.uy,.wx,.wy) is the Lagrange density function. If the body forcesare neglected, we haveL(ux, uy, wx, wy,.ux,.uy,.wx,.wy) = U(ux, uy, wx, wy,.ux,.uy,.wx,.wy)(6)From the partial integration of Equation (5) in y direction, we can obtain�C12εxx + C11εyy − R(wxx + wyy)�y=±h = σyy��y=±h,�2C66εxy + R(wyx − wxy)�y=±h = σxy��y=±h,�K1wyy + K2wxx + R(εxx − εyy)�y=±h = Hyy��y=±h,�K1wxy − K2wyx − 2Rεxy�y=±h = Hxy��y=±h.(7)Thus, the homogeneous boundary conditions areσyy = 0, σxy = 0, Hyy = 0, Hxy = 0, at y = ±h.(8)Let the displacement vector beq = (ux, uy, wx, wy)T(9)where the superscript T represents the transpose. Then the Lagrange density function canbe written as L(q,.q), from which the dual variable p required by Hamiltonian form isderived asp = ∂L∂.q =C11.ux + C12∂uy∂y + R(.wx + ∂wy∂y )C66( ∂ux∂y +.uy) + R(.wy − ∂wx∂y )K1.wx + K2∂wy∂y + R(.ux − ∂uy∂y )K1.wy − K2 ∂wx∂y + R( ∂ux∂y +.uy)= (σxx, σyx, Hxx, Hyx)T(10)By Equations (2), (3) and (10), the Hamiltonian dual equation of the problem can beobtained as.v = Hv(11)Crystals 2022, 12, 6364 of 9where H =� ABC−AT�,A =0a1 ∂∂y0a2 ∂∂y− ∂∂y0a3 ∂∂y00a4 ∂∂y0a5 ∂∂y00a6 ∂∂y0,B =b10b200b30b4b20b500b40b6,C =00000a7 ∂2∂y20a8 ∂2∂y200a9 ∂2∂y200a8 ∂2∂y20a10 ∂2∂y2, and v = (ux, uy, wx, wy, σxx, σyx, Hxx, Hyx)T. AT is theadjoint operator matrix of A. ai(i = 1, 2, · · · , 10) and bi(i = 1, 2, · · · , 6) in the matricesare the elastic constants (see Appendix A). H satisfies JHJ = HT is a Hamiltonian op-erator matrix [18], where J =�0I4−I40�in which I4 is the 4 × 4 unit matrix. FromEquations (3) and (11), Equation (8) can be written asσyy = −a7∂uy∂y − a8∂wy∂y − a1σxx − a4Hxx = 0 ,σxy = 0 ,Hyy = −a8∂uy∂y − a10∂wy∂y − a2σxx − a5Hxx = 0 ,Hxy = −a9 ∂wx∂y − a3σyx − a6Hyx = 0 .(12)at y = ±h.2.3. Symplectic Analysis and EigenvectorsLetv(x, y) = X(x)Y(y)(13)Substituting Equation (13) into Equation (11), we obtain.X(x) = µX(x)(14)and the eigenvalue equationHY(y) = µY(y)(15)where µ and Y(y) are the eigenvalue and the corresponding eigenvector, respectively. Zeroeigenvalue of the problem exists because the fact that boundaries at both sides (y = ±h) arefree. There are Jordan form eigen-solutions of different orders for the eigen-solutions of zeroeigenvalue. The solution to the problem can be expressed by Jordan form eigen-solutions.Now the problem is to solve the zero-eigenvalue equation.The eigenvalue equation isHY(y) = 0(16)when µ = 0. Solving Equation (16) yields the eigen-solutions of zero eigenvalueY(0)1= (1, 0, 0, 0, 0, 0, 0, 0)T,(17)Y(0)2= (0, 1, 0, 0, 0, 0, 0, 0)T,(18)Y(0)3= (0, 0, 1, 0, 0, 0, 0, 0)T,(19)Y(0)4= (0, 0, 0, 1, 0, 0, 0, 0)T.(20)These eigenvectors are the solutions of the original Equation (11) with boundaryconditions (8). Letv1 = Y(0)1 , v2 = Y(0)2 , v3 = Y(0)3 , v4 = Y(0)4 .(21)Crystals 2022, 12, 6365 of 9Next, solve the Jordan form eigen-solutions of zero eigenvalue.2.3.1. The First-Order Jordan Form Eigen-SolutionsThe governing equations for finding the first-order eigen-solutions of Jordan form areHY(1)i= Y(0)i(i = 1, 2, 3, 4)(22)The solutions areY(1)1= (0, a1y, 0, −a4y, −a7, 0, a8, 0)T,(23)Y(1)2= (−y, 0, 0, 0, 0, 0, 0, 0)T,(24)Y(1)3= (0, −a2y, 0, a5y, a8, 0, −a10, 0)T,(25)Y(1)4= (−a3y, 0, a6y, 0, 0, 0, 0, −a9)T.(26)These eigen-solutions are not directly the solutions of the original problem. Thefirst-order Jordan form solutions of the original problem arev5 = Y(1)1+ xY(0)1 ,(27)v6 = Y(1)2+ xY(0)2 ,(28)v7 = Y(1)3+ xY(0)3 ,(29)v8 = Y(1)4+ xY(0)4 .(30)2.3.2. The SecondOrder Jordan Form Eigen-SolutionsConsider equationsHY(2)i= Y(1)i(i = 1, 2, 3, 4)(31)When i = 2, the solution of Equation (31) isY(2)2=�0, −12a1y2, 0, 12a4y2, a7y, 0, −a8y, 0�T(32)When i = 1,3,4, no solutions of Equation (31) exist due to the fact that solutionscannot satisfy the boundary conditions (8) at the same time. Hence, these chain of Jordanform eigen-solutions are terminated. The 2nd-order Jordan form solution of the originalproblem isv9 = Y(2)2+ xY(1)2+ 12x2Y(0)2(33)2.3.3. The Third-Order Jordan Form Eigen-SolutionsSolving equationHY(3)2= Y(2)2(34)gives the eigen-solutionY(3)2=�a11h2y + a12y3, 0, a13h2y + 12a4y3, 0, 0, 12a7(h2 − y2), 0, a14h2 + a15y2�T(35)in which the constants ai(i = 11, 12, · · · , 15) can be found in Appendix A. The 3rd-orderJordan form solution of the original problem can be composed in the same wayv10 = Y(3)2+ xY(2)2+ 12x2Y(1)2+ 16x3Y(0)2(36)Crystals 2022, 12, 6366 of 9It can be proven that there are no other high-order Jordan form solutions.Up to here, the expressions of the general solution of Equation (11) can be written asv =10∑i=1mivi=m1 + m5x − (m6 + a3m8)y − m9xy + m10�a11h2y − 12x2y + a12y3�m2 + m6x + (a1m5 − a2m7)y + m9�12x2 − a12 y2�+ m10�16x3 − a12 xy2�m3 + m7x + a6m8y + m10�a13h2y + a42 y3�m4 + m8x − (a4m5 − a5m7)y + a42 m9y2 + a42 m10xy2−a7m5 + a8m7 + a7m9y + a7m10xy12a7m10�h2 − y2�a8m5 − a10m7 − a8m9y − a8m10xy−a9m8 + m10�a14h2 + a15y2�(37)The constants mi(i = 1, 2, · · · , 10) in Equation (37) can be determined according tothe specific problem and boundary conditions. Then the stresses and displacements of theproblem can be obtained.3. Bending of Decagonal Quasicrystal Cantilever Beam with Concentrated LoadAs an application of the symplectic approach for quasicrystal elasticity, an analyticalsolution for bending of a decagonal quasicrystal cantilever beam is discussed. Consider adecagonal quasicrystal cantilever beam with length l, width 2h and thickness b, which isunder concentrated load P at the free end as pictured in Figure 1.Figure 1. Decagonal quasicrystal cantilever beam.The boundary conditions can be expressed asσyy = σyx = Hyy = Hyx = 0, for y = ±h,σxx = Hxx = 0, b� h−h σyxdy = −P, for x = 0,(38)ux = uy = wx = wy = 0, ∂uy∂x = 0, for y = 0 and x = l.Substituting Equation (37) into the displacement and stress boundary conditions inEquation (38) respectively, we havem2 = −l3P2a7bh3 , m4 = − 3b4lP8bh , m6 =3l2P4a7bh3 ,m1 = m3 = 0 .(39)andm8 = 3b4P8bh , m10 = −3P2a7bh3 ,m5 = m7 = m9 = 0 .(40)Crystals 2022, 12, 6367 of 9Thus, the phonon and phason stresses are obtained asσxx = − 3P2bh3 xy , σyy = 0 , σxy = σyx =3P4bh3 (y2 − h2) ,Hxx =3a8P2a7bh3 xy , Hyy = 0 ,Hxy = 3a3P8bh3 (y2 − h2) , Hyx = − 3a15P2a7bh3 (y2 − h2) .(41)The displacements are obtained asux =3P4a7bh3 x2y − 3a12P2a7bh3 y3 −3P4a7bh3 l2y − 3b3P4bh y ,uy =3a1P4a7bh3 xy2 −P4a7bh3 x3 +3P4a7bh3 l2x −2P4a7bh3 l3 ,wx = − 3b4P8bh3 y3 + 9b4P8bh y ,wy = − 3b4P8bh3 xy2 + 3b4P8bh x − 3b4P8bh l .(42)Equation (41) shows that the expressions of the phonon stresses are exactly same asthe stresses of the well-known classical elasticity theory [27].The coupling constant R has great influence on the mechanical behaviors of QCs,and it has not been measured yet. Next, we consider the influence of the coupling elasticconstant of the phonon-phason field on the displacements of phonon field and phason field.The phonon and phason elastic constants of the two-dimensional decagonal quasicrystalsare C11 = 234.33GPa, C12 = 57.41GPa, K1 = 122GPa and K2 = 24GPa [28]. Take otherparameters as P = 200KN, l = 1m, h = 0.08m and b = 0.1m.ux = ux/�10−3m�,uy = uy/�10−3m�,wx = wx/�10−4m�and wy = wy/�10−4m�are normalized displacements of phonon field and phason field, respectively. Figures 2and 3 show that the displacements of phonon field and phason field both increase withthe increase of the coupling constant R. The phonon field displacement is one order ofmagnitude larger than the phase field displacement.Figure 2. Normalized displacements of phonon field versus y at x = 0.2.Crystals 2022, 12, 6368 of 9Figure 3. Normalized displacements of phason field versus y at x = 0.2.4. ConclusionsThe unified framework of the symplectic approach for quasicrystal elasticity prob-lems is established. The problem is reduced to the zero eigenvalues with their Jordanforms, which are important in applications. Through working out an eigen-problem of thesymplectic dual system, the solution of the Hamiltonian dual equation is obtained. Theanalytical solutions are obtained in a rigorous step-by-step manner, which is fundamentallydifferent from the classical semi-inverse method with pre-determined trial functions. Thestress and displacement can be calculated together. The symplectic approach is effectiveand provides a new channel for the research of quasicrystal elasticity theory.Author Contributions: Writing—original draft and editing, G.L.; Resources, Methodology andwriting—review, L.L. All authors have read and agreed to the published version of the manuscript.Funding: This research was funded by the National Natural Science Foundation of China (Nos.11962026, 12002175, 12162027 and 62161045), Natural Science Foundation of Inner Mongolia (Nos.2020MS01018, 2021MS01013), Scientific Research Project in Colleges and Universities of Inner Mongo-lia (NJZY22519) and Foundation of Basic Scientific Research Project of Inner Mongolia AgriculturalUniversity (JC2020002).Data Availability Statement: Not applicable.Conflicts of Interest: The authors declare no conflict of interest.Appendix Aa1 = − C12K1+R2C11K1−R2 , a2 = R(K2−K1)C11K1−R2 , a3 = R(K1−K2)C66K1−R2 ,a4 = R(C11+C12)C11K1−R2 , a5 = − C11K2−R2C11K1−R2 , a6 = C66K2−R2C66K1−R2 ,a7 = 2a4b4 , a8 = − a3a4b4 , a9 = a3(1+a6)b4, a10 = a2(a5−1)b2,a11 = −(C11+C12)(2C66K21−(K1+K2)R2)2(C11K1−R2)(C66K1−R2), a12 = 2C11K1+C12K1−R26(C11K1−R2),a13 = (C11+C12)R(C66(−3K1+K2)+2R2)2(C11K1−R2)(C66K1−R2),a14 = − (C11+C12)(K1−K2)R(2C66K1−C66K2−R2)2(C11K1−R2)(C66K1−R2),a15 = 3(C11+C12)(K1−K2)R2(C11K1−R2),b1 =K1C11K1−R2 , b2 = −RC11K1−R2 , b3 =K1C66K1−R2 ,b4 = −RC66K1−R2 , b5 =C11C11K1−R2 , b6 =C66C66K1−R2 .Crystals 2022, 12, 6369 of 9References1.Shechtman, D.; Blech, I.; Gratias, D.; Cahn, J.W. Metallic phase with long-range orientational order and no translational symmetry.Phys. Rev. Lett. 1984, 53, 1951–1953. [CrossRef]2.Hu, C.Z.; Wang, R.H.; Ding, D.H. Symmetry groups, physical property tensors, elasticity and dislocations in quasicrystals. Rep.Prog. Phys. 2000, 63, 1–39. [CrossRef]3.Suck, J.B.; Schreiber, M.; Häussler, P. Quasicrystals: An Introduction to Structure, Physical Properties and Applications; Springer:Berlin/Heidelberg, Germany, 2002.4.Li, X.F.; Fan, T.Y. Elastic analysis of a mode II crack in a decagonal quasi-crystal. Chin. Phys. B 2002, 11, 266–271.5.Gao, Y.; Ricoeur, A.; Zhang, L.L. Plane problems of cubic quasicrystal media with an elliptic hole or a crack. Phys. Lett. A 2011,375, 2775–2781. [CrossRef]6.Radi, E.; Mariano, P.M. Stationary straight cracks in quasicrystals. Int. J. Fract. 2010, 166, 105–120. [CrossRef]7.Yang, L.Z.; Gao, Y.; Pan, E.; Waksmanski, N. An exact solution for a multilayered two-dimensional decagonal quasicrystal plate.Int. J. Solids Struct. 2014, 51, 1737–1749. [CrossRef]8.Li, X.Y. Elastic field in an infinite medium of one-dimensional hexagonal quasicrystal with a planar crack. Int. J. Solids Struct.2014, 51, 1442–1455. [CrossRef]9.Li, L.H.; Liu, T.G. Decagonal quasicrystal plate with elliptic holes subjected to out-of-plane bending moments. Phys. Lett. A 2014,378, 839–844. [CrossRef]10.Lazar, M.; Agiasofitou, E. Fundamentals generalized elasticity and dislocation theory of quasicrystals: Green tenser, dislocationkey-formulas ang dislocation loops. Philos. Mag. 2014, 94, 4080–4101. [CrossRef]11.Zhao, X.F.; Li, X.; Ding, S.H. Two kinds of contact problems in three-dimensional icosahedral quasicrystals. Appl. Math. Mech.2015, 36, 1569–1580. [CrossRef]12.Lazar, M.; Agiasofitou, E. Eshelbian mechanics of novel materials: Quasicrystals. J. Micromech. Mol. Phys. 2016, 1, 1640008.[CrossRef]13.Wang, X.; Schiavone, P. Elastic field near the tip of an anticrack in a decagonal quasicrystalline material. Appl. Math. Mech. 2020,41, 401–408. [CrossRef]14.Sun, T.Y.; Guo, J.H.; Pan, E. Nonlocal vibration and buckling of two-dimensional layered quasicrystal nanoplates embedded in anelastic medium. Appl. Math. Mech. 2021, 42, 1077–1094. [CrossRef]15.Zhong, W.X. A New Systematic Methodology for Theory of Elasticity; Dalian University of Technology Press: Dalian, China, 1995. (InChinese)16.Feng, K. On difference scheme and symplectic geometry. In Proceedings of the Beijing Symposium Differential Geometry and DifferentialEquations; Science Press: Beijing, China, 1985; pp. 42–58.17.Zhong, W.X. Plane elasticity problem in strip domain and Hamiltonian system. J. Dalian Univ. Technol. 1991, 31, 373–384.(In Chinese)18.Yao, W.A.; Zhong, W.X. Symplectic Elasticity; Higher Education Press: Beijing, China, 2002. (In Chinese)19.Leung, A.Y.T.; Xu, X.S.; Zhou, Z.H.; Wu, Y.F. Analytic stress intensity factors for finite elastic disk using symplectic expansion.Eng. Fract. Mech. 2009, 76, 1866–1882. [CrossRef]20.Liu, Y.M.; Li, R. Accurate bending analysis of rectangular plates with two adjacent edges free and the others clamped or simplysupported based on new symplectic approach. Appl. Math. Model. 2010, 34, 856–865. [CrossRef]21.Leung, A.Y.T.; Zheng, J.J.; Lim, C.W.; Zhang, X.C.; Xu, X.S.; Gu, Q. A new symplectic approach for piezoelectric cantilevercomposite plates. Comput. Struct. 2008, 86, 1865–1874. [CrossRef]22.Xu, X.S.; Gu, Q.; Leung, A.Y.T.; Zheng, J.J. A symplectic eigensolution method in transversely isotropic piezoelectric cylindricalmedia. J. Zhejiang Univ.-Sci. A 2005, 6, 922–927.23.Zhao, L.; Chen, W.Q. Symplectic analysis of plane problems of functionally graded piezoelectric materials. Mech. Mater. 2009, 41,1330–1339. [CrossRef]24.Jiménez, F.; Ober-Blöbaum, S. Fractional Damping through Restricted Calculus of Variations. J. Nonlinear Sci. 2021, 31, 46.[CrossRef]25.Xu, C.H.; Leng, S.; Zhou, Z.H.; Xu, X.S.; Deng, Z.C. Accurate and straightforward symplectic approach for fracture analysis offractional viscoelastic media. Appl. Math. Mech. 2022, 43, 403–416. [CrossRef]26.Koyunbakan, H. The Transmutation Method and Schrödinger Equation with Perturbed Exactly Solvable Potential. J. Comput.Acoust. 2009, 17, 1–10. [CrossRef]27.Timoshenko, S.P.; Goodier, J.N. Theory of EIasticity; Higher Education Press: Beijing, China, 2013. (In Chinese)28.Fan, T.Y. Mathematical theory and methods of mechanics of quasicrystalline materials. Engineering 2013, 5, 407–448. [CrossRef] | RESEARCH GATE |
23Analysis of femicide cases in Campinas, SP, Brazil, from 2018 to 2019 through the ecological model of violenceAbstract Lethal violence against women is a complex and multidimensional phenomenon in which a wide number of factors intersect and converge to make a femicide happen at a specific time and place. The main factors that contrib-uted to the occurrence of femicides in the city of Campinas were identified from January 2018 to December 2019. Interviews were conducted with family members, friends, neighbors, witnesses, and health agents about 24 femicides using the verbal autopsy technique. The autopsies were supplemented, when possible, with information from the media and clinical autopsy reports. For the data analysis process, narratives of the cases were carried out, recovering the most important aspects of the verbal autopsies and organizing the factors found in the four levels of the ecological model of violence used by the World Health Or-ganization: individual, relational, community, and social. The analysis was structured in cate-gories following a deductive approach. Starting from particular cases delimited in time (2018 and 2019) and in space (municipality of Campinas) it is expected to understand the phenomenon of femicide in its broadest dimension.Key words Gender-based violence, Femicide, Gender analysis in health, Gender perspective Mônica Caicedo-Roa (https://orcid.org/0000-0001-7313-8260) 1 Ricardo Carlos Cordeiro (https://orcid.org/0000-0002-0437-1066) 2DOI: 10.1590/1413-81232023281.09612022EN1 Programa de Pós-Graduação em Saúde Coletiva, Faculdade de Ciências Médicas, Universidade Estadual de Campinas. R. Tessália Vieira de Camargo 126, Cidade Universitária Zeferino Vaz. 13083-887 Campinas SP Brasil. monicacaicedoroa@gmail.com 2 Departamento de Saúde Coletiva, Faculdade de Ciências Médicas, Universidade Estadual de Campinas. Campinas SP Brasil. ArtiCle24Caicedo-Roa M, Cordeiro RCintroductionFemicides are multicausal, frequent, and com-plex events. Every day, an average of 137 deaths of women are accounted for by a member of their family worldwide1. Most cases occur in domes-tic spaces and are the responsibility of intimate partners or male relatives2. Brazil is considered a country with high levels of violence against wom-en. In 2019, an average of 13 femicides were re-ported daily3. During the last five years, the coun-try has registered an increase in femicide figures, particularly during the COVID-19 pandemic4. Multiple variables participate in femicide, and not all have been fully explored. Conceptual models help understand the phenomenon of le-thal violence against women. The World Health Organization (WHO) implemented the ecolog-ical model of violence in the World Report on Violence and Health5,6. The model, introduced in the 1970s, has been applied to the analysis of child abuse, youth violence, intimate partner violence, and abuse of older adults. The model explores the relationship between individual and contextual factors considering violence as the re-sult of several layers of influence on behavior at four levels of analysis5:Individual level: refers to the characteristics of the person that increase the possibility of being a victim or perpetrator of violence. It includes biological, historical, demographic, personal, and behavioral factors (personality, low school performance, substance abuse, and history of a violent relationship, among others). Relational level: encompasses close social re-lationships (partners, intimate partners, and fam-ily members) that increase the risk of violence. Community level: integrates the contexts in which relationships are embedded (schools, workplaces, and neighborhoods) and identifies characteristics of these scenarios. Social level: considers the broader factors that influence a climate favorable to violence or reduce inhibitions against it, such as norms that support violence to resolve conflicts. It also in-cludes health, educational, economic, and social policies that keep levels of inequality high. The ecological framework considers the dif-ferent causes of violence and the interaction of family, community, and social factors, which are also influenced by cultural and economic values. Notably, this model cannot contemplate all the elements of the reality it intends to mod-el. It is an analytical strategy for understanding the complexity of the phenomenon of lethal vi-olence against women. The choice of this model responds to recognizing the participation of a broad set of factors that intersect and converge so that femicide occurs in a specific temporal and spatial context. This study aims to analyze, through the eco-logical model of violence, the cases of femicide that occurred in 2018 and 2019 in the municipal-ity of Campinas, SP, Brazil. MethodThe study was conducted in the municipality of Campinas, with approximately 1.2 million in-habitants, located approximately 100 kilometers northwest of the municipality of São Paulo. Case research enables the analysis of single events to understand the overall context through a deduc-tive approach. It is understood that an individual can be seen as a manifestation of a totality built from social relationships in a specific historical context7. Through a partnership with the munic-ipal Department of Health, death certificates and addresses of all residents who died by homicide (n = 38) from January 1, 2018, to December 31, 2019, were obtained. Among the cases, by con-sensus of the field group, those where the com-ponent of gender inequality characterizing femi-cides was identified were selected.As this study analyzed cases of femicide that had already occurred, the data were obtained by performing verbal autopsies with family mem-bers, friends, acquaintances, and neighbors in the areas where the deceased women lived, inter-acted, and worked8. A verbal autopsy is an indi-rect method for estimating a cause of mortality. It collects, codes, and systematizes mortality infor-mation in epidemiological and public health sur-veillance studies8. The verbal autopsy’s objective was to complement the information and expand the understanding of the characteristics of deaths by femicide. Previous studies on femicides used this technique9,10.The information was collected through field-work, moving the research team to the living, working, and interaction surroundings of the de-ceased women, which implied an approximation of the living conditions of the researched subjects. Autopsies were performed 15 days after death to respect the mourning period of family members and acquaintances. The cases’ more personal and subjective character was not captured; instead, an attempt was made to approximate the victim’s liv-ing conditions as closely as possible.25Ciência & Saúde Coletiva, 28(1):23-36, 2023Using verbal autopsies, information from the media, and clinical autopsies as a reference, nar-ratives were elaborated, retrieving the most sig-nificant elements of the victims’ lives and deaths for each level of the ecological model. Subse-quently, they were organized by frequencies and categories. The Research Ethics Committee of the School of Medical Sciences of the Universi-ty of Campinas (Unicamp) approved the study under CAAE: 04005118.9.0000.5404 and 29654720.6.0000.5404. Consent forms were pro-vided to participants. The names used are ficti-tious, and any information that could reveal the identity of the victim or perpetrator of the femi-cide has been removed. resultsIn total, 24 cases of women living in Campinas who died by femicide in 2018 and 2019 were an-alyzed. Chart 1 describes the sociodemographic characteristics of the victims, the probable trig-ger, and the mechanism used to cause the death. Chart 2 shows the elements and categories com-posing each level. These were schematically orga-nized in Figure 1. In this analysis, the relational level proved to be the most relevant for femicide, followed by the individual, social, and commu-nity levels.The death of a woman is not an isolated event. Due to the relevance of their social role, death considerably affects the lives of people who depend on them. In the 24 cases of femicide analyzed, in addition to the deaths of women, three deaths of abusers occurred. They commit-ted suicide after the femicide or died during the aggression. Furthermore, 45 children lost their mothers. Discussion The ecological model of violence makes it pos-sible to identify the relevant factors for femicide on a case-by-case basis. Some factors may be cross-sectional, appear at more than one level, and have greater or lesser importance depending on the case. The factors intersect to increase the risk of femicide. Violence against women must be understood as a macro phenomenon inserted in a global, community, and relational context and ends in individual death. The main elements of each level are discussed below. relational levelThis level is crucial because violent relation-ships and inequality of power between genders can be located there. The relational level inte-grates two individualities: the woman and the perpetrator of the femicide. Only in their inter-action does the relational level emerge. This level is formed differently from the simple junction of individual levels and operates with its dynamics. The most important categories at this level were: Violent relationshipWorldwide, 30% (95%CI 27.8-32.2) of wom-en in a relationship reported having experienced physical and/or sexual violence by an intimate partner throughout their lives11,12, with a lifetime prevalence of violence ranging 15-71%13. Brazil-ian studies reported prevalence values from 27.4 to 57.6%14,15. Women who experience domestic violence experience multiple episodes and vari-ous types of abuse.Most women analyzed in this study had a re-lationship with a violent man (79%), supporting the very high prevalence of exposure to violence among victims of femicide16. Violence is a re-source partners use to ensure obedience, main-tain control, and favor the continuity of relation-ships. It usually starts in low grades and increases in severity and frequency. Violence behaves cy-clically17: periods of apparent calm and well-be-ing alternate with stressful situations that trigger aggression. In the repetition of the cycle of vio-lence, the aggressions become more substantial and can end in femicide. A study in 10 countries, including Brazil, showed that when a woman has been a victim of violence inflicted by her partner, it is very likely that an act of serious violence will subsequent-ly occur. Most violent acts are not isolated inci-dents, following a pattern of continued abuse13. Studies with women victims of attempted femi-cide estimated that 67% had a history of abuse before the attempted murder18. History of an abusive relationshipThe history of a violent relationship appeared in 25% of the cases. Previous theories suggest-ed that women victims of violence tend to get involved in new violent relationships. Thus, the history of a violent relationship is a risk factor for femicide16. 26Caicedo-Roa M, Cordeiro RCShort relationshipThe known pattern of violent relationships describes that relationships become increasingly violent until they reach femicide19. However, four of the femicide cases studied contradict the pat-tern. In Valentina’s case, the relationship lasted four months, Valeriana and Orquídea’s grand-daughter three months, and in Carina’s case, it was only a month. The fact that these femicides took place in such short-term relationships is a matter of concern because it reduces the possi-bility of women seeking support and ending the relationship safely, demanding greater agility of institutions to protect the victims’ lives. relationship with a married manTwo cases of femicide occurred as a result of relationships with married men. Manifestations of intense violence were seen before the discov-ery of male infidelity. Maintaining a relationship Chart 1. Sociodemographic characteristics of victims, motivation, and mechanisms of femicide cases, Campinas 2018-2019.Sociodemographic characteristics of femicide casesProbable motivation/trigger and mechanisms used to cause deathAlana: 38-year-old, White, with graduate studiesThe partner was jealous for suspecting that she was in another relationship. She was shot 16 times.Antonieta: 49-year-old, Brown, with a college degree, married, and a machine operator in a fishing net companyRape. She suffered traumatic brain injury with a pipe bar.Carina: 25-year-old, White, single, highway service attendantShe disagreed with her boyfriend. She suffered stab wounds to the chest, jaw, eyebrow, and lip, and lethal wounds to her lung and heart (duration of relationship: one month).Dorelia: 75-year-old, Brown, uneducated, married, and bartender in a neighborhood barShe had a disagreement over money with her son, who beat and pushed her, causing TBI.Eva: 41-year-old, White, self-employedDisagreement over the end of the relationship, he assaulted her with a stab wound to the right side of the chest so that she would learn not to challenge him. Flávia: 40-year-old, Black, married, and recycling pickerCouple’s disagreement. She was beaten and burned with gasoline.Franciele: 34-year-old, White, with a college degree, married, and an administrative employee at a grocery storeThe partner did not want to accept the end of the relationship and was jealous of a supposed new partner. She suffered strangulation.Jeannete: 19-year-old adolescent, Brown, with incomplete elementary education, single, linked to drug traffickingShe had links and debts with the drug trade. She suffered 31 punctures to the chest and abdomen, and signs of sexual assault.Majô: 29-year-old, Black, marriedRape. She suffered mechanical asphyxia and strangulation.Marcela: 13-year-old adolescent, Brown, with incomplete elementary education, in a common-law marriage, and buffet service attendantCouple’s disagreement. He did not want her to leave the house. She suffered a gunshot wound to the thigh with massive blood loss.Martina: 26-year-old, White, with incomplete higher education, single, and student in the health areaThe trigger of aggression is unknown. She suffered strangulation and asphyxiation from her boyfriend, who used a cell phone cord.Martha: 32-year-old, White, with incomplete elementary education, single, and a waitressRelationship termination and suspected of having relationships with other men. She suffered a beating and assault with a knife in the cervical region.Miranda: 40-year-old, Brown, with incomplete elementary education, married, and service attendant at an ice cream shopRape. She suffered stab wounds and was later beheaded.Nicolaza: 50-year-old, White, with incomplete higher education, single, and cleaning ladyShe was suspected of being pregnant, and the abuser did not want his wife to know about the infidelity. She suffered hanging and a blow to the head.it continues27Ciência & Saúde Coletiva, 28(1):23-36, 2023with a married man can also be a factor in fem-icide, especially when pregnancy is suspected or confirmed. Behavioral issues of the abuser Death threat Being threatened with death is a risk factor for femicide (OR = 7.36; 95%CI 2.99-18.11)20. Rating scales identified this threat as a sign of risk20-23. Six women in the research were threatened with death, and two filed a police report, but the restraining order was offered in none of the cases. In Ramona’s case, she and her family were threatened with death if they report-ed the attacks. Problematic consumption of psychoactive substances (PAS) by the abuser Studies on the pattern of alcohol and drug use in murders and attempted murders of women by their intimate partners have shown a strong relationship between substance use and violence24. A systematic review found that sub-stance abuse by the abuser, including alcohol and other drugs, increases the risk of femicide by 85% (OR = 1.85; 95%CI 1.19-2.86)20. Alcohol acts by disinhibiting behavior. People under the influence of psychoactive substances may show aggressive behavior. Sociodemographic characteristics of femicide casesProbable motivation/trigger and mechanisms used to cause deathNeusa: 54-year-old, White, with incomplete high school education, single, and traderRelationship termination. She suffered a gasoline burn on 95% of the body surface.Olivia: 64-year-old, White, separated, and retiredShe refused to hand over the pension money to her grandson. She suffered a stabbing wound, with a neck cut.Orquídea: 82-year-old, White, with incomplete elementary education, widowed, and retiredShe suffered multiple blows with a knife, fork, and pruning shears from an abuser who mistook her for the woman (victim’s granddaughter) with whom he had just ended a relationship (duration of relationship: three months).Ramona: 31-year-old, White, with incomplete high school education, married, and service attendant at a post officeShe was in a relationship with a highly violent partner who regularly physically abused her. She was beaten with a stick and kicked in the head.Samanta: 28-year-old, Brown, with incomplete higher education, separated, and an attendant at a liquor storeRelationship termination. She was hanged. There was a suicide simulation.Sílvia: 43-year-old, Black, with incomplete high school education, single, cleaning lady, and sex workerSuspected of being HIV+. She suffered hanging, beating, and body burn.Thaisa: 24-year-old, White, single, and a sex workerShe had a child with the abuser, and he did not want his wife to know about the infidelity. She was beaten, and her body was buried for approximately one month.Taciana: 21-year-old, Black, single, and treasurer at a grocery storeRelationship termination. She suffered 15 bullet impacts.Valentina: 31-year-old, Brown, with incomplete higher education, single, and service attendant at a car washDisagreement for money with her partner. She suffered cervical, thoracic, and abdominal stab wounds (duration of relationship: four months).Valeriana: 32-year-old, Brown, single, and cookShe had sexual intercourse with other partners. She suffered a beating and assault with a knife in the neck from her partner (duration of relationship: three months).Source: Authors.Quadro 1. Características sociodemográficas das vítimas, motivação e mecanismos dos casos de feminicídios, Campinas 2018-2019.28Caicedo-Roa M, Cordeiro RC Controlling behavior of the partner Controlling behavior includes a wide variety of restrictive attitudes, such as: Limit the places that the woman frequents and the time they stay in them, such as healthcare services or visits to family/friends; limit access to social networks; limit or prevent activities such as working or studying; decide on the clothes, make-up, food, and drink she can wear or ingest; send controlling messages by cell phone or email insis-tently; restrict access to money, bank account, or other property owned by the woman or couple.Chart 2. Elements and categories of WHO ecological model levels present in femicide cases, Campinas 2018-2019.level (n)Category (n)level elements (n)Relational (97)Violent relationship (37)Relationship with a violent man (19). History of an abusive relationship (6). Short relationship (4). Relationship with a married man (2). Multiple abusers (2). Others (4).Behavioral issues of the abuser (25)Be threatened with death (6). Problematic consumption of PAS by the abuser (4). Controlling behavior of the partner (4). Threat or attempted suicide by the abuser (3). Previous disagreement with the abuser (3). History of depression of the abuser (2). Others (3).Separation and termination of relationship (12)Relationship termination (7). Difficulty for the abuser to accept the end of the relationship (2). History of relationship termination (2). Others (1).Poor support network (10)Poor support network (5). Conflicting family relationships (2). Others (3).Financial difficulties of the abuser (4)Unemployment of the abuser (2). Others (2).Sexual assault (3)Suffer sexual assaults (3).Underestimation of the risk of death (3)Availability and access to firearms for abusers (2). Others (1).Prison (3)Relationship with a man with a history of imprisonment (2). Others (1).Individual (96) Gender (24)Female gender identity (24).Economic conditions (14)Low socioeconomic level (10). Unhealthy physical environment (2). Others (2).Origin (11)Origin outside São Paulo (11).PAS consumption (10)Problematic consumption of PAS (10).Educational conditions (9)Low level of education (7). School drop-out (2).Be alone (6)Be alone at home (5). Others (1).Age (5)Age (Adolescent 1; Elderly 4).Family conditions (5)Exposure/witnessing childhood violence (3). Others (2).Pregnancy (3)Suspected/pregnancy of an abuser (3).Race (3)Black race (3).Occupation (2)Sex work (2).Social (80)Misogyny (40)Misogynist culture (19). High social tolerance for violent behavior against women (17). Lack of mechanisms to deal with emotional issues of dependence and relationship termination (2). Others (2).Chauvinism (27)Chauvinist culture (24). Culture that values monogamy (2). Others (1).Violence and lack of social cohesion (8)Violent culture (2). Death as a strategy of revenge and conflict resolution (2). Lack of social cohesion (2). Others (2).Prejudice (5)Stigma and discrimination against people who use PAS (2). Others (3).it continues29Ciência & Saúde Coletiva, 28(1):23-36, 2023Figure 1. Ecological model categories applied to femicide cases.Source: Authors.3. Society4. Community1. Relationships2.Individual1. Misogyny2. Chauvinism1.Economic conditions2. Violent behavior of the abusers3. Relationship with justice institutions4. Community with drug trafficking and organized crime violence1. Violent relationship2. Behavioral issues of the abuser3. Separation and termination of the relationship4. Poor support network1. Gender2. Economic conditions3. Origin4. PAS consumptionThe behavior of this type is usually accom-panied by justifications such as jealousy, distrust of the woman’s word, or an apparent desire to protect her. With the massive use of electronic devices, other forms of control over women are becoming more frequent25. Alana and Samanta were watched through their phones. Alana re-ceived calls incessantly and was asked to sleep with the phone camera focusing on her image. In Samanta’s case, the attacker asked for the device location to be constantly sent. A partner exhibit-ing a controlling behavior is a risk factor for fem-icide (OR = 5.60; 95%CI 4.41-7.13)20.level (n)Category (n)level elements (n)Community (64)Economic conditions (14)Lack of employment opportunities (6). Low socioeconomic status (4). Neighborhood without paved roads (2). Others (2).Violent conduct of the abusers (14)Availability and access to firearms for abusers (5). Use of violence as a means of conflict resolution (3). Lack of community cohesion (2). Others (4).Relationship with justice institutions (11)Prison system focused on punishment and not on the resocialization of prisoners (5). Others (5).Violence due to drug trafficking and organized crime (8)Community with violence due to drug trafficking and organized crime (6). Others (2).Psychoactive substances (6)Lack of an effective system for rehabilitating and reintegrating people with problematic consumption of psychoactive substances (4). Widespread consumption of psychoactive substances in the community (2).Support networks for women (4)Lack of support networks for women in situations of violence (2). Lack of mechanisms to deal with emotional issues of dependence and relationship termination (2).Source: Authors.Chart 2. Elements and categories of WHO ecological model levels present in femicide cases, Campinas 2018-2019.30Caicedo-Roa M, Cordeiro RC Threat or attempted suicide by the abuser In three cases, the abusers threatened to commit suicide before the end of the relationship (cases of Orquídea, Taciana, and Alana). Psycho-logical pressure is used as a way to keep women in relationships. Faced with the threat of suicide by an author of violence, women are recommended to avoid contact with the individual and that they notify the suicide prevention services to manage the crisis. Thus, the abuser can receive support in the face of the suicidal threat, and the women are not exposed to situations of potentially lethal physical confrontation for them. idea of betrayal or infidelitySome femicides are justified by the abusers as a response to betrayal. In Franciele’s case, af-ter ending the relationship, the husband kills her, attacks the alleged lover with two shots, and lat-er commits suicide for not accepting the end of the relationship and believing that he had been betrayed. Alana received 16 gunshots because her partner found messages with hearts on her phone. Valeriana was brutally beaten for having sex with other partners. The idea of betrayal to justify killing a woman is associated with the conception of ownership over their bodies, actions, and feelings. Theories such as sexual property have suggested that the probability of violence increases when the man believes he has a right over the woman and her reproductive capacities. In this context, the loss of such control triggers aggression20. Other the-ories focus on the male idea of the right to au-thority, where violence is seen as an instrument to limit female independence, reinforced by the idea that the woman belongs to him. termination of relationshipThe end of a violent relationship is a trigger for femicide (OR = 2.33; 95%CI 1.64-3.30)20,26. Nicolaidis et al. conducted a study with women who had survived femicide attempts, finding that in 73% of cases, the woman had tried to leave the relationship18. In this research, six of the women died after trying or ending a violent relationship. The murders occurred shortly after the end. The first year after separation is the period of greatest risk26, the first three months being critical. However, ending the relationship does not guarantee the cessation of violence. Aggressions and threats persist. There are multiple reasons women do not want to end relationships: fear of more severe abuse or death, lack of economic support, concern for children, emotional depen-dence, lack of support networks, hope for change from the abuser, and associated prejudice with the separation, among others27. The following items stand out among the most common predictors of termination: the na-ture of the violence, the woman’s life history, so-cial and psychological factors, external resources, and coping strategies28. The end of an abusive re-lationship is not usually a linear process, usually with many ups and downs. Sometimes victims intend to leave relationships for reasons other than violence, such as alcohol or drug abuse by partners, financial problems, or infidelity rather than recognizing the risk of their death18.Generally, the quality of life of abused wom-en improves after separation, but permanence in a relationship lasts an average of 11 years29. The process of breaking up an abusive relationship is not a simple matter. Women in this situation should be supported with available legal resourc-es. Terminations of violent relationships must be assisted by competent domestic violence per-sonnel; otherwise, they may increase the risk of death.The restraining order is a resource available to women who end a relationship. It notifies the abuser about the protection of the State and lim-its the possibilities of contact under penalty of imprisonment. Brazilian law also provides that divorce applications do not require the willing-ness of both parties, which facilitates the termi-nation of legally established ties with the abuser. Poor support networkWomen who experience violence have pre-carious support networks. 2002 WHO study reported that approximately 80% of the women interviewed in Brazil had talked about physical abuse with someone. However, many women avoid talking about the experience of violence, and when they do, they usually talk to family, friends, or religious leaders13. Traditionalist conceptions about the role of women, chauvinist attitudes, distrust of wom-en’s words when they denounce violence, ten-dency to blame the victim or justify the abuser, and feelings of shame and guilt prevent women from exposing situations of violence and seeking help. The dichotomy between public and private makes women vulnerable, placing them in envi-ronments where it is more challenging to offer protection. Victims of moderate and severe physical vi-olence seek institutions or authorities but do so 31Ciência & Saúde Coletiva, 28(1):23-36, 2023when they can no longer tolerate the aggressions or are seriously injured13. In the face of risk situ-ations for physical integrity, it is recommended to have an escape route planned to protect them-selves and subsequently report it to the author-ities. Women victims of domestic violence who do not find support networks in the community, school, legislation, or society become victims of femicide.Support networks are vital for adolescents be-cause they lack sufficient personal and economic resources to face violence. Adolescents like Mi-randa and Jeannete, aged 13 and 19, had poor support networks. Young women are increasing-ly experiencing situations of violence by intimate partners, showing that violence starts early13,30. A study with participants from five US cities found that almost 50% of women who survived femicide attempts underestimated the danger-ousness of the situation, did not recognize that their lives were at risk, and were surprised by the abuser’s attack18. The fact that women do not recognize that they are at risk of death and that perpetrators do not have a single pattern rein-forces the need to train people who carry out in-terventions or have contact with women victims of violence—such as healthcare workers, lawyers, psychologists, social workers, and others—to be aware of the seriousness of the situation and the risk of death for the woman. individual levelFemale gender identity, origin outside São Paulo, problematic consumption of PAS, and low socioeconomic and schooling level were the rele-vant factors for femicide at this level. GenderOne essential but controversial and complex point is identifying the gender component in fe-micide cases. In the research carried out by the Gender Nucleus of the Public Ministry of São Paulo, this question was posed31. Cases are only classified as femicides when there is a relation-ship or evidence of domestic violence between the victim and the abuser; other murders are not typified as such, which implies under-classifica-tion and underreporting of cases. Women are primarily victims of violence in relationships, and while they can resort to violent behavior against abusers, their position in the relationship itself is disadvantageous. Women sometimes use violence to defend themselves but rarely initiate physical aggression13. economic conditionsWomen with low economic conditions suffer violence more frequently32. High levels of poverty can increase tension and conflict in relationships. Couples dissatisfied with their finances are more likely to resort to violence32. On the other hand, the abuser being employed has been identified as a protective factor against femicide (OR = 0.50; 95%CI 0.36-0.70)20.Economic deprivation and marginalization caused by gender inequities influence femicides. Women who lack financial independence may find it more difficult to leave abusive relation-ships19. Women with better financial resources are more likely to leave their aggressive part-ners28. Economic independence makes them less willing to maintain violent relationships, given the low perception of retribution and the high emotional cost. Women in situations of violence spend their energy on safety for themselves and their chil-dren. Violence prevents them from engaging in work activities, causes absenteeism, and causes significant difficulties in maintaining or chang-ing jobs. In addition to providing economic in-come, working increases women’s self-esteem and freedom. Having one’s own money provides security and a feeling of self-sufficiency33. OriginMigrant women face factors that increase their vulnerability, such as isolation, distance, separation from the family, racism, xenopho-bia, and precarious and informal work relation-ships34. Migration studies with a gender focus have shown the relevance of incorporating differ-ences by sex in the analysis of migratory flows35. Migrant women are more exposed to structural inequities, lower economic resources, lack of em-ployment options, and divorce36. PAS consumption by the victimThe victim’s use of PAS increases the risk of femicide (OR = 2.56; 95%CI 1.78-3.67). Studies have indicated the consumption of substances by women as a factor that favors their victimization20. educational conditionsThere is an inverse relationship between education and violence, suggesting that educa-tion has a protective effect. Women with higher schooling level are more likely to choose their partners, decide on marriage, and have greater autonomy and control over resources within the marriage13,15. Better educational conditions are 32Caicedo-Roa M, Cordeiro RCclosely related to better economic positions, af-fecting violence exposure. Femicides of pregnant women, with suspected pregnancy, and postpartum womenSocially, the role of the mother is one of the most valued. The prevalence of violence in preg-nancy is estimated to reach up to 45%37-39. Of the autopsies performed, one was of a woman mur-dered during pregnancy and another during the puerperium. An international study of women assaulted during pregnancy reported that in 90% of cases, the attacks came from the biological fa-ther of the unborn child13. Pregnancy has been identified as a risk condition for femicide40. Men who abuse women during pregnancy are more at risk of committing femicide (OR = 3.93; 95% CI 2.99-5.18)20. Social levelAt the social level, the misogynist and chau-vinist culture stood out, in addition to the high social tolerance of violent behavior against wom-en. Latin American societies such as Brazil have been influenced mainly by traditionalist and reli-gious thought models.Misogyny and chauvinismMisogyny is a term that describes hatred, contempt, or discriminatory prejudice against women. It is a concept of social character man-ifested in relationships. It can be present as ex-clusion, hostility, discrimination, subordination, violence, and the maintenance of male privilege over females. Misogyny is typical of environ-ments where women are perceived as violating patriarchal norms and repressed through hostile reactions. Thus, misogyny refers to control over women, which perpetuates their subordination41. The hatred and control over women are based on religious, naturalistic/biological, and cultural beliefs that pass through different generations in education and socialization practices. Ideas such as women’s lesser intelligence and their propensi-ty for evil are ingrained in some cultures. Women in patriarchal and misogynistic societies cannot exercise their freedom, self-determine their exis-tence, and enjoy fundamental rights, being pun-ished and attacked for their status as women. Over the last few years, the issue of violence against women has been visible. Issues such as dignity, integrity, respect, well-being, and valu-ing the feminine are increasingly present in po-litical, academic, social, and family debates as a response against the system of values and beliefs that shape unequal relations between genders. Men with traditionalist thoughts about gen-der roles are likelier to engage in violence against their partners to ensure obedience and submis-sion. In affective relationships, the control of women begins with the control and devaluation of the feminine. Violence often escalates into sexual violence; women are required to have unwanted sexual contact or perform sexual acts without their approval. Then, physical aggression appears. All behaviors respond to the purpose of guaranteeing female obedience and submission for the benefit of the male counterpart42. Violence and lack of social cohesionSocial organization refers to the ability of community residents to regulate behavior. Com-munities with social isolation find it more chal-lenging to establish agreements on the behavior of residents and non-residents. The lack of ties can generate an apathetic attitude towards situ-ations of domestic violence, e.g., avoiding calling the police or confronting the abuser, making vi-olent men feel free to attack their partners with impunity. Limiting a woman’s contact with other peo-ple is part of the control exercised by partners, reducing the likelihood of exposing the violent behavior. In turn, isolation increases female de-pendence on men. Women with strong social support are more likely to find help leaving an abusive partner. Social disapproval restrains ag-gressive male behavior32. Community level The cases studied show that femicides affect both women with favorable and precarious eco-nomic conditions. Victims of domestic violence rarely rely on their neighbors to help resolve per-sonal issues. Individuals with medium and high incomes are less likely to intervene in situations of domestic violence than individuals with lower incomes43. At this level, the following stand out: relationship with justice institutionsA national survey showed that among wom-en who were victims of serious aggression, 22.2% sought an official body such as women’s police stations, the military police, or hotlines30. Re-porting domestic violence is rare. In Ramona’s case, the abuser threatened to kill her and her family if a complaint was made. 33Ciência & Saúde Coletiva, 28(1):23-36, 2023The social response to domestic violence re-duces femicides. Studies have addressed this is-sue by testing the exposure reduction hypothesis, which advocates that services or interventions that limit contact between the victim and the abusive partner reduce the likelihood of abuse and violence44. Among the interventions are the restraining orders and arrest of male abusers. Men may take revengeful attitudes towards im-posed measures, particularly when they feel the measures prevent them from exercising their authority. Thus, they can react violently when restraining orders violate their “right” to domi-nation or control. Women avoid seeking assistance and legal protection services due to the perception of the complexity of the legal process. This situation occurs particularly with poor women with low education and low economic possibilities to as-sume the costs of legal processes. Clear police procedures for domestic violence situations and trained units are needed. Traditional police offi-cers may show minor sensitivity to situations of domestic violence or consider it trivial, especially when faced with the ambivalence of denuncia-tors, who denounce the aggressions but do not definitively end the abusive relationships. Drug trafficking and organized crimeMurders related to human trafficking, drug trafficking, and organized crime affect women for reasons of gender. Gang research has shown that while male gang members experience higher victimization rates, female members are exposed to sexual violence by both opposing gangs and male members in their gangs. Women linked to male gang members are at greater risk of serious violence than women without such ties1,45. Final considerations Analyzing the cases of femicide in Campinas through the ecological model of violence gave us an expanded understanding of the phenome-non of lethal violence against women. The most relevant level for femicide is the relational one, in which four categories stand out: the relation-ship with a violent man, behavioral issues of the abuser partner, the end of the relationship, and a deficient support network. For the individual level, gender, economic conditions, origin, and problematic PAS consumption are relevant. On the other hand, at the social level, misogyny and chauvinism are relevant. Finally, at the commu-nity level, the precarious economic conditions in the community, the problematic relationship with the justice institutions, and the community violence derived from drug trafficking and orga-nized crime are present. This analysis shows that several elements contribute to the occurrence of femicide. Women seeking care services or reporting violence need to be correctly oriented. It is es-sential to assess the risk of femicide to offer legal and social measures to preserve their integrity, guarantee their security, and improve their qual-ity of life. The reporting is an initial step towards interdicting the cycle of violence. However, the actions cannot be limited to the promotion of the complaint. They must be accompanied by a structure of a material, psychological, and securi-ty support for women and their dependents. If the denunciators are not welcomed and ori-ented, they will return to situations of violence due to economic and emotional dependence or lack of personal, community, and social resourc-es. Women go through a period that requires strengthening internal mechanisms and seeking external help to end the situation of vulnerability. Economic issues appeared at both the indi-vidual and community levels and undoubtedly contributed to the occurrence of femicide cases. Societies structured in economic, racial, and gen-der inequality environments have a higher inci-dence of femicide. 34Caicedo-Roa M, Cordeiro RCCollaborationsM Caicedo-Roa and RC Cordeiro: participated in all phases of the manuscript elaboration, in the conception of the research design, analysis of the text contents, the writing of the article, and its critical review. Acknowledgments To professor Lourdes Maria Bandeira (Depart-ment of Sociology at UnB) in memoriam for her considerations and contributions throughout the research.35Ciência & Saúde Coletiva, 28(1):23-36, 2023FundingThe research was financed through a scholarship granted by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (Capes) code 001. Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) partially financed the field-work (2018/07162-0).references1. United Nations Office on Drugs and Crime. Global Study on Homicide. Gender-related Killing of Women and Girls. Viena: UN; 2018. 2. World Health Organization (WHO). Pan American Health Organization (PAHO). Understanding and addressing violence against women. Femicide. Geneva: WHO; 2012.3. Instituto de Pesquisa Econômica Aplicada (Ipea). Fó-rum Brasileiro de Segurança Pública. Atlas da Violên-cia. Brasília: Ipea; 2019.4. Fórum Brasileiro de Segurança pública. Anuário Brasileiro de Segurança Pública [Internet]. 2021. [acessado 2022 ago 3]. Disponível em: https://forum-seguranca.org.br/wp-content/uploads/2021/10/anua-rio-15-completo-v7-251021.pdf 5. Krug EG. Relatório mundial sobre violência e saúde. Genebra: OMS; 2002.6. Dahlberg LL, Krug EG. Violência: um problema global de saúde pública. Cien Saude Colet 2006; 11(Supl.):1163-1178.7. Heloani R, Capitão CG. Formulações gerais sobre o objeto de pesquisa em Psicologia: estudo de caso. In: Heloani R, Souza RMB, Rodrigues RRJ, organi-zadores. Sociedade em transformação: estudo das re-lações entre trabalho, saúde e subjetividade. Londrina: EDUEL; 2007. p. 17-35.8. World Health Organization (WHO). Verbal autopsy standards: ascertaining and attributing cause of death; Geneva: WHO; 2007.9. Caicedo-Roa M, Cordeiro RC, Martins ACA, Faria PH. Femicídios na cidade de Campinas, São Paulo, Brasil. Cad Saude Publica 2019; 35(6):e00110718.10. Caicedo-Roa M, Nascimento JL, Bandeira LM, Cor-deiro RC. Queima às bruxas: feminismo e feminicí-dios íntimos por queimadura em uma metrópole bra-sileira. Cien Saude Colet 2022; 27(2):525-534. 11. World Health Organization (WHO). Department of Reproductive Health and Research, London School of Hygiene and Tropical Medicine, South African Medi-cal Research Council. Global and regional estimates of violence against women: prevalence and health effects of intimate partner violence and non-partner sexual violence; Geneva: WHO; 2013.12. United Nations (UN). The world’s women 2015: trends and statistics. New York: UN; 2015.13. Organización Mundial de la Salud (OMS). Estudio multipaís de la OMS sobre salud de la mujer y violen-cia domestica contra la mujer. Ginebra: OMS; 2002.14. Silva M, Neto G, Figueiroa J, Cabral Filho J. Violen-ce against women: prevalence and associated fac-tors in patients attending a public healthcare service in the Northeast of Brazil. Cad Saude Publica 2010; 26(2):264-272.15. Santos IB, Leite FMC, Amorim MHC, Maciel PMA, Gigante DP. Violência contra a mulher na vida: estudo entre usuárias da Atenção Primária. Cien Saude Colet 2020; 25(5):1935-1946.16. Campbell J, Glass N, Sharps PW, Laughon K, Bloom T. Intimate partner homicide: review and implications of research and policy. Trauma Violence Abuse 2007; 8(3):246-269.36Caicedo-Roa M, Cordeiro RC17. Lucena K, Deininger L, Coelho H, Monteiro A, Vian-na R, Nascimento J. Analysis of the cycle of domestic violence against women. J Human Growth Develop 2016; 26(2):139-146.18. Nicolaidis C, Curry MA, Ulrich Y, Sharps P, McFarla-ne J, Campbell D, Gary F, Laughon K, Glass N, Cam-pbell J. Could we have known? A qualitative analysis of data from women who survived an attempted ho-micide by an intimate partner. J Gen Intern Med 2003; 18(10):788-794. 19. Reckdenwald A, Parker KF. Understanding the chan-ge in male and female intimate partner homicide over time: a policy-and theory-relevant investigation. Fe-minist Criminology 2011; 7(3):167-195.20. Spencer CM, Stith SM. Risk factors for male perpe-tration and female victimization of intimate partner homicide: a meta-analysis. Trauma Violence Abuse 2020; 21(3):527-540.21. Campbell JC, Webster DW, Glass N. The danger as-sessment validation of a lethality risk assessment instrument for intimate partner femicide. J Interpers Violence 2009; 24(4):653-674.22. Brasil. Conselho Nacional do Ministério Público (CNMP). Formulário de avaliação de risco FRIDA: um instrumento para o enfrentamento da violência domés-tica contra a mulher. Brasília: CNMP; 2019. 23. Echeburúa E, Fernández-Montalvo J, Corral P, López-Goñi JJ. Assessing risk markers in intimate partner femicide and severe violence: a new assessment ins-trument. J Interpers Violence 2009; 24(6):925-939.24. Sharps P, Campbell JC, Campbell D, Gary F, Webster D. Risky mix: drinking, drug use, and homicide. NIJ Journal 2003; 250:8-13.25. Campeiz AB, Carlos DM, Campeiz AF, Silva JL, Freitas LA, Ferriani MGC. Violence in intimate rela-tionships from the point of view of adolescents: pers-pectives of the Complexity Paradigm. Rev Esc Enferm USP 2020; 54:e03575.26. Campbell JC, Webster D, Koziol-McLain J, Block C, Campbell D, Curry MA, Gary F, Glass N, McFarlane J, Sachs C, Sharps P, Ulrich Y, Wilt SA, Manganello J, Xu X, Schollenberger J, Frye V, Laughon K. Risk factors for femicide in abusive relationships: results from a multisite case control study. Am J Public Health 2003; 93(7):1089-1097. 27. Ellsberg M, Peña R, Herrera A, Liljestrand J, Winkvist A. Candies in hell: women’s experiences of violence in Nicaragua. Soc Sci Med 2000; 51(11):1595-1610.28. Anderson DK, Saunders DG. Leaving an abusive part-ner: an empirical review of predictors, the process of leaving, and psychological well-being. Trauma Vio-lence Abuse 2003; 4(2):163-191.29. Alsaker K, Moen BE, Kristoffersen K. Health-related quality of life among abused women one year after leaving a violent partner. Social Indicators Research 2008; 86(3):497-509.30. Fórum Brasileiro de Segurança Pública. Visível e invisível: a vitimização de mulheres no Brasil [Inter-net]. 2019. [acessado 2021 out 13]. Disponível em: https://forumseguranca.org.br/wp-content/uplo-ads/2019/02/relatorio-pesquisa-2019-v6.pdf31. Núcleo de Gênero Ministério Público do Estado de São Paulo. Raio X do feminicídio em São Paulo: é pos-sível evitar a morte [Internet]. 2018. [acessado 2021 jul 7]. Disponível em: http://www.mpsp.mp.br/portal/page/portal/Nucleo_de_Genero/Feminicidio/RaioX-FeminicidioC.PDF32. Van Wyk JA, Benson ML, Fox GL, DeMaris A. De-tangling individual-, partner-, and community-level correlates of partner violence. Crime & Delinquency 2003; 49(3):412-438.33. Alsaker K, Moen BE, Baste V, Morken T. How has li-ving with intimate partner violence affected the work situation? A qualitative study among abused women in Norway. J Fam Violence 2016; 31:479-487.34. Bertoldo J. Migração com rosto feminino: múltiplas vulnerabilidades, trabalho doméstico e desafios de políticas e direitos. Katálysis 2018; 21(2):313-323.35. Peres RG, Baeninger R, organizadores. Migração fe-minina: um debate teórico e metodológico no âm-bito dos estudos de gênero. Anais do XVIII Encontro Nacional de Estudos Populacionais. 2012 nov 19-23, Águas de Lindoia, Brasil. Associação Brasileira de Es-tudos Populacionais; 2012.36. Sabri B, Campbell JC, Messing JT. Intimate partner homicides in the united states, 2003-2013: a compari-son of immigrants and nonimmigrant victims. J Inter-pers Violence 2021; 36:4735-4757.37. Campbell JC, Lewandowski LA. Mental and phy-sical health effects of intimate partner violence on women and children. Psychiatr Clin North Am 1997; 20(2):353-374.38. Gharacheh M, Azadi S, Mohammadi N, Montazeri S, Khalajinia Z. Domestic violence during pregnancy and women’s health-related quality of life. Glob J He-alth Sci 2016; 8(2):27-34.39. Conceição HN, Coelho SF, Madeiro AP. Prevalência e fatores associados à violência por parceiro íntimo na gestação em Caxias, Maranhão, 2019-2020. Epidemiol Serv Saude 2020; 30(2):e2020848.40. Campbell JC, Lewandowski LA. Mental and physical health effects of intimate partner violence on women and children. Psychiatr Clin North Am 1997; 20:353-374.41. Manne K. Down girl: the logic of misogyny. New York: Oxford University Press; 2017. 42. Conopoima Y. El femicidio como resultado de la educación patriarcal. Universidad y Sociedad 2019; 11(4):118-123.43. Edwards KM, Mattingly MJ, Dixon KJ, Banyard VL. Community matters: intimate partner violence among rural young adults. Am J Community Psychol 2014; 53(1-2):198-207.44. Dugan L, Rosenfeld R, Nagin DS. Exposure reduction or retaliation? The effects of domestic violence resour-ces on intimate-partner homicide. Law Society Review 2003; 37(1):169-198.45. Manjoo R. Report of the Special Rapporteur on vio-lence against women, its causes and consequences. United Nations; 2012.Article submitted 09/04/2022Approved 15/08/2022Final version submitted 17/08/2022Chief editors: Romeu Gomes, Antônio Augusto Moura da SilvaThis is an Open Access article distributed under the terms of the Creative Commons Attribution LicenseBYCC | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/343121072Sputtered-Growth of High-Temperature Seed-Layer Assisted β-Ga2O3ThinFilm on Silicon-Substrate for Cost-Effective Solar-Blind PhotodetectorApplicationArticle in ECS Journal of Solid State Science and Technology · July 2020DOI: 10.1149/2162-8777/aba7fdCITATIONS14READS3932 authors:Some of the authors of this publication are also working on these related projects:Organic Photovoltaics View projectMetal oxide nanostructures View projectKanika AroraStanford University13 PUBLICATIONS 387 CITATIONS SEE PROFILEMukesh KumarIndian Institute of Technology Ropar81 PUBLICATIONS 2,367 CITATIONS SEE PROFILEAll content following this page was uploaded by Kanika Arora on 26 July 2020.The user has requested enhancement of the downloaded file.ECS Journal of Solid StateScience and Technology ACCEPTED MANUSCRIPT • OPEN ACCESSSputtered-growth of high-temperature seed-layer assisted β-Ga2O3 thinfilm on silicon-substrate for cost-effective solar-blind photodetectorapplicationTo cite this article before publication: Kanika Arora et al 2020 ECS J. Solid State Sci. Technol. in press https://doi.org/10.1149/2162-8777/aba7fdManuscript version: Accepted ManuscriptAccepted Manuscript is “the version of the article accepted for publication including all changes made as a result of the peer review process,and which may also include the addition to the article by IOP Publishing of a header, an article ID, a cover sheet and/or an ‘AcceptedManuscript’ watermark, but excluding any other editing, typesetting or other changes made by IOP Publishing and/or its licensors”This Accepted Manuscript is © 2020 The Author(s). Published by IOP Publishing Ltd.. As the Version of Record of this article is going to be/has been published on a gold open access basis under a CC 4.0 licence, this AcceptedManuscript is available for reuse under the applicable CC licence immediately.Everyone is permitted to use all or part of the original content in this article, provided that they adhere to all the terms of the applicable licencereferred to in the article – either https://creativecommons.org/licenses/by/4.0/ or https://creativecommons.org/licenses/by-nc-nd/4.0/Although reasonable endeavours have been taken to obtain all necessary permissions from third parties to include their copyrighted contentwithin this article, their full citation and copyright line may not be present in this Accepted Manuscript version. Before using any content from thisarticle, please refer to the Version of Record on IOPscience once published for full citation and copyright details, as permissions may be required.All third party content is fully copyright protected and is not published on a gold open access basis under a CC licence, unless that is specificallystated in the figure caption in the Version of Record.View the article online for updates and enhancements.This content was downloaded from IP address 213.182.195.10 on 26/07/2020 at 19:46For Review OnlySputtered-growth of high-temperature seed-layer assisted β-Ga2O3 thin film on silicon-substrate for cost-effective solar-blind photodetector applicationJournal:ECS Journal of Solid State Science and TechnologyManuscript IDJSS-100373.R1Manuscript Type:Research PaperDate Submitted by the Author:08-Jul-2020Complete List of Authors:Arora, Kanika; IIT Ropar, PhysicsKumar, Mukesh; IIT Ropar, PhysicsKeywords:Solar-blind photodetector, Seed layer, Silicon substrate, optimised parameters, β-Ga2O3 https://mc04.manuscriptcentral.com/jss-ecsECS Journal of Solid State Science and Technologyccepted ManuscriptFor Review Only1 Sputtered-Growth of High-Temperature Seed-Layer Assisted β-Ga2O3 Thin-Film on Silicon Substrate for Cost-Effective Solar-Blind Photodetector Application Kanika Arora and Mukesh Kumar z Department of Physics, Indian Institute of Technology Ropar, 140001, Punjab, India zE-mail: mkumar@iitrpr.ac.in Abstract β-Ga2O3 thin films was grown on cost-effective p-Si(100) substrate by sputtering technique. The evolution of crystalline structure with growth parameters revealed that the gallium oxide thin film grown on the high-temperature seed layer and various optimised growth parameters like sputtering power, deposition pressure and pre-substrate annealing has been proved extremely beneficial in exhibiting excellent crystalline quality. However, the direct growth of β-Ga2O3 on Si substrate with seed-layer was found to be amorphous in nature. The discussion about the critical role of varied growth conditions were carried in detail. The photoresponse of the optimized device showed a photoresponsivity of 95.64 AW-1 and a corresponding quantum efficiency of 4.73x104 % at moderate bias under 250 nm illumination which is higher than most of the devices being reported on planar β-Ga2O3 solar-blind photodetectors deposited on high cost substrates. Moreover, the device showed the high transient response at moderate as well as at self-bias mode with good reproducibility and stability. The rise and decay time of the photodetector at self-powered mode was found to be in millisecond (58.3 ms/34.7 ms). This work paves the alternative way towards the fabrication of β-Ga2O3 solar-blind photodetector on cost-effective substrate and compatible with mature Si technology. This paper is part of the JSS Focus Issue on Gallium Oxide Based Materials and Devices II. Introduction Page 1 of 26https://mc04.manuscriptcentral.com/jss-ecsECS Journal of Solid State Science and Technology123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960ccepted ManuscriptFor Review Only2 Photodetectors are one of the key components in many optoelectronic devices, which transduce optical into electrical information. There are three components of the solar spectrum—ultraviolet, visible, and infrared—which are collectively referred as optical radiation. The UV radiation further divides itself into three parts as UV-A (320nm-400nm), UV-B (280nm-320nm) and UV-C (200-280nm). At sea level, approximately 6% of the optical radiation is UV of which UVC is basically absent as it cannot penetrate the earth’s protective atmosphere. The photodetectors operating in this region could minimise the chances of false radiation by detecting a very weak signal under intense earthly background, strengthening the ability of resistance to environmental interference. Hence they can early detect the signals often call early warning signals before any catastrophic events take place and has wide range of military and civil applications such as the missile tracking, flame detection in petrol filling stations, oil spill detection from leaking tankers in sea, volcanic eruptions detection, UV astronomy, shrinkage of the ozone layer and others1-5. Henceforth these detectors have continuously provided the motivation to develop the technologies which can tolerate increasingly high temperatures and harsh environments and also at the same time can commercialise for real-world applications that would demand the production of large scale wafer processing by employing millions of miniature pixels into large arrays.2-4 To have photodetector working in the solar blind region (200-280nm) we need wider band gap semiconductors6 like β-Ga2O32-4, 7, ZnMgO8-10, ZnGa2O411, Zn2GeO412, AlxGa1−xN13, In2Ge2O714, LaAlO315, BN16, and diamond17. Materials like nitrides and diamond are sensitive only to a limited range of UVC light, also they have complex heterostructures which involve multipart fabrication process, hence restricting their widespread applications. While for other materials, it is difficult to produce single-crystal of these materials owing to phase-separation, therefore single semiconductor material that can be easily fabricated would be extremely beneficial. In particular, UV photodetectors fabricated from wide band-gap semiconductors such as gallium oxide (β-Ga2O3) have attained an increasing attention in recent years because of their high thermal (only form of monoclinic metallic oxide that is stable from its room temperature to its melting point i.e. 1780º C) and chemical stability as well high tolerance of electric field determining its possibility to work in hostile environments. The major strength of Page 2 of 26https://mc04.manuscriptcentral.com/jss-ecsECS Journal of Solid State Science and Technology123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960ccepted ManuscriptFor Review Only3 gallium oxide material based solar-blind photodetectors is the high thermal and chemical stability and potentially high electric field tolerance and efficient control over harsh environment due to the reasons stated above. The choice of material will, therefore, show a crucial role for further development.3, 5 Numerous studies were conducted concerning the growth of thin films from GO and the manufacturing process of solar-blind UV detectors. Polycrystalline GO thin film deposited on sapphire substrates using the sol-gel process.18 GO thin film deposited on c-plane sapphire substrates had been produced using plasma-assisted MBE 19. GO films were also made with a mist chemical vapor deposition 20 and sputtering21 and some other methods. The manufacturing process of materials and device technology would become more commercially appealing if low cost GO preparation methods were adopted. The sputtering technique is much more appropriate for industrial production compared to the growth methods described above. The GO thin films were generated by many researchers using sputtering technique on quartz and sapphire and their structural and optical properties were studied.21-29 Nonetheless, most materials reported by R.F sputtering have issues in achieving high quality crystal, high substrate cost and high performance using sputteing techniques as they have studied the effect of not more than two parameter at a time and thus they partially solved the problem. Since this can have serious consequences, it is imperative to better understand the effect of multiparamters at a time. From this viewpoint, the framework to high efficiency based devices based on a GO needs to be addressed. Apart from that, so far, mostly gallium oxide based UV detector ever reported was fabricated on sapphire substrates30 and very few have been reported on silicon substrate31. The major disadvantage of the sapphire is its high cost. The fabrication of low-cost photodetector is always demanded the growing interest for commercialization. The growth of high-quality β-Ga2O3 based device on Si substrates is of huge interest in terms of low cost, well-developed technology, large availability, excellent crystalline quality, and good electrical as well as thermal conductivity. However, the development of high-quality gallium oxide with the silicon substrate is challenging due to difficulties such as large lattice mismatch and thermal expansion coefficient between the gallium oxide and silicon which results in high tensile stress and Page 3 of 26https://mc04.manuscriptcentral.com/jss-ecsECS Journal of Solid State Science and Technology123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960ccepted ManuscriptFor Review Only4 dislocation density. These factors critically affect the crystallinity and photo generated carrier dynamics in β-Ga2O3 which finally hinders the carrier mobility, thermal stability, selectivity, responsivity, and sensitivity henceforth degrade the efficiency of the photodetector. Therefore, the growth of electronic quality β-Ga2O3 and an interface engineering of absorbing layer with cost effective substrate is critically required. In order to have the influence of device structure for the high-performance detector, it is very important to understand the device physics. By using thin-film epitaxy approach the misfit strain imposed by a substrate can be strain down to few percent. Such strain-layer epitaxy approach can be used to improve the functionalities of thin films32 33. In this work, we approach the problem with a broader perspective and engineered the interface of β-Ga2O3/p-Si (100) by using all-together multi-parameter approach of R.F magnetron sputtering.Sputtering parameters like sputtering pressure, power and other parameters were optimized to achieve a high-quality β-Ga2O3 thin film for high-performance solar-blind photodetector application. To re-crystallize Si wafer, it was annealed in a vacuum and to accommodate tensile strain, high-temperature seed layer is incorporated followed by post-annealing and sequential re-growth to provide a good template for the epitaxial like growth of β-Ga2O3. Experimental Section The β-Ga2O3 thin films were fabricated using radio frequency magnetron sputtering from a 99.9999 % pure 2” diameter Ga2O3 target. Firstly, p-Si (100) and cleaned sequentially via ultrasonication by submerging into the de-ionized water, acetone and IPA for 30 mins. For optical measurement, gallium oxide thin film is also deposited on quartz substrate. In order to remove impurities like native oxides on substrates, p-Si (100) substrates prior to deposition were dipped into HF solution (5: 95 :: HF: Deionised water) for approximately 1 min. Finally, the cleaned substrates were blown off in nitrogen and loaded into the high vacuum sputtering chamber. The chamber loaded with samples was pumped down to a base pressure of 8x10-7 Torr. During the entire deposition, the flow rate of argon gas (50 sccm) is entered in the sputtering system with the concentration described in Table 1 and substrate was rotated at 4 Page 4 of 26https://mc04.manuscriptcentral.com/jss-ecsECS Journal of Solid State Science and Technology123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960ccepted ManuscriptFor Review Only5 rpm to enhance the film uniformity. To optimise β-Ga2O3 on a silicon substrate for high crystalline quality, the effect of 5 parameters such as sputtering power, growth pressure, substrate annealing, effect of seed layer and multiparameter combination effect were studied in details. The detail growth conditions are listed in Table 1 with Fig. 1 (a-c) to show the description of growth temperature profile. Figure S1 of supplementary information shows the thickness of β-Ga2O3 thin film deposited on Si substrate which is found to be 353±2.35 nm. In Fig.1 (a) growth temperature profile represents the sample were deposited without seed layer while in Fig. 1(b) sample were deposited without substrate heating, Fig. 1 (c) represents the samples that are deposited with combination effect of temperature profile of Fig. 1 (a, b) temperature growth profile. Fig. 1 (d) represents the schematic of the fabrication of β-Ga2O3 thin films under various optimized conditions. TABLE I. The growth parameters of gallium oxide film at four different conditions [deposition pressure, sputtering power, without substrate heating and without seed layer] and the film at the optimised parameter Growth Parameters S. no. Sample Power (W) PAr (mTorr) Substrate heating Seed–layer-assisted growth 1 A 100 5 Yes Yes 2 B 75 3 Yes Yes 3 C 100 3 No Yes 4 D 100 3 Yes No 5 E 100 3 Yes Yes The surface morphologies of Ga2O3 films were investigated using a Bruker scanning probe microscope (SPM) in tapping mode. The structural properties were studied using Grazing Incident X-ray diffraction (GIXRD) using PANalytical X’pert Pro multipurpose diffractometer equipped with high-intensity CuKα (λ=0.1504nm) radiation. We have Page 5 of 26https://mc04.manuscriptcentral.com/jss-ecsECS Journal of Solid State Science and Technology123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960ccepted ManuscriptFor Review Only6 optimised the Omega for maximum intensity and thus fixing it to 1.5º with a step size of 0.04º and 1.5 sec time per step. To study the optical properties of the film, the quartz substrate was also put along with silicon substrate during depositions. Ultraviolet-visible (UV-VIS) transmission spectrum was taken using PerkinElmer spectrophotometer. Device fabrication and measurements The interdigitated metal contact (Cr/Au) having a thickness (80nm/120nm) has been patterned. Prior to electrical measurements, we carried the rapid thermal annealing on a hot plate at 250 °C for 20 mins in air ambient to ensure good contact of electrodes with bottom film. The electrical and photoresponse measurements of the photodetector were performed using a Bentham PVE (300) and monochromator (TMc 300) comprises of 75W xenon arc lamp, a lock-in amplifier, and chopper connected to Keithley 6430 digital current meter. The spectral photo-responses of the devices were measured in the range of wavelength from 250 nm to 1100 nm. The input power intensities at different wavelength were calibrated through the standard Si photodiode. All the measurements were done under ambient conditions in the closed and optically dark chamber to block the interference of surrounding light. Results and discussion Structural characterization and surface topography The crystal structure of each film was investigated with the help of GIXRD to avoid complete load of the signal by the substrate34 . All the peaks shown here are in accordance with JCPDS (041-1103) data as shown in Fig. 2 (a). All samples show common diffraction peaks of (400), (111), (-601), (-603) and (-204) which corresponding to β-Ga2O3 phase located at 30.07º, 34.17º, 44.7º, 59.1º, and 64.1º, respectively. Comparing Sample A with Sample E deposited at varied deposition pressure, it is observed that the peak of β-Ga2O3 phase enhanced with reduce in deposition pressure. This enhancement in crystallinity may be due to the high energy of adatoms depositing on the substrate at lower deposition pressure 35, 36. Therefore, adatoms deposited at 3 mtorr pressure are energetically favourable having high surface mobility due to increase in the mean free path in comparison with 5 mtorr pressure. Hence crystalline peak of gallium oxide deposited at 3 mtorr increases due to increase in crystallinity of the film. Page 6 of 26https://mc04.manuscriptcentral.com/jss-ecsECS Journal of Solid State Science and Technology123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960ccepted ManuscriptFor Review Only7 Next, to see the effect of sputtering power on the crystallinity of the film, we compared sample B and sample E which are deposited at different sputtering powers. The sample B deposited at 75 W power shows very feeble peaks of β-Ga2O3 phase, the film is either microcrystalline or amorphous in nature. Whereas, sample E deposited at power 100W has intense and sharp β-Ga2O3 peaks. At higher sputtering power argon species are highly energised and translates their kinetic energy to the adatoms by momentum transfer. The adatoms which are already residing on the surface of the film get knocked and diffuse to nearby voids due to their enhanced impact surface mobility. This may improve the film quality by reducing voids and resulting the film growth along in-plane orientation. In contrast, at low sputtering power, the adatoms reaching the surface do not have enough impact on adatoms already present at the surface, hence, they grow in out of plane orientation 37. We have also studied the effect of substrate temperature on film growth. On comparing samples C and E in Fig. 2(c), it is observed that peak intensity is weak for sample C which is grown with no substrate heating. We conclude that substrate temperature also has substantial effect on the crystallinity of the film. Heating silicon substrate in vacuum results in reformation of the crystallinity of silicon thus reduces the structural disorder (defects, scratches, etc.). It suppresses the presence of irregularities due to impurities at the topmost surface38 which offer hindrance to the smooth growth of gallium oxide film. Finally, the effect of seed layer was observed by comparing samples D and E. It is clear from Fig. 2 (d) that sample D without seed layer showed no sharp peak. It may be attributed to deterioration of the film due to inhomogeneous strain and low quality crystal epitaxy. Whereas, seed layer growth at high temperature in sample E provides nucleation sites of gallium oxide that have strong bonding with Si substrate. Therefore, next re-growth is on nucleation sites of gallium oxide forming the epitaxial coverage of Ga2O3 on silicon. Hence, seed layer guidance reduces lattice mismatch which enhances the crystallinity. This explains the presence of much stronger peak for sample E than sample D39. The above analysis confirms that Sample E possesses good-quality and high crystallinity with respect to other Ga2O3 thin films. AFM analysis was also performed to study the morphology of the gallium oxide thin films. Page 7 of 26https://mc04.manuscriptcentral.com/jss-ecsECS Journal of Solid State Science and Technology123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960ccepted ManuscriptFor Review Only8 The mean square roughness was found to be 4.20 nm, 6.79 nm, 4.55 nm, 0.308 nm and 9.66 nm for samples A, B, C, D and E, respectively. No specific trend was observed, however homogeneous distribution without any cracks and voids were found as shown in Fig. 2(b). Optical characterization The optical spectra of β-Ga2O3 thin film grown on quartz substrate are shown in Fig. 3. The Fabry-Perot oscillations in transmission spectra reveal good interference quality that reflects uniform and homogenous film growth. The average optical transmission is more than 85% in the visible edge to the deep UV region. Optical bandgap (Eg) of samples are estimated using a formula 2gah= A(h-E )where α represents the absorption coefficient, A is proportional constant, and h is the incident photon energy. The bottom right inset of Fig. 3 shows the Tauc plot of different samples. The band gap was calculated by extrapolating the linear portion of the curve on energy-axis. The optical bandgap of different samples are ranging between 4.41-4.60 eV. The bandgap of β-Ga2O3 lies in the range 4.52-4.59 eV according to literature.40, 41 The increase in bandgap indicates a shift from non-stochiometry to stochiometry and thus an improvement in crystallinity. The difference in growth conditions promote the production of near-stochiometric GO thin films through partial filling of the structural defects which decreases localized state concentration and thus increases the optical bandgap. For hν < 4.41 eV, the exponential tail of the spectrum and the small, discrete tail at about (3.37eV-3.80 eV) is due to oxygen vacancies which introduce deep levels resulting in electronic transition from Ga-4s states in the conduction band minimum and O-2p states in the valence band maximum42. This poor energy absorption can also be related to changes within the energy gap in impurity states. Additional electrons are located at empty 4s orbitals of gallium atoms closer to the vacancy of oxygen and thereby localized states grow in the band gap and therefore drop in band gap energy. Thus the decrease in the optical bandgap shows a decrease of the structural order and the stoichiometry of the thin films. Electrical characterization Photodetector was fabricated on the optimised sample i.e., Sample E, for the performance Page 8 of 26https://mc04.manuscriptcentral.com/jss-ecsECS Journal of Solid State Science and Technology123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960ccepted ManuscriptFor Review Only9 analysis. Fig. 4 (a) shows the I-V characteristics under dark and 254 nm DUV light illumination with optical input power 31 µWcm-2 . It is observed that there is a sharp increase of photocurrent with bias voltage for DUV illumination in comparison to dark. At wavelength 254 nm and Poptical = 31 µWcm-2, the photodetector exhibits Ilight = 6.32 µA. Inset of Fig. 4 (a) represents the schematic of fully fabricated β-Ga2O3 device characteristics under DUV light. Fig. 4 (b) shows the relationship between linear dynamic ranges (LDR) which gives information about signal-to-noise ratio. It is determined by the relation given by equation (1): 20 loglightDarkILDRI (1) where Ilight is the photocurrent at 254 nm illumination and Idark is the current at the dark condition. It can be seen that the highest LDR is 9.54 dB at 4.4 V indicating high signal-to-noise ratio. Photoresponse of the device is a very important paramter to determine the performance of the photodetector at different wavelengths, under 1 V, 2 V, 3 V, 4 V and 5 V biasing conditionsas shown in Fig. 4(c). The responsivity of photodetector is determined by the relation given as equation (2)41, 43-46: lightdarkopticalIIRP (2) Where R represents the responsivity and Poptical as the optical power intensity of the incident light. As depict from Fig. 4 (c) the responsivity increases as bias voltage increases from 1 V to 5 V, due to more efficiently charge separation and collection. A sharp edge cut-off response for sample E at 250nm, 5 V bias (Poptical = 17 nW/cm2) can be clearly observed, that indicates the device has higher spectral selectivity towards the deep UV region. The spectral response of our photodetector is compared with other recently reported solar-blind photodetector in Table 2. The responsivity of 95.64 AW-1 was achieved under 5 V bias which is much higher than most of the devices being reported, demonstrating the good reliability to grow these photodetectors on cost effective Si substrate. Photoresponse from underneath silicon substrate is not detected in our device which is indicating good insulation of our β-Ga2O3 film from silicon. Fig. 4 (d) shows the variation of external quantum efficiency (EQE) with a Page 9 of 26https://mc04.manuscriptcentral.com/jss-ecsECS Journal of Solid State Science and Technology123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960ccepted ManuscriptFor Review Only10 wavelength under various bias voltages ranging from 1 V - 5 V which is another important factor to determine the performance of photodetector and is given by equation (3)47, 48: RhcEQEe (3) where R is the responsivity, h is Planck’s constant, c is the speed of light, e is the electron charge and 𝝀 is the wavelength of incident optical light. Similarly, as illustrate from Fig. 4 (d) that the EQE increases with increase in voltage from 1 V to 5 V due to increase in charge carriers’ collection efficiency with bias. It is observed that photodetector device exhibits a higher EQE of 4.73x104 % at 250 nm, 5 V bias. This implies our photodetector has high internal gain making it viable to work in solar-blind region. Figure 5 (a-b) shows the transient response measured by periodically switching on/off 254 nm DUV light repeatability at zero bias for optimized sample E. The optimised sample based β-Ga2O3 photodetector shows the magnificent performance at self-powered mode. The presence of back-to-back Schottky-like diode in dark with zero bias voltage prevents charge carriers from crossing the Au electrode/Ga2O3 interface, that corresponds a reasonably very low dark current of 1.96 pA. On illumination with DUV light (intensity 31 µWcm−2) at zero bias, a sharp rise and decay in photocurrent was observed as seen in Figure 5 (a). This depicts that the β-Ga2O3 device fabricated by multiple parameter approach shows great reproducibility and high reliability for solar-blind photodetector application. The rise and decay time of the β-Ga2O3 solar-blind photodetector is calculated by fitting the equation 12//0ttIIAeBe44, 49, 50on single cycle of transient photoresponse of sample E. The rise and decay time is found to be 58.3 ms and 34.7 ms respectively, as shown in Fig. 5 (b). Such high photo-response speed at zero bias is depicted based on theory of energy band. Owing to the structural disparity between Au and β- Ga2O3, β-Ga2O3 energy bands bends upward at the interface of source and drain sides and the Schottky barrier forms. Besides the variation in work functions, Schottky barrier height formation is well known to be greatly influenced by surface interface states and adsorbent sites. The interface states will occur as the Ga2O3 contact surface with the Au electrode breaks the lattice periodicity where the probability of point of contacts between the Au and β-Ga2O3 varies. We identified asymmetric features of Page 10 of 26https://mc04.manuscriptcentral.com/jss-ecsECS Journal of Solid State Science and Technology123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960ccepted ManuscriptFor Review Only11 I-V indicating different Schottky barrier height (SBH) at both ends. We evaluated tunneling-related energy (E00) to thermal energy (kBT) with the following relationship in order to calculate the reasonable approximation of the prevailing transportation process in the MSM DUV photodetector: 1/200*2sqNEm (1) where the charge carrier, carrier density, effective mass of charge and dielectric constants are the q, N, m *, and ɛs respectively. For kBT>>E00, the tunneling process is dominated by thermionic emissions. In this instance, m m* =0.34m0 51 and ɛs =10 52 for Ga2O3, and its carrier density N for Ga2O3, because of its highly resistant nature, seems to be in the scale of 1015-1016 cm-3 53. When these values are inserted, one might anticipate that its E00 is far below kBT within the range of 0.17-1.7 meV. It is therefore reasonable to suggest that perhaps the thermionic emissions at the Au/Ga2O3 interface dominate the Schottky behaviour. The SBH difference for MSM Ga2O3, can be computed using the dark I-V spectra. Due to the fact that our MSM structure has two consecutively Schottky barriers connected to a coplanar surface, we can therefore use the thermionic emission model equations described below. Figure S2 of supporting details shows the fitted data. 12211222exp() 1exp() 1*exp()*exp()TESTESTEB TESTEB TESTEqVqVIIIkTkTqIAA TkTqIAA TkT (2) where ISTE, A*, q, k, ΦB1TE, ΦB2TE and h are the thermionic emission saturation current, Richardson constant, charge of electron, Boltzmann constant, Schottky barrier heights (SBH) of either coplanar electrodes and Planck 's constant, respectively. The barrier level at the source-drain in forward as well as reverse biases are 0.812 eV and 0.861 eV respectively. This shows that such Au/Ga2O3 contact interaction is an asymmetrical. Due to the formation of states at either side of the interface, defects as well as adsorbent states at the source-drain sides, the differences observed in either side of the SBH for Ga2O3 thin film is cultivated. This allows Page 11 of 26https://mc04.manuscriptcentral.com/jss-ecsECS Journal of Solid State Science and Technology123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960ccepted ManuscriptFor Review Only12 a strong photoconducting function to be achieved at even zero voltage. Such asymmetric behavior with similar electrodes was also recently evaluated by Cui. et. al. and showed the impact of MSM β- Ga2O3 metal oxide photodiode on asymmetric SBH with relation to oxygen flow during growth.53 The band bending builds up a different source-drain interface and develops a built-in electric field that eventually led to the efficient pair separation of e/hs. Recently, and Qinwei. et. al. has also observed a ZnS self-powered, asymmetrical functional photodetector using symmetric Ag electrodes for both ends. 54 We have also performed the transient photo-response measurements with 254 nm DUV light source at moderate 5 V bias for Sample E, the results are shown in Fig. 5 (c). Fig. 5 (c) shows the transient photo-response for the optimized sample with a gradual rise and decay photoconductive behaviour. The response curve as seen from Fig. 5 (d) can be analyze properly by fitting bi-exponential equation having different time constants. The bi-exponential equation is represented as: 12ttoIIAeBewhere, Io represents the steady-state photo-current, A and B represents maximum amplitude of exponential constants, t is the time and τ1 and τ2 are the time constants in association with the exponentials. Using the above-mentioned equation,49 the rise time and decay times at 5 V bias are estimated to be 76 ms/ 113.2 ms and 25 ms/ 369 ms, respectively is shown in Fig. 5 (d). As indicated in Fig. 5 (c), we measured the PD with a DUV source of 254 nm in 5 V and it exhibits that applying moderate bias, the overall comparative on/off gain decreases. An externally applied bias voltage appears to dominate the electronic transfer into/from interface states. The increase in dark current at a large bias affected by plenty of carriers liberated from the trapping sites and a rapid drift speed which might attributed to the non-radiative charge carrier loss which thus leads to increment in the dark current. This finding may be due to self-heating at such a high rate, that would not only significantly raise charge carrier dispersion but it also increases the odds of recombination. 55-57. However, the on/off ratio of the photodetectors exceeded 103 at 0 V and a light intensity of 17 µW cm−2 as can be seen from Figure 5 (a) which is still greater than many reported devices. This issue has been widely observed within the research community, for example, Song et. al..58 reported high on/off ratio greater than 104 at 0 V and a light intensity of 17 mW cm−2, Page 12 of 26https://mc04.manuscriptcentral.com/jss-ecsECS Journal of Solid State Science and Technology123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960ccepted ManuscriptFor Review Only13 however, at higher bias they achieved on/off ratio was nearly 3.5 at bias of 0.1 V. Whereas, the slow response time indicates that the charge carrier trapping and releasing process is slow, resulting the slow-response inclination at 5 V bias. TABLE 2. Comparison of device characteristics of the β-Ga2O3 based photodetectors on different substrates reported. Structure Substrate Method Electrode R (A/W) τr/τd (s) Ref. GO FLs β-Ga2O3 Exfoliation Ni/Au 0.0124 @ 10 V 0.86/17.63 59 GO TFs Sapphire PLD Ni/Au 0.903 @ 5 V – 60 GO NWs Sapphire MOCVD Ti/Al/Ti/Au 0.000375 @ 5 V – 61 GO TFs Sapphire PAMBE Au/Ti 0.037 @ 10 V 91/36 19 GO NBs Au/Si CVD Cr/Au 39.2 @ 30 V ≪ 0.02 62 GO TFs Sapphire RF Ti/Al 70 @ 10 V – 63 GO TFs Sapphire MOCVD Au 17 @ 20 V – 64 GO TFs Sapphire MOCVD Ni/Au 20 @ 5 V 0.05/0.03 65 GO TFs Sapphire MOCVD Ti/Al 0.76 @ 20 V – 66 GO TFs β-Ga2O3 LMBE Ti/Au 0.05 @ 20 V 0.3 67 GO NBs Al2O3 CVD – 851 @ 5 V – 68 GO TFs Al2O3 RF Ti/Au 0.893 @ 10 V 4/14.2 69 GO TFs Sapphire MOCVD Ni/Au 17,000 @ 20 V 0.64/0.38 70 Page 13 of 26https://mc04.manuscriptcentral.com/jss-ecsECS Journal of Solid State Science and Technology123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960ccepted ManuscriptFor Review Only14 GO NRs FTO-glass Hydro. Ti/Au 0.00108 @ 0 V – 71 GO TFs Glass ALD Pt 45.11 @ 10 V 32/78 72 α/β-GO / graphene NRs FTO-glass Hydro. – 0.00026@ 0 V 0.54 73 GaN/ Sn: GO Al2O3 PLD Ag 3.05 @ 0 V 74 GO Diamond Spin-coating Ti/Pt/Au 12 @ -15 V 0.56/0.065 75 GO TFs Si RF Cr/Au 95.64 @ 5 V 58.3/ 34.7 This work Notes:- FLs: Flakes, TFs: thin films, LMBE: Laser molecular beam epitaxy, PLD: Pulsed laser deposition, MOCVD: metal-organic CVD, NBs: Nanobelts, RF: Radio frequency magnetron sputtering, NRs: Nanorods, Hydro.: Hydrothermal, ALD: Atomic layer deposition, NPl: Nanoplates. Conclusions Finally, we condense few concepts and ideas that may give direction to a new generation of solar-blind photodetectors. The integration of β-Ga2O3 as solar-blind photodetector with conventional high-tech Si technology is one of the most exciting applications. Incorporating seed layer with various other parameters associated with the sputtering system strongly enhanced the crystallinity of the β-Ga2O3 on Si substrate. We found a critical role of multiple parameters in heteroepitaxially grown layers to achieve high-performance devices. The use of seed layer and thermal cyclic growth has been proved extremely beneficial in enchancing the crystalline quality. The photodetector with optimised β-Ga2O3 on Si substrate yield the responsivity 95.64 AW-1 and EQE 4.73x104 % at 5 V which is higher than most of the devices being reported in various other high-cost substrates. The reported achievements of this study suggest that high quality β-Ga2O3 based deep UV photodetectors can be fabricated on Si substrate at low costs. The growth of β-Ga2O3 on Si may open new prospects to accelerate the Page 14 of 26https://mc04.manuscriptcentral.com/jss-ecsECS Journal of Solid State Science and Technology123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960ccepted ManuscriptFor Review Only15 device performance to a next level. Acknowledgments We sincerely acknowledge funding through interdisciplinary Solar PV project from Indian Institute of Technology Ropar and Defence Research and Development Organisation (ERIP/ER/201612009/M/01/1714). Page 15 of 26https://mc04.manuscriptcentral.com/jss-ecsECS Journal of Solid State Science and Technology123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960ccepted ManuscriptFor Review Only16 References 1. P. Schreiber, T. Dang, T. Pickenpaugh, G. A. Smith, P. Gehred, and C. W. Litton, in "Optoelectronics' 99-Integrated Optoelectronic Devices", p. 230-248. International Society for Optics and Photonics, 1999. 2. K. Arora, D. P. Singh, P. Fischer, and M. Kumar, Advanced Optical Materials, n/a (n/a), 2000212. 3. K. Arora, N. Goel, M. Kumar, and M. Kumar, Acs Photonics, 5 (6), 2391-2401 (2018). 4. N. Kumar, K. Arora, and M. Kumar, Journal of Physics D: Applied Physics, 52 (33), 335103 (2019). 5. X. Chen, F. Ren, S. Gu, and J. Ye, Photonics Research, 7 (4), 381-415 (2019). 6. E. Monroy, F. Omnès, and F. Calle, Semiconductor Science and Technology, 18 (4), R33 (2003). 7. Y. Qin, S. Long, Q. He, H. Dong, G. Jian, Y. Zhang, X. Hou, P. Tan, Z. Zhang, and Y. Lu, Advanced Electronic Materials, 5 (7), 1900389 (2019). 8. M. Fan, K. Liu, Z. Zhang, B. Li, X. Chen, D. Zhao, C. Shan, and D. Shen, Applied Physics Letters, 105 (1), 011117 (2014). 9. M. M. Fan, K. W. Liu, Z. Z. Zhang, B. H. Li, X. Chen, D. X. Zhao, C. X. Shan, and D. Z. Shen, Applied Physics Letters, 105 (1), 011117 (2014). 10. H. Chen, P. Yu, Z. Zhang, F. Teng, L. Zheng, K. Hu, and X. Fang, Small, 12 (42), 5809-5816 (2016). 11. Y. Teng, L. X. Song, W. Liu, Z. Y. Xu, Q. S. Wang, and M. M. Ruan, Journal of Materials Chemistry C, 4 (15), 3113-3118 (2016). 12. X. Zhou, Q. Zhang, L. Gan, X. Li, H. Li, Y. Zhang, D. Golberg, and T. Zhai, Advanced Functional Materials, 26 (5), 704-712 (2016). 13. E. Cicek, R. McClintock, C. Y. Cho, B. Rahnema, and M. Razeghi, Applied Physics Letters, 103 (19), 191108 (2013). 14. L. Li, P. S. Lee, C. Yan, T. Zhai, X. Fang, M. Liao, Y. Koide, Y. Bando, and D. Golberg, Advanced Materials, 22 (45), 5145-5149 (2010). 15. E.-J. Guo, H.-B. Lu, M. He, K.-J. Jin, and G.-Z. Yang, Appl. Opt., 49 (29), 5678-5681 (2010). 16. A. Soltani, H. A. Barkad, M. Mattalah, B. Benbakhti, J.-C. D. Jaeger, Y. M. Chong, Y. S. Zou, W. J. Zhang, S. T. Lee, A. BenMoussa, B. Giordanengo, and J.-F. Hochedez, Applied Physics Letters, 92 (5), 053501 (2008). 17. Y. Koide, M. Liao, and J. Alvarez, Diamond and Related Materials, 15 (11), 1962-1966 (2006). 18. Y. Kokubun, K. Miura, F. Endo, and S. Nakagomi, Applied physics letters, 90 (3), 031912 (2007). 19. T. Oshima, T. Okuno, and S. Fujita, Japanese Journal of Applied Physics, 46 (11R), 7217 (2007). 20. H. Ito, K. Kaneko, and S. Fujita, Japanese Journal of Applied Physics, 51 (10R), 100207 (2012). 21. Y. Liao, S. Jiao, S. Li, J. Wang, D. Wang, S. Gao, Q. Yu, and H. Li, CrystEngComm, 20 (1), 133-139 (2018). 22. S. Jiao, H. Lu, X. Wang, Y. Nie, D. Wang, S. Gao, and J. Wang, ECS Journal of Solid State Science and Technology, 8 (7), Q3086 (2019). 23. A. Hassanien, A. Atta, M. El-Nahass, S. I. Ahmed, A. A. Shaltout, A. M. Al-Baradi, A. Alodhayb, and A. Kamal, Optical and Quantum Electronics, 52 (4), 1-16 (2020). 24. M.-Q. Li, N. Yang, G.-G. Wang, H.-Y. Zhang, and J.-C. Han, Applied Surface Science, 471 694-702 (2019). 25. C.-C. Wang, S.-H. Yuan, S.-L. Ou, S.-Y. Huang, K.-Y. Lin, Y.-A. Chen, P.-W. Hsiao, and D.-S. Wuu, Journal of Alloys and Compounds, 765 894-900 (2018). Page 16 of 26https://mc04.manuscriptcentral.com/jss-ecsECS Journal of Solid State Science and Technology123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960ccepted ManuscriptFor Review Only17 26. S. Li, S. Jiao, D. Wang, S. Gao, and J. Wang, Journal of Alloys and Compounds, 753 186-191 (2018). 27. H. Li, P. Chen, S. Yuan, T. Huang, S. Zhang, and D. Wuu, IEEE Photonics Journal, 11 (6), 1-8 (2019). 28. H. Li, S.-H. Yuan, T.-M. Huang, H.-J. Chen, F.-H. Lu, S. Zhang, and D.-S. Wuu, Journal of Alloys and Compounds, 823 153755 (2020). 29. H. Wu, Y. Huang, Y. Zhi, X. Wang, X. Chu, Z. Chen, P. Li, Z. Wu, and W. Tang, Optical Materials Express, 9 (3), 1394-1403 (2019). 30. S. Nakagomi and Y. Kokubun, Journal of Crystal Growth, 349 (1), 12-18 (2012). 31. X. Guo, N. Hao, D. Guo, Z. Wu, Y. An, X. Chu, L. Li, P. Li, M. Lei, and W. Tang, Journal of Alloys and Compounds, 660 136-140 (2016). 32. Y. Tang, Y. Zhu, X. Ma, A. Y. Borisevich, A. N. Morozovska, E. A. Eliseev, W. Wang, Y. Wang, Y. Xu, and Z. Zhang, Science, 348 (6234), 547-551 (2015). 33. K. J. Choi, M. Biegalski, Y. Li, A. Sharan, J. Schubert, R. Uecker, P. Reiche, Y. Chen, X. Pan, and V. Gopalan, Science, 306 (5698), 1005-1009 (2004). 34. L. Dupré, T. Gorisse, A. L. Lebranchu, T. Bernardin, P. Gentile, H. Renevier, and D. Buttard, Nanoscale Research Letters, 8 (1), 123 (2013). 35. K.-Y. Chan and B.-S. Teo, Microelectronics Journal, 37 (10), 1064-1071 (2006). 36. K.-Y. Chan and B.-S. Teo, Journal of Materials Science, 40 (22), 5971-5981 (2005). 37. M. Santhiya, K. S. Pugazhvadivu, L. Balakrishnan, and K. Tamilarasan, AIP Conference Proceedings, 1731 (1), 080077 (2016). 38. A. Ishizaka and Y. Shiraki, Journal of The Electrochemical Society, 133 (4), 666-671 (1986). 39. S. Hu, P. W. Leu, A. F. Marshall, and P. C. McIntyre, Nat Nano, 4 (10), 649-653 (2009). 40. T. K. O. Vu, D. U. Lee, and E. K. Kim, Journal of Alloys and Compounds, 806 874-880 (2019). 41. X. Liu, C. Yue, C. Xia, and W. Zhang, Chinese Physics B, 25 (1), 017201 (2015). 42. L. Dong, R. Jia, B. Xin, B. Peng, and Y. Zhang, Scientific Reports, 7 40160 (2017). 43. Nisika, K. Kaur, K. Arora, A. H. Chowdhury, B. Bahrami, Q. Qiao, and M. Kumar, Journal of Applied Physics, 126 (19), 193104 (2019). 44. K. Kaur, K. Arora, B. Behzad, Q. Qiao, and M. Kumar, Nanotechnology, 30 (6), 065706 (2018). 45. W. Y. Kong, G. A. Wu, K. Y. Wang, T. F. Zhang, Y. F. Zou, D. D. Wang, and L. B. Luo, Advanced materials, 28 (48), 10725-10731 (2016). 46. D. Guo, H. Liu, P. Li, Z. Wu, S. Wang, C. Cui, C. Li, and W. Tang, ACS applied materials & interfaces, 9 (2), 1619-1628 (2017). 47. S. Rajamani, K. Arora, A. Konakov, A. Belov, D. Korolev, A. Nikolskaya, A. Mikhaylov, S. Surodin, R. Kryukov, and D. Nikolitchev, Nanotechnology, 29 (30), 305603 (2018). 48. S. Rajamani, K. Arora, A. Belov, D. Korolev, A. Nikolskaya, Y. Usov, D. Pavlov, A. Mikhaylov, D. Tetelbaum, and M. Kumar, IEEE Sensors Journal, 18 (10), 4046-4052 (2018). 49. N. Kumar, K. Arora, and M. Kumar, Semiconductor Science and Technology, 34 (5), 055004 (2019). 50. K. Arora, V. Kumar, and M. Kumar, arXiv preprint arXiv:1809.10724, (2018). 51. H. He, R. Orlando, M. A. Blanco, R. Pandey, E. Amzallag, I. Baraille, and M. Rérat, Physical Review B, 74 (19), 195123 (2006). 52. M. Passlack, E. Schubert, W. Hobson, M. Hong, N. Moriya, S. Chu, K. Konstadinidis, J. Mannaerts, M. Schnoes, and G. Zydzik, Journal of applied physics, 77 (2), 686-693 (1995). 53. S. Cui, Z. Mei, Y. Zhang, H. Liang, and X. Du, Advanced Optical Materials, (2017). 54. Q. An, X. Meng, K. Xiong, and Y. Qiu, Scientific Reports, 7 (1), 4885 (2017). 55. D. Guo, H. Shi, Y. Qian, M. Lv, P. Li, Y. Su, Q. Liu, K. Chen, S. Wang, and C. Cui, Semiconductor Page 17 of 26https://mc04.manuscriptcentral.com/jss-ecsECS Journal of Solid State Science and Technology123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960ccepted ManuscriptFor Review Only18 Science and Technology, 32 (3), 03LT01 (2017). 56. P. Li, H. Shi, K. Chen, D. Guo, W. Cui, Y. Zhi, S. Wang, Z. Wu, Z. Chen, and W. Tang, Journal of Materials Chemistry C, 5 (40), 10562-10570 (2017). 57. Z. Wu, L. Jiao, X. Wang, D. Guo, W. Li, L. Li, F. Huang, and W. Tang, Journal of Materials Chemistry C, 5 (34), 8688-8693 (2017). 58. W. Song, X. Wang, H. Chen, D. Guo, M. Qi, H. Wang, X. Luo, X. Luo, G. Li, and S. Li, Journal of Materials Chemistry C, 5 (44), 11551-11558 (2017). 59. S. Oh, C.-K. Kim, and J. Kim, Acs Photonics, 5 (3), 1123-1128 (2017). 60. F.-P. Yu, S.-L. Ou, and D.-S. Wuu, Optical Materials Express, 5 (5), 1240-1249 (2015). 61. W. Weng, T. Hsueh, S.-J. Chang, G. Huang, and S. Hung, IEEE transactions on nanotechnology, 10 (5), 1047-1052 (2011). 62. L. Li, E. Auer, M. Liao, X. Fang, T. Zhai, U. K. Gautam, A. Lugstein, Y. Koide, Y. Bando, and D. Golberg, Nanoscale, 3 (3), 1120-1126 (2011). 63. L. Qian, X. Liu, T. Sheng, W. Zhang, Y. Li, and P. Lai, Aip Advances, 6 (4), 045009 (2016). 64. G. Hu, C. Shan, N. Zhang, M. Jiang, S. Wang, and D. Shen, Optics express, 23 (10), 13554-13561 (2015). 65. D.-S. Wuu, S.-L. Ou, R.-H. Horng, P. Ravadgar, T.-Y. Wang, and H.-Y. Lee, in "Oxide-based Materials and Devices III", Vol. 8263, p. 826317. International Society for Optics and Photonics, 2012. 66. P. Guo, J. Xiong, X. Zhao, T. Sheng, C. Yue, B. Tao, and X. Liu, Journal of Materials Science: Materials in Electronics, 25 (8), 3629-3632 (2014). 67. Q. Feng, L. Huang, G. Han, F. Li, X. Li, L. Fang, X. Xing, J. Zhang, W. Mu, and Z. Jia, IEEE Transactions on Electron Devices, 63 (9), 3578-3583 (2016). 68. R. Zou, Z. Zhang, Q. Liu, J. Hu, L. Sang, M. Liao, and W. Zhang, Small, 10 (9), 1848-1856 (2014). 69. Y. Peng, Y. Zhang, Z. Chen, D. Guo, X. Zhang, P. Li, Z. Wu, and W. Tang, IEEE Photonics Technology Letters, 30 (11), 993-996 (2018). 70. F. Alema, B. Hertog, O. Ledyaev, D. Volovik, G. Thoma, R. Miller, A. Osinsky, P. Mukhopadhyay, S. Bakhshi, and H. Ali, Physica status solidi (a), 214 (5), 1600688 (2017). 71. S. Wang, K. Chen, H. Zhao, C. He, C. Wu, D. Guo, N. Zhao, G. Ungar, J. Shen, and X. Chu, RSC advances, 9 (11), 6064-6069 (2019). 72. S. H. Lee, S. B. Kim, Y.-J. Moon, S. M. Kim, H. J. Jung, M. S. Seo, K. M. Lee, S.-K. Kim, and S. W. Lee, Acs Photonics, 4 (11), 2937-2943 (2017). 73. C. Wu, C. He, D. Guo, F. Zhang, P. Li, S. Wang, A. Liu, F. Wu, and W. Tang, Materials Today Physics, 12 100193 (2020). 74. D. Guo, Y. Su, H. Shi, P. Li, N. Zhao, J. Ye, S. Wang, A. Liu, Z. Chen, and C. Li, ACS nano, 12 (12), 12827-12835 (2018). 75. H. Kim, S. Tarelkin, A. Polyakov, S. Troschiev, S. Nosukhin, M. Kuznetsov, and J. Kim, ECS Journal of Solid State Science and Technology, 9 (4), 045004 (2020). Page 18 of 26https://mc04.manuscriptcentral.com/jss-ecsECS Journal of Solid State Science and Technology123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960ccepted ManuscriptFor Review Only19 List of Figures Figure 1 Figure 1. Temperature profile during growth for different samples mentioned in Table 1. (a) Sample D (b) Sample C (c) Sample A, Sample B, and Sample E (d) Schematic representation of the fabrication of β-Ga2O3 under various optimised conditions. Page 19 of 26https://mc04.manuscriptcentral.com/jss-ecsECS Journal of Solid State Science and Technology123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960ccepted ManuscriptFor Review Only20 Figure 2 Figure 2 (a) GIXRD omega-2 scans of different samples grown on p-type Si (100) using growth conditions same as mentioned in table above. (b)Atomic force microscopy (3.5×3.5μm2) two-dimensional images (A) Sample A (B) Sample B (C) Sample C (D) Sample D and (E) Sample E Page 20 of 26https://mc04.manuscriptcentral.com/jss-ecsECS Journal of Solid State Science and Technology123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960ccepted ManuscriptFor Review Only21 Figure 3 Figure 3. Shows the optical transmission spectra of samples A, B, C, D and E grew on quartz substrate Inset shows the combined tauc plot of the samples A, B, C, D, and E. Page 21 of 26https://mc04.manuscriptcentral.com/jss-ecsECS Journal of Solid State Science and Technology123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960ccepted ManuscriptFor Review Only22 Figure 4 Figure 4. (a) Linear I-V characteristics of the β-Ga2O3 device under dark and 254 nm illumination (inset shows the schematic of characterization of fully fabricated β-Ga2O3 device) (b) Linear dynamic range of photodetector at different bias. (c) Comparison of photoresponsivity versus wavelengths at different biases voltages and (d) Comparison of external quantum efficiency versus wavelengths at different biases. Page 22 of 26https://mc04.manuscriptcentral.com/jss-ecsECS Journal of Solid State Science and Technology123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960ccepted ManuscriptFor Review Only23 Figure 5 Figure 5. (a) Transient current response of sample E at 0 V bias by switching 254 nm ON/OFF periodically in a repeatable manner. (b) Experimental fitting of transient response to calculate rise and decay time of the device. (c) Transient current response of sample E at 5 V bias by switching 254 nm ON/OFF periodically in a repeatable manner (d) Experimental fitting of transient response to calculate rise and decay time of the device. Page 23 of 26https://mc04.manuscriptcentral.com/jss-ecsECS Journal of Solid State Science and Technology123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960ccepted ManuscriptFor Review Only24 List of tables TABLE I. The growth parameters of gallium oxide film at four different conditions [deposition pressure, sputtering power, without substrate heating and without seed layer] and the film at the optimised parameter Growth Parameters S. no. Sample Power (W) PAr (mTorr) Substrate heating Seed–layer-assisted growth 1 A 100 5 Yes Yes 2 B 75 3 Yes Yes 3 C 100 3 No Yes 4 D 100 3 Yes No 5 E 100 3 Yes Yes Page 24 of 26https://mc04.manuscriptcentral.com/jss-ecsECS Journal of Solid State Science and Technology123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960ccepted ManuscriptFor Review Only25 TABLE 2. Comparison of device characteristics of the β-Ga2O3 based photodetectors on different substrates reported. Structure Substrate Method Electrode R (A/W) τr/τd (s) Ref. GO FLs β-Ga2O3 Exfoliation Ni/Au 0.0124 @ 10 V 0.86/17.63 59 GO TFs Sapphire PLD Ni/Au 0.903 @ 5 V – 60 GO NWs Sapphire MOCVD Ti/Al/Ti/Au 0.000375 @ 5 V – 61 GO TFs Sapphire PAMBE Au/Ti 0.037 @ 10 V 91/36 19 GO NBs Au/Si CVD Cr/Au 39.2 @ 30 V ≪ 0.02 62 GO TFs Sapphire RF Ti/Al 70 @ 10 V – 63 GO TFs Sapphire MOCVD Au 17 @ 20 V – 64 GO TFs Sapphire MOCVD Ni/Au 20 @ 5 V 0.05/0.03 65 GO TFs Sapphire MOCVD Ti/Al 0.76 @ 20 V – 66 GO TFs β-Ga2O3 LMBE Ti/Au 0.05 @ 20 V 0.3 67 GO NBs Al2O3 CVD – 851 @ 5 V – 68 GO TFs Al2O3 RF Ti/Au 0.893 @ 10 V 4/14.2 69 GO TFs Sapphire MOCVD Ni/Au 17,000 @ 20 V 0.64/0.38 70 GO NRs FTO-glass Hydro. Ti/Au 0.00108 @ 0 V – 71 GO TFs Glass ALD Pt 45.11 @ 10 V 32/78 72 Page 25 of 26https://mc04.manuscriptcentral.com/jss-ecsECS Journal of Solid State Science and Technology123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960ccepted ManuscriptFor Review Only26 α/β-GO / graphene NRs FTO-glass Hydro. – 0.00026@ 0 V 0.54 73 GaN/ Sn: GO Al2O3 PLD Ag 3.05 @ 0 V 74 GO Diamond Spin-coating Ti/Pt/Au 12 @ -15 V 0.56/0.065 75 GO TFs Si RF Cr/Au 95.64 @ 5 V 58.3/ 34.7 This work Notes:- FLs: Flakes, TFs: thin films, LMBE: Laser molecular beam epitaxy, PLD: Pulsed laser deposition, MOCVD: metal-organic CVD, NBs: Nanobelts, RF: Radio frequency magnetron sputtering, NRs: Nanorods, Hydro.: Hydrothermal, ALD: Atomic layer deposition, NPl: Nanoplates. Page 26 of 26https://mc04.manuscriptcentral.com/jss-ecsECS Journal of Solid State Science and Technology123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960ccepted ManuscriptView publication stats | RESEARCH GATE |
43Case reportIctal asystole with reduced cardiac sympathetic function in new-onset symptomatic epilepsyJulia Matzen 1*, Friedhelm C. Schmitt 1*, Michael C. Kreissl 2, Jürgen Voges 3,4, Hans-Jochen Heinze 1,4, Imke Galazky 11 Department of Neurology, University of Magdeburg, Magdeburg, Germany2 Department of Radiology and Nuclear Medicine, University Hospital Magdeburg, Magdeburg, Germany3 Department of Stereotactic Neurosurgery, University of Magdeburg, Magdeburg, Germany4 Leibniz Institute for Neurobiology (LIN), Magdeburg, GermanyCorrespondenceDr. med. Julia Matzen, MDOtto-von-Guericke University, Department of Neurology Leipziger Str. 44, D-39120 Magdeburg, Germanye-mail: julia@matzen-net.de Phone +49 391 67 15 001Fax +49 391 67 15 032SUMMARYIntroduction. So far, cardiac sympathetic dysfunction, demonstrated in pharmacoresistant epilepsy pa-tients with ictal bradycardia or asystole by I-123 metaiodobenzylguanidine (I-123 MIBG) imaging has been attributed to repeated occurrence of seizure activity.Aim. Discussion of the mechanisms of cardiac sympathetic dysfunction associated with ictal asystole un-der consideration of a case with new onset epilepsy.Materials and methods. We describe the occurrence of a cardiac asystole during a complex-partial sei-zure in an antiepileptic-drug-naïve patient with new-onset symptomatic epilepsy.Results. MIBG imaging showed reduced tracer accumulation in cardiac sympathetic nerve endings in this patient with right parietotemporal glioblastoma.Discussion and Conclusion. To our knowledge, this is the first report of impaired cardiac sympathetic function in new-onset symptomatic epilepsy without antiepileptic drug treatment. MIBG imaging should be considered in patients with ictal bradycardia or asystole.Keywords: ictal asystole • MIBG imaging • symptomatic epilepsy • glioblastomaReceived April 7, 2018 Accepted for publication on-line February 8, 2019 Published on-line February 16, 2019INTRODUCTIONInvolvement of the autonomic nervous system often occurs during epileptic seizures. Frequent autonom-ic symptoms include for example sweating, pupil dil-atation or tachycardia. In contrast to the latter, ictal bradycardias or even asystoles are rare. In retrospec-tive analyses of data from presurgical video-electro-encephalography-monitoring, ictal asystoles of 5 to 60s duration were found in 0.27 to 0.59% of patients (Kerling et al., 2009; Rocamora et al., 2003; Schuele et al., 2007). Nevertheless, ictal bradycardias and asysto-les are of relevance for the affected patients, since they usually go along with syncopes, often accompanied by falls and risk of injury. A higher risk of sudden unex-pected death in epilepsy (SUDEP) in these patients has been widely discussed but could not be proven yet. The pathophysiology underlying ictal bradyarrhythmias is not entirely clear, however, a key role is attributed to an imbalance in the autonomic nerve system (Sforza et al., * Both authors contributed equally.Copyright and photocopying by Foundation of Epileptology, 2019Journal of Epileptology • 2019 • 27 • 43–47 • 10.21307/jepil-2019-00144Julia Matzen et al.2014). In patients with long-lasting temporal lobe ep-ilepsy (TLE), however, cardiac sympathetic denerva-tion has been shown by means of I-123 metaiodoben-zylguanidine (MIBG) imaging, a scintigraphic method using an analogue of norepinephrine, that is retained in sympathetic nerve endings. Among the investigated patients with TLE, the finding was significantly more pronounced in those experiencing asystoles with their seizures (Druschky et al., 2001; Kerling et al., 2009).AIMCase presentation and discussion of the mechanisms of cardiac sympathetic dysfunction associated with ic-tal asystole under consideration of a case with new on-set epilepsy.MATERIAL AND METHODHere, we report on an antiepileptic-drug-naïve patient suffering ictal asystoles as the sole seizure manifestation of new-onset symptomatic epilepsy due to glioblasto-ma of the right parietotemporal lobe, in whom MIBG scanning demonstrated marked sympathetic denerva-tion of the heart.RESULTSCaseA 71-year old male patient was admitted to our de-partment with recurrent glioblastoma WHO °IV of the right parietotemporal lobe for brachytherapy. The tumour was diagnosed 11 months before by MRI and histopathological analysis after a hemianopsia to the left side had been found during a routine ophthalmo-logical examination. First-line therapy had consisted of incomplete surgical resection with implantation of carmustine wafers and subsequent radiotherapy with concomitant and adjuvant temozolomide chemother-apy. The decision for implantation of iodine-125-seeds was made because of further progression of the tumour. Clinical neurological and neuropsychological exami-nation revealed homonymous hemianopsia to the left, reduced alertness and mild to moderate executive dys-function. There was no history of epileptic seizures in terms of overt clinical seizure manifestation. Howev-er, the patient reported on two events with loss of con-sciousness preceded by nausea and dizziness shortly be-fore the diagnosis of tumour. Otherwise, medical his-tory revealed no relevant diseases, especially no his-tory of cardiovascular or Parkinson’s disease; the pa-tient was not taking any regular medication besides the above mentioned chemotherapeutics.Figure 1. (A) EEG showing right temporal (T4) delta slowing (B) a seizure pattern evolves consisting of rhythmic 3/s activity, no clinical abnormalities (C) 50 s later: ongoing seizure pattern, patient complains an olfactory sensation, bradycardia (D) asystolia, generalised EEG flattening, patient unconscious (E) asystolia ends spontaneously after 23 s, clinically hypoxic myoclonus (F) normalisation of EEG pattern, patient awake.45Sympathetic cardiac function in ictal asystoleDuring preoperative routine-EEG, the patient sud-denly complained about a funny smell, then became un-responsive with staring followed by loss of conscious-ness with loose muscle tone and single bilateral asym-metric jerks before awakening. The patient was quick-ly reoriented and was amnestic for the episode. The EEG (fig. 1) initially showed continuous right fronto-temporal delta-slowing maximal at F4 without epilep-tiform discharges, ECG showed sinus rhythm with 72 bpm. Subsequently, rhythmic right temporal theta-ac-tivity developed 10 s before clinical seizure manifesta-tion which included bradycardia. The patient became asystolic 66 s after clinical seizure onset, cardiac ac-tivity reoccurred after 23 s. Flattening of the EEG last-ed 26 s (fig. 1).Subsequent cardiological examination including chest X-ray, 24 h Holter-ECG and transthoracal echo-cardiography was unremarkable. MIBG imaging re-vealed a low myocardial MIBG uptake (heart/mediasti-num ratio 1.6 after 15 min and 1.8 after 4 hours post injection (fig. 2) reflecting a decreased post-ganglion-ic cardiac catecholamine reuptakeAfter the patient exhibited another asystole requir-ing cardiac resuscitation during his hospital stay, a car-diac pacemaker was implanted and antiepileptic drug therapy was initiated (levetiracetam 1500 mg daily). The patient did not experience any asystoles or other seizure manifestations until his death 8 months later.DISCUSSIONIctal asystole is a rare manifestation of seizures most-ly of temporal origin (Rocamora et al., 2003; Schuele et al., 2007; Tenyi et al., 2017). The exact pathomecha-nisms underlying this seizure mediated cardiac brad-yarrhythmia are unclear. However, in patients with temporal lobe epilepsy (TLE), cardiac sympathetic in-nervation has been shown to be compromised using MIBG imaging (Druschky et al., 2001; Kerling et al., 2009) a finding that is significantly more pronounced in those TLE patients exhibiting ictal asystoles (Kerling et al., 2009). According to the literature there are two hypothesis concerning the pathophysiological mean-ing of this finding: It has been proposed that enhanced sympathetic outflow from brain regions activated dur-ing seizure propagation leads to a compensatory down-regulation of sympathetic nerve endings (Druschky et al., 2001, Kerling et al., 2009). Also, an inherent distur-bance of cardiac sympathetic innervation accompany-ing neuronal migration disorders in TLE patients has been discussed (v. Manitius-Robeck et al., 1998). How-soever caused, deficiency of cardiac sympathetic inner-vation probably limits counterregulation of vagal acti-vation thus facilitating cardiac bradyarrythmias.In its clinical appearance, the case presented here is typical for ictal asystoles as depicted in the literature: the patient experiences an aura, in this case in the form of a smelling sensation, followed by a complex-par-Figure 2. Planar (123)I metaiodobenzylguanidine scans with low myocardial tracer uptake (heart/ mediastinum ratio 1.6 after 15 min and 1.8 after 4 hours post injection). This is often found in deficient post-ganglionic cardiac catecholamine uptake in idiopathic Parkinson´s disease (left: anterior view; right: posterior view).46Julia Matzen et al.tial seizure with staring and unresponsiveness. Both the semiology and the ictal EEG (fig. 1) are suggestive for a temporal lobe seizure. He then loses conscious-ness due to cerebral hypoperfusion and subsequently myoclonic jerks before awakening as an expression of cerebral reperfusion (Duplyakov et al., 2014). MIBG scanning in our patient showed a low post-ganglion-ic sympathetic innervation in a comparable degree as described in the literature (Druschky et al., 2001; Ker-ling et al., 2009). In the aforementioned studies, how-ever, patients with ictal asystole and pathological find-ings on MIBG imaging had been suffering from chron-ic pharmacoresistant epilepsy either of unknown etiol-ogy or due to long-lasting epileptogenic lesions such as hippocampal sclerosis or low grade tumours. In con-trast, the patient reported here suffered from new-on-set symptomatic epilepsy, which had not been treated with antiepileptic drugs yet. Contrary to our case re-port, Tenyi et al. (2017) found in their metaanalysis of 156 patients that the group of new onset ictal asystole patients (defined as occurrence of ictal asystole with-in a year) had a significant higher preponderance of fe-male patients and patients with heart disease.For temozolomide, which the patient had been tak-ing in the course of radiochemotherapy, cardiac side effects have not been described. There is one patient reported by Druschky et al. (2001) who had been suf-fering from TLE due to hippocampal gliosis for only one year, but in contrast to our patient she had a high seizure frequency of 15/month and was taking carba-mazepine 1200 mg daily, a drug that is known to po-tentially cause bradycardia and other cardiac arrhyth-mias. In the other 21 patients-described by Druschky et al. (2001) – mean duration of epilepsy was 21.8 years, and 10 patients were taking carbamazepine. The find-ing of a compromised sympathetic innervation of the heart in a patient with a newly acquired epilepsy with ictal asystoles challenges both of the above mentioned hypotheses: neither compensatory downregulation of sympathetic nerve endings due to repeated, seizure-driven, enhanced cerebral sympathetic outflow nor combined malformations of the peripheral and central-nervous system can explain the occurrence of ictal asys-toles on the basis of post-ganglionic sympathetic den-ervation of the heart in this case. Furthermore, a medi-cation side effect is excluded in this drug-naive patient. Interestingly, only 18% of the patients with new-onset ictal asystole in the cohort of Tenyi et al. (2017), were pretreated with AEDs. Also, this group had a signifi-cant higher proportion of epileptogenic lesions in MRI, so that it could be speculated that in these acute symp-tomatic subgroups of patients ictal asystole is a marker of different pathological mechanism compared to the “typical”, long-lasting epilepsy patient, who are more prone to SUDEP.CONCLUSIONDysfunctional cardiac sympathetic innervation should also be considered in patients with sporadic seizure oc-currence, with a recent seizure onset and without intake of drugs with potential cardiac side effects. For further discrimination of the pathomechanism of cardial au-tonomic seizures, MIBG imaging can be helpful in pa-tients with this seizure type.CONFLICT OF INTERESTSAll authors have no conflict of interest: neither the au-thors nor the author’s institution have a financial or other relationship with other people or organizations that may have influenced the presented work.ACKNOWLEDGEMENTSWe thank Dr. med. J. Ruf for providing the results and images of figure 2.REFERENCESDuplyakov D., Golovina G., Lyukshina N., Surkova E., El-ger C.E., Surges R.: Syncope, seizure-induced bradycardia and asystole: Two cases and review of clinical and pathophysiologi-cal features. Seizure, 2014, 23: 506–511. doi: 10.1016/j.seizu-re.2014.03.004Druschky A., Hilz M.J., Hopp P., Platsch G., Radespiel-Tröger M., Druschky K. et al.: Interictal cardiac autonomic dysfunc-tion in temporal lobe epilepsy demonstrated by (123)I metaio-dobenzylguanidine-SPECT. Brain, 2001, 124: 2372–2382. doi: 10.1093/brain/124.12.2372Kerling F., Dütsch M., Linke R., Kuwert T., Stefan H., Hilz M.J.: Relation between ictal asystole and cardiac sympathe-tic dysfunction shown by MIBG-SPECT. Acta Neurol. Scand., 2009, 120: 123–129. doi: 10.1111/j.1600-0404.2008.01135.xManitius-Robeck S.v., Schüler P., Feistel H., Platsch G., Stefan H.: Iktuale Synkopen – Kardiale sympathische Innerva-tionsstörung als Ursache? [Ictal syncopes. Cardiac sympathe-tic innervation disorder as the etiology?]. Nervenarzt, 1998, 69: 712–716.Rocamora R., Kurthen M., Lickfett L., von Oertzen J., Elger C.E.: Cardiac asystole in epilepsy: Clinical and neurophysiolo-gical features. Epilepsia, 2003, 44: 179–185.47Sympathetic cardiac function in ictal asystoleSchuele S.U., Bermeo A.C., Alexopoulos A.V., Locatelli E.R., Burgess R.C., Dinner D.S., Foldvary-Schaefer N.: Vid-eo-electrographic and clinical features in patients with ic-tal asystole. Neurology, 2007, 69: 434–441. doi: 10.1212/01.wnl.0000266595.77885.7fSforza E., Pichot V., Gschwind M., Roche F., Maeder-In-gvar M.: Bradycardia and asystole during generalised interic-tal EEG discharges. Epileptic Disord., 2014, 16: 506–509. doi: 10.1684/epd.2014.0710Tenyi D., Gyimesi C., Kupo P., Horvath R., Bone B., Barsi P. et al.: Ictal asystole: A systematic review. Epilepsia, 2017, 58: 356–362. doi: 10.1111/epi.13644• | RESEARCH GATE |
Amyloid-b peptide 37, 38 and 40 individually andcooperatively inhibit amyloid-b 42 aggregation†Gabriel A. Braun,‡a Alexander J. Dear,abcd Kalyani Sanagavarapu,§aHenrik Zetterbergefgh and Sara Linse*aThe pathology of Alzheimer's disease is connected to the aggregation of b-amyloid (Ab) peptide, which invivo exists as a number of length-variants. Truncations and extensions are found at both the N- and C-termini, relative to the most commonly studied 40- and 42-residue alloforms. Here, we investigate theaggregation of two physiologically abundant alloforms, Ab37 and Ab38, as pure peptides and in mixtureswith Ab40 and Ab42. A variety of molar ratios were applied in quaternary mixtures to investigate whethera certain ratio is maximally inhibiting of the more toxic alloform Ab42. Through kinetic analysis, we showthat both Ab37 and Ab38 self-assemble through an autocatalytic secondary nucleation reaction to formfibrillar b-sheet-rich aggregates, albeit on a longer timescale than Ab40 or Ab42. Additionally, we showthat the shorter alloforms co-aggregate with Ab40, affecting both the kinetics of aggregation and theresulting fibrillar ultrastructure. In contrast, neither Ab37 nor Ab38 forms co-aggregates with Ab42;however, both short alloforms reduce the rate of Ab42 aggregation in a concentration-dependentmanner. Finally, we show that the aggregation of Ab42 is more significantly impeded by a combination ofAb37, Ab38, and Ab40 than by any of these alloforms independently. These results demonstrate that theaggregation of any given Ab alloform is significantly perturbed by the presence of other alloforms,particularly in heterogeneous mixtures, such as is found in the extracellular fluid of the brain.IntroductionAlzheimer's disease (AD), a progressive neurodegenerativedisease, is the most common cause of dementia worldwide,afflicting nearly 50 million people.1 Although symptomatictreatments for AD exist, there are currently no available meansof slowing or reversing the progression of the disease.2–4Recently, the anti-amyloid b (Ab) antibody aducanumab5–7 wasapproved for clinical use by the US Food and Drug Adminis-tration (FDA),8 and other promising antibody candidates are inthe pipeline, but their clinical efficacy remains uncertain. Thisis due, in part, to the fact that the etiology of AD remains poorlyunderstood. The most common theory of the progression of ADis the amyloid cascade hypothesis, which contends that aggre-gation of amyloid b peptide (Ab) is the critical pathologicalevent responsible for triggering the onset of the disease, withself-assembly of tau an important second step.9,10 In particular,dispersed Ab oligomers, rather than the deposited plaques, arewidely understood to be the primary cytotoxic forms of theseaggregates,3,11 although brils may still play a signicant role,directly12 or indirectly through generation of toxic oligomersthrough secondary nucleation.13,14 While the cascade model hasbeen challenged in recent years,15–17 it is nevertheless commonlyaccepted that the aggregation of Ab plays a critical part in theinitiation and progression of AD.18–20 Understanding the Abaggregation process, its underlying microscopic steps, and theirconnection with toxicity, is essential for the development ofeffective treatments and better diagnostic tools.21Ab is produced from the transmembrane amyloid precursorprotein (APP). In the so-called amyloidogenic pathway for pro-cessing of APP, sequential cleavages of the protein by b- and g-secretases result in the release of Ab into the extracellularuid.22,23 There is some variability in the cleavage sites by bothsecretases, leading to the presence of Ab length-variants, oraBiochemistry and Structural Biology, Lund University, Lund, Sweden. E-mail: sara.linse@biochemistry.lu.sebDepartment of Cell Biology, Harvard Medical School, Boston, MA, USAcPaulson School of Engineering and Applied Science, Harvard University, Cambridge,MA, USAdDepartment of Chemistry, University of Cambridge, Cambridge, UKeDepartment of Psychiatry and Neurochemistry, Institute of Neuroscience andPhysiology, The Sahlgrenska Academy at the University of Gothenburg, M¨olndal,SwedenfClinical Neurochemistry Laboratory, Sahlgrenska University Hospital, M¨olndal,SwedengDepartment of Neurodegenerative Disease, UCL Institute of Neurology, Queen Square,London, UKhUK Dementia Research Institute at UCL, London, UK† Electronicsupplementaryinformation(ESI)available.SeeDOI:10.1039/d1sc02990h‡ Current address: University of California, San Francisco, Program in Chemistryand Chemical Biology, San Francisco, CA, United States.§ Current address: Wren Therapeutics Sweden AB, Lund, Sweden.Cite this: Chem. Sci., 2022, 13, 2423All publication charges for this articlehave been paid for by the Royal Societyof ChemistryReceived 2nd June 2021Accepted 22nd January 2022DOI: 10.1039/d1sc02990hrsc.li/chemical-science© 2022 The Author(s). Published by the Royal Society of ChemistryChem. Sci., 2022, 13, 2423–2439 | 2423ChemicalScienceEDGE ARTICLEOpen Access Article. Published on 07 February 2022. Downloaded on 3/14/2022 5:09:18 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence.View Article OnlineView Journal | View Issuealloforms, that are generally between 37 and 43 residueslong.24–27 Of these alloforms, the 40-residue length-variant APP672-711, Ab1–40, is the predominant form found in vivo, makingup almost 60% the Ab peptide present in cerebrospinal uid(CSF).28,29 The next most abundant alloforms are Ab1–38, Ab1–42,and Ab1–37, which make up approximately 15%, 10% and 8% ofAb in CSF, respectively (although the concentration of Ab1–42 inCSF decreases signicantly in AD patients, which is proposed tobe the result of peptide sequestration in plaques30,31). N-terminal length variants are observed both with extended25and truncated32 N-termini, in which case extension retards33and truncation accelerates34 aggregation.Of the C-terminal Ab length-variants found most commonlyin vivo, the 40- and 42-residue alloforms are the most aggrega-tion prone and are also found in the highest abundance in thecharacteristic AD plaques.35 As a result, the aggregation of thesepeptides has been extensively characterized, while the aggre-gationofshorter,lessaggregation-pronealloformshasremained relatively less studied.36 Using a chemical kineticsapproach, the microscopic steps underlying the in vitro aggre-gation of Ab have previously been identied from global anal-ysesoflargesetsofdataoverrangesofpeptideconcentration,37,38 showing that Ab peptides, ending at residue40 and 42, respectively, aggregate through a nucleation-and-growth pathway in which the formation of new aggregatesoccurs predominantly through the autocatalytic secondarynucleation on the surface of existing brils.14 However, theaggregation behavior of Ab peptides is highly sensitive to solu-tion conditions, including ionic strength39,40 and pH,41 and tothepresenceofforeignsurfaces,suchasphospholipidmembranes42,43 and nanoparticles.44,45 Indeed, in the complexmilieu of cerebrospinal uid, it has been found that, while Ab42aggregates through a secondary-nucleation-dominated mecha-nism, the rate of this bril-catalyzed nucleation is signicantlyreduced relative to that of the peptide in pure buffer.46,47 Addi-tionally, the aggregation of Ab peptides is perturbed by the co-existence of Ab peptides of different lengths. For example, ithas previously been shown that, while Ab42 and Ab40 aggregateto form homomolecular brils, monomeric Ab42 strongly cata-lyzes the aggregation of Ab40.48 Furthermore, Ab42 has beenshown to both co-aggregate with and cross-seed Ab peptideswith N-terminal extensions ranging from 5–40 residues inlength.33 Consequently, it is important to account for theinteractions between Ab alloforms when considering the in vivobehavior of these peptides.Although Ab alloforms ending before residue 40 are lesscommonly found in plaques in AD patients and are not directlyimplicated in the onset of AD, these shorter alloforms arenevertheless important factors to consider in the disease-related aggregation of longer alloforms. Formation of oligo-meric co-assemblies might decrease the rate of aggregation bypeptides with a higher intrinsic nucleation rate, and accelera-tion the aggregation of peptides with a lower intrinsic nucle-ation rate. Indeed, there are indications that these shorteralloforms play an important role in vivo by modulating patho-genic Ab1–42 aggregation. It has been shown these shorteralloforms are not inherently toxic, but are in fact neuro-protective in a dose-dependent manner and capable of reducingAb1–42 deposition.49,50 It has been proposed that this neuro-protective behavior could be caused by inhibition of Ab42 olig-omerization by the shorter alloforms51,52 or by impededoligomer conversion to brillar structure,53 as has previouslybeen observed for Ab40.54,55 The interactions and cross-reactivitybetween Ab alloforms other than Ab1–42/Ab1–40 have, to ourknowledge, not been carefully studied.In this study, we investigate the interactions and cross-reactivity between four Ab naturally abundant alloforms, Ab37,Ab38, Ab40, and Ab42, with an aim to unravel any cross-catalysisor co-aggregation processes (Fig. 1). Working in vitro and usinghighly puried peptides to maximize reproducibility, we char-acterize their interactions primarily through kinetic aggregationassays using the uorescent dye thioavin T (ThT), the quantumyield of which reports quantitatively on aggregate massconcentration when properly calibrated.56 We investigate notonly the effect that each alloform has on the aggregation of theothers in binary mixtures, but also consider three-peptide andfour-peptide mixtures in order to better understand howdifferent ratios of Ab alloforms affects the overall rate ofconversion of monomer to amyloid bril. In these experiments,we ask whether any particular ratio of the shorter length vari-ants will maximally affect the aggregation of Ab42.ResultsThe aggregation of Ab(M1-37), Ab(M1-38), Ab(M1-40), andAb(M1-42) (Fig. 2, referred to herein as Ab37, Ab38, Ab40, andAb42, respectively) in binary, trinary, and quaternary mixtureswas investigated using in vitro kinetic assays, analytical highperformanceliquidchromatography(HPLC),andcryo-transmissionelectronmicroscopy(cryo-TEM).First,theFig. 1Possible modes of cross-catalysis through (a) heterogeneous primary nucleation in co-oligomers or (b) on fibril surfaces. (c) Co-aggregation is here defined as the process leading to any form of aggregates composed of more than one peptide alloform.2424 | Chem. Sci., 2022, 13, 2423–2439© 2022 The Author(s). Published by the Royal Society of ChemistryChemical ScienceEdge ArticleOpen Access Article. Published on 07 February 2022. Downloaded on 3/14/2022 5:09:18 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence.View Article Onlineaggregation of Ab37 or Ab38 in the absence of other peptides wasinvestigated. Next, the aggregation of two-peptide systems,including either Ab37 or Ab38 together with either Ab40 or Ab42,wasinvestigated.Additionally,two-peptidecross-seedingexperiments were also performed. Subsequently, three-peptidesystems composed of Ab40, Ab42, and either Ab37 or Ab38 werestudied. Finally, the aggregation behavior of four-peptidesystem, including all four alloforms, was investigated. Allexperiments were performed in 20 mM sodium phosphatebuffer, 200 mM ethylenediaminetetraacetic acid (EDTA), 0.02%NaN3, at pH 7.4, as all alloforms aggregate fast enough to giverelatively reproducible data under these conditions. The incor-poration of an N-terminal Met residue has no observable effecton the ssNMR spectrum of Ab42 brils and aggregation mech-anism,57 and was chosen because it allows for expression “as is”with easy purication without the complication of tags andproteases.58Independent aggregation of Ab37 and Ab38Before studying the aggregation of Ab37 and Ab38 in hetero-molecular mixtures, the aggregation of each peptide indepen-dent of the others was rst investigated (Fig. 3). Ab37 and Ab38displayed greatly reduced aggregation propensity relative toAb40 and Ab42. In identical conditions and at the low end of theconcentration range studied here, Ab40 aggregates within twohours, while Ab42 aggregates within a fraction of an hour.37,38,41However, the aggregation propensity among these naturallyoccurring Ab alloforms is evidently not directly correlated topeptide length, as Ab38 was found here to aggregate much moreslowly than Ab37.The aggregation curve shapes for both Ab37 and Ab38 bothdisplayed long lag phases, during which little aggregation wasobservable above the signal-to-noise ratio, followed by distinctgrowth phases, in which most aggregate growth happenedrapidly. This distinctive curve shape, with a sharp transitionbetween lag and growth phases, is characteristic of an aggre-gation mechanism driven by a self-replicating, bril-catalyzedprocess. The concentration dependence of the rate of aggrega-tion of each peptide was clearly shown by the half-time plots(Fig. 3c), in which the time at which half of the total peptide hadaggregated was plotted versus starting monomer concentration.The scaling exponent, g, of a power function tted to the half-time data provides valuable insight into the underlying aggre-gation mechanism.59,60 Aggregation mechanisms driven byfragmentationandsecondarynucleation,thetwoself-replicating processes observed in amyloid-related aggregation,display characteristic g values: for fragmentation-dominatedsystems, g ¼ �0.5, while for secondary-nucleation-dominatedsystems g # �1.0, with the exact value depending on the reac-tion order of secondary nucleation and the balance betweenprimary and secondarynucleation, andwhether ornotsecondary nucleation saturates.38,61 For both Ab37 and Ab38, g ¼�1.0, indicating that the aggregation of both peptides is drivenby secondary nucleation. This is not surprising, as the aggre-gation of both Ab40 and Ab42 has been shown to be dominatedFig. 2Amino acid sequences of the peptides investigated in this work.Fig. 3The aggregation kinetics of pure solutions of (a) Ab37 and (b) Ab38 in 20 mM sodium phosphate buffer, pH 7.4, with 200 mM EDTA, 0.02%NaN3, and 20 mM ThT. Three technical replicates are shown for each concentration. (c) Double-logarithmic plot of the aggregation half-timeversus initial monomer concentration, [m]0, for Ab37 (purple) and Ab38 (blue), where the half-time is the time at which the fluorescence signal ishalf of its final plateau value. Each dashed line shows a fitted power function with scaling exponent, g. Points show the mean of the threetechnical replicates, with error bars showing the standard deviation.© 2022 The Author(s). Published by the Royal Society of ChemistryChem. Sci., 2022, 13, 2423–2439 | 2425Edge ArticleChemical ScienceOpen Access Article. Published on 07 February 2022. Downloaded on 3/14/2022 5:09:18 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence.View Article Onlineby this same process.37,38 The high variability among technicalrepeats at low concentrations prevents data tting to generatea quantitative description of the aggregation process for eitherpeptide, as has previously been done for Ab40 and Ab42.37,38,62However, such comprehensive characterization is not requiredfor this study, for which the interactions between different Aballoforms, rather than the isolated behavior of any one peptide,is the primary focus.Aggregation kinetics of mixtures of Ab40 with Ab37 or Ab38In kinetic assays of the aggregation process in binary mixturesof Ab40 and either Ab37 or Ab38, only a single transition wasobserved (Fig. 4). This contrasts with the aggregation behaviorof monomeric Ab40/Ab42 mixtures, which produced a double-sigmoidal prole that was found to correspond to the self-assembly of each peptide on distinct timescales into segre-gated, homomolecular brils.48 The effects of each peptide onthe aggregation of the other were studied by holding theconcentration of one peptide constant while varying theconcentration of the other. The addition of Ab40 to solutions ofAb37 or Ab38 decreased the lag time before the single observeduorescence transition (Fig. 4a and b). Conversely, witha constant concentration of Ab40, the addition of either Ab37 orAb38 resulted in a delay in aggregation (Fig. 4c and d). In bothcases, the addition of the shorter alloform to Ab40 results in anincrease in the intensity of ThT uorescence, indicating that theshorter alloforms are aggregating concurrently with Ab40, ratherthan at a later timepoint. The modulation of the kinetics ofaggregation for both peptides in both Ab37/Ab40 and Ab38/Ab40mixtures, as well as the fact that a monophasic aggregationprole is observed at all molar ratios tested despite the differingtimescales of the independent aggregation of Ab37 and Ab38compared to that of Ab40, suggest that these peptides may co-aggregate to form some level of mixed aggregates. However,ThT uorescence assays, which report only on total aggregatemass concentration, cannot distinguish co-aggregates fromcoexisting homomolecular brils.The ultrastructure of the aggregates formed by thesepeptides, both alone and in mixtures, was investigated usingcryo-TEM (Fig. 5). Images were collected of brils produced byeach peptide individually, as well as by those formed by equi-molarAb37/Ab40andAb38/Ab40mixtures.Sampleswereprepared in the same manner as for the kinetic assays, withbril formation monitored by ThT uorescence. The sampleswere frozen on grids once the nal uorescence plateau wasreached. All of the brils formed from a pure peptide solutiondisplay relatively consistent morphology. Moreover, thesehomomolecular brils were morphologically similar betweenpeptides: the brils formed by Ab37, Ab38, and Ab40 had averageFig. 4Aggregation kinetics of two-peptide systems composed of Ab40 with (a & c) Ab37 and (c & d) Ab38 in 20 mM sodium phosphate buffer, pH7.4, with 200 mM EDTA, 0.02% NaN3, and 20 mM ThT. The effect of Ab40 at varying concentrations on the aggregation of (a) Ab37 or (b) Ab38.Aggregation kinetics for Ab40 alone at the same concentrations are shown in Fig. S1.† The effect of varying concentrations of (c) Ab37 or (d) Ab38on the aggregation of Ab40. The extended time-courses for (c) and (d) are shown in Fig. S2.†2426 | Chem. Sci., 2022, 13, 2423–2439© 2022 The Author(s). Published by the Royal Society of ChemistryChemical ScienceEdge ArticleOpen Access Article. Published on 07 February 2022. Downloaded on 3/14/2022 5:09:18 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence.View Article Onlinehelical half-pitch periods of 118 � 30 nm, 131 � 40 nm, and 125� 22 nm and average bril diameters of 19 � 6 nm, 15 � 5 nm,and 18 � 4, respectively. Notably, the morphology of the brilsformed by each of these peptides was signicantly differentthan that of the brils formed by Ab42, which have a helical halfpitch of 31 � 17 nm.48 The brils formed from binary mixtures,on the other hand, displayed distinct morphology. For bothAb37/Ab40 and Ab38/Ab40 mixtures, brils were observed withhelical half-pitch periodicities that did not correspond to thebrils formed independently by either of the peptides present inthe reaction mixture. In both cases, the average bril half-pitchperiod (160 � 90 nm for Ab37/Ab40 and 183 � 106 for Ab38/Ab40)was both signicantly longer and more variable than for any ofthe homomolecular brils (Fig. 5b). Additionally, the averagewidth of the brils formed in each binary mixture (21 � 10 nmfor Ab37/Ab40 and 23 � 12 for Ab38/Ab40) was both larger andmore variable than for either of the constituent peptides,although this was statistically signicant only for Ab38/Ab40(Fig. 5c). The presence in both binary systems of brils withmorphology distinct from those of homomolecular brilsformed by either constituent peptide provided further evidencethat some level of co-aggregation had occurred both betweenAb37 and Ab40 and between Ab38 and Ab40.Aggregation kinetics of mixtures of Ab42 with Ab37 or Ab38Following the investigation of the aggregation behavior of Ab40with Ab37 and Ab38, similar experiments were performed on two-peptide systems with Ab42 (Fig. 6). Starting from monomericAb37/Ab42 or Ab38/Ab42 mixtures, two distinct transitions wereobserved in the curves as monitored by ThT uorescence. Suchbiphasic aggregation proles were previously observed for theaggregation of a mixture of Ab40 and Ab42, which was found tocorrespond to aggregation into separate brils.48 This indicatesthat the nature of the interactions of these shorter alloformswith Ab42 are quite different than with Ab40. The biphasic prolesuggests that two separate aggregation processes are occurringon discrete timescales, consistent with the formation ofhomomolecular brils.There were nevertheless signicant interactions between thetwo peptides present in both Ab37/Ab42 and Ab38/Ab42 mixtures,as evidenced by the changes to the timescale of aggregation ofboth peptides in the mixture. The presence of Ab42 stronglyaccelerated the aggregation of both Ab37 and Ab38 (Fig. 6a and b,respectively) by dramatically shortening their lag phases.Notably, each of Ab37 and Ab38 clearly inhibited the aggregationof Ab42 (Fig. 6c and d, respectively), although this effect is lessdramatic than the reciprocal acceleration of both Ab37 and Ab38aggregation by Ab42.The aggregation of Ab42 is known to proceed through threedistinct processes: primary (non-bril-catalyzed) nucleation,secondary (bril-catalyzed) nucleation, and elongation.37 Toinhibit aggregation, an effector must perturb one or more ofthese processes. Inhibition of each of the processes affects thekinetics of aggregation in a distinct manner,14,63 meaning thatFig. 5The morphology of fibrils formed by Ab37, Ab38, and Ab40 in pure and equimolar binary solutions. (a) Representative cryo-TEM images offibrils. Fibrils were formed in 20 mM sodium phosphate buffer, pH 7.4, with 200 mM EDTA, 0.02% NaN3, and 20 mM ThT. (b) Quantification of fibrilnode-to-node distance (half-pitch period). Representative grey scale profiles used for helical half-pitch period measurements are shown inFig. S3.† (c) Quantification of fibril diameter. Plots represent >50 measurements, taken from at least 15 different fibrils. **p # 0.01, ***p # 0.001,and ****p # 0.0001 by one-way ANOVA followed by Dunnett's post-hoc test.© 2022 The Author(s). Published by the Royal Society of ChemistryChem. Sci., 2022, 13, 2423–2439 | 2427Edge ArticleChemical ScienceOpen Access Article. Published on 07 February 2022. Downloaded on 3/14/2022 5:09:18 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence.View Article Onlinethe change in the Ab42 curve shape in the presence of each of theshorter alloforms can, in principle, be used to determine whichprocess is being inhibited. Using the AmyloFit program,59a model of secondary-nucleation-dominated aggregation was tto the data shown in Fig. 6c and d; the parameters corre-sponding to the rate of each process were, one-by-one, allowedto vary while all other parameters were held constant (Fig. 7).This analysis shows that the observed changes to the curveshape of Ab42 aggregation in the presence of both Ab37 and Ab38is poorly modeled by inhibition of primary nucleation (Fig. 7aand d, respectively). However, the change to the Ab42 aggrega-tion curve shape, although signicant, is too small for thekinetic analysis to distinguish between inhibition of secondarynucleation (Fig. 7b and e) and inhibition of elongation (Fig. 7cand f). Nevertheless, this tting indicates that the inhibition ofAb42 aggregation by both Ab37 and Ab38 involves interactionbetween the shorter peptide and the Ab42 bril, as bothsecondarynucleationandelongationarebril-catalyzedprocesses.Cross-seeding of Ab37 and Ab38 with Ab40 and Ab42Having studied the aggregation in systems composed of two Aballoforms, both starting in a monomeric state, we next investi-gated whether any of these peptide mixtures displayed cross-seeding behavior. In aggregating systems driven by a bril-catalyzed secondary nucleation process, as is the case for allAb alloforms studied here, the addition of pre-aggregated seedbrils to a solution of monomer can accelerate aggregation byproviding a catalytic surface for nucleation from the start of thereaction process. The self-seeding of both Ab42 and Ab40 is wellFig. 6Aggregation kinetics of two-peptide systems composed of Ab42 with (a & c) Ab37 and (b & d) Ab38 in 20 mM sodium phosphate buffer, pH7.4, with 200 mM EDTA, 0.02% NaN3, and 20 mM ThT. (a and b) Increasing Ab42 concentration leads to accelerated aggregation of Ab37 and Ab38,respectively. Aggregation kinetics for Ab42 alone at the same concentrations are shown in Fig. S1.† (c and d) Increasing concentrations of Ab37 andAb38, respectively, decrease the rate of Ab42 aggregation; in both, the upper panel shows the first, low-intensity fluorescence transition and thelower panel shows the full biphasic aggregation profile.2428 | Chem. Sci., 2022, 13, 2423–2439© 2022 The Author(s). Published by the Royal Society of ChemistryChemical ScienceEdge ArticleOpen Access Article. Published on 07 February 2022. Downloaded on 3/14/2022 5:09:18 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence.View Article Onlineestablished.37,38,64 Although Ab40 and Ab42 have been shown todisplay some cross-seeding behavior, these effects are onlyevident at high seed concentrations (0.5 and 1 mM Ab40 seedadded to 2 mM Ab42 monomer) and are much weaker than thoseobserved for self-seeding.48In cross-seeding reactions, interactions between the mono-meric peptide and the surface of the seed bril can result indivergent kinetic outcomes.65 The seed bril can serve asasurfaceforheterogenousprimarynucleation,whereinmonomer adsorption by the bril increases local monomerconcentration, thereby catalyzing aggregation.66,67 Conversely,higher-affinity monomer adsorption by the bril surface leadsto monomer sequestration, thus reducing the overall rate ofaggregation, particularly during the bril growth phase.48,67 Inprevious cross-seeding reactions between Ab40 and Ab42, Ab42seeds were shown to have no effect on Ab40 aggregation;although high concentrations of Ab40 seeds (25–50%) wereshown to affect Ab42 aggregation, these effects were muchweaker than those of the self-seeding of Ab42.48Here, we found that the aggregation of Ab37 is catalyzed bythe presence of seeds of any of the four peptides (Fig. 8a). Theeffects were most dramatic for the self-seeding of Ab37 (purple)and Ab38 (blue), although they were nevertheless signicant inall three cross-seeding cases. For the cross-seeding of Ab37 withAb40 brils (green), the addition of seeds not only decreased theduration of the lag phase, but also reduced the slope of thecurve during the aggregate growth phase. In contrast, eitherAb38 (blue) or Ab42 (red) seed shortened the lag phase withminimal effects on the curve shape. This reduction in lag phasewithout a concomitant change in the curve shape suggests thatthe likely cause of the acceleration is the provision of a catalyticsurface for the heterogeneous primary nucleation of Ab37.However, at such long lag times, variability between experi-mental repeats is inevitable and prevents explicit verication ofthis proposed mechanism through tting to kinetic models.Notably, the catalytic effect of Ab42 brils on Ab37 aggregationwas smaller than that of monomeric Ab42 at comparableconcentrations (Fig. 6a), indicating that monomer–monomerinteractions are a signicant factor in the catalysis of Ab37aggregation by Ab42. The fact that some cross-seeding behaviorwas observed, however, indicates that interactions betweenmonomeric Ab37 and brillar Ab42 do also occur. This isconsistent with the conclusions drawn from mechanistic ttingto the data of Ab42 aggregation in the presence of monomericAb37 (Fig. 7a–c).As with Ab37, the acceleration of Ab38 aggregation wasstronger in the self-seeded case than for cross-seeding (Fig. 8b).At high concentrations of Ab38 seed, there was essentially no lagphase, with signicant quantities of brils formed almostinstantaneously. Having already found that Ab38 brils can seedthe aggregation of monomeric Ab37 (Fig. 8a), the reciprocaleffect is also observed, with the addition of Ab37 seeds acceler-ating Ab38 aggregation. For the cross-seeding of Ab38 with bothAb40 and Ab42 seed brils, a clear concentration-dependentdecrease in the lag time was observed; in both cases, therewas no perturbation to the aggregation curve shape.For Ab40, clear self-seeding behavior was observed (Fig. 8c),consistentwithpreviousndings.38,48Additionally,clearreduction in the duration of the lag time was observed with theaddition of both Ab37 and Ab38 seed brils, particularly atFig. 7Fitting different models of inhibition to aggregation kinetics of 3 mM Ab42 with varying concentrations of (a–c) Ab37 and (d–f) Ab38; the dataare the same as are shown in Fig. 6c and d, respectively. The models used here correspond to inhibition of (a & d) primary nucleation, (b & e)secondary nucleation, and (c & f) elongation processes. (g) Mean residual errors for fitting shown in (a–f), normalized to the error for the best fitfor each data set. Errors for Ab37/Ab42 fitting are shown in purple and errors for Ab38/Ab42 fitting are shown in blue.© 2022 The Author(s). Published by the Royal Society of ChemistryChem. Sci., 2022, 13, 2423–2439 | 2429Edge ArticleChemical ScienceOpen Access Article. Published on 07 February 2022. Downloaded on 3/14/2022 5:09:18 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence.View Article Onlinerelatively high seed concentrations. As is the case for both Ab37and Ab38, however, this cross-seeding effect was again weakerthan the self-seeding effect. Similar to what was observed for theseeding of Ab37 monomer with Ab40 brils, the addition of Ab37brils to Ab40 monomer affected not only the lag phase but alsothe slope of the curve during the aggregate growth phase.Indeed, at low concentrations of Ab37 seed brils, this inhibitoryeffect was more signicant than the reduction of the lag phase,resulting in an increase of the aggregation half-time relative tounseeded Ab40.Clear self-seeding behavior was observed for Ab42, as hasbeen reported previously;37,48 however, no cross-seeding catal-ysis was observed with either Ab37 or Ab38 seed brils (Fig. 8d).Even at high concentrations of Ab37 and Ab38 seed brils, boththe timescale and curve shape of Ab42 aggregation werecompletely unperturbed, indicating that these foreign aggre-gates are inert with respect to monomeric Ab42, neitherproviding a surface for heterogeneous nucleation nor seques-tering monomer. This suggests that the inhibition of Ab42aggregation observed in the presence of Ab37 or Ab38 (Fig. 6c andFig. 8Self- and cross-seeding experiments for monomeric (a) Ab37, (b) Ab38, (c) Ab40, and (d) Ab42. In all panels, the color of the framecorresponds to the identity of the monomeric peptide and the color of the kinetic curves corresponds to the identity of the seed peptide, withAb37, Ab38, Ab40, and Ab42 represented by purple, blue, green, and red, respectively. Seed concentrations are given as a molar percentage of themonomer concentration. The rightmost panel in each row shows the aggregation half-time as a function on seed percentage. Ab37, Ab38, andAb40 monomer solutions were prepared to a peptide concentration of 10 mM, and Ab42 was prepared to a monomer concentration of 3 mM. Allexperiments were performed in 20 mM sodium phosphate buffer, pH 7.4, with 200 mM EDTA, 0.02% NaN3, and 20 mM ThT.2430 | Chem. Sci., 2022, 13, 2423–2439© 2022 The Author(s). Published by the Royal Society of ChemistryChemical ScienceEdge ArticleOpen Access Article. Published on 07 February 2022. Downloaded on 3/14/2022 5:09:18 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence.View Article Onlined, respectively) was caused by monomeric or oligomeric, ratherthan brillar, species. This is especially relevant given that, atthe point in time at which Ab42 aggregates, both Ab37 and Ab38will be in a primarily monomeric state, due to the relatively slowaggregation of both peptides.Aggregation kinetics of three-peptide mixturesHaving characterized the effects that the presence of eithermonomeric or brillar species of each Ab alloform had onaggregation kinetics of the others, we next studied the behaviorof Ab37 or Ab38 in three-peptide monomeric mixtures with Ab40and Ab42. For both the Ab37/Ab40/Ab42 and the Ab38/Ab40/Ab42mixtures, two discrete transitions in ThT uorescence wereobserved, separated by an intermediate plateau phase (Fig. 9aand c, respectively). This is consistent with two aggregationprocesses happening on distinct timescales. The intensity of thesecond transition in ThT uorescence increased with increasingconcentration of Ab37 or Ab38, indicating that both peptidesaggregate as part of the second, slower process. Based on theprevious monomeric two-peptide aggregation kinetic assays(Fig. 4 and 6), it is most likely that the rst uorescence tran-sition here represents the aggregation of Ab42, with the secondtransition showing the co-aggregation of Ab40 and Ab37 or Ab38.To analyze the identity and composition of the peptideaggregates that have formed during each transition, analyticalHPLC was used to determine the concentration of each peptideover the time-course of the aggregation of an equimolar three-peptide mixture (Fig. 10). To characterize the composition ofbrillar species, aliquots were removed during the initial lagphase (tL) and each plateau phase (tP1 and tP2) (Fig. 10b and c formixtures including Ab37 and Ab38, respectively). The brilspresent in these aliquots were isolated by centrifugation,washed to remove any oligomers associated with the brilsurface, then dissolved and injected on an HPLC instrument. InFig. 9Aggregation of three-peptide systems composed of 3 mM each Ab40 and Ab42, with varying concentrations of (a and b) Ab37 or (c and d)Ab38 in 20 mM sodium phosphate buffer, pH 7.4, with 200 mM EDTA, 0.02% NaN3, and 20 mM ThT. Panels (a) and (c) show the aggregation kineticscurves. Panels (b) and (d) show the half-time of Ab42 aggregation plotted as a function of Ab37 and Ab38 concentration, respectively. The pointsrepresent the average of the three technical triplicates, with the error bars showing the standard deviation.© 2022 The Author(s). Published by the Royal Society of ChemistryChem. Sci., 2022, 13, 2423–2439 | 2431Edge ArticleChemical ScienceOpen Access Article. Published on 07 February 2022. Downloaded on 3/14/2022 5:09:18 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence.View Article Onlineboth three-peptide mixtures, it was found that the brilspresent in the intermediate plateau phase were primarilycomposed of Ab42, while the brils in the nal plateau phasewere composed of all three peptides. These data are supportedby the complementary analysis of the peptide composition ofthe supernatants (Fig. 10d and e), for which additional aliquotswere taken during each aggregate growth phase (tG1 and tG2).For both Ab37/Ab40/Ab42 and Ab38/Ab40/Ab42 mixtures, it wasfound that the concentration of monomeric Ab42 decreasedduring the rst transition in ThT uorescence and was unde-tectable by the intermediate plateau phase. During the secondtransition phase, the concentration of both Ab40 and either Ab37or Ab38 decreased. At the nal plateau phase, the majority ofpeptides were converted to brils and the monomer concen-tration was below the detection limit of analytical HPLC. Thisindicates that, in these three-peptide mixtures, the behavior ofeach peptide is consistent with the behavior to in two-peptidesystems, with Ab42 forming brils during the rst transitionand Ab40 co-aggregating with Ab37 or Ab38 during the secondtransition.Despite aggregating separately from Ab42, the presence ofeither Ab37 or Ab38 perturbed the kinetics of Ab42 aggregation, aswas previously observed in two-peptide systems (Fig. 6c and d).The half-time of Ab42 aggregation increased with increasingconcentration of either Ab37 or Ab38 (Fig. 9b and d, respectively),with the half-time of Ab42 aggregation dened here as the timeat which the uorescence intensity reached half-way betweenthe initial baseline and the intermediate plateau. This showsthat Ab37/Ab42, Ab38/Ab42, and Ab40/Ab42 (ref. 48) interactionsobserved in two-peptide systems persist in three-peptidesystems.Fig. 10Alloform composition in fibrillar and monomeric species over the time-course of aggregation of trimolecular mixtures of 3 mM eachAb40, Ab42, and either (b & d) Ab37 or (c & e) Ab38, as analyzed using analytical HPLC. (a) A representative kinetic aggregation curve, with the timepoints at which aliquots were removed marked with red lines. (b & c) The composition of fibrillar species during the lag phase and each plateauphase. (d & e) The composition of monomeric species during the lag phase, both transition phases, and both plateau phases. Kinetic experimentswere performed in 20 mM sodium phosphate, 200 mM EDTA, 0.02% NaN3, pH 7.4, with 20 mM ThT.2432 | Chem. Sci., 2022, 13, 2423–2439© 2022 The Author(s). Published by the Royal Society of ChemistryChemical ScienceEdge ArticleOpen Access Article. Published on 07 February 2022. Downloaded on 3/14/2022 5:09:18 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence.View Article OnlineAggregation kinetics of four-peptide mixturesFinally, the aggregation behavior in monomeric mixtures of allfour alloforms was studied. First, the aggregation of 3 mM Ab42in mixtures with varying concentrations of Ab37, Ab38, and Ab40at an equimolar ratio was characterized (Fig. 11). For all four-peptide mixtures, two distinct transitions in ThT uorescenceintensity were observed. The intensity of the second transition(which was absent in a solution containing only Ab42) scaledwith increasing concentration of the Ab37/Ab38/Ab40 mixture,while the intensity of the rst transition did not. This indicatesthat all three of the shorter Ab alloforms may form co-aggregates, while Ab42 aggregates rst, forming separatebrils. As would be expected given the behavior observed inthree-peptide mixtures, the aggregation of Ab42 was perturbedby the presence of shorter Ab alloforms despite the sloweraggregation of these peptides, with increasing concentrations ofthe short-alloform mixture leading to longer Ab42 aggregationhalf-times (Fig. 11b).We next characterized the relative effects of each of theshorter Ab alloforms on the aggregation of Ab42, with an aim todetermine whether specic ratios of these alloforms wouldmaximally affect Ab42 aggregation. For this, the aggregation of 3mM Ab42 was studied in the presence of a constant totalconcentration (6 mM) of mixtures of Ab37, Ab38, and Ab40, withthese three peptides in varying ratios (Fig. 12). For each ratiotested,onlytwouorescencetransitionswereobserved,providing further evidence that the three shortest alloforms co-aggregate irrespective of their ratio. Both the timescale andintensity of the second transition varied based on the ratio ofAb37, Ab38, and Ab40, due to the differing aggregation propen-sities and ThT quantum yielded of each peptide. The aggrega-tion of Ab42 was found to be inhibited by all solutions tested,irrespective of the peptide ratios, although to varying extents.Included in this study were 6 mM each Ab37, Ab38, and Ab40alone, to allow for side-by-side comparison of the degree towhich each alloform inhibits Ab42 aggregation. Of the three,Ab38 inhibited the aggregation of Ab42 to the greatest degree,followed by Ab37 and Ab40. This was found to be reproduciblebetween repeats of the whole experiment. Notably, mixtures ofall three peptides produced a stronger inhibitory effect than anysingle peptide alone. Although the relative strength of inhibi-tion of these different mixtures was not found to be highlyreproducible between repeats of the experiment, mixtures withhigher concentrations of Ab38 generally exhibited strongerinhibitory effects than did mixtures with low concentrations ofAb38. Still, Ab38 alone was not the most effective, with thestrongest inhibition observed at the 2 : 3 : 1 and 1 : 4 : 1 ratiosof Ab37 : Ab38 : Ab40.DiscussionThe results of this study points to signicant cross-reactivity inquaternary mixtures of Ab42 and shorter alloforms, beyondthose observed in binary mixtures.48 Notably, both monomericAb37 and Ab38 appear to interact with Ab42 in much the samemanner as does monomeric Ab40. Most fundamentally, all threeof the shortest alloforms aggregate independently of Ab42,giving rise to two distinct uorescence transitions. Further-more, similar reciprocal kinetic effects are observed in binaryFig. 11(a) The aggregation of 3 mM Ab42 with varying concentrationsof a 1 : 1 : 1 mixture of Ab37/Ab38/Ab40. (b) The half-time of theaggregation of Ab42 shown in (a) versus the total concentration of thethree shorter alloforms. Kinetic experiment was performed in 20 mMsodium phosphate buffer, pH 7.4, with 200 mM EDTA, 0.02% NaN3, and20 mM ThT.Fig. 12Aggregation kinetics of 3 mM Ab42 with 6 mM of Ab37/Ab38/Ab40 mixtures in varying molar ratios. (a) The first ThT transition, correspondingto the aggregation of Ab42 is shown; the full aggregation profiles are shown in Fig. S4.† (b) The half-times of Ab42 aggregation in (a), versus theconcentration of Ab37 (green panel), Ab38 (blue panel), and Ab40 (red panel). The color of each point corresponds to the curves of the same colorin (a). (c) Triangular heat-map representation of the half-time of Ab42 aggregation depending on the Ab37/Ab38/Ab40 concentrations. Experimentswere performed in 20 mM sodium phosphate buffer, pH 7.4, with 200 mM EDTA, 0.02% NaN3, and 20 mM ThT.© 2022 The Author(s). Published by the Royal Society of ChemistryChem. Sci., 2022, 13, 2423–2439 | 2433Edge ArticleChemical ScienceOpen Access Article. Published on 07 February 2022. Downloaded on 3/14/2022 5:09:18 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence.View Article Onlinemonomeric mixtures of Ab42 with each of the three shorteralloforms. As was found here for both Ab37 and Ab38, the pres-ence of monomeric Ab42 accelerates the aggregation of Ab40 bysignicantly shortening the lag phase,48 although this effect issignicantly more pronounced with Ab37 and Ab38, which bothaggregate much more slowly than Ab40. Additionally, the pres-ence of any of the three shortest alloforms serves to inhibit theaggregation of Ab42. In each case, this effect is weaker than thereciprocal Ab42-induced acceleration of aggregation of theshorter alloform.The cross-seeding between Ab42 and each of the shorteralloforms showed a stronger dependence on alloform length. Asreported by Cukalevski et al., the addition of Ab42 seeds to Ab40monomers results in minimal change to the kinetics of Ab40aggregation.48 This contrasts with the cross-seeding interac-tions between Ab42 seeds and both Ab37 and Ab38 monomers, forwhich a clear Ab42-concentration-dependent decrease in theaggregation half-time of the shorter alloform was observed inboth cases. Furthermore, Cukalevski et al. reported that theaddition of high concentrations of Ab40 seeds to Ab42 mono-mers leads to a moderate reduction of both the lag time and therate of Ab42 aggregation during the bril growth phase (theseconverse effects largely cancel out, leaving the half-time of Ab42essentially unchanged).48 This behavior differs from the inter-actions observed here for the cross-seeding of Ab42 monomerswith either Ab37 or Ab38, for which even high seed concentra-tions had no effect on the kinetics of Ab42 aggregation.The results of this study serve to extend our understanding ofinteractions between Ab alloforms by investigating the aggre-gation behavior of tri- and tetra-molecular mixtures. The factthat Ab37, Ab38, and Ab40 each affected Ab42 aggregation toa different extent indicates that inhibition of Ab42 aggregationby these shorter alloforms is not simply due to non-specicinteractions with Ab42, but rather includes some level ofsequence specicity. That Ab38 is a more potent inhibitor ofAb42 aggregation than are either Ab37 or Ab40 indicates that thiseffect is not directly correlated to alloform length. Likewise, thisinhibition does not appear to be correlated with aggregationpropensity, as Ab37 and Ab40 inhibited Ab42 aggregation toa similar extent, despite Ab40 displaying a much higher intrinsicaggregation rate than Ab37. Determination of the variousfactors, including peptide length and sequence, that affect therelative inhibitory potency of these shorter alloforms towardsAb42 aggregation would provide greater physical insight into theinteractions that underlie this inhibition.Notably, the results of this study indicate that, in addition tobeing alloform-specic, the effect of short Ab alloforms on theaggregation of Ab42 is inherently cooperative. Were this inhi-bition non-cooperative, it would be expected that the decreasein Ab42 aggregation would simply scale with the relativeconcentrations of the short alloforms. Instead, a mixture ofthese shorter alloforms affected the rate of Ab42 aggregationmore signicantly than any single alloform independently, withthe strongest inhibition seen at the 2 : 3 : 1 and 1 : 4 : 1 Ab37-: Ab38 : Ab40 ratios. While the underlying molecular interac-tions that give rise to this cooperative inhibition are not readilyevident, these results further buttress the conclusion thatinhibition of Ab42 aggregation by shorter Ab alloforms issequence specic. One possible explanation for inhibitionobserved in certain quaternary mixtures may be the formationof mixed oligomers that cannot readily convert to brils andthat compete for catalytic sites on the bril surface with moreconversion-competent homomolecular oligomers.68 AlthoughAb42 in heteromolecular mixtures aggregates at reduced ratecompared to pure Ab42, it seems to promote nucleation of Ab37,Ab38 and Ab40 in these mixed oligomers, and aer Ab42 hasformed homomolecular brils, these three peptides seem formjoint brils with different morphology than any of the purebrils of Ab37, Ab38 and Ab40.When studied alone as single peptides at a range ofconcentrations, Ab38 appears to aggregate much more slowlythan Ab37. The origin of this difference is not clear but may bea consequence of primary or secondary nucleation being lesseffective with the extra Gly residue at the C-terminus of Ab38. Ingeneral, longer sequences aggregate more slowly if the addi-tionallengthdifferenceliesinanon-amyloidogenicsegment.33,69,70These results suggest that in the brain extracellular uids,where a number of different short Ab alloforms are present, thedisease-associated aggregation of Ab1–42 may in fact be modu-lated by these short alloforms and, furthermore, that the rate ofAb1–42 aggregation may be sensitive to specic ratios of thesealloforms. The relevance of these ndings is further under-scored by the disease-promoting effect of familial AD-causingpresenilin mutations, which lead to a loss of Ab1–37 and Ab1–38whilst affecting neither Ab1–40 nor Ab1–42 production.71ConclusionsIn this study, we show that the Ab alloforms Ab37 and Ab38 bothaggregate through a secondary nucleation driven process toform b-sheet rich brils. The aggregation of both of thesepeptides is comparable to the well-characterized aggregationprocesses of Ab40 and Ab42, although the aggregation of theshorter two alloforms is signicantly slower, which is consistentwith their reduced aggregation propensity in vivo. Furthermore,we nd that the aggregation of all four Ab alloforms is highlysensitive to the presence of other alloforms. Both Ab37 and Ab38co-aggregate with Ab40, forming brils that are ultrastructurallydistinct from the brils formed by either of the constituentpeptides alone. Furthermore, the kinetics of aggregation inthese binary mixtures is highly dependent on the ratio of Ab37/Ab40 or Ab38/Ab40 present in solution. In contrast to their co-aggregation with Ab40, we nd that both Ab37 and Ab38 aggre-gate independently from Ab42. There is nevertheless signicantkinetic modulation between Ab42 and both Ab37 and Ab38, withthe shorter alloforms impeding the aggregation of Ab42 ina concentration-dependentmannerand Ab42reciprocallyaccelerating the aggregation of the shorter alloforms. Finally,we nd that the aggregation of Ab42 is sensitive to the specicratio of shorter alloforms present, indicating a level of sequencespecicity in these interactions; we furthermore demonstratethat a mixture of Ab37, Ab38, and Ab40 more potently affects Ab422434 | Chem. Sci., 2022, 13, 2423–2439© 2022 The Author(s). Published by the Royal Society of ChemistryChemical ScienceEdge ArticleOpen Access Article. Published on 07 February 2022. Downloaded on 3/14/2022 5:09:18 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence.View Article Onlineaggregation than any one of the shorter alloforms alone, indi-cating that this inhibition is inherently cooperative.These results demonstrate that interactions between Aballoforms signicantly modulate the aggregation behavior ofthese peptides in vitro. As the presence of Ab38 has been shownto attenuate Ab42 bril formation and deposition in vivo,49,50 it islikely that the interactions described here are pertinent to thebiological behavior of these peptides. Indeed, short Ab allo-forms, which make up as much as 90% of the Ab peptidepresent in the extracellular uid of the brain, may play a path-ologically relevant role in moderating the aggregation of Ab42.These results are relevant not only to understanding the path-ogenesis of AD, but also to the development of AD therapeutics.As modulation of g-secretase activity is increasingly beingproposed as a promising route to AD therapeutics,72–74 anunderstandingofhowdifferentAballoforms—andhowdifferent combinations of these alloforms—affect pathologicalAb42 aggregation is becoming increasingly important.Materials and methodsExpression and purication of Ab peptidesThe genes encoding wild-type Ab(M1-42), Ab(M1-40), Ab(M1-38),and Ab(M1-37) (referred to herein as Ab42, Ab40, Ab38, and Ab37,respectively) were produced by overlapping PCR and cloned intothe PetSac vector.58 The peptides were expressed in E. colistrains BL21 Star (DE3) pLysS (for Ab42, Ab40, and Ab37) or BL21-Gold (DE3) pLysS (for Ab38). Cells were cultured in LB mediumwith 50 mg L�1 ampicillin and (for BL21 Star cells) 30 mg L�1chloramphenicol. Well-isolated bacterial colonies were used toinoculate 50 mL cultures grown in 250 mL baffled asks at 37 �Cwith 130 rpm shaking for ca. 7 h. Once an OD of 0.7–1.0 wasreached, 500 mL of these cultures were then added to 500 mL LBmedium (containing the relevant antibiotics, as detailed above)in 2 L baffled asks, which were grown at 37 �C with 125 rpmshaking for ca. 15 h. Cells were then harvested by centrifugationat 6000 � g for 10 min at 4 �C.The peptides were isolated from inclusion bodies aer iter-ative sonication and centrifugation. The inclusion bodies weredissolved in 8 M urea in 10 mM Tris–HCL, pH 8.5, 1 mM EDTA,and the peptide puried by ion exchange chromatography ona DEAE cellulose resin, as previously described.58 The purity ofthe eluted fractions was assessed by SDS-PAGE. The purestfractions were pooled. Monomer isolation and removal ofinclusion-body-bound proteins was achieved by ltration using30 kDa MWCO lters, and size exclusion chromatography ona 26 � 600 mm Superdex 75 column. The peptide was thenaliquoted, frozen and lyophilized, and stored at �20 �C.Monomer isolation for kinetic assaysTo isolate pure monomer, Ab peptides were subjected to size-exclusionchromatography(SEC)immediatelybeforepreparing aggregation kinetic assays. Aliquots of puried,lyophilized Ab were dissolved in 1.0 mL 6 M GuHCl, injected ona Superdex 75 10/300 GL column, and eluted with an isocratic0.7 mL min�1 ow of 20 mM sodium phosphate buffer, pH 7.4,with 200 mM EDTA and 0.02% NaN3. Typical retention timeswere between 19 and 21 minutes for all peptides. The center ofthe monomer peak was collected on ice using low-bindingEppendorf tubes. The concentration of puried peptide wasdetermined by integrating the absorbance at 280 nm of thecollected peak, using 3280 ¼ 1400 L mol�1 cm�1 (based on thepresence of a single tyrosine residue in each alloform).Concentration determination in this manner is accurate within�20%.Preparation of unseeded kinetic assays for pure peptidesA dilution series was prepared from the SEC-puried peptide.First, thioavin T (ThT, Calbiochem) was added to the samplefrom a concentrated stock solution (ltered through 0.2 mmlter) to a nal concentration of 20 mM. The dilution series wasthen prepared from this peptide solution, using 20 mM sodiumphosphate buffer, pH 7.4, with 200 mM EDTA, 0.02% NaN3, and20 mM ThT (at this concentration, ThT uorescence was foundto scale linearly with aggregate concentration for all alloforms,Fig. S5†). All samples were prepared in low-binding tubes on ice(Axygen). The samples were loaded from low to high peptideconcentration, with ve wells per sample and 90 mL per well,into a 96-well, half-area, clear bottomed, PEG-coated, blackpolystyrene plate (Corning 3881), which was subsequentlysealed. The whole setup was repeated at least three times foreach peptide.Preparation of multimolecular unseeded kinetic assaysAll four peptides were puried separately by SEC, as describedabove. ThT was added from a concentrated stock solution toeach of the puried peptides to a nal concentration of 20 mM.Additionally, a dilution solution of 20 mM sodium phosphatebuffer, pH 7.4, with 200 mM EDTA, 0.02% NaN3, and 20 mM ThTwas prepared. The puried peptide solutions and the dilutionsolution were combined in varying quantities to give the desiredmolar ratios and total concentrations. The samples were thenloaded from low to high concentration of the most aggregation-prone peptide in the mixture, with four to ve 90 mL wells persample, into a 96-well plate (Corning 3881), which was subse-quently sealed. The whole setup was repeated at least threetimes for each peptide combination.Preparation of self- and cross-seeded kinetic assaysSeed brils were made from SEC-puried peptide, with brilformation monitored by ThT uorescence, as described above.Seed brils were removed once the uorescence plateau wasreached, approximately 20 hours for Ab37, 48 hours for Ab38, 5hours for Ab40, and 1 hour for Ab42. These seed brils were thenadded in varying concentrations to aliquots of a solution of SEC-puried monomeric peptide, to which ThT had been added toa nal concentration of 20 mM. For all seeded experiments, theaggregation of a set of unseeded control samples was run inparallel to the seeded reactions. Four technical replicates ofeach sample were then plated into a 96-well plate (Corning3881). The whole setup was repeated at least twice for eachpeptide combination.© 2022 The Author(s). Published by the Royal Society of ChemistryChem. Sci., 2022, 13, 2423–2439 | 2435Edge ArticleChemical ScienceOpen Access Article. Published on 07 February 2022. Downloaded on 3/14/2022 5:09:18 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence.View Article OnlineKinetic assaysThT uorescence was measured using a FLUOstar Omega orFLUOstar Optima plate reader (BMG Labtech) at 37 �C underquiescent conditions. Fluorescence was measured through thebottom of the plate using an excitation lter of 440 nm and anemission lter of 480 nm, with measurements taken every 60–120 s (depending on the aggregation propensity of the alloformbeing studied).Kinetic analysisAll kinetic analyses were performed using the free onlineprogram AmyloFit (www.amylot.ch.cam.ac.uk).59 To facilitatehalf-time determination and tting of various mechanisticmodels, uorescence data were normalized, using both a zero-point offset (in case of seeded experiments) and end-pointnormalization. For data sets with multiple uorescence transi-tions, only the rst transition, corresponding to Ab42 aggrega-tion, was normalized; for this, the intermediate plateau phasewas used as the aggregation endpoint.Aggregation half-times were determined by performinga linear t to the normalized ThT uorescence data for theperiod in which the signal intensity was between 0.4 and 0.6.The half-time was taken to be the time at which the linear tequals 0.5. When applicable, the scaling exponent was deter-mined by plotting the half-time versus the relevant startingmonomer concentration and tting to this data a powerfunctiont1/2 ¼ a[m]g0where t1/2 is the half-time, [m]0 is the starting monomerconcentration, a is a proportionality constant, and g is thescaling exponent.The tting of mechanistic models was used to assess whichof the microscopic Ab42 aggregation steps are perturbed in Ab42/Ab38 and Ab42/Ab37 binary mixtures. For this, the data from thesample containing only Ab42 was tted rst, using the inte-grated rate law½M�½M�N¼ 1 ��1 � ½M�0½M� 0�e�kNt ��B� þ CþektBþ þ Cþekt � Bþ þ CþB� þ Cþ�kN2kkNfor which the parameters are denedk ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi2½m�0kþ½m�0n2k21 þ ½m�0n2�KMsl ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi2kþkn½m�0ncqC� ¼ kþ½P�0k� kþ½M�02½m�0kþ� l22k2kN ¼ 2k+[P]NkN ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffikN2 � 2CþC�k2pB� ¼ kN � kN2kwere [m]0 is the initial monomer concentration; [P]0 is the brilnumber at the start of the reaction; [P]N is the bril number atequilibrium, when the reaction has reached completion (see ref.38 for the detailed expression of [P]N); [M]0 is the bril massconcentration at the start of the reaction; [M]N is the bril massconcentration at equilibrium; kn, k2, and k+ are the rateconstants for primary nucleation, secondary nucleation, andelongation, respectively; KM is the saturation constant forsecondary nucleation; and nc and n2 are the reaction orders ofprimary and secondary nucleation, respectively.For this tting, the primary and secondary nucleation reac-tion orders (nc and n2, respectively) both set to 2, based onpreviously reported data for Ab42 under similar conditions,37,41and the initial seed concentration was set to 0. The rateconstants k+, kn, and k2 (corresponding to elongation, primarynucleation,andsecondarynucleation,respectively)weresimultaneously tted to the data, thus accounting for any minordifferences in the starting monomer concentration. For theanalysis of each binary aggregation data set, three different tswere performed with one rate constant as a variable parameterand the other two rate constants xed to the values obtained forAb42 alone. The comparison of the three ts reveals whetherselective perturbation of a given process could accurately modelthe observed inhibition of Ab42 aggregation.Monomer depletion by HPLCThe samples used for monomer depletion studies wereprepared and incubated as described above, with aggregationmonitored by ThT uorescence. All samples were prepared toa concentration of 3 mM each Ab42, Ab40, and either Ab38 or Ab37and plated in 100 mL aliquots. One sample was removed intoa low-binding tube (Axygen) at the start of incubation, withadditional samples removed during the initial lag phase, therst uorescence transition, the intermediate plateau phase,the second uorescence transition, and the nal plateau phase.In addition to the removed samples, two additional sampleswere le in the plate to monitor the entire aggregation process.Upon removal, each sample was centrifugated at 20 000 � gfor 5 min at room temperature to sediment the aggregatedpeptide. Following centrifugation, the upper 50 mL of thesupernatant was removed and injected on an analytical HPLCinstrument (Agilent 1100), equipped with a C8 column (AgilentZORBAX 300SB, 200 � 4.6 mm), run at a ow rate of 1 mL min�1at 70 �C. The mobile phases used were (solvent A) 0.1% formicacid in water and (solvent B) 0.1% formic acid in acetonitrile. Alinear gradient of 15–30% B over 10 min, followed by an iso-cratic ow of 30% B for an additional ve min was used. The2436 | Chem. Sci., 2022, 13, 2423–2439© 2022 The Author(s). Published by the Royal Society of ChemistryChemical ScienceEdge ArticleOpen Access Article. Published on 07 February 2022. Downloaded on 3/14/2022 5:09:18 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence.View Article Onlineapproximate retention times for each peptide were as follows:Ab42 ¼ 13.7 min, Ab40 ¼ 12.0 min, Ab38 ¼ 10.6 min, Ab37 ¼10.8 min.Fibril composition by HPLCSamples were prepared and incubated as described above, withaggregation monitored by ThT uorescence. All samples wereprepared to a concentration of 3 mM each of Ab42, Ab40, andeither Ab38 or Ab37, and plated in 125 mL aliquots. Samples wereremoved at the start of incubation, during the intermediateplateau phase, and the nal plateau phase. At each time point,the contents of two wells were removed and combined to givea total sample volume of 250 mL. In addition to the removedsamples, two additional samples were le in the plate tomonitor the aggregation process.Upon removal, samples were ltrated through 0.2 mm spinlters (VIVASPIN 500) at 15 000 � g for 5 min at room temper-ature to trap the brils. The lter-trapped brils were thenwashed ve times with 500 mL ultrapure water (to a total washvolume of 10x the sample volume) to remove any monomeric oroligomeric peptide on the bril and lter surface. Aer eachwash, the sample was centrifugated at 15 000 � g for 5 min atroom temperature, leaving a retentate volume of ca. 25 mL. Aerthe nal wash, the retentate was mixed in a 1 : 3 ratio with 6 MGuHCl to dissolve the brils for 15 min before being injected onan analytical HPLC instrument. HPLC was performed asdescribed above.Cryo-EMFor Cryo-EM samples, monomer isolation for all peptides wasperformed as described above. Solutions were prepared witha total monomeric peptide concentration of 10 mM and incu-bated in the same manner as for the kinetic assays, withaggregation monitored by ThT uorescence. Samples weretaken once the nal uorescence plateau was reached. Speci-mens were prepared for imaging in a controlled environmentvitrication system (CEVS) to maintain stable temperature andto minimize solution loss. The sample was prepared as a thinliquid lm, <300 nm thick, on lacey carbon lmed copper grids.This was plunged into liquid ethane at �180 �C to vitrify thesample; this minimizes water crystallization as well as compo-nent segmentation and rearrangement, thus maintaining orig-inal microstructures. The vitried sample was stored underliquid nitrogen until imaged. The grid was transferred into theelectron microscope (JEM 2200FS) using a Fischione Model2550 cryo transfer tomography holder. The microscope wasequipped with an in-column energy lter (Omega lter), thusallowing for zero-loss imaging. The acceleration voltage was 200kV and zero-loss images were recorded digitally with a TVIPSF416 camera using SerialEM under low dose conditions andwith a 30 eV energy selecting slit in place.To quantify bril ultrastructure, the node-to-node distanceand bril diameter were measured using ImageJ (version 2.1.0/1.53c, NIH). For each sample, at least 50 measurements wereperformed, measuring at least een total brils from vedifferent images, each taken from different parts of the samplegrid. Measurements are presented in boxplots with Tukeywhiskers, with the median represented by the center line, thebox containing the 25th–75th percentiles, and individuallyplottedpointsrepresentingstatisticaloutliers.Statisticalsignicance was assessed using a one-way ANOVA followed byDunnett's post-hoc test, with p < 0.05 considered signicant. Allstatistical analysis was performed in GraphPad Prism (version9.0.2).Data availabilityAll data presented in this article will be made available uponreasonable request.Author contributionsH. Z. and S. L. designed the study. G. A. B., K. S. and S. L. per-formed the experiments. G. A. B. and A. J. D. analyzed data. G. A.B. wrote the paper with input from all co-authors.Conflicts of interestH. Z. has served at scientic advisory boards for Eisai, Denali,Roche Diagnostics, Wave, Samumed, Siemens Healthineers,Pinteon Therapeutics, Nervgen, AZTherapies and CogRx, hasgiven lectures in symposia sponsored by Cellectricon, Fujirebio,Alzecure and Biogen, and is a co-founder of Brain BiomarkerSolutions in Gothenburg AB (BBS), which is a part of the GUVentures Incubator Program (outside submitted work).AcknowledgementsCryo-EM images were obtained with expert assistance fromAnna Carnerup (Lund University). This work was funded by theSwedish Research Council (grants #2015-00143 to S. L. and#2018-02532 to H. Z.), the European Research Council (#681712toH.Z.),SwedishStateSupportforClinicalResearch(#ALFGBG-720931 to H. Z.), the Lindemann Trust Fellowship (A.J. D.), the Novo Nordisk Foundation (#NNF19OC0054635 to S.L.), and the Fulbright U.S. Student Program (G. A. B). H. Z. isa Wallenberg Scholar.References1 C. A. Lane, J. Hardy and J. M. Schott, Eur. J. Neurol., 2018, 25,59–70.2 Developing therapeutics for Alzheimer's disease: progress andchallenges, ed., M. S. Wolfe, Elsevier/AP, Academic Press isan imprint of Elsevier, Amsterdam; Boston, 2016.3 A. B. Reiss, H. A. Arain, M. M. Stecker, N. M. Siegart andL. J. Kasselman, Rev. Neurosci., 2018, 29, 613–627.4 J. L. Cummings, G. Tong and C. Ballard, J. Alzheimer's Dis.,2019, 67, 779–794.5 J. Sevigny, P. Chiao, T. Bussi`ere, P. H. Weinreb, L. Williams,M. Maier, R. Dunstan, S. Salloway, T. Chen, Y. Ling,J. O'Gorman, F. Qian, M. Arastu, M. Li, S. Chollate,M. S. Brennan, O. Quintero-Monzon, R. H. Scannevin,© 2022 The Author(s). Published by the Royal Society of ChemistryChem. Sci., 2022, 13, 2423–2439 | 2437Edge ArticleChemical ScienceOpen Access Article. Published on 07 February 2022. Downloaded on 3/14/2022 5:09:18 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence.View Article OnlineH. M. Arnold, T. Engber, K. Rhodes, J. Ferrero, Y. Hang,A. Mikulskis, J. Grimm, C. Hock, R. M. Nitsch andA. Sandrock, Nature, 2016, 537, 50–56.6 S. B. Haeberlein, Presented in part at the 12th Clinical Trials onAlzheimer's Disease (CTAD) Congress, San Diego, CA, 2019.7 S.Linse,T.Scheidt,K.Bernfur,M.Vendruscolo,C. M. Dobson, S. I. A. Cohen, E. Sileikis, M. Lundqvist,F. Qian, T. O’Malley, T. Bussiere, P. H. Weinreb, C. K. Xu,G.Meisl,S.R.A.Devenish,T.P.J.KnowlesandO. Hansson, Nat. Struct. Mol. Biol., 2020, 27, 1125–1133.8 “FDA Grants Accelerated Approval for Alzheimer's Drug.” June2021.https://www.fda.gov/news-events/press-announcements/fda-grants-accelerated-approval-alzheimers-drug. Accessed October 2021.9 J. Hardy and D. Allsop, Trends Pharmacol. Sci., 1991, 12, 383–388.10 J. Hardy and G. Higgins, Science, 1992, 256, 184–185.11 E. N. Cline, M. A. Bicca, K. L. Viola and W. L. Klein, J.Alzheimer's Dis., 2018, 64, S567–S610.12 M. G. Iadanza, M. P. Jackson, E. W. Hewitt, N. A. Ranson andS. E. Radford, Nat. Rev. Mol. Cell Biol., 2018, 19, 755–773.13 S. Linse, Biophys. Rev., 2017, 9, 329–338.14 M. T¨ornquist, T. C. T. Michaels, K. Sanagavarapu, X. Yang,G. Meisl, S. I. A. Cohen, T. P. J. Knowles and S. Linse,Chem. Commun., 2018, 54, 8667–8684.15 R. M. Ransohoff, Science, 2016, 353, 777–783.16 V. Calsolaro and P. Edison, Alzheimer's Dementia, 2016, 12,719–732.17 F. Kametani and M. Hasegawa, Front. Neurosci., 2018, 12, 25.18 G. S. Bloom, JAMA Neurol., 2014, 71, 505.19 G.K.Gouras,T.T.OlssonandO.Hansson,Neurotherapeutics, 2015, 12, 3–11.20 D. J. Selkoe and J. Hardy, EMBO Mol. Med., 2016, 8, 595–608.21 P.Arosio,M.Vendruscolo,C.M.DobsonandT. P. J. Knowles, Trends Pharmacol. Sci., 2014, 35, 127–135.22 Y. Zhang, R. Thompson, H. Zhang and H. Xu, Mol. Brain,2011, 4, 3.23 C. Haass, C. Kaether, G. Thinakaran and S. Sisodia, ColdSpring Harbor Perspect. Med., 2012, 2, a006270.24 K.G.Mawuenyega,T.Kasten,W.SigurdsonandR. J. Bateman, Anal. Biochem., 2013, 440, 56–62.25 N. Kaneko, R. Yamamoto, T.-A. Sato and K. Tanaka, Proc. Jpn.Acad., Ser. B, 2014, 90, 104–117.26 J. Reinert, B. C. Richard, H. W. Klai, B. Friedrich,T. A. Bayer, J. Wiltfang, G. G. Kovacs, M. Ingelsson,L. Lannfelt, A. Paetau, J. Bergquist and O. Wirths, ActaNeuropathol. Commun., 2016, 4, 24.27 G. Brinkmalm, W. Hong, Z. Wang, W. Liu, T. T. O'Malley,X.Sun,M.P.Frosch,D.J.Selkoe,E.Portelius,H. Zetterberg, K. Blennow and D. M. Walsh, Brain, 2019,142, 1441–1457.28 M. Bibl, B. Mollenhauer, H. Esselmann, P. Lewczuk,H.-W.Klai,K.Sparbier,A.Smirnov,L.Cepek,C. Trenkwalder, E. R¨uther, J. Kornhuber, M. Otto andJ. Wiltfang, Brain, 2006, 129, 1177–1187.29 J. Wiltfang, H. Esselmann, M. Bibl, A. Smirnov, M. Otto,S. Paul, B. Schmidt, H.-W. Klai, M. Maler, T. Dyrks,M. Bienert, M. Beyermann, E. R¨uther and J. Kornhuber, J.Neurochem., 2002, 81, 481–496.30 D. Galasko, L. Chang, R. Motter, C. M. Clark, J. Kaye,D. Knopman, R. Thomas, D. Kholodenko, D. Schenk,I. Lieberburg, B. Miller, R. Green, R. Basherad, L. Kertiles,M. A. Boss and P. Seubert, Arch. Neurol., 1998, 55, 937.31 P.E.Spies,M.M.Verbeek,T.vanGroenandJ. A. H. R. Claassen, Front. Biosci., Landmark Ed., 2012, 17,2024–2034.32 S. Zampar, H. W. Klai, K. Sritharen, T. A. Bayer, J. Wiltfang,A. Rostagno, J. Ghiso, L. A. Miles and O. Wirths, Neuropathol.Appl. Neurobiol., 2020, 46, 673–685.33 O. Szczepankiewicz, B. Linse, G. Meisl, E. Thulin, B. Frohm,C. Sala Frigerio, M. T. Colvin, A. C. Jacavone, R. G. Griffin,T. Knowles, D. M. Walsh and S. Linse, J. Am. Chem. Soc.,2015, 137, 14673–14685.34 T. Weiffert, G. Meisl, P. Flagmeier, S. De, C. J. R. Dunning,B.Frohm,H.Zetterberg,K.Blennow,E.Portelius,D. Klenerman, C. M. Dobson, T. P. J. Knowles and S. Linse,ACS Chem. Neurosci., 2019, 10, 2374–2384.35 M. P. Murphy and H. LeVine, J. Alzheimer's Dis., 2010, 19,311–323.36 A. Vandersteen, E. Hubin, R. Sarroukh, G. De Baets,J.Schymkowitz,F.Rousseau,V.Subramaniam,V.Raussens,H.Wenschuh,D.WildemannandK. Broersen, FEBS Lett., 2012, 586, 4088–4093.37 S. I. A. Cohen, S. Linse, L. M. Luheshi, E. Hellstrand,D. A. White, L. Rajah, D. E. Otzen, M. Vendruscolo,C. M. Dobson and T. P. J. Knowles, Proc. Natl. Acad. Sci. U.S. A., 2013, 110, 9758–9763.38 G. Meisl, X. Yang, E. Hellstrand, B. Frohm, J. B. Kirkegaard,S. I. A. Cohen, C. M. Dobson, S. Linse and T. P. J. Knowles,Proc. Natl. Acad. Sci. U. S. A., 2014, 111, 9384–9389.39 A. Abelein, J. Jarvet, A. Barth, A. Gr¨aslund and J. Danielsson,J. Am. Chem. Soc., 2016, 138, 6893–6902.40 G.Meisl,X.Yang,C.M.Dobson,S.LinseandT. P. J. Knowles, Chem. Sci., 2017, 8, 4352–4362.41 G. Meisl, X. Yang, B. Frohm, T. P. J. Knowles and S. Linse, Sci.Rep., 2016, 6, 18728.42 R. Sabat´e, A. Espargar´o, L. Barbosa-Barros, S. Ventura andJ. Estelrich, Biochimie, 2012, 94, 1730–1738.43 A. K. Srivastava, J. M. Pittman, J. Zerweck, B. S. Venkata,P. C. Moore, J. R. Sachleben and S. C. Meredith, ProteinSci., 2019, 28, 1567–1581.44 C. Cabaleiro-Lago, F. Quinlan-Pluck, I. Lynch, S. Lindman,A. M. Minogue, E. Thulin, D. M. Walsh, K. A. Dawson andS. Linse, J. Am. Chem. Soc., 2008, 130, 15437–15443.45 K. Ezzat, M. Pernemalm, S. P˚alsson, T. C. Roberts, P. J¨arver,A. Dondalska, B. Bestas, M. J. Sobkowiak, B. Lev¨anen,M.Sk¨old,E.A.Thompson,O.Saher,O.K.Kari,T.Lajunen,E.SverremarkEkstr¨om,C.Nilsson,Y. Ishchenko, T. Malm, M. J. A. Wood, U. F. Power,S. Masich, A. Lind´en, J. K. Sandberg, J. Lehti¨o, A.-L. Spetzand S. EL Andaloussi, Nat. Commun., 2019, 10, 2331.46 E. R. Padayachee, H. Zetterberg, E. Portelius, J. Bor´en,J. L. Molinuevo, N. Andreasen, R. Cukalevski, S. Linse,2438 | Chem. Sci., 2022, 13, 2423–2439© 2022 The Author(s). Published by the Royal Society of ChemistryChemical ScienceEdge ArticleOpen Access Article. Published on 07 February 2022. Downloaded on 3/14/2022 5:09:18 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence.View Article OnlineK. Blennow and U. Andreasson, Brain Res., 2016, 1651, 11–16.47 R.Frankel,M.T¨ornquist,G.Meisl,O.Hansson,U. Andreasson, H. Zetterberg, K. Blennow, B. Frohm,T. Cedervall, T. P. J. Knowles, T. Leiding and S. Linse,Commun. Biol., 2019, 2, 365.48 R. Cukalevski, X. Yang, G. Meisl, U. Weininger, K. Bernfur,B. Frohm, T. P. J. Knowles and S. Linse, Chem. Sci., 2015,6, 4215–4233.49 B. D. Moore, J. Martin, L. de Mena, J. Sanchez, P. E. Cruz,C.Ceballos-Diaz,T.B.Ladd,Y.Ran,Y.Levites,T. L. Kukar, J. J. Kurian, R. McKenna, E. H. Koo,D. R. Borchelt, C. Janus, D. Rincon-Limas, P. Fernandez-Funez and T. E. Golde, J. Exp. Med., 2018, 215, 283–301.50 M. O. Quartey, J. N. K. Nyarko, J. M. Maley, J. R. Barnes,M.A.C.Bolanos,R.M.Heistad,K.J.Knudsen,P.R.Pennington,J.Buttigieg,C.E.DeCarvalho,S. C. Leary, M. P. Parsons and D. D. Mousseau, Sci. Rep.,2021, 11, 431.51 E. Portelius, U. Andreasson, J. M. Ringman, K. Buerger,J.Daborg,P.Buchhave,O.Hansson,A.Harmsen,M. K. Gustavsson, E. Hanse, D. Galasko, H. Hampel,K. Blennow and H. Zetterberg, Mol. Neurodegener., 2010, 5, 2.52 E.Portelius,B.VanBroeck,U.Andreasson,M. K. Gustavsson, M. Mercken, H. Zetterberg, H. Borghysand K. Blennow, J. Alzheimer's Dis., 2010, 21, 1005–1012.53 T. C. T. Michaels, A. ˇSari´c, S. Curk, K. Bernfur, P. Arosio,G. Meisl, A. J. Dear, S. I. A. Cohen, C. M. Dobson,M. Vendruscolo, S. Linse and T. P. J. Knowles, Nat. Chem.,2020, 12, 445–451.54 A. Jan, O. Gokce, R. Luthi-Carter and H. A. Lashuel, J. Biol.Chem., 2008, 283, 28176–28189.55 M. M. Murray, S. L. Bernstein, V. Nyugen, M. M. Condron,D. B. Teplow and M. T. Bowers, J. Am. Chem. Soc., 2009,131, 6316–6317.56 S. G. Bolder, L. M. C. Sagis, P. Venema and E. van der Linden,Langmuir, 2007, 23, 4144–4147.57 R. Silvers, M. T. Colvin, K. K. Frederick, A. C. Jacavone,S. Lindquist, S. Linse and R. G. Griffin, Biochemistry, 2017,56, 4850–4859.58 D. M. Walsh, E. Thulin, A. M. Minogue, N. Gustavsson,E. Pang, D. B. Teplow and S. Linse, FEBS J., 2009, 276,1266–1281.59 G. Meisl, J. B. Kirkegaard, P. Arosio, T. C. T. Michaels,M.Vendruscolo,C.M.Dobson,S.LinseandT. P. J. Knowles, Nat. Protoc., 2016, 11, 252–272.60 G. Meisl, T. C. T. Michaels, S. Linse and T. P. J. Knowles, inAmyloid Proteins, ed. E. M. Sigurdsson, M. Calero and M.Gasset, Springer New York, New York, NY, 2018, vol. 1779,pp. 181–196.61 G. Meisl, L. Rajah, S. A. I. Cohen, M. Pfammatter, A. ˇSari´c,E.Hellstrand,A.K.Buell,A.Aguzzi,S.Linse,M. Vendruscolo, C. M. Dobson and T. P. J. Knowles, Chem.Sci., 2017, 8, 7087–7097.62 A. J. Dear, G. Meisl, T. C. T. Michaels, M. R. Zimmermann,S. Linse and T. P. J. Knowles, J. Chem. Phys., 2020, 152,045101.63 A. Munke, J. Persson, T. Weiffert, E. De Genst, G. Meisl,P. Arosio, A. Carnerup, C. M. Dobson, M. Vendruscolo,T. P. J. Knowles and S. Linse, Proc. Natl. Acad. Sci. U. S. A.,2017, 114, 6444–6449.64 C. J. Sarell, P. G. Stockley and S. E. Radford, Prion, 2013, 7,359–368.65 M.I.Ivanova,Y.Lin,Y.-H.Lee,J.ZhengandA. Ramamoorthy, Biophys. Chem., 2021, 269, 106507.66 S. Chia, P. Flagmeier, J. Habchi, V. Lattanzi, S. Linse,C. M. Dobson, T. P. J. Knowles and M. Vendruscolo, Proc.Natl. Acad. Sci. U. S. A., 2017, 114, 8005–8010.67 S. Bondarev, K. Antonets, A. Kajava, A. Nizhnikov andG. Zhouravleva, Int. J. Mol. Sci., 2018, 19, 2292.68 F. Hasecke, C. Niyangoda, G. Borjas, J. Pan, G. Matthews,M. Muschol and W. Hoyer, Angew. Chem., Int. Ed., 2021,60, 3016–3021.69 J. C. Kessler, J.-C. Rochet and P. T. Lansbury, Biochemistry,2003, 42, 672–678.70 S. Abeln and D. Frenkel, PLoS Comput. Biol., 2008, 4,e1000241.71 C. Arber, J. Toombs, C. Lovejoy, N. S. Ryan, R. W. Paterson,N. Willumsen, E. Gkanatsiou, E. Portelius, K. Blennow,A. Heslegrave, J. M. Schott, J. Hardy, T. Lashley, N. C. Fox,H. Zetterberg and S. Wray, Mol. Psychiatry, 2020, 25, 2919–2931.72 S. Mekala, G. Nelson and Y.-M. Li, RSC Med. Chem., 2020, 11,1003–1022.73 M. S. Wolfe, Molecules, 2021, 26, 388.74 G. Yang, R. Zhou, X. Guo, C. Yan, J. Lei and Y. Shi, Cell, 2021,184, 521–533.e14.© 2022 The Author(s). Published by the Royal Society of ChemistryChem. Sci., 2022, 13, 2423–2439 | 2439Edge ArticleChemical ScienceOpen Access Article. Published on 07 February 2022. Downloaded on 3/14/2022 5:09:18 PM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence.View Article Online | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/358931367Revista Maya de Geociencias - Marzo 2022Article · March 2022CITATIONS0READS4544 authors:Some of the authors of this publication are also working on these related projects:Paleomagnetism and tectonics of the Trans-Mexican Volcanic Belt View projectFinal emplacement of the Santa Cruz ophiolite in eastern Guatemala View projectLuis ValenciaInstituto Politécnico Nacional33 PUBLICATIONS 0 CITATIONS SEE PROFILEBernardo García-AmadorUniversidad Nacional Autónoma de México50 PUBLICATIONS 68 CITATIONS SEE PROFILEJoshua RosenfeldIndependent Researcher75 PUBLICATIONS 209 CITATIONS SEE PROFILEClaudio BartoliniConsulting Exploration Americas57 PUBLICATIONS 112 CITATIONS SEE PROFILEAll content following this page was uploaded by Claudio Bartolini on 01 March 2022.The user has requested enhancement of the downloaded file.PIONEROS DELAS GEOCIENCIASFUNDADORES DELAGEOLOGÍA MODERNAPIONEROS DELAPALEONTOLOGÍAPROPIEDADESBÁSICAS EN UN YACIMIENTOMARZO2022R E V I S T A D E G E O C I E N C I A SNOTAS GEOLÓGICASTEMAS DE INTERÉSPIONEROS DE LAS GEOCIENCIASMARZO2022R E V I S T A D E G E O C I E N C I A SRevista Maya: Revista Maya de Geociencias que(RMG) nace del entusiasmo de profesionistas con la inquietudde difundir conocimientos relacionados con la academia,inves�gación, la exploración petrolera y Ciencias de la Tierraen general.El obje�vo principal de la revista es proporcionar un espacioa todos aquellos jóvenes profesionistas que deseen dar aconocer sus publicaciones. los fundadores de la revista sonLuis Angel Valencia Flores, Bernardo García Amador y ClaudioBartolini.Otro de los obje�vos de la Revista Maya de Geociencias esincen�var a profesionales, académicos, e inves�gadores, apar�cipar ac�vamente en beneficio de nuestra comunidadjoven de geociencias.La Revista tendrá una publicación mensual, por medio de unarchivo PDF, el cuál será distribuido por correo electrónico ycompar�do en las redes sociales. Esta revista digital no �enefines de lucro. La RMG es internacional y bilingue. Si deseaspar�cipar o contribuir con algún manuscrito, por favorcomunícate con cualquiera de los editores.Las notas geológicas �enen como obje�vo el presentarsíntesis de trabajos realizados en México y en diferentespartes del mundo por jóvenes profesionales y pres�giososgeocien�ficos. Son notas escencialmente de divulgación,con resultados y conocimientos nuevos, en beneficio denuestra comunidad de geociencias. Estas notas no estánsujetas a arbitraje.Revista Maya: The Revista Maya de Geociencias(RMG) springs from the enthusiasm of professionals witha desire to distribute knowledge related to academicresearch, explora�on for resources and geoscience ingeneral.The main objec�ve of the RMG is to provide a place foryoungprofessionalswhowishtodistributetheirpublica�ons. The founders of the Revista are Luis ÁngelValencia Flores, Bernardo García and Claudio Bartolini.Afurtherobjec�veoftheRMGistoencourageprofessionals, academicians and researchers to ac�velypar�cipate for the benefit of our community of younggeoscien�sts.The RMG is published monthly as a PDF file distributed byemail and shared through social media. This digitalmagazine has no commercial aim. It is interna�onal andbilingual (Spanish and English). If one wishes to par�cipateor contribute a manuscript, please contact any of theeditors.The geological notes aim to synthesize work carried out inMexico and other parts of the world both by youngprofessionals and pres�gious geoscien�sts. These notesare produced principally to reveal new understandings forthe benefit of our geoscien�fic community and are notsubjected to peer review.*Es importante aclarar, que las opiniones científicas, comerciales, culturales,sociales etc., no son responsabilidad, ni son compartidas o rechazadas,por los editores de la revista.Portada de la revista: fracturas de sal, en un lago ubicado en Irán, alsur de Isfahán. Autor: Fotogra�a de Francisco Javier Vega.Revista de divulgaciónGeocientíficaEDITORESBernardo García-Amador es candidato a doctoren Ciencias de la Tierra por la UNAM. Su geo-pasión es entender las causas y consecuenciasde la tectónica. Actualmente se encuentra enproceso de graduarse deldoctorado, con untrabajo que versa en la evolucion tectónica deNicaragua (Centroamérica). Además imparte elcurso de tectónica en la Facultad de IngenieríadelaUNAM.RecientementeBernardohapublicado parte de su trabajo de doctorado enlas revistas Tectonics y Tectonophysics, ademásde ser coautor de otros artículos científicos dedis�ntos proyectos.bernardo.garcia@ingenieria.unam.eduLuis Angel Valencia Flores (M.C.). IngenieroGeólogo y Maestro en Ciencias en Geología,egresado de la Escuela Superior de Ingeniería yArquitectura-Unidad Ticomán. Ha trabajado enel IMP, Pemex Ac�vo Integral Litoral de Tabasco,Schlumberger, Paradigm Geophysical, ComisiónNacionaldeHidrocarburos,AspectEnergyHoldings LLC, actualmente es académico del IPN(posgradoylicenciatura)ylaUNAM(licenciatura)impar�endolasmateriasdeEvaluación de formaciones, Caracterización deyacimientos,Geologíadeyacimientos,Geoquímica, entre otras del ramo petrolero.Cuenta con experiencia de 20 años trabajandoendiversosproyectosdeplaneaciónyperforación de campos, pozos costa afuera,petro�sica, geomodelado y caracterización deyacimientos entre ellos: Cantarell, Sihil, Xanab,Yaxche, Sinan, Bolon�ku, May, Onixma, Faja deoro, campos de Brasil, Bolivia y Cuba. ComoDirector General Adjunto en la CNH fue parte delequipo editor técnico en la generación de losAtlas de las Cuencas de México, par�cipó comoponente del Gobierno de México en eventospetroleros de Canadá, Inglaterra y EstadosUnidos.EsTechnicalAdvisordelCapítuloestudian�l de la AAPG-IPN.luis.valencia.11@outlook.comJosh Rosenfeld (Ph.D.). He obtained an M.A.from the University of Miami in 1978, and a Ph.D.from Binghamton University in 1981. Josh joinedAmoco Produc�on Company as a petroleumgeologist working from 1980 to 1999 in Houston,Mexico and Colombia. Upon re�ring fromAmoco, Josh was employed by Veritas DGC un�l2002 on explora�on projects in Mexico. He hasbeen a member of HGS since 1980 and AAPGsince 1981, and currently does geology from hishome in Granbury, Texas.jhrosenfeld@gmail.comClaudio Bartolini (Ph.D.) is presently a seniorexplora�on advisor at Petroleum Explora�onConsultants Americas. He has more than 25years of experience in both domes�c andinterna�onal mining and petroleum explora�on,mainly in the United States and La�n America.Claudio is an associate editor for the AAPGBulle�n and he has edited several books on thepetroleum geology of the Americas. He is aCorrespondent member of the Academy ofEngineering of Mexico.Claudio was made an Honorary Member of theAAPG in 2022 in recogni�on of his service to theAssocia�on, and his devo�on to the science andprofession of petroleum geology.bartolini.claudio@gmail.comSalvador Ortuño Arzate received his M. Sc.from the Na�onal Autonomous University ofMexico (UNAM) and his Ph.D. from theUniversité de Pau and Pays de l'Adour (UPPA)in France. He has been a researcher at theIns�tuto Mexicano del Petróleo and theIns�tut Français du Pétrole, focusing his workon the Explora�on Petroleum field. Salvadorhas published several papers and a book, “ElMundo del Petróleo” (Petroleum´s world),examining and shedding light on the history ofpetroleum and the implica�ons for the society.Also, he has worked as an advisor for severaluniversi�es and na�onal corpora�ons. Lastly, hehas served as faculty and has taught differentcourses at the Secretariat of Na�onal Defenseand at the Engineering School of U.N.A.M.soaortuno@gmail.comElingenierocubanoHumbertoÁlvarezSánchez culmina 54 años como geólogo.RealizóestudiosenlaCordilleradeGuaniguanico y en su premontaña y en losmacizos metamórficos, volcánicos y ofiolí�cosde Cuba central. Autor de 18 formaciones ylitodemasdelaestra�gra�acubana.Descubridor del único depósito industrial defosforitas marinas de Cuba. Miembro de lasubcomisiónJurásicodelprimerLéxicoEstra�gráficodeCuba.ComoCountryManager y Senior Geologist de compañíascanadienses, panameñas y de Estados Unidos,dirigióexploracionesencomplejosdelPaleozoico-Mesozoico en tres Estados deBrasil, en los greenstone belts de Uruguay;Andes de Perú y complejos volcánicos deHonduras y Panamá y otros países. Miembrode la Comisión Ministerial “Ad Honorem” delPlan Maestro de Minería de Panamá, fueConsultor Senior del Banco Interamericano deDesarrollo para el proyecto geocien�fico delpaís. Formely Miembro del Consejo Cien�ficodeGeologyWhithoutLimits.FormerlyRepresentanteparaAméricaCentraldelServicioGeológicodelaGranBretaña.Re�rado en Panamá, se ocupa de redactarestudios sobre la geología de Cuba.geodoxo@gmail.comRamón López Jiménez es un geólogo con 14años de experiencia en inves�gación y envarios sectores de la industria y serviciospúblicos. Es un especialista en obtención dedatos en campo, su análisis y su conversión adiversos productos finales. Ha trabajado enEEUU,Mexico,Colombia,ReinoUnido,Turquía y España. Su especialidad es lasedimentología marina de aguas profundas.Actualmenterealizainves�gaciónenafloramientos an�guos de aguas someras yprofundas de México, Turquía y Marruecos encolaboraciónconen�dadespúblicasyprivadas de esos países. Es instructor decursos de campo y oficina en arquitectura deyacimientos de aguas profundas y tectónicasalina por debajo de la resolución sísmica.r.lopez.jimenez00@aberdeen.ac.ukCOLABORADORESJoséAntonioRodríguezArteagaesuningeniero geólogo con 31 años de experienciaen inves�gación de geología de terremotos yriesgogeológico,asociadoonoalasismicidad. Es especialista en sismologíahistóricaehistoriadelossismosenVenezuela,recibiendoentrenamientoprofesional en Geomá�ca Aplicada a laZonificación de Riesgos, Bogotá, Colombia.En sus inicios profesionalesy por 5 añosconsecu�vos,fuegeólogodecampo,trabajando en prospección de yacimientosminerales no- metálicos en la región centrooccidental de Venezuela. Tiene en su haber comoautor, coautor o coordinador, tres libros dedicadosa la catalogación sismológica del siglo XX, alpensamiento sismológico venezolano y un Atlasgeológico de la región central del país, preparado demanera conjunta con la Escuela de Geología, Minasy Geo�sica de la Universidad Central de Venezuela.Actualmente prepara un cuarto texto sobre losestudios de un inquieto naturalista alemán del sigloXIX y sus informes para los terremotos destructoresen Venezuela de los años 1812, 1894 y 1900.rodriguez.arteaga@gmail.comMarisol Polet Pinzón Sotelo es IngenieraGeólogaegresadadelaUniversidadAutónomadeGuerreroyMaestraenCienciasGeológicasdelaUniversidadAutónoma de Nuevo León; ha colaborado enproyectos de inves�gación en el noroeste deMéxico,siendoautoraycoautoradepublicaciones cien�ficas; cuenta con 8 añosdeexperienciaenexploracióndehidrocarburosenPEMEXExploraciónyProducción.SehadesarrolladoenelmodeladodesistemaspetrolerosenProyectos de aguas profundas y someras enel norte del Golfo de México.poletpinzon@gmail.comRafaelGuardadoesgraduadoenlaUniversidaddeOrienteen1970comoIngenieroGeólogo.Cursóestudiosdeespecialización en la Universidad Minera de StPetersburgoenRusia,an�guaU.R.S.S.,1972-1974.DefendióeldoctoradoenGeología en 1983. Es Académico Titular de laAcademia de Ciencias de Cuba, ProfesorTitular,ProfesorConsulyanteyProfesorEmerito de la Universidad De Moa. OrdenCarlos J. Finlay. Ha publicado más de 70 ar�culos,y es Tutor de tesis de Doctorado y maestrías. Harecibido múl�ples premios y dis�nciones, y es unprofesor reconocido en Cuba y el exprenjero en laIngeniería Geológica, la Reducción de los RiesgosGeológicosyelenfrentamientoalCambioClimá�co.rafaelguardado2008@gmail.comJon Blickwede egresó de la Universidad deTu�s en Boston, Massachuse�s, EEUU con unBachillerato en Ciencias de la Tierra en 1977.Entró a la Universidad de New Orleans,Louisiana en 1979, donde hizo su tésis deMaestría en Geología sobre la FormaciónNazas en la Sierra de San Julián, Zacatecas,México. Jon comenzó su carrera en 1981,trabajando por 35 años como geólogo deexploración petrolera para varias compañiastal como Amoco, Unocal, y Statoil. Realizóproyectos de geología sobre EEUU, México,Centroamerica y el Caribe para estas empresas.Durante 2018, Jon fundó la empresa TeyraGeoConsul�ng LLC (www.teyrageo.com), dondeestárealizandounproyectodecrearafloramientos digitales y excursiones geológicasvirtuales en EEUU y México, u�lizando imágenestomados con su drone, integrados con otrosdatos geoespaciales.jonblickwede@gmail.comInstrucciones básicas para los autoresApreciables autores, al someter su material para la publicación en laRevista Maya de Geociencias, por favor mantener los siguienteslineamientos editoriales de su manuscrito al momento de enviarlo alequipo editorial y colaboradores:Semblanzas: 3 páginas máximo.Notas sobre pioneros de las geociencias: 4 páginas máximo.Los "temas de interés para la comunidad": 4 páginas máximo.Notas geológicas: 10 páginas máximo.Nuestro agradecimiento a Manuel Arribas, un gran fotógrafo y excelente diseñador gráficoEspañol, por la creación del nuevo logotipo de la Revista Maya de Geociencias y sus indicacionespara la compaginación de la misma.https://manuelarribas.es/Miguel Vazquez Diego Gabriel, es estudiantede la carrera de Ingeniería Geológica en laUniversidad Nacional Autónoma de México(Facultad de Ingeniera), sus principales areasde interés a lo largo de la carrera han sido latectónica, geoquímica y mineralogía. Es unentusiasta de la divulgación cien�fica, sobre todoen el área de las Ciencias de la Tierra.diegogabriel807@gmail.comSaúl Humberto Ricardez Medina es pasantede Ingeniería Geológica, miembro ac�vo delcapítulo estudian�l de la AAPG del Ins�tutoPolitécnicoNacional,par�cipóenelXCongreso Nacional de Estudiantes de Cienciasde la Tierra como Expositor del trabajo“Análisis de Backstripping de la Cuenca SalinadelIstmo”.Actualmente,seencuentratrabajandoensutesisdelicenciaturarelacionadaaiden�ficaryreconocersecuenciassedimentariaspotencialmentealmacenadorasdehidrocarburosenlascuencas del sureste.ricardezmedinasaulhumberto@gmail.com“Hay una fuerza motriz más poderosa que elvapor, la electricidad y la energía atómica: lavoluntad.”EinsteinSemblanzas………………………………………………………………..8Pioneros delasGeociencias……………………………………………..12Resúmenesdetesisypublicaciones……………………………………..16Loslibros recomendados………………………………………………..27Temasdeinterés………………………………………………………….29Fotografíasdeafloramientos/microscopio…………………….………46Notasgeológicas………………………………………………………….49MisceláneosMuseosdehistorianatural………………………………………………81Lacasadelosarrecifes…………………………………………………..82Tesisselectasdel2020-UNAM………………………………………..83LaCavernadelarte……………………………………………………….84Nuevolibro (Moisés Dávila Serrano)……………………………………..86Glosario detérminosgeológicos………………………………….……87Geo-caricatura (Wilmer Pérez Gil).................................................88CristalesgigantesdeNaica,Chihuahua,México………………………89DeCubaaCanarias,entrelíneasycolores……………………………..90Asociaciones geológicashermanas….………….…….….…………..……….…….91CONTENIDO MARZO20228SEMBLANZASMarie Tharp was a pioneering American geologist andoceanographic cartographer who helped create the firstscien�ficmapoftheAtlan�cOceanfloor.Herrevolu�onary work revealed the detailed topography andmul�-dimensional geographic landscape of the oceanbo�om and revealed the presence of a con�nuous ri�valley along the axis of the Mid-Atlan�c Ridge, a shi� inthinking that led to the acceptance of the theories of platetectonics and con�nental dri�.She was born in 1920, and spent her youth accompanyingher father, a soil surveyor for the U.S. Department ofAgriculture,givingheranearlyintroduc�ontomapmaking. But young Tharp had no interest in pursuing asimilar career as this was understood to be “men’s work.”She graduated from Ohio University in 1943 with a degreein English and music. The military recruitment of youngmen in WWII gave women an opportunity to exploreprofessions typically reserved for men, and having taken ageology class in college, she was recruited to theUniversity of Michigan’s petroleum geology program. Shecompleted her master’s degree, and then a secondmasters in math from the University of Tulsa. She took ajob at an oil company in Tulsa but found it unsa�sfying asshe was confined to an office where she collected mapsand data for the men going into the field. She moved toNew York City and found dra�ing work as one of the firstwoman at the Lamont Geological Observatory, now theLamont-DohertyEarthObservatoryatColumbiaUniversity. It was here she met Bruce Heezen and wouldeventually work exclusively with him plo�ng the oceanfloor.For the first two decades of their collabora�on, Heezencollected data aboard research ships, while Tharp drewmaps from that data, since women were s�ll excludedfrom working aboard ships at that �me. Working withpens, ink and rulers, Tharp drew the underwater details,longitude degree by la�tude degree, described bythousands of sonar readings taken by researchers, andindependently used data collec�on from the Woods HoleOceanographic Ins�tu�on’s R/V Atlan�s. She painstakinglyaligned sounding profiles from Atlan�s and other vessels,crea�ng a total of approximately six profiles stretchingwest-to-eastacrosstheNorthAtlan�c.Herworkrepresented the first systema�c a�empt to map the en�reocean floor. She was eventually able to join a datacollec�on research expedi�on in 1968.Piecing together maps that they made, she and Heezenrevealed a 40,000-mile underwater ridge girdling theglobe. With the discovery, they laid the founda�on forlater work that showed the sea floor spreads from centralridges and that the con�nents are in mo�on with respectto one another — a revolu�onary and controversial theoryamong geologists at the �me, but crucial to thedevelopment of plate tectonic theory.A�er Heezen died in 1977, Tharp focused her energy oncomple�ng a comprehensive view of the world's oceans.The World Ocean Floor map was published later that yearby the Office of Naval Research and is s�ll in wide usetoday.Marie Tharp: (1920 - 2006)9R E V I S T A D E G E O C I E N C I A SIn 1978 Tharp and Heezen (posthumously) were awardedthe Hubbard Medal, the Na�onal Geographic Society’shighest honor, joining the ranks of explorers anddiscoverers like Ernest Shackleton, Louis and Mary Leakyand Jane Goodall. In 1998, Tharp was honored during the100th anniversary of the Library of Congress' Geographyand Map Division. The following year, she was recognizedby Woods Hole Oceanographic Ins�tu�on with the MarySears Woman Pioneer in Oceanography Award. In 2001she was honored by her home ins�tu�on with the Lamont-Doherty Heritage Award.Tharp died in 2006 at age 86. Her pioneering work foreverchanged the way people think about and study the richtopography of the ocean floor.Source:h�ps://web.whoi.edu/womens-comm/marie-tharp-pioneering-mapmaker-of-the-ocean-floor/How One Brilliant Woman Mapped the Secrets of the Ocean Floor | Short Film Showcase:h�ps://www.youtube.com/watch?v=vE2FK0B7gPo10R E V I S T A D E G E O C I E N C I A SKrystyna Piotrowska: (1938 - 2022)Con gran tristeza, el 26 de Enero de 2022, a la edad de 83años,falleciólaProfesoraDoctoraKrystynaPiotrowska.La doctora Krystyna Piotrowska nació el 5 de Julio de 1938en Łuck, en la actual Ucrania. Al inicio de la guerra, regresócon sus padres a Varsovia, donde completó su educaciónsecundaria y comenzó sus estudios en la Facultad deGeología de la Universidad de Varsovia.En Febrero de 1962, obtuvo el �tulo de Master enGeología (en estra�gra�a y exploración) sobre la base desu tesis �tulada: "Budowa geologiczna Wyżniej Świstówki"("Estructura geológica de la Alta Świstówka"), elaboradabajo la supervisión del profesor Edward Passendorfer. Enese mismo año, se incorporó al Departamento deTectónica y Cartogra�a Geológica de la Facultad deGeologíadelaUniversidaddeVarsoviaysusinves�gaciones se centraron en la geología de los MontesTatra y las Montañas Pieniny, con especial énfasis en elanálisis estadís�co mesoestructural y en los estudiosfotogeológicos.En 1969 defendió su tesis doctoral ( premiada por el Rectorde la Universidad de Varsovia) con el �tulo "Análisisfotointerpreta�vo cualita�vo de las estructuras disyun�vasdel basamento cristalino de los Montes Tatra Polacos". Enlos úl�mos años de su trabajo en la Facultad de Geologíaimpar�óconferenciassobretectónicaycartogra�ageológica.En1971comenzósuac�vidadcien�ficaenelDepartamento de Ciencias Geológicas (más tarde Ins�tutode Ciencias Geológicas) de la Academia de Ciencias dePolonia, donde se le ofreció par�cipar en un proyectointernacional co-ejecutado por la Academia de Ciencias dePolonia y la Academia de Ciencias de Cuba, des�nado a lapreparación del Mapa Geológico de Cuba en la escala1:250.000. En los años 1971-1975 Krystyna Piotrowskarealizó inves�gaciones y trabajos cartográfico-geológicossobre la elaboración del mapa geológico de la provincia dePinar del Río, en el occidente de Cuba y en los años1977-1982 supervisó una expedición del equipo polaco-cubano que trabajaba en el mapa de la provincia deMatanzas en el centro-occidente de la República de Cuba.En el transcurso de los trabajos de inves�gación ycartográficos relacionados con dicho proyecto, junto conlas especialistas cubanos, se realizaron mapas geológicos ytectónicos en las escalas 1:250.000 y 1:100.000; perfiles ysondeos documentados, así como el reconocimiento de lascaracterís�cas tectónicas de Cuba, la unificación de lalitoestra�gra�a de las unidades formales, la iden�ficación11R E V I S T A D E G E O C I E N C I A Sdenuevasunidadesymilesdedeterminacionesmicropaleontológicas y macrofaunís�cas. El resultado deestos trabajos fue también una tesis de KrystynaPiotrowska �tulada "Estructuras del manto del oeste deCuba a par�r del ejemplo de la Sierra de los Órganos",sobre cuya base se le concedió en 1978 el �tulo de Doctoraen Ciencias Naturales. En los años 1994-2002 KrystynaPiotrowska trabajó simultáneamente en el Ins�tuto deGeogra�a de la Academia Santacrucensis (Świętokrzyskie)como jefa del Departamento de Fundamentos de laGeología. Impar�ó clases con estudiantes de Ciencias de laTierra y fue supervisora de numerosas tesis de master ylicenciatura.El primero de Enero de 1995 Krystyna Piotrowska comenzóa trabajar en el Ins�tuto Geológico Estatal (PIB), donde enlos años 2002-2006 dirigió el Departamento de Cartogra�aGeológica. Durante sus 20 años de trabajo en el Ins�tuto,fue responsable de las hojas del Mapa Geológico Detalladode Polonia a escala 1:50.000 y del Mapa Geológico dePolonia a escala 1:200.000, y puso en marcha numerososproyectos, como los mapas geológicos y turís�cos de losparques nacionales y paisajís�cos. También es coautoradel Diccionario de Unidades Litoestra�gráficas de Poloniay fue una de las editoras cien�ficas del Mapa Geológico dePolonia a escala 1:500.000. Su papel fue fundamental en elinicio de la aplicación de métodos de geo�sica superficial yen el desarrollo de modelos geológicos en el Ins�tuto,como co-creadora del modelo espacial de la estructurageológica profunda de Polonia. En los años 2005-2015dirigió su mayor proyecto en el Ins�tuto: trabajos sobre elMapa Geológico Detallado de los Montes Tatra a escala1:10.000, reuniendo a un grupo de geólogos de Varsovia,Cracovia, Wrocław y geólogos del Servicio GeológicoEslovaco (Ins�tuto Geológico Estatal Dionýz Štúr).El proyecto a largo plazo culminó con la publicación de 25hojas, incluidas 16 hojas limítrofes entre Polonia yEslovaquia. El efecto de la cooperación extranjera fuetambién su coautoría del Mapa Geológico-Educa�vo de losMontes Tatra en la escala 1:50.000 y la producción delMapa Geológico de los Montes Tatra para el Atlas delParque Nacional. Durante estos trabajos, se ocupóespecialmentedelosjóvenesgeólogosyapoyóac�vamente sus trabajos y estudios durante todo elproceso de cartogra�a geológica.En el año 2000, Krystyna Piotrowska recibió del Presidentede la República de Polonia el �tulo de Catedrá�ca deCiencias de la Tierra. Por sus logros, fue condecorada conla Cruz Dorada al Mérito (2005), la Insignia Dorada delIns�tuto Geológico Polaco (2013) y la insignia honorífica alMérito de la Geología Polaca.Fue miembro de muchos Consejos Cien�ficos, entre ellos,el Consejo de la Facultad de Matemá�cas y CienciasNaturales de la Universidad Tecnológica de Silesia, elComité de Ciencias Geológicas de la Academia Polaca deCiencias, la Comisión para la Aceptación de EstudiosCartográficos, el Consejo Editorial del Geological Quarterly,el Comité Editorial de Trabajos del PGI y los trabajos delIns�tuto de Geogra�a de la Academia Santacrucensis(Świętokrzyska), el Consejo de Programa del Atlas de losMontes Tatra y la Comisión de Unificación y Redacción delMapa Geológico de Cuba. Como parte de su ac�vidad,asumió la evaluación de los logros cien�ficos y la emisiónde opiniones sobre las solicitudes de otorgamiento de�tulos cien�ficos, así como la revisión de estudios yar�culos cartográficos y fue tutora de dos tesis para elgrado de doctor.KrystynaPiotrowskadedicósuvidaprofesionalprincipalmente a sus amadas montañas Tatra y a lageología de Cuba Su dedicación a la tectónica de losMontes Tatra, desde su época universitaria se extendióhasta la prac�ca del alpinismo realizando escaladas deinvierno y verano en equipos femeninos. Fue una personacomprome�da y entregada casi hasta el final de sus años.Siempre alegre y amistosa, invitó a cooperar a geólogos devarios centros de inves�gación. Se unió en la búsqueda dedescubrimientosconjuntos,organizóconferenciascien�ficas y numerosas reuniones en el campo. Le gustabay apreciaba el trabajo con los jóvenes. Entrenó a geólogoscubanos que colaboraron con ella en el levantamientogeológico de Cuba. Siempre llena de energía y carisma,desarrolló en sus alumnos curiosidad por la geología ysensibilidad por la naturaleza.Durante muchos años, compar�ó su pasión y trabajoprofesional en Polonia y en Cuba con su esposo, tambiéngeólogo y cartógrafo, el Dr. Jerzy Piotrowski, finado añosantes.La misa funeral por la Doctora Krys�na Piotrowska serealizó el 4 de Febrero de 2022 a las 12:00 en la iglesia deSt. Borromeo y después se le llevó a la tumba familiar enStare Powązki.Traducción del polaco por Humberto Álvarez Sánchez,revisada por el Doctor Andrzej Pszczółkowski; sobre labase de la Nota necrológica publicada el 31 de enero de2022 por el Ins�tuto Geológico Polaco-Ins�tuto Nacionalde Inves�gación. En homenaje a la vida de la DoctoraKrys�na Piotrowska.h�ps://www.researchgate.net/scien�fic-contribu�ons/Krystyna-Piotrowska-8383886612R E V I S T A D E G E O C I E N C I A SJohn F. Dewey“The con�nents stay on top like a scum.... It stays on topand squashes and scrunches and makes a mountain belt.”Geologists must o�en think in staggeringly long �me spansandjumpmentallybetweenspa�alscales,bothmicroscopic and macroscopic. So, for fun, some do asgeologist John F. Dewey has for the past three decades:shrink �me and space to a more human scale by buildingmodel railways. Dewey has been tooling with an extensivemodel of a spiraling Swiss Alpine railway, housed now inthe garage of his California home. Here, boxcars are about6 inches long, but for Dewey the focus is on rocks, not rails.“I try to get the scenery to look like the geology ofSwitzerland,” Dewey says. “You can see the rocks �l�ng inthe correct direc�on up in the mountainside. The sceneryis very serious, actually. The geology is correct.”Dewey brings the same a�en�on to detail in his researchon orogeny, the geological process of mountain building.“The earth generates enormous complexity on a smallscale,” he says. “The great trick is taking that complexity,analyzing and synthesizing it, and then pulling it throughinto the larger scale.” This perspec�ve—seeing themountain for the rocks, one might say—is one of thehallmarks of Dewey's research. It led first to his pursuit ofa unifying theory for the forma�on of mountain belts inIreland and the eastern United States, and then to hisgroundbreaking proposals in 1968 of how plate tectonicscould explain orogeny. Like the mountain belts he studies,Dewey's career has straddled the Atlan�c Ocean, withacademic posts in both his na�ve England and the UnitedStates, where he is now Dis�nguished Professor ofGeology at the University of California, Davis (Davis, CA).In his Inaugural Ar�cle, published in this issue of PNAS,Dewey presents the results of research underway since hiselec�on to the Na�onal Academy of Sciences in 1997 (1).In the paper, he reports on processes that can buildmountains in surprisingly short geologic �mes, which heillustrates with detailed evidence from the Irish mountainsthat he has been studying since his doctoral work 47 yearsago.Simultaneous ObsessionsAs a child in London during World War II and its a�ermath,Deweyexcelledatsports—boxing,rugby,cricket,gymnas�cs, even the high jump and javelin. “I wasobsessed with sport; I really was,” he says, “but it was clearthat I was not good enough to get to the top in any ofthem. And I was really ambi�ous. I wanted to be the top ofsomething.” At age 16, he found a field in which he wouldlead. During a holiday in Devon, England, he learned a bitof field geology with his great uncle, Henry Dewey, ageologist with the Bri�sh government. Intrigued, theyoung Dewey considered studying geology for a career andfound that his plan met with the approval of hishousemaster at Bancro�'s School (Essex, U.K.), an amateurgeologist named John Hayward. Immediately, academicstudies became a “simultaneous obsession,” along withsports, Dewey says. He a�ended Queen Mary, Universityof London, where he graduated with a first-class geologydegree in 1958.PIONEROSDE LASGEOCIENCIAS13R E V I S T A D E G E O C I E N C I A SDewey was less interested in the booming fields of miningand petroleum applica�ons and more a�racted toacademic geology, where he could pursue the sort ofteaching he enjoyed during his frequent lectures toschoolchildren. He performed graduate work at ImperialCollege London, where he studied the Irish Caledonianmountain belt, which would come to dominate hissubsequent research career. A�er earning his doctorate in1960, Dewey worked as a lecturer at the University ofManchester (Manchester, U.K.) and the University ofCambridge (Cambridge, U.K.).Linking Oceans and MountainsThe late 1960s marked a turning point both in Dewey'scareer and in the field of geology. In 1967, Dewey took a 3-month sabba�cal to work at the Lamont GeologicalObservatory(nowtheLamont–DohertyEarthObservatory) in New York. He studied the Appalachian/Caledonian belt, which comprised the other tails of theIrish Caledonides that he had inves�gated during hisdoctoral work. These mountains are part of a range thatcurves from the eastern United States around toGreenland, Norway, and Scotland. With the help of a largemap and extensive research in the Columbia University(New York) library, Dewey plo�ed the geologic details ofthe en�re Caledonian system, struggling to make sense ofit.As Dewey worked toward a unifying explana�on fororogeny, the then-new theory of plate tectonics wasbeginning to take hold in the geology world. Researchershad been trying to solve the puzzle of why rocks from theocean were so young, roughly 160 million years andyounger, versus the 4-billion-year-old rocks from thecon�nents. Geologists at Lamont and elsewhere began torealize that deep ridges in the ocean were constantly beingpulled apart by strong geological forces, with new oceaniccrust forming at the axial crack and carried outward as if ona conveyer belt, a phenomenon called seafloor spreading.The old oceanic crust is then pulled back down into theearth's interior and destroyed at oceanic trenches, orsubduc�on zones.Gradually, plate tectonics began to explain the puzzlingobserva�ons from the oceans. Dewey, however, broughtthe theory ashore and used it to solve the puzzles ofmountain belts. Through his research, he began to realizethat there existed mountain belts of different rock typesthat could be explained only by the opening and closing ofan ocean between them.Plate tectonics could address this ques�on directly.Researchers saw that con�nental plates sat on top of theocean's rocks because they are much lighter and weaker.“The con�nents stay on top like a scum. So when onecon�nent collides with another, it won't go down thetubes” and into the earth, as an oceanic plate would,Dewey says. “It stays on top and squashes and scrunchesand makes a mountain belt.” While at Lamont, Deweyfilled in the details of seafloor spreading, collidingcon�nents, and mountain building. “They just exploded,”he says of the papers he wrote during this �me (2, 3).“That was what got me going in plate tectonics. I wasalways interested not just in the details of geology but howall the details put together to give you the bigger picture.”From Ink to ComputersDewey feels lucky to have been at Lamont during that�me, when renowned scien�sts such as Walter C. PitmanIII and Lynn R. Sykes were there. Plus, he says, “I got to loveAmerica. It was a cultural delight. America was open andlively and a freewheeling place. It was a free and exci�ngperiod,” he says. In 1970, Dewey took a posi�on at StateUniversity of New York at Albany. The next 12 years werefabulous, Dewey says. “I had superb students, greatcolleagues, and we just did a huge amount of research. Wechanged the way people thought about the world.” Un�lthe early 1970s, he studied old mountain belts, such as the400-million-year-old Appalachian system. A�er a while, hebegan extending his research to younger mountain beltssuch as the Alps, Himalayas, and Andes, where signs of theongoing mountain building were s�ll evident (4–6).When it comes to choosing research topics, Dewey says, “Ijust follow my nose. If something really interests me muchmore than what I'm doing now, I just drop what I'm doingnow and start on something else.” He cites as a majorinfluence in this philosophy the late Sir Edward “Teddy”Bullard, who told him not to follow the advice of “olderpeople” and instead just pursue interests without fear offailure. His thinking has always been quite visual, he says—another bonus in geology. He believes strongly in the valueof drawing maps, which necessitates going into the field,looking at the rocks in immense detail, and then being ableto see the larger picture. Early in his career, Dewey drewmaps by hand, with stencils and inks, but by the late 1990she was happily using ILLUSTRATOR (Adobe Systems, San14R E V I S T A D E G E O C I E N C I A SJose, CA) to create his maps on the computer. “I was a bitlate coming to computers, really,” he says.The American WestA�er a decade of research in New York, Dewey began tofeel “a bit foot-loose again” and entertained thoughts ofmoving elsewhere. Courted for years by various Englishuniversi�es, he finally accepted a posi�on as chair of theDepartment of Geological Sciences at the University ofDurham (Durham, U.K.) in 1982. In 1986, he moved againto serve as chair of the Department of Earth Sciences atthe University of Oxford. In his first 5–6 years at Oxford,Dewey was happy as he carried out more research on theBri�sh and Irish Caledonides (7–9). Yet dissa�sfac�on withBri�sh academic culture caught up with him before long. “Igot �red of a closed, class-based society,” he says. At theage of 63, he examined the career choices in front of him.“I thought I was going downhill a bit intellectually, and so,come 2001, I thought, `Well, do I really want to re�re andbecome an old emeritus in Oxford, or do I want to breakout again?”' he says. “So I broke out again.”“Just how long does it really take to make a mountain?”This �me, he headed to California, accep�ng a Professor ofGeology posi�on at the University of California, Davis, in2001. It was “an immense intellectual rejuvena�on,” hesays, as he began new research projects in the SierraNevada and Rocky mountain ranges (10). He skied theslopes surrounding Lake Tahoe, but, most importantly, hewas back in an academic environment that challengedhim.Quick-Cooking MountainsDewey's PNAS Inaugural Ar�cle (1), which exemplifies hisresearch style, involves massive data synthesis and springsfrom his line of detailed work on the Irish Caledonides,which he has studied for 47 years (7–11). The research “issomething I'm known for, and, although it soundsimmodest, nobody else is really capable of wri�ng thispaper,” he says. The research concerns the forma�on anddestruc�on of mountain belts and looks to resolve theques�on, “Just how long does it really take to make amountain?” The answer may be surprising, Dewey says.“People think of mountain building as a long, slowprocess,” on the order of 50 million years, he says. Yetwhen he synthesized the sedimentologic, mineral, andgeochronologic data for the Irish Caledonides, the modelpointed to something more rapid. “These mountains weremade and destroyed in 10–12 million years,” Dewey says,“and that's really a new concept.” In the case of the IrishCaledonides, this arc of islands on the edge of a denseoceanic plate began colliding with a con�nental plateroughly 475 million years ago. The islands essen�allyclimbed on top of the con�nent, Dewey says, forming themountain range. The mountain-building process was cutshort, however, because the oceanic plate was heavyenough to sink beneath the con�nent, which formed anew subduc�on zone and relieved the collision pressuresrather quickly. In contrast, two colliding con�nents have noavenue of escape, and so they build in pressure andcon�nue to form mountains over many more millions ofyears.According to Dewey, this work might spur a revisita�on ofother mountain belts. In par�cular, one aspect of orogenymay need to be reexamined: the metamorphism and heattransfer in rocks. In tradi�onal models, the rocks indeveloping mountains heat up gradually by conduc�onunder the forces of collision. “It is now quite clear that allthose models are completely wrong,” he says. “Theserocks heated up very, very quickly. The only way you can dothat is shove into them something very hot, or intrudesome very hot liquid,” like magma. “Basically, I'm nowconvinced, and I think I can prove it in many orogenic beltsof all ages around the world, that metamorphism takesplace very, very fast, and all these conduc�ve modelscannot be right.”Importance of SweatIn October of this year, Dewey plans to give a talk at theannual mee�ng of the Geological Society of America in SaltLake City on one of his favorite topics: balancing detail andthe big picture in geology and geophysics. On the onehand, geologists have always tended to “get ratherobsessed with a very small scale, the par�cular, theminu�ae, the details,” Dewey says, “and geophysiciststend to think on whole-earth scales, because they can'tdeal with the immense complexity that we have on a smallscale. So there's a terrible spa�al and temporal gap, and Ithink we need to get over that.”Dewey believes the answer lies in more intensely focusedfield work and in “ge�ng sweaty. The truth resides in therocks,” he says. “The rocks don't lie.” He says studies of therocks “are difficult, and it's hard work if you're reallyge�ng at the maps. But if you have intelligent maps aboutrocks, you can do a lot.” He points to another role model inthis respect, geologist Robert Shackleton. Abstract models15R E V I S T A D E G E O C I E N C I A S“canrunawaywithus,”Deweysays.“They'remathema�cally great fun to do, but you mustn't everbelieve them.”Although Dewey plans to re�re in the not-too-distantfuture, he says he will never be able to leave geologyen�rely. He and his wife, Molly, plan to split their �meamong California, Ireland, England, South Africa, and NewZealand. Dewey says he hopes to stay fit through his wide-ranging athle�c regimen of fishing, swimming, cycling,playing cricket and tennis, and walking the hills of hisbeloved mountains. And, even if he does not leave hishouse, he can s�ll immerse himself in Alpine geology. “Inormally do a �ny bit here and there on the model,” hesays, “but, when I re�re next year, I'll spend a tremendous6 months and just have a real bash of the railway.”Source: https://www.pnas.org/content/102/43/15283Researchgate: https://www.researchgate.net/pro�ile/John-Deweyhttps://www.geosociety.org/awards/06speeches/sgt.htmhttps://www.jsg.utexas.edu/news/2017/11/john-dewey-thoughts-from-one-of-the-founding-fathers-of-the-theory-of-plate-tectonics-on-geology-education-and-his-own-illustrious-career/GSA Award: h�ps://www.geosociety.org/awards/06speeches/sgt.htmJohn Dewey of Cambridge University discusses theorigins of the Appalachian-Caledonian mountainrange at the History of the Earth’s Crust Symposium.Credit: Barre� Gallagher16PUBLICACIONESMario E. Manrique OviedoTESIS &RESÚMENESDeterminación de la densidad poblacional dehuellasdefisiónnuclearespontáneasenobsidianasdeldomoriolí�cosanAntonio,ubicado en fundo Michihuasi en el DistritoCaylloma, provincia de Caylloma de la región deArequipa, PerúUniversidad Nacional De San Agus�n de ArequipaTesis de Maestría para optar el �tulo profesional de Ingeniero Geólogo, 2021Asesor: Dr. Fredy Fortunato García ZúñigaResumenEl presente trabajo realizado en el domo riolí�co San Antonio, a proximidades del Fundo Michihuasi, distrito y provinciade Caylloma, departamento de Arequipa, está situado a una al�tud de 4650 m.s.n.m., presenta un relieve variado, planoy ligeramente ondulado en la parte superior, las laderas al norte presentan poca pendiente y hacia el sur son abruptas ylimitan con el borde de la caldera de Caylloma.Las rocas más an�guas corresponden a sedimen�tas marinas del Grupo Yura seguida por rocas volcánicas del Terciario delos grupos Tacaza y Barroso. El Grupo Tacaza con�ene yacimientos epitermales de baja, media y alta sulfuración y el GrupoBarroso está formado por intrusiones �po domo - lava y domos de naturaleza andesí�ca a riolí�ca.Del caparazón vitreo del domo riolí�co San Antonio se extrajeron 2 muestras de roca in situ, se seleccionaroncuidadosamente 7 obsidianas que fueron preparadas y analizadas en el Laboratorio de Física de la Universidad Nacionalde San Agus�n con el fin de evidenciar huellas de fisión nuclear. Estas obsidianas fueron atacadas con ácido fluorhídricoa una concentración de 45% y a una temperatura constante de 15.1C° y a diferentes intervalos de �empo. La muestraobsidiana codificada como C-11 con un �empo de ataque de 240 segundos es la que expone mejores resultados dehuellas de Fisión nuclear siendo la densidad poblacional promedio de 8.236 ± 0.510 x 10-3 (huellas/cm2) y su diámetropromedio de 15.929 ± 0.445 μm.Figura 1.1. Localización de lazona de estudio.17R E V I S T A D E G E O C I E N C I A STomogra�a eléctrica para exploración geohidrológica en un sector del acuífero La Misión, Baja CaliforniaCentro de Inves�gación Cien�fica y de Educación Superior de Ensenada, Baja California (CICESE)Maestría en Ciencias en Ciencias de la Tierra con orientación en Geo�sica Aplicada, 2021Sustentante: Paul Alexander Calderón PulidoDirector de tesis: Dr. Marco Antonio Pérez FloresResumenEl acuífero de La Misión, ubicado en una porción Noroeste del estado de Baja California, ha cumplido con el obje�vo deproveer agua a algunas poblaciones del estado.En el transcurso del �empo se ha realizado una variedad de estudios en el acuífero, los cuales han inferido un decrementoen la disponibilidad y calidad del agua, mo�vo de este estudio. Se escogió una sección del acuífero donde existe contactocon un tributario para ejecutar un levantamiento geoeléctrico u�lizando métodos de corriente directa. Se extrajeron datosde 15 perfiles u�lizando las configuraciones tetrapolares: Wenner, Schlumberger y Dipolo-Dipolo. Como la vegetación esabundante, los perfiles no pudieron realizarse en línea recta y tomaron caminos irregulares. Se extrajeron 2 muestras deagua a lo largo de los límites del área para medir sus propiedades �sicas y relacionarlas con los resultados. Se eliminarondatos inconsistentes, se calculó el factor geométrico con precisión al usar el posicionamiento de los electrodos con GPS yse aplicó una inversión en dos y tres dimensiones para generar modelos de la zona de estudio. Se escogió el mejor modeloy se concluyó la existencia de 4 grupos: dos flujos de agua, una zona de mezcla y una zona conductora. Las mediciones deresis�vidad de las muestras de agua coinciden con los respec�vos grupos declarados del modelo final.Figura 1.1. Ubicación de La Misión, Baja California, mapa generado por ArcGIS.18R E V I S T A D E G E O C I E N C I A SProspección geo�sica con GPR y sísmica en la zona arqueológica de Peralta, Guanajuato, MéxicoIns�tuto Politécnico Nacional.Tesis para obtener el �tulo de: Ingeniero Geo�sico, 2021Sustentantes: Cruz Cruz Ildefonso Xocoyotzin y Cruz Segundo AlejandroDirector de tesis: Ing. Miguel Gómez San�ago.Resumen.En el presente trabajo se aborda el uso de técnicas de exploración geo�sica con la finalidad de ayudar al estudio de la zonaarqueológica de Peralta en estado de Guanajuato, tal es el caso del so del Radar de Penetración Terrestre RTP o GPR porsus siglas en ingles ha sido usado en estudio de si�os arqueológicos por su alta resolución, la rapidez al momento derealizar el levantamiento y la visualización en campo de los datos con un procesamiento preliminar. La sísmica derefracción la cual nos ayuda a iden�ficar las diferentes velocidades de propagación de onda en el subsuelo y así determinardiferentes capas de dentro del mismo, así como espesores de dichas capas. Y por úl�mo el uso de la técnica de AnálisisMul�canal de Ondas Superficiales (MASW) es un método que analiza las propiedades de propagación de las ondas sísmicassuperficiales (Vs) y la cual nos sirve para determinar la estabilidad del subsuelo. En este trabajo se realizó sobre laestructura principal de la zona arqueológica denominada Recinto de los Gobernantes tomando un total de 16 radargramasy 2 volúmenes de adquisición con GPR sobre dicha estructura, además de 6 perfiles de refracción alrededor de la zona deestudio y 17 puntos de atribución para la técnica de MASW. Cada uno de los datos de adquisición se procesaron enso�wares especializados. Al término del procesamiento de los datos en los resultados se encontraron posibles objetos deinterés arqueológico, así como la determinación de diferentes espesores de los materiales de construcción y laprofundidad a la cual estría el nivel del suelo original sobre el cual se comenzó la construcción de la estructura aquíestudiada. También se logró realizar una correlación entre dos de los perfiles de refracción que se tomaron dentro de laestructura de estudio con los resultados obtenidos en el volumen de adquisición tomados en el mismo si�o.Figura 1.1. Localización de la zonaarqueológica Peralta, Guanajuato. [mapagenerado en QGIS]19R E V I S T A D E G E O C I E N C I A SCaracterización litoestra�gráfica de la Caliza Chimeco y la Formación Mapacheen el Cerro Boludo, Municipio de Chila de las Flores, Puebla, México.Universidad Nacional Autónoma de MéxicoTesis para obtener el �tulo de: Ingeniero Geólogo, 2020Sustentante: Alex Altamirano CabelloDirectora de tesis: Dra. Mayumy Amparo Cabrera RamírezResumen.El área de estudio se localiza en el municipio Chila de las Flores, definido entre los límites sur de Puebla y norte de Oaxaca,es parte de la Provincia Fisiográfica Sierra Madre del Sur y de la Subprovincia Cordillera Costera del Sur. Se realizó unestudio estra�gráfico, sedimentológico y petrográfico de la Caliza Chimeco (Oxfordiano) y la Formación Mapache(KimmeridgianoTitoniano), ambas expuestas en la periferia del Cerro Boludo, el cual es parte del poblado Santa Ana LosVázquez.Los estudios estra�gráficos, sedimentológicos y petrográficos de las unidades consisten en el análisis de caracterís�caslitológicas, estudio de sus respec�vas estructuras sedimentarias, el análisis de secciones delgadas, elaboración desecciones sedimentarias (una por unidad estra�gráfica) e interpretación de ambientes de depósito de la Caliza Chimeco yla Formación Mapache. La Caliza Chimeco se encuentra cons�tuida por margas, margas con ooides, limolita calcárea y,según la clasificación de Dunham (1962) calizas grainstone con ooides y, según la clasificación de Folk (1959) oomicrita. Enconjunto, estas litologías forman un total de 63.15 m de sección sedimentaria expuesta en la periferia del Cerro Boludo.Por otra parte, la Formación Mapache está cons�tuida según la clasificación de Dunham (1962) por calizas grainstone conooides, calizas grainstone con bioclastos y calizas grainstone con pellets; calizas packstone con ooides, calizas packstonecon bioclastos y calizas packstone con pellets y, según la clasificación de Folk (1959): ooesparita, bioesparita, pelesparita,oomicrita, biomicrita y pelesparita, respec�vamente. Las calizas conforman un total de 132 m de sección sedimentariaexpuesta en la periferia del Cerro Boludo. Lo que resulta en un total de 195.15 m.Las caracterís�cas litoestra�gráficas de la Caliza Chimeco indica un ambiente de depósito de aguas someras áridas–semiáridas, es decir, un ambiente litoral, específicamente en subambientes backshore, berma, foreshore y shoreface.Respecto a la Formación Mapache, sus caracterís�cas litoestra�gráficas indican un ambiente de depósito shoreface osublitoral con comunicación intermitente al mar abierto (iniciada en la sección superior de la Caliza Chimeco) que permi�ael desarrollo de abundante fauna marina (pelecípodos, gasterópodos, equinodermos, foraminíferos y braquiópodos).Figura 1.1. Ubicación del área deestudio respecto a la división estataly municipal. Adaptado de Bibliotecadigital de mapas, Editado de GoogleEarth.20R E V I S T A D E G E O C I E N C I A SAcoplamiento del rumbo geoeléctrico a invariantes de rotación del tensorde impedancias magnetotelúrico en presencia de distorsión galvánica.Centro de Inves�gación Cien�fica y de Educación Superior de Ensenada, Baja California (CICESE).Tesis para cubrir parcialmente los requisitos necesarios para obtener el grado de Doctor en Ciencias en Ciencias de la Tierracon orientación en Geo�sica Aplicada, 2021Sustentante: Rocío Fabiola Arellano CastroDirector de tesis: Dr. Enrique Gómez Treviño.Resumen.El método magnetotelúrico se ha vuelto una técnica muy empleada en años recientes dado que permite es�marimpedancias electromagné�cas de la Tierra mediante la medición de ondas electromagné�cas que �enen origen naturaly que ocurren en un rango de frecuencias muy amplio. Sin embargo, es�maciones de las impedancias transversal eléctricay transversal magné�ca del tensor magnetotelúrico bidimensional pueden no estar acopladas con la dirección del rumbode la estructura, teniendo además implicaciones significa�vas cuando se trata con distorsiones galvánicas. Entre losmétodos u�lizados actualmente se pueden obtener impedancias sin distorsión combinando una ecuación cuadrá�ca conel tensor de fase. En la terminología de Groom-Bailey, la ecuación cuadrá�ca proporciona amplitudes y fases que soninmunes al giro o twist y el tensor de fase proporciona fases inmunes tanto al twist como a la cizalla o shear. Por otro lado,se pueden obtener los ángulos de rumbo sin distorsión u�lizando el enfoque de Bahr o el tensor de fase. En principio, estoes todo lo que se necesita para proceder a una interpretación bidimensional. Sin embargo, las impedancias resultantesignoran el rumbo porque son invariantes bajo la rotación del sistema de coordenadas y, si van a estar relacionadas con unrumbo geológico, deben estar acopladas a una dirección par�cular. Esta es una ambigüedad adicional a la ya conocida de90 grados que surge en los métodos clásicos de determinación del rumbo y que debe resolverse de forma independiente.En este trabajo u�lizamos el modelo de descomposición de Groom-Bailey para resolver la ambigüedad al traer de vueltael acoplamiento entre impedancias y rumbo en presencia de distorsiones galvánicas. Nuestro enfoque es un híbrido entrelos métodos numéricos y analí�cos existentes que reduce el problema a una decisión binaria. Presentamos tres algoritmos,dos de ellos aún requieren op�mizar el ajuste a los datos y el tercero solo necesita una comparación de fases. Los tresrealizan un seguimiento de los posibles cruces de las curvas de fase, lo que proporciona una solución clara y única. Usamosdatos sinté�cos y de campo para ilustrar el desempeño de los tres esquemas.Figura 1.1. Esquema de unlevantamiento del métodomagnetotelúrico. Se u�lizanbobinas para medir loscampos magné�cos y paresde electrodos para loseléctricos.21R E V I S T A D E G E O C I E N C I A SEstudio hidrogeoquímico del sistema de manan�ales al Poniente y Sur de la Ciudad de Mexico.Universidad Nacional Autónoma de México.Tesis para obtener el grado de Maestría en Ciencias de la Tierra (Aguas subterráneas), Febrero 2021Sustentante: Sandra Paola Blanco GaonaDirector de tesis: Dr. Oscar A. Escolero Fuentes (Ins�tuto de Geología)Resumen.La complejidad del flujo y la evolución química en los sistemas de aguas subterráneas los hacen inherentemente di�cilesde entender. Como enfoque para entender los sistemas de flujo de agua subterránea se ha propuesto dividirlos encomponentes, de manera convencional se dividen en flujos locales y regionales. Los sistemas de flujo locales son de granimportancia en la actualidad debido a la escala de �empo estos, la cual es suficientemente corta para reaccionar a lasac�vidades humanas, por ello es posible evaluar el efecto de acciones humanas en el suministro de agua y plantear unages�ón sostenible del recurso hídrico.El sistema de manan�ales estudiado se ubica al suroeste de la Cuenca de México, la mayoría de los manan�ales descargandentro del denominado Suelo de Conservación donde la importancia de este radica en los servicios ambientales quegenera a la Ciudad de México. La relevancia de la inves�gación reside en que los manan�ales hacen parte de un sistemade flujo local, el cual se caracteriza por un flujo ac�vo que refleja las variaciones anuales de las condiciones climá�casactuales, con ello la escala de �empo que afecta este �po de flujo es suficientemente corta como para reaccionar a lasac�vidades humanas.Se estableció un modelo hidrogeoquímico del funcionamiento de los manan�ales; para ello se aplicó el enfoquemul�criterio combinando diferentes técnicas hidrogeológicas; las cuales fueron monitoreo, modelación hidrogeoquímicae isotopos. Mediante el monitoreo de parámetros fisicoquímicos en 14 manan�ales durante el año 2019, se iden�ficó lainfluencia del clima en los manan�ales.Como proceso anterior a la modelación se analizaron 41 datos químicos e isotópicos de muestra de agua de manan�al,con el propósito de determinar el origen y la circulación del agua subterránea; con ello se iden�ficó una evolución naturaly una evolución por contaminación en las aguas. La evolución por composición natural se caracteriza por manan�ales pocomineralizados (conduc�vidades menores a 150 µS/cm) y la evolución atribuida por contaminación se caracteriza por elaumento de la mineralización debido al aumento de la influencia antropogénica.Posteriormente, a par�r de diagramas bivariantes de las concentraciones iónicas e isotopos estables, se logró iden�ficarlos principales procesos �sicos y químicos que controlan las caracterís�cas iónicas de los manan�ales. Los diferentesprocesos iden�ficados fueron evapoconcentración, disolución de CO2, intercambio iónico, hidrolisis y precipitación desilicatos. Para validad los procesos en los manan�ales, se aplicó modelación hidrogeoquímica, para ello se realizó unamodelación directa para concentrar los iones por evaporación y una modelación inversa para cuan�ficar todos losprocesos iden�ficados. Los procesos mas relevantes se llevan a cabo en la zona no saturada, es decir concentración porevaporación y disolución de CO2.Por úl�mo, la información isotópica se comparó con la Línea Mundial de las aguas meteóricas y con la Línea local de lasaguas meteóricas, indicando que el agua subterránea deriva de precipitaciones locales.Con base al gradiente isotópico se presenta una clasificación de los manan�ales, donde se observa una coherenciaespacial, química y geológica. En las partes altas los manan�ales sufren evaporación desde la superficie debido a lainfluencia climá�ca directa, y no presentan mayor evolución química; en contraste, en las partes bajas no se presenta unainfluencia climá�ca directa sino desfasada, además, un aumento en las concentraciones iónicas.Se concluye una influencia del clima en los manan�ales que se amor�gua a medida que el flujo de agua alcanza mayorprofundidad. Se establece un componente corto, medio y largo del sistema de flujo local.22R E V I S T A D E G E O C I E N C I A SEl primero y el segundo �enen en común manan�ales que descargan dentro del suelo de conservación y cuya evoluciónquímica se da por procesos naturales. Estos manan�ales están dentro de la formación de las Cruces y se diferencianporque el componente corto presenta una mayor influencia del clima. El componente corto se presenta sobre el derramedel Xitle, con poca presencia de suelo (litosoles), una alta influencia antropogénica y un aumento en la profundización delflujo de agua.Figura 1.1. Localización de la zona de estudio, figura parte superior Norteamérica y Centroamérica, figura delmedio se observa la Cuenca de México con sus limites y dentro de ella la Ciudad de México, en color gris elsuelo urbano y en color verde el suelo de conservación. Por úl�mo, en la parte inferior se observa la ubicaciónde los manan�ales, donde se puede observar que en su gran mayoría descargan en la Sierra de las Cruces. Seobservan 3 estaciones del programa Red de Deposito Atmosférico (REDDA), las estaciones ExconventoDesierto de los Leones (EDL), Eco guardas Ajusco (EAJ) y San Nicolas Totolapan (STN); del programa Red deMeteorología y Radiación Solar (REDMET) se ubican dos estaciones dentro del suelo urbano, Ajusco Medio(AJIM) y Pedregal (PED). Finalmente se observa la estación meteorológica en la zona del Ajusco.23R E V I S T A D E G E O C I E N C I A SDeterminación del estado de equilibrio químico fluido-roca del yacimiento geotérmicoCerro Prieto-I, Baja California, a par�r de la geotermometría de solutosCentro de Inves�gación Cien�fica y de Educación Superior de Ensenada, Baja California (CICESE)Tesis para cubrir parcialmente los requisitos necesarios para obtener el grado de Maestro en Ciencias: 2022Sustentante: Annely Reyna AvilezDirector de tesis: Dr. Efraín Gómez AriasResumenEl estado de equilibrio en un sistema geotérmico se puede determinar a par�r de la interacción fluido-roca que se lleva acabo en el yacimiento, donde la composición química de los fluidos (vapor y agua) dependerá del �empo de interaccióncon la roca y de la temperatura, lo que resulta un cambio en la composición mineralógica de la roca (alteración mineral)del yacimiento. La temperatura del yacimiento se puede es�mar a par�r del uso de geotermómetros, los cuales sonecuaciones analí�cas desarrollados a par�r de la composición química de fluidos y del estado de equilibrio químico queexiste entre la interacción fluido-roca. Cuando hay variación en las condiciones de equilibrio, la es�mación de latemperatura con la geotermometría puede ser equivocada. En el presente trabajo, se realizó un análisis termodinámico-químico del estado de equilibrio fluido-roca de la sección-I del campo geotérmico de Cerro Prieto, a par�r de informaciónde la composición química de fluidos de pozos. Se desarrolló un programa (en lenguaje fortran) para el cálculo de lasac�vidades iónicas y coeficientes de ac�vidad para los principales cons�tuyentes iónicos disueltos en las aguasgeotérmicas (Na+, K+, Mg2+ y Ca2+) y a par�r de diagramas de estabilidad mineral se observó que los fluidos geotérmicos delcampo, �enden al equilibrio entre los minerales microclina (feldespatos-K) y albita (plagioclasa-Na). El mineral quegobierna al Ca2+ en el fluido es la wairakita. Con esto se determinó que el CGCP-I se encuentra en cuasi-equilibrio a unatemperatura de 280 C. Además, se desarrollaron geotermómetros basados en ac�vidades iónicas y relaciones deconcentración para Na/K. Las temperaturas obtenidas por medio de los geotermómetros desarrollados en este trabajopresentan errores normalizados de ±10% con respecto a la temperatura del yacimiento. Por ello, al comparar sudesempeño con los geotermómetros clásicos, muestran una reducción en la incer�dumbre de las es�maciones de lastemperaturas del yacimiento. Este trabajo propone que, con el desarrollo e implementación de geotermómetros basadosen la composición química del campo geotérmico y las temperaturas de fondo de pozo (BHT), se tendría un mejor controly un monitoreo más eficaz en la es�mación de las temperaturas del yacimiento y con ello complementar estudios paraiden�ficar nuevas áreas con potencial de producción en el campo geotérmico.Figura 1. Zona de estudio. El campogeotérmico de Cerro Prieto (CGCP) estáubicado dentro del Saltón Trough (quese ex�ende desde el Mar de Saltón alsur de California, EE.UU., hasta el Golfode California, México)a 21 km alsureste de la ciudad de Mexicali. La zonageotérmica del Valle de Mexicali selocaliza dentro de la cuenca del Saltón,que abarca desde el Saltón Sea en laporción sur del estado de California(EE.UU.) hasta el Golfo de California yforma parte del sistema tectónico deSanAndrés,formadoporfallastranscurrentescondesplazamientonormales, de dirección general NO-SE.24R E V I S T A D E G E O C I E N C I A S25R E V I S T A D E G E O C I E N C I A S26R E V I S T A D E G E O C I E N C I A Sh�ps://store.aapg.org/detail.aspx?id=132027R E V I S T A D E G E O C I E N C I A SEl libro recomendadoh�p://acading.org.ve/info/publicaciones/libros/pubdocs/Libro_VARGAS_20_anos_despues.pdf28R E V I S T A D E G E O C I E N C I A SEl libro recomendadoh �p s : / / w w w . a c a d e m i a . e d u / 4 2 2 4 2 7 8 6 /LOS_DILUVIOS_EN_LAS_MONTA%C3%91AS_DE_LA_CORDILLERA_DE_LA_COSTA_EDUARDO_R%C3%96HL_Corregido_VERSION_DIGITAL29TEMASDE INTERÉSLAS MARCAS DE LA II GUERRA EN TERRITORIO DE VENEZUELA.Y LOS INICIOS DE LA GEOLOGÍA PETROLERA LOCALJosé Antonio Rodríguez ArteagaColaborador de la RevistaIntroducciónEn 1939, estalla la II Guerra Mundial y Venezuela es seriamente afectada por la marina alemana-nazi la cual inicia la guerraen aguas de la región Atlán�ca y Caribe, RAC, denominada operación Wes�ndien para impedir que el petróleo extraído yrefinado en la región llegue a los aliados que se ubican en Europa (Straka, 2016: 309).Es así que montada dicha operación naval y militar, la región frente a la costa nacional, Operación Neuland, lleva consigoun con�ngente de submarinos que son apostados a lo largo de la costa nacional, cuya misión más cercana y certera parael ataque evitaba que cualquier petrolero proveniente de occidente y/o del oriente venezolano, tomase rumbo a lasrefinerías de Aruba y Curazao para su debido procesamiento.Tal situación provoca para 1942 una sensible disminución de nuestra producción petrolera afectando al país con unaimportante merma en sus ingresos fiscales (Straka, op. cit.) para el Estado, mas por la intervención progresiva de la marinaamericana los ataques irán disminuyendo y la producción se recuperará de 400 a 700 mil barriles/día para el año ´44.Esta, reseña muestra en apretada síntesis algunas de las acciones puntuales emprendidas por la Standard Oil Company-Venezuela en sus instalaciones para contrarrestar el ataque alemán.En el diario devenir de los campos petrolerosorientales, estuvieron involucrados dos geólogos que en aquel �empo prestaron sus servicios profesionales. Luego de lacon�enda, los conocimientos y experiencia adquirida serán aprovechadas en las labores docentes dictando materias enla Escuela de Geología de la Universidad Central de Venezuela, en sus inicios.Producción petrolera in situEn Venezuela y por mucho �empo la búsqueda de petróleo se realizó de manera empírica, y a comienzos del siglo XXfueron establecidas campañas de exploración en las que par�ciparían de manera ac�va y primordial la ac�vidadexploratoria de la geología y la geo�sica con interés petrolero.Hacia los años 1870 la exploración, basada en la localización de “manaderos de petróleo“o menes, serán los responsablesdel surgimiento de un conflicto interno en el que se involucraría el Estado y las empresas privadas mo�vada a laapropiación de la renta que de ello derivaba y el surgimiento de un preámbulo histórico de conflictos por la renta del oronegro, los cuales no serán discu�dos en estas líneas (Ramírez, 2007). Pretendemos eso sí, ubicar en contexto amplio,aquellos elementos que inician la exploración y explotación de petróleo nacional confrontados a futuro con los conflictosde la II Guerra Mundial, tarea nada nada fácil ante la mul�plicidad de fechas variables que manejan diversos autores,siendo necesario un arreglo cronológico lo mejor adaptado para un ineludible trabajo a futuro.En 1875, como preámbulo, la población de San José de Cúcuta, Colombia es destruida por un terremoto que ocasiona enla localidad de Rubio, estado Táchira, Venezuela y en terrenos de La Alquitrana, nombre por demás elocuente, la aparicióngeneral de fracturas de las que manó en forma violenta un material viscoso y negro: petróleo, iniciándose así el capítulode la historia bituminosa venezolana con la presencia de Manuel Antonio Pulido Pulido, dueño de las �erras, quien30R E V I S T A D E G E O C I E N C I A Saprovecha este inesperado “reventón” para fundar una compañía con capital 100% nacional, dedicándose a su explotaciónque en 1878, cons�tuirá la primera empresa de la región andina, la Petrolia del Táchira, que extraerá y refinará “crudo”nacional hasta 1934, fecha de ex�nción de la concesión, a tan solo 5 años del iiciada la II Guerra (Ramírez Vera, op. cit.).Previo a los acontecimientos de descubrimiento y producción de ese año no solo será exclusividad de la la región andina,sino que en el occidente nacional 25 años antes serán descubiertos y puestos en producción los pozos Zumaque I y II.Al respecto, (Frechilla y Texera, 2005) señalan la Caribbean Petroleum Company como la responsable de los pozosZumaque I y II en la Costa Oriental del Lago de Maracaibo responsables del descubrimiento del gigantesco campo MeneGrande,el cual iniciaría la explotación intensiva de petróleo, con apenas un mes después de haberse iniciado la I GuerraMundial. Como se coligue, la presencia petrolera, a nivel nacional en sus primeros �empos estuvo ín�mamente ligada cona las 2 Guerras mundiales ocurridas.Previo a todo esto ello, en 1912, se emprendían los primeros estudios geológicos sistemá�cos en suelo nacional bajo lasórdenes del geólogo estadounidense Ralph Arnold, quien junto a un grupo de 52 entusiastas colegas, entre quienes seencontraba, el venezolano Pedro Ignacio Aguerrevere, dieran a conocer los hallazgos de petróleo, revelando el potencialpetrolífero que el país tenía.Arnold y su grupo trabajaron por encargo de la empresa General Asphalt,subsidiaria de la Caribbean PetroleumCompanny responsable del descubrimiento del campo de Mene Grande 2 años después. Además del informe, la aventurapetrolera y sus hallazgos, casi diarios de viaje aparecerían publicados en 1960 luego que éste explorador falleciera aprincipios de enero de 1946; acompañaban en el texto los señores Macready y Barrington, y llevaría por �tulo: First Big OilHunt:Venezuela 1911-1916. Todos los pazos conducentes al hallazgo petrolero venezolano.En 2008, una versión en español será la 2da reeditción con el �tulo: Venezuela Petrolera: Primeros pasos, 1911-1916, cuyapresentación llegando hasta allí, la publicación de esa especial aventura.En la neutralidad de la guerraEn los inicios de la II Guerra Mundial (IIGM), Venezuela mantuvo aparentes signos de neutralidad, rumor que se corríaentre el personal de Estándar Oil Venezuela por su posición contraria al conflicto bélico pero tras el ataque a Pearl Harborrompen relaciones diplomá�cas con alemanes, italianos y japoneses.Balance de daños en VenezuelaUna revisión intensiva del buques petroleros frente a costas nacionales, muestra como fueron atacados y hundidos, losbuques Pedernales (Foto) y Arkansas, amén del San Nicolás, Tía Juana, San Rafael, Oranjestad, Rafaela y Monagas (Foto).Este úl�mo de hundido por un torpedo U-502 a tan solo a 40 millas de la Península de Paraguaná, estado Falcón, occidentevenezolano y provocando el fallecimiento de algunos tripulantes connacionales de los 19 venezolanos que formaban partede la tripulación, amén de otros de nacionalidad no-venezolana que cons�tuían el personal del buque. Ello ocasionóciertamente la inquietud a nivel mundial con seria repercusiones para la Alemania- nazi.Dos atrevidos geólogos y su trayectoriaEl personal ubicado en Caripito, era netamente norteamericano, (Virgil Winkler, com. pers., 1995), exis�endo algunosgeólogos en la población de Quiriquire, pero cuyos datos permanecen silenciados. Sin filiaciones de su par�cipación en laIIGM, su presencia con�núa en dicha población con el añadido de algunos datos de asombro: Estos dos geólogos VirgilWinkler y Alfonso Kroboth, sin ningún vínculo con la Alemania nazi de entonces una vez iniciada la con�enda se dio de baja(Winkler, com. pers).De las vivencias personales del primero y sus vivencias se destacarían:1- Un alistamiento fallido en el ejército, obteniendo la clasificación 4D,”no apto para el servicio por su visión limitada y suspies planos”; 2- Su presencia como personal joven desconocedor de la región, y en calidad de sus�tuto del Dr. Harvey D.Borger Jefe del Laboratorio de Caripito. Borger, quien sin aparente mo�vo se separaría del cargo en forma “temporal”,regresando a EE.UU. y no retornar a Venezuela. De allí la enorme responsabilidad que asumió Winkler al aceptar el cargo.31R E V I S T A D E G E O C I E N C I A SFuera de la ru�na de trabajo que puedise imperar en un campo petrolero, Virgil Winkler ya integrado a la vida profesionalen los laboratorios geológicos existentes en Caripito señaló un punto poco natural en él, así, 4- los buques torpedeados enlas costas de Venezuela regularmente avistados por los hidroaviones de la empresa iniciarían una necesaria prác�ca alsobrevolar toda la costa en procura de avistar y rescatar a las víc�mas de los naufragios, muchos de ellos con ac�tudes casipsicó�cas, embebidos en crudo y flotando en el mar. Esta situación se vio con detenimiento cercano a la región orientaldel país. 5- El sobrevuelo era una tarea “mul�propósito”: Su uso regular era mo�vada a la adquisición de fotogra�as aéreaspara su empleo en la cartogra�a geológica geológicos, pero igual actuar en labores humanitarias en el rescate detripulación afectada con el ataque de la noche anterior. Winkler se mantuvo en su cargo tras la guerra hasta sufallecimiento y entre los necesarios e ineludibles cambios en la empresa. Por su parte, Kroboth, renunció a su cargo en1942 encontrando cargo docente en 1943 durante la existencia del Ins�tuto de Geología en la avenida San Mar�n deCaracas. Su paso por la empresa y la II Guerra Mundial, sigue en un completo misterio.Bibliogra�aArnold, Ralph, Macready, George y Barrington, Thomas, 1960, First Big Oil Hunt: Venezuela 1911-1916, 353 pp.Arnold, Ralph, Macready, George y Barrington, Thomas, 2008, (Pérez Marchelli, Héctor trad.), Fundación Editorial Trilobita,Caracas, 370 pp.Douglas C. Ramírez Vera, 2007, Mene en Venezuela: el surgimiento del conflicto por la renta del petróleo, preámbulohistórico a la coyuntura actual (1917 a 1936), Análisis polí�co, Bogotá, Colombia, 59: 24-45.Mar�n Frechilla, Juan José y Texera, Yolanda (comp.), 2005, Petróleo nuestro y ajeno: la ilusión de modernidad, Colección CienciasSociales. CDCH-UCV, Caracas.Rodríguez Arteaga, José Antonio, 2022, Breve Historia de la Geología de Venezuela en treinta y dos personajes. Bole�n deHistoria de las Geociencias en Venezuela. Sociedad de Historia de las Geociencias en Venezuela, Bole�n 137 (Enero2022, en prensa).Straka, Tomás, comp. 2016, La nación petrolera: Venezuela 1914-2014. Universidad Metropolitana, (Academia Nacional dela Historia eds.), Caracas, 338 pp.José Antonio Rodríguez Arteagarodriguez.arteaga@gmail.comBuque Petrolero Monagas hundido a su salida a las refinerías de Aruba, y /o Curazao para serrefinado el crudo y enviado a los Aliados en Europa,II Guerra Mundial.32R E V I S T A D E G E O C I E N C I A SHablemos sobre li�o: Parte IIICompilado porSaúl Humberto Ricardez MedinaEl li�o se ha catalogado como uno de los elementos crí�cos o casi crí�cos en varios estudios recientes debido a la rele-vancia que están adquiriendo las tecnologías verdes. Es uno de los 15 minerales comodi�es estudiados por los paísesmiembros de Inventario Internacional de Minerales Estratégicos (ISMI), quienes han concluido que podría volverse unelemento cri�co dependiendo del nivel de crecimiento de nuevos usos, como las baterías de vehículos híbridos y eléc-tricos. En la década de 1970 el Consejo Nacional de Inves�gación de Estados Unidos crea el subpanel de Li�o encargadode evaluar las reservas y recursos de li�o disponibles en el mundo, concluyo que las reservas y los recursos de todos los�pos de ocurrencia del li�o eran de 10.6 Mt. En ese momento se tenía es�mado que las salmueras del Salar de Atacamacontenían 4.3 Mt de li�o.El suministro actual de li�o proviene de depósitos de pegma�ta y salmueras en cuencas endorreicas. Siendo Australia,Brasil, China, Portugal y Zimbabwe los principales productores en pegma�tas, mientras que Argen�na, Chile, China y Es-tados Unidos los principales productores en salmueras endorreicas. Los otros �pos de depósito �enen potencial comocontribuyentes futuros. Hoy en día, estas reservas son las más importantes porque abastecen a las dos operaciones, queproducen una can�dad sustancial del carbonato de li�o del mundo.Comercio MundialEl li�o es comercializado globalmente en una gran variedad de forma como concentrados, carbonato, óxidos e hidróxidosy cloruro, bromuro y metal.Los principales países exportadores de carbonato de li�o, óxidos e hidróxidos de li�o en 2014 se muestran en la Fig. 1.Aunque hay algunas exportaciones significa�vas de óxidos e hidróxidos, sobre todo de Estados Unidos de América y decarbonato. En 2014 Chile exportó 49 387 toneladas de carbonato de li�o o casi el 60 por ciento de todo el carbonato deli�o exportado en ese año. Esto se compara con solo el 14 por ciento del segundo mayor exportador, que fue Argen�na.EE. UU. fue el mayor exportador de óxidos e hidróxidos de li�o, con 8 885 toneladas o el 35 % del total de óxidos e hidróxi-dos exportados en 2014. China fue el segundo mayor exportador de óxidos e hidróxidos con el 24 % del total.Países productoresde LitioProducción MineraReservas20202021Estados UnidosWW750,000Argentina5,9006,2002,200,000Brasil1,4201,5005,700,000Chile21,50026,0009,200,000China13,30014,0001,500,000Portugal34890060,000Zimbabwe4171,200220,000Otros países--2,700,000Total, mundial(redondeado)82,500100,00022,000,000Tabla 1: Producción minera mundial y sus reservas en toneladas métricas Modificado de: (USGS). (enero 31,2022). W de Withheld: Dato retenido para evitar la divulgación de datos de propiedad de la empresa.33R E V I S T A D E G E O C I E N C I A SFigura 1: Mayores países exportadores de carbonato de li�o y oxido e hidróxido de li�o en 2014 Tomado de:(BGS) (agosto 2, 2016).Comercio MundialLas importaciones en 2011, 2012 y 2013 superaron las 80 000 toneladas, un aumento del 26% en 2011 en comparacióncon antes de la recesión de 2008, pero las importaciones cayeron por debajo de las 80 000 toneladas en 2014. También semuestra una caída similar en 2009 para las importaciones de li�o óxidos e hidróxidos, pero, al igual que con las exportacio-nes, no es tan dramá�co. Las importaciones en 2014 han aumentado a más de 20,000 toneladas, Las tendencias recientesen las importaciones de carbonato de li�o se muestran en la Figura 2.En la Figura 3 se muestran los principales países importadores de carbonato de li�o y óxidos e hidróxidos de li�o en 2014.Tal vez sea sorprendente que China no haya sido el mayor importador de carbonato de li�o en 2013; más bien, la Repúblicade Corea fue la más grande con importaciones de 14 981 toneladas o el 20 por ciento del carbonato de li�o total en eseaño. Sin embargo, China fue el segundo más grande (13 601 toneladas o 18 por ciento), seguido por Japón (12 108 tone-ladas o 16 por ciento) y EE. UU. (9 626 toneladas o 13 por ciento del total). El mayor importador de óxidos e hidróxidos deli�o en 2014 fue Japón, que importó 6 667 toneladas o el 28 % del total importado a nivel mundial, seguido de Bélgica (2388 toneladas o el 10 %) y la República de Corea (1 965 toneladas u ocho por ciento del total). Europa depende de las34R E V I S T A D E G E O C I E N C I A SFigura 2: Tendencias globales de importación de carbonato de li�o y oxido e hidróxido de li�oTomado de: (BGS) (agosto 2, 2016).Figura 3: Mayores países importadores de carbonato de li�o y oxido e hidróxido de li�o en 2014Tomado de: (BGS) (agosto 2, 2016).35R E V I S T A D E G E O C I E N C I A Simportaciones tanto de carbonato de li�o como de óxidos e hidróxidos de li�o y recibió la mayoría de sus importacionesen 2014 de Chile. De las 2 812 toneladas de óxidos e hidróxidos de li�o importados por los 28 miembros de la Unión Euro-pea (de países fuera de la UE), el 37 % provino de Chile, mientras que otro 33 % provino de Rusia. De las 10 664 toneladasde carbonato de li�o importadas por la UE-28 (desde fuera de la UE), el 84 % provino de Chile.UsosEl li�o �ene un amplio uso en el mercado, pero sus tres formas básicas son en mena o concentrado, metal y compuestosquímicos manufacturados.Las menas y concentrados son usados en la industria del vidrio, cerámica y porcelana. La Petalita, Lepidolita y la Amblygo-nita puede ser usada sin requerir algún �po de beneficio. El li�o es ú�l porque crea condiciones favorables para la nu-cleación interna. En comparación con el sodio, favorece la cristalización externa. Debido a su pequeño radio iónico y altaintensidad de campo, el li�o adiciona una alta resistencia mecánica y al choque térmico, así como una buena resistenciaquímica. En estas aplicaciones, el li�o también se introduce en forma de carbonato de li�o.En su forma de metal el li�o es elemento solido más ligero teniendo un peso atómico de 6.94 y una densidad de 0.53 g/cm3a 20°C. El li�o metálico se u�liza en la síntesis de bu�l li�o. En la metalurgia no ferrosa, la alta reac�vidad del li�o con losgases es u�lizada para capturar oxígeno y azufre, convir�éndolos en compuestos estables. El li�o también se usa enaleaciones de li�o aluminio y li�o magnesio, donde adiciona resistencia a altas temperaturas, mejora la elas�cidad y au-menta la resistencia a la tracción.A pesar del alza en la demanda de aleaciones de aluminio y li�o en la industria aeronáu�ca comercial y militar ésta no hacumplido con las previsiones de crecimiento que se tenían al 2006.El carbonato de li�o, además de su consumo por parte de la industria cerámica, se está u�lizando cada vez más en lasceldas de reducción de aluminio, donde mejora la conduc�vidad del baño fundido al reducir la temperatura de operacióny da como resultado una mayor producción. También es muy eficaz para reducir las emisiones de flúor al retener el flúorcomo fluoruro de li�o en el baño fundido. El consumo de carbonato de li�o ha disminuido constantemente desde su puntomáximo a fines de la década de 1970 y, aunque la tasa de disminución se ha desacelerado en los úl�mos años, aún se prevéque el consumo en el sector del aluminio disminuya aún más en los próximos años. En forma purificada, el carbonato deli�o se usa en el tratamiento quimioterapéu�co del trastorno bipolar.El hidróxido de li�o se empleó por primera vez como ingrediente en las baterías alcalinas de almacenamiento donde supresencia aumentó la vida ú�l de las celdas. En la década de 1940, Clarence Earl desarrolló una grasa especial al hacerreaccionar hidróxido de li�o con ácidos grasos, descubrió una nueva aplicación. Se encontró que la grasa resultante reteníasu viscosidad en un amplio rango de temperatura y permanecía estable en presencia de agua.Las grasas a base de li�o se han conver�do en un estándar producto en los campos de lubricación militar, industrial y au-tomotriz. El hidróxido de li�o anhidro puede absorber grandes can�dades de dióxido de carbono, una propiedad que seha u�lizado en la regeneración del aire sistema de mando Apolo y módulos lunares, y en nuevas aplicaciones en la industriaminera para equipos de respiración ar�ficial.Las salmueras de cloruro de li�o y bromuro de li�o �enen bajas presiones de vapor y se u�lizan en sistemas de refri-geración por absorción. El cloruro de li�o también cons�tuye el material de alimentación para la producción de li�o me-tálico.El fluoruro de li�o se u�liza principalmente como fundente en esmaltes, vidrios y vidriados, y en soldadura y soldadurafuerte.Además de ser el metal más liviano, el li�o también es el metal más electronega�vo y, por lo tanto, es ideal para usar enmuchas aplicaciones de baterías. Las caracterís�cas de las baterías de li�o incluyen alta densidad de energía, alto voltajede funcionamiento, amplio rango de temperatura de funcionamiento y larga vida ú�l. Las aplicaciones actuales incluyenmarcapasos cardíacos, hardware militar, cámaras, respaldo de memoria de computadora, relojes y equipos de mediciónen la industria de extracción de petróleo. El crecimiento futuro en esta área puede incluir el uso de baterías de li�o recarga-bles para herramientas eléctricas portá�les y equipos electrónicos y de comunicaciones e incluso como fuente de alimen-36R E V I S T A D E G E O C I E N C I A SUsos de Litio en EUA 2021Baterías 74%Ceramica y Vidrio 14%Grasas y lubricantes 3%Material de fundición 2%Polimeros 2%Tratamiento de aire 1%Otros 4%Figura 4: Usos de Li�o en Estados Unidos de América en 2021 Modificado de: (USGS). (enero 31, 2022).tación para vehículos eléctricos y el corazón ar�ficial total. La aplicación de mayor valor para las baterías de li�o siguesiendo los productos de consumo portá�les secundarios, especialmente las computadoras portá�les (con un contenido de30-40g de Li) y los teléfonos celulares (con un contenido de 2-3g de Li).Bibliogra�aAlatorre, A. (2018) “Minerales y Rocas Industriales”, Ins�tuto Politécnico Nacional, Escuela Superior de Ingeniería y Arqui-tectura ESIA Ticoman Unidad Ciencias de la Tierra. Ciudad de México. Octubre 15, 2018 [apuntes de clase].Bradley, Dwight, and McCauley, Andrew, (2013). A preliminary deposit model for lithium-cesium-tantalum (LCT) peg-ma�tes Noviembre 3, 2021, de U.S. Geological Survey Open-File Report 2013–1008, 7 p., h�ps://doi.org/10.3133/ofr20131008.Bri�sh Geological Survey (BGS). (agosto 2, 2016). Mineral profile – Lithium. Noviembre 3, 2021, de Bri�sh Geological Sur-vey Si�o web: h�ps://www.bgs.ac.uk/news/mineral-profile-lithium/Kogel, J., Trivedi, N., Barker, J., & Krukowski, S. (2006). Industrial Minerals & Rocks, Commodi�es, Markets and Uses. 7thEdi�on Book, . Estados Unidos de America: Society for Mining, Metallurgy and Explora�on, (SME).Serbia Energy & Mining. (marzo 19, 2021). ICT metal mining case study at jadarite mine in Serbia. noviembre 6, 2021, deSerbia Energy & Mining Si�o web: h�ps://serbia-energy.eu/ict-metal-mining-case-study-at-jadarite-mine-in-serbia/United States Geological Survey (USGS). (diciembre 19, 2017). Cri�cal mineral resources of the United States—Economicand environmental geology and prospects for future supply. noviembre 6, 2021, de United States Geological Survey Si�oweb: h�ps://pubs.er.usgs.gov/publica�on/pp180237R E V I S T A D E G E O C I E N C I A S60 AÑOS DE LA ENSEÑANZA DE LAINGENIERIA GEÓLOGICA EN CUBA.Dr. Rafael Guardado Lacaba,Colaborador de la Revistarafaelguardado2008@gmail.comResumenEnero de 1962, el Gobierno Revolucionario de Cuba,realiza la Reforma Universitaria, la cual genera la creaciónde nuevas carreras entre ellas la de Ingeniería Geológica eIngeniería de Minas en la Universidad de Oriente yLicenciatura en Geología en la Universidad de la Habana, yposteriormente Ingeniería Geo�sica en la CUJAE.La carrera de Ingeniería Geológica es promotora deinves�gaciones en la geología del país, con per�nencia,impacto y consecuencia tecnológica en función de losintereses del desarrollo socioeconómico del territorio, laregión, y la nación.La aplicación de los planes de estudios c, C prima, d y e,fortaleció la formación de un profesional de perfil ampliocon una mayor preparación en los conocimientos yhabilidades de geología en relación con los anterioresplanes de estudio.Se hace énfasis en los conocimientos de la ingenieríageológica, el desarrollo de la capacidad de trabajo enequipo, el dominio de las técnicas de dirección, lainforma�zación, el trabajo en red, entendimiento de losprocesos polí�cos, la capacidad de comunicación, en lossistemas de ingeniería, y la exigencia de un aprendizajecon�nuo. Es una carrera acreditada de excelencia basadaen los elementos de evaluación de: per�nencia e impactosocial, profesores, estudiantes, infraestructura, currículo.SummaryJanuary 1962, the Revolu�onary Government of Cubacarries out the University Reform, which generates thecrea�on of new careers, including Geological Engineeringand Mining Engineering at the Universidad de Oriente anda Bachelor of Geology at the University of Havana, andlater Geophysical Engineering at CUJAE.The Geological Engineering career promotes research inthecountry'sgeology,withrelevance,impactandtechnological consequence based on the interests of thesocioeconomic development of the territory, the region,and the na�on.The applica�on of the study plans c, C prima, d and e,strengthened the forma�on of a professional with a broadprofile with greater prepara�on in the knowledge and skillsof geology in rela�on to the previous study plans.Emphasisisplacedonknowledgeofgeologicalengineering, the development of teamwork skills, masteryof management techniques, computeriza�on, networking,understanding of poli�cal processes, communica�on skills,intheengineeringsystems,andthedemandforcon�nuous learning. It is an accredited career based on theevalua�on elements of: relevance and social impact,teachers, students, infrastructure, curriculum1. Introducción. La carrera de Ingeniería Geológicaescenario de referencia.En enero de 1962, el Gobierno Revolucionario lleva a cabola Reforma Universitaria, importante momento históricoque permite determinar las tendencias en el desarrollo dela Educación Superior en Cuba. La Reforma Universitariageneró la creación de nuevas carreras entre ellas la deLicenciatura en Geología en la Universidad de la Habana eIngenieríaGeológicaeIngenieríadeMinasenlaUniversidad de Oriente. En 1964 ante la necesidad deprofesionales de Ciencias de la Tierra y considerando elcreciente desarrollo minero metalúrgico de la regiónoriental se unifican en la Universidad de Oriente las dosescuelas de geología con un perfil de ingeniería. Al mismo�empo se crea en el ISPJAE la carrera de IngenieríaGeo�sica.En esta etapa fue decisiva el apoyo del Ins�tuto de Minasde Leningrado, de la Academia de Minas de Freiberg, delIns�tuto de Minas de Eslovaquia y Praga, y otrasuniversidades,cuyacolaboracióncondujoalaespecialización, en Ingeniería Geológica e Hidrogeología,Prospección de Yacimientos de Minerales Sólidos, yProspección de Yacimientos de Petróleo y Gas.En estos años surgen y se desarrollan los cursosuniversitarios para trabajadores, en 1965 se inicia el PlanExtramuros de la Universidad de Oriente, en Moa y Nicarose crean las primeras unidades docentes del país.Par�endo del principio de que no hay desarrollo de laeducación superior sin inves�gación, desde la segundamitad de la década del 60 se trabaja para lograr unaadecuadainterrelacióndocencia-inves�gación–producción, emplear del modo más eficiente posiblepotencialcien�ficodeprofesoresyestudiantesuniversitarios, con el doble obje�vo de elevar el nivel y lacalidad y de la propia docencia universitaria y de contribuirdirectamente a mejorar las condiciones económicas ysociales del país.38R E V I S T A D E G E O C I E N C I A SEn 1973, se inicia la formación de ingenieros geólogos enla Filial Universitaria de Minas de Matahambre, con elmismo plan de estudio que el de la Escuela de IngenieríaGeología de la Universidad de Oriente, las filialesuniversitarias de Nicaro y de Moa como un apoyo a laindustria en la formación de profesionales. Ante laimportancia estratégica del desarrollo universitario.Julio de 1976, surge el Ministerio de Educación Superior yelIns�tutoSuperiorMineroMetalúrgicodeMoa,Universidad que unió a toda la enseñanza de la geología, laminería, la metalúrgica y las demás ingenierías y cienciasafines, que respondían al ver�ginoso desarrollo que enestos años en el nordeste de Holguín. La construcción de lanuevaplantadeníquelErnestoCheGuevara,modernización de Nicaro, y un nuevo Programa deDesarrollo Integral Minero Metalúrgico.En 1981 se inicia el desarrollo de los planes de estudio B deIngeniería Geológica. Este plan con mejor estructuramantenía algunas deficiencias tales como:• Existencia de un ciclo básico que no respondía en todoslos contenidos a las necesidades de la carrera.• Predominio de la enseñanza teórica en las disciplinas yasignaturas de la carrera.• No se precisan las tareas que debe ser capaz de enfrentarel geólogo para dar solución a los problemas profesionalesmás generales del sistema laboral de la profesión.• Incomprensión de los obje�vos como categoría rectoradel proceso docente en cada uno de los niveles en que éstese desarrolla.• Poco fondo de �empo para la ejecución de las prác�casdocentes y de producción.• Los programas analí�cos no garan�zaban siempre laconsecu�vidad de los conocimientos.En el inicio de la década del 80 la dirección del paísteniendo en cuenta la tendencia de integración de lasciencias propone la formación de un profesional de amplioespectro, que proponen la creación de un egresado deperfil amplio con una sólida formación en ciencias básicas.En 1985, a solicitud del MINBAS, fundamentado en lasnecesidades de incrementar la can�dad de graduados engeología y geo�sica en occidente en par�cular para laexploración del petróleo, el MES aprueba la reapertura deingenieros geólogos en la Facultad de Tecnología de laUniversidad de Pinar del Río y la CUJAE.La nueva visión del Comandante en Jefe de la aplicacióndel plan de estudio C en la década del 90, fortaleció laformación de un profesional de perfil amplio y con unamayor preparación en los conocimientos y habilidades degeología en relación a los anteriores planes de estudio. Elplan de estudio C se desarrolló en el Ins�tuto SuperiorMinero Metalúrgico de Moa y en la Universidad de Pinardel Río.Desde el curso 1992-1993, se desarrolla un conjunto deacciones para garan�zar que la ciencia y la técnicacons�tuyan en obje�vo central del trabajo de la EducaciónSuperior, para obtener resultados de importancia y deincidencia económica y social en los plazos más brevesposibles, y que posibilite conver�r a las universidades encentros de inves�gación cien�fica.ActualmentelacarreradeIngenieríaGeológicaespromotora de inves�gaciones de impacto, para resolverproblemas de la geología del país, con per�nencia, yconsecuencia tecnológica en función de los intereses deldesarrollo socioeconómico del territorio, la región y elpaís, todo lo cual se manifiesta en el sa�sfactoriocumplimiento de sus elevados compromisos en los planesde resultados y su generalización, en los recursos que vanapareciendo para sostener la inves�gación, en los registrosy patentes de nuevos productos y tecnologías.El modelo cubano de la carrera de Ingeniería Geológica�ene como estrategia clave la flexibilidad organiza�va, lacooperación nacional e internacional. La aplicación delconcepto de perfil amplio en las carreras, la educación depostgrado, la educación con�nua, adquiere par�cularrelevancia, pues es en este cuarto nivel donde alcanzan laespecialidad los profesionales, lográndose además unaestrechavinculacióndelpostgradoconeltrabajocien�fico-técnico. En el aspecto organiza�vo, se hancreado grupos de trabajo cien�fico docentes, por año ydisciplinas, grupos mul�disciplinarios en función deobje�vos complejos y concretos lo que ha permi�do unanueva concepción de desarrollar la formación en valoresen carrera elevando la calidad del graduado de geología.En los úl�mos cursos se producen avances en la labordocente-educa�va en la carrera de Ingeniería Geológica enCuba, como consecuencia de diversos factores, entre loscuales podemos iden�ficar como de primera importancialos siguientes:•La consolidación gradual del enfoque integral para lalabor educa�va y polí�co ideológica.•La madurez pedagógica y cien�fica del claustro,sustentada en el crecimiento del número de doctores,en un sistema integral de superación de los39R E V I S T A D E G E O C I E N C I A Sprofesores que se encuentra en vías degeneralización, y en su incues�onable pres�gio antelos estudiantes.•El fortalecimiento de las alianzas con el ServicioGeológico y las empresas geológicas-mineras del país,la industria cubana del níquel, del petróleo, elMinisterio de la Construcción, centros deinves�gación y los órganos de gobiernos territoriales,provinciales y el país. El desarrollo de la basematerial, en especial de computadoras. Elperfeccionamiento del sistema de aseguramientobibliográfico.•Las unidades docentes en Empresas Geológicas y losterritorio y el país.Ello se refleja, entre otros aspectos, en la sa�sfacción delos estudiantes con la calidad del proceso de formación,reflejadas en las encuestas realizadas durante los tresúl�moscursos,laaprobacióndelasinspeccionesgenerales prac�cada a la carrera y la aprobación comocarrera acreditada por la Agencia de Acreditación por elMinisterio de Educación Superior MES.2. NUEVAS TENDENCIAS EN LA ENSEÑANZA DE LAINGENIERÍA GEOLÓGICAEn los úl�mos años, se ha observado un creciente interéspor la Ingeniería Geológica, la reduccion de riesgosgeodinámicos geoclimá�cosl y ambiental, provenientetanto del medio geológico y su explotación como delimpacto del hombre sobre el ambiente, de ahí que lacaracterizaciónyevaluacióndelambienteseasignifica�vamente diferente según los casos por lo quedebemos de esclarecer las relaciones precedentes de estesaber geológico.La Ingeniería Geológica se define como la aplicación de lasciencias geológicas al diseño, la minería, la extracción delpetróleo y gas, y la construcción de estructuras deingeniería. El ingeniero geólogo es un ambientalistaformado para reconocer y entender el significado de lascondiciones geológicas y su influencia en los diseños deingeniería. Los proyectos que requieren un adecuadoconocimiento de Ingeniería Geológica cubren un amplioespectro, abarcando desde los residuos peligrosos, lacimentación de presas, los túneles, las canteras, y laexplotación de los recursos minerales. En la actualidad yobligadas por la adaptación a los cambios sociales yeconómicos, las universidades donde se enseña IngenieríaGeológica,hanreaccionado,introduciendonuevoscontenidos, como la cartogra�a geológica digital, lossensores remotos, sistemas de Información geográfica, elambiente, los análisis de riesgos geológicos y antrópicos, lainforma�zación y los modelos digital del terreno, elpatrimoniogeológicominero,losnuevosmétodosgeo�sicos, e incluso otras disciplinas en el campo de laenergía o la industria mineral.Elmayorbeneficioobtenidoestribaenelperfeccionamiento y la revisión y modernización de losplanes de estudio de la carrera de Ingeniería Geológicageneral e integral, la cual ha adquirido mucha relevancia,ofreciendo al alumno una formación más versá�l ymejores posibilidades para la ampliación de su perfilprofesional; tales como ingeniería del petróleo, ingenieríacivil, ambiental, el patrimonio geológico minero, elordenamiento territorial, el análisis de riesgos , así comode la propia Ingeniería de Minas.3. LA CARRERA DE INGENIERIA GEOLÓGICA Y EL SIGLOXXI.En estos 60 años los cambios presentados en la Geologiason extraordinario. La convergencia en el marco de lasdisciplinas(geodinámica,geoquímica,geo�sica,Yacimientos minerales geo informá�ca etc.) ha generadonuevos saberes geológicos en el estudio e interpretaciónde los diversos y complejos procesos y fenómenosgeológicos. Pero además el avance de las nuevastecnologías y la informá�ca han surgido propiciandoavances en el saber cien�fico de estas disciplinas.Los desa�os y oportunidades en el siglo XXI incluyen unatecnología inteligente de ordenadores que exigen delusuario ser más crea�vo, cambios constantes en elambiente de laboral que obligan al trabajo en equipo;numerosos grupos sociales que exigen como principiosbásicos de diseño los relacionados con el medio ambiente,la salud y la seguridad, todo lo cual exigirá ingenieros conalta capacitación intelectual; además de la educaciónenfocada hacia la ciencia. El énfasis se deberá poner sobrela integración de funciones, trabajo en equipo, redes,tareas mul�disciplinares y reciclaje durante la vidaprofesional.Las actuales especializaciones de las ingenierías y lasbarreras que, ar�ficialmente, crean los departamentos yfacultades en las universidades se conver�rán en obsoletasy, la Ingeniería Geológica deberá estar encajada en unamatriz de tareas obje�vas complejas. La carrera deberáconsiderarse como un proceso docente - educa�vo –cien�fico - cultural donde el producto final es el graduado,el ingeniero geólogo.Elgraduadodebetenerdentrodesuformaciónprofesional una mejor comprensión de su perfil valora�voingeniero geológico el cual contempla las siguientesac�vidades:•Sen�do de lo que le rodea: sensible a las diferencias40R E V I S T A D E G E O C I E N C I A Sculturales, aspectos medioambientales y principiosé�cos, así como alerta ante las oportunidades delmercado.•Con fundamentos sólidos: educado a través de losfundamentos teórico y prác�cos de los avances de laciencia y la tecnología; teniendo una perspec�vahistórica de los avances de la ciencia que puedentener impacto en la ingeniería y preparado para uncon�nuo aprendizaje a lo largo de su vida profesional.•Mentalidad técnica abierta: comprendiendo que losproblemas reales de la vida son mul�disciplinarios einterdisciplinarios, y previendo soluciones en uncontexto de varias alterna�vas y probabilidades.Trabajando con varias disciplinas y entrenado en lamodelización de sistemas.Eficazentrabajoengrupo:cooperandoenunaorganización de trabajos individuales dirigidos hacia unobje�vo crea�vo común, eficaz en la comunicación oral yescrita; decidido a buscar y u�lizar consejos de expertos,conocedor del valor del �empo y comprendiendo lasdiversas facetas de las operaciones de negocio: ges�ón,marke�ng, financiación y costes, leyes, recursos humanos,serviciosy,especialmente,lacalidad.¾Versá�l:solucionador de problemas, capaz de tomar decisiones;innovador en el desarrollo de productos y servicios.Orientado al cliente: buscador de la sa�sfacción de losclientes, asegurando la ra�o coste / eficacia en un mercadoglobal.Las nuevas generaciones de ingenieros están orientados alos futuros desa�os: la destreza en la comunicación, oral ysobre todo escrita, así como el arte de transmi�r lainformaciónasudes�natario,elconocimientointerdisciplinario;esmuyrecomendablequelaexperiencia geotécnica y geoquímica, geo�sica, seancombinadasconelconocimientodelaIngenieríaGeológica y con el análisis con modelos numéricosresueltos por ordenador.La�tulacióndeingenierogeólogosa�sfaceadecuadamente estos requerimientos. La habilidad demarcar claramente los obje�vos, a corto y largo plazo, es lomás importante para uno mismo y para el proyecto, asícomo buscar las oportunidades para alcanzar dichosobje�vos. El entusiasmo por el trabajo propio y por susretos conlleva el éxito de nuestro desarrollo profesional. Elconocimiento de la historia de nuestra especialidad, susimplicaciones sociales, tradiciones, así como de losprincipales casos prác�cos reales, nos ayudan a disfrutarde nuestro trabajo y a realizar contribuciones mássignifica�vas y de mayor contenido.4. LA BÚSQUEDA DE LA CALIDAD DEL GRADUADO ENINGENIERÍA GEOLÓGICA.En un contexto dinámico y compe��vo, la calidad de laenseñanza universitaria se ha conver�do en uno de lospilares para alcanzar el éxito, un producto de calidad: elingeniero geólogo que hoy formamos, es mucho másexigente que en �empos pasados. Pensamos que en elmomento que estamos viviendo es crucial saber primerodónde estamos y hacia dónde nos movemos, para juzgarmejor que debemos hacer. Acciones asociadas con lacalidad total en el proceso docente educa�vo involucran:• Dis�nguir proyectos potenciales de desarrollosfuturos graduados integrales de Ingeniería Geológica.• poner atención estricta a los procesos de enseñanzaaprendizaje en la formación de valores• priorizar y enfocar la atención sobre problemascien�fico técnicos.• El trabajo en red, en grupos de trabajo, equipos•Serequieredeldominiodehiloconductortecnológico que es el proceso de informa�zación ycomunicación, pero no confundirse, este es solo un“instrumento”• Priorizar la organización, las técnicas de dirección, eldominiodelidioma,enfocarseenelsistemacorpora�vo.• Desarrollar una cultura integral de alta calificaciónprofesional Modelo tradicional - mejorar la calidad dela educación para sa�sfacer la demanda de mercado.Nueva visión - mejorar la calidad de la educación paracrear nueva oferta de mercado (innovación, ciencia ytecnología).4.1. La acreditación de la carrera de Ingeniería Geológicaun nuevo paso a la calidad del profesional en la Geologíade Cuba. La educación superior en América La�na enfrentacambiosimportantesen,latransformaciónyreorganización de los sistemas de educación superior,especialmente el surgimiento de nuevas universidades,primordialmente privadas, las formas y modalidades derelación entre los gobiernos y las ins�tuciones; y lavaloración de la educación universitaria por la sociedad,grupos empresariales y gobiernos.Históricamente las universidades nunca fueron evaluadas,excepto, interna y externamente, para cues�onar supoli�zación (o pasiva enajenación), y sus modalidades degobierno autónomo y de democra�zación. Hasta hacepocos años no exis�an sistemas de información y procesosde evaluación de la calidad de los servicios educa�vos (ode la inves�gación). Y aún hoy día la cultura y los procesos41R E V I S T A D E G E O C I E N C I A Sde evaluación son bastante incipientes. La acreditación porsu parte, es el acto por el cual el Estado adopta y hacepúblico el reconocimiento que realiza a través del ConsejoNacional de Acreditación del MES, para comprobar losresultados que una ins�tución educa�va presenta sobre lacalidad de sus programas académicos, su organización yfuncionamiento, y el cumplimiento de su función social. Laacreditación de las carreras universitarias es un acto por elcual el Estado adopta y hace público el reconocimientoque las comisiones de acreditación hacen de la evaluaciónexterna que se efectúa a una ins�tución o carrerauniversitariasobrelacalidaddesusprogramasacadémicos, su organización y funcionamiento y elcumplimiento de su función social, dicho en otras palabrases el tes�monio que da el estado cubano sobre la calidadde un programa (Acreditación de Programas) o ins�tución(Acreditación Ins�tucional)), con base en un procesoprevio de auto evaluación en el cual intervienen lains�tución, las comunidades académicas y el CNA. Losprogramas académicos (carreras) universitarios deberíanadoptar tres procesos básicos:1.Proceso con�nuo de "prospec�va y planificaciónestratégica": se debe establecer un mecanismo paraanalizar y ges�onar cada programa académico, cadacarrera como una unidad auto contenida, como unproducto.1.Proceso de garan�a de calidad académica: Requierede una amplia par�cipación, canales efec�vos decomunicación, la determinación y aceptación deresponsabilidad formal, la sistema�zación deinformación (medición e indicadores de desempeño),y un compromiso ins�tucional para la capacitación yel desarrollo del personal.1.Proceso de evaluación de programas académicos: sedebe evaluar si los profesionales egresados soncompetentes, si pueden insertarse produc�vamenteen el mercado laboral, si están equipados pararesponder a las condicionantes sociales y del medioambiente, si los contenidos teóricos y prác�cos sonrelevantes, si las técnicas de aprendizaje son las másapropiadas, si las unidades académicas son eficientes,si los sistemas de información y los textos u�lizadosson actualizados y adecuados, si existe unavinculación con las empresas productoras de bienes yservicios, si el nivel y par�cipación estudian�l sonidóneos, etc.El establecimiento de la acreditación responde a unanecesidaddelascomunidadeseduca�vasdelasins�tuciones universitarias, la cual es la del mejoramientode la calidad del servicio prestado. El modelo de calidad dela carrera de Ingeniería Geológica sirve de punto dereferencia para que las ins�tuciones nacionales y/ointernacionales puedan hacer un diagnós�co del grado decalidad de su ges�ón, mediante la auto evaluación.Mediante esta auto evaluación se iden�fican los puntosfuertes y las áreas de mejora y se reconocen las carenciasmás importantes de forma que puedan sugerirse planes demejora.La carrera de Ingeniería Geológica da un gran peso altrabajo de auto evaluación; y la define la como la prác�capermanente dirigida a mejorar la calidad del plan deestudio de la carrera de Ingeniería en Geología. Por esto,respondiendo a las polí�cas de perfeccionamiento, haasumido el liderazgo del proceso y viene propiciando lapar�cipación y el compromiso de la comunidad académicapara garan�zar el cumplimiento de los obje�vos trazadosen la formación en valores y la calidad total universitaria.En Cuba los estándares correspondientes a la formacióndel profesional estarán dado en las siguientes variables decalidad: per�nencia e impacto social, el claustro deprofesores, los estudiantes, y la infraestructura, eldesarrollo del currículo.CONCLUSIONESLa enseñanza de Ingeniería Geológica afronta nuevos retosen este inicio de siglo XXI. La gran preocupación paraenfrentar con éxitos los profundos cambios en losprocesos económicos y ambientales que lleva consigo lacompe��vidad( nacional e internacional ) está enlograr actuaciones cada vez con enfoques de calidad totalque permita graduados de geología mas integrales ,eficientes y eficaces.En la actualidad se adoptan nuevas metodologías deenseñanza aprendizaje, nuevos me�dos y procedimientosdidác�cos que requieren de formación, capacitación yentrenamiento de los recursos humanos. Se requiera deun cambio radical de ac�tud de los actores económicos(gobierno,empresarios,geólogos,geo�sicos,geoquímicas,geotécnicos,trabajadores,etc.)haciarelacionesmancomunadasponiendoénfasisenelmejoramiento de la imagen nacional de la educación de lageología integral, cer�ficada por sus productos: elgraduado en ingeniería geológica, compe�r de formacoopera�va. En este sen�do, la educación en la esfera dela Geología de manera con�nua es el instrumentoadecuado, formación del profesional en Ingeniería enGeología, superación postgraduada de los profesionales,formación académica de master y doctores.Elprocesodeuniversalizacióndelaenseñanzauniversitaria en Cuba, representa un nuevo modeloeduca�vo que contempla que aprender, mas que un42R E V I S T A D E G E O C I E N C I A Sobje�vo, es un proceso permanente que permite preparary capacitar, en todos los niveles y sectores, recursoshumanos de la geología capaces de crear, desarrollar yemplear,cadadíamás,nuevastecnologíasyprocedimientos administra�vos modernos.La internacionalización de las ac�vidades de la ingenieríageológica que antes se inscribían solo a la esfera nacionalo local se presenta ahora en red, empresa – red,universidad – red. Esto se ha conver�do en un modo deactuación en la que debemos insertarnos, ejemplo de ello(CyTED,ALFA,UNESCO,CEPAL,etc.)quesehatransformado en el “ sujeto que hace cultura “, en elsen�do de que empieza ar�cular el sistema de valores enlos mas distantes rincones del planeta.Los nuevos métodos de dirección por valore que hoy sedesarrollan en la Carrera de Ingeniería Geológica, deestructura de comunicación interna, la organización dereaccionarantelosdesa�osyoportunidades(planificación, ges�ón, estrategias, alianzas, métodos deescenarios, etc.) están basados en el manejo sistémico deinformación que requiere de personas capacitadas connuevas concepciones de empresa, de organización, deltrabajo en equipo y del trabajo integrado.Se requiere de un nuevo hilo conductor geo tecnológicoque es el proceso de informa�zación en la Geología, quenodebemosconfundirporqueesteessolounaherramienta promovedora una tendencia al cambio en lages�ón universitaria de los departamentos, facultades,ins�tuciones, empresas geológicas, etc.El desafió que se abre por delante en la educación de laIngeniería Geológica en Cuba consiste en integrar lacapacidad de trabajar en equipos, de mo�var, de mirar deconjunto a la carrera en sus múl�ples dimensionessociales, a su alianza con las empresas geológicas,hidrogeológicas, geotécnicas, petroleras, etc., a su vez dereconocer las propias ap�tudes con hones�dad, sinarrogancia, de conocer el personal con se trabaja, dees�mular a todos de manera real, de mo�var inicia�vas ycapacidad de emprender, de prestar atención y dar laimportancia a la capacitación y superación del personal, deresponderalossubordinados,decrearambientepar�cipa�vo y de tener en fin, una visión global y unhorizonte estratégico del desarrollo sostenibleREFERENCIAS BIBLIOGRAFICA.E. Mar�nez y M. Letelier (eds.) (1997), Evaluación yacreditación universitaria - Metodologías y experiencias(Nueva Sociedad/UNESCO/O.U.I./USACH, Caracas, pp.9-44).CUBA/MINISTERIO DE EDUCACION SUPERIOR (1990),Reglamento de inspección de la educación superior (MES,Cuba, 55 p.).CUBA / MINISTERIO DE EDUCACION SUPERIOR, 2003,SISTEMA DE EVALUACION Y ACREDITACION DE LASCARRERA UNIVERSITARIAS. Presentación, Patrón decalidad de las carreras universitarias (SEA – CU 01 ), GUIAPARA LA EVALUACION DE LA CARRERA.(SEA –CU 02), Reglamento para la acreditación yevaluación de carreras universitarias (SEA – CU 03)Resolución No 116/ 2002 del MES. La Habana Cuba.Plan de Estudios E del ingeniero geólogo. 2018 ComisiónNacional de la carrera de ingeniería Geologica Moa MES.RafaelGuardadoesgraduadoenlaUniversidaddeOrienteen1970comoIngenieroGeólogo.Cursóestudiosdeespecialización en la Universidad Minera de StPetersburgoenRusia,an�guaU.R.S.S.,1972-1974.DefendióeldoctoradoenGeología en 1983. Es Académico Titular de laAcademia de Ciencias de Cuba, ProfesorTitular,ProfesorConsulyanteyProfesorEmerito de la Universidad De Moa. OrdenCarlos J. Finlay. Ha publicado más de 70 ar�culos,y es Tutor de tesis de Doctorado y maestrías. Harecibido múl�ples premios y dis�nciones, y es unprofesor reconocido en Cuba y el exprenjero en laIngeniería Geológica, la Reducción de los RiesgosGeológicosyelenfrentamientoalCambioClimá�co.rafaelguardado2008@gmail.com43R E V I S T A D E G E O C I E N C I A SREGISTROS GEOFÍSICOS DE POZOLuis Ángel Valencia FloresEditor de la revistaUn registro de pozo es una representación digital o analógica de una propiedad �sica que se mide contra la profundidad.La obtención de éste se realiza a par�r de una sonda que va recorriendo la trayectoria del pozo y de un equipo superficialque traduce la información enviada por la sonda a través de un cable que se registra en una cinta magné�ca o películafotográfica.BREVE HISTORIA DE LOS REGISTROS GEOFI�SICOS.Los primeros experimentos de prospección del subsuelo se llevaron a cabo en 1912 por Conrad Schlumberger, yconsis�eron en enviar una corriente eléctrica introduciéndola al subsuelo entre dos varillas metálicas, y en dibujar sobreun plano las líneas de potencial constante observadas en la superficie. Las formas de éstas indicaron la naturaleza yconfiguración geométrica de los dis�ntos cuerpos atravesados por el campo eléctrico.El primer registro eléctrico se realizó en el año de 1927, en el pequeño campo petrolero de Pechelbronn, Alsacia, provinciadel noreste de Francia. Este registro, una gráfica única de la resis�vidad eléctrica de las formaciones rocosas atravesadaspor el pozo, se realizó por el método de "estaciones". El instrumento de medición de fondo (llamado sonda), se detenía enintervalos periódicos en el agujero, se hacían mediciones, y la resis�vidad calculada se trazaba manualmente en unagráfica. Este procedimiento se repe�́a de estación en estación hasta que se grabara todo el registro.En 1912 Conrad Schlumberger, u�lizando un equipo muy básico, grabó el primer mapa decurvas de potencial en su finca cerca de Caen, Normandía44R E V I S T A D E G E O C I E N C I A SEl registro de resis�vidad eléctrica se introdujo comercialmente en Venezuela, en el año de 1929, Estados Unidos y Rusiay, un poco más tarde, en las Indias Orientales Holandesas. Rápidamente se reconoció en la industria petrolera la u�lidadde la medición de la resis�vidad para propósitos de correlación y para la iden�ficación de las capas potenciales portadorasde hidrocarburo.Primer registro eléctrico de AméricaPara el año de 1931, la medición del potencial espontáneo (SP) se incluyó con la curva de resis�vidad en el registroeléctrico. En ese mismo año, los hermanos Schlumberger, Marcel y Conrad, perfeccionaron un método de registrocon�nuo y se desarrolló el primer trazador gráfico.En 1932, los hermanos Schlumberger introdujeron al mercado un nuevo �po de arreglo, con el cual mejoraron la calidadde la curva, conocida como normal, debido a que detallaba bien los límites de las capas, pero disminuía el radio deinves�gación. Asimismo, en 1934 desarrollaron otros disposi�vos, como la curva normal larga, con los cuales trataron detener un mayor radio de inves�gación. La conjunción de estos tres disposi�vos junto con el registro de potencial naturalintegra lo que hoy se conoce como registro eléctrico (convencional).Los primeros registros en pozos perforados en México fueron tomados a par�r de 1936 por las Cías. Royal Dutch Shell y laBri�sh Petroleum, usando equipos patentados por Schlumberger, operados manualmente. Se registraban tres curvas, unade potencial natural y dos de resis�vidad, una denominada normal con espaciamiento corto de 0.4 m. y otra inversa con45R E V I S T A D E G E O C I E N C I A Sespaciamiento mayor a 5.70 m. que permi�an diferenciar mejor las formaciones, su contenido era conocer el grado deinvasión del lodo de perforación. Los geólogos de la Cía. Aguila (SHELL) entrenaban a los ingenieros mexicanos.En 1939 se comenzó a trabajar con las propiedades radiac�vas de las rocas. En esa época el conocimiento de estapropiedad era muy pobre y se u�lizaba el registro de rayos gamma para determinar cualita�vamente la litología(arcillosidad de las rocas) y para la correlación geológica, la ventaja de este instrumento fue que se pudo tomar en agujeroademado y llegó a ser de gran u�lidad en pozos donde nunca se había corrido un registro.Para el año de 1941 se observó que los registros eléctricos eran inadecuados para encontrar zonas porosas en calizasmasivas y se introdujo el registro de neutrones, el cual basa su principio en emi�r neutrones que interactúan con elhidrógeno de la formación, relacionándolo directamente con la porosidad.En 1948, se u�lizó corriente alterna para registrar pozos que se corrían con lodos base aceite, a esto se le denominóregistro de inducción. Como el problema no implicaba invasión por filtrado de lodo acuoso, la profundidad de inves�gaciónde esta herramienta era reducida. Desarrollos más modernos de tal registro han permi�do también su uso en lodos conagua dulce. La profundidad de inves�gación del registro de inducción se ha incrementado gradualmente con los años, a finde minimizar el efecto de invasión y de capas laterales.Un problema que se presentaba fue la falta de información en la vecindad del pozo, pero con la introducción del registromicrolog en 1949, se pudo obtener el factor de formación de las rocas en el lugar de la medición, por medio de electrodoscon un espaciamiento muy pequeño, los cuales tendrían contacto con las paredes del pozo a través de un pa�n. Con esteinstrumento se pudo conocer también el diámetro del agujero y determinar zonas en las que exis�era enjarre. En esemismo año se desarrolló el registro laterolog, que trata de medir la resis�vidad verdadera de la formación en lodos muysalados o conduc�vos. El principio se basa en enfocar la corriente por medio de un gran número de electrodos, con elpropósito de tener mayor resolución ver�cal en capas delgadas. Por estas fechas también se comenzaron a medir laspropiedades acús�cas de las rocas, cuan�ficando el �empo de tránsito de estas mismas y relacionándolo con su porosidad.En la actualidad es una herramienta indispensable en la evaluación de formaciones.En 1952 se desarrollaron los registros micro enfocados con el propósito de medir la resis�vidad de la zona lavada o lainfluencia de la resis�vidad del filtrado de lodo.Otra herramienta que infiere la porosidad se desarrolló en 1953, con el nombre de registro de densidad, dicha herramientapudo medir la densidad de la roca en el pozo, ayudando al geo�sico a determinar tanto las variaciones de ésta con laprofundidad en la prospección gravimétrica como la porosidad de la misma. La herramienta ha evolucionado en la formade realizar las lecturas y úl�mamente se u�liza el registro litodensidad, el cual da ya una idea de la litología.En 1958 apareció́ el registro de proximidad para determinar directamente la resis�vidad de la zona lavada. En nuestros díasse cuenta con herramientas que miden, de una forma casi directa, los valores de resis�vidad verdadera de la formación. Elregistro de doble inducción (1963), que es una composición de varios �pos de sondas del arreglo de inducciónconvencional, mide a dos radios de inves�gación, lo mismo que el registro doble laterolog, ocupándose en formacionescompactas.En 1970 dio comienzo la era del procesamiento de los registros. El uso de las computadoras hizo posible analizar con mayordetalle la información brindada por los registros. Existen centros de procesos localizados en lugares estratégicos a dondellega la información por teléfono, radio o satélite. En las unidades superficiales se han acondicionado computadoras paradar una interpretación rápida en el lugar de la obtención del registro. Al mismo �empo, las herramientas de registro secombinan para obtener la mayor información posible en una sola corrida.Material bibliográfico compilado por Luis Valencia, obtenido de:•Evolución de los registros de resis�vidad y su aplicación en la es�mación de la saturación de fluidos (agua ehidrocarburos)", Carlos Yair Sánchez Corona, Tesis de licenciatura, UNAM.•Schlumberger Web page.•Viro Cosultoría, manual de interpretación de registros geo�sicos de pozo.46R E V I S T A D E G E O C I E N C I A SBase del Glaciar de Jamapa, Pico de Orizaba. Uno de los úl�mos glaciares en alta montaña de México.Perforando Nucleos de Hielo en el Glaciar de Jamapa. Fotogra�as provistas porel Dr. Alejandro Carrillo Chávez.Travesía en el Glaciar de Ayoloco, Iztaccihuatl.47R E V I S T A D E G E O C I E N C I A SAglomerado de bloques y bombas cementadas por tobas basál�cas en contacto con tobas ácidasde la Formación La Yeguada (Provincia de Veraguas. Foto de H. Álvarez, 2012). Los elementos en lafoto son de orden métrico. Foto de Humberto Álvarez Sánchez, Colaborador de la Revista.Brechas de Ignimbritas de la Formación La Yeguada (Provincia de Veraguas. Foto de H. Álvarez, 2012).La Formación "La Yeguada" se distribuye ampliamente en las cumbres de la Cordillera Central de Panamá. Setrata de una unidad potente piroclás�ca cons�tuida por ignimbrí�cas y brechas explosivas con grandesbloques, estra�ficadas en bancos de grandes espesores de composición ácida, dací�ca a riolita y edades de 13a 9 Ma. La máxima ac�vidad de esta fase volcánica marca la fase final del ciclo miocénico. Foto de HumbertoÁlvarez Sánchez, Colaborador de la Revista.48R E V I S T A D E G E O C I E N C I A SA nosotros losestudiantes degeología nos gustamucho realizar lasprác�cas decampo, porquetenemos laoportunidad detomar muchasfotogra�as deestructurasgeológicas,montañas y deafloramientos.Eres estudiante de geología y �enes fotogra�asde afloramientos de tu área de estudioo de viajes de campo?Comunícate conSaúl Humberto Ricardez Medinaricardezmedinasaulhumberto@gmail.comquien está a cargo de organizar esta información.49NOTASGEOLÓGICASThe Greatest Oil Well in HistoryJon BlickwedeTeyra GeoConsul�ng LLCHouston, Texaswww.teyrageo.comIt was a big, early 20th century gusher… but it wasn’t thefamous Spindletop well drilled in 1901 near Beaumont,Texas, or even any of the many super-prolific oil wells ofthe Middle East.The Greatest Oil Well in History is a story mostly forgo�en,as it lay hidden for some 70 years in the archives of my firstemployer Amoco Produc�on Company. During the 1990’s,when I worked on a series of joint technical projectsinvolving Amoco, Pemex and the Ins�tuto Mexicano delPetróleo, my friend and former Amoco colleague JoshRosenfeld found a dog-eared copy of a 1922 Pan AmericanPetroleum & Transport Company (predecessor to Amoco)publica�on en�tled Mexican Petroleum in a dusty cornerof the company library in Houston. He decided to make anice li�le reprint of a por�on of it to distribute at theAmoco booth at a number of Mexican industry events. Thesubject of Josh’s excerpt from the book was Cerro Azul #4,drilled in the Tampico-Misantla Basin of Mexico during late1915/early 1916 by Pan American. By the way, thePresident of Pan American Petroleum & Transport waslegendary California oilman Edward L. Doheny, whose lifestory inspired Upton Sinclair’s 1920’s novel Oil! and muchlater the 2007 Oscar-winning movie There Will Be Blood.Valve completely over the well, only nine daysa�er the blowout.The well at its maximum flow before being closedin. At an es�mated 598 �, the gusher would haveexceeded the height of the Marathon oil tower inHouston.50R E V I S T A D E G E O C I E N C I A SCerro Azul #4 was drilled with a cable-tool rig into thekars�fied Albian-Cenomanian rudist reef complex of the ElAbra Forma�on, in the northwest part of an ellip�cal beltof oil discoveries that came to be known as the Faja de Oro(or, as it’s inaccurately expressed in English, the “GoldenLane”). The well blew out on 10 February 1916 at a depthof around 1,700 feet, and wasn’t brought under controlun�l nine days later. During that nine-day period, thegusher steadily grew to around 700 feet in height, andbecause of high winds caused by a cold front uncommonthis far south in the Gulf, much of the oil was blown up totwo miles from the wellsite. Apart from the drilling toolsand steel cable being shot far from the wellbore by thegusher, it was reported that stalac�tes and stalagmitesassociated with the cavernous porosity of the El Abrareservoir were ejected at high velocity along with the oil.Because of the great quan�ty of oil flowing uncontrolledfrom the well, an a�empt was made to collect as much ofit as possible by digging a series of trenches away fromthe wellsite leading into surface pits. The volume of oilcould then be calculated by measuring its flow ratethrough the trenches along with the width and depth ofthe flow. By this means, Cerro Azul #4 was es�mated tohave reached a flow rate of 260,858 barrels per day (nottaking into account the oil blown away from the gusher inthe wind) on the day before it started to be broughtunder control. This (minimum) es�mate far outstrips themaximum recorded flow rate of any other individual well,anywhere in the world.It should be emphasized that, although Doheny led thecompany that drilled Cerro Azul #4, the great discoverywould not have been made were it not for the informedop�mism, persistence, and belief in the oil poten�al ofthe Tampico region by Pan American’s Chief GeologistEzequiel Ordoñez, now regarded as one of the giants ofthe Mexican petroleum industry.There are many other fascina�ng details of the Cerro Azul#4 story, such as the fact that it took only nine days forthe wellsite crew to bring the blowout under control,despite having to forge on-site some of the componentsof the valve assembly. Perhaps most amazing of all is thatCerro Azul #4 is s�ll producing today, or at least as of theThe men who closed in well Cerro Azul #4.51R E V I S T A D E G E O C I E N C I A SJon Blickwede egresó de la Universidad de Tu�s en Boston,Massachuse�s con un Bachillerato en Ciencias de la Tierra en 1977.Durante su programa del Bachillerato, fue introducido a la geología deMéxico cuando trabajó un verano como asistente a los geólogos ytopógrafos en las minas de plata en Real de Catorce, S.L.P.Luego trabajó para la United States Geological Survey en Denver,Colorado en el Departamento de Recursos de Uranio y Torio.Entró a la Universidad de New Orleans, Louisiana en el año 1979, dondehizo su tésis de Maestría en Geología sobre la Formación Nazas en laSierra de San Julián, Zacatecas, México.Jon comenzó su carrera en la industria petrolera en 1981, trabajandosiempre como geólogo de exploración, para varias compañias tal comoAmoco, Unocal, y úl�mamente la petrolera estatal noruega Statoil, dela cual se jubiló en el año 2017. Ha realizado proyectos de geologíaregional de Mexico, Centroamerica y el Caribe para todas estasempresas.Durante 2018, Jon fundó la empresa Teyra GeoGonsul�ng LLC(www.teyrageo.com), basado en Houston, Texas.Ahora Jon está realizando unos proyectos geológicos enfocados en laregión del Golfo de México, incluyendo un proyecto de crearafloramientos digitales y excursiones geológicas virtuales en EEUU yMéxico, u�lizando imágenes ortofotográficos de drones, integradoscon otros datos geoespaciales como imágenes de satélite, mapasgeológicos, etc.PanAmericanChiefGeologistEzequiel Ordoñéz.Cerro Azul #4 well in the early 2000’s.early 2000’s when the last accompanying photo was taken.The en�re story of the birth of this world champion can befound in the aforemen�oned 1922 publica�on MexicanPetroleum, available on the web at Google Books.So some ques�ons naturally arise: was Cerro Azul #4 theglobal extreme outlier, never to be repeated? Or couldthere be any analogues out there s�ll wai�ng to bediscovered, perhaps by you and your team? Without thegusher, of course…52R E V I S T A D E G E O C I E N C I A SMÉTODOS NO CONVENCIONALES DE PROSPECCIÓN DEPETRÓLEO Y GAS: KARSTS HIPOGÉNICOS DE CUBAOCCIDENTALLeslie F. Molerio-LeónInversiones Gamma, S.A.PO Box 6249, Habana 6, CP 10600, La Habana, CubaE-mail: especialistaprincipal@gmail.comRESUMENLa presencia de karsts hipogénicos formados por lamigración de gases desde yacimientos gasopetrolíferosprofundos pueden cons�tuir evidencias indirectas dereservas de petróleo y gas. Los llamados “oil-field karsts”(karst de campos de petróleo) descubiertos y definidos afines de los 90 en Estados Unidos cons�tuyen un métodocomplementario no convencional de exploración cuyoalcance se describe en este ar�culo.INTRODUCCIÓNEl cavernamiento hipogénico es un �po de desarrollo deconductossubterráneosenelkarst;estoes,deespeleogénesis, que no está vinculado con sistemas dedrenaje superficial o subterráneo. Durante años no fueconsiderado importante debido a que no producíacambios notables en el paisaje cársico, a diferencia delsingené�co o epigené�co, que son los karsts clásicos; sinembargo, en los úl�mos 40 años va adquiriendo unasignificación cada vez mayor.Ello ha sido debido a que muchos de los grandes sistemascavernarios del mundo están asociados a este �po deprocesos (Carlsbad, Lechuguilla en Nuevo México, EstadosUnidos), a su importancia como ambientes extremos parala vida (Cueva de Vila Luz, Tabasco, México)y a larelevanciaquehatomadosuasociaciónconlaprospección de ciertos minerales ú�les y, en par�cular, enla exploración de petróleo y gas, para la que se haacuñado, incluso, el término de “karst de campos depetróleo” (oil-field karst) propuesto por C.A. Hill en 1987,a raíz de los descubrimientos de la asociación entre laporosidad secundaria debida a la migración de H2S en elcampo Yates, en Texas y la reevaluación de los sistemas decavernas del Campo Dollarhide, en Nuevo México, ambosen Estados Unidos y asociados a yacimientos de petróleo ygas.Este ar�culo pasa revista a los estudios de este �po queestamos llevando a cabo en los karsts de Cuba Occidental,a raíz del descubrimiento de ligeras evidencias dehipogenismo asociado a migración de gases (Molerio yGrau, 2011; Molerio, 2013a). La modelación para laprospección comenzó a elaborarse a principios de ladécada del 2000 (Molerio, 2003, 2004) sobre la base deestudios en cuevas de la Sierra de los Órganos. En términosdemejoramientodesuconocimientoyges�ón,recientementeseesbozaronalgunasideasdeaprovechamientosostenibleyproteccióndeestosrecursos (Molerio, 2019).RECONOCIMIENTOSEn las fases iniciales de la inves�gación se buscóasesoramiento para evaluar la componente biogénica deestas cuevas, y la asistencia de mis compañeras MaríaOfelia Orozco y María Rodríguez, fue fundamental.Información complementaria e intercambios con JulioHernández, Dave Tavares, Miguel Gala, Silvia Valladares,Arthur Palmer y Andrés Ros fueron fundamentales paramejorar el conocimiento del tema. Trabajos de campo ymuestreos específicos para esta inves�gación en varias deCuba fueron compar�dos con nuestros compañerosEsteban Grau, Vladimir Otero, Brigi�e Pileta, Oriol Chávez,Antonio González, Marian Alonso, Mario Guerra y con Ana,mi compañera.KARSTS HIPOGÉNICOS Y LOS PROCESOS DE FORMACIÓNLa Tabla 1 resume los procesos de formación de losdis�ntos �pos de karsts. En ar�cular, la hipogénesis esesencialmente un conjunto de procesos de mezcla deaguas cuya fuente de acidez no guarda relación conprocesos en superficie. Las caracterís�cas fundamentalesde los �pos de conductos formados por los procesoshipogené�cos se resumen en la Tabla 2. Se supone querepresentan entre el 10 y 15% de todas las cuevasconocidas y forman cavernas debido a los siguientesmecanismos (Palmer, 2012; Molerio, 2021):•Por mezcla de aguas en las costas marí�mas (incluyelas cuevas de margen costero, “Flank Margin Caves”de Mylroie y Carew, 1990, 1995; las de “Efecto Salino”de Montoriol y Assens, 1957 y las Indirectas Abrasivasde Molerio, 2013b)•Cuevas termales•Mezclas de aguas profundas con aguas superficiales•Disolución por aguas con H2S y cuevas formadas porácido sulfúrico53R E V I S T A D E G E O C I E N C I A S•Oxidación de minerales de hierro•Paragénesis•Incepción o inserciónLa erosión inversa de Maucci (1952) es esencialmente unproceso de mezcla de aguas que �ene lugar en la zona nosaturada, entre el epikarst y la zona de fluctuación del nivelfreá�co. Ácidos de origen profundo asociados a procesosígneos y metamórficos �enen potencial para el desarrollode cavernamiento, pero ello no está explicado aun demanera sa�sfactoria. El flujo convec�vo Curl (1966) es otroproceso espeleogené�co en el que el patrón de desarrolloha sido inducido por el potencial de densidad. Otros �posgené�cos, como las cuevas asociadas a inundaciones,tampoco serán documentados aquí.TipoZona hidrodinámicaTipo deacuíferoControleshidráulicosSistemas �ísico-químicosControles�ísico-químicosProcesos dominantesen el control de la tasade disoluciónSingenéticosNo saturada o vadosaLibreIntercambio conprocesos subaéreos(efecto de mezcla)H2CO3-CaCO3CinéticosMezcladeaguas(fundamentalmenteefectosalinoydeinsaturación por mezclaagua dulce/agua saladaH2CO3-CaCO3CinéticosEfectos combinados deagua en movimiento ymezcla de aguas.H2S-H2SO4Balance demasasEfecto de mezclaEpigenéticosEpikarstZona no saturada ovadosaZonaSaturadaoFreáticaLibreIntercambio conprocesos subaéreos(agua en movimientoy mezcla de aguas).Lacapacidaddedisolución se atenúaenladireccióndel�lujoSistemas mixtos:H2CO3-CaCO3 yH2S-H2SO4Controlesmixtos: cinéticoybalancedemasasEfectos combinados deagua en movimiento ymezcla de aguasHipogenéticosCirculación profundaCon�inado, semi-con�inado osemilibreSin intercambio conprocesos subaéreos.LacapacidaddedisoluciónesIndependiente del�lujoH2S-H2SO4Balance demasasIrrupciones deagresividad limitadas entiempo y espacio:Efecto de mezclaOxidación del H2SEnfriamiento de aguastermales ascendentesMetamor�ismoReducción de sulfatosMaduración dehidrocarburosDedolomitizaciónTabla 1. Procesos de carsificación y espeleogénesis (Molerio et al., 2004)Fig. 1. Área de estudios actuales54R E V I S T A D E G E O C I E N C I A SProcesosPatrones genéticos y características morfológicasMezcla de aguas en las costasmarítimas(“FlankMarginCaves”–FMC-,de“EfectoSalino”–ES-e IndirectasAbrasivas –IA-Se encuentran típicamente en las zonas costeras carbonatadas jóvenes con altaporosidad primariaPoca o ninguna relación de las cuevas con el relieve super�icialDominantemente horizontales y, las más antiguas se presentan en nivelessuperpuestos, asociados a las posiciones del nivel del mar �luctuante duranteelCuaternarioyalineadashorizontalmenteenzonasqueestuvieronoriginalmente al nivel del marLos planos de desarrollo correlacionan bien con los niveles del mar durante elpasado geológico recienteEl patrón de desarrollo dominante es laberíntico, vinculado con la porosidadde la roca y, en menor grado, el agrietamientoPlanta laberíntica, con salones lobulares, unidos por coalescencia o por pasajesestrechos asociados a clasti�icaciónEn la super�icie se presenta campos de lapiés libres, de crestas agudasLas cuevas en la línea de costa apenas presentan espeleotemasLos derrumbes, de tipo graviclásticos predominan y dominan los sedimentosno cohesivos del tipo de arenas y cantos redondeadosPresentan fuertes huellas de abrasión mas no de scallops u otras formaserosivas asociadas al agua corrienteLos per�iles son francamente horizontalesEl desarrollo actual es más activo en la zona de intercambio entre el mar y elagua subterránea de menor salinidadPor lo común de poca extensiónCuevas termalesPatrones de desarrollo vertical o subvertical laberínticoLa rami�icación aumenta hacia arriba en el corteLa fuente de CO2 activa es profunda y suele estar asociada a actividad ígnea o adescomposición metamór�ica de las rocas carbonatadasSe asocian con paleokarstsPatrón de desarrollo controlado por el agrietamiento con fuerte interconexiónentre las fracturasLas cortezas parietales de calcita son las espeleotemas más comunes, perotambién se encuentran depósitos de cuarzo y otros silicatos; en este caso, lasgalerías suelen ser verticales y planasLos sedimentos son ricos en elementos poco comunes en otras cuevas (Ba, Tl,Pb, As) y abundan los óxidos de hierro y manganesoLas galerías inferiores suelen estar rellenas de sedimentos provenientes de ladeposición por aguas ascendentesEn el área suelen encontrarse manantiales termalesMezclas de aguas profundascon aguas super�icialesSe trata de pequeños bolsones que suelen presentarse aisladosPer�iles irregulares del techo y el pisoAbundantes formas erosivas del tipo de cúpulasOcurren en calizas muy fracturadasAbundan depósitos de óxidos de hierro y manganesoLos sedimentos arcillosos son tanto de super�icie como profundos, en estecaso, provenientes de las etapas iniciales de disolución de los carbonatosDisolución por aguas con H2SSuelen estar asociadas a los yacimientos gasopetrolíferosLas cuevas activas expiden fuerte olor a sul�hídrico en ciertos sitios y en lasaguasLos sedimentos son ricos en sulfatos, particularmente de yeso y minerales dehierroCuevas formadas por ácidosulfúricoSe forman por la mezcla del sulfuro de hidrógeno ascendente con aguassomeras donde se oxidó en o cerca del nivel de las aguas subterráneasPatrones de desarrollo rami�icados, de planta irregular, laberínticos ydispuestos en varios niveles superpuestosLas galerías pueden ser tanto horizontales como verticales y no exhiben unpatrón único ya que también pueden encontrarse grietas que se estrechanhacia arribaGuardan poca o ninguna relación con el relieve super�icialSecciones transversales irregulares o galerías sin continuidadBolsones y cúpulas se encuentran frecuentemente en la bóveda, que sueletener forma de cúpulaBajo pH de las aguasLa mayor parte del volumen de las cuevas se encuentra sobre el nivel freáticoLas galerías que descargan hacia la super�icie generalmente disminuyen elárea de la sección transversal en la dirección del �lujoTabla 2. Caracterís�cas generales de los conductos formados por cavernamiento hipogénico (compiladode Montoriol y Assens, 1957; Núñez, 1957; Curl, 1966; Lowe, 1992; Lowe y Gunn, 1997: Palmer, 2012;Molerio, 2013b, 2021; Ros et aL., 2016).55R E V I S T A D E G E O C I E N C I A SOxidación de minerales dehierroPatrón rami�icado en forma de redNo suelen disponerse en niveles superpuestosLa correlación con el relieve super�icial es muy baja o inexistenteDominan depósitos de arena dolomíticaLas formas cenitales y parietales de disolución incluyen cúpulas y bolsonescomo elementos dominantesSuele presentar depósitos masivos de yesoSe encuentran asociadas muchas veces a depósitos de sideritaParagénesisSección estrecha, encañonada, más amplia en las partes altas, dominantementevertical, más alta que anchaAbundante relleno por sedimentos arcillosos con nichos y rampas por encimadel relleno de sedimentosLas formas erosivas parietales de tipo de meandros pueden mostrar scallopsde crestas suavizadas y las ranuras de los meandros se inclinan hacia abajo, endirección del �lujoSuelen ser indicadores de lazos freáticos descendentesSon indicadoras de inundaciones, a las que se asocia su desarrolloIncepción o inserciónCavidades primarias, primigenias o protocuevas que no forman redes, sinohorizontes aislados asociados a los planos de estrati�icaciónPatrones horizontales de desarrollo y sección irregularDesarrollo próximo a zonas capaces de generar reacciones de tipo REDOXLos procesos no siempre están asociados al �lujo gravitacional sino también alas mareas, capilaridad o difusión iónicaConductos o pozos sin continuidad aparentePROSPECCIÓN EN LOS KARSTS DE CAMPOS DE PETRÓLEOLa mayor parte de las cuevas formadas por ácido sulfúricoresultan del ascenso de sulfuro de hidrógeno desdehorizontes profundos que se oxidó cerca del nivel de lasaguas subterráneas (Fig. 2: véase también a Jagnow et al,2000). En su absoluta mayoría están asociadas a campospetroleros, de ahí que se haya acuñado el término dekarsts de campos de petróleo (oil-field karsts; Hill, 1987,1995; Palmer y Hill, 2005) para este �po de cuevas (Fig. 2;Lowe, 1999). En los úl�mos años, una profusa literatura hadescrito los resultados de las inves�gaciones en este �pode karsts y, los procesos ac�vos detectados en algunascuevas, como la de Villa Luz (Hose y Pisarowicz, 1999),entre otras, han confirmado que no se trata de procesosinac�vos, sino que �enen lugar con especial intensidad enestos momentos.Fig. 2. Procesos de espeleogénesis hipogénica debidas a migración de H2S desde fuentes profundas(Según Palmer y Hill, 2005; Palmer, 2012)56R E V I S T A D E G E O C I E N C I A SEl problema directo de la inves�gación se resume acon�nuación:•Existen procesos de espeleogénesis asociados acondiciones artesianas en el territorio cubano•Los paleokarsts son, en todos los casos, preneógenos, aunque se han encontrado evidencias depaleokarsts o, al menos de cavernas holofósiles enrocas del Mioceno medio•La porosidad de cavernamiento y los sistemas decavernas confinados o formados bajo condicionesartesianas no guardan relación con el relievemoderno superficial, lo que indica la falta deintercambio con procesos subaéreos•Los patrones de cavernamiento de patrónagrietamiento en determinados contextoshidrológicos y geológicos muestran pocas variacionesen su geometría, excepto en los casos en que hanocurrido procesos colaterales de mezclaEl problema inverso se describe del modo siguiente:•El cavernamiento mixto epigené�co-hipogené�co encarsos de campos de petróleo (Oil-Field Karst)cons�tuye una evidencia de la migración de H2Sdesde los reservorios a la superficie y es unamanifestación indirecta de la presencia dehidrocarburos en el subsuelo•La migración de H2S es indicadora de pobrespropiedades aislantes de los sellos (que en el casopresentado en esta contribución se trata de Claystonecon fragmentos de pedernal y caliza de la FormaciónVega Alta del Paleoceno Superior –Eoceno Inferior).En el área de estudio se reconocen estos rasgos:•Desarrollo de un intenso y rápido cavernamiento(porosidad secundaria; Fig. 3)•Dominio de patrones circulares de cavernamiento aescala local y regional y en menor grado lineales,pero, sobre todo, mixtos (Figs. 4 y 5)•Preeminencia de morfologías de efecto de mezcla(Figs. 6-8)•Independencia del cavernamiento respecto a lasposiciones de las zonas de recarga/descarga,incluyendo captura de drenaje superficial•Permeabilidad de fracturas modificada por patroneslaberín�cos asociados a grandes salones de mezcla•Enriquecimiento de hierro (pirita, goethita, hema�ta)en las rocas y las aguas•Desarrollo de niveles de cavernamientoindependientes o casi independientes de la oscilacióndel nivel de base local/regional•Asociación con los yacimientos gasopetrolíferosonshore•El cavernamiento mixto epi-hipogené�co estárestringido a los sedimentos del Mioceno Superior-PliocenoLos eventuales procesos espeleogené�cos por ácidosulfúrico en Boca de Jaruco, Bellamar, Santa Catalina-Carboneras,SantaMarta–LaCachurrayVaradero,permiten suponer que durante la parte alta del Miocenomedio y superior se produjo una migración del H2S –probablemente en forma acuosa– desde el yacimientogasopetrolífero hacia la superficie. La mezcla de estasaguas con las del acuífero superior, somero, ricas en O2 yque forman parte de un karst de �pos singené�co yepigené�co provocó, al menos, un episodio de mezcla deaguas que produjo un cavernamiento (espeleogénesis)mixto epi-hipogené�co. Donde este proceso estuvoausente, solamente se desarrolló un karst de �poepigené�co clásico.Fig. 3. Distribución del cavernamientoen la margen derecha del tercioinferior de la cuenca del río Jaruco,donde se han explorado 18 cavidadesconcentradas en 3 km2 (Otero et al.,2021)57R E V I S T A D E G E O C I E N C I A SEn la actualidad, el proceso de migración parece estardetenido en los si�os inves�gados. Las aguas profundasdel yacimiento �enen un pH alrededor de 9 y con valorespor encima de 7, el H2S forma un ácido débil comoresultado de su disociación en H+ y HS-. La combinacióncon aguas ricas en HCO3-y Ca2+ causa la sobresaturación decalcita en los si�os de reducción de sulfato y, por tanto, lasaguas se neutralizan y el sistema es inerte.Fig. 4. Patrones de cavernamiento lineales y mixtos(compilación cortesía de Esteban Grau y Ángel Graña).Fig. 5. Patrón de cavernamiento circular (compilacióncortesía de Esteban Grau y Ángel Graña).Fig. 6. Morfologías hipogénicas en Cinco Cuevas (FotoVladimir Otero).Fig. 7. Morfologías hipogénicas en la Cueva de Las Muelas(Cueva Lehmann) (Foto Vladimir Otero).Fig. 8. Morfologías hipogénicas en laCueva de Don Mar�n (Foto del autor).58R E V I S T A D E G E O C I E N C I A SREFERENCIASCurl, R.L. (1966): Cave Conduit Enlargement by Natural Convec�on. Cave Notes, 8(1): 4-8Hill, C.A. (1987): Geology of Carlsbad Cavern and other caves in the Guadalupe Mountains, New Mexico and Texas. NewMéxico Bur. Mines and Mineral Res. 150:Hill, C.A. (1995): H2S-related porosity and sulfuric acid oil field karst. In: Budd, D.A., A.H. Saller, P.M. Harris (eds):Unconformi�es and porosity in Carbonate strata. AAPG, 63:301-306.Hose, L.D., J.A. Pisarowicz (1999): Cueva de Villa Luz, Tabasco, Mexico: Reconnaissance study of an ac�ve sulfur springcave and ecosystem. Journal Caves and Karst Studies. 61(1):13-21Jagnow, D.H., C.A. Hill, D.G. Davis, H.R. DuChene, K. I. Cunningham, D. E. Northup, J. Michael Queen (2000): History of theSulfuric Acid Theory of speleogenesis in the Guadalupe Mountains, New Mexico. Journal Caves and Karst Studies.62(2):54-59Lowe, D.J. (1992): The origin of Limestone caverns: An Incep�on Horizon Hypothesis. PhD Thesis. ManchesterMetropolitan Univ. Council Natl Acad. Awards, 512:Lowe, D.J. (1999): The role of speleogenesis in the development of hydrocarbon and mineral deposits. Comm.Hydrogeology and Speleogenesis, UIS.Lowe, D.J., J. Gunn (1997): Carbonate speleogenesis: An incep�on horizon hypothesis. Acta Carsologica, 26(2): 457-488Maucci W. (1952): L'ipotesi del l'erosione inversa come contributo allo studio della speleogenesi. Boll.Soc.Adria�ca Sc.Nat., 46: 1-60.Molerio León, L.F. (2003): Simulación matemá�ca de los procesos de carsificación, cavernamiento (espeleogénesis) ymigración de hidrocarburos en sistemas hipogené�cos. VI Taller Internacional Informá�ca y Geociencias, La Habana, 6:Molerio León, L.F. (2004): Procesos de cavernamiento (espeleogénesis) en sistemas hipogené�cos. Ing. Hidr. y Ambiental, La Habana,XXV (2):39-43Molerio León, L.F. (2013a): Evidencias de carsificación y cavernamiento mixto epi-hipogené�co en la CoberturaNeoautóctona de la Franja de Crudos Pesados del Norte de La Habana-Matanzas. Mapping La�no. 2 Sep�embre 2013,17:h�p://mappingla�no.com/blog/2013/09/02/evidencias-de-carsificacion-y-cavernamiento-mixto-epi-hipogene�co/Molerio León, L.F. (2013b): Resumen de la Tipología Hidrogeológica del Karst Cubano. Aragonito, No. 1, Suplemento delPeriódico Digital Espeleológico El Explorador, Cuba:1-25Molerio-León, L.F. (2019): Ges�ón de ecosistemas cársicos hipogené�cos y epi-hipogené�cos. I Simposio sobre Carso,Cuevas y Ambientes Subterráneos. XII Convención Internacional sobre Medio Ambiente y Desarrollo, La Habana, SC-004NOTA FINALLa línea de inves�gación es promisoria. La discriminaciónadecuada de los procesos de control de espeleogénesishipogénica mediante exploración directa, básicamente,puedencons�tuireventualmenteunrecursometodológico adicional para la prospección de petróleo ygas onshore. Las inves�gaciones desarrolladas en la úl�madécada por el autor han permi�do iden�ficar un grupo desistemas de cuevas en los Yacimientos de Boca de Jaruco yVaradero, en la Franja de Crudos Pesados del Norte de LaHabana-Matanzas que parecen sa�sfacer algunas de laspropiedades y patrones de las cuevas hipogénicasformadas por H2S o por ácido sulfúrico.En tanto otras no se cumplen, la inves�gación también seha orientado a la simulación matemá�ca de los procesosepigené�cos modificando, para el karst hipogénico, unmodelo construido por este autor, basado en el desarrollode los procesos termodinámicos de no equilibrio. Losresultados de la modelación conjugados con la exploraciónde campo han conducido a la iden�ficación de formashipogénicas del �po señalado, en zonas donde hoy no selleva a cabo o es muy limitada la prospección de petróleoy gas onshore.59R E V I S T A D E G E O C I E N C I A SMolerio León, L.F. y E. Grau González (2011): Episodios hipogené�cos de carsificación y espeleogénesis en el territorio deLa Cachurra-Santa Marta, Matanzas. El Explorador, Periódico digital espeleológico, Cuba, No. 89, Sep�embre 30, 2011:1-4.Molerio León, L.F., E. Balado Piedra, R. Fernández Ortega, R. Gu�érrez Domech, E. Jáimez Salgado, J. R. Fagundo Cas�llo,J. B. González Tendero, R. Lavandero Illera, J. Mar�nez Salcedo, M. Condis, L. F. De Armas, J. L. Clinche Crego, J. PajónMorejón, E. Dalmau Hevia, T. Crespo Díaz, A. Graña González, E. Vento Canosa, M.G. Guerra Oliva, A. Romero Emperador,M. C. Mar�nez Hernández, A. Mar�nez Zorrilla (2004): El Mundo Subterráneo. Universidad para Todos. Edit. Academia,Ciudad de La Habana, 31:Montoriol Pous, J., J. Assens Caparrós (1957): Sobre el papel desempeñado por el efecto salino en la génesis de ciertascavidades kárs�cas desarrolladas en las líneas de costa. Rev. Ciencias, Univ. Oviedo, VII(1):81-88Mylroie, J.E., J.L. Carew (1990): The Flank Margin Model for Dissolu�on Cave Development in Carbonate Pla�orms. EarthSurface Processes and Landforms, 15:413-424.Mylroie, J.E., J.L. Carew (1995): Karst development on carbonate islands. In/ Budd, D.A., P.M. Harris, A. Saller:Unconformi�es and Porosity in Carbonate Strata (Eds.). Memoir 63, American Associa�on of Petroleum Geologists, Tulsa,OK, pp. 55–76Núñez Jiménez, A. (1957): La Espeleología y sus progresos. Revista Soc. Cien�fica de Espeleología, I(1), abril, Santa Clara,Cuba, (sin paginar)Otero Collazo, V., A. González Ramón, L. Molerio León, O. Chávez Bonora, M. Alonso Mar�nez (2021): Sobre laespeleogénesis de las cavidades de Boca de Jaruco, Mayabeque-Cuba. Primeros resultados de las campañas de2019-2020. Bole�n SEDECK, 16, 15: h�ps://www.leg.mn.gov/docs/2013/other/130624/vol2.pdfPalmer, A.N. (1991): Origin and morphology of limestone caves. Geol. Soc. Amer. Bul.103:1–21Palmer, A.N. (2012): Geología de cuevas. Cave Books, Guangdon, China, 502:Palmer, A.N., C.A. Hill (2005): Sulfuric Acid Caves. in Culver, D.C. y W.B. White (2005): Encyclopedia of Caves. Elsevier, 654:Ros Vivancos, A., J. L. Llamusí Latorre, J. Sánchez Ortega, F. Gázquez Sánchez, J. M. Calaforra Chordi (2016): HypogenicLeslieMolerioLeónesgeólogodeYacimientosMineraleseHidrogeólogo-Hidrotécnico, MSc. en Hidrología Isotópica, Ex-Director delServicio Hidrológico de la República de Cuba, Especialista Principal enProyectos e Ingeniería en INVERSIONES GAMMA; Acreditación Profesionalpara la dirección y ejecución de obras subterráneas; Auditor Ambiental;Experto para diferentes agencias de Naciones Unidas; Miembro de laAcademia de Ciencias de Nueva York y de 38 sociedades cien�ficasinternacionales, cubanas y extranjeras. Conferencista y docente enuniversidades de La�noamérica y Europa; tutor o cotutor de 35 tesis degrado, maestrías y doctorados. Ha par�cipado en unos 480 proyectos de suespecialidad en 47 países. Tiene publicados unos 340 ar�culos en revistasespecializadas y es autor o coautor de 17 libros y folletos. Ha par�cipadoen unos 135 congresos cien�ficos. Posee numerosas condecoraciones ypremios por su labor profesional (Orden Carlos J. Finlay del Consejo deEstado de la República de Cuba, el Premio Jesús Fco. de Albear “A la Obrade la Vida” (Sociedad Cubana de Geología) y el Premio InternacionalEspeleo 2018 (España).h�ps://orcid.org/0000-0001-9667-3258h�ps://www.researchgate.net/profile/Leslie-Molerio-Leonh�ps://www.linkedin.com/in/lesliemolerio241147/60R E V I S T A D E G E O C I E N C I A SProgress on revision of Belize’s Mesozoicand Cenozoic chronostra�graphyDavid T. King, Jr. and Lucille W. PetrunyGeosciences, Auburn University, Auburn, Alabama 36849 USAAbstractThe Mesozoic and Cenozoic chronostra�graphy of Belize, which pertains to both the northern and southern onshore basinsof the country (Corozal and Belize, respec�vely), has been based up to now upon the work of early geological pioneers inBelize. During the 1950s, these early workers studied and named most of the informal stra�graphy of the former Bri�shcolony and made age es�mates for all the known stra�graphic units. These age es�mates were likely reasonable at the�me, but those same es�mated ages for Belize’s stra�graphic units have been repeated in government and corporatereports, on geological maps, and in most published papers without further significant inves�ga�on for the past ~ 70 years.Star�ng in 2003, in a series of papers addressing the sedimentology, stra�graphy, and deposi�onal environments of nearlyall the forma�ons of northern Belize, plus the Toledo forma�on of southern Belize, our research group has presentedmodern facies analyses, and – since 2016 – provided revised ages of many of the stra�graphic units in both Belize basins.A revised chronostraigraphy for Belize is important for future inves�ga�ons; and we have found that most original agees�mates are in substan�al error. In the present paper, our research group’s progress on revision of the chronostra�graphyof Belize’s two onshore sedimentary basins is reviewed, and we note where addi�onal stra�graphic work is needed.Introduc�onThe history of Belize stra�graphy is briefly summarized here star�ng with the earliest pioneering geological work. Duringthe 1920s, colonial geologist Leslie Ower published three reports on the geology (including stra�graphy) of Bri�shHonduras (Belize), which became the star�ng point for inves�ga�ons of that country’s geology and geological resources(Ower, 1927; 1928a; and 1928b). The simplis�c and incomplete view of Belize geology from Ower was not sufficient,however, to promote con�nued geological explora�on of Belize, therefore during the 1950s, Giovanni Flores, anexplora�on geologist with the Bahamas Explora�on Company, Ltd., spent considerable �me in the field and produced adetailed company report on the surface stra�graphy of northern Belize (i.e., the Corozal basin and its overlying strata;Flores, 1952a). He also authored a short report published in the AAPG Bulle�n (Flores, 1952b). It was, however, Flores’company report, including his hand-drawn geological map and type-wri�en narra�ve that became the standard referencefor northern Belize stra�graphy for decades therea�er. His age es�mates and informal stra�graphic names have beenrepeated in the years since his work in all government reports, all geological maps, and most of the few published paperson northern Belize stra�graphy. During the 1970s, exploratory drilling in northern Belize revealed two subsurface-onlyforma�ons, the Hill Bank (a.k.a. Hillbank) and overlying Yalbac forma�ons, which were not known to Flores. Theseforma�ons were described first and named by Richard Bryson (1975) in a corporate report for Anschutz OverseasCorpora�on. The history of northern Belize stra�graphic nomenclature has been reviewed in detail by King et al. (2004),who published a paper on the state of knowledge of the stra�graphy of northern Belize (‘Stra�graphy of Belize, north ofthe 17th parallel’), including both surficial and subsurface stra�graphic units. However, that paper repeats the sameerroneous age es�mates from older literature.¹Purdy et al. (2003), Petersen et al. (2012), and some other papers and reports from the 1990s depict a “Belize forma�on” in the Corozalbasin as well. This “Belize forma�on,” if it were a valid name, would consist of the Iguana Creek forma�on, Orange Walk group, andRed Bank group. However this name is not recognized as a forma�on name in northern Belize. Cornec does not use this forma�onname on any of his maps and it was not men�oned by Flores (1952a, b). Thus, we do not regard the “Belize forma�on” as a valid namefor the strata in northern Belize, and we discourage its use with regard to northern Belize stra�graphy.²As noted in Gill et al. (2018), the stra�graphic posi�on of the stra�graphic break or hiatus is actually a few meters above the contactbetween the Hill Bank and overlying Yalbac (i.e., within the lowermost beds of the Yalbac forma�on) according to proprietary well logsaccompanying the samples analyzed by Gill (2017). Gill et al. (2018) suggested that the forma�on contact is moved to the level of thestra�graphic break or hiatus in future work, and we have done so for the present paper.³This 2-m carbonate bed, which was first described by Jones (2014) is situated at the top of an outcrop located at mile marker 34 onthe Hummingbird highway, near the village of St. Margaret. The age of this carbonate bed clearly shows that it is much older than, andtherefore unlikely to be a part of, the overlying Hill Bank forma�on (Gill et al., 2018). We speculate here that this carbonate bed maybe part of the northern Belize equivalent of the San Ricardo forma�on of southern Belize (see Figure 2), which is an upper part of thesouthern basin’s Todos Santos forma�on.61R E V I S T A D E G E O C I E N C I A SIn southern Belize, colonial geologist C. G. Dixon (1957) described the main stra�graphic units in that area (i.e., the Belizebasin). Strata above the Toledo were not discussed by Dixon, but Nair (1987) produced a government report in which hefirst proposed and named the youngest stra�graphic unit in the basin, the Belize Forma�on¹, which lies above the Toledoand is subsurface only. Perhaps owing to subsequent interest in explora�on drilling in southern Belize, the subsurfacestra�graphy of southern Belize was reviewed in substan�al detail by Edward Purdy and two other geologists (Purdy et al.(2003).A modern geological map of Belize was first produced by Jean Cornec (1985; 1986) not long a�er Belize became anindependent na�on. Cornec eventually produced a series of updated geological maps for the Geology and PetroleumDepartment in Belmopan (Cornec, 1998; 2003; 2005; 2013; and 2015). All these maps have used essen�ally the samestra�graphic nomenclature and biostra�graphic age determina�ons presented by Flores and Dixon, and other coevalpublica�ons have essen�ally followed the same early pioneers’ chronostra�graphic framework. Figure 1 shows the 2015geological map by Jean Cornec to which the approximate loca�on of the Corozal and Belize basins have been added by usin the present paper.Figure 1 - Geological map of Belizeby Jean Cornec (2015); used withpermission.Theapproximate,schema�coutlinesofthetwobasins, Corozal and Belize, havebeen added by us to this map. Agesshown on this map are those ofCornec, and are not the same as therevised ages noted in the presentreport.Abbrevia�ons used in thenorthern Corozal basin and vicinity:J – Jurassic; UK – Upper Cretaceous;Et – Early Ter�ary (sic); Ltrb – LowerTer�ary (sic) Red Bank group; Ltow–LowerTer�aryOrangeWalkgroup;andQ–Quaternary.Abbrevia�ons used in the southernBelize basin: J – Jurassic K –Cretaceous; T – Ter�ary (sic); and Q– Quaternary.The same informalstra�graphic units on this map, andessen�ally the same ages used inthis map, are shown in Figure 2.This map and its separate legendcan be obtained on the web page ofBelize’sGeologyandPetroleumDepartment.62R E V I S T A D E G E O C I E N C I A SFigure 2 shows the previously accepted stra�graphy and ages of the northern (Corozal) and southern (Belize) sedimentarybasins of Belize (modified – and corrected – from an analysis by Purdy et al., 2003). On this figure, there are bulleted notes(red dots) showing where our research group’s work has found evidence for needed changes in the chronostra�graphy ofthe extant lithostra�graphic units. Please note that the informal lithostra�graphy and original chronostra�graphy of Figure2 are consistent with the 2015 geological map in Figure 1. However, the geochronologic (numerical) ages indicated onFigure 2 have been added by us, and they match with the current version of the Interna�onal Chronostra�graphic Chart(Cohen et al., 2021).Figure 2. Chronostra�graphy and correlated informal lithostra�graphy of northern Belize (Corozal basin) and southernBelize (Belize basin) according to previous workers. Bulleted points (red dots) with numbers indicate the stra�graphic levelwhere new age determina�ons have been made. Corresponding red-dot numbers below the diagram briefly describethose new age determina�ons, which are reviewed in the present report. Legend of symbols is at the bo�om, as arefootnotes regarding the stra�graphic nomenclature of Cretaceous-Paleogene boundary (KPgB) units. This diagram iscorrected and modified from Purdy et al. (2003); see also their cited references. The stra�graphic units have no consistentver�cal scale and thicknesses are not related to those actually known. Unconformi�es in the stra�graphic sec�ons areschema�c. Ages given on the chronostra�graphic column (le� side) are from the current version of the Interna�onalChronostra�graphic Chart (Cohen et al., 2021). The unit “Quaternary reef limestones, alluvium, and soils” has been addedto both basins with regard to field observa�ons of the authors (see text).63R E V I S T A D E G E O C I E N C I A SChronostra�graphic analysisThe following are some major discrepancies that we have discovered in the previously accepted chronostra�graphicrela�onships, which are bulleted in Figure 2. Regarding the Corozal basin, King’s graduate student Karena Gill (Gill, 2017;Gill et al., 2018) used stron�um-ra�o analysis of proprietary drill cu�ngs to establish the actual age range of thesubsurface-only Yalbac and Hill Bank forma�ons (the method used follows that described in McArthur et al., 2001; 2012).The clas�c and carbonate Hill Bank forma�on was found to be middle Norian (Upper Triassic) to upper Sinemurian (LowerJurassic), and the Yalbac was found to be upper Valanginian (Lower Cretaceous) to upper Turonian (Upper Cretaceous). Apreviously unknown hiatus (i.e., disconformity) between the Yalbac and Hill Bank forma�ons² spanning approximately 60million years (upper Sinemurian to upper Valanginian) was thus revealed by Gill et al. (2018). Further, a carbonate bed³from the upper part of the Margaret Creek forma�on was found to be Middle Triassic (Ladinian; Gill et al., 2018), whichsuggests that yet another hiatus may exist between the Hill Bank and Margaret Creek. Gill et al. (2018) also presentedstron�um-isotopic evidence that the basal part of the Barton Creek forma�on is lower Coniacian. Thus, a rela�vely smallstra�graphic break (disconformity) between the Yalbac and the overlying Barton Creek forma�ons, as had been suggestedby previous work, was found to span only a small part of the upper Turonian to lower Coniacian. In younger strata ofnorthern Belize, King’s graduate student Sandor Ricke�s (King et al., 2018; Ricke�s, 2020; Ricke�s et al., 2021) reportedindex calcareous nannofossils that confirmed a late Paleocene to early Eocene (lower Thane�an to upper Ypresian)assemblage for the Red Bank group of northern Belize. Of special importance in the present report is the age of the baseof the Red Bank, which was established by Ricke�s (2020) and Ricke�s et al. (2021) using drill cu�ngs as being upperPaleocene (specifically, lower Thane�an). Previously, the Red Bank was thought to range from Miocene to Pliocene (Flores,1952a; Cornec, 2015). Table 1 reviews the new age determina�ons for northern Belize stra�graphic units.Stratigraphic unitNew age of baseNew age of topAge-determinationmethodReference(s)Red Banklower Thanetian(upper Paleocene)upper Ypresian(lower Eocene)CalcareousnannofossilsKing et al. (2018);Ricketts et al.(2021)Orange WalkTBD*TBD*Iguana CreekTBD*TBD*Doubloon BankTBD*TBD*El CayoTBD*TBD*KPg boundaryend-Cretaceous(66 m. y.)end-Cretaceous(66 m. y.)Sr-ratio analysisOcampo et al.(1996)Barton Creeklower Coniacian(Upper Cretaceous)latestMaastrichtian(Upper Cretaceous)Sr-ratio analysisBase - Gill et al.(2018); Top –Ocampo et al.(1996)Yalbacupper Valanginian(Lower Cretaceous)upper Turonian(Upper Cretaceous)Sr-ratio analysisGill et al. (2018)Hill Bankmiddle Norian(Upper Triassic)upper Sinemurian(Lower Jurassic)Sr-ratio analysisGill et al. (2018)Margaret Creek?Ladinian(Middle Triassic)Sr-ratio analysisGill et al. (2018)Table 1. Summary of new chronostra�graphy for the northern (Corozal) basin of Belize, as discussed in thetext. TBD* means to be determined in the future; *These stra�graphic units must be Paleocene accordingto the KPg boundary and Red Bank age-bracket, as discussed in the text. Query (?) means unknown.64R E V I S T A D E G E O C I E N C I A SThere are four lithostra�graphic units situated between either (1) the Cretaceous-Paleogene (KPg) boundary interval innorthern Belize, which has a well-established age of 66 m.y., or (2) the top of the Barton Creek forma�on, which has anestablished stron�um-isotope age of approximately 66 m.y. as well (Ocampo et al., 1996), and the overlying Red Bankgroup (now known to be upper Paleocene at its base). These four units are, in stra�graphic order, the El Cayo group,Doubloon Bank forma�on, Iguana Creek forma�on, and Orange Walk group. All four of these stra�graphic units must,therefore, be Paleocene owing to the above-noted age bracke�ng (Figure 3). Even so, addi�onal work needs to be done onthese stra�graphic units to be�er understand the diverse sedimenta�on and age rela�onships among these four units andtheir cons�tuent facies during this rela�vely narrow �me span of Belize’s geological history.Figure 3. Revised correla�on of the global chronostra�graphic chart and the informal lithostra�graphy of northern Belize(Corozal basin), as described in the present report. The segment of the global chronostra�graphic chart shown here is fromthe current version of the Interna�onal Chronostra�graphic Chart (Cohen et al., 2021); used with permission. Eachstra�graphic unit is placed adjacent to the chronostra�graphic chart according to the new age-determina�ons describedin the present report. A query (?) means that the age of the base of the adjacent stra�graphic unit shown is not known.Two major stra�graphic breaks (marked hiatus) are shown; they represent two ‘great unconformi�es’ of northern Belize.Two rela�vely minor stra�graphic breaks (marked hiatus) are shown as well. The interval noted Paleocene* consists of fourstra�graphic units, as shown, which must be Paleogene owing to the fact that they are age-bracketed by the KPgB interval(end-Maastrich�an) and the base of the Red Bank group, which is known to be upper Paleocene (specifically, lowerThane�an; see discussion in the text).65R E V I S T A D E G E O C I E N C I A SIn the southern Belize basin, King’s graduate student Jason Fisher (Fisher et al., 2016; Fisher, 2017) used index calcareousnannofossils for age determina�on, and found that much of the outcropping Toledo forma�on of southern Belize has arela�vely narrow age range (middle Paleogene, specifically Selandian). Previously, the Toledo was thought to range fromLate Cretaceous to Oligocene (Purdy et al., 2003; Cornec, 2015). Keller et al. (2003), however, noted that the KPg boundaryinterval of southern Belize, an interval up to 25 m thick (called the Sepur forma�on by them), directly underlies the Toledoand they used planktonic foraminifera to establish that this KPg interval is lowermost Paleocene (lower Danian). Fromtheir work, it is not clear if the overlying Toledo forma�on is in conformable contact with the KPg boundary interval withinthe Belize basin, but the Toledo is clearly younger than the southern Belilze KPg boundary interval. Clearly, the KPg-Toledostra�graphic rela�onship needs to be further inves�gated, and the age of the directly overlying lower beds of the Toledoneeds to be determined.Figure 3 shows the stra�graphy of the Corozal basin and the presently known ages of the stra�graphic units of that basinin comparison to the per�nent part of the Interna�onal Chronostra�graphic Chart (Cohen et al., 2021). Because new ageda�ng for the Belize basin is limited to the Toledo forma�on (Fisher et al., 2016; Fisher, 2017) and the KPg boundaryinterval (i.e., the Sepur forma�on of Keller et al., 2003), a similar chronostra�graphic analysis for that basin is notpresented at this �me. We think that all the southern Belize forma�ons deserve a campaign of close study so that thesouthern basin’s stra�graphy and its geological history are much be�er understood.Belize’s informal lithostra�graphyThe following is a brief summary of lithostra�graphic research that has been published by King and Petruny or by King andhis Belizean graduate students. The order below follows the informal⁴ stra�graphic units of, first, the Corozal basin ofnorthern Belize, and then the southern Belize basin. This sec�on emphasizes the new findings regarding Belize’slithostra�graphy since the days of Flores’ and Dixon’s reports.Regarding the Margaret Creek forma�on, King and Petruny (2013) described the petrology and deposi�onal environmentsof this clas�c (red-bed) unit (i.e., braided fluvial systems and humid alluvial fans) and its thickness distribu�on, which istrough-like in form and trends northeast from the Maya Mountains (including drilled thicknesses up to 180 m). Later, Kinget al. (2019) reported on the detrital zircon age distribu�on from the Margaret Creek forma�on, which included anAppalachian suite of ages related to zircons derived from weathering of the underlying Upper Paleozoic Santa Rosa groupof the Maya Mountains. The Upper Paleozoic Santa Rosa group, in turn, also had an Appalachian (specifically Grenvillian)sedimentary source from the �me when Belize was s�ll part of North America (King et al., 2019), hence there are somedetrial zircons with ages over 1 b.y. in those sediments. Figure 4 shows outcrops of Margaret Creek forma�on at theinformal ‘type sec�on’ at St. Margaret’s Village.The Hill Bank and Yalbac forma�ons were studied by Karena Gill (Gill, 2017) and published by Gill et al. (2018). The paperby Gill et al. (2018) was the first openly published report on these important oil-bearing forma�ons in Belize, whichincluded deposi�onal facies analysis (clas�c coastal plain deposits to shallow shelfal carbonates in the Hill Bank and sabkhato shallow marine facies in the Yalbac), as well as the stron�um-ra�o age determina�ons as noted in the sec�on above.By virtue of its wide chronostra�grahic span, the substan�al stra�graphic break or hiatus (revealed by the gap in thestron�um-ra�o values presented in Gill et al., 2018) is likely a “great unconformity” within the Corozal basin, and is soindicated in Figure 3.⁴All stra�graphic units in Belize should be considered informal. Both the Interna�onal Stra�graphic Guide, 2nd edi�on (Salvador, 1994)and the North American Stra�graphic Code (NACSN, 2005) state that formal units must be published in a recognized scien�fic mediumand must include a statement of intent to designate that formal unit. On both of these points, all stra�graphic units in Belize fail toqualify as formal units. The original descrip�ons of Belize stra�graphic units, whose names remain in informal use today, trace back tohand-typed (and subsequently photocopied) documents (i.e., not recognized scien�fic media). Therefore, all stra�graphic units inBelize are actually “provisional informal names,” as discussed in the Interna�onal Stra�graphic Guide, 2nd edi�on (Salvador, 1994).Subsequent publica�ons in recognized scien�fic media (e.g., King et al., 2004) have referred to these units, their descrip�ons, and insome instances their informal ‘type locali�es,’ but these publica�ons have not formalized the units with statements of intent. Further,it is best if formal stra�graphy is organized by the geological en��es within a given country such as the local ‘geological survey’ orsimilar, not academic authors from other places. Owing to these relevant issues, we use the lower case “f” on forma�ons and “g” ongroups to make clear the con�nued informal status of these units and to be consistent with our previous work (e.g., King et al., 2004,and references therein).66R E V I S T A D E G E O C I E N C I A SThe Barton Creek forma�on and its various shallow water carbonate deposi�onal environments were described by Kingand Petruny (2014; Figure 5A, B). King and Petruny (2014) did not address age analysis of the Barton Creek, however,Ocampo et al. (1996) previously reported latest Maastrich�an stron�um-isotope ra�os for the upper beds of the BartonCreek (directly below the Chicxulub ejecta of the northern Belize KPg boundary interval) at Albion Island, Belize. Further,Vega et al. (1997) has reported Late Cretaceous crabs and other megafauna from uppermost Barton Creek at Albion Island.Subsequently, Gill et al. (2018) reported stron�um-ra�o age dates consistent with lower Coniacian to upper Campanianwithin the lower por�on of the Barton Creek in the Spanish Lookout area, and thus established the Barton Creek basal ageas lower Coniacian.Several researchers have studied the overlying Cretaceous-Paleogene (K-Pg) boundary (i.e., the Albion forma�on ofnorthern Belize and the coeval Cayo diamic�te of central Belize), including Ocampo et al. (1996), Pope et al. (1999; 2005),Keller et al. (2003), and King and Petruny (2003; 2015; and 2020). In northern and central Belize, this ~ 10-15 m thickinterval consists of direct ejecta from the Chicxulub impact, which is a few 100 km away in Mexico (Figure 5C). In southernBelize, the KPg boundary interval is mixed with marine sediment layers and together comprises the ~ 25-m thick Sepurforma�on (according to Keller et al., 2003).In northern Belize, between the KPg boundary interval and the Red Bank group, there are four informal stra�graphic units(Figures 2 and 3): in order of age (oldest first), they are – El Cayo group, Doubloon Bank forma�on, Iguana Creek forma�on,and the Orange Walk group. Apparently, the shallow shelfal carbonates of the El Cayo group and the cherty Doubloon Bankforma�on have not been described in any openly published source since the days of Giovanni Flores. Therefore, not muchmore can be said with specificity about their origin; and clearly they should studied more fully and their stron�um-ra�oages determined. The overlying Iguana Creek forma�on, which is a breccia and conglomerate of possible alluvial fan origin(that appears to derive from an episode of upli� in the Maya Mountains), is found mainly in western Belize. The youngerOrange Walk group of northern Belize, which is more widely distributed across northern Belize than the Iguana Creek, hasbeen studied by our research group.Figure 4. Outcrops of the Margaret Creek forma�onon the Hummingbird highway near the village of St.Margaret. Image (A) shows an outcrop that is near theoriginal ‘type locality’ of Flores (1952a) for theMargaret Creek forma�on, looking northwest. Image(B), from another part of the same outcrop in A,shows the channel deposits of several shallow,braided streams that carried sediment from theancientMayaMountains,andtheintercalatedoverbank sediments. Outcrop in (A) is approximately25 m high; 15-m scale is indicated in (B). Modifiedfrom figures in King and Petruny (2014).67R E V I S T A D E G E O C I E N C I A SFigure 5. Example outcrops of three stra�graphic units ofnorthern Belize. Images (A) and (B) show limestones ofthe Barton Creek forma�on. Image (A) was taken byJason Fisher in the valley of Barton Creek (used withpermission); (B) shows a high-wall at the road materialsquarry on Albion Island in the Rio Hondo (modified froma figure in King et al., 2004). Persons for scale in (A);outcrop in (B) is approximately 20 m in height. Image (C)shows the Albion forma�on (KPgB interval) lying uponthe Barton Creek forma�on at the road materials quarryon Albion Island. KPgB is indicated by white line. Vehiclefor scale. Image (D) shows three barrier island facieswithin the Orange Walk group, as exposed near the townof San Pablo on the Northern highway. Persons for scale.Modified from a figure in King et al. (2003).Among strata of the Orange Walk group in northern Belize, King et al. (2003) reported on the sedimentary faciesrela�onships, including barrier island sands, lagoonal limestones, and patch-reef deposits, and their molluscan andcoralline paleoecology. This work showed that barrier islands and associated coeval reef systems of the Orange Walk hada west-northwest deposi�onal strike, based upon surface outcrop analysis. The work by King et al. (2003) did not, however,address biostra�graphic age of the Orange Walk. Figure 5D shows an outcrop of the barrier island faces of the Orange Walkgroup near the town of San Pablo on the Northern highway.The Cenozoic Red Bank group was studied by King’s student Sandor Ricke�s (Ricke�s, 2020), and two papers werepublished on the petrology, biostra�graphic age, and deposi�onal environments (i.e., �dal flats and shallow marineestuaries) of the Red Bank as sampled in surface outcrops (King et al., 2018) and drill cu�ngs (Ricke�s et al., 2021). Usingwell logs, Ricke�s (2020) and Ricke�s et al. (2021) also described a dis�nc�ve, lower Red Bank zone of infiltra�on of claysinto underlying, karsted carbonate strata (the Barton Creek forma�on) in the Spanish Lookout area. Figure 6 shows anoutcrop of some estuarine clay faces of the Red Bank group in the parking lot of a restaurant in Spanish Lookout.Above the Red Bank group, there are no well-defined stra�graphic units (only Quaternary alluvium, cave deposits,shoreline deposits, and reefal limestones occupy this unnamed, younger realm of Belize stra�graphy). In northern Belize,this unnamed interval was noted and described briefly by King et al. (2004). In southern Belize, the existence of thisunnamed interval is based only on King’s field observa�ons and inferences from geological reports by others.Even though there is a possibility for as-yet unknown stra�graphic units to exist above, or as lateral equivalents to, the RedBank group in northern Belize, evidently there is no extant northern Belize equivalent of the Belize forma�on of southernBelize (as has been suggested by Purdy et al., 2003, by Petersen et al., 2012, and in some unpublished reports). Therefore,it would appear (as noted on Figure 3) that there is a second substan�al hiatus or ‘great unconformity’ in northern Belizebetween the top of the Red Bank and the base of the overlying unnamed Quaternary stra�graphic unit(s). The Toledoforma�on needs a detailed study, both sedimentologic and biostra�graphic, using drill cu�ngs and well logs, so that thewhole of the forma�on can be understood, and its age range determined.68R E V I S T A D E G E O C I E N C I A SFigure 6. Outcrop (with three close-upviews) of some of the estuarine clayfacies of the Red Bank group, which areexposed in the parking lot of WesternDairiesrestaurant,SpanishLookout,Belize. Stars indicate where the close-upviews (A), (B), and (C) were obtained.Close-up views show the lateral varia�onin clay facies color and texture. Modifiedfrom a figure in King et al. (2018).In southern Belize, the Toledo forma�on, which has been the topic of King’s graduate student’s (Jason Fisher’s) thesis(Fisher, 2017) and three related published papers, consists of deep water clas�c fan facies and other associated tectonicallyaffected and related sedimentary deposits (Fisher and King, 2015; 2016; and 2019). Figure 7A, B shows an outcrop of thecoarser facies in a submarine fan complex of the Toledo group cropping out near the town of Dump on the Southernhighway; and 7C shows an outcrop of Toledo distal fan facies along the new Mile-14 highway in southern Belize.Calcareous nannofossils from the Toledo forma�on cropping out along the Southern and Mile-14 highways in southernBelize indicate that the Toledo exposed there is middle Paleocene (Selandian; Fisher et al., 2016), even though the Toledoforma�on lies upon KPg boundary ejecta (according to Keller et al., 2003). Fisher’s new age determina�on may call intoques�on the previous broad assignment of Toledo as ‘Upper Cretaceous to Oligocene’ (e.g., Cornec, 2015), and mayindicate that there is a gap between the KPg boundary ejecta-bearing interval (called Sepur forma�on by Keller et al., 2003)and the base of Toledo.Southern Belize forma�ons below the Toledo (see Figure 2), including (1) the Todos Santos (a red-bed unit thought to bethe southern equivalent of the northern Belize Margaret Creek forma�on, and its ‘carbonate member,’ the San Ricardoforma�on), (2) the Coban forma�on (Lower and Upper parts, and the Punta Gorda ‘member’), and (3) the Campurforma�on (and its upper ‘breccia member,’ La Cumbre), have not been studied significantly since the report of Dixon(1957) and the review of the southern basin by Nair (1987). All authors that we are aware of generally relate Todos Santosto the northern Belize red beds of the Margaret Creek, and they generally relate the sequence of Coban-Campur to the HillBank-Yalbac-Barton Creek sequence of northern Belize. Whether this is correct from a chronostra�graphic point of viewremains to be seen. These southern Belize forma�ons would be much be�er understood if there were addi�onal study,including assessment of their age rela�ons.69R E V I S T A D E G E O C I E N C I A SFigure 7. Outcrops of the Toledoforma�on of southern Belize. Theupper images (A) and (B) show acoarse-grained proximal fan faciesof the Toledo, which is exposed ina small quarry adjacent to theSouthern highway near the townof Dump. Images (A) and (B) areright and le� side views of thesame outcrop (the images overlapslightly). Person for scale. Image(C) shows an example of distal fanfacies of the Toledo forma�on asexposed during 2015 on the newlyconstructedMile-14highway.Vehicle for scale.Also in southern Belize, according to Purdy et al. (2003), the deposi�onal environments of the youngest stra�graphic unitthere, the Belize forma�on, range from shallow shelfal carbonates to shoreline and deeper water clas�cs; and thethickness of the forma�on varies very widely from place to place. The forma�on does not crop out at the surface and hasbeen penetrated by only a few wells in southern Belize. There is a notable dearth of informa�on about this forma�onbeyond the paper by Purdy et al. (2003) and an older government report by Nair (1987). The wide range of lithologies andlateral varia�on of this forma�on suggests that the Belize forma�on may in fact be a group of several forma�ons.Addi�onal work should be focused on this stra�graphic unit as well.Future workWithout a correct chronostra�graphic framework for Belize, all geological interpreta�ons that rely on the previouslypublished age-dates for the informal lithostra�graphy for the country are likely to be incorrect. In geology, it is cri�callyimportant to know the ages of rocks – and the ages for the sedimentary forma�ons of Belize’s northern and southernbasins are no excep�on.Understanding sedimentary origin (deposi�onal environments) and geological ages of Belize’s stra�graphic units hasimplica�ons for tectonic interpreta�ons of the area, economic explora�on (e.g., minerals and hydrocarbons), andunderstanding how regional tectonics and global eustasy has affected this part of the Central America area over �me.70R E V I S T A D E G E O C I E N C I A SBelize has a small petroleum produc�on from two oil fields, which were discovered and drilled by Belize Natural Energy ofSpanish Lookout, Belize, and there are numerous oil seeps in the country (reviewed by King and Petruny, 2012), whichsuggests future produc�on poten�al. Also, Belize has a significant minerals industry, mainly related to dolomite mining,cement produc�on, and precious metals. Basic informa�on like the petrology and ages of the stra�graphic units of Belizetends to enhance the future for these economic endeavors.In the coming years, it is our research group’s plan to con�nue the descrip�on and age-determina�on of the stra�graphicunits of both the Corozal and Belize basins in pursuit of an accurate chronostra�graphy and thus a complete geologicalhistory for Belize.AcknowledgementsWe thank the staff of the Geology and Petroleum Department of the government of Belize for the invaluable help theyhave provided over the past 25 years for our research group’s work in Belize. Also, we sincerely thank the management ofBig Creek Group of Companies, Independence, Belize, for their financial and logis�cal support of many of our fieldexperiences in Belize. Karena Gill’s thesis research, including stron�um-ra�o analyses, and the work on the detrital zirconsof the Margaret Creek forma�on, was funded by Belize Natural Energy, Spanish Lookout, Belize. The Office of the VicePresident for Research at Auburn University as also provided support for two field campaigns in Belize.ReferencesBryson, R. S., 1975, Stra�graphy problems of northern Belize: Denver, Colorado, Anschutz Overseas Corpora�on, 22p.Dixon, C. G., 1956, Geology of Southern Bri�sh Honduras with notes on adjacent areas: Belize City, Bri�sh HondurasGovernment Prin�ng Office, 85p.Cohen, K. M., Harper, D. A. T., Gibbard, P. L., 2021, ICS Interna�onal Chronostra�graphic Chart, v. 2021/10: Interna�onalCommission on Stra�graphy, IUGS. [www.stra�graphy.org; visited: 2022-Jan-21].Cornec, J. H., 1985, Notes on the provisional geological map of Belize at the scale 1:250,000. Petroleum Office, Belmopan,Belize, 22 pp.Cornec, J. H., 1986, Provisional geological map of Belize, scale 1:250,000. Petroleum Office, Belmopan, Belize.Cornec, J. H., 1998, Geology map of Belize, scale 1:1.250,000: Belmopan, Belize, Geology and Petroleum Office.Cornec, J. H., 2003, Geology map of Belize, 1:750,000: Belmopan, Belize, Geology and Petroleum Office.Cornec, J. H., 2005, Geology map of Belize with bathymetry (1:750,000): Geology and Petroleum Office, Belmopan, Belize.Cornec, J. H., 2013, Geology map of Belize: Geology and Petroleum Office, Belmopan, Belize.Cornec, J. H., 2015, Geology map of Belize and Geology map legend: Geology and Petroleum Office, Belmopan, Belize.Fisher, J. D., 2017, Sedimentology of the Toledo forma�on, Belize basin, Central America, unpublished Masters’ thesis,Auburn University, Auburn, Alabama, 116p.Fisher, J. D., and D. T. King, Jr., 2015, Stra�graphy of the Toledo forma�on, Belize Basin, southern Belize: Gulf CoastAssocia�on of Geological Socie�es Transac�ons, v. 65, p. 107-123.Fisher, J. D., and D. T. King, Jr., 2016, Lower Cenozoic fan deposits in southern Belize: Gulf Coast Associa�on of GeologicalSocie�es Transac�ons, v. 65, p. 181-195.Fisher, J. D., and D. T. King, Jr., 2019, Carbonate facies within siliclas�c submarine fan deposits, Lower Cenozoic Toledoforma�on, southern Belize: petrography and provenance: GCAGS Journal, v. 8, p. 153-169.Fisher, J. D., D. T. King, Jr., and R. O. B. P. Da Gama, 2016, Submarine fan complex facies of the Paleogene Toledo forma�onin southern Belize: Geological Society of America, Southeastern Sec�on Mee�ng, Abstracts with Programs, v. 48,no. 3, paper no. 28-1.Flores, G., 1952a, Geology of northern Bri�sh Honduras: Bulle�n of the American Associa�on of Petroleum Geologists, v.36, p. 404-413.Flores, G., 1952b, Summary report of the preliminary geological studies of the area N of 17o N la�tude, Bri�sh Honduras:Freeport, Bahamas, Bahamas Explora�on Company, 35 p.Gill, K. K., 2017, Stra�graphy and sedimentary petrology of the Hillbank and Yalbac forma�ons, Corozal Basin, Belize:unpublished Masters’ thesis, Auburn University, Auburn, Alabama, 103p.Gill, K. K., D. T. King Jr., H. Zou, and F. Smith, 2018, Sedimentary facies analysis and stron�um-isotope stra�graphy of theHillbank and Yalbac forma�ons, Corozal Basin, Belize: Gulf Coast Associa�on of Geological Socie�es Transac�ons,v. 68, p. 229-254.Jones, A., 2014, The geology of the southernmost Corozal basin: unpublished Undergraduate Thesis, University of the WestIndies, 81p.Keller, G., W. S�nnesbeck, T. Ada�e, B. Holland, D. Stüben, M. Har�ng, C. de Leon, J. de la Cruz, 2003, Spherule deposits inCretaceous-Ter�ary boundary sediments in Belize and Guatemala: Journal of the Geological Society, v. 160, p. 783-795.71R E V I S T A D E G E O C I E N C I A SKing, Jr. D. T., K. O. Pope, and L. W. Petruny, 2004, Stra�graphy of Belize, north of the 17th parallel: Gulf Coast Associa�onof Geological Socie�es Transac�ons, v. 54, p. 289-304.King, Jr., D. T., and L. W. Petruny, 2001, Stra�graphy and sedimentology of Cretaceous-Ter�ary boundary breccia depositsin Belize, Central America: Geologia y Mineria IV (edicion especial electronica CD-ROM de la Revisita Cubana‘Ciencias de la Tierra y del Espacio’), paper no. 084, 14p.King, Jr., D. T., and L. W. Petruny, 2003, Stra�graphy and sedimentology of Cretaceous-Ter�ary boundary breccia depositsat Albion Island, Belize, in Koeberl, C., and F. Mar�nez-Ruiz, eds., Impact markers in the stra�graphic record: Berlin,Springer-Verlag, p. 203-228.King, Jr., D. T., and L. W. Petruny, 2012, Northern Belize’s onshore petroleum stra�graphy, structure, and oil seeps: GulfCoast Associa�on of Geological Socie�es Transac�ons, v. 62, p. 227-242.King, Jr., D. T., and L.W. Petruny, 2013, Stra�graphy of the Margaret Creek Forma�on, Corozal Basin, northern Belize: GulfCoast Associa�on of Geological Socie�es Transac�ons, v. 63, p. 275-283.King, Jr., D. T., and L. W. Petruny, 2014, Stra�graphy of the Barton Creek Forma�on, Corozal Basin, northern Belize: GulfCoast Associa�on of Geological Socie�es Transac�ons, v. 64, p. 215-228.King, Jr., D. T., and L. W. Petruny, 2015, Correla�on of northern Belize’s Cretaceous-Paleogene (‘KT’) boundary sec�ons: GulfCoast Associa�on of Geological Socie�es Transac�ons, v. 65, p. 463-473.King, Jr., D. T., and L. W. Petruny, 2020, Chicxulub target stra�graphy and ejecta: Insights from northern Belize: GeoGulfTransac�ons, v. 70, p. 143–151.King, Jr., D. T., J. Cornec, L. W. Petruny, D. Milham, B., Holland, S. Ricke�s, N. R. Myers, R. D. Weber, and R. George, 2018,Sedimentary characteris�cs and geological history of the Cenozoic Red Bank group, northern Belize: Gulf CoastAssocia�on of Geological Socie�es Transac�ons, v. 68, p. 269-284.King, Jr., D. T., H. Zou, K. K. Gill, L. W. Petruny, and F. Smith, 2019, Detrital zircons from the Margaret Creek forma�on,Corozal basin, northern Belize: Gulf Coast Associa�on of Geological Socie�es Transac�ons, v. 69, p. 221-231.King, Jr., D. T., L. W. Petruny, and K. O. Pope, 2003, Shallow-marine facies of the Orange Walk group, Miocene-Pliocene,northern Belize (Central America): Gulf Coast Associa�on of Geological Socie�es Transac�ons, v. 53, p. 384-397.McArthur, J. M., R. J. Howarth, and G. A. Shields, 2012, Chapter 7, Stron�um isotope stra�graphy: in Gradstein, F. M., J. G.Ogg, M. Schmitz, and G. Ogg, eds., The geological �me scale, v. 1 and 2, Elsevier, p. 127-144.McArthur, J. M., R. J. Howarth, and T. R. Bailey, 2001, Stron�um isotope stra�graphy: LOWESS version 3: best fit to themarine Sr-isotope curve for 0-509 Ma and accompanying look-up table for deriving numerical age: The Journal ofGeology, v. 109, p. 155-170.Nair, K. M., 1987, Stra�graphy of Belize basin: Geology and Petroleum Office, Ministry of Natural Resources, Belmopan,Belize, 92p.NACSN (North American Commission on Stra�graphic Nomenclature), 2005, North American Stra�graphic Code: AmericanAssocia�on of Petroleum Geologists Bulle�n, v. 89, p. 1547-1591.Ocampo, A. C., K. O. Pope, and A. G. Fischer, 1996, Ejecta blanket deposits of the Chicxulub crater from Albion Island,Belize, in Ryder, G., D. Fastovsky, and S. Gartner, eds., The Cretaceous-Ter�ary event and other catastrophes inEarth history: Boulder, Colorado, Geological Society of America Special Paper 307, p. 75-88.Ower, L. H., 1927, Features of Bri�sh Honduras: The Geographical Journal, v. 70, no. 4, p. 372-386.Ower, L. H., 1928a, The silica lines of Bri�sh Honduras: Geological Magazine, v. 65, p. 507-510.Ower, L. H., 1928b, Geology of Bri�sh Honduras: Geological Magazine, v. 36, p. 494-509.Petersen, H. I., B. Holland, H. P. Nyto�, A. Cho, S. Piasecki, J. de la Cruz, J., and J. H. Cornec, 2012, Geochemistry of crudeoils, seepage oils and source rocks from Belize and Guatemala: indica�ons of carbonate-sourced petroleumsystems: Journal of Petroleum Geology, v. 35, p.127-163.Pope, K. O., A. C. Ocampo, A. G. Fischer, F. J. Vega, D. E. Ames, D. T. King, Jr., B. W. Fouke, R. J. Wachtman, and G. Kletetschka,2005, Chicxulub impact ejecta deposits in southern Quintana Roo, México, and central Belize: in Kenkman, T., F. P.Hörz, and A. Deutsch, eds., Large Meteorite Impacts III: Geological Society of America Special Publica�on 384, p.171-190.Pope, K. O., A. C. Ocampo, A. G. Fischer, W. Alvarez, B. W. Fouke, C. L. Webster, F. J. Vega, J. Smit, A. E. Fritsche, and P. Claeys,1999, Chicxulub impact ejecta from Albion Island, Belize: Earth and Planetary Science Le�ers, v. 170, p. 351-364.Purdy, E.G., Gischler, E., and Lomando, A. J., 2003, The Belize margin revisited – Origin of Holocene antecedent topography:Interna�onal Journal of Earth Science, v. 92, p. 552-572.Ricke�s, S. O., 2020, Nature, age, and origin of the informal Lower Cenozoic Red Bank group, northern Belize: unpublishedMasters’ thesis, Auburn University, Auburn, Alabama, 145p.Ricke�s, S., D. T. King, Jr., N. R. Meyers, Sr., and D. Larsen, 2021, Upper Paleocene to Lower Eocene clay deposits of the RedBank group, northern Belize, Central America: GeoGulf Transac�ons, v. 71, p. 225-239.Salvador, A., ed., 1994, Interna�onal Stra�graphic Guide, 2nd ed.: Interna�onal Union of Geological Sciences andGeological Society of America, Boulder, Colorado, 214p.72R E V I S T A D E G E O C I E N C I A SDavid T. King, Jr. Dr. King's academic research interests are relatedto stra�graphy and Earth history.In par�cular, he studies the effects ofasteroid and comet impact upon Earth history and the stra�graphic record. Inthe area of applied research interests, his interests are petroleum explora�onand deep subsurface waste disposal.He is an author of many scien�ficpapers, including “Shallow marine-impact origin for the Wetumpka structure(Alabama, USA)”, which was published in Earth and Planetary Science Le�ers(2002). With Hal Levin, he is co-author of an historical geology textbook, TheEarth through Time, 11th ed. (2017). From 1997-2000, Dr. King served on theNorth American Commission on Stra�graphic Nomenclature. He currentlyserves on the Alabama Board of Licensure for Professional Geologists as therepresenta�ve of his state’s academic community, and he is a licensed PG inthe states of Alabama, Louisiana, and Texas.Researchgate: h�ps://www.researchgate.net/profile/David-King-JrLucille W. Petruny. Dr. Petruny’s academic research interests are related toEarth history and its connec�on to astronomy, including terrestrial impactstructures and their ejecta, par�cularly Wetumpka impact crater in Alabamaand the Chicxulub impact’s ejecta blanket in Belize, Central America. She hascompleted field geology and drill-core studies of impacts and their ejecta invarious parts of the U.S., and has done field work in Belize and adjacentMexico.Dr. Petruny was a research scien�st for two NASA-funded core-drilling projects with respect to impact structures – Chesapeake Bay (2005)and Wetumpka (2009). She is an author of many scien�fic papers andabstracts, including “Stra�graphy of Belize, north of the 17th parallel” and“Chicxulub target stra�graphy and ejecta: Insights from northern Belize,” bothof which were published in the GCAGS Transac�ons during 2004 and 2020,respec�vely.Researchgate:h�ps://www.researchgate.net/scien�fic-contribu�ons/L-W-Petruny-200481589873R E V I S T A D E G E O C I E N C I A SEstudios Glaciológicos en MéxicoAlejandro Carrillo-Chavez1*, Hugo Delgado Granados2,Lorenzo Vázquez Selem3,Guillermo On�veros-Gonzalez11CentrodeGeocienciasUNAMCampusJuriquilla,Querétaro2 Ins�tuto de Geo�sica UNAM CU, Ciudad de México3 Ins�tuto de Geogra�a UNAM CU, Ciudad de México* Correspondencia a: ambiente@geociencias.unam.mxIntroducción:Las grandes masas de hielo del planeta Tierra (la CriósferaTerrestre) han atraído la curiosidad de la humanidad desdehace cientos o miles de años. En algunas ocasiones estaatracción ha sido con propósitos de búsqueda de recursosnaturales, en otros como un desa�o depor�vo, y desde1894 con la fundación de la Comisión Internacional deGlaciares en Suiza se inicia la inves�gación cien�fica formalde los glaciares. Por ejemplo, los registros históricosindican que en el año 1518, Diego de Ordaz subió el glaciarde volcan Popocatépetl para hacer observaciones en elcrater y explorar por azufre, pero fue Francisco Montañoquien,poco�empodespués,extrajoazufreparapropósitos bélicos a encomienda de Hernán Cortes. El 8de agosto de 1786 Jacques Balmat y Michel Paccardhicieron historia al lograr el primer ascenso al Mont Blancen los Alpes, iniciando así el montañismo moderno.También tenemos las épicas expediciones al Polo Sur aprincipios del siglo pasado de Roald Amundsen, RobertSco� y Ernest Shackleton. Pero no fue sino hasta el año de1958 (año Geo�sico Internacional) que se dio un gran pasoal instrumentar las inves�gaciones cien�ficas de Glaciaresen Antár�ca. Actualmente la inves�gación glaciar puedeser dividida en Glaciología (Química, Física y Biológica) yGeologíaGlacial(Geocronología,Cartogra�a,Geomorfología, entre otras).Existe hielo glacial en extensas zonas Polares comoAntár�ca y Groenlandia sobre basamento con�nental(glaciares con�nentales), y en caso de Antár�da seex�enden grandes masas de hielo sobre el océano (en elPolo Norte existen grandes masa de hielo flotante sobreagua del Océano Ár�co que cambian de forma y tamañoconforme a la temporada. Por otro lado, el hielo tambiénse puede acumular (en menores can�dades comparadocon los glaciares con�nentales) en zonas de alta montaña(Glaciares de Montaña).El hielo glacial se forma cuandola acumulación de nieve-hielo permanece por encima dela línea de equilibrio (aproximadamente la cota a par�r dela cual la temperatura promedio anual es de menos de0°C) a lo largo de un año o más, permi�endo que la nieverecién caída y/o el hielo acumulado por diversos procesosatmosféricos (escarcha, helada, ven�sca, granizo, nieve,etc.)setransformeenhielodeglaciar,procesometamórficoenelqueloscristalesdenievesetransforman para eliminar la presencia de aire dentro de laestructura cristalina del hielo. Los glaciares �enen unrégimen de acumulación (alimentación) y pérdida (fusión,ablación) de nieve y hielo, y crecen cuando es mayor laacumulación que la fusión (ablación), y su desaparición escuando se dan las condiciones contrarias (ablación mayorque acumulación).Las grandes masas polares de glaciares con�nentales sonun verdadero control del clima de la Tierra, influyendo enla circulación de corrientes marinas, el nivel del mar entodo el planeta y actúan como un termostato. Por otrolado, los glaciares de montaña pueden aportar excelenteinformación sobre variaciones locales y regionales delclima(avancesyretrocesosglaciares),fuentesdepar�culas solidas en hielo (origen natural vs. aporteantropogénico) e incluso sobre la variación de ac�vidadsolar en la zona en donde se desarrolla el glaciar. Es decir,a la la�tud en la cual se desarrolla un glaciar, desdela�tudes altas cercanas a los polos comoel MonteMcKinley, en Alaska, y hasta el Chimborazo y el An�sana enel Ecuador.En general, las condiciones que rigen elfenómeno glacial van a estar dominadas por la la�tud y laal�tud. La nieve precipitada en la zona de acumulación delglaciar de montaña con�ene todas las par�culas de polvo(de tamaño micrométrico en la mayoría de los casos)presentes en la atmósfera en ese momento.De esta manera se precipitan y preservan en el hieloglaciar cenizas de erupciones volcánicas, cenizas deincendios forestales, aerosoles marinos (cloro, sodio,sulfato, carbonato, metales pesados, etc.). Pero tambiénse precipitan y acumulan productos de la ac�vidadhumana: cenizas de plantas industriales que quemanpetróleo (carbón, metales pesados), productos resultadode la combinación de gases y aerosoles como CO2 y SO2entre otros productos residuales de la combus�ón decombus�bles fósiles en medios de transporte (aviones,autos, camiones, etc.) y ac�vidad industrial en general.En breve, el hielo glacial puede conservar un excelenteregistro de las par�culas sólidas (micrométricas) presentesen la atmósfera. Por supuesto, el registro en �empo(cronológico) es mayor conforme el glaciar es mayor entamaño y duración, siempre y cuando se comprenda elrégimen glacial local.Interesantemente, los primeros estudios del hielo glaciar,que se realizaron a principios del siglo XX, estuvieronenfocados a estudios petrológicos del hielo glacial, por unlado el estudio mineralógico de la transformación de nieve74R E V I S T A D E G E O C I E N C I A Sa hielo de glaciar, y por otro lado analizar los cambios defase del agua que permiten explicar la transformación delvapor de agua a hielo y sus diferentes variantes cristalinas.Posteriormente, las inves�gaciones glaciares se enfocarona describir caracterís�cas �sicas de los glaciares (tamaño,espesor, forma, etc.). Después, se empezaron cuan�ficarbalances de masas en las zonas de acumulación y ablación,velocidad del glaciar, etc. A con�nuación, se empezaron aperforar los primeros núcleos de hielo para analizar elregistro de par�culas solidas en hielo y determinar lascaracterís�casquímicasdelhieloenAntár�cayGroenlandia. Posteriormente, se empezaron a extraer yanalizar núcleos de hielo en las grandes zonas montañosas(Himalaya, Alaska, Alpes, Andes). Los Glaciares de losAndes (principalmente en la Cordillera Blanca en la zonacentro-norte del Perú) son de vital importancia puesto quese localizan en la zona tropical del hemisferio sur (9ºLa�tud Sur). En base a su ubicación geográfica (la�tud), losglaciarespuedenserpolares,dela�tudesaltas,templados, híbridos o tropicales. Los glaciares híbridos(intermedios entre zonas templadas y tropicales ) y losglaciares tropicales son muy interesantes porque registrancambiosclimá�cos,atmosféricosydeoscilacionestropicales como la Oscilación del Niño (ENOS, El NiñoOscilación del Sur) y cambios de temperatura en el OcéanoPacíficodelsuresteysuimpactoencondicionesatmosféricas regionales). Adicionalmente también seregistran todas las par�culas atmosféricas naturales yantropogénicas mencionadas anteriormente.A la fecha existe muy buena documentación de laGlaciología Física yGeoquímica Glaciar en la zona deCordillera Blanca.Por supuesto que el sistema natural que controla tamaño,forma movimiento y composición química de un glaciar esmucho mas complejo. Se puede tener toda una discusiónsobre el cambio climá�co, gases efecto invernadero,calentamiento global, variaciones en ac�vidad solar,variabilidadclimá�calocalytemporal,impactoantropogénico, etc. Pero, por el momento no trataremosestos temas y nos enfocamos a una breve descripción delos estudios glaciares en México.Glaciología Física en México:Por otro lado, los únicos glaciares híbridos-tropicales(�enen un comportamiento combinando entre glaciarestemplados y glaciares tropicales) del hemisferio norte denuestro planeta se encuentran en México a una la�tud de~19º N (los glaciares de Colombia y Venezuela están enhemisferio norte, pero son básicamente tropicales).En1958,JoséLuisLorenzo,pionerodeestudiosdeGlaciología en México, describió 12 glaciares (o lenguasglaciares) en el Iztaccíhuatl, 8 (lenguas glaciares) en el Picode Orizaba, 3 en el Popocatepetl (arriba de la cota de los4,500 m), con una extensión total de 9.5 km2 (área que fuecorregida posteriormente). Para finales de los 80s variaslenguas glaciales en el Pico de Orizaba e Iztaccíhuatlhabían desaparecido y los glaciares de las tres montañasmásaltasdelpaíshabíandisminuidosutamañoFigura 1. Si�os de perforación en Antár�ca (izquierda; Glaciar Con�nental) y en Altas Montañas (derecha). Mapastomados de Bethan Davis, 2014 y Lonnie Thompson, 2017.75R E V I S T A D E G E O C I E N C I A Sconsiderablemente. Desde 1986 Hugo Delgado y suequipo de estudiantes del Ins�tuto de Geo�sica, UNAM,describió y documentó procesos de balance de masas,disminución del tamaño de los glaciares en México yactualizó la base de datos de José Luis Lorenzo de 1958.Para 2019, Jorge Cortés-Ramos, Hugo Delgado-Granados,C. Huggel y Guillermo On�veros-González reportan laevolución de glaciar del Pico de Orizaba, el mas grande delos que aun existen en México, y la relacionan con cambiosen la temperatura, precipitación y variaciones del balancede energía superficial. También analizaron imágenes desatélite y se produjo un modelo digital de elevaciónindicando los cambios del glaciar desde 1958 y hasta 2017.Por otro lado, desde el año 2008 Lorenzo Vázquez-Selem,del Ins�tuto de Geogra�a UNAM, ha documentado lahistoria de los glaciares de México desde el máximo de laúl�ma era Glacial (hace unos 20,000 años) y hasta laPequeña Era del Hielo (PEH), basándose en el estudio delas morrenas (depósitos glaciares y su datación conisótopos cosmogénicos 36Cl y datos palinológicos). LorenzoVázquez ha establecido una cronología glacial para losavances y retrocesos de los glaciares de varias montañasen México. De par�cular interés resulta la reconstrucciónde los glaciares del Iztaccíhuatl durante la PEH, periodocomprendido entre los siglos XV y XIX, en dondepredominaron condiciones más frías y que favorecieron elavance de los glaciares en las montañas de todo elmundo).Glaciología Química en México:El Centro de Geociencias, UNAM Campus, Juriquilla estádesarrollado estudios de Glaciología Química en nievereciente y hielo glaciar an�guo. Cómo se mencionóanteriormente,laspar�culassolidasmicrométricas(polvo) y algunos compuestos que circulan en la atmósferacomo aerosoles, se precipitan en la nieve (los gases comoel CO2 y SO2 se disuelven) de alta montaña y quedanpreservadas en el hielo glaciar. A la fecha, se estánperforando núcleos de hielo somero (un metro) en losglaciares de los glaciares del Pico de Orizaba y delFigura 2. Variaciones de extensión delGlaciar del Pico de Orizaba de 1958 a 2017.Imagen tomada de Jorge Cortés-Ramos,HugoDelgado-Granados,C.HuggelyGuillermo On�veros-González, 2019.76R E V I S T A D E G E O C I E N C I A SFigura 3. Izquierda; Si�os en los que Margarita Caballero, Socorro Lozano-García, Lorenzo Vázquez-Selem y Beatriz Ortegareportan evidencias prehistóricas de glaciares en México. Derecha Reconstrucción de la extensión de los glaciares delIztaccíhuatl durante la Pequeña Edad del Hielo. Imágenes tomadas de Margarita Caballero, Socorro Lozano-García, LorenzoVázquez-Selem y Beatriz Ortega (2010) y Lorenzo Vázquez-Selem, 2008.Iztaccíhuatl con el obje�vo de determinar concentracionesde metales pesados en el hielo y otros compuestos(Proyecto PAPIIT-UNAM). Los metales pesados pueden serde origen natural y/o antropogénico, como ya se indicoanteriormente.Existentécnicasquímicasactuales(isotopía estable de metales pesados) que permitendiferenciar el origen de estos metales pesados en elregistro glaciar. Se pretende u�lizar estas técnicas deQuímica Isotópica en estas inves�gaciones de la UNAM. Ala fecha se �enen datos de concentraciones de algunosmetales como Zn (promedio de 0.013 mg/l), Pb (promediode 0.000,1 mg/l), y iones como F (promedio 415 mg/l), SO4(0.800 mg/l), entro otros elementos.También existenelementos como el Berilio, que se encuentra en laatmósfera (en par�cular un isótopo del Berilio, el Berilio10que se forma por interacción de rayos cósmicos conelementos en la atmósfera, de manera similar a laformación del Carbono 14) y que se acumula en el hieloglaciar. Información sobre las concentraciones de Berilio10(10Be) en el hielo nos indican de manera indirecta lavariación de la intensidad de la ac�vidad solar.Estodebido a que la ac�vidad solar afecta directamente anuestra atmósfera y su interacción con el campomagné�co solar crea un escudo magné�co que protege ala Tierra de rayos cósmicos (las auroras boreales, yaustrales, son el fenómeno mas evidente de la interacciónde la ac�vidad solar con el campo magné�co de la Tierra).En breve, la cuan�ficación de concentraciones de 10Be enFigura 4. Izquierda: Trabajo de campo para perforación y obtención de núcleos de hielos en glaciar del Pico deOrizaba (Foto: Juan Carlos Gómez de la Fuente “Jano”). Derecha: Personal y Laboratorio de Espectrometría deMasas con Acelerador (AMS) del Ins�tuto de Física UNAM, CU, que se u�liza para medir concentraciones de 10Be(Foto de la Pagina de Laboratorio (h�p://laboratorios.fisica.unam.mx/background?id=11).77R E V I S T A D E G E O C I E N C I A Shielo glaciar nos permite conocer las variaciones en laac�vidadsolar.Este�podeinves�gacionesdeGeoquímica Glacial en glaciares de México (zona híbrida-tropical la�tud norte) nos permi�rá determinar cambios(impactos) ambientales (natural vs. antropogénico) y devariaciones de la ac�vidad solar. Los datos de lainformación obtenida sobre metales pesados, otros ionesy de 10Be será comparada con datos de la Cordillera Blanca(Glaciares tropicales la�tud sur), y de otras zonasmontañosasadiferentesla�tudes,einclusoconinformación de Antár�ca y Groenlandia. De esta manera,la UNAM con�nua con la generación de informacióncien�fica de punta que pretende contribuir al mejorconocimiento de los recursos y su uso sustentable, no solodel territorio nacional, sino de todo nuestro planeta.Conclusiones:Los Glaciares son procesos geológicos indispensables en lahistoria de nuestro planeta durante toda su historia.Aunque vivimos en un periodo de la Tierra en que losglaciares han disminuido considerablemente su tamaño enlos úl�mos miles de años, estos aun juegan un papelfundamental en fenómenos de la Tierra, tales como:reguladores de clima; controladores de nivel del mar;registro ambiental (natural y antropogénico) y climá�co;registro de ac�vidad solar y como almacén y fuente deagua para el ser humano en muchas partes del planeta.Durante los úl�mos 40 años años, prác�camente todos losglaciares de nuestro planeta, han experimentado unadramá�ca disminución, y los glaciares híbridos-tropicalesde nuestro país no son la excepción. Desde la década de1950 y hasta nuestros días se han desarrollado en Méxicoimportantesinves�gacionesdeGeologíaGlaciar,CronologíaGlaciaryGeoquímicaGlaciar.Estasinves�gaciones son muy importantes porque, aun ycuando los glaciares mexicanos desparezcan en poco�empo, el registro ambiental y climá�co que se puedeobtener de ellos es de gran valor en el entendimiento deprocesos glaciares en la�tudes de 19º N, y de esta manera,comparar la información generada con glaciares en altasla�tudes, zonas tropicales, y aun con glaciares polares. Unmejor entendimiento de los procesos y recursos naturalesde nuestro planeta conlleva a un mejor desarrollosustentable del recurso de nuestro planeta Tierra.Agradecimientos:Queremos expresar nuestro agradecimiento al “GrupoUNAM de apoyo para la perforación de núcleos de hielo enPico de Orizaba e Iztaccihuatl”: Jaime Carrera, DoraCarreón, Erandi Cerca, Gabriela Ponce “Gaby”, LorenzoOr�z “Lencho”, Juan Carlos Gómez de la Fuente “Jano”,Samael Oliver “Sama”, Julio Zacatzi, Daniela Montaño“Dany”, Jehú Hinojosa, Eber Ramirez “Gerber”, JonatánHortelano “Jona”, Libertad Barrón, Luis Acosta yRaúlFigura 5. Parte del “Grupo UNAM de apoyo para la perforación de núcleos de hielo en Pico de Orizaba e Iztaccihuatl”78R E V I S T A D E G E O C I E N C I A SGómez. El Trabajo actual de “Geoquímica de Núcleos dehielo” es posible gracias a los recursos del ProyectoUNAM-PAPIITIN110421“Concentracionesyfraccionamiento Isotópico de Zn y Hg en agua de lluvia yhielo glaciar de alta montaña: procesos geoquímicos,fuentes y trayectorias de metales en el centro de México”.Bibliogra�a:Caballero, Margarita, Lozano-García, Socorro, Vázquez-Selem, Lorenzo y Ortega Beatriz,2010. Evidencias decambio climá�co y ambiental en registros glaciales y encuencas lacustres del centro de México durante el úl�momáximoglacial.Bole�ndelaSociedadGeológicaMexicana. Vol 62, No. 3, pp. 359-377.Cortés-Ramos J., Delgado-Granados, H., Huggel, C. andOn�veros-González, G., 2019. Evolu�on of the largestglacier in Mexico (Glaciar Norte) since the 50s: factorsdriving glacier retreat. GEOGRAFISKA ANNALER: SERIES A,PHYSICAL GEOGRAPHY 2019, VOL. 101, NO. 4, 350–373 .Delgado-Granados, H., Brugman, M., 1995. Monitoreo delos glaciares del Popocatépetl. Volcán Popocatépetl,Estudios Realizados durante la Crisis de 1994–1995.CENAPRED-UNAM, México, pp. 221–244.Delgado-Granados, H., 1997. The glaciers of Popocatépetlvolcano(Mexico):Changesandcauses:QuaternaryInterna�onal, 43/44, 53–60.Delgado-Granados, H., Piedad-Sánchez, N., CárdenasGonzález,L.,2001.SulfurdioxideemissionsfromPopocatépetl volcano (Mexico): case study of a high-fluxpassively-degassingerup�ngvolcano.JournalofVolcanology and Geothermal Research 108, 107–120.Delgado-Granados, H., Julio Miranda, P., Álvarez, R.,Cabral-Cano, E., Cárdenas González, L., Correa Mora, F.,Luna Alonso, M., Huggel, C., 2005. Study of Ayoloco Glacierat Iztaccíhuatl volcano (Mexico): hazards related tovolcanic ac�vity — ice cover interac�ons. Zeitschri� fürGeomorphologie,SpecialIssueonVolcanicGeomorphology and Hazards 140, 181–193.Thompson, L. 2017. Past, Present, and Future of GlacierArchives from World’s Highest Mountains. Proceedings ofthe American Philosophical Society. Vol. 161, No. 3, pp.226-243.Vazquez Selem, L. 2008. Huellas del Pasado en las altasmontañas. Revista Ciencia y Desarrollo, CONACyT, Julio2008, pp. 40-47.Vázquez Selem, L. 2011. Las glaciaciones en las montañasMayaGeocienciasdeRevista79R E V I S T A D E G E O C I E N C I A SDr. Alejandro Carrillo-Chávez. Ingeniero Geólogo del Ins�tuto Politécnico Nacional,Maestría en La Universidad de Cincinna�, y Doctorado en la Universidad deWyoming. Inicio su trabajo en el Ins�tuto Mexicano del Petroleo y después inicio vidaacadémica en la Universidad Autónoma de Baja California Sur. En 1998 ingreso al aUnidad Inves�gación en Ciencias de la Tierra (UNICIT) UNAM, Campus Juriquilla(actual Centro de Geociencias). Su trabajo inicial fue sobre petrogra�a ígnea ymetamórfica. En academia inicio dando clases de petrología ígnea y metamórfica.Actualmente es Tutor del Posgrado en Ciencias de la Tierra UNAM. Su maestría fuesobre yacimientos minerales metálicos y su doctorado sobre geoquímica ambiental.Actualmente sus lineas de inves�gación son: Metales Pesados en Medio Ambiente,Hidrogeoquímica, Geoquímica Isotópica de Metales Pesados e Hidrogeoquímica deSalmueras Petroleras. A la fecha es responsable de un Proyecto UNAM sobreConcentraciones de metales e isotopía estable de Zn y Hg en agua de lluvia, nieve ynúcleos de hielo en glaciares mexicanos. ambiente@geociencias.unam.mxDr. Hugo Delgado-Granados. Ingeniero Geólogo de la Facultad de Ingeniería de laUniversidad Nacional Autónoma de México. Estudios de maestría y doctorado en laFacultad de Ciencias de la Universidad Tohoku del Japón. Su carrera académica haestado siempre vinculada a la UNAM, impar�endo clases desde 1983 deGeodinámica como profesor defini�vo en la Facultad de Ingeniería. Sus trabajos deGlaciología iniciaron en el Ins�tuto de Geogra�a de 1983 a 1984 conjuntamente conJosé I. Lugo Hubp. Colaboro con Jaime H. Urru�a Fucugauchi en el Laboratorio dePaleomagne�smo y Geo�sica Nuclear de 1984 a 1986 realizando estudios depaleomagne�smo en rocas volcánicas. Desde 1991 es inves�gador del Departamentode Sismología y Vulcanología del Ins�tuto de Geo�sica UNAM. Desde 1985 es elCorresponsal mexicano del Servicio Mundial de Monitoreo de Glaciares de laAsociación Internacional de Ciencias Hidrológicas de la UNESCO. Actualmente es elCoordinador del Grupo de Trabajo de Hielos y Nieves de La�noamérica y el Caribe delPrograma Hidrológico Internacional de la UNESCO y representante mexicano ante elComité Cien�fico de Inves�gación Antár�ca (SCAR), designado por la AcademiaMexicana de Ciencias y forma parte del Comité Cien�fico Asesor de la Secretaría deGobernación para el estudio de la ac�vidad del volcán Popocatépetl. Es responsablede los sistemas de vigilancia remota de emanaciones de gas en el Popocatépetl, asícomo del monitoreo de sus glaciares. hgd@igeofisica.unam.mxDr. Lorenzo Vázquez-Selem. Licenciado en Geogra�a del Colegio de Geogra�a,Facultad de Filoso�a y Letras, UNAM. Maestría en Geogra�a, División de Estudios dePosgrado, UNAM. Maestría en levantamientos de suelos (Master of Science in SoilSurvey). Interna�onal Ins�tute for Aerospace Survey and Earth Sciences (ITC),Enschede, Países Bajos. Doctorado en Geogra�a, Arizona State University, Tempe,Arizona, E.U.A. (Glacial Chronology of Iztaccihuatl Volcano, Centrla Mexico: A Recordof Environmental Change on the Border of the Tropics). Líneas de Inves�gación: 1.Geomorfología glacial e historia de las glaciaciones en las montañas de México. 2.Geomorfología volcánica. 3. Relaciones entre geoformas, génesis de suelos y erosiónde suelos. 4. Aplicaciones de la dendrocronología en la geomorfología. ActualmenteInves�gador Titular del Ins�tuto de Geogra�a UNAM. lselem@igg.unam.mxDr. Guillermo On�veros-González. Licenciado en Física BUAP;Maestría en elIns�tuto de Geo�sica UNAM (Balance de energía en la superficie del glaciar norte delvolcán Citlaltépetl); Doctorado en el Ins�tuto de Geo�sica UNAM (Estudios de ladinámica glacial del “glaciar norte” del volcán Citlaltépetl).Actualmente comoPosDoctorante en el Centro de Geociencias UNAM, Campus Juriquilla, bajo ladirección del Dr. Alejandro Carrillo-Chávez con el proyecto: “Concentraciones deCarbón Negro en nieve superficial del Glaciar del Pico de Orizaba y su Impacto en elalbedo y balance de masas del Glaciar.” gon�verosg@gmail.com80MISCELÁNEOSXaman Ek , Dios de la Estrella PolarLa quinta deidad más común en los códices es Xaman Ek, el dios de la estrella polar, que aparece 61veces en los tres manuscritos. Se le representa siempre con la cara de nariz roma y pintas negraspeculiares en la cabeza. No �ene más que un jeroglífico de su nombre, su propia cabeza, que se hacomparado a la del mono. Esta cabeza, con un prefijo diferente al de su nombre, es también el jeroglíficodel punto cardinal norte, lo cual �ende a confirmar su iden�ficación como dios de la estrella polar. Lanaturaleza de su aparición en los manuscritos indica que ha de haber sido la personificación de algúncuerpo celeste, importante.En algún lugar se habla de Xamán Ek como del "guía de los mercaderes", y bien puede haberlo sido,puesto que la estrella polar es la única estrella fija que se observa en las la�tudes del Petén y Yucatán,que no cambia radicalmente de posición durante el año. Se dice también que los mercaderes le ofrecíancopal (pom) en los altares que se ven a la orilla de los caminos. Era una deidad benévola; se la encuentraasociada con el dios de la lluvia y era probablemente el patrono del día Chuen.81R E V I S T A D E G E O C I E N C I A SThe Dinosaur House:New ZelandHaz click en la imagen82R E V I S T A D E G E O C I E N C I A SLa casa de los arrecifeshttps://www.ceupe.com/blog/que-es-un-arrecife-de-coral.htmlhttps://www.oceano.org/en/thematic-pages/the-coral/?gclid=Cj0KCQiArt6PBhCoARIsAMF5wajucemYJYvdP1ulZa_XkcxdQ4flMYXd1qQFWrEzsQ0BgoUgyd-bTOsaAh0tEALw_wcBhttps://www.nationalgeographic.org/encyclopedia/reef/https://www.livescience.com/40276-coral-reefs.htmlhttps://www.coris.noaa.gov/about/what_are/https://www.impacttravelalliance.org/ocean-conservation-a-travelers-complete-guide/?gclid=Cj0KCQiArt6PBhCoARIsAMF5wajIhfzFrYVFolbug748UJvzwcK1A7-xO52-89Yy6b2ek2OWGMJ0K8MaAsPPEALw_wcBhttps://www.youtube.com/watch?v=ZiULxLLP32s83R E V I S T A D E G E O C I E N C I A STesis selectas presentadas la Universidad Nacional Autónoma de México en 2020Compilacion realizada por Miguel Vazquez Diego Gabriel, Colaborador de la RevistaMagnitud local y parámetros de atenuación para el estado de Hidalgo.Vega Baez, María JimenaModelación de la propagación de la onda elás�ca en medios heterogéneos por el método de haces gaussianos.Vera Chávez, NainSimulaciones numéricas de intrusiones magmá�cas para estudiar el origen geodinámico del Campo VolcánicoLos Humeros.Bayona Ordoñez, Andrés DavidCartogra�a, anatomía y facies de la ignimbrita Xál�pan asociada al colapso de la caldera de Los Humeros(Puebla); evidencias de la mayor erupción de la faja volcánica trans-mexicana e implicaciones.Cavazos Álvarez, Jaime AlbertoCaracterización petro�sica de la unidad sedimentaria jurásica Cahuasas como posible almacenador de CO2.Cortés Prado, César AlbertoEstudio metalogené�co de la porción SW del Distrito Minero Tata�la-Las Minas, Veracruz.Hernández Aviles, GeovannyEstudio neotectónico de la falla agua fría, en la parte central del cinturón volcánico trans-mexicano.Lagunas Ocón, Alma DeliaCharacteriza�on of geochemical and environmental processes controling the stable isotope and trace elementcomposi�on of drip water and farmed calcite in río secreto karst cave, located in the Yucatán Península, México.Lases Hernández, María FernandaEstudio geológico del sector centro y sur del graben de Juchipila, Jal., Zac., México.Mar�nez Reséndiz, Emma VanesaRelación entre episodios tectono-magmá�co cretácico-tardío oligocéno temprano y el desarrollo de múl�pleseventos mineralizantes de plata - oro en el distrito minero San Dimas, Sierra Madre.Montoya Lopera, Paula AndreaEstudio geomorfológico y paleosismológico de la falla Ameca, Jalisco.Núñez Meneses, Andrés DavidAnálisis hidráulico-mecánico de flujo de agua subterránea y respuesta de la fase sólida por medio de modeladoacoplado.Ochoa González, Gil HumbertoLímite de provincias paleoproterozoicas Yavapai y Mazatzal en el noroeste de Sonora y suroeste de Arizona:estudios de U-Pb y Hf en zircones de rocas proterozoicas y laramídicas.Reategui Palomino, Walter VladimirPaleoclima�c and paleoenvironmental history of the northern mesoamerican fron�er.Wogau Chong, Kurt Heinrich84Cuento: Aurora(Autor: Claudio Bartolini)Aurora nació en un barrio muy pobre, al igual que millones de niños la�noamericanos, cuyo futuro está marcado por lafalta de oportunidades. Creció en una familia de once hermanos, con padres alcohólicos. La escuela primaria a la queasis�a no tenía normas de higiene, por lo que tres o cuatro veces al mes llegaba a casa con piojos, y se daba a la tarea dedespiojarse con una peineta especial que usaban los once hermanos. En una ocasión, al observar la peineta con mayordetalle, se dio cuenta que había un piojo enorme, casi del tamaño de un chícharo. Al acercar su cara para apreciarlo mejor,le dio la impresión de que el piojo hacía gestos, intentando comunicarse con ella. Efec�vamente, cuando lo acercó más asu cara, escuchó la voz ladina del piojo que decía: “Hola, me llamo Quantum, y soy un piojo que viaja a través del �empo”.Estoy enterado de que tu sueño es ser pintora y estudiar historia del arte y de las bellas artes, y yo te ayudaré a cumplir tusueño.Si tú quieres, le dijo, mientras vas a la escuela, yo te impar�ré clases durante los próximos años, solo �enes queponerme al lado de tu oreja, y desde allí me dedicaré en cuerpo y alma, para que termines tus estudios con brillantez. Laúnica condición es me permitas succionar un poco de tu sangre cada día, para sobrevivir. La niña estuvo de acuerdo con elpiojo, y aprendió cómo cuidar de él, sobre todo no dejar que su madre la espulgara, porque lo encontraría fácilmente.Aurora adquirió una habilidad sorprendente para eliminar todos los piojos de su peineta, dejando siempre a un lado aQuantum, quien se posicionaba como de costumbre, al lado de la oreja de la niña, para con�nuar con sus ac�vidadeseduca�vas.Quantum poseía un conocimiento profundo de la historia de la humanidad y del futuro de la misma. Le bastaba lamínima can�dad de sangre del cuero cabelludo de Aurora, para mantenerse vivo y con�nuar con su función de profesor.Sabía que una de las cosas que Aurora más deseaba en el mundo era la de ser pintora, por lo que consideró fundamental,el incluir cursos sobre la vida de los pintores más famosos y controversiales de la historia del arte. La niña, por su cuenta,era como una esponja, aprendía de manera impresionante, y paso a paso se conver�a en una experta en la historia del artey de las bellas artes. Quantum estaba muy orgulloso de su alumna, ya que a pesar de su mala alimentación, dedicaba�empo y alma a sus estudios. Incluso, cuando el piojo se enteraba de que ella no había desayunado en casa, se absteníade chuparle su ración de sangre, porque le daban remordimientos.Quantum le enseñaría historias de primera mano, pues el había convivido con la mayoría de los pintores de otrasépocas, y conocía sus vidas en detalle. Las primeras clases versaron sobre las vidas de Leonardo da Vinci y Miguel ÁngelBounarro�, los grandes pintores italianos del Renacimiento. Le contó que da Vinci fue un gran genio, y que además dearquitecto, fue escritor, pintor, anatomista y cien�fico-inventor entre otros. Sus dos grandes pinturas fueron la Mona Lisay la Úl�ma Cena. Miguel Ángel por su cuenta, fue un gran arquitecto, escultor y pintor, otro de los grandes ar�stas delRenacimiento. A pesar de la diferencia de edades, entre ambos, eran grandes amigos, y compar�an no solamente suconocimiento ar�s�co, sino también su amor por el hachís. En 1508, por encargo del Papa Julio II, Miguel Ángel inició lapintura de la bóveda de la Capilla Six�na del Va�cano, en Roma. Lo que el Papa nunca se imaginó, es que Miguel Ángelinvitó, a hurtadillas, a su amigo da Vinci para que le ayudara con el proyecto. Miguel Ángel pintaba durante el día, y da Vincillegaba por la noche, y ambos pintaban hasta llegada la madrugada. Los amigos pintores consumían tanto hachís, quedurante horas hablaban con Dios y con los angelitos. La agudización de sus sen�dos, los sen�mientos de felicidad ysa�sfacción, y la percepción de que el �empo pasa más lento, los mantenía muy produc�vos, y finalmente terminaron depintar la Capilla Six�na en 1512.Quantum le contó que él vivía feliz al lado de los dos afamados pintores, par�cularmente con da Vinci, quien letenía tanta confianza, que comúnmente le permi�a que se paseara por su larga barba, pero en uno de esos paseos, se vioen serios problemas, porque da Vinci y Miguel Ángel estaban fumando sendos carrujos de hachís, y todo el humo fue aparar a los pulmones de Quantum, quien empezó a alucinar que montaba a un perro enorme y rico en sangre, viviendopor un corto �empo, la vida excéntrica de un piojo millonario.En esos años, vividos bajo los efectos de la droga, dejaron también en las pinturas, miles de mensajes secretos queresultaron ser visiones del futuro, incluyendo dibujos de los robots que, en el futuro, se enviarían a Marte; la estructuradel genoma humano, y el COVID-19, entre tantos más.Una de las historias que emocionó grandemente a Aurora, fue la historia de Goya, el pintor español, quien a pesarde sufrir esquizofrenia aguda, pintó varios cientos de obras de individuos, parejas, mujeres, niños y grupos de personas.Fue muy famoso entre la aristocracia y la monarquía de su �empo. Las pinturas negras de Goya en las paredes del comedory el cuarto de espera de su casa, son tes�gos fieles de la etapa final de su vida, en la que, agobiado por su enfermedad85mental, se veía rodeado de demonios y entes maléficos de color negro, reflejando terror e histeria, agobiado por suenfermedad mental. En su mente enferma, los oscuros demonios eran su familia, y con ellos discu�a acaloradamente todolo relacionado con el arte y la pintura.Le explicó con finos detalles, que los cientos de pinturas de personajes realizadas por Goya a través de su vida,nunca exis�eron, y que el pintor jamás tuvo modelos o gente que posara para él. Todas las imágenes, en sus cuadros, serecrearon en su mente, como resultado de su esquizofrenia crónica y deterioro de un mundo real que ya no exis�a. Pegadofuertemente al cuero cabelludo del pintor, Quantum vivió una de sus experiencias mas negras y horríficas, el día en quevio en la recién terminada pintura de Goya, que estaba un piojo inmerso en un charco de sangre. Lo había detectado ysabía quien era, así que huyó despavorido. A pesar de una vida de sufrimiento mental, Goya fue sin duda, uno de losgrandes pintores de los siglos diecisiete y dieciocho.El piojo Quantum tenía claro que el �empo es infinito, pero también sabía que cuando los humanos pasan a travésde él, su �empo de vida es muy limitado. Dado a que el piojo ya se había trasladado al futuro, sabía en qué año y a quéedad moriría Aurora, y por ese mo�vo se empeñaba en enfocarse en la educación de la chica, para que sus sueños sehicieran realidad.Los cursos incluyeron desde luego al pintor Vincent van Gogh, quien fue una de las figuras del arte más influyenteen el siglo diecinueve. Las voces con las que conversaba en cada momento de su vida, le decían que de no lograr la famacon sus pinturas, tendría que suicidarse. Van Gogh realizó más de dos mil pinturas con temas de paisaje, retrato yautorretrato, u�lizando colores atrevidos y dramá�cos, que fueron la base del arte contemporáneo. El pintor tuvoperiodos muy graves de demencia, con episodios psicó�cos y de alucinaciones, por los que estuvo internado en varioshospitales psiquiátricos. Sus pinturas son el reflejo de una realidad deformada e irreal, producto de su locura, y de las vocesen su cerebro que dirigían su pincel.Quantum fue tes�go presencial de la locura de van Gogh. Él estuvo en el cuero cabelludo del pintor el día en quese suicidó, dándose un �ro en el pecho. El piojo tuvo mucha suerte, pues, de haberse disparado van Gogh en la sien, lohubiera hecho añicos. Van Gogh había hecho una apuesta con una voz que siempre tuvo en su cerebro, con quien desa�oque si no lograba ser un pintor famoso, se quitaría la vida, y al final, cumplió con su reto al pie de la letra. De hecho, a lolargo de su obra, pintó dos mil cien cuadros, sin haber recibido en vida reconocimiento alguno.Aurora estaba fascinada cuando el piojo le describió la vida de la pintora surrealista Leonora Carrington, la hija deun rico industrial de tex�les, quien se fue de casa muy joven, desoyendo las órdenes de su rígido padre. A los veinte añosse hizo amante del pintor surrealista Max Ernst, quien fuera arrestado por los nazis durante la segunda guerra mundial.Intentando ayudarlo a escapar, Leonora viajó a España, donde sufrió un colapso nervioso y fue internada en un hospitalpsiquiátrico, donde fue diagnos�cada con ansiedad paralizante y alucinaciones profundas. Durante su estancia en elhospital, fue tratada con fuertes dosis de Cardiazol y Luminal, que le terminaron de afectar su cerebro. Quantum no sedespegó de ella durante esa crisis psicó�ca, pero tuvo que aguantar los delirios amorosos de la pintora, pues muy seguidolo confundía con su amante, y quería hacer el amor con él.Tiempo después, Leonora escapó a México, donde tuvo una vida muy produc�va como pintora. Los medicamentosadministrados en España habían afectado seriamente su cerebro, y su sen�do de la realidad había sido bloqueado,solamente podía realizar pinturas surrealistas. A par�r de entonces, la pintora vivió en un inframundo del cual nunca pudoescapar. Su obra constaba de un mundo de criaturas fantás�cas, fantasmas, y con cierta magia, que caracterizaron su es�lode pintura. El mural “El mundo mágico de los mayas” en el Museo Nacional de Antropología, es una representación fiel,de un cerebro que pudo captar y pintar lo que no ven los ojos humanos. Leonora Carrington vivió noventa y cuatro largosaños, atormentada por sus propios fantasmas; sin embargo, fue una de las grandes pintoras surrealistas del siglo veinte.En el transcurso de la educación de Aurora, durante una reunión del grupo transgresor del �empo, se le dio aQuantum la oportunidad de ser otro animal, pero declinó la oferta; prefirió seguir siendo un piojo, quien había tenido laoportunidad de apoyar a miles de niños pobres en el mundo, incluso, había iniciado su carrera profesional, siendo un piojocon residencia en los primeros Homo sapiens que migraron desde el centro de África hacia Europa hace aproximadamente300,000 años.Pasaron los años y se acercaba el momento de la despedida. Aurora era ya una persona adulta, y estaba más quelista para empezar su vida profesional como pintora y maestra de arte. Se había graduado con honores en la Academia deBellas Artes. Quantum estaba melancólico, tendría que empezar una nueva vida, quién sabe adónde y en qué �empo. Laeducación que Quantum impar�ó a Aurora a través de los años rindió frutos. Aurora Reyes Flores fue una de las primeraspintoras de México, y la primera muralista mexicana. Pintó su primer mural “Atentado a los maestros rurales” en 1936, y“Mujeres en guerra” en 1937. Realizó también un sinnúmero de pinturas a través de su vida, siendo los murales su mayorpasión. Entre 1960 y 1972 pintó cuatro murales, y en 1978, terminó su sexto mural en la Casa de Cortés en Coyoacán.Cuando se inauguró este úl�mo mural, el piojo Quantum estaba allí, celebrando y aplaudiendo sumamente feliz, desde laoreja de otra niña pobre, su nueva alumna.86R E V I S T A D E G E O C I E N C I A SMoisés Dávila SerranoColega geólogo mexicano, escribió recientemente el siguiente libro,que está a la venta en Amazon en versión digital:Rupestre ModernoRupestre Moderno es una novela sobre un misterioso hombre que habita en secrecía en lasgrutas de la Península de Yucatán.Unpardeantropólogosdescubrenaccidentalmenteevidenciadesuexistencia.Se proponen encontrarlo sin saber que cuando lo consigan sus vidas se habrán transformadoirremediablemente.Durante la trama se describen las bellezas naturales del subsuelo de esa región del SuresteMexicano y los fenómenos que le dan origen.Al término de la obra, el lector habrá recibido su�lmente información real sobre la culturayucateca, su historia y evolución geológica, detalles sobre los an�guos pobladores de laregión y las bases del concepto de desarrollo sostenible.h�ps://www.amazon.com/-/es/Mois%C3%A9s-D%C3%A1vila-Serrano-ebook/dp/B09RVJ5MFJ/ref=sr_1_1?__mk_es_US=%C3%85M%C3%85%C5%BD%C3%95%C3%91&crid=2KR5V5JS070GP&keywords=modern+rupestre&qid=1644203700&sprefix=rupestre+moderno%2Caps%2C113&sr=8-187R E V I S T A D E G E O C I E N C I A SGlosario de términos geológicosCompilado por:E.P Saul Humberto Ricardez MedinaEsta compilación selecta de términos geológicos que u�lizan regularmente los profesionistas de las Ciencias de laTierra �ene la intención de apoyar a aquellos estudiantes que requieran de una referencia sobre el tema.Petro�sica: rama de las geociencias en encargada del estudio de las propiedades �sicas y fisicoquímicas de las rocas ysu relación entre los fluidos que almacenan su movimiento a través del medio poroso del yacimiento.Poro: Es el espacio vació entre par�culas que conforman una roca en el cual permite que se almacenen fluidos, estosporos pueden estar conectados o aislados.Porosidad: Es la relación de espacio vacío o volumen poroso, entre el volumen total de la roca, el resultado se expresaen una fracción de 0 a 1 y/o en porcentaje.Porosidad efec�va: Es el porcentaje de espacio vació (poros) conectados entre sí, facilitando la circulación de fluidospor las rocas.Porosidad no efec�va: Es el porcentaje de espacio vacío (poros) que no están conectados entre sí, esto imposibilita lacirculación de fluidos entre el espacio vacío.Porosidad absoluta o total: Es el porcentaje del espacio vacío total, sin importar si los poros esta interconectados ono, es decir es la sumatoria de la porosidad efec�va y la no efec�va.Porosidad primaria o intergranular: Es la porosidad que �ene origen durante el proceso de deposición de lossedimentos o materia que da origen a la roca, esto generalmente se presenta en rocas sedimentarias como lasareniscas.Porosidad secundaria o inducida: Es la porosidad originada por algún agente externo a la sedimentación (fracturas,dolomi�zación, lixiviación) y li�ficación del material para la formación de rocas.Porosidad por fractura o diastrofismo: Es la porosidad formada por acción de fuerzas tectónicas.Porosidad por dolomi�zación: Es la porosidad que se origina por una sus�tución de una molécula de calcio por unade magnesio (las calizas se convierten en dolomías), lo cual las vuelve mas porosas.88R E V I S T A D E G E O C I E N C I A SM.Sc. Wilmer Pérez Gil (Pinar del Río, Cuba, 1983) es Ingeniero Geólogo egresado de laUniversidad de Pinar del Río "Hermanos Sáiz Montes de Oca" en 2010. A par�r de 2012 ejercecomo docente en el Dpto. de Geología, perteneciente a la Facultad de Ciencias Técnicas de lareferida casa de altos estudios. Imparte asignaturas en pregrado como Geología General,Fotogra�a y Dibujo Geológico Básico, Rocas y Minerales Industriales, entre otras disciplinas. Desde2011 se desempeña como responsable de Eventos y Asuntos Editoriales de la Sociedad Cubana deGeología, en la filial de la provincia de Pinar del Río. A inicios de 2021 crea el proyecto"Geocaricaturas", grupo público de Facebook para la promoción del conocimiento de las cienciasde la Tierra, con una perspec�va educa�va a través del humor inteligente. Buena parte de lascaricaturas de temá�ca geológica que conforman esta inicia�va gráfica se han publicado ensecciones de geohumor de revistas como Ciencias de la Tierra (Chile), y Tierra y Tecnología(España). Desde finales del propio 2021 es miembro del LAIGEO o Capítulo La�noamericano deEducación de las Geociencias (IGEO, por sus siglas en inglés), donde se presenta como responsabledel Proyecto "GeoArte en América La�na y el Caribe". Posee varios geopoemas y geocuentosdedicados a la geología, algunos publicados y otros aún inéditos, donde fusiona literatura, cienciae imaginación.89R E V I S T A D E G E O C I E N C I A SCompilado por Nimio Tristán,Geólogo,Houston, TexasCristales gigantes de Naica, Chihuahua (México)Las maravillas de nuestro planeta van mucho más allá de lo que podemos ver en el horizonte. Así como existen bosquesfrondosos o islas de colores inimaginables, también hay tesoros en las entrañas del suelo. Uno de los paisajes subterráneosmás impactantes es precisamente la Cueva de los Cristales, una caverna plagadas de formaciones minerales de selenitaque alcanzan los 12 metros de largo y relucen tonos claros comparables a los de la Luna.La Cueva de los Cristales, descubierta por casualidad Todo empezó a finales del siglo XVIII, cuando un grupo de minerosdivisó una veta en el suelo a 130 km al sur de la ciudad de Chihuahua. Ese fue el inicio de la Mina de Naica, de la cual seextrajeron plata, zinc y sobre todo plomo durante más de 200 años, hasta su cierre en 2015 debido a una inundación. Enel 2000, los espeleólogos Eloy y Javier Delgado, que trabajaban a 290 metros de profundidad taladrando la roca, dieroncon una cámara que estaba junto a un depósito natural de metales. Al ingresar, los hombres vieron cristales que rebasabanlos 10 metros de largo; había algunos que atravesaban este espacio, de 20 x 35 metros, de suelo a techo y pesaban hasta55 toneladas.h�ps://www.youtube.com/watch?v=yYepZlLcp3sh�ps://cen.acs.org/physical-chemistry/geochemistry/Naicas-crystal-cave-cap�vates-chemists/97/i6h�ps://mexicotravelchannel.com.mx/ecoturismo/20201105/naica-cristales-gigantes-chihuahua/90R E V I S T A D E G E O C I E N C I A SDe Cuba a Canarias, entre líneas y colores(crónica virtual publicada en el grupo de facebook de la Filial Pinar del Río de la Sociedad Cubanade Geología)Por: Wilmer Pérez Gil (24-10-2021), a propósito de celebrarse el "Día del TrabajadorGeólogo-Minero" en Cuba.El pasado 20 de octubre de 2021, el Ilustre Colegio de Geólogos Españoles (ICOGE), presentó la inicia�va "Geosolidarioscon La Palma. Del Mal país al Buen País". La misma fue presentada en Madrid, España, por su vicepresidente, el Dr. Sr.Carlos García Royo y la Sra. Cris�na Sapalski Roselló, presidente y vicepresidenta respec�vamente de la ONG "Geólogosdel Mundo", asociaciónsin fines de lucro fundada en 1999. Dicha organización anunció en conferencia de prensa en eseentonces, su proyecto de ayuda humanitaria con el obje�vo de recaudar fondos y mejorar la calidad de vida de lapoblación palmera, afectada trás la erupción del volcán Cumbre Vieja, que al cierre de esta crónica con�nuabaarrojando lava y materiales incandescentes.A través de la comunicación vía gmail, el Sr. García Royo, le propone al Prof. Wilmer Pérez Gil, docente delDepartamento de Geología de la Universidad de Pinar del Río "Hermanos Sáiz Montesde Oca" en Cuba, una ilustraciónpara promover e iden�ficar el proyecto de ayuda humanitaria anunciado por la referida organización nogubernamental. El joven profesor cubano, quién llevaba varios meses colaborando con la revista "Tierra y Tecnología"de España, para la difusión de su proyecto personal de geocaricaturas en la península Ibérica, accede a colaborarmodestamente con el proyecto de La Palma.Lailustraciónfuerealizadaendosdías,ypresentadaensoporteacrílicoDINA3enlapropiaruedade prensa ese propio día20 de octubre (ver pósters e ilustración adjuntos), con copias para los periodistas presentes. En la imagen, a la izquierdadel póster, se observa a la Sra. Sapalski Roselló,sosteniendo la placa de acrílico con la ilustración plasmada, realizada porel profesor cubano, mientras es entrevistada por la Televisión Española y Radio Nacional de España. La imagen sepresentó en un programa televisivo de ese país (abajo a la izquierda), con la intención de divulgar el proyecto de ayudahumanitaria para los habitantes de La Palma.Hacemos extensiva la solidaridad con la población de La Palma, Canarias. Cualquier información sobre el proyecto"Geosolidarios con La Palma" para el envío de donaciones, favor de comunicarsecon el Dr. Sr. Carlos García Royo,Presidente de la ONG "Geólogos del Mundo", a través del correo electrónico:•geologosdelmundo@icog.es•cgroyo@hotmail.comA su vez, pueden visitar su si�o web en redes sociales. Contribuye con tu granito de arena a levantar a La Palma de lascenizas y a mi�gar el impactodejado por la ac�vidad erup�va en la población de la isla. SÚMATE!91R E V I S T A D E G E O C I E N C I A SAsociación de Geólogos y Geo�sicosEspañoles del Petróleoh�ps://aggep.org/Como parte de las actividades de difusión de nuestrarevista de geociencias, Tenemos una relación de buena fey amistad con las Escuelas, sociedades y asociacionesgeológicas en otros países del mundo.Sociedad Geológica de Españah�ps://sociedadgeologica.org/Sociedad Cubana de Geologíah�p://www.scg.cu/Sociedad Dominicana deGeologíah�p://sodogeo.org/Universidad Tecnológicadel Cibao Oriental,República Dominicanah�ps://uteco.edu.do/http://cbth.uh.edu/MAYAR E V I S T A D E G E O C I E N C I A S¿QUIERES COLABORAR CON NOSOTROS?ENVÍANOS UN CORREOA:luis.valencia.11@outlook.com; bernardo.garcia@ingenieria.unam.eduPiezadeMayapán,Yucatán.INAH.MUSEO REGIONAL DEANTROPOLOGÍAView publication stats | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/362822021“You can only help them save the patient once they trust you”:Clinicianperspectives and theories of use of a pediatric emergency teleconsultationprogramArticle · August 2022DOI: 10.1016/j.ssmqr.2022.100150CITATIONS0READS276 authors, including:Some of the authors of this publication are also working on these related projects:Public health surveillance View projectIntegrating Methadone and HIV Treatment Services in Tanzania View projectAmber MehmoodUniversity of South Florida101 PUBLICATIONS 606 CITATIONS SEE PROFILEJunaid Abdul RazzakWeill Cornell Medical College241 PUBLICATIONS 5,044 CITATIONS SEE PROFILEHuba AtiqChildlife Foundation5 PUBLICATIONS 51 CITATIONS SEE PROFILECarlos Castillo-SalgadoJohns Hopkins Bloomberg School of Public Health3 PUBLICATIONS 1 CITATION SEE PROFILEAll content following this page was uploaded by Amber Mehmood on 29 August 2022.The user has requested enhancement of the downloaded file.“You can only help them save the patient once they trust you”: Clinicianperspectives and theories of use of a pediatric emergencyteleconsultation programArmaan A. Rowther a,*, Amber Mehmood b, Junaid A. Razzak c, Huba Atiq d,Carlos Castillo-Salgado e, Haneefa T. Saleem aa Department of International Health, Johns Hopkins Bloomberg School of Public Health, 615 N. Wolfe St., Baltimore, MD, 21205, USAb College of Public Health, University of South Florida, 13201, Bruce B. Downs Blvd Tampa, Florida, 33612, USAc Department of Emergency Medicine, New York Presbyterian Weill Cornell Medicine, New York, NY, 10065, USAd ChildLife Foundation, 172-B, 1st Floor, Najeeb Center, Shahra e Quaideen, P.E.C.H.S, Block 2, Karachi, 75400, Pakistane Department of Epidemiology, Johns Hopkins Bloomberg School of Public Health, 615 N. Wolfe St., Baltimore, MD, 21205, USAA R T I C L E I N F OKeywords:TelemedicineRemote consultationQualitative methodsPediatric emergency medicinePakistanA B S T R A C TWhile telemedicine applications in low- and middle-income countries are growing rapidly, few studies address thesocial and organizational factors of implementation essential to sustaining clinician engagement. This study aimsto explore clinician experiences and perspectives of a pediatric emergency teleconsultation support program inSindh, Pakistan, and theories of use behind its implementation in government hospitals. We conducted in-depthinterviews with 20 stakeholders of the teleconsultation program, including program administrators (n ¼ 3),teleconsultants (n ¼ 7) and on-site physicians and nurses (n ¼ 10) purposively sampled for maximum variationacross perceived fidelity and acceptability on a preliminary questionnaire. Interview questions and probes weredesigned to elicit rich perspectives on communication structures, enabling and constraining factors, and overalleffects of teleconsultation on clinical routines and quality of care. Transcripts were analyzed thematically usingcombined inductive and deductive coding. We found that, behind the technical adaptation to using telemedicinein the resuscitation room, providers perceived a dynamic reconfiguration of professional roles and routines ofconsultation according to opportunities and constraints associated with mutual confidence or trust, constructionsof distance, subjectivity of information, and the interface of technological artifacts with medical practice. De-scriptions of communication patterns revealed an unstable tension between two competing theories of use pre-mised on either inherent contextual differences or presumed asymmetries in expertise. Long-term sustainability oftelemedicine applications to pediatric emergency care requires serious consideration of how such conflictingtheories of use and their associated assumptions about end-user needs crystallize in practice and affect clinicianengagement and perceived benefits to patient care in the future.1. IntroductionAlthough telemedicine (TM), or “healing at a distance,” has existed invarious forms for over a century (Strehle & Shabde, 2006; Wootton et al.,2017), the rapid advancement of telecommunication technology anddramatic cost reductions in recent decades have led to an explosivegrowth of TM applications globally, including in low- and middle-incomecountries (LMICs), as a strategy for improving availability of timely andappropriate health care services (Perednia & Allen, 1995; Wootton &Bonnardot, 2015; World Health Organization, 2010). The most salientadvantage put forward by TM proponents, whether policymakers or cli-nicians, is its ability to remove barriers of distance and improve thequality, equity, and efficiency of health care through better access tospecialist providers where limited by numerical inadequacy or unevengeographic distribution (Craig & Petterson, 2005; NHS Executive, 1998).Despite rapid growth and high levels of interest in TM from governmentsand industry alike, its impact on processes and quality of care remainspoorly understood. This is in part due to what McLean et al. (2013)describe in their systematic overview of TM evaluations as a failure bymost studies to adequately consider and include contextual details of TM* Corresponding author. Johns Hopkins School of Medicine, 1830 E. Monument Street, Suite 2-300, Baltimore, MD , 21205, USA.E-mail address: armaan.rowther@jhu.edu (A.A. Rowther).Contents lists available at ScienceDirectSSM - Qualitative Research in Healthjournal homepage: www.journals.elsevier.com/ssm-qualitative-research-in-healthhttps://doi.org/10.1016/j.ssmqr.2022.100150Received 17 December 2021; Received in revised form 14 August 2022; Accepted 15 August 2022Available online 20 August 20222667-3215/© 2022 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).SSM - Qualitative Research in Health 2 (2022) 100150interventions, in turn making it “difficult to disentangle the contributionsof technological and human/organizational factors on the outcomes re-ported.“ (McLean et al., 2013) Moreover, only one of every fourgovernment-sponsored telehealth programs has undergone any form ofevaluation, and the role of clinicians in research is often limited toquantitative measures of end-user acceptance or satisfaction with TM(Kissi et al., 2020; Scott Kruse et al., 2018; Wade et al., 2014; WorldHealth Organization, 2017). An improved understanding of the effects ofTM will require qualitative elucidation of the human and organizationalfactors of implementation, including influence on clinical roles, re-sponsibilities, and decision-making (Mathijs, 2007, p. 7), and is essentialto realizing the full potential of TM to enhance patient care and clinicaloutcomes.Synchronous applications are among the longest established forms ofTM yet have come to be far outnumbered by asynchronous uses, such asremote interpretation of medical images and other store-and-forwardprograms considered less difficult or costly to implement, especially inresource-limited settings (Krupinski et al., 2002; Wootton & Bonnardot,2015; Zaylor, 1999). As such, appropriation has been slower in clinicalcontexts such as pediatric emergency medicine (PEM), where real-timevideoconferencingcapabilitiesarecrucialtoenablinginteractivecommunication and accurate clinical interpretations (Bregstein et al.,2021; Gattu et al., 2016; Sauers-Ford et al., 2019). With the advancementof telecommunication technologies and added urgency of the COVID-19pandemic, applications of TM have expanded rapidly and include diverseforms such as teleconsultation (e.g., synchronous consultation byvideoconferencing or phone), tele-expertise (e.g., asynchronous writtenor verbal specialist opinion about a clinical case), and telemonitoring(e.g., use of medical measurement tools that allow recording and transferof health data to medical centers or professionals for regular trending orproblem-based detection) (Dorsey & Topol, 2020; Hollander & Carr,2020). With critical shortages in trained PEM specialists and substantialdisparities in access to high-quality care, including in high-incomecountries,pediatricemergencyTMholdsimmensepotentialforaddressing these challenges as videoconferencing technology becomesmore accessible (Medicine, 2007; Uscher-Pines & Kahn, 2014). Oneparticularly promising TM care delivery strategy is real-time, video--assisted teleconsultation between providers, namely pediatric emergencyhealth workers in rural or remote areas and PEM specialists locatedoff-site, which may facilitate not only narrowed quality gaps but alsopoint-of-care education (Foster et al., 2020; Ray et al., 2017; WorldHealth Organization, 2018). Professional-centered TM technologies suchas teleconsultation, tele-expertise, and telemonitoring are increasinglyused to provide rural or underserved hospitals with access to pediatricand emergency care specialists (Dharmar et al., 2013; Marcin et al., 2004;Mueller et al., 2014). However, as with TM programs in general, PEMapplications are largely pilot programs with little evaluative data avail-able, and end-user involvement is often limited to surveys on feasibility,acceptability, or satisfaction without thick description or rigorous anal-ysis of the clinician experience or integration into practice routines(Dharmar et al., 2013; Gattu et al., 2016; Heath et al., 2009; Lehoux et al.,2002; Mair & Whitten, 2000; Nicolini, 2006). These gaps in existingpediatric emergency TM evidence point to a broader paucity of studiesthat transcend the technological black box of TM to include the humandimensions that make practice possible and influence success or failure ofimplementation (May & Ellis, 2001; Whitten et al., 2000).Nearly all (99%) of the estimated 6.3 million children less than 5years of age who die each year live in LMICs, and 29% of those deathsoccur in South Asia (Liu et al., 2015; Wang et al., 2014). While up to athird of sick children in these countries are critically ill and requirereferral to a hospital, the emergency and critical care services available tothem remain one of the weakest components of health systems in LMICs(Kalter et al., 1997; Nolan et al., 2001; Razzak & Kellermann, 2002).Even when severely ill and injured children are received by staffedemergency departments (EDs) instead of ill-equipped hospital wards oroutpatient facilities, the quality of care is often inadequate or untimely,contributing to half of the child deaths in LMIC hospitals occurringwithin 24 h of presentation (Molyneux, 2001; Nolan et al., 2001).Improving quality of pediatric emergency care through the rapid iden-tification, efficient prioritization, and appropriate treatment of childrenwith acute, severe, and reversible disease is essential to closing the widegap between international goals for reducing child mortality and stag-nating or worsening under-five mortality rates in many LMICs (Baker,2009; Liu et al., 2016; UNICEF, 2015). The urgency of this need to reducechild mortality is greatest in countries like Pakistan that face a ‘doubledisease burden’ of infectious diseases on top of the health challenges of arapidly industrializing society such as injuries (Boutayeb, 2006; Jafaret al., 2013). This situation has contributed to Pakistan having one of thehighest under-five mortality rates in the world: 74 deaths per 1,000 livebirths, according to the most recent Demographic and Health Survey(DHS) (National Institute of Population Studies - NIPS/Pakistan and ICF,2019).Despite Pakistan's constitutional guarantees of health care provisionby federal and provincial governments, historical barriers to healthsector improvements and reforms have included ethnic and politicalpolarization and the prioritization of national security in the face ofrepeated domestic terrorist attacks, prolonged military entanglements inKashmir and Afghanistan, and high levels of violent crime in megacitieslike Karachi (Ahmed & Nisar, 2010; Bhutta et al., 2013; Khan, 2010).Complex security issues, natural disasters, and political fragility havefurther contributed to a high burden of disease, limited access to careamong rural populations and the urban poor, and staff shortages andabsenteeism in large segments of the Pakistani health system, includingemergency medical services (Bhutta et al., 2013; Closser, 2010; Ghum-man & Horney, 2016; Kirsch et al., 2012; Umer et al., 2009; Watkins,2016). Public sector hospitals nonetheless remain the most accessiblesource of pediatric emergency care to the large portions of Pakistan'spopulation unable to afford private urban facilities for their children(Hussain et al., 1997; Mehdi, 1996).The vast majority of available research on Pakistan's emergency caresystem are focused on Karachi, the nation's largest city and among themost populous in the world (population of 16.8 million), where PEDsreceive a high proportion of critically ill children (CIA, 2020). At one ofKarachi's largest PEDs, 8% of patients present in the highest triage acuity,and overall mortality is as high as 13% (Habib & Khan, 2018). Contrib-uting factors may include delays in presentation and treatment forchildhood illnesses, whether due to transportation barriers after referralto Karachi from remote areas, failures to recognize danger signs requiringtreatment escalation, or hesitation rooted in perceived poor quality ofcare at public hospitals, phenomena that have been observed in previousmaternal and child health research in regions of South Asia (Aftab et al.,2018; de Zoysa et al., 1998; Qureshi et al., 2016; Razzak et al., 2001; Syedet al., 2008). Challenges facing those who attend tertiary care facilities inPakistan include very high patient volumes exacerbated by frequentovercrowding, ambulance diversions, and long wait times that createadditional barriers to receiving timely and effective emergency care(Habib & Khan, 2017; Razzak et al., 2008; Rehmani, 2004). Shortages inadequately trained PEM specialists and low adherence to standard clin-ical guidelines for child health and emergency care represent key gaps inhealth system capacities in Pakistan and many other LMICs to deliverhigh-quality care to acutely ill children (Bickler & Rode, 2002; Bryceet al., 2003; Nolan et al., 2001; Razzak et al., 2015). Consistently poorperformance on health indicators such as under-five mortality, in thecontext of a strained pediatric emergency care system, has contributed toa growing impetus behind the development of innovative strategies forimproving access and quality of PEM care.In this study, we sought to further the available evidence for TMapplications in LMICs to strengthen pediatric emergency care by exam-ining clinician experiences and perspectives of a teleconsultation pro-graminSindh,Pakistan,andthetheoriesofusebehinditsimplementation as a quality improvement intervention in governmenthospitals. We employed in-depth interviews and qualitative thematicA.A. Rowther et al.SSM - Qualitative Research in Health 2 (2022) 1001502analysis to understand impacts of TM on clinical roles and performanceas well as emergent factors within the social and structural context of thepediatric EDs of Pakistani government hospitals.2. Theoretical framework2.1. TM theory of useIn our exploration of clinician experiences and perspectives of a pe-diatric emergency teleconsultation program, we applied a conceptualframework adapted from the TM Theory of Use framework (Fig. 1)derived and presented by Lehoux et al. (2002) (Lehoux et al., 2002).The model was based on qualitative investigation of clinicalcommunication routines in the Quebec Inter-Regional TelemedicineNetwork through interviews with physicians in six medical specialtiessupplemented by observations of video-recorded consultations. Lehouxand her colleagues' efforts to develop the framework was in response towhat Rashid Bashshur (1997) referred to as the ‘technological dilem-ma’—the danger for TM to be “a solution looking for problems”, pro-ducing ill-fitting, inefficient applications to clinical settings—and theneed to reduce the lag between TM practices and the research informingthem (Bashshur, 1997). The TM Theory of Use framework aims to helpbridge such gaps by describing an approach for researchers to articulateintended versus actual uses of telehealth, with attention to often over-looked non-financial costs to users (Lehoux et al., 1999, pp. 29–38; Pel-letier-Fleury et al., 1997).A ‘theory of use’, according to Lehoux et al. (2002), describes anunderlying logic or rationale for TM and the associated “manner in whichclinicians are expected to use it in their daily activities” (Lehoux et al.,2002). Their focus on breaches between theory and actual practice re-sembles elements of organizational learning literature, including argu-ments by Argyris and Schon (1974) about incongruence betweenespoused theories of action and lived “theories-in-use” hindering pro-fessional effectiveness (Argyris & Schon, 1974; Denis & Lehoux, 2013). Aunique strength of the TM Theory of Use, however, is its framing ofclinical communication routines shaped by TM as not only mental modelsbut also social practices, resulting in a framework for studying tele-consultation that draws heavily from concepts of a sociological theory,namely Giddens' Structuration Theory.An important and distinguishing feature of Giddens' (1984) socialtheory is his insistence on the knowledgeability and reflexivity of agentswho can observe, monitor, reflect on, and evaluate their actions andthose of others and thereby modify social practices within structures ofconstraint (Giddens, 1984). Applying Giddens' duality of structures,Lehoux et al.’s (2002) TM Theory of Use model posits clinicians asknowledgeable, reflexive agents and each TM program as a structurethat, rather than operating “behind the backs” of users, is embedded inand shaped by knowledge and practices of providers recursively (Lehouxet al., 2002). The result is an approach to TM that views its introductionand utilization as a process of sociotechnical change, one that requiresattention to opportunities and constraints relevant to human and socialfactors such as trust, power, and gender.2.2. Sociotechnical perspective of TMBetween the determinism of decision theorists and the interpretivismof institutionalists, a sociotechnical perspective has emerged within thelast three decades of science and technology studies that emphasizes thereciprocal influence between technological innovation and its socialcontext (Akrich, 1992; Barley, 1986; DeSanctis & Poole, 1994; Jasanoffet al., 2001; Mintrom, 2015, pp. 12–21; Prasad, 1993; Sismondo, 2010).This perspective, which is represented in multiple TM theoreticalframeworks including the TM Theory of Use, advocates for researchwhich also considers how people incorporate technological innovationssuch as teleconsultation or tele-expertise and the emergent interactionsthat are mediated by new technologies when socially situated in appli-cation (Lehoux et al., 2002; Sicotte & Lehoux, 2003; Trupia et al., 2021;Whitten et al., 2000). Contentions of this perspective include that tech-nologies may embody intentions of their design but also require work byusers to stabilize as assemblages of human and technical elements and toassume practical meanings that only emerge once introduced in practice(Nicolini, 2006). There is a paucity of studies that examine TM-in-usefrom such a perspective across the evaluative literature about TM andhealth care delivery, which has been limited in its attendance to thedetails of how TM programs are translated to use in clinical practicebeyond piloting or initial implementation stages May et al., 2003.While teleconsultation or tele-expertise programs are often designedbased on assumed asymmetries of expertise—where specialists canconfirm diagnoses or advise treatment options remotely—interveningtechnologies may modify communication routines by introducing novelforms of interdependence in the attending–consultant relationship, asphysical distance from the patient renders the teleconsultants' role reliantFig. 1. Teleconsultation in use: conceptual framework adapted from Lehoux et al. (2002) TM theory of use and essential features of TM as described by Mortet al. (2003).A.A. Rowther et al.SSM - Qualitative Research in Health 2 (2022) 1001503on attending physicians’ clinical data gathering, reporting, and evenprocedural skill (Deldar et al., 2016; Mathieu-Fritz & Esterle, 2013;Sicotte & Lehoux, 2003). Point-of-care education as an explicit objectiveof teleconsultation also complicates clinical communication by extendingits function beyond mere information transfer (Ray et al., 2017). Tele-consultation systems have the potential, therefore, to not only facilitateaccess to specialist advice but also cause substantial changes in clinicalcommunication patterns as integration of the TM system in practice re-quires negotiations of power and trust among clinicians at either end ofthe call (May et al., 2003).Such aspects of adaptation and normalization for TM systems reflectsone example of how the technical and social facets of teleconsultation areintimately intertwined, such that social context moderates the effects ofany TM application, including the structures and logic intrinsic to itsdesign (Mathieu-Fritz & Gaglio, 2018). The resulting reconfiguration ofscripts in routines and interactions of physicians by a TM program,described by Nicolini as “the work to make telemedicine work,” can be anuncertain, unpredictable, and often invisible process, the outcome ofwhich may only be fully appreciated by close study of TM appropriationin practice (Gherardi, 2010; Nicolini, 2006; Trupia et al., 2021). Forexample, in her study of webcams in telehealth, Pols found that theintervening technology magnified characteristics of social relationshipsby modifying relational distance between end users (Pols, 2011). More-over, the novelty of TM contributes to relative instability of anynormative sequences, which are negotiated but not embedded intoinstitutional modes of talk for participants, as found by Pappas, Seale,and others in conversation analyses of the opening, physical exam, anddiagnostic interactions of telecardiology, including interprofessional in-teractions involving on-site providers (Pappas et al., 2019; Pappas &Seale, 2009, 2010). As such, studies of communication patterns andpractice-based research of teleconsultation are imperative to advancingour understanding of how professional-centered TM technologies thatare routinized in health care delivery affect clinical team communication,a prerequisite to influencing clinical decision making and resulting pa-tient management (Deldar et al., 2016).3. Materials and methods3.1. Study settingInterviewees were recruited from five participating public hospitals inurban or peri-urban areas of Karachi and from the teleconsultation pro-gram under study, initiated in 2016 through a public-private partnershipbetween the provincial government of Sindh and the private nonprofitorganization ChildLife Foundation (CLF). The setting for this researchwas Sindh Province, one of four administrative provinces of Pakistan, atCLF-administered pediatric emergency departments (PEDs) concentratedin the megacity of Karachi. CLF was launched with a mission of providingquality and affordable healthcare to the children of Pakistan in 2010 inpartnership with the Government of Sindh. In 2016, CLF introduced thefirst TM center in Karachi, Pakistan, using synchronous video-assistedteleconsultation and monitoring to support public hospitals. The ratio-nale for the CLF teleconsultation program was to help address humanresource challenges in pediatric emergency care, most notably theshortage and inequitable distribution of PEM specialists, who aregeographically concentration in a limited number of urban areas inPakistan. Goals of the program include improving health worker per-formance by providing expert teleconsultation to attending providers atremote sites as well as detecting unsafe conditions, improving adherenceto evidence-based protocols, and helping avoid preventable morbidityand mortality of children in the PED.As a digital health intervention, the CLF TM program falls underCategory 2.4 of the World Health Organization (WHO) classificationsystem (World Health Organization, 2018): consultation on case man-agement among health care workers. The program also involves com-ponents of decision support through protocol checklists and health careprovider communication. The basic design of the CLF teleconsultationprogram is as follows: each CLF-supported PED has a resuscitation roomforcriticallyillpediatricpatients.Ataminimum,onephys-ician—primarily responsible for clinical evaluation and treatment—ispresent alongside one or more nurses who are tasked with inputtingclinical data into the electronic medical record (EMR) and assisting withpatient care. In addition, based on the clinical severity and indication, thepatient's real-time telemetry data is monitored and recorded. Locatedwithin each PED resuscitation room is a TM camera and phone to facil-itate the synchronous video-assisted teleconsultations between theon-site care team and the remote PEM specialist with additional trainingand/or certification in advanced pediatric life support and TM supportdelivery. Teleconsultants are located at a remote TM center in Karachiand have display monitors that accommodate up to six camera views andcomputers at each station with access to telemetry data and some otherclinical information available from the EMR. Teleconsultants also haveremote controls that can pan, tilt, or zoom the camera at any of the PEDs.The TM program provides round-the-clock access to remote specialistswho can provide both consultation, such as advice on diagnosticreasoning or treatment decisions, and monitoring through the TM cam-eras, which facilitates visual assessment of patients and evaluation ofproviders' clinical practices (e.g., CPR performance).The tasks involved in the TM program include consultation (e.g., aphysician at a remote site requests a second opinion or assistance with apatient case), intervention (e.g., the teleconsultant contacts the attendingprovider about a concern or unsafe condition detected while moni-toring), virtual rounding (at periodic intervals, the teleconsultant hasclinical staff provide updates on admitted patients), and evaluative ob-servations of system and provider performance (e.g., whether vitalmonitoring is connected and functioning, adherence to CPR guidelines,observation of best practices for infection control, etc.). For the tele-consultant to accomplish these tasks, the system is required to facilitatetimely communication, clear visibility of the field by camera, and accessto telemetry and other clinical data. An additional requirement of thesystem is that the communication modality not interfere with or delayprompt treatment by on-site providers during pediatric emergencies.3.2. Study settingSemi-structured in-depth interviews were conducted with 20 stake-holders of the PEM teleconsultation program, including teleconsultants(n ¼ 7), on-site physicians and nurses (n ¼ 10), and administrators (n ¼3), from November 2019 to January 2020. Interviews were conducted inUrdu or English (according to participant preference) by the first author,a bilingual Pakistani American man with doctoral-level training in publichealth and qualitative research methods, undergraduate medical edu-cation at a US medical school, and a decade of research experience thathas included extensive work with emergency health workers and patientsin Pakistan.Teleconsultants and on-site providers were purposively sampled fromamong respondents to a brief structured questionnaire to maximizesample variation across level of perceived fidelity and acceptability of theteleconsultation program (Steckler et al., 2002). The questionnaire wasadministered to a convenience sample of 44 TM end-users from amongday- and night-shift workers at each site and included Likert-typeresponse choices for level of agreement with eleven statements relatedto: burden/ease of use, satisfaction with TM experience, attitude towardintervention, opportunity costs of use, appropriateness of use, strength ofaudio and visual connection, and the perceived improvement to qualityand accessibility of care (Appendix 1). Questionnaire participants weregrouped into response categories according to indicated level ofperceived fidelity and acceptability, and recruitment for subsequentin-depth interviews was targeted to include TM consultants and on-siteproviders across all response categories until saturation was reached.This study was funded by a grant from the Johns Hopkins UniversityCenter for Global Health, and ethical approval for this research wasA.A. Rowther et al.SSM - Qualitative Research in Health 2 (2022) 1001504obtained from the Institutional Review Board (IRB) of the Johns HopkinsBloomberg School of Public Health in Baltimore, Maryland, and from theIRB of Dow University of Health Sciences in Karachi, Pakistan. Verbalinformed consent was obtained from questionnaire respondents, whilethose recruited for in-depth interviews provided written consent. In-terviews lasted between 30 min to an hour and were conducted in Urduand English. Initial interview topics included: understanding of the TMprogram's purpose and appropriate use; extent of engagement with tel-econsultation; events or influences in the program setting that eitherfacilitate or impede teleconsultation; and subjective effects of TM supporton personal motivation, clinical competence, and provider confidence.Interviews were digitally audio-recorded with consent of participantsand translated to English so that all final transcripts were in English priorto analysis.3.3. Data analysisData analysis was started during data collection to help informquestions and probes of subsequent interviews with emergent themes inan iterative manner and to effectively determine when saturation wasreached. Analytic memo writing continued throughout this process untilall in-depth interviews were coded and results finalized. Thematicanalysis involved reading and re-reading of transcripts several times andthe use of an initial subset of transcripts for open coding and determiningemergent themes (Salda~na, 2021). These were then organized underdeductive categories drawn from essential features of TM described byMort et al. (2003) and the TM Theory of Use derived by Lehoux et al.(2002) (Fig. 1).. (Lehoux et al., 2002)The TM Theory of Use applies sociological constructs of structurationto explain the human and organizational dimensions of TM's use incontext that exist beneath the technological interface, such as explicit orimplicit assumptions about the needs of users and the resultingopportunities or constraints associated with incorporation into clinicalroutines (Giddens, 1984). The use of structuration theory to study TMgenerally and teleconsultation in particular provides a holistic perspec-tive that emphasizes reciprocal influence between the technical innova-tion and clinical context (Sicotte & Lehoux, 2003). Under such aperspective, technologies such as TM are acknowledged to embody aimsof their designers but also produce emergent interactions and meaningsonce integrated by users into routine practice and situated in a particularclinical and social setting (May et al., 2003). End users, who were in thiscase attending providers and remote teleconsultants, are recognized asagents responsible for the often invisible work required for TM normal-ization (Akrich, 1995; Trupia et al., 2021).The conceptual framework was refined for relevance to additionalcodes or themes as they emerged and was used to develop naming anddefinitions for a common thematic map (Fig. 2). Codes were applied tointerview transcripts using the Dedoose computer-aided qualitative dataanalysis software (Dedoose, Version 8.2.14, Los Angeles, CA, USA). An-alytic memos from throughout coding and data collection, frequentrevisiting of the original interview transcripts, and periodic researchmeetings to discuss emergent subthemes and negative findings were usedto generate the final thematic analysis and key results (see Fig. 3).4. ResultsThe sample of TM end-users included teleconsultants, on-site pro-viders, and program administrators affiliated with CLF and employed ateither the TM aggregation site or one of five government hospitals inKarachi or surrounding peri-urban areas (Table 1). Length of exposure tothe teleconsultation program ranged from two months to three years(from the time of program inception), and clinical roles of the on-siteproviders included nurses, junior medical officers, and senior consul-tants. Teleconsultants either completed or were completing PediatricFig. 2. Common thematic map: categories of conceptual framework for teleconsultation in use combined with emergent themes and subthemes derived fromclinician interviews.A.A. Rowther et al.SSM - Qualitative Research in Health 2 (2022) 1001505Emergency Medicine certification. The gender distribution of partici-pants was two males and nineteen females, which reflected the overallgender distribution of providers and staff at CLF.The study findings are organized according to seven sub-themes, twoof which describe competing but coexistent theories of use that emergedfrom participant accounts of the assumed user needs and underlyingcommunication structures in the teleconsultation program (Symmetry:Cameras and Context; Asymmetry: Experience and Expertise). The sub-sequent five subthemes represent areas of perceived opportunities andlimits for TM based on clinicians’ descriptions of the enabling and con-straining contextual factors specific to emergency pediatric care inPakistan. Four of these are deductively defined features of TM (Distance,Information, Medical Practice, and Technological Artifacts) while onewas inductively derived (Confidence/Trust).4.1. TM theories of use: communication structures and users’ needsTwo distinct communication structures and associated theories of useemerged from participant descriptions of the purpose, appropriate use,and mechanisms of improved quality of care underlying the pediatricemergencyteleconsultationprogramingovernmenthospitalsofPakistan. In one theory of use, the contributions of teleconsultants relymore on differences of context than on seniority, since being distancedfrom mental, physical, and emotional burdens of a busy public ED allowsthem to review transmitted clinical data and findings more calmly orFig. 3. Contextual factors of teleconsultation that weigh toward symmetric vs asymmetric structures of provider communication and their corresponding theoriesof use.Table 1Demographic profile of participants (n ¼ 20).Participant*Role*Months of TM ExposureFacility SettingGenderOP1MO6Peri-urbanFemaleOP2MO2Peri-urbanFemaleOP3RN24Peri-urbanFemaleOP4RN18UrbanFemaleOP5RN36UrbanFemaleOP6SC5Peri-urbanFemaleOP7SC9UrbanFemaleOP8SC4Peri-urbanFemaleOP9SC3UrbanFemaleOP10SC5Peri-urbanFemaleTC1PEM14TMFemaleTC2PEM3TMMaleTC3PEM3TMFemaleTC4PEM18TMFemaleTC5PEM15TMFemaleTC6PEM21TMFemaleTC7PEM6TMFemalePA1PEM24N/AFemalePA2Admin18N/AFemalePA3Admin36N/AMale*Participant role: OP ¼ On-site Provider; TC ¼ Teleconsultant; PA ¼ ProgramAdministrator.*Clinical Role: MO ¼ Medical Officer; RN ¼ Registered Nurse; SC ¼ SeniorConsultant; PEM ¼ Pediatric Emergency Medicine Physician.A.A. Rowther et al.SSM - Qualitative Research in Health 2 (2022) 1001506thoroughly than would be possible for on-site providers.In the emergency room, you feel more stressed out standing in front of onechild's crying mother [while also] trying to apply ‘CPAP’ [continuouspositive airway pressure] to another crying child. [ …] That all cuts offwhen you are sitting here in telemedicine and you are just focused on what'shappening and what needs to be done. (TC6)In another theory of use, TM consultants assume a hierarchical role ofsenior physicians, whose greater expertise enables them to refine on-siteproviders’ diagnostic reasoning and correct their erroneous managementdecisions.The difference between my duties and theirs? I am more responsible thanthem because they see us on a senior level. They trust us […] They do theimmediate management, while my role is giving that expert opinion. (TC3)Participants from all the end user categories described contextualfactors of TM implementation that were perceived as supporting onetheory of use over the other, either by leveling the field between on-siteand teleconsultants or by stacking the program toward the seniority orsuperior clinical competence of TM consultants (Fig. 1). While someproviders emphasized one as preferred or prevalent over the other, boththeories of use emerged in accounts from across the range of participantroles, with the tenuous balance between these communication structuresemerging as a central theme of the data.4.1.1. Symmetry: Cameras and ContextWhile often acknowledging a general reliance on authority conferredby seniority in the ED, on-site providers drew a distinction between thisand the unique position of TM physicians with their ability to view allpatients and activity in the pediatric resuscitation room by camera(Table 2, Quote 1). This allowed TM physicians to not only offer clinicalinstructions but also provide additional monitoring as “an extra set ofeyes” in the crowded, complicated, and chaotic environment of pediatricEDs, where patients may inadvertently be neglected (Table 2, Quote 2).The TM environment was viewed by teleconsultants and on-site pro-viders alike as relatively calm, less physically demanding, and allowingmore focus in its computer-based practice.Emphasis on the contextual advantages of TM physicians, such as theability to readily access reference manuals, review clinical protocols, orconsult other seniors—rather than on the deficient clinical knowledge orskills of on-site providers—was portrayed by both teleconsultants andprogram administrators as an integral part of teleconsultant training.It's about them [teleconsultants] having the attitude of, “They are in adifferent situation than we are, so they're in the chaos of the moment,Table 2Key interview excerpts on TM theories of use and related communication structures.TopicQuote#TypeaQuotationSymmetry: Cameras and Context1OPIn the ED, we follow the orders of our senior doctors as it is, whereas the difference with telemedicine is that they are covering all the patients, able towatch them and us and the doctors. If, God forbid, a patient is getting neglected, we immediately receive a call from them: “That patient is having fits[convulsions]." [ …] They remind us about our patients. (OP3)2TCOne idea is that the telemedicine physician is an extra set of eyes because, you know, there are some challenges specifically in an emergency setting, likethe volume of patients, the chaos of the environment, the social problems that come with patients into the room. By not having to deal with those, youprovide sort of an extra set of eyes and can tell the doctor if, for example, there's a patient having fits in the other corner of the room, a patient they wouldotherwise not notice or know about. (TC6)3OPSometimes we're busy with patients, right? But it's not like they call us again and again, annoying us when this child is getting ambu bagged or while thatkid needs an ETT [endotracheal tube]. We talk to them when we're free. They check on us to see if we're free and then call us back. (OP4)4TCWe can only see what's on the screen, that's all, so we have to ask the nurse and the doctor. They know and do everything because they are in touch withthe actual patient. (TC2)5TCMost of the corrections I make, most of my interventions as a TM doctor, it's just to get the ground doctor to adhere to existing guidelines. The protocol isthe protocol. It's clear-cut. (TC3)6OPWhile discussing a case with them, they'll ask me, “Oh, did you ask about this? Did you examine that?” And that helps us because we are in such a hurry,with so many patients to look after, and every patient in P1 is very sick. So, we might skip over some important points, and they remind us by askingwhether we've performed this exam or looked for something specific. So, we'll do that and then be able to better manage the patient. (OP2)7OPFor me, sometimes in cases like poisonings, I need to remember the dosages or something but don't have time to recheck. Then I can go back and calltelemedicine and ask them, “Please confirm this dose for me.” […] In terms of protocols and dosages, they can still help even if they are junior. They canhelp because those are just reference manuals or values. (OP7)8PAThey will fill out one per person and indicate whether they were following protocols. This is as per the 'PALS' [Pediatric Advanced Life Support]protocols, and all the fields are PALS-based. So, were they following the ‘PAT’ triangle? Were they following infection control practices? Also, did theycheck the airway? Is the compression rate and depth in CPR fine? Is the ambu bagging and chest rise fine? As much as can be observed, they put it in here,okay?” (PA1)Asymmetry: Experience and Expertise9OPWhen I'm stuck as to what to do next, I call them. They give me advice: “If you haven't done this, then do that, or try this.” [ …] So even though I'm asenior doctor here, I do get stuck sometimes, so if I can contact another senior [then] they can give me expert advice about my patient. (GD3)10TCThe knowledge that I have gained through my clinical experience, that is indicated right now to those at the junior level, to the medical officers and to thestaff. I can correct them, and I can teach them. Like, change your behavior, change your medicine, the drug or dose, change your treatment plan becausethis isn't going to be beneficial for the patient. That is something that you need to be a superior for. (TC5)11TCIf I saw that patient, and he deteriorated and expired, then he is also my responsibility! And I need to think about what it was, like, what did I miss? Whatmore could I have done? So, I feel every patient which I or any TM physician has seen, even once from that screen, then becomes our responsibility, andthat patient is our patient. (TC6)12TCWe also get judged. It's not like we just get seated here and allowed to do what we want. No, there is also surveillance here. What we do and how weoperate is being monitored, we get tested, our performance is being judged, and at the monthly meetings and M&M conference, we are told, you did thisbut you should have done that. (TC7)13PAI think that the people sitting in telemedicine are sometimes junior. So, it's not much of a benefit to call them because I know more than they do. That'ssometimes a drawback. They should be senior doctors in the telemedicine room. (OP7)14PAThe true legacy in the long-term, that is also sustainable, is the impact made possible by TM. And to achieve that vision, we need the best of the best,because their true role her is that of coaching. (PA3)15TCAt these other sites, they don't have as many resources, as many diagnostic tests available, so we can make a difference because we are practicingaccording to experience […] according to what we've studied and our experiences, we know the possible causes and available treatments. (TC7)16TCSometimes something is going on that is not in the protocol […] what happens is that the patient is not relieved and is in continuous distress. In that case,I would not abide by, or I would get away from the protocol, and I would try to save the patient. (TC3)a OP ¼ On-site Provider, TC ¼ Teleconsultant, PA ¼ Program Administrator.A.A. Rowther et al.SSM - Qualitative Research in Health 2 (2022) 1001507where anyone could make that mistake.” We’re not telling them that there'salways a deficiency in the other provider, but rather that something needsto be done for the patient. (PA1)Such an attitude supported a symmetric communication structure inthe teleconsultation process, in which the TM physician is sympathetic tothe contextual constraints of the on-site provider. Based on estimatesfrom both on-site and teleconsultants, the teleconsultants would initiatecalls at least as often as they receive calls for consultation. TM-initiatedcalls include follow-up on prior cases or attempts to draw attention tochildren who, based on their observations by camera or telemetry, mayrequire intervention or investigation. The timing and content of the callsalso depended on availability and priorities of on-site providers, whomay be engaged in patient care and could request the teleconsultant tocall back later (Table 2, Quote 3).In accordance with a symmetric, context-based theory of use, severalteleconsultants described their role as complementary with and depen-dent on that of the on-site provider. This sense of mutual dependence wasreinforced by teleconsultants’ inherent reliance on the reports, physicalexams, procedural skills, and adherence to advice of on-site providers(Table 2, Quote 4). Several on-site providers also expressed indepen-dence and primacy in their clinical decisions, with an option but not anobligation to follow teleconsultant advice.In terms of my clinical decisions, it's like, I'm the primary physician. I cantake their advice, but I'm not compelled to follow that. What I know is goodfor the patient, I can do that. (OP1)Finally, protocolization of care, posited by administrators to underpinimprovements in quality, also supported balance in TM encounters byproviding a mutually accessible point of reference, namely evidence-based guidelines in which all providers were trained (Table 2, Quote 5).While reinforced by such contextual factors, the symmetric theory ofuse was also rooted in and reproduced by some key assumptionsexpressed by both users and administrators about clinicians’ needs andhow teleconsultation improves the quality of pediatric emergency care.The first assumption was that the scope of TM involvement should spanall stages of care provision, beginning with verification of the patienthistory and physical exam findings (Table 2, Quote 6). In end-user ac-counts, this was premised on the contextual constraints of the resusci-tation rooms creating a need for TM to offer overburdened on-siteproviders with supervision at each step of assessment and treatment fornearly every patient, rather than only select advanced consultations.We wear a lot of hats. It's not just, this is the most knowledgeable expertwho is there for the most complicated cases. No, that's not the case. In somecases, it could be something as basic as, did you check for this examfinding?" (TC7)A related assumption voiced by clinicians was that TM monitoring foradherence to clinical protocols, including appropriate timings, dosages,and other reference-based guidelines, was a key opportunity and mech-anism for quality improvement by teleconsultants (Table 2, Quote 7).These assumptions were reproduced by the design of the TM program,with administrators assigning teleconsultants to complete multipleassessment forms regarding on-site provider performance in comparisonto existing clinical guidelines (Table 2, Quote 8). Such requirements foradditional documentation were described as a major adaptation inclinician routine to the TM quality improvement strategy, one that in turndrew teleconsultants’ attention to monitoring and verification of protocoladherence.4.1.2. Asymmetry: experience and expertiseWhile acknowledging the numerous contextual factors and features ofimplementation that support a symmetric TM communication structure,many clinicians, both junior and senior, gave accounts of a contrastingbut concurrent TM theory of use based on asymmetry in clinicalcompetence and hierarchy in authority. Teleconsultants were portrayedin these instances as providing crucial expert advice for severe or com-plex cases that otherwise exceeded the limited knowledge of on-siteproviders (Table 2, Quote 9). The teleconsultant's ability to providesuch advanced clinical support was described as both a requisite andresult of her very unique position on the medical team; nearly all par-ticipants endorsed a notion that teleconsultants should ideally be “se-nior” clinicians, but some also suggested that the depth and breadth ofpatient exposure in TM shifts provided exceptional experience that inturn made them more qualified as experts.Being senior makes the process easier […] but also, for example, if I wouldhave managed patients in just one hospital or ICU, I might have managedsix or seven patients with diabetic ketoacidosis in the whole year. But hereon telemedicine, there are seven hospitals where I’m seeing patients, right?So, in a month, I might see just as many cases, so it adds up to a lot ofexperience. (TC6)Greater experience and expertise were thus described as inherent tothe teleconsultants’ role, and the perceived superiority of their skills andknowledge, whether explicitly acknowledged or indirectly demon-strated, enabled them to not only monitor and advise but also correct andinstruct on-site providers (Table 2, Quote 10).Apart from its consonance with traditional norms of clinical authorityat Pakistani public teaching hospitals, an asymmetric TM theory of usewas reinforced by alignment with both the internal and externalaccountability felt by teleconsultants for ensuring quality of care. Inter-nally, teleconsultants described feelings of anxiety and duty in relation tooutcomes of patients referred to TM (Table 2, Quote 11). Externally, theyfelt a burden of answerability for the quality of care.The downside is that we are given more responsibility in our duties than thebedside physician. Because when I give a second opinion, then I’m tellingthem to do this, and oftentimes that’s what they are doing. The wholeresponsibility is on me then, not them. If somebody asks for answers, theycan say, I asked her and she told me this. So, I have more pressure on me.(TC5)Such a hierarchy of accountability was reinforced by the design of theTM program, which included continuous monitoring and evaluation notonly of on-site providers by teleconsultants by way of doctor performancechecklists, but also of teleconsultants by program administrators(Table 2, Quote 12). The primary forms of teleconsultant assessmentwere by retrospective review of TM aggregation site recordings and bymorbidity and mortality conferences examining provider notes andmissed opportunities for TM intervention.One clear consequence of the asymmetric theory of use on TMcommunication patterns was a conviction among most teleconsultantsand on-site providers alike that, despite their roles being mutuallydependent, teaching and benefit was unidirectional: “We correct them,and they learn from us, that's the main thing” (TC3). Moreover, thebenefit of doctors or nurses consulting TM was often portrayed ascontingent on teleconsultants' level of experience (Table 2, Quote 13).On-site providers thus gave accounts of being selective in engaging TMbased on seniority.I use teleconsultation for other patients when there is no one senior presentin our emergency department. I would consult the person most senior, who Ithink is the telemedicine doctor. And that’s usually at night […] Based onthe way they’re talking on the phone, talking about the case, I just infer ifthey are [senior]. (OP2)Due to uncertainty about the exact identity or background of theproviders in TM, several on-site providers acknowledged trying to inferteleconsultants’ levels of seniority based on interactions and, in turn,weighing their advice accordingly. The asymmetric theory of use thusadded a layer of investigation and negotiation to TM communicationpatterns, beyond the mere exchange of information and advice.The influence of seniority in reconstituting such patterns ofA.A. Rowther et al.SSM - Qualitative Research in Health 2 (2022) 1001508asymmetry in the TM process was based on a few central assumptionsabout TM user needs. Among administrators responsible for programdesign and implementation, one such assumption was that sustainable,long-term improvements to pediatric emergency care quality wouldrequire teleconsultation to serve as a teaching tool, which in turnnecessitated that teleconsultants be seen as above on-site providers inclinical skill and authority (Table 2, Quote 14). Another relatedassumption among the teleconsultants was that resource constraints atremote sites, such as limited diagnostic imaging or labs, further height-ened the need for teleconsultants to provide the expert knowledge andexperience needed to ensure correct diagnostic reasoning and treatmentdecisions (Table 2, Quote 15). In contrast, on-site providers placed lessemphasis on limitations in human and facility resources at remote sites,instead suggesting that managing the complications that occur despitecorrect diagnosis and initial treatment was the basis of TM's most sig-nificant impact on quality of care.What aspect of patient care is most improved? It’s not diagnostic reasoning.The diagnosis we can make, but it’s managing the secondary issues. Like, ifthe child goes into a complication, or for example the ECG pattern isabnormal and I’m stuck like, what is happening? So that is where theycome to our help. (OP6)Shared across clinician accounts was a notion of standardized pro-tocol adherence as necessary but at times not sufficient for closing gaps inquality of care and improving outcomes (Table 2, Quote 16). Thesediffering assumptions about TM user needs and mechanisms of impactalso converged in their result of inserting greater asymmetry in howclinicians approached TM in routine practice, forming a counterpoise tothe more symmetric, context-based TM theory of use.4.2. Enabling and constraining properties of teleconsultation4.2.1. Confidence/trustThe persistent importance of teleconsultant seniority for many endusers was explained by some in terms of confidence or ability to rely onTM support (Table 3, Quote 1). In either theory of use, the degree ofconfidence and trust—in the expertise of the teleconsultant, in the reli-ability of on-site providers’ physical exams and reports, or in theappropriateness of treatment protocols to individual patients or facilityresources—shaped and constrained the TM process and its perceivedutility for pediatric emergency care in Pakistan. Based on clinician ac-counts, providers at either end of TM consultation must negotiate anddetermine confidence in the process at several overlapping levels.The extent to which on-site providers trust the teleconsultant wasidentified by nearly all participants as crucial for clinician engagementand perceived utility of the program. Requisites for trust included notonly perceived competence but also a communicated sense of partnership(Table 3, Quote 2). In concurrence with the symmetric theory of use,many on-site providers insisted that this can and should include pointingout mistakes (Table 3, Quote 3), while the teleconsultants explained thatreceptiveness to correction required that they avoid sounding harsh orjudgmental.You can only help them save the child once they trust you, and once theyknow that we are there to help them and not there to judge them […] Attimes, they are resistant because they feel we are judging them. (OP8)Thus, more evaluative components of TM—seen by administrators asessential to long-term improvement of quality—posed risks to aconstructiveteleconsultant-providerrelationshipifimplementeddemandingly, according to teleconsultants, many of whom confirmedfeeling frustrated with frequent cases of seemingly uninterested or un-cooperative on-site providers.You obviously feel hurt because your job is to help them […] So when youget ignored or the doctor is not cooperative, it’s frustrating because you arenot there and can’t examine or treat the patient. (TC1)Teleconsultants also held reservations about the degree to which on-site provider reports could be relied upon (Table 3, Quote 4). Strategiesfor overcoming such uncertainty included asking on-site physicians torepeat exam maneuvers as teleconsultants watched by camera or exer-cising greater caution by ordering treatments to be implemented incre-mentally, thereby supplementing their diagnosis using a test oftreatment. A recommendation frequently voiced by administrators, tel-econsultants, and on-site providers for improving mutual confidence andsense of partnership among clinicians was to create greater opportunityfor on-site providers to become acquainted and familiar with physiciansstaffing the TM aggregation site (Table 3, Quote 5). Rationales for thisstrategy included increasing awareness of teleconsultants’ backgroundssuch as their level of training and experience, establishing a moreTable 3Key interview excerpts about enabling and constraining contextual factors of TMin pediatric emergency care: Confidence/Trust and Distance.TopicQuote#TypeaQuotationConfidence/Trust1OPI do think the most important thing, the biggest factorin terms of my level of trust in the telemedicine doctor,is their level of seniority. (OP8)2PAThere has to be an element of mutual respect, anelement of mutual understanding […] It's not a giventhat you will have a good relationship with the on-sitephysicians, but I think the important thing is yourattitude. That helps build comradery. (PA1)3OPSome of them question my judgement or knowledge[…] But I don't think about it like that. For me, thepatient and the quality of care is more important thanfeeling embarrassed. (OP1)4TCI would say that most of the time they have missedsome part of the physical exam. You can say, like,seventy or eighty percent of the time the misssomething basic […] I think they are trained to do it,but they miss it. (TC4)5PAMany providers have suggested this idea where maybemonthly or bimonthly there should be a meeting whereone or two TM physicians can visit the ground site andmeet the whole team. […] Seeing each other face-to-face makes a big difference. (PA1)6TCThere is not a hundred percent compliance. They'resort of following, you know, whatever the commonknowledge that's passed on to them is, and they areresistant to change. (TC6)Distance7TCIf you do completely just telemedicine, and you're notin touch with the ER or in touch with the clinic side, itwould become difficult for you to understand theiremotions and feelings, and the panic they're sufferingfrom. And you'll be out of habit from being away. (TC3)8TCThat is why we have rotations! It keeps us in touch withpatients and our clinical skills […] Because we do getthis feeling or need for meaning and motivation. Untilwe touch the patient, or until we examine the patientourselves, we won't have that internal satisfaction ofhaving seen and cared for the patient (TC7)9TCAs a woman, I have to anticipate that, like, domesticresponsibilities are something that can make workingin a place like the emergency department difficult, verydifficult. And telemedicine offers an opportunity to stillcontinue practicing medicine when other femalephysicians might be forced to exit the workforce. (TC6)10OPI feel reassured by the fact that there's someonewatching. Not only that they are watching but that wereceive their call and advice […] We feel verysupported by telemedicine being there. (OP3)11PATelemedicine is a quick, efficient way of getting newprotocols to the rest of the public hospitals andproviding the data […] In the public sector right now,there's not any surveillance. No surveillance, noresearch, nothing! But we can provide them. We caneven see the changes in diseases patterns (PA2)a OP ¼ On-site Provider, TC ¼ Teleconsultant, PA ¼ Program Administrator.A.A. Rowther et al.SSM - Qualitative Research in Health 2 (2022) 1001509personal connection between clinicians by reducing the anonymity ofteleconsultants, and finally improving receptivity to teleconsultantintervention or advice by creating a greater sense of alliance andcomradery.At a level more proximate than inter-provider interactions mediatedby TM between sites, teleconsultation was also shaped by clinicians’confidence in themselves and their own abilities. For teleconsultants, thisincluded concerns about retention of clinical skills despite distance fromdirect patient care, motivating several to continue in-person shifts whiledoing TM duties.I don’t only do telemedicine shifts […] Like in a week, I would do threedays here, and the remaining four days, I would do my clinics. In this way,I stay in touch and know the beside practice as well. (TC3)Apart from losing clinical skills, teleconsultants expressed anxious-ness about their ability to remain vigilant yet calm while having tomonitor up to twelve hospital sites simultaneously. At a more distal level,teleconsultants had limited confidence in health system functioning atremote sites, including concerns about the unavailability of diagnostictests and tools, use of outdated local treatment protocols (Table 3, Quote6), and limited opportunity to meaningfully improve outcomes due todelays in patient presentation.Among on-site providers, for whom the proximate level of confidencewas in themselves and the remote health facilities where they worked,the most commonly mentioned doubts were in their ability to manage thehigh volume of patients who would present. Overcrowding of both urbanand peri-urban hospitals limited on-site providers perceived ability to notonly provide timely patient treatment and monitoring but also initiateand participate in TM consultations.Because we have so many patients […] Mainly, it’s the flow of patients thatmakes telemedicine hard to use at times. That can make it harder to usetheir help or communicate with them. [OP2]At a more distal level, on-site providers conveyed a high level ofconfidence in teleconsultants’ abilities to accurately evaluate patientsremotely, with technical functioning of TM equipment appearing lessoften as a concern than the level of experience of the teleconsultant.4.2.2. DistanceWhile respondents of all participant roles, especially administrators,endorsed the TM program's advantages with respect to overcomingspatial separation of clinicians and patients over the hundreds of kilo-meters between facilities and accompanying delays in care provision,alternative conceptualizations of distance also emerged from clinicianaccounts of TM. The first was related to practical distance between theteleconsultant and hands-on clinical medicine, a distance that could limither skills and ability to connect with on-site providers (Table 3, Quote 7).The major strategy used by teleconsultants to maintain clinical knowl-edge and motivation derived from providing direct patient care wassimply to maintain in-person shifts (Table 4, Quote 8). The second majorconceptualization of distance was a social, gendered distance faced bywomen and overcome in part by working as teleconsultants.A second motivation was the feeling of treating patients in remote sites inareas far away where I, being female and being in the cultural norms ofPakistan, would not be allowed to travel or go to even for my job or treatingpatients. (TC6)In addition to overcoming felt constraints against traveling alone todistant or remote hospitals, female teleconsultants identified TM as anopportunity to practice medicine and serve the mostvulnerable patients despite gender-specific barriers to working in theED such as domestic duties after marriage or the social and physicalburdens of pregnancy and motherhood (Table 3, Quote 9). Finally, thereexisted a notion of mental and emotional distance, as discussed under thetheme Symmetry: Cameras and Context, which enabled teleconsultants toexercise greater focus by being physically removed from the chaos andclinical uncertainty of the resuscitation room.Closely related to the theme of confidence, distance emerged as amajor factor connecting the TM program to strengthening of Pakistan'shealth system and workforce. With respect to the distance betweenproviders, virtual proximity of the teleconsultant enabled by videocon-ferencing technology allowed them to reach health workers in under-served areas who may otherwise have little support, low morale, andhigh rates of turnover.By providing this extra support, they feel more comfortable and supported,which helps retain health workers […] For the doctors there, the differenceis that they feel, we are not alone, we have support. (TC7)On-site providers almost universally endorsed feeling supported or atleast comforted by the virtual presence of teleconsultants, even whenTable 4Key interview excerpts about enabling and constraining contextual factors of TMin pediatric emergency care: Information, Technological Artifacts, and MedicalPractice.TopicQuote#TypeaQuotationInformation1OPTheir ability to evaluate patients by camera is amazing.They can see the size of the pupil of the child who isbeing discussed, and they can see every detail. If I'mstuck with an ECG, like what is it or what to do […]then they can see the ECG and then tell me. (OP5)2TCAt times, a subtle examination finding or a singleexamination finding becomes very important,especially when we are differentiating something likeseptic shock and cardiogenic shock. […] the doctortells us, we have given a bolus and now the child ishaving creps, but what I'm seeing from here, on myscreen, makes me feel the child is dehydrated. He hassinking eyes, and he has tachycardia, so he needs morefluid. At times, it gets very challenging to decidewhether to follow what I'm seeing and feel like doing orto follow what they are telling me. (TC6)3TCThe options on the form are either yes, no, or notapplicable. So. Like, if someone had not given me thisform and asked me, you observed that CPR so howwould you rate it, that would be very subjective. Butnow, this because very objective performance data.(PA1)Technological Artifacts4TCThe idea is to put these cameras all over the country.This is more economical. This is more feasible. And wecan have more nurses, but we can't have that manymore doctors […] What we are doing is nothinginnovative. We are just making use of existingtechnologies. The cameras and videoconferencingtechnology is easy. It's the human resources that arechallenging. (PA3)5OPTelemedicine is very easy for us. It's just a matter ofcalling them. That's it, just pressing three buttons onthe phone. (OP4)Medical Practice6OPWe are omitted from the counseling part of patientcare, and that is a very important part of management.The on-site physician is undertaking that process alsowhile we watch through cameras, in addition to themdoing all the physical things and managing everything[ …] Only when they communicate with us do we cometo know. Visually we can still see that patient, but wecannot ask any questions of the parent like, what is thechild's history or what exactly happened to this child.(TC6)7TCLike, the doctor's counseling, the small touch on theparent's hand, everything, this all helps a lot. So, wecan see those things a little, but I think the motivationyou feel is less, you can say. The motivation is a littleless for us in telemedicine (TC4)a OP ¼ On-site Provider, TC ¼ Teleconsultant, PA ¼ Program Administrator.A.A. Rowther et al.SSM - Qualitative Research in Health 2 (2022) 10015010acknowledging their supervisory or evaluative role (Table 3, Quote 10).On the other hand, with respect to the distance between teleconsultantsand patients, remoteness of the teleconsultant was portrayed asproviding a broader view of population health and epidemiologic trendsto inform practices and protocols, particularly in the absence of publicsurveillance systems (Table 3, Quote 11). Clinicians thus portrayed TM ascreating distance between teleconsultants and the practice setting yetsimultaneously facilitating proximity, socially or symbolically as well aspractically, in their routines and relationships with consulting providersand patients.4.2.3. InformationInformation or data was another factor that served as both anenabling and constraining property of teleconsultation in the pediatricED according to clinicians. At the level of patient data, TM and on-siteproviders both expressed high levels of confidence in what they consid-ered ‘objective’ processes, most notably what senior teleconsultantscould glean from properly placed telemetry probes and camera-mediatedobservation (Table 4, Quote 1). The latter was often discussed as reliablemeans of confirming or correcting clinical assessments, with high videoresolution seen as central to the utility of TM support among respondentsof all participant roles.Teleconsultants and on-site providers differed, however, in theirviews of which other sources or types of information should be consid-ered subjective and by extension less reliable. For teleconsultants, inac-curacies and subjective elements of on-site providers’ reports regardingpatient history and physical exam findings represented key limitations intheir ability to provide effective support (Table 4, Quote 2). Reasonsoffered by teleconsultants for their suspecting inaccuracies or omissionsin on-site provider reports included a lack of experience appreciatingsubtle physical exam findings, lapses in assessment due to forgetting orbeing too busy, efforts to fit information into preconceived diagnoses,and most importantly, an alleged resistance to the questioning or inter-ference of teleconsultants.Sometimes the on-site doctors are not cooperative. Like, they don’t like to,they’re reluctant to talk to us […] When we face a doctor who is reluctantor who is not cooperating, we try to speak politely […] If we ask themsomething in a harsh way, obviously they will also answer curtly. (TC2)Recommendationsforimprovinginformationqualityincludedworkshops on exam maneuvers for on-site providers and communicationtraining to standardize data reporting protocols and reduce the percep-tion of teleconsultants as impolite or harsh. On-site providers, in contrast,expressed the greatest degree of uncertainty in the evaluative data pro-duced at the provider and facility level based on teleconsultants'completion of performance checklists. Many admitted to having little tono familiarity with or exposure to this data, while some others suggestedthe performance data were simplistic and not reflective of contextualchallenges on the ground. In contrast, program administrators and tele-consultants emphasized the assessment forms’ simplicity in arguing fortheir relative objectivity and value (Table 4, Quote 3).4.2.4. Technological artifacts and medical practiceThe final subthemes that emerged from clinician accounts of enablingand constraining factors were the technological artifacts involved in TMimplementation and their intersection with the reconfiguring of medicalpractice. With respect to technology, teleconsultants expressed a strongdesire to expand the range of devices used to transmit data to their ag-gregation site—digital fundoscopic images, electronic stethoscopes, orpoint-of-care ultrasounds—but these were viewed as limited by the lackof sufficient expertise and facility resources for implementation at remotesites. Dissemination of technological artifacts related to TM was none-theless considered more efficient and feasible than increasing the numberof trained physicians in remote regions (Table 4, Quote 4). Among tele-consultants, there was consensus on the relative simplicity and ease ofuse of TM technology after an initial phase of familiarization and trainingwith the tools, namely the computer systems, monitor display options,and remote controls for the cameras. The primary challenge posed bytechnological artifacts was related to time, namely delays attributed toemploying them.When we focus the cameras on the patient, we have to zoom using thebuttons on the remote, so this can take some time, to focus on them. And thedoctors and consultants are calling us, but it takes us time to put the patienton the screen and zoom. So, that delay can take up time. (TC2)While on-site providers were largely in agreement about TM'saccessibility and convenience (Table 4, Quote 5), this hinged on their nothaving to view a monitor or position the camera, whose movements froma central location in the resuscitation room were exclusively controlledby the teleconsultant. Although connectivity problems and interruptionswere described as relatively frequent, occurring as often as once or twicea week, this was considered by clinicians to be routine, quickly resolvedby technical support, and not a critical barrier to use.While the technological artifacts of TM were considered essential tothe teleconsultants' virtual presence at the bedside, their detachmentfrom key aspects of the more traditional clinical encounter—includingnon-physical interactions such as eliciting details of the patient history orcounseling the child's family—were viewed as important reconfigura-tions of medical practice and resulting from constraints and opportunitiestied to TM technology's design and application (Table 4, Quote 6). Evenamong teleconsultants who described themselves as more accountablefor quality of care, their perceptions about the nature and relative diffi-culty of their role as teleconsultants was shaped by having the bird's-eyeview—literal and metaphorical—of patient care and being separatedfrom the words and bodies of patients.They [on-site providers] are the ones that are more challenged, with abigger duty than us, because they are telling us, giving us the story of thatpatient, touching the patient, talking to the attendants, and we are taking abroader view of that same patient. (TC5)This opportunity to directly witness yet not physically intervene inpatient care was recognized by both on-site providers and teleconsultantsas unique to the teleconsultant as compared to an in-person or by-phoneconsultant. Teleconsultants expressed particular frustration with theinability to rescue providers facing procedural complications while sta-bilizing the patient, such as difficult intravenous or interosseus lines andfailed intubations. The reconfiguration of medical practice by TM tech-nology was an important element of clinician experience, causing a fewof the teleconsultants to admit feeling less motivation or meaning in theirwork (Table 4, Quote 7).5. DiscussionTo our knowledge, this study represents the first in-depth qualitativeinvestigation of how features and implementation of a pediatric emer-gency teleconsultation program are experienced by end-users andstakeholders in a LMIC setting. We found that, behind the technicaladaptation to using videoconferencing technology in the pediatricresuscitation room, providers perceived a dynamic reconfiguration ofprofessional roles and routines of consultation according to various op-portunities and constraints they associated with TM. Descriptions ofcommunication patterns enacted by on-site providers and remote tele-consultants betrayed an unstable tension between two competing the-ories of use premised on either inherent contextual differences orpresumed asymmetries in expertise between clinicians at each end(Table 5). The long-term sustainability of TM applications for improvingaccess to and quality of pediatric emergency care requires seriousconsideration of not only economic costs of dissemination, maintenance,and training for TM technologies but also how such conflicting theoriesof use and their associated assumptions about end-user needs crystallizeA.A. Rowther et al.SSM - Qualitative Research in Health 2 (2022) 10015011in practice and affect perceived utility and clinician engagement of TM inthe future.Prior qualitative and mixed-methods research on clinician attitudestoward using pediatric emergency teleconsultation in high-incomecountries have found lack of perceived benefit and poor integration inexisting work routines to consistently be among the most prominentbarriers to appropriation among clinicians (Haimi et al., 2018; Kim et al.,2019; Ray et al., 2017; Sauers-Ford et al., 2019; Uscher-Pines & Kahn,2014). The present study furthers our understanding of such barriers bydemonstrating how the formation of work routines and mutual confi-dence for a novel pediatric emergency teleconsultation program inPakistan is contested and negotiated by providers at multiple levels.According to our findings, clinicians and administrators who viewedteleconsultation as beneficial nonetheless held widely divergent ideasabout how TM improves quality of care, the appropriate scope of TMinvolvement, which patient cases should be referred, and where au-thority and accountability falls, all differences rooted in competing the-ories of use. Such misalignments in the scripts embedded in TM designversus actual practice may cause the work necessary to integrate TM intoclinical routines to exceed perceived benefits and contribute to programfailure (Nicolini, 2006). Recommendations for improving the perceivedbenefits and acceptance of TM among clinicians from prior studies inhigh-income settings include increasing awareness of evidence forimproved outcomes with TM (Mueller et al., 2014), legitimation of TMthrough its representation as usual care or routine in protocols (Ray et al.,2017), and strengthening the relationships between providers at eitherend of the TM interface (Wade et al., 2014). The last of these recom-mendations was strongly echoed by participants of this study, whoendorsed a common desire for greater familiarity and improved rapportbetween providers, in particular the teleconsultants at the aggregationsite in Karachi and the local providers at peri-urban or remote hospitalsites in broader Sindh. Our results regarding factors that influence thenegotiation of trust and establishment of mutual provider confidence inthe teleconsultation process extends and expands upon prior researchindicating the centrality of trust in the teleconsultation process in theoryand practice (Lehoux et al., 2000).Our findings regarding the tension between asymmetric and sym-metric theories of use among end-users are consistent with prior researchemploying social or processual views of TM implementation. Forexample, Sicotte and Lehoux's (2003) detailed case studies of a Canadianteleconsultation system suggest that the benefits of TM are paradoxicallypremised on expertise asymmetry as its rationale and yet, due to theremoteness of the consultant, have a strong reliance on credible infor-mation from on-site providers, in turn tending to demand more narrowexpertise gaps or high quality information technology to compensate forlarger deficits (Sicotte & Lehoux, 2003). Similarly, our results point toteleconsultant seniority—a proxy for expertise asymmetry—as a centralfactor for clinician acceptance of a pediatric emergency TM program inPakistan. This is despite broad acknowledgement among participantsthat potential benefits of TM in overcrowded pediatric EDs of Pakistanigovernment hospitals extend far beyond providing advanced diagnosticsupport for complex cases, which is itself constrained by teleconsultants'distance and dependence on local provider reports of uncertainreliability. We find that, for such high-acuity clinical contexts andresource-limited settings as this, the relative importance of patientevaluations such as physical examination and patient history serves toamplify TM's sensitivity to previously observed physician anxieties aboutTM relying on information from a human intermediary and thus framedas less objective (Lehoux et al., 2002). It may also amplify the necessityfor trust between attending clinician and remote specialist in overcomingbarriers to teleconsultation in clinical practice (Lehoux et al., 2000).This study was limited to attitudes and perspectives of urban or peri-urban providers in Pakistan during a relatively early phase of pediatricemergency teleconsultation implementation. Although this sampleincluded participants with up to three years of exposure to TM beginningfrom the program's inception, our findings regarding conflicting theoriesof use may not be fully transferrable to more mature or long-standing TMapplications in pediatric emergency care or those with more specificscripting or standards for virtual presence (Foster et al., 2020). Moreover,the contextual factors and various opportunities or constraints of TMacknowledged in clinician narratives of teleconsultation may hold moreor less relevance for providers in more rural settings or those outside ofPakistan and South Asia. For this study setting however, the inclusion ofon-site providers and teleconsultants allowed for triangulation of data tostrengthen the rigor of our analysis, while purposive sampling to captureperspectives of providers with maximum variation across reported pro-gram fidelity and acceptability helped enhanced the completeness of ourinvestigation. Participants were predominantly women, which whileconsistent with the overall gender distribution of pediatric emergencycare providers at CLF and affiliated hospital sites may have influencedthe particular themes which emerged and reached saturation in thisstudy, and its unknown how results may have differed with inclusion ofmore men. Lastly, our chosen conceptual model for informing the designof the interview guide and thematic analysis placed emphasis on pro-fessional routines and clinical communication with less focus on medi-colegal issues of liability, financial issues of reimbursement, or regulatoryissues such as licensing or credentialing. While a few participants choseto discuss perspectives related to credentialing and salary structure forteleconsultants, these potentially sensitive topics were only probedfurther when raised by the participant and did not ultimately reachsaturation in the data analysis.A strength of this study is its application of a conceptual frameworkgrounded in social science theory to understand the complex, dynamicprocess of TM's organizational appropriation and individual integrationinto clinical routines of pediatric emergency care, an approach whichhelps capture the social as well as technical properties of teleconsultationfrom clinicians' perspectives and experiences (DeSanctis & Poole, 1994;Giddens, 1984; Hagendijk, 1990; Lehoux et al., 2002; Prasad, 1993).Moreover, our study on a novel application of teleconsultation inPakistan provides further evidence supporting the importance of exam-ining both technical and social aspects of TM implementation in evalu-ative as well as formative research (May et al., 2003). For example,clinician narratives in this study revealed social and symbolic construc-tions of the ‘distance’ introduced by technological artifacts in TM,consistent with prior analyses of how TM reorders rather than erasesspatial dimensions of healthcare, with implications related to providerconfidence, accountability, motivation, and gender (Oudshoorn, 2012).While TM's impact on women's access to care in South Asia has beeninvestigated previously (Parajuli & Doneys, 2017), our findings suggestthat TM's ability to circumvent gendered travel restrictions is also rele-vant for women physicians. This points to the potential for TM to reducegender-based barriers against Pakistani women's continued employmentin medical practice after completion of undergraduate medical educa-tion, a phenomenon well-documented in existing literature but virtuallyunexplored in relation to TM (Gupta et al., 2020; Masood, 2019; Mohsin& Syed, 2020; Mumtaz et al., 2018). Our results add to the growingrecognition of TM as a means of affording control and flexibility neces-sary to advance professional equity for female physicians (Nadkarni &Mittal, 2021), who experience higher rates of burnout and bear a greaterTable 5Assumptions regarding teleconsultation as quality improvement and clinicians asend users underlying each TM theory of use.Symmetric, Context-Based TM Theoryof UseAsymmetric, Seniority-Based TM Theoryof UseComprehensive coverage byteleconsultantsFocused advising by teleconsultantsCertainty of definitive diagnosis andevidence-based treatmentUncertainty of undifferentiated patientsand clinical complicationsProtocol adherence at all sitesProtocol limitations at remote facilitiesTM as quality assurance for patient careTM as teaching tool for provider capacitybuildingA.A. Rowther et al.SSM - Qualitative Research in Health 2 (2022) 10015012share of familial responsibility (Brubaker, 2020; Ly & Jena, 2018).Another strength of the study was its use of iterative thematic analysis tocapture inductive themes from the interviews such as the aforementionedgendered conceptualization of distance and emergent uses of TM that gobeyond aiding in diagnostic problem solving or decision-making forclinical care (Oudshoorn, 2012). For example, in many participant ac-counts, TM also functioned as a potential tool for conducting epidemio-logic surveillance in low-resource contexts, for providing reassuranceand supporting the morale of health workers in remote areas, or forfacilitating interactive medical education for on-site providers. The use ofTM as a point-of-care education tool for pediatric emergency care pro-viders introduces potential complications in assessing long-term efficacyof the technology, since successful application for knowledge transferand capacity building would presumably lead to diminished utility overtime, a challenge for TM evaluations discussed previously in the litera-ture (Sicotte & Lehoux, 2003).In summary, this study has explored key clinician perspectives, the-ories of use, and communication patterns emerging from the appropria-tionofapediatricemergencyteleconsultationprogramandsociotechnical integration into routine practice within PEDs of Pakistanigovernment hospitals. Respondents described two theories of use andassociated communication structures that differ in their positions along acontinuum of relational asymmetry: one model premised on hierarchicalauthority and expertise of teleconsultants, and another model focused onovercoming contextual challenges faced by on-site providers (e.g.,overcrowding, frequent interruptions, patient acuity, etc.) through TM-mediated partnership. In either theory of use, clinician engagementwas portrayed as constrained by the degree of mutual confidence ortrust—primarily in the expertise of the teleconsultant and in the reli-ability of the on-site providers' physical exams and reports—establishedacross intervening distances, technological artifacts, and the dynamicreconfiguration of professional roles and practices conventional tomedical consultations by TM. For teleconsultants, who were predomi-nantly women in the program under study, an important emergent socialopportunity and function of TM was transcending proscriptive normsagainst women's travel and other gender-based barriers to womenpracticing PEM in underserved areas of Pakistan.The communication patterns described by clinicians as resulting fromthe social and technical adaptation to teleconsultation in the pediatricresuscitation room revealed a tenuous balance between symmetric versusasymmetric theories of use that were not predictably or neatly distributedby clinical role. This balance includes invisible work by providers tomake the necessary organizational, interactional, and practical changesrequired to then begin to stabilize what are arguably experimental ap-plications of TM in practice (Trupia et al., 2021). Attention to suchformative processes is crucial to understanding how and why TM systemsfail so often in terms of normalization toward routinized embedding inclinical practice (May et al., 2003). This will require additionalprocess-orientedqualitativeevaluationsofhowTMsystemsaredesigned, developed, implemented, and evaluated by users, a researchagenda that is distinct but also complementary to ongoing efforts toevaluate the efficacy and cost-effectiveness of digital health interventions(World Health Organization, 2016).Given the paucity of literature on thehuman dimensions that make TM implementation feasible and successful(May & Ellis, 2001; Whitten et al., 2000), realizing its full potential forimproving quality and access to care will require further studies toadvance our understanding of factors influencing clinician engagement,emergent TM theories of use, and how deployment of TM technologyreshapes routines and relationships among providers as well as betweenpatients and providers.Ethical approvalWe obtained written informed consent from all participants. To pro-tect confidentiality, we use code names and remove or alter identifyingdetails about participants or clinical encounters when presenting find-ings. Ethical approval for this research was obtained from the Institu-tional Review Board (IRB) of the Johns Hopkins Bloomberg School ofPublic Health in Baltimore, Maryland, and from the IRB of Dow Uni-versity of Health Sciences in Karachi, Pakistan.Declaration of competing interestThe authors declare that they have no known competing financialinterests or personal relationships that could have appeared to influencethe work reported in this paper.Appendix 1. Emergency Department Physician/Nurse Survey (English)Strongly AgreeDisagreeNeutralAgreeStrongly AgreeN/AI had a meaningful clinical encounter using telemedicine support.The equipment was easy to use by providers for consultation.The process of contacting and receiving telemedicine support was acceptable.I felt I was understood by the health care provider providing telemedicine support.I felt comfortable receiving the information through telemedicine.The consult was better performed through telemedicine than over the phone.Telemedicine helped to provide better care for my patient.Telemedicine facilitated earlier transfer of my patient to receive definitive care.Telemedicine support improved my confidence, knowledge, and skills to care for my patient.Telemedicine improved access to specialized care.I am satisfied with the support of the telemedicine provider in managing my patient.ReferencesAftab, W., Shipton, L., Rabbani, F., Sangrasi, K., Perveen, S., Zahidie, A., et al. (2018).Exploring health care seeking knowledge, perceptions and practices for childhooddiarrhea and pneumonia and their context in a rural Pakistani community. BMCHealth Services Research, 18, 1–10.Ahmed, F., & Nisar, N. (2010). Public-private partnership scenario in the health caresystem of Pakistan. EMHJ-Eastern Mediterranean Health Journal, 16(8), 910–912.Akrich, M. (1992). The de-scription of technical objects. MIT press.Akrich, M. (1995). User representations: Practices, methods and sociology. Pinter.Argyris, C., & Schon, D. A. (1974). Theory in practice: Increasing professional effectiveness.Jossey-bass.Baker, T. (2009). Pediatric emergency and critical care in low-income countries. PediatricAnesthesia, 19, 23–27.Barley, S. R. (1986). Technology as an occasion for structuring: Evidence fromobservations of CT scanners and the social order of radiology departments.Administrative Science Quarterly, 78–108.Bashshur, R. L. (1997). Critical issues in telemedicine. Telemedicine Journal, 3, 113–126.Bhutta, Z. A., Hafeez, A., Rizvi, A., Ali, N., Khan, A., Ahmad, F., et al. (2013).Reproductive, maternal, newborn, and child health in Pakistan: Challenges andopportunities. The Lancet, 381, 2207–2218.Bickler, S., & Rode, H. (2002). Surgical services for children in developing countries.Bulletin of the World Health Organization, 80, 829–835.A.A. Rowther et al.SSM - Qualitative Research in Health 2 (2022) 10015013Boutayeb, A. (2006). The double burden of communicable and non-communicablediseases in developing countries. Transactions of the Royal Society of Tropical Medicine& Hygiene, 100, 191–199.Bregstein, J., Fenster, D., & Lame, M. (2021). Pediatric emergency telemedicine: Anuntapped resource for vulnerable populations during COVID-19. Pediatric EmergencyCare, 37(4), 230–231.Brubaker, L. (2020). Women physicians and the COVID-19 pandemic. JAMA, 324,835–836.Bryce, J., El Arifeen, S., Pariyo, G., Lanata, C. F., Gwatkin, D., Habicht, J.-P., et al. (2003).Reducing child mortality: Can public health deliver? The Lancet, 362, 159–164.CIA. (2020). Central intelligence agency world factbook South Asia: Pakistan. The worldfactbook. Central Intelligence Agency.Closser, S. (2010). Chasing polio in Pakistan: Why the world's largest public health initiativemay fail. Vanderbilt University Press.Craig, J., & Petterson, V. (2005). Introduction to the practice of telemedicine. Journal ofTelemedicine and Telecare, 11, 3–9.Deldar, K., Bahaadinbeigy, K., & Tara, S. M. (2016). Teleconsultation and clinical decisionmaking: A systematic review. Acta Informatica Medica, 24, 286–292.Denis, J. L., & Lehoux, P. (2013). Organizational theories. Knowledge Translation in HealthCare, 308–319.DeSanctis, G., & Poole, M. S. (1994). Capturing the complexity in advanced technologyuse: Adaptive structuration theory. Organization Science, 5, 121–147.Dharmar, M., Romano, P. S., Kuppermann, N., Nesbitt, T. S., Cole, S. L., Andrada, E. R.,et al. (2013). Impact of critical care telemedicine consultations on children in ruralemergency departments. Critical Care Medicine, 41, 2388–2395.Dorsey, E. R., & Topol, E. J. (2020). Telemedicine 2020 and the next decade. The Lancet,395, 859.Foster, C. C., Macy, M. L., Simon, N.-J., Stephen, R., Lehnig, K., Bohling, K., et al. (2020).Emergency care connect: Extending pediatric emergency care expertise to general emergencydepartments through telemedicine. Academic Pediatrics.Gattu, R., Teshome, G., & Lichenstein, R. (2016). Telemedicine applications for thepediatric emergency medicine: A review of the current literature. Pediatric EmergencyCare, 32, 123–130.Gherardi, S. (2010). Telemedicine: A practice-based approach to technology. HumanRelations, 63, 501–524.Ghumman, U., & Horney, J. (2016). Characterizing the impact of extreme heat onmortality, Karachi, Pakistan, June 2015. Prehospital and Disaster Medicine, 31, 263.Giddens, A. (1984). The constitution of society: Outline of the theory of structuration. Univ ofCalifornia Press.Gupta, A., Dogar, M. E., Zhai, E. S., Singla, P., Shahid, T., Yildirim, H. N., et al. (2020).Innovative telemedicine approaches in different countries: Opportunity for adoption,leveraging, and scaling-up. Telehealth and Medicine Today, 5(1).Habib, M. I., & Khan, K. M. (2017). Overcrowding and possible solutions for a busypaediatric emergency department. Journal of Pakistan Medical Association, 67,1398–1403.Habib, M. I., & Khan, K. M. A. (2018). Profile and outcomes of critically ill children in alower middle-income country. Emergency Medicine Journal, 35, 52–55.Hagendijk, R. (1990). Structuration theory, constructivism, and scientific change. Theoriesof science in society, 43–66.Haimi, M., Brammli-Greenberg, S., Waisman, Y., & Baron-Epel, O. (2018). Physicians'experiences, attitudes and challenges in a pediatric telemedicine service. PediatricResearch, 84, 650–656.Heath, B., Salerno, R., Hopkins, A., Hertzig, J., & Caputo, M. (2009). Pediatric critical caretelemedicine in rural underserved emergency departments. Pediatric Critical CareMedicine, 10, 588–591.Hollander, J. E., & Carr, B. G. (2020). Virtually perfect? Telemedicine for COVID-19. NewEngland Journal of Medicine, 382, 1679–1681.Hussain, R., Lobo, M. A., Inam, B., Khan, A., Qureshi, A. F., & Marsh, D. (1997).Pneumonia perceptions and management: An ethnographic study in urban squattersettlements of Karachi, Pakistan. Social Science & Medicine, 45, 991–1004.Jafar, T. H., Haaland, B. A., Rahman, A., Razzak, J. A., Bilger, M., Naghavi, M., et al.(2013). Non-communicable diseases and injuries in Pakistan: Strategic priorities. TheLancet, 381, 2281–2290.Jasanoff, S., Markle, G. E., Peterson, J. C., & Pinch, T. (2001). Handbook of science andtechnology studies. Sage publications.Kalter, H., Schillinger, J., Hossain, M., Burnham, G., Saha, S., De Wit, V., et al. (1997).Identifying sick children requiring referral to hospital in Bangladesh. Bulletin of theWorld Health Organization, 75, 65.Khan, N. (2010). Beyond crisis: Re-Evaluating Pakistan. New Delhi: Routledge.Kim, J. W., Tiyyagura, G., & Langhan, M. (2019). A qualitative analysis of generalemergency medicine providers' perceptions on pediatric emergency telemedicine.Pediatric Emergency Care, 35, 856–861.Kirsch, T. D., Wadhwani, C., Sauer, L., Doocy, S., & Catlett, C. (2012). Impact of the 2010Pakistan floods on rural and urban populations at six months. PLoS currents, 4.Kissi, J., Dai, B., Dogbe, C. S., Banahene, J., & Ernest, O. (2020). Predictive factors ofphysicians' satisfaction with telemedicine services acceptance. Health InformaticsJournal, 26, 1866–1880.Krupinski, E., Nypaver, M., Poropatich, R., Ellis, D., Safwat, R., & Sapci, H. (2002).Chapter 2: Clinical applications in telemedicine/telehealth. Telemedicine Journal andE-Health, 8, 13–34.Lehoux, P., Sicotte, C., Denis, J.-L., Berg, M., & Lacroix, A. (2000). Trust as a keycomponent in the use of teleconsultation. Annals of the Royal College of Physicians andSurgeons of Canada, 33, 482–487.Lehoux, P., Sicotte, C., Denis, J.-L., Berg, M., & Lacroix, A. (2002). The theory of usebehind telemedicine:: How compatible with physicians' clinical routines? SocialScience & Medicine, 54, 889–904.Lehoux, P., Sicotte, C., & Lacroix, A. (1999). Theory of use behind telehealth applications.The impact of telemedicine on health care management. IOS Press.Liu, L., Oza, S., Hogan, D., Chu, Y., Perin, J., Zhu, J., et al. (2016). Global, regional, andnational causes of under-5 mortality in 2000–15: An updated systematic analysis withimplications for the sustainable development goals. The Lancet, 388, 3027–3035.Liu, L., Oza, S., Hogan, D., Perin, J., Rudan, I., Lawn, J. E., et al. (2015). Global, regional,and national causes of child mortality in 2000–13, with projections to inform post-2015 priorities: An updated systematic analysis. The Lancet, 385, 430–440.Ly, D. P., & Jena, A. B. (2018). Sex differences in time spent on household activities andcare of children among US physicians, 2003-2016. In Mayo clinic proceedings (pp.1484–1487). Elsevier.Mair, F., & Whitten, P. (2000). Systematic review of studies of patient satisfaction withtelemedicine. BMJ, 320, 1517–1520.Marcin, J. P., Nesbitt, T. S., Kallas, H. J., Struve, S. N., Traugott, C. A., & Dimand, R. J.(2004). Use of telemedicine to provide pediatric critical care inpatient consultationsto underserved rural Northern California. The Journal of Pediatrics, 144, 375–380.Masood, A. (2019). Influence of marriage on women's participation in medicine: The caseof doctor brides of Pakistan. Sex Roles, 80, 105–122.Mathieu-Fritz, A., & Esterle, L. (2013). Changes in professional practices as a result ofmedical teleconsultations. Interprofessional cooperation and the delegation of tasks.Revue Française de Sociologie, 54, 303–329.Mathieu-Fritz, A., & Gaglio, G. (2018). In search of the sociotechnical configurations oftelemedicine. R�eseaux, 27–63.Mathijs, S. (2007). Introduction to telemedicine (2nd ed.). International Journal ofIntegrated Care.May, C., & Ellis, N. T. (2001). When protocols fail: Technical evaluation, biomedicalknowledge, and the social production of ‘facts’ about a telemedicine clinic. SocialScience & Medicine, 53, 989–1002.May, C., Harrison, R., Finch, T., MacFarlane, A., Mair, F., & Wallace, P. (2003).Understanding the normalization of telemedicine services through qualitativeevaluation. Journal of the American Medical Informatics Association, 10, 596–604.McLean, S., Sheikh, A., Cresswell, K., Nurmatov, U., Mukherjee, M., Hemmi, A., et al.(2013). The impact of telehealthcare on the quality and safety of care: A systematicoverview. PLoS One, 8, Article e71238.Medicine, I.o (2007). Emergency care for children: Growing pains. Washington, DC: TheNational Academies Press.Mehdi, I. J. (1996). Emergency medicine in Pakistan. Annals of Emergency Medicine, 27,84–88.Mintrom, M. (2015). Herbert A. Simon, administrative behavior: A study of decision-makingprocesses in administrative organization. The Oxford Handbook of Classics in PublicPolicy and Administration.Mohsin, M., & Syed, J. (2020). The missing doctors—an analysis of educated women andfemale domesticity in Pakistan. Gender, Work and Organization, 27, 1077–1102.Molyneux, E. (2001). Paediatric emergency care in developing countries. The Lancet, 357,86–87.Mueller, K. J., Potter, A. J., MacKinney, A. C., & Ward, M. M. (2014). Lessons from tele-emergency: Improving care quality and health outcomes by expanding support forrural care systems. Health Affairs, 33, 228–234.Mumtaz, R., Qazi, S. H., & Sajjad, S. (2018). Balancing profession, family and culturalnorms by women dentists in Pakistan. Open Journal of Social Sciences, 6, 154.Nadkarni, A., & Mittal, L. (2021). Can telehealth advance professional equity for womenin medicine? Telemedicine and e-Health, 27(12), 1344–1345.National Institute of Population Studies - NIPS/Pakistan and ICF. (2019). Pakistandemographic and health survey 2017-18 (Islamabad, Pakistan, and Rockville,Maryland, USA).NHS Executive. (1998). Information for health: An information strategy for the modern NHS1998-2005: A national strategy for local implementation. NHS/Department of Health.Nicolini, D. (2006). The work to make telemedicine work: A social and articulative view.Social Science & Medicine, 62, 2754–2767.Nolan, T., Angos, P., Cunha, A. J., Muhe, L., Qazi, S., Simoes, E. A., et al. (2001). Qualityof hospital care for seriously ill children in less-developed countries. The Lancet, 357,106–110.Oudshoorn, N. (2012). How places matter: Telecare technologies and the changing spatialdimensions of healthcare. Social Studies of Science, 42, 121–142.Pappas, Y., & Seale, C. (2009). The opening phase of telemedicine consultations: Ananalysis of interaction. Social Science & Medicine, 68, 1229–1237.Pappas, Y., & Seale, C. (2010). The physical examination in telecardiology andtelevascular consultations: A study using conversation analysis. Patient Education andCounseling, 81, 113–118.Pappas, Y., Vseteckova, J., Mastellos, N., Greenfield, G., & Randhawa, G. (2019).Diagnosis and decision-making in telemedicine. Journal of patient experience, 6,296–304.Parajuli, R., & Doneys, P. (2017). Exploring the role of telemedicine in improving accessto healthcare services by women and girls in rural Nepal. Telematics and Informatics,34, 1166–1176.Pelletier-Fleury, N., Fargeon, V., Lano�e, J.-L., & Fardeau, M. (1997). Transaction costseconomics as a conceptual framework for the analysis of barriers to the diffusion oftelemedicine. Health Policy, 42, 1–14.Perednia, D. A., & Allen, A. (1995). Telemedicine technology and clinical applications.JAMA, 273, 483–488.Pols, J. (2011). Wonderful webcams: About active gazes and invisible technologies.Science, Technology & Human Values, 36, 451–473.Prasad, P. (1993). Symbolic processes in the implementation of technological change: Asymbolic interactionist study of work computerization. Academy of ManagementJournal, 36, 1400–1429.A.A. Rowther et al.SSM - Qualitative Research in Health 2 (2022) 10015014Qureshi, R. N., Sheikh, S., Khowaja, A. R., Hoodbhoy, Z., Zaidi, S., Sawchuck, D., et al.(2016). Health care seeking behaviours in pregnancy in rural Sindh, Pakistan: Aqualitative study. Reproductive Health, 13, 75–81.Ray, K. N., Felmet, K. A., Hamilton, M. F., Kuza, C. C., Saladino, R. A., Schultz, B. R., et al.(2017). Clinician attitudes toward adoption of pediatric emergency telemedicine inrural hospitals. Pediatric Emergency Care, 33, 250–257.Razzak, J. A., Baqir, S. M., Khan, U. R., Heller, D., Bhatti, J., & Hyder, A. A. (2015).Emergency and trauma care in Pakistan: A cross-sectional study of healthcare levels.Emergency Medicine Journal, 32, 207–213.Razzak, J. A., Cone, D. C., & Rehmani, R. (2001). Emergency medical services and culturaldeterminants of an emergency in Karachi, Pakistan. Prehospital Emergency Care, 5,312–316.Razzak, J. A., Hyder, A. A., Akhtar, T., Khan, M., & Khan, U. R. (2008). Assessingemergency medical care in low income countries: A pilot study from Pakistan. BMCEmergency Medicine, 8, 1–8.Razzak, J. A., & Kellermann, A. L. (2002). Emergency medical care in developingcountries: Is it worthwhile? Bulletin of the World Health Organization, 80, 900–905.Rehmani, R. (2004). Emergency section and overcrowding in a university hospital ofKarachi, Pakistan. Journal of Pakistan Medical Association, 54, 233.Salda~na, J. (2021). The coding manual for qualitative researchers. SAGE PublicationsLimited.Sauers-Ford, H. S., Hamline, M. Y., Gosdin, M. M., Kair, L. R., Weinberg, G. M.,Marcin, J. P., et al. (2019). Acceptability, usability, and effectiveness: A qualitativestudy evaluating a pediatric telemedicine program. Academic Emergency Medicine, 26,1022–1033.Scott Kruse, C., Karem, P., Shifflett, K., Vegi, L., Ravi, K., & Brooks, M. (2018). Evaluatingbarriers to adopting telemedicine worldwide: A systematic review. Journal ofTelemedicine and Telecare, 24, 4–12.Sicotte, C., & Lehoux, P. (2003). Teleconsultation: Rejected and emerging uses. Methods ofInformation in Medicine, 42, 451–457.Sismondo, S. (2010). An introduction to science and technology studies. Chichester: Wiley-Blackwell.Steckler, A. B., Linnan, L., & Israel, B. (2002). Process evaluation for public healthinterventions and research. San Francisco, CA: Jossey-Bass.Strehle, E., & Shabde, N. (2006). One hundred years of telemedicine: Does this newtechnology have a place in paediatrics? Archives of Disease in Childhood, 91, 956–959.Syed, U., Khadka, N., Khan, A., & Wall, S. (2008). Care-seeking practices in South Asia:Using formative research to design program interventions to save newborn lives.Journal of Perinatology, 28, S9–S13.Trupia, D. V., Mathieu-Fritz, A., & Duong, T. A. (2021). The sociological perspective ofusers' invisible work: A qualitative research framework for studying digital healthinnovations integration. Journal of Medical Internet Research, 23, Article e25159.Umer, M., Sepah, Y. J., Shahpurwala, M. M., & Zafar, H. (2009). Suicide bombings:Process of care of mass casualties in the developing world. Disasters, 33, 809–821.UNICEF, U. (2015). Levels and trends in child mortality. New York: UNICEF.Uscher-Pines, L., & Kahn, J. M. (2014). Barriers and facilitators to pediatric emergencytelemedicine in the United States. Telemedicine and e-Health, 20, 990–996.Wade, V. A., Eliott, J. A., & Hiller, J. E. (2014). Clinician acceptance is the key factor forsustainable telehealth services. Qualitative Health Research, 24, 682–694.Wang, H., Liddell, C. A., Coates, M. M., Mooney, M. D., Levitz, C. E., Schumacher, A. E.,et al. (2014). Global, regional, and national levels of neonatal, infant, and under-5mortality during 1990–2013: A systematic analysis for the global burden of diseasestudy 2013. The Lancet, 384, 957–979.Watkins, K. (2016). The state of the world's children 2016: A fair chance for every child. ERIC.Whitten, P., Davenport Sypher, B., & Patterson, J. D. (2000). Transcending the technologyof telemedicine: An analysis of telemedicine in North Carolina. Health Communication,12, 109–135.Wootton, R., & Bonnardot, L. (2015). Telemedicine in low-resource settings. Frontiers inPublic Health, 3, 3.Wootton, R., Craig, J., & Patterson, V. (2017). Introduction to telemedicine. CRC Press.World Health Organization. (2010). Telemedicine: Opportunities and developments inmember states. Report on the second global survey on eHealth. World HealthOrganization.World Health Organization. (2016). Monitoring and evaluating digital health interventions: Apractical guide to conducting research and assessment.World Health Organization. (2017). Global diffusion of eHealth: Making universal healthcoverage achievable: Report of the third global survey on eHealth. World HealthOrganization.World Health Organization. (2018). Classification of digital health interventions v1. 0: Ashared language to describe the uses of digital technology for health. World HealthOrganization.Zaylor, C. (1999). Clinical outcomes in telepsychiatry. Journal of Telemedicine andTelecare, 5, 59–60.de Zoysa, I., Bhandari, N., Akhtari, N., & Bhan, M. K. (1998). Careseeking for illness inyoung infants in an urban slum in India. Social Science & Medicine, 47, 2101–2111.A.A. Rowther et al.SSM - Qualitative Research in Health 2 (2022) 10015015View publication stats | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/216646852Anthropogenic arsenic menace in Delhi Yamuna Flood PlainsArticle in Environmental Earth Sciences · January 2012DOI: 10.1007/s12665-011-1072-2CITATIONS39READS5326 authors, including:Some of the authors of this publication are also working on these related projects:Morphometric characteristics as indicators of active tectonics in Sikkim (Tista River, eastern Himalaya): View projectNuclear Structure study in mass A ~ 140 region View projectChandra Shekhar DubeyK. R. Mangalam Institutions of Higher Education97 PUBLICATIONS 1,858 CITATIONS SEE PROFILEBhupendra K MishraAKS University, Satna9 PUBLICATIONS 165 CITATIONS SEE PROFILEDericks Praise ShuklaIndian Institute of Technology Mandi79 PUBLICATIONS 927 CITATIONS SEE PROFILER. P. SinghInter University Accelerator Centre1,004 PUBLICATIONS 8,277 CITATIONS SEE PROFILEAll content following this page was uploaded by Dericks Praise Shukla on 20 May 2014.The user has requested enhancement of the downloaded file.1 23Environmental Earth Sciences ISSN 1866-6280Volume 65Number 1 Environ Earth Sci (2012) 65:131-139DOI 10.1007/s12665-011-1072-2Anthropogenic arsenic menace in DelhiYamuna Flood PlainsChandra S. Dubey, BhupendraK. Mishra, Dericks P. Shukla, RavindraP. Singh, M. Tajbakhsh & PurushottamSakhare1 23Your article is protected by copyright andall rights are held exclusively by Springer-Verlag. This e-offprint is for personal use onlyand shall not be self-archived in electronicrepositories. If you wish to self-archive yourwork, please use the accepted author’sversion for posting to your own website oryour institution’s repository. You may furtherdeposit the accepted author’s version on afunder’s repository at a funder’s request,provided it is not made publicly available until12 months after publication.ORIGINAL ARTICLEAnthropogenic arsenic menace in Delhi Yamuna Flood PlainsChandra S. Dubey • Bhupendra K. Mishra •Dericks P. Shukla • Ravindra P. Singh •M. Tajbakhsh • Purushottam SakhareReceived: 13 December 2009 / Accepted: 20 April 2011 / Published online: 6 May 2011� Springer-Verlag 2011AbstractArsenic, one of the most poisonous chemicalelements, was analyzed in the waters of the host of the2010 Commonwealth Games, i.e., New Delhi. The studyrevealed shocking outcomes with arsenic concentrationswell beyond the safe limits set by WHO, and a maximumconcentration up to 180 ppb was found in the groundwater.Analysis of around 120 water samples collected exten-sively along the Yamuna Flood Plain showed that morethan 55% had arsenic contamination beyond the WHOlimit of 10 ppb. The maximum value of arsenic in coal andfly ash from Rajghat coal-based thermal power plant con-tained 200 and 3,200 ppb, respectively. Moreover, theore petrography of coal samples shows the presence ofarsenopyrite mineral. Maximum concentration of arseniccontamination is found within a 5-km radius from powerplants. In the perspective of Delhi, arsenic contamination ispurely anthropogenic due to coal-based thermal powerplants, which had already shown toxic arsenic, fluorine andChina-type coal effects. The presence of such power plantsin coal field locations, e.g., West Bengal and Bangladesh,could release the arsenic due to combustion in superther-mal power plants, thus accentuating the arsenic concen-tration besides the natural arsenic coming from the forelandbasins of the Himalaya in Indian sub-continent.KeywordsArsenic � Water � Coal and fly ash �Coal-based thermal power plant � DelhiIntroductionArsenic is associated with many problems, being a naturalor anthropogenic contaminant in areas where human sub-sistence is at risk. It comprises a large number of chemicalspecies with widely differing toxicities (Jain and Ali 2000;Smedley 2006) and producing non-specific pathologies,which are hard to reverse and socially disabling. Anthro-pogenic contaminants with varying toxic and deleteriouscontaminants have been studied in detail in and aroundthe national capital region by various workers such asKrishnamurti (1987), Ram et al. (1997), Mehra et al.(1998), Singh (2001), Dubey et al. (2002), Goel et al.(2004), Jain (2004), Kumar et al. (2006, 2009) andSrivastava and Ramanathan (2008).The study incorporates investigations in arsenic con-tamination located in parts of NCT-Delhi, the capital ofIndia in the vicinity of Rajghat and Badarpur coal-basedthermal power plant with a total generation capacity of1,085 MW. Arsenic, being a very coalphile element, hasstrong affinity to coal matter, i.e., organic and (or) inor-ganic, but obligatorily authigenic. There are instances ofmajor dust pollution around power stations from fly ashdispersal. The main method of disposal of fly ash fromthe power stations is by mixing with water; the resultantslurry is pumped through pipes to ash disposal ponds.C. S. Dubey � B. K. Mishra � D. P. Shukla (&) � R. P. SinghDepartment of Geology, Center for Advanced Studies(Environment and Neotectonic Group), University of Delhi,Delhi-07, Indiae-mail: dericks.82@gmail.comC. S. Dubeye-mail: csdubey@gmail.comM. TajbakhshDepartment of Watershed Management,Birjand University, Birjand, IranP. SakhareImpact Assessment Division (Industries),Ministry of Environment & Forest, Government of India,Lodhi Road, New Delhi-03, India123Environ Earth Sci (2012) 65:131–139DOI 10.1007/s12665-011-1072-2Author's personal copyThe supernatant from these ponds is discharged into RiverYamuna (Mehra et al. 1998). Lalwani et al. (2004, 2005,2006) analyzed arsenic in groundwater samples of Delhiand found values as low as 17 ppb at Rainy Well and amaximum value of 100 ppb at Kotla Mubarakpur, leadingto the assumption that chemical-rich garbage dumping intothe open landfills or industrial waste is the possible sourceof contamination of the groundwater of Delhi.The investigated area (Fig. 1), forming a part of Surveyof India topographic sheet No.53H/2/NE and 53H/2/SE,lies between latitudes 28�300 3000 to 28� 4304800N and lon-gitudes E 77�1500000 to 77�1805700. The total area is about1,474.88 km2 comprising about 1,033.28 km2 rural and441.60 km2 urban areas. The altitude varies from 165 mnear the Yamuna Flood Plains to a maximum of 315 m atthe Aravalli ridges.It is well established that coal-containing arsenopyritesused without treatment causes major outbreak of cancerousdiseases, as in the case of the Guizhou Province, InnerMongolia,China.Moreoverasmentionedearlier,researchers have found arsenic contamination in the Delhiregion; hence, in the present study an attempt is made tocarry out the arsenic mapping and to know the source ofarsenic contamination in parts of Delhi and Yamuna FloodPlain.Sampling and methodsGroundwater and surface samples were collected fromdifferent sites as indicated in the boxes in Fig. 1 during themonths of May–June (pre-monsoon) and end of August(post-monsoon) in two successive seasons in the year 2007.Groundwater samples were collected from hand pumps,tube wells and power pumps, while surface water sampleswere collected from Yamuna River water, drain water,effluent drain water and fly ash slurry water. Water fromhand pumps and power pumps are generally used fordrinking and various household purposes, while tube wellwater is used for irrigation purposes by people living nearthe riverbank. Approximately, 120 samples were collectedand their temperature, electrical conductivity (EC) and pHwere measured in the field itself using digital instrumentsas shown in Table 1. These samples were analyzed in thefield for arsenic contamination using Wagtech DigitalArsenator (WDA), whose detection limit is as low as2 ppb. This digital arsenator has a correlation of 97% withgraphite furnace atomic absorption spectrometry (GF-AAS) Swash (2003) and is certified by the Sri RamInstitute, Delhi and Kanpur University, Kanpur. Sankarar-amakrishnan et al. (2008) concluded that the WagtechDigital Arsenator is suitable for arsenic testing (up toFig. 1 Map showing theinvestigated area in and aroundYamuna Flood Plain. Boxes A,B and C show the area of higharsenic concentrations132Environ Earth Sci (2012) 65:131–139123Author's personal copyTable 1 Table showing the concentration of arsenic in ppb, temperature, depth, electrical conductivity (EC) and pH in water samples of Delhi Area in the vicinity of Yamuna River. Sampleshaving arsenic concentration higher than WHO limit of 10 ppb are mentioned here out of a total of 98 samples (for more explanation, see the text)S. no.Sample no.LatitudeLongitudeSource of waterDepth (m)Temperature (�C)EC (mS/cm)pHAs (ppb)1BP22050733N 28�30024.400E 77�19000.100Hand pump (GW)12.19228.916.9107 (180*)2GC12050701N 28�38038.900E 77�1604.3600Hand pump (GW)10.66820.41.26.8104 (88*)3GC12050702N 28�38042.1600E 77�1604.500Hand pump (GW)10.66823.51.26.282 (86**)4GC12050703N 28�38042.1600E 77� 160 4.500Hand pump (GW)9.14426.31.3691 (86*)5GC12050704N 28�38043.4700E 77�1601.3600Hand Pump (GW)10.66825.71.36.1486GC12050705N 28�38047.8400E 77�1603.4800Hand pump (GW)10.66825.51.26.277 (69**)7GC12050706N 28�38049.3600E 77�1601.12Hand pump (GW)10.66825.51.26.259 (46*)8GC12050707N 28�38049.3600E 77�1601.1200Tube well (GW)18.28825.51.26.2279GC12050709N 28�39016.0600E 77�15056.5800Hand pump (GW)12.19225.31.26.273 (101*)10GC12050710N 28�3908.4500E 77�1607.0400Power pump (GW)30.4827.51.16.42211GC15050712N 28�38040.300E 77�16014.100Hand pump (GW)18.28826.31.36.42612GC15050713N 28�38047.600E 77�16019.500Drain water (SW)–28.41.56.14713GC15050713AN 28�38047.600E 77�16019.500Hand pump (GW)9.14427.51.56.11114GC15050714N 28�38035.800E 77�16012.300Hand pump (GW)12.19226.61.26.52115GC15050716N 28�38059.200E 77�16019.300Hand pump (GW)18.28829.21.16.591(101*) (98**)16GC15050717N 28�39000.800E 77�16022.400Hand pump (GW)15.2427.11.65.81717GC15050718AN 28�39000.900E 77�16022.500Hand pump (GW)15.2426.41.26.44418GC15050719N 28�38034.000E 77�16013.800Power pump (GW)39.62427.40.76.89319GC260507110N 28�38049.900E 77�16013.500Hand pump (GW)12.19225.61.17.6110 (94*) (92**)20GC26050741N 28�38049.700E 77�16012.600Hand pump (GW)13.71627.41.37.14521IP18050730N 28�37011.100E 77�15019.200Power pump (GW)6.09629.62.95.89122IP18050731N 28�3700.5300E 77�15019.500Hand pump (GW)9.14427.636.33423MV300507114N 28�35046.600E 77�16025.200Hand pump (GW)12.19225.31.16.9110 (135*) (139**)24MV30050762N 28�36009.600E 77�16027.900Hand pump (GW)12.19224.71.17.21825MV30050763N 28�36004.800E77�16019.700Hand pump (GW)12.19226.41.17.41626MV30050764N 28�35050.000E 77�16012.500Hand pump (GW)12.19227.21.16.891 (93*) (97**)27MV30050768N 28�36003.000E 77�16048.700Power pump (GW)24.38428.60.87.62728NG190807127N 28�4204800E 77�1500.000Hand pump (GW)15.2426.41.56.12529NG190807128N 28�4201200E 77�1503600Hand pump (GW)12.19227.51.26.52030RW28050744N 28�37044.300E 77�15036.600Hand pump (GW)12.19225.41.47.12931RW28050745N 28�37048.300E 77�15036.800Hand pump (GW)12.19225.31.27.11332RW28050747N 28�37048.300E 77�15036.800Hand pump (GW)12.19226.11.37.12633RW28050748N 28�37055.600E 77�15047.500Tube well (GW)18.28838.517.33934RW28050749N 28�37055.600E 77�15047.500Hand pump (GW)12.19226.31.26.873 (96*) (92**)35RW28050750N 28�37055.700E 77�15047.500Tube well (GW)15.2437.21.16.939Environ Earth Sci (2012) 65:131–139133123Author's personal copy100 ppb) in field having a correlation of 0.95 with labo-ratory measurements. The standard samples of arsenicprepared at the Sri Ram Institute were also checked by thetwo different Wagtech Digital Arsenators in the Geo-chemistry Labotatory of the Department of Geology,University of Delhi, which gave similar correlations asabove.The bottles were rinsed several times with water, filteredon site, completely filled with the same water, then acidi-fied with distilled HNO3 to maintain a pH value of 2 andtightly closed to minimize any contact with air. The sam-ples were taken in duplicate and placed in 250 and 500-mLpolypropylene plastic bottles for further analysis byGF-AAS at the Department of Geology, University ofDelhi and the Sri Ram Institute, Delhi.In the vicinity of the study area, there are two coal-basedthermal power plants, which have been running for the last40 years. Their effluent water, fly ash water, fly ash andcoal used for power generation were also collected andtested for arsenic contamination. Coal samples were ana-lyzed by reflected light microscopy to ascertain mineralogyand arsenic-bearing minerals.Approximately, 1 g of an air-dried coal/fly ash wasweighed into a pre-weighed 30-mL porcelain crucible andthoroughly mixed with 1.5 g of Eschka Mixture and cov-ered with an additional 1.5 g. The sample was heated in acold muffle furnace at 500�C for 1 h, and the temperaturewas increased to 750�C for 3 h. After cooling, 30 mL ofhot deionized water was added to a 150-mL beaker, and5 mL of concentrated HCl was slowly and carefully addedwhile stirring. Then, 15 mL of concentrated HCl wasadded in three 5 mL portions and the solution was cooledand transferred to a 100-mL polypropylene volumetricflask. The volume was diluted with deionized water asrequired for analysis by GF-AAS. These samples wereanalyzed by GF-AAS with the background correctionsystem and Strichart recorder and hydride generationapparatus for producing the hydrides of arsenic at the ShriRam Institute, Delhi.Results and discussionGroundwater analysisArsenic contamination in groundwater became a high-profile problem in recent years due to the use of deep tubewell for water supply in the Ganga Delta, causing seriousarsenic poisoning to a large number of people (Mandal andSengupta 2006). Li et al. (2007) found that over 137 mil-lion people in more than 70 countries are probably affectedby arsenic poisoning of drinking water. It is now apparentthatarseniccontaminationofgroundwaterisquiteTable 1 continuedS. no.Sample no.LatitudeLongitudeSource of waterDepth (m)Temperature (�C)EC (mS/cm)pHAs (ppb)36RW28050752N 28�37049.300E 77�15039.600Hand pump (GW)12.19220.31.46.64137TP16050722N 28�38034.500E 77�15033.000Hand pump (GW)9.14426.91.56.375 (78*)38TP16050723AN 28�3804100E 77�15023.100Fly ash slurry water (SW)–31.21.16.93539TP16050724N 28�38026.800E 77�15024.700Hand pump (GW)10.66826.71.36.775 (76*)40TP16050725N 28�38018.700E 77�15034.300Hand pump (GW)12.19232.81.96.31441TP16050726AN 28�3800.5300E 77�15025.400Effluent drain water (SW)–33.41.36.24842TP29050753N 28�3800.6500E 77�15022.300Effluent drain water (SW)–28.717.15543TP29050754N 28�3800.6500E 77�15022.300Effluent drain water (SW)–29.317.22744TP29050756N 28�3802.300E 77�15024.600Effluent drain water (SW)–28.21.16.74345TP29050757N 28�37043.800E 77�15012.200Yamuna riverwater (SW)–28.51.56.81146WB02060776N 28�43028.600E 77�1307800Hand pump (GW)9.14427.51.56.12947WB02060777N 28�43028.100E 77�13038.200Hand pump (GW)18.28826.61.26.51148WB02060779N 28�430 28.100E 77�13038.200Hand pump (GW)9.14427.51.56.12349WB02060782N 28�43022.4300E 77�1304000Hand pump (GW)15.2426.61.26.52050WB02060784N 28�43016.200E 77�1309800Hand pump (GW)15.2427.51.26.223The concentration indicated in brackets with * shows post-monsoon sample analysis, while ** shows samples analyzed by AAS134Environ Earth Sci (2012) 65:131–139123Author's personal copyextensive in India (Chakraborti et al. 1998; Pandey et al.1999; Jain 2002; Acharyya et al. 2005; Shukla et al. 2010).Samples were collected in two different seasons from theYamuna Flood Plain area to insure the quality results.Arsenic is observed in most of the groundwater samples.More than 75% of these samples were contaminated witharsenic concentrations, whereas more than 55% of thesampleshaveconcentrationsabovetheWHOlimit(Table 1).Rainy Well area is located at the opposite side of theRajghat coal-based thermal power plant where watersamples show high arsenic concentration. In one watersample (N 28�37055.600, E 77�15047.500) at a depth of10.192 m, arsenic concentration was 73 ppb, which roseto 96 ppb after the monsoons. At another location (N28�37011.100, E 77�15019.200) near Indraprastha coal-basedthermal power plant, the arsenic concentration was foundto be 91 ppb, as indicated in Fig. 2a. Lalwani et al. (2004)reported 0.0170 ppm of arsenic in the Rainy Well area.A maximum arsenic concentration of 110 ppb wasdetected in a hand pump having approximate depth of12.19 m, near the Pushta bus stand (N 28�38049.900, E77�16013.500) in Geeta Colony area as shown in Fig. 2b; itis still being used for drinking purposes. High arsenic(104 ppb) is observed in hand pump water sample near theBhamb Road. Most of the groundwater samples of this areaare contaminated. Water sample collected from Shri BalajiTemple (12-Block in Geeta Colony) by hand pump aftermonsoons had an arsenic value of 101 ppb (N 28�38059.200,E 77�16019.300).Mayur Vihar Phase-I near Akshardham Temple, wherethe land is used for agricultural purposes, mainly for veg-etable cultivation, showed arsenic contamination as highas 135 ppb (N 28�35046.600, E 77�16025.200) as shownin Fig. 2c. The same sample gave an arsenic value of139 ppb, as analyzed by GF-AAS at the Sri Ram Institute,Delhi, depicting a high correlation factor (0.97) for thereliability of analyzed data. The survey showed stomach,gastric and gastrointestinal problems reported by the peo-ple who drank hand pump water in high arsenic-contami-nated areas. Moreover, this contaminated groundwater isused for agricultural purposes in all vegetable farms in andaround the Yamuna Flood Plain.In the Badarpur area, a sample collected from oppositeto the Badarpur coal-based thermal power plant across thecanal(N28�30024.400,E77�19000.100)hasarsenicFig. 2 Map showing sample locations: (a) Rajghat TPP, IndraprasthaTPP, Rainy Well area, (b) Geeta Colony and (c) Mayur Vihar area inthe vicinity of Yamuna River, depicting high arsenic concentration, asindicated in values in ppb over respective bars. Yellow color barsindicate concentration above 50 ppb, magenta color bars have20–50 ppb, while green color bars indicate less than 20 ppbconcentrationEnviron Earth Sci (2012) 65:131–139135123Author's personal copyconcentration of 107 ppb with a groundwater level depth of12.192 m in the summer season (May), which rose to180 ppb post-monsoon season The pH (6.5–7) of the waterwas close to near neutral and showed more mobility andleaching during the monsoon to post-monsoon period withrise in water table.Surface water and fly ash slurry analysisThe surface water samples from different locations ofYamuna showed comparatively lower concentration ofarsenic from 10 to 11 ppb except at one location at a drainnear Geeta colony (47 ppb) whose water is utilized foragricultural purposes.The artificial drains from these power plants that flushout waste material and effluent water from the plant toYamuna River were analyzed. The sampling and analysisof this effluent water from these power plants (Rajghat andIndraprastha) show arsenic concentration varying from 42to 55 ppb. (N 28�3800.6500, E 77�15022.300) (Fig. 2a).The fly ash slurry from the fly ash pond in the vicinity ofpower plants was analyzed and showed 35 ppb of arsenicconcentration (N 28�3804100, E 77�15023.100). Krishnamurti(1987) conducted an EDXA on the fly ash samples fromIndraprasthapowerstation,depictingarangeof0.091–1.328% elemental arsenic and predicted the coalused to be the major source of arsenic pollution. Theeffluents from Indraprastha power plant showed lowerarsenic contamination (10 ppb). Since the effluents and flyash slurry from coal-based power plants had arsenic con-centration, the coal and fly ash from Rajghat and Indrap-rastha power plant was analyzed for identification of thesource of arsenic contamination in the area.Analysis of coal and fly ash using GF-AAS:The coal samples from Rajghat and Badarpur power plantswere analyzed by atomic absorption spectroscopy (GF-AAS) and yielded 230 and 200 ppb of arsenic, respec-tively. The GF-AAS analysis of fly ash from Rajghat powerplant ash pond (N 28�3804100, E 77�15023.100) depicted thepresence of high concentration of arsenic (3,200 ppb)(Table 2). During coal combustion in power plants, most ofthe initial arsenic in coal volatilizes into the gaseous phase.In the widely used combustion of pulverized coal, most ofthe arsenic-bearing micro-mineral phases escape into gas-eous and particulate phases and only a minor part of arsenicclay remains in the bottom ash. The dominant fraction ofescaping arsenic is in fly ash. Because 97–99% of the flyash is collected by electrostatic precipitators, the atmo-spheric emission of arsenic (solid phase and gaseous) isusually assumed to be rather minor (10–30% from initialarsenic in coal). However, fly ash disposal creates somedifficult environmental problems because it is potentiallytoxic in natural waters and soils. The As leaching rate fromash disposal is greatly controlled by the ash chemistry. Inthe natural environment, arsenic can be readily leachedfrom acid (SiO2 rich) bituminous coal ashes (Yudovich andKetris2005).Analysis of coal samples by reflected light microscopyCoal petrography was done using reflected light micros-copy of polished thin sections. Most of the coal sampleswere identified as low rank coals (e.g., lignite and bitu-minous), consisting of plant matter such as mosses andother bog plant materials.These coal samples show the presence of arsenic-bearingsulfide minerals such as pyrite and arsenopyrite (Fig. 3a, b)within the organic matrix of coal. Pyrite is in the form ofbright yellowish, euhedral cubes within the organic matrixof coal. It is isotropic, pleochroic and has perfect cubiccleavage. Arsenopyrite is bright creamy white with a tint ofbrass yellow in reflected light and has an orthorhombicshape. It is strongly anisotropic with polarized light and hasvery high reflectivity and lamellar twinings.There is strong regional variability of arsenic distribu-tion due to geologic variability of the individual coalbasins. For example, bituminous coals in Eastern Germany,Czech Republic and SE China are enriched in arsenic,whereas the coals in South Africa or Australia are muchdepleted in metal concentration compared to other coals. Ingeneral, some relationship exists between the arseniccontent and its mode of occurrence in coals. Typically, athigh arsenic content, sulfide sides dominate (pyrite andother more rare sulfides), whereas at low arsenic content,organic arsenic dominates, both being authigenic (Yudo-vich and Ketris 2005). The arsenic data for coals in India isnegligible except for few detailed works (Priyadarshi2004), and more research is needed to understand arsenicTable 2 Arsenic concentration of coal and fly ash samples from Rajghat and Badarpur power plant analyzed by AASS. no.Sample no.LatitudeLongitudeTypes of sampleAs (ppb)1TP18050732N 28�37043.800E 77�15012.200Coal2302BP22050736N 28�30029.500E 77�18052.000Coal2003TP16050723N 28�38036.400E 77�15026.100Fly ash3,200136Environ Earth Sci (2012) 65:131–139123Author's personal copyrelease from Indian coalfields. As it is observed that withincreasing coal rank sorbed exchangeable arsenate contentdecreases, the presence of low rank coal from Bina (M.P.)has high arsenic content.Dispersion of arsenic from the Rajghat power plantThe arsenic values, which depict the dispersion and con-tamination of arsenic in different directions, were plottedfor the study area. The previous section reveals that arsenicfrom Rajghat coal-based thermal power plant pollutes theRiver Yamuna and its flood plain. Thus considering themto be the source of this contamination, arsenic dispersion isfound out to be easterly with Geeta Colony being highlyaffected as shown in Fig. 2. The arsenic values increase asone goes away from the Yamuna River in an NE direction.Since Yamuna River flows in the SE direction, arsenicshould be dispersed along the flow of river, which is con-trary to the results obtained. The reason for dispersion inthis NE direction is attributed to sludge gates being presentat the Vikas Marg Bridge at ITO. The mobility and infil-tration of arsenic in nearby sediments and groundwater ofYamuna Flood Plain area are controlled by operation ofthese sludge gates in different seasons. These gates areused to maintain the continuous flow of the river. It hasbeen observed that these gates are closed for most of themonths of the year. This causes arsenic to mobilize in theopposite direction of the river flow. On the western side ofthe river, Delhi quartzites are present, which does not allowthe dispersion in the western direction and hence the dis-persion is in the eastern direction, which gets highlyaffected due to accumulation and dispersion of arsenic inthis area. Thus, we can conclude that the arsenic problempersisting in Geeta Colony is due to the untreated effluentscoming out of the Rajghat power plant.Flux of arsenic in the Yamuna Flood PlainsThere are two major artificial draining effluents fromRajghat and Indraprastha power station into the YamunaRiver. The flux is estimated based on the discharge rate,water run-off and concentration of arsenic mixing into theYamuna River per year. The calculation is done with thehelp of field data and formula mentioned below. Dischargerate formula used is given below. The discharge unit ism3/s.Q ¼ A � VTheoretically,Q = Streamdischarge(volume/time),A = cross sectional area, and V = flow velocity.The flux was calculated by the following equation(Chapman 1992).U ¼ZCðtÞQðtÞdtTheoretically, flux (U) is in mass per unit time, and usuallyta-1 is derived from the continuous measurement of bothwater discharge Q (m3 s-1) and concentrations C (mass pervolume, usually mg L-1) between time t1 and t2.The wastewater effluents from these power plants havebeen drained into the Yamuna River for more than40 years. Water samples from these drains were analyzedat different locations in pre- and post-monsoon seasons.Various parameters such as arsenic concentration, dis-charge rate, flow velocity of drain, cross-sectional area andflux were measured and calculated for the two drain sites,i.e., Rajghat and Indraprastha (Table 3). Using the above-mentioned formula for flux, it is estimated that 5.5 and1.96 Tonnes/year of arsenic have been dispersed fromRajghat and Indraprastha power plants, respectively, in theYamuna River. Chronic arsenic poisoning from higharsenic-containing coal in China is an example of long-Fig. 3 a Pictomicrograph showing the presence of arsenopyrite grain(Aspy) in the coal sample of Rajghat power plant. b Pictomicrographshowing the presence of pyrite grain (Py) in the coal sample ofRajghat power plantEnviron Earth Sci (2012) 65:131–139137123Author's personal copyterm exposure and its health hazards (Liu et al. 2002; Yuet al. 2006). The arsenic environment fluxes from the coal-based thermal power plants show that arsenic has pollutedthe Yamuna Flood Plains, which is used for vegetablefarming especially edible roots of dietary vegetables(Brake et al. 2004).ConclusionFor the first time, we report the source and origin ofanthropogenic arsenic contamination in parts of Delhi withspecial emphasis on the Yamuna Flood Plains. The studyarea of Geeta Colony, Mayur Vihar, Wazirabad, Nigam-bodh Ghat, Kotla Mubarakpur and Rajghat, Indraprasthaand Badarpur coal-based thermal power plants show thepresence of arsenic content above the permissible limit ofthe World Health Organisation (WHO) and the Bureau ofIndian Standards (BIS). The surface water samples acrossthe Yamuna Flood Plain, and drain water of Rajghat andIndraprastha thermal coal-based power plants also showhigh values of arsenic contamination; 5.5 Tonnes/year ofarsenic is discharged into the Yamuna River from theRajghat power plant.Thee area is the future hub of the CommonwealthGames. The ground and surface waters with arsenic con-tamination have contaminated vegetables grown in theYamuna Flood Plains posing serious health hazards to thelocals of the area. At the first instance, arsenic mappingwill be an important input to observe the levels of problemand its management, especially in the vicinity of coal fieldsand coal-based thermal power plants of India.ReferencesAcharyya SK, Shah BA, Ashyiya ID, Pandey Y (2005) Arseniccontamination in groundwater from parts of Ambagarh-Chowkiblock, Chhattisgarh, India: source and release mechanism.Environ Geol 49:148–158Brake SS, Jensen RR, Mattox JM (2004) Effects of coal fly ashamended soils on trace element uptake in plants. Environ Geol45:680–689Chakraborti D, Samanta G, Mandal BK, Roy Chowdhury T, ChandaCR, Biswas BK, Dhar RK, Basu GK, Saha KC (1998) Calcutta’sindustrial pollution: Groundwater arsenic contamination in aresidential area and sufferings of people due to industrial effluentdischarge—an eight-year study report. Curr Sci 74(4):346–355Chapman D (1992) Water quality assessments—a guide to use ofbiota, sediments and water in environmental monitoring, secondedition� 1992, 1996 UNESCO/WHO/UNEPDubey CS, Pattanayak SK, Sharma BK, Sirohi A (2002) Groundwaterpollution in the satellite city Faridabad, India. In: Dubey CS andSaklani PS (eds) Geoindicators and related environmentalstudies: focus on India (KB Powar Volume). Pilgrims Publishers,ISBN: 81-7769-187-2, Kathmandu, Nepal, pp 47–56Goel R, Kapoor SK, Misra K, Sharma RK (2004) Removal of arsenicfrom water by different adsorbents. Indian J Chem Technol11:518–525Jain CK (2002) Arsenic contamination in ground water: Indianscenario. Indian J Environ Health 44:238–243Jain CK (2004) Metal fractionation study on bed sediments of RiverYamuna. India Water Res 38:569–578Jain CK, Ali I (2000) Arsenic: occurrence, toxicity and speciationtechniques. Water Res 34:4304–4312Krishnamurti CR (1987) The cycling of arsenic, cadmium, lead andmercury in India. In: Hutchison TC, Meema KM (eds) Lead,mercury, cadmium and arsenic in the environment. Wiley, India,pp 315–319Kumar M, Ramanathan AL, Rao MS, Kumar B (2006) Identificationand evaluation of hydrogeochemical processes in the ground-water environment of Delhi, India. Environ Geol 50:1025–1039Kumar M, Sharma B, Ramanathan AL, Rao MS, Kumar B (2009)Nutrient chemistry and salinity mapping of the Delhi aquifer, India:source identification perspective. Environ Geol 56:1171–1181Lalwani S, Dogra TD, Bhardwaj DN, Sharma RK, Murty OP, Vij A(2004) Study on arsenic level in groundwater of Delhi usinghydride generator accessory coupled with atomic absorptionspectrophotometer. Indian J Clin Biochem 19:135–140Lalwani S, Dogra TD, Bhardwaj DN, Sharma RK, Murty OP (2005)Efficacy of Filter Plant in Removal of Arsenic from Water ofYamuna River. Indian J Commun Med 30(4):140–141Lalwani S, Dogra TD, Bhardwaj DN, Sharma RK, Murty OP (2006)Study on arsenic level in public water supply of Delhi usinghydride generator accessory coupled with atomic absorptionspectrophotometer. Indian J Clin Biochem 21:70–76Li C, Feng S, Shao Y, Jiang L, Lu X, Hou X (2007) Effects of arsenicon seed germination and physiological activities of wheatseedlings. J Environ Sci 19(6):725–732Liu J, Zheng B, Aposhian HV, Zhou Y, Chen ML, Zhang A, WaalkesMP (2002) Chronic Arsenic Poisoning from Burning High-Arsenic-Containing Coal in Guizhou, China. Environ HealthPerspect 110(2):119–122Mandal A, Sengupta D (2006) An assessment of soil contaminationdue to heavy metals around a coal-fired thermal power plant inIndia. Environ Geol 51:409–420Mehra A, Farago ME, Banerjee DK (1998) Impact of fly ash fromcoal fired power station in Delhi, with particular reference tometal contamination. Environ Monit Assess 50:15–35Pandey PK, Khare RN, Sharma S, Sar SK, Pandey M, Binayke P(1999)Arsenicosisanddeterioratinggroundwaterquality:unfolding crisis in central east India region. Current Science77:686–693Priyadarshi N (2004) Distribution of arsenic in Permian Coals ofNorthKaranpura coalfield,Jharkhand. JGeol Soc India63:533–536Table 3 The arsenic flux to Yamuna River from Rajghat andIndraptrastha power plantsS. no.Parameters measuredRajghatIndraprastha1LatitudeN 28�3800.6500N 28�37034.6002LongitudeE 77�15022.300E 77�15018.2003Arsenic value (ppb)55104Discharge rate (m3/s)3.276.235Speed of water (m/s)0.480.926Flux (tonnes/year)5.51.96138Environ Earth Sci (2012) 65:131–139123Author's personal copyRam S, Mishra HP, Zaidi A, Baranawal M (1997) Groundwaterpollution studies on eastern bank of Yamuna River, Delhi. GeolSurv India Special Publication 48(1):197–201Sankararamakrishnan N, Chauhan D, Nickson RT, Tripathi RM,Iyengar L (2008) Evaluation of two commercial field test kitsused for screening of groundwater for arsenic in Northern India.Sci Total Environ 401(1–3):162–167Shukla DP, Dubey CS, Singh NP, Tajbakhsh M, Chaudhry M (2010)Sources and controls of arsenic contamination in Rajnandgaonand Kanker District, Chattisgarh Central India. J Hydrol395:49–66Singh M (2001) Heavy metal pollution in freshly deposited sedimentsof the Yamuna River (the Ganges River tributary): a case studyfrom Delhi and Agra urban centres, India. Environ Geol40(6):664–671Smedley PL (2006) Sources and distribution of Arsenic in groundwater and aquifers. In: Proceedings of seminar on Arsenic inGround Water—a world problem, vol 5, pp 4–32Srivastava SK, Ramanathan AL (2008) Geochemical assessment ofgroundwater quality in vicinity of Bhalswa landfill, Delhi, India,using graphical and multivariate statistical methods. EnvironGeol 53:1509–1528Swash P (2003) Field Evaluation of the Wagtech Arsenator. RoyalSchool, Imperial College, London. http://www.physics.harvard.edu/*wilson/arsenic/measurement/field-eval-wagtech-arsenator.htm. Accessed 25 October 2009Yu G, Sun D, Zheng Y (2006) Health effect of exposure to naturalarsenic in groundwater and coal in China: an overview ofoccurrence. Environ Health Perspect 115:636–642Yudovich YE, Ketris MP (2005) Arsenic in coal: A review. Int J CoalGeol 61:141–196Environ Earth Sci (2012) 65:131–139139123Author's personal copyView publication stats | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/355153838A comparative study of outdoors thermal comfort in the Mediterranean climateArticle · March 2021CITATIONS0READS2514 authors:Some of the authors of this publication are also working on these related projects:[IEA Annex 80] Resilient Cooling for Buildings View project[RénoBâti] La rénovation du bâti résidentiel en Belgique View projectKahina LabdaouiBadji Mokhtar - Annaba University7 PUBLICATIONS 41 CITATIONS SEE PROFILESaid MazouzUniversité Larbi Ben Mhidi47 PUBLICATIONS 216 CITATIONS SEE PROFILEShady AttiaUniversity of Liège221 PUBLICATIONS 4,161 CITATIONS SEE PROFILEJacques TellerUniversity of Liège276 PUBLICATIONS 2,779 CITATIONS SEE PROFILEAll content following this page was uploaded by Shady Attia on 08 October 2021.The user has requested enhancement of the downloaded file.32FOCUS AbstractManystudiesexploredthermalcomfortinseveralclimatezonesandurbanenvironmentssinceitcaninfluencethequalityoflifeandurbanspaces'use.ThisstudyproposestoassessthermalcomfortbycombiningtwosoftwareRayMandandEnvi-met,intheMediterraneanclimate(Csa).ThermalcomfortisevaluatedbythePhysiologicalEquivalentTemperature(PET)andvalidatedinAnnaba,Algeria,characterizedbyCsa,withinHeatingDegreeDays(HDD=1200-1800),CoolingDegreeDays(CDD=1100-1700).Fouroutdoorspaces(2Courtsand2squares)havebeenevaluatedinsummer2017.Everyurbanareawasdividedintogreed(4m*4m),includingmeasurementpoints.TheresultsshowedastrongcorrelationbetweenairtemperatureT(a)andPET.Thetreescoolingeffectcontributestoimprovingthermalcomfort,especiallyduringthehothoursoftheday.ThedifferencesinT(a)reached4°Cand3°CforMeanradianttemperature(Tmrt)atnoon,consideredoneoftheday'shottesthours.Keywords:Heatingdegreedays(HDD)andcoolingdegreedays(CDD),Mediterraneanclimate(Csa),PhysiologicalEquivalenttemperature(PET),Envi-met,RayManIntroductionDuringthelastfewyears,specialattentionhasbeenpaidtothermalcomfortintheoutdoorurbanenvironmenttoimprovecitizenhealthandwell-beingwithinnewsustainablestrategies[1–6].Manystudiestriedtodefineoutdoorthermalcomfortconditions[2,7–12].Thermalcomfortisreferredtoastheusers'satisfactionlevelregardingthethermalenvironment[5,13].Understandingthemicroclimateenvironmentcharacteristicsisnecessary,asitinfluencescitizens'comfortandoutdooractivities[7].AccordingtoAceroandHerranz-Pascual[14],themeanradianttemperature(Tmrt),relativehumidity,airtemperature,windspeedarenecessarytocalculateoutdoorthermalcomfort,besidesmetabolicrateandclothinginsulation[15].ENVI-metandRaymanareapprovedsoftwaretocalculateoutdoorthermalcomfort.ENVI-metisfoundoncomputationalfluiddynamics(CFD)andthermodynamics,whileRayManis3Dimensionradiationmodel.SomestudiesfocusedondevelopingForparagraphtitlesuseCambria,size8boldanduniversalindicescapableofmeasuringthermalcomfortinahotandcoldclimate[5,7].ManyindicesHavebeensuccessfullyappliedtourbanareaforthermalcomfortassessment(e.g.,Perceivedtemperature(Pt)[5,16],standardeffectivetemperature(SET*)[5,17],anditsoutdoorvariant,OUT-SET*[18].TheUniversalThermalClimateIndex(UTCI)[19],thePhysiologicallyEquivalentTemperature(PET)[20,21]andpredictedmeanvote(PMV)[5,22–24].ThemostusedindicesforoutdoorthermalperceptionarePET,PMV,UTCI,andSET*[5].Theseindexesareappropriateforcalculatingheatingandcoldstress.However,SET*ismorequalifiedforhotclimates,whilePETandUTCIareprincipallyhiredandverifiedinhotandcoldclimates[25].ThePETIndexwasinitiatedinWesternandMiddleEurope[7,8,26].Itwastestedandvalidatedinseveralclimateszones[20,25,27,28],andexploredindifferentoutdoorenvironments[2,11,28–34].Themicroclimateconditionshaveanimportantimpactonoutdooractivities.Sunnyorshadyenvironmentssignificantlyaffectpeople'stimeduration,encouragingthemtostayorleave[35].Urbangreenspacesprincipallyproducedirectshadingandcoolingbyevapotranspiration,minimizingtheairandsurfacetemperatureandmakingalocalcooling[36–38].Manystudiestriedtoobtainacomprehensivediscernmentofthedifferencebetweensunlightandshadedenvironments.Matzarakisetal.[20]recorded15°CinPETdifferencevalueswithintheunshadedandtree-shadedarea.Johansson[39]exploredurbangeometry'simpactonoutdoorthermalcomfortwithinthehot,dryclimate.TheresultsshowedthatPETintheshadedanddeepurbancanyonarecoolerthatsallowone(20°Cofdifference)awingtothelowairtemperatureandmeanradianttemperature[15].Basedonthecurrentliterature,onlyafewworksexploredthermalcomfortbycombiningtwosoftwaresuchasEnvi-metandRaymanandvalidatedbyinsitumeasurementswithintheCsaclimate.ThepresentedmethodallowedmoreaccuratePETvaluesbycombiningtwosoftware(Envi-metandRayman)andvalidatedbyin-situmeasurementswithintheCsaclimate.Thismethodconcernstheentireurbanspacesareaandnotonlysomemeasurementpoints.Theeffectofthegreencoveronmicroclimaticparametersandthermalcomfortindexwasdefinedpreciselyaccordingtotheday'shours.Thisstudyaimedto(1)EvaluatethermalcomfortinfoururbanspacescharacterizedbydifferentgreencoverwithintheMediterraneanclimate.(2)Analyzetheimpactofgreenspacesonairtemperature,meanradianttemperature(Tmrt).KahinaLabdaouiSaidMazouzShadyAttiaJacquesTellerACOMPARATIVESTUDYOFOUTDOORSTHERMALCOMFORTINTHEMEDITERRANEANCLIMATEFig.1 – Conceptual framework of thermal comfort assessment using Calibration and other methods. 33 FOCUS 2.MethodTheproposedmethodisappliedinAnnabacity(Algeria)characterizedbytheMediterraneanClimate(Csa)accordingtotheKöppen[40]classificationand(HDD=1200-1800,CDD=1100-1700).Thisstudycomparesfoururbanspaces(twocourtsandtwosquares)locatedinAnnabacitycentreandElBouni(Annaba'ssuburban).Themethodologyincludedthreemainmethods:Datacollection,Dataprocessingandfinally,Dataanalysisandsynthesis.2.1DatacollectionEveryspacewasdividedintoagrid(4m*4m).Themeasurementpointswereselectedonthisgrid(10measurementpointsineachcourtwhileeverysquarehad5measurementpoints).WeusedtheLm8000tool(Thermo-Anemometer,Hygrometer,Thermometer&Illuminometer),ataheightof1.10mfromthegroundformeasuringairtemperature,relativehumidityandwindvelocity.However,theywerecontinuouslymeasured(everytwohours)duringtwosummer'sdaysfrom8amto8pm.2.2DataprocessingWeusedthehighspatialandtemporalresolutionmicroclimatemodelENVI-met.Thathasbeenwidelyusedandvalidatedforassessingbuiltenvironments.However,thismodelwasvalidatedagainsthuman-biometeorologicalmeasurementsanddemonstratedgoodperformancewhensimulatingandcalibratingtheurbanthermalenvironmentintermsofairtemperature(Ta)andhumanheatstress[41–43].WevalidatedthesimulatedresultscomparedtothemeasuredT(a)[8],highlightingagoodcorrelationbetweensetdata.Thatallowedhavingcalibrateddata(airtemperature,windvelocity,relativehumidity,andTmrt)fortheentireurbanspacesarea.ThisstudyadoptedtheRayManmodeltocalculatePET.ItcalculatesthePET-basedonairtemperature,airhumidity,windspeed,meanradianttemperature,humanclothing,andactivity[7,32,34].RayManusedcalibratedresultsasinputdatatocalculatePET,besidesestimatingtherelatedskyviewfactor(SVF).2.3DataAnalysisandSynthesisThisstepisrelatedtotheResultsection;thecalibrateddatasuchasT(a)andTmrtbesidesPEThavebeenanalyzed,differencesinT(a),Tmrt,andPETwerefoundinthefouroutdoorspaces.Figure1summarizesPhysiologicalEquivalentTemperature(PET),andfinally(3)exploringtheHeatingDegreedaysandCoolingDegreedayscriteriapresentanaddedvaluecomparedtopreviousstudies.2.4StudycaseThefieldstudywasconductedinAnnaba,Algeria,positionedbetween(Lattitude36°54'North,Longitude:7°46'East,Sealevel:+5m).AnnabaischaracterizedbytheHotSummerMediterraneanclimate(Csa),accordingtoKöppen[40]classification.Court1andsquare1areinAnnabacitycentre.Adensegreencovercharacterizesthesetwospaces.However,Square2andCourt2arelocatedInAnnaba'ssuburbanarea(ElBouni).Square2hasnotrees,whileCourt2includestwoparts,onewithtreesandthesecondonewithoutgreencoverage(Fig1).3.ResultsThissectionfirstpresentsacomparativeanalysisofcalibrateddataintheselectedoutdoorenvironmentssincetheysignificantlyimpactoutdoorthermalcomfort.PETisanalyzedindetailsconcerningmicroclimaticdataandtimeoftheday.ThelastpartoftheresultsincludestheimpactofgreencoveronPET,T(a)andTmrt.TheresultsshowedthatT(a),TmrtandPETarequitecloseonsomemeasurementpoints.Havinganaveragevalueofrelativemicroclimaticdataoneachurbanspaceallowsachievingbetteranalysisresults.Everysquarehasonerepresentativepoint,P1inSquare1andP1'insquare2.WhileCourt1hasthreerepresentativepoints(P1,P2,P3),andCourt2hastwopoints(P4,P5).3.1.1ComparativeanalysisdatainselectedsquaresThereisasignificantdifferenceinT(a)onthesquarescale(Table1).Theairtemperatureinsquare1islowerthansquare2atalmostallhoursoftheday.Thehighestdifferencewasrecordedat10amandnoon,+3.17°Cand3.27°C,respectively.Indeed33.54°Cand34.75°Cwererecordedatsquare2against30.36°Cand31.46°Cinsquare1(Table1).TheresultalsoshowedthedifferenceinTmrtvalueswhereTmrtinsquare1waslowerthansquare2foralmostthehoursoftheday(10am,noon,6pmand8pm)(Fig2).Asignificantdifferencewasrecordedatnoonatapproximately3°CwhenTmrtinsquare2was21.96°Ccomparedtosquare1with19.34°C(Table1).RegardingPET,anoticeabledifferenceof3.3°Cwasrecordedatnoon;PETinsquare1was27.08°Cand30.38°Cinsquare2(Fig2).However,thelowestPETvalueswerenotedsimultaneously(8amand8pm)(Table1andFig2).Indeed20°Cand25°Cwererecordedatsquare2,and20°Cand24°Catsquare1(8amand8pm).3.1.2ComparativeanalysisdatainselectedCourtsTheresultshighlightedasignificantdifferenceincalibrateddataincomparisonoftwoCourts.ConsideringT(a),Court1ischaracterizedbylowertemperatureagainstCourt2foralmostthehoursoftheday(10am,noon,2pm,6pmand8pm)(Fig2),thetemperature'sdifferencesweremorenoticeableat10amvaryingbetween3.16°Cand3.8°Cbesidesnoon,whenthedifferenceinthetemperaturereached3.9°C(Table2).However,Court2recordedlowertemperature(20°C-21°C)at8amincomparisontoCourt1(24°C-25°C).TmrtinCourt1werelowerthanCourt2foralmosthoursoftheday(10am,noon,2pm,and8pm).Themostsignificantdifferencewasrecordedat2pmfor3.19°C.However,at8am.TmrtinCourt1(14°C)washigherthanCourt2(13°C)(Table2).RegardingPET,Court1recordedalowertemperaturethanCourt2forthemajorityofthetime.Thenoticeabledifferencewasrecordedat10am(2.08°C);however,theresultshighlightedthatbothCourtsreachedthelowestvaluesatthesametime(8amand8pm)(Table2andFig2).3.2AnalysisoftheexistingcorrelationsbetweenthedifferentthermalcomfortindicesTheresultsshowastrongcorrelationbetweenthermalcomfortindicesintheselectedoutdoorenvironments.Thepositivecorrelationsoutdoorenvironmentsarepresentedinthefollowingequations: T(a)=0.883PET+7.012whereR2=0.96(1)T(a)=0.831Tmrt+13.61whereR2=0.752(2)PET=0.990Tmrt+6.489whereR2=0.865.(3)Equation(1)showsastrongpositivecorrelationbetweenT(a)andPET,whereR2isequivalentto0.96(Fig3).Equation(2)andequation(3)alsodefinedastrongconnectionbetweenT(a)andTmrtaswellasPETandTmrtwhereR2isequalto0.752and0.865,respectively(Fig4andFig5).3.3ImpactofgreencoveronthermalcomfortindicesCourt1andSquare1arecharacterizedbyadensegreencoverwith81%and84%,respectively,makingitsuitabletohaveatree-shadedareaduringallhoursoftheday.ThosespacesarealsocharacterizedbyaminimumvalueofSVF(0.35-0.45inCourt1and0.35insquare1)(Table3),whileCourt2hasonly43%ofgreencover,whichimpliesamediumshadeinparthavingtrees(SVFinthegreenareaisTimeSquare1Square2T(a)TmrtPETT(a)TmrtPET8am.25.18 °C 14.07°C 20.5°C 26.2°C 13.02°C 19.62°C 10am.30.36°C 18.73°C 25.9°C 33.5°C 20.06°C 27.12°C Noon.31.46°C 19.34°C 27.1°C 34.7°C 21.96°C 30.38°C 2pm.33.48°C 22.38°C 29.8°C 31.8°C 20.82°C 27.68°C 4pm.34.73°C 25.09°C 31.2°C 33.2°C 24.06°C 29.3°C 6pm.31.25°C 23.32°C 27.9°C 31.8°C 23.75°C 28.4°C 8pm.27.57°C 19.81°C 23.9°C 28.8°C 20.26°C 24.6°C Tab.1–Calibrated data and PET values in selected squares 34FOCUS equivalentto0.48)(Table3).However,square2 hasnogreencover,andtheskyviewfactorisequalto0.97,sothespaceisopen.SelectedurbanspacesGreencoverarea(m2)GreencoverSVFbinthegreenareaCourt111145.6681%0.35-0.45Court21995.8243%0.48Square12923.5584%0.35Square2--- Tab.3 – Green cover percentage and (SVF) in the selected area. SVF b Sky View Factor is calculated using RayMan Software. 4.DiscussionandConclusionThepresentedworkisacomparativeresearchoffouroutdoorurbanspaceshavingthesamemorphology(TwosquaresandtwoCourts)within(Csa)andcharacterizedbyHDD=1200-1800,CDD=1100-1700.Thisstudyisbasedonmicroclimaticmeasurements,calibrateddata(Envi-met),andcalculatingPETusingRaymansoftware.Thisstudyvalidatedsimulatedresultsbycomparingthemeasuredandmodelledairtemperaturevalue[44],Usingthemodelledapproach[44,45]allowedcurrentPETresults.4.1TheeffectofgreencoveronmicroclimaticparametersandPET4.1.1.AirtemperatureTheresultshighlightednoticeabledifferencesinT(a)atnoon(34.43°Cforsquare2and31.46°Cforsquare1).Inotherterms,adifferenceof3.27°Cinairtemperaturerelatedtothehottesthoursoftheday.Leeetal.[41]definedshadingbytreescanopiesasacrucialfactor-basedahuman–biometeorologicalperspective[46]inurbanopenspaces.Theresultsalsoshowedasignificantdifferenceinairtemperature(3.9°C)attheCourtslevelduringthehottesthoursoftheday(noon).Court1ischaracterizedbyadensegreencover(81%),withanimportantcrowndimensionoftreesthanCourt2(43%),wherethetreeshaveasmallercrown.Urbanspaceswithdensevegetationcoverconsiderablyaffectcoolingeffectandthermalcomfort[47,48].Sky–viewfactor(SVF)ortheareaenvelopedbytreecanopy[49],whichtheminimumvalueswererelatedtotreescharacteristicssuchastreestype,sizecrownandthelocationofgreenTimeCourt1Court2T(a)(°C)Tmrt(°C)PET(°C)T(a)(°C)Tmrt(°C)PET(°C)8am.24.56-25.09 14.11 20.12-20.6 20.13-20.98 12.90 19.76-19.82 10am.29.68-30.26 18.58-18.74 24.6-25.7 33.42-33.48 19.62-19.84 26.7-26.68 Noon.31.46-31.63 19.22-19.57 26.77-27.26 34.45-34.6 21.62-21.8 29.78-28.92 2pm.33.39-33.61 21.71-22.34 29.53-29.75 33.39-34.2 24.06-24.91 29.7-29.5 4pm.34.69-34.77 24.61-25.07 31.32-31.6 33-34.69 23.81-23.93 29.08 6pm.31.2-31.28 23.27-23.34 28.07-28.33 31.2-31.52 23.01-23.54 28 8pm.27.5-27.64 19.74-19.9 24-24.46 27.52-29.33 20.10-20.18 24.72-24.94 Tab.2 – Calibrated data and PET values in selected Courts Fig.3– Correlation between T(a) and PET Fig. 2 – Comparison of calibrated data (T(a),Tmrt and PET in squares and courts a:Squarescale,b:Courtscale.Fig.4 – Correlation between T(a) and Tmrt Fig.5 – Correlation between PET and Tmrt 35 FOCUS cover,affectthelocalmicroclimate[44]andespeciallyduringthehothoursoftheday.4.1.2Meanradianttemperature(Tmrt)andthermalcomfortIndex(PET)TheresultshighlightedaconsiderabledifferenceinCourts(3.19°C)andSquares(2.62°C)duringthehottesthoursofthedays(noonand2pm).ThelowestvaluesofTmrtwererelatedtotreesshadedarea[50,51].However,thesunnyarearecordedalwaysahighTmrt(Fig2).Thenumberoftrees,crownwidth,anddimensionssignificantlyreduceTmrt[29,46,51,52].TheresultsshowedthatPETincourt1andsquare1arelowerthansquare2andcourt2foralmosthoursoftheday.Thedifferencesrecordedinsquareswereequivalentto3.3°Cand2.08°Catnoon,and10am,whichemphasizestheroleoftreescanopyinimprovingthermalcomfortduringthehottesthoursoftheday[29,46,51,52].TheresultsshowedthattheminimalvaluesofPETvarybetween20°C-25°Cat8amand8pmCohenetal.[7]defined20≤PET≤26asthethermalcomfortaverageinTelAvivwiththeMediterraneanclimate(Csa).Inthisregard,wecanconfirmthatsquare1andCourt2arethermallycomfortablecomparedtosquare2andCourt2foralmosthoursoftheday.Adensegreencoverinoutdoorurbanspacesimprovesthermalcomfortbyreducingairtemperatureandmeanradianttemperature;thevaluescanreach4°CforT(a)and3°CforTmrtduringthehottesthour'sdayinsummer.ThisstudycomparesfouroutdoorspaceswithsimilarmorphologyanddifferentgreencoverconfigurationinAnnabacity,Algeria.Thecalibrationprocessallowedaccurateresults.Thefindingshighlighteddifferencesinmicroclimaticparameters(airtemperatureandMeanradianttemperature)duringthehoursoftheday.However,treesandtheirsizemayreducethehumanthermalstressbymorethan3°C,improvingsignificantlythermalcomfortrangesduringthehottesthoursofthedayinsummer.Thisstudy'sresultscanbeappliedinCsaClimate(HDD=1200-1800,CDD=1100-1700).Significantfactorstobeconsideredarethetypeoftrees,size(widthofthecrown),thegreenspatialconfigurationofoutdoorenvironments,andthesurroundingbuildings'heights.Finally,Itwillbenecessarytoconsidertheusers'thermalsensationintheoutdoorurbanspaceastheyarethemostexposedtotheheat.Usingaquestionnairesurveyandcalibratedresultssuchasairtemperature,TmrtbesidesPETasathermalcomfortindexcouldinterestfutureresearch.Moreover,comparingtheoutdoorthermalcomfortrangeintheMediterraneanareaallowedustoidentifyaprecisethermalcomfortrange.REFERENCES[1] Givoni,B,Noguchi,M.,Saaroni,H.,Pochter,O.,Yaacov,Y,andFeller,N,'Outdoorcomfortresearchissues.',EnergyandBuildings,vol.1,pp.77–86,2003.[2]I.KnezandS.Thorsson,'Influencesofcultureandenvironmentalattitudeonthermal,emotionalandperceptualevaluationsofapublicsquare',IntJBiometeorol,vol.50,no.5,pp.258–268,May2006,doi:10.1007/s00484-006-0024-0.[3]M.Nikolopoulou,N.Baker,andK.Steemers,'Thermalcomfortinoutdoorurbanspaces:understandingthehumanparameter',SolarEnergy,vol.70,no.3,pp.227–235,2001,doi:10.1016/S0038-092X(00)00093-1.[4]M.NikolopoulouandS.Lykoudis,'Thermalcomfortinoutdoorurbanspaces:AnalysisacrossdifferentEuropeancountries',BuildingandEnvironment,vol.41,no.11,pp.1455–1470,Nov.2006,doi:10.1016/j.buildenv.2005.05.031.[5]O.Potchter,P.Cohen,T.-P.Lin,andA.Matzarakis,'Outdoorhumanthermalperceptioninvariousclimates:Acomprehensivereviewofapproaches,methodsandquantification',ScienceofTheTotalEnvironment,vol.631–632,pp.390–406,Aug.2018,doi:10.1016/j.scitotenv.2018.02.276.[6]J.SpagnoloandR.deDear,'Afieldstudyofthermalcomfortinoutdoorandsemi-outdoorenvironmentsinsubtropicalSydneyAustralia',BuildingandEnvironment,vol.38,no.5,pp.721–738,May2003,doi:10.1016/S0360-1323(02)00209-3.[7]P.Cohen,O.Potchter,andA.Matzarakis,'HumanthermalperceptionofCoastalMediterraneanoutdoorurbanenvironments',AppliedGeography,vol.37,pp.1–10,Feb.2013,doi:10.1016/j.apgeog.2012.11.001.[8]M.H.Elnabawi,N.Hamza,andS.Dudek,'ThermalperceptionofoutdoorurbanspacesinthehotaridregionofCairo,Egypt',SustainableCitiesandSociety,vol.22,pp.136–145,Apr.2016,doi:10.1016/j.scs.2016.02.005.[9]R.-L.Hwang,T.-P.Lin,andA.Matzarakis,'Seasonaleffectsofurbanstreetshadingonlong-termoutdoorthermalcomfort',BuildingandEnvironment,vol.46,no.4,pp.863–870,Apr.2011,doi:10.1016/j.buildenv.2010.10.017.[10] N.Kántor,L.Égerházi,andJ.Unger,'Subjectiveestimationofthermalenvironmentinrecreationalurbanspaces—Part1:investigationsinSzeged,Hungary',IntJBiometeorol,vol.56,no.6,pp.1075–1088,Nov.2012,doi:10.1007/s00484-012-0523-0.[11] D.Lai,D.Guo,Y.Hou,C.Lin,andQ.Chen,'StudiesofoutdoorthermalcomfortinnorthernChina',BuildingandEnvironment,vol.77,pp.110–118,Jul.2014,doi:10.1016/j.buildenv.2014.03.026.[12] A.Tseliou,I.X.Tsiros,S.Lykoudis,andM.Nikolopoulou,'Anevaluationofthreebiometeorologicalindicesforhumanthermalcomfortinurbanoutdoorareasunderrealclimaticconditions',BuildingandEnvironment,vol.45,no.5,pp.1346–1352,May2010,doi:10.1016/j.buildenv.2009.11.009.[13] ASHRAEStandard,'AmericanSocietyofHeating,RefrigeratingandAir-ConditioningEngineers',vol.55,no.2004,2004.[14] J.A.AceroandK.Herranz-Pascual,'Acomparisonofthermalcomfortconditionsinfoururbanspacesbymeansofmeasurementsandmodellingtechniques',BuildingandEnvironment,vol.93,pp.245–257,Nov.2015,doi:10.1016/j.buildenv.2015.06.028.[15] S.Watanabe,K.Nagano,J.Ishii,andT.Horikoshi,'Evaluationofoutdoorthermalcomfortinsunlight,buildingshade,andpergolashadeduringsummerinahumidsubtropicalregion',BuildingandEnvironment,vol.82,pp.556–565,Dec.2014,doi:10.1016/j.buildenv.2014.10.002.[16] H.Staiger,G.Laschewski,andA.Grätz,'Theperceivedtemperature–aversatileindexfortheassessmentofthehumanthermalenvironment.PartA:scientificbasics',IntJBiometeorol,vol.56,no.1,pp.165–176,Jan.2012,doi:10.1007/s00484-011-0409-6.[17] R.R.Gonzalez,Y.Nishi,andA.P.Gagge,'Experimentalevaluationofstandardeffectivetemperatureanewbiometeorologicalindexofman'sthermaldiscomfort',IntJBiometeorol,vol.18,no.1,pp.1–15,Mar.1974,doi:10.1007/BF01450660.[18] Pickup,JanddeDear,R,'Anoutdoorthermalcomfortindex(OUT_SET*)-partI-themodelanditsassumptions.InBiometeorologyandurbanclimatologyattheturnofthemillenium.',2000,vol.99,pp.279–283.[19] Jendritzky,G,deDear,R,andHavenith,G,'UTCIwhyanotherthermalindex?',InternationalJournalofBiometeorology,vol.3,pp.421–428,2012.[20] A.Matzarakis,H.Mayer,andM.G.Iziomon,'Applicationsofauniversalthermalindex:physiologicalequivalenttemperature',InternationalJournalofBiometeorology,vol.43,no.2,pp.76–84,Oct.1999,doi:10.1007/s004840050119.[21] H.MayerandP.Höppe,'Thermalcomfortofmanindifferenturbanenvironments',TheorApplClimatol,vol.38,no.1,pp.43–49,1987,doi:10.1007/BF00866252.[22] S.Berkovic,A.Yezioro,andA.Bitan,'Studyofthermalcomfortincourtyardsinahotaridclimate',SolarEnergy,vol.86,no.5,pp.1173–1186,May2012,doi:10.1016/j.solener.2012.01.010.[23] C.R.deFreitasandE.A.Grigorieva,'Acomprehensivecatalogueandclassificationofhumanthermalclimateindices',IntJBiometeorol,vol.59,no.1,pp.109–120,Jan.2015,doi:10.1007/s00484-014-0819-3.[24] D.Lai,C.Zhou,J.Huang,Y.Jiang,Z.Long,andQ.Chen,'Outdoorspacequality:AfieldstudyinanurbanresidentialcommunityincentralChina',EnergyandBuildings,vol.68,pp.713–720,Jan.2014,doi:10.1016/j.enbuild.2013.02.051.[25] E.Johansson,S.Thorsson,R.Emmanuel,andE.Krüger,'Instrumentsandmethodsinoutdoorthermalcomfortstudies–Theneedforstandardization',UrbanClimate,vol.10,pp.346–366,Dec.2014,doi:10.1016/j.uclim.2013.12.002.[26] A.MatzarakisandH.Mayer,'AnotherKindofEnvironmentalStress:ThermalStress.WHOCollaboratingCentreforAirQualityManagementandAirPollutionControl',NEWSLETTERS,vol.18,pp.7–10,1996.[27] Á.Gulyás,J.Unger,andA.Matzarakis,'Assessmentofthemicroclimaticandhumancomfortconditionsinacomplexurbanenvironment:Modellingandmeasurements',BuildingandEnvironment,vol.41,no.12,pp.1713–1722,Dec.2006,doi:10.1016/j.buildenv.2005.07.001.[28] S.Thorsson,T.Honjo,F.Lindberg,I.Eliasson,andE.-M.Lim,'ThermalComfortandOutdoorActivityinJapaneseUrbanPublicPlaces',EnvironmentandBehavior,vol.39,no.5,pp.660–684,Sep.2007,doi:10.1177/0013916506294937.[29] F.Ali-ToudertandH.Mayer,'Thermalcomfortinaneast–westorientedstreetcanyoninFreiburg(Germany)underhotsummerconditions',Theor.Appl.Climatol.,vol.87,no.1–4,pp.223–237,Jan.2007,doi:10.1007/s00704-005-0194-4.[30] H.Andrade,M.-J.Alcoforado,andS.Oliveira,'Perceptionoftemperatureandwindbyusersofpublicoutdoorspaces:relationshipswithweatherparametersandpersonalcharacteristics',IntJBiometeorol,vol.55,no.5,pp.665–680,Sep.2011,doi:10.1007/s00484-010-0379-0.[31] I.Charalampopoulos,I.Tsiros,A.Chronopoulou-Sereli,andA.Matzarakis,'AnalysisofthermalbioclimateinvariousurbanconfigurationsinAthens,Greece',UrbanEcosyst,vol.16,no.2,pp.217–233,Jun.2013,doi:10.1007/s11252-012-0252-5.[32] T.-P.Lin,K.-T.Tsai,C.-C.Liao,andY.-C.Huang,'Effectsofthermalcomfortandadaptationonparkattendanceregardingdifferentshadinglevelsandactivitytypes',BuildingandEnvironment,vol.59,pp.599–611,Jan.2013,doi:10.1016/j.buildenv.2012.10.005.36FOCUS [33] T.-P.LinandA.Matzarakis,'TourismclimateandthermalcomfortinSunMoonLake,Taiwan',IntJBiometeorol,vol.52,no.4,pp.281–290,Mar.2008,doi:10.1007/s00484-007-0122-7.[34] A.Matzarakis,F.Rutz,andH.Mayer,'Modellingradiationfluxesinsimpleandcomplexenvironments—applicationoftheRayManmodel',IntJBiometeorol,vol.51,no.4,pp.323–334,Feb.2007,doi:10.1007/s00484-006-0061-8.[35] L.ChenandE.Ng,'Outdoorthermalcomfortandoutdooractivities:Areviewofresearchinthepastdecade',Cities,vol.29,no.2,pp.118–125,Apr.2012,doi:10.1016/j.cities.2011.08.006.[36] D.Armson,P.Stringer,andR.Ennos,'Theeffectoftreeshadeandgrassonsurfaceandglobetemperaturesinanurbanarea',UrbanForestry&UrbanGreening,vol.11,pp.245–255,Jan.2012,doi:10.1016/j.ufug.2012.05.002.[37] F.Kong,H.Yin,C.Wang,G.Cavan,andP.James,'Asatelliteimage-basedanalysisoffactorscontributingtothegreen-spacecoolislandintensityonacityscale',UrbanForestry&UrbanGreening,vol.13,no.4,pp.846–853,2014,doi:10.1016/j.ufug.2014.09.009.[38] N.Weber,D.Haase,andU.Franck,'ZoomingintotemperatureconditionsinthecityofLeipzig:Howdourbanbuiltandgreenstructuresinfluenceearthsurfacetemperaturesinthecity?',ScienceofTheTotalEnvironment,vol.496,pp.289–298,Oct.2014,doi:10.1016/j.scitotenv.2014.06.144.[39] E.Johansson,'Influenceofurbangeometryonoutdoorthermalcomfortinahotdryclimate:AstudyinFez,Morocco',BuildingandEnvironment,vol.41,no.10,pp.1326–1338,Oct.2006,doi:10.1016/j.buildenv.2005.05.022.[40] Köppen,‘Grundrissderklimakunde’,1931.[41] H.Lee,H.Mayer,andL.Chen,'ContributionoftreesandgrasslandstothemitigationofhumanheatstressinaresidentialdistrictofFreiburg,SouthwestGermany',LandscapeandUrbanPlanning,vol.148,pp.37–50,Apr.2016,doi:10.1016/j.landurbplan.2015.12.004.[42] M.Taleghani,D.J.Sailor,M.Tenpierik,andA.vandenDobbelsteen,'Thermalassessmentofheatmitigationstrategies:ThecaseofPortlandStateUniversity,Oregon,USA',BuildingandEnvironment,vol.73,pp.138–150,Mar.2014,doi:10.1016/j.buildenv.2013.12.006.[43] Z.WuandL.Chen,'Optimizingthespatialarrangementoftreesinresidentialneighborhoodsforbettercoolingeffects:Integratingmodelingwithin-situmeasurements',LandscapeandUrbanPlanning,vol.167,pp.463–472,Nov.2017,doi:10.1016/j.landurbplan.2017.07.015.[44] C.Skelhorn,S.Lindley,andG.Levermore,'Theimpactofvegetationtypesonairandsurfacetemperaturesinatemperatecity:AfinescaleassessmentinManchester,UK',LandscapeandUrbanPlanning,vol.121,pp.129–140,Jan.2014,doi:10.1016/j.landurbplan.2013.09.012.[45] S.E.Gill,J.F.Handley,A.R.Ennos,andS.Pauleit,'AdaptingCitiesforClimateChange:TheRoleoftheGreenInfrastructure',builtenviron,vol.33,no.1,pp.115–133,Mar.2007,doi:10.2148/benv.33.1.115.[46] H.Lee,J.Holst,andH.Mayer,'ModificationofHuman-BiometeorologicallySignificantRadiantFluxDensitiesbyShadingasLocalMethodtoMitigateHeatStressinSummerwithinUrbanStreetCanyons',AdvancesinMeteorology,vol.2013,pp.1–13,2013,doi:10.1155/2013/312572.[47] F.Aram,E.HiguerasGarcía,E.Solgi,andS.Mansournia,‘Urbangreenspacecoolingeffectincities’,Heliyon,vol.5,no.4,p.e01339,Apr.2019,doi:10.1016/j.heliyon.2019.e01339.[48] P.Cohen,O.Potchter,andA.Matzarakis,'DailyandseasonalclimaticconditionsofgreenurbanopenspacesintheMediterraneanclimateandtheirimpactonhumancomfort',BuildingandEnvironment,vol.51,pp.285–295,May2012,doi:10.1016/j.buildenv.2011.11.020.[49] L.Shashua-BarandM.Hoffman,'Vegetationasaclimaticcomponentinthedesignofanurbanstreet:Anempiricalmodelforpredictingthecoolingeffectofurbangreenareaswithtrees',EnergyandBuildings,vol.31,pp.221–235,Apr.2000.[50] M.FahmyandS.Sharples,'OnthedevelopmentofanurbanpassivethermalcomfortsysteminCairo,Egypt',BuildingandEnvironment,vol.44,no.9,pp.1907–1916,Sep.2009,doi:10.1016/j.buildenv.2009.01.010.[51] H.Lee,H.Mayer,andL.Chen,'ContributionoftreesandgrasslandstothemitigationofhumanheatstressinaresidentialdistrictofFreiburg,SouthwestGermany',LandscapeandUrbanPlanning,vol.148,pp.37–50,Apr.2016,doi: 10.1016/j.landurbplan.2015.12.004.[52] M.TaleghaniandU.Berardi,'Theeffectofpavementcharacteristicsonpedestrians'thermalcomfortinToronto',UrbanClimate,vol.24,pp.449–459,Jun.2018,doi:10.1016/j.uclim.2017.05.007. UNEÉTUDECOMPARATIVEDUCONFORTTHERMIQUEÀL'EXTÉRIEURDANSLECLIMATMÉDITERRANÉENRésuméDenombreusesétudesontexploréleconfortthermiqueextérieurdansdifférentsclimatsafindecomprendresonimpactsurlaqualitédevieetl’utulisationdesespacespublics.Cetteétudeproposedemesurerleconfortthermiqueparlatempératurephysiologiqueéquivalente(PET),àl’aidedel’outilRayMan,suivantunprocessusdecalibrationréaliséparlelogicielEnvi-met,etvalidéeàAnnaba,Algérie.UnevilleméditerranéennedontDegré-jourdechauffeDJC=1200-1800etDegré-jourfroidDJF=1100-170.Quatreespacesextérieurs(2courset2places)ontétéévaluésenété2017.Chaqueespaceurbainétaitdiviséenunegrille(4m*4m),oùsesituentlespointsdemesure.Lesrésultatsontmontréunefortecorrélationentrelatempératuredel’airT(a)etPET.Lesarbresoptimisentl’effetderefroidissement,etcontribuentàl’améliorationduconfortthermiqueenparticulierpendantlesheureschaudesdelajournée.Eneffet,lesdifferencesdetemperaturess'avèrenttrèsimportante,4°CpourlaT(a)et3°Cpourlatemepratureradiante(Tmrt)àmidi,unedesheureslespluschaudesdelajournée.Motsclés: ClimatMediteranian(Csa),Degré-jourdechauffe(DJC),Degré-jourfroidDJF,TempératurePhysiologiqueEquivalente(PET),Envi-met,RayMan1.IntroductionCesdernièresannées,leconfortthermiqueaconnuuneattentionparticulièredanslesespacesurbainsafind’améliorerlasantéetlebienêtredescitoyensselondenouvellesstratégiesdurables[1-6].Denombreusesétudesonttentédedéfinirlesconditionsduconfortthermiqueextérieur[2,7–12],considérécommeleniveaudesatisfactiondesusagerspourl’environnementthermique[5,13].Ceciditl’environnementmicroclimatiquesauneinfluencesurlesactivitéextérieuresdescitoyens[7].SelonAceroandHerranz-Pascal[14]latempératureradiantemoyenne(Tmrt),l’humiditérelative,latempératuredel’air,lavitesseduventsontnécessairespourcalculerleconfortthermiqueextérieur,enplusdutauxmétaboliqueetdel'isolationdesvêtements[15].ENVI-metandRaymansontdeslogicielsapprouvéspourlecalculeduconfortthermiqueextérieurENVI-metestbasésurladynamiquedesfluidesnumérique(CFD)etlathermodynamique,tandisqueRayManestunmodèlederayonnementtridimensionnel.Beaucoupd’étudessesontfocaliséessurledéveloppementd’indicesuniverselscapablesdemesurerleconfortthermiquedanslesclimatschaudsetfroids[5–7].Denombreuxindicesontétéappliquésavecsuccèsenmilieuurbain(ex:latempératureperçue(Pt)[5,16],latempératureeffectivestandard(SET*)[5,7]etsavarianteextérieure,OUT-SET*[18].L'indiceuniverseldeclimatthermique(UTCI[19],latempératurephysiologiquementéquivalente(PET)[20,21]etlavoixmoyenneprédite(PMV)[5,22–24].PETaétéinitiéenEuropeoccidentaleetcentrale[7,8,26].Ilaététestéetvalidédansplusieurszonesclimatiques[20,25,27,28],etexplorédansdifférentsenvironnementsextérieurs[2,11,28-34].Lesconditionsmicroclimatiquesontunegrandeinfluencesurlesactivitésenpleinair.Unenvironnementensoleilléouombragéaffectedemanièresignificativeladuréedetempsdespersonnes,lesincitantsàresterouàpartir[35].Lesespacesvertsurbainsproduisentprincipalementdel'ombredirecteetunrefroidissementparévapotranspiration,minimisantlatempératuredel'airetdelasurfaceparlacréationd’unrefroidissementlocal[36–38].Denombreusesétudesonttentéd'obtenirundiscernementcompletdeladifférenceentreespacessoleillésetombragés.Matzarakisetal.[20]ontenregistré15°CdedifférencedansPETentrelesdeuxzones(ensoleilléeetlazoneombragéepardesarbres).Johansson[39]aétudiél'impactdelagéométrieurbainesurleconfortthermiqueextérieurdansleclimatchaudetsec.Lesrésultatsontmontréunedifférence(20°C)enPET,lescanyonsurbainsombragésetprofondssontplusfraisquelescanyonspeuprofonds.Enraisondelafaibletempératuredel'airetdelatempératureradianteMoyenne[15].Surlabasedelalittératureactuelle,peudetravauxontexploréleconfortthermiquedansleclimatCsaparlacombinaisondedeuxlogiciels(Envi-metetRayman),dontlerésultatestvalidépardesmesuresinsitu.Laméthodeprésentéeapermisl’acquisitiondevaleursplusprécisesduPET,encombinantdeuxlogiciels(Envi-metetRayman)quisontvalidéespardesmesuresin-situdansleclimatdelaCsa.Cetteméthodeconcernel'ensembledelazonedesespacesurbainsetnonseulementcertainspointsdemesure.L'effetdelacouverturevégétalesurlesparamètresmicroclimatiquesetl'indicedeconfortthermiqueaétédéfinienfonctiondesheuresdelajournée.Cetteétudeviseà(1)évaluerleconfortthermiquedansquatreespacesurbainscaractérisésparunevariétédecouverturevégétaledansleclimatméditerranéen.(2)Analyserl'impactdesespacesvertssurlatempératuredel'air,etlatempératuremoyenneradiante(Tmrt)etTempératurephysiologiqueéquivalente(PET),(3)explorerlescritèresdedegrés-joursdechauffageetdedegrés-joursderefroidissementprésententunevaleurajoutéeparrapportauxétudesprécédentes.2.MéthodeLaméthodeproposéeestappliquéedanslavilled’Annaba(Algérie),caractériséeparleclimatMéditerranéen(Csa)selonlaclassificationdeKöppen[40],ainsiqueHDD=1200-1800,CDD=1100-1700.L’objetdecetteétudeportesurlacomparaisonde4espacespublics(2courtset2places)àAnnabacentreetElBouni(banlieued’Annaba).Ceciditlaméthodologiecomprends:Collectedesdonnées,TraitementdesdonnéesetenfinAnalyseetsynthèsedesdonnées.2.1Collectededonnées:Chaqueespaceaétédiviséenunegrille(4m*4m).Lespointsdemesureontétésélectionnéssurcettegrille,(10pointsdemesuredanschaqueCourtet5danschaqueplace).Nousavonsutilisél’appareilLM8000(Thermo-Anémomètre,37 FOCUS Hygromètre,ThermomètreetIlluminometre),àunehauteurde1.10mdusol,pourmesurercontinuellementlatempératuredel’air,l’humiditérelativeetlavitesseduvent,durantdeuxjournéesd’étéde8hà20h(chaquedeuxheures).2.2Traitementdesdonnées:NousavonsutilisélemodèleEnvi-metàhauterésolutionspécialeettemporelle,quiaétélargementutiliséetvalidédansl’évaluationdesenvironnementsbâtis.LemodelsimuléaétévalidéparrapportàlaT(a)mesurée,mettantenévidenceunebonnecorrélationentrelesdeuxdonnées.Cequiapermisl’obtentionde04donnéescalibrées(T(a),vitesseduvent,humiditérelativeetTmrt)pourlatotalitédesespacessélectionnés.CetteétudeaaussiadoptélelogicielRayManpourlecalculeduPET,basésurT(a),humiditérelative,vitesseduvent,Tmrt,lesvetementshumainsetactivités[7,32,34].Raymanautilisélesdonnéescalibréescommedonnéesd’entréepourlecalculdePETainsiquel’évaluationduFacteurdevuedeciel(SVF).2.3Analysedesdonnées:Cetteétapeestliéeàlasection«Résultats»,oùlesdonnéescalibréestellesqueT(a),TmrtainsiquePETontétéanalysées.Fig.1résumelaméthodeappliquéeàl’échelledesespacesurbainspourl’évaluationduconfortthermique.3.Casd’étudeCetteétudeaétémenéeàAnnaba,Algérie,situéeentre(Lattitude36°54'Nord,Longitude:7°46'Est,Niveaudelamer:+5m).Annabaestcaractériséeparleclimatméditerranéenchaudd'été(Csa),selonlaclassificationdeKöppen[40].Lecours1etlaplace1setrouventdanslecentrevilled'Annaba,ilssontcaractérisesparUnecouverturevégétaledense.Enrevanche,laplace2etlecours2sontsituésdanslazonesuburbained'Annaba(ElBouni).Laplace2estdépourvued'arbres,tandisquelecours2comprenddeuxparties,lapremièrecontientdesarbresetlasecondeestsanscouverturevégétale(Fig1).3.1ResultatsCettesectionprésented'aborduneanalysecomparativedesdonnéescalibréesdanslesespacesurbainssélectionnés,vuleurimpactsignificatifsurleconfortthermiqueextérieur.PETestanalyséendétailenfonctiondesdonnéesmicroclimatiquesetdel'heuredelajournée.Ladernièrepartiedesrésultatscomprendl’effetdelacouverturevégétalesurlaPET,T(a)etTmrt.LesvaleursPETsontassezprochesàcertainspointsdemesure.Avoirunevaleurmoyennedesdonnéesmicroclimatiquesrelativessurchaqueespaceurbainpermetd'obtenirdemeilleursrésultatsd'analyse.Chaqueplaceaunpointreprésentatif,P1danslaplace1etP1'danslaplace2.Alorsquelecours1atroispointsreprésentatifs(P1,P2,P3),etlecours2disposededeuxpoints(P4,P5).3.1.1Analysecomparativedesdonnéesdanslazonesélectionnée.LesrésultatsontmontréunedifférencesignificativedeT(a)àl'échelledesplaces(Tableau1).T(a)danslaplace1estinférieure(T(a)delaplace2pourlamajoritédesheuresdelajournée.Ladifférencelaplusimportanteaétéenregistréeà10hetàmidi,(+3,17°Cet3,27°C)respectivement.LesrésultatsontégalementmontréladifférencedanslesvaleursdeTmrt,oùlesTmrtdelaplace1étaientinférieursàceuxdelaplace2pourlamajoritédesheuresdelajournée(10h,12h,18het20h)(Fig2).LesvaleursdeTmrtdanscours1étaientinférieursàceuxdecours2pourlapluspartdesheuresdelajournée(10h,midi,14het20h).Ladifférencelaplussignificativeaétéenregistréeà14hpour3,19°C.Cependant,à8h:00Tmrtaucours1(14°C)étaitplusélevéqueceluidelacours2(13°C)(Tableau2).EncequiconcernelaPET,lecours1aenregistréunetempératureplusbassequecours2pourlapluspartdutemps,ladifférencelaplusnotableaétéenregistréeà10heures(2,08°C).Cependant,lesrésultatsontmisenévidencequelestroisvariablesclimatiquesontatteintlesvaleurslesplusbassesàl’échelledesquatreespacesaumêmemoment(8h:00et20h:00)(Tableau2etFig2).3.2AnalysedescorrélationsexistantesentrelesdifférentsindicesdeconfortthermiqueLesrésultatsontmontréunefortecorrélationentrelesindicesdeconfortthermiquedanslesespacesurbainssélectionnés.Cescorrélationspositivessontprésentéesdansleséquationssuivantes:T(a)=0,883PET+7,012oùR2=0,96(1)T(a)=0,831Tmrt+13,61oùR2=0,752(2)PET=0,990Tmrt+6,489oùR2=0,865.(3)L'équation(1)montreunefortecorrélationpositiveentreT(a)etPET,oùR2estéquivalentà0,96(Fig3).L'équation(2)etl'équation(3)ontégalementdéfinituneforteconnexionentreT(a)etTmrtainsiquePETetTmrtoùR2estégalà0,752et0,865,respectivement(Fig4etFig5).3.3ImpactdelacouverturevégétalesurlesindicesdeconfortthermiqueLecours1etleplace1sontcaractérisésparunecouverturevégétaledenseavec81%et84%,respectivement,cequipermetd'avoirunezoneombragéepardesarbresduranttoutedelajournée.CesespacessontégalementcaractérisésparunevaleurminimaledeSVF(0,35-0,45danslacour1et0,35danslaplace1)(Tableau3),tandisquelecours2n'aque43%decouvertureverte,cequiimpliqueuneombremoyennedanslapartieayantdesarbres(SVFdanslazoneverteestéquivalentà0,48)(Tableau3).Cependant,laplace2n'apasdecouvertureverte,etlefacteurdevueducielestégalà0,97,cequidéfinitunespacetotalementouvert.4.DiscussionetconclusionLetravailprésentéestunerecherchecomparativedequatreespacesurbainsayantlamêmemorphologie(Deuxplacesetdeuxcours)dansleclimatCsa,etcaractérisésparDJC=1200-1800,DJF=1100-1700.Cetteétudeestbaséesurdesmesuresmicroclimatiques,desdonnéescalibrées(Envi-met),etlecalculdelaPETàl'aidedulogicielRayman.Cetteétudeavalidélesrésultatssimulésencomparantlavaleurdelatempératuredel'airmesuréeetmodélisée[44].L'utilisationdel'approchemodélisée[44,45]apermisd'obtenirlesrésultatsplusprecisdelaPET.4.1L'effetdelacouverturevégétalesurlesparamètresmicroclimatiquesetPET4.1.1.Températuredel'airLesrésultatsontsoulignédesdifférencessignificativesdeT(a)àmidi(34,43°Cpourlaplace2et31,46°Cpourlaplace1).End'autrestermes,unedifférencede3,27°CdanslaT(a)serapportaitauxheureslespluschaudesdelajournée.Leeetal.[41]ontdéfinil'ombrageparlacanopéedesarbrescommeunfacteurcrucialbasésuruneperspectivebiométéorologiquehumainehumaine[46]danslesespacesouvertsurbains.Lesrésultatsontégalementmontréunedifférencenotabledelatempératuredel'air(3,9°C)auniveaudescourspendantlesheureslespluschaudesdelajournée(midi).Lecours1estcaractériséeparunecouverturevégétaledense(81%),avecunedimensionimportantedelacouronnedesarbresqueleCours2(43%),oùlesarbresontunecouronnepluspetite.Lesespacesurbainsavecunecouverturevégétaledenseaffectentconsidérablementl'effetderefroidissementetleconfortthermique[47,48].Lefacteurdevueduciel(SVF)relatifàlazoneenveloppéeparlacanopéedesarbres[49],dontlesvaleursminimalessontliéesauxcaractéristiquesdesarbrestellesqueletyped'arbre,latailledelacouronneetl'emplacementdelacouverturevégétale,affectelemicroclimatlocal[44]enparticulierpendantlesheureschaudesdelajournée.4.1.2Températureradiantemoyenne(Tmrt)etindicedeconfortthermique(PET)LesrésultatsontmisenévidenceunedifférenceconsidérableentrelesCours(3,19°C)etlesplaces(2,62°C)pendantlesheureslespluschaudesdelajournée(midiet14h).LesvaleurslesplusbassesdeTmrtétaientliéesàlazoneombragéeparlesarbres[50,51].Cependant,lazoneensoleilléeatoujoursenregistréuneTmrtélevé(Fig2).Lenombred'arbres,lalargeurdelacouronneetlesdimensionsréduisentsignificativementlaTmrt[29,46,51,52].LesrésultatsontmontréquelesvaleursdePETdanscours1etlaplace1sontinférieuresàcellesducours2etplace2pendantlamajoritédesheuresdelajournée.Lesdifférencesenregistréesdanslesplacesétaientéquivalentesà3,3°C(àmidi)et2,08°C(10h),cequisoulignelerôledelacanopéedesarbresdansl'améliorationduconfortthermiquependantlesheureslespluschaudesdelajournée[29,46,51,52].LesrésultatsontmontréquelesvaleursminimalesdePETvariententre20°C-25°Cà8het20hCohenetal.[7]ontdéfini20≤PET≤26commelamoyennedeconfortthermiqueàTelAvivayantunclimatméditerranéen(Csa).Decefait,nouspouvonsconfirmerquelaplace1etlecours1sontthermiquementplusconfortablesparquelaplace2etàlecours2pendantlamajoritédesheuresdelajournée.UnecouverturevégétaledensedanslesespacesurbainsextérieursamélioreleconfortthermiqueenréduisantlaT(a)etlaTmrt;..Ceciditlesvaleurspeuventatteindre4°CpourT(a)et3°CpourTmrtpendantlesheureslespluschaudesdelajournéeenété.Cetteétudeacomparéquatreespacesextérieursayantlamêmemorphologieavecuneconfigurationvégétaledifférentedanslavilled'Annaba,enAlgérie.Leprocessusdecalibrationapermisd'obtenirdesrésultatsprécis.Lesrésultatsontmisenévidencedesdifférencesdanslesparamètresmicroclimatiques(T(a)etTmrt)durantlesheuresdelajournée.Cependant,lesarbresetleurtaillepeuventréduirelestressthermiquehumaindeplusde3°C,améliorantdemanièresignificativelesplagesdeconfortthermiquependantlesheureslespluschaudesdelajournéeenété.LesrésultatsdecetteétudepeuventêtreappliquésauclimatCsaetdanstoutelazonecaractériséeparDJC=1200-1800,DJF=1100-1700.Lesfacteurssignificatifsàprendreencomptesontletyped'arbres,lataille(largeurdelacouronne),laconfigurationspatialevertedesenvironnementsextérieursetlahauteurdesbâtimentsenvironnants.Enfin,ilseranécessairedeprendreencomptelasensationthermiquedesusagersdansl'espaceurbainextérieur,carilssontlesplusexposésàlachaleur.L'utilisationd'uneenquêteparquestionnaireetlesrésultatscalibrés(ex:T(a),Tmrt)etPETpourraitfaireobjetdefuturesrecherches.View publication stats | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/356664782Bi-Fractal Characterization of the Pore Network of Tight SandstoneArticle in Frontiers in Earth Science · November 2021DOI: 10.3389/feart.2021.751944CITATIONS0READS787 authors, including:Some of the authors of this publication are also working on these related projects:National Natural Science Foundation for Young Scientists of China (41802148) View project2017YFC0603106 View projectZezhang SongChina University of Petroleum - Beijing23 PUBLICATIONS 108 CITATIONS SEE PROFILEZhao JunyiChina University of Petroleum - Beijing4 PUBLICATIONS 1 CITATION SEE PROFILEAll content following this page was uploaded by Zezhang Song on 01 December 2021.The user has requested enhancement of the downloaded file.Bi-Fractal Characterization of the PoreNetwork of Tight SandstoneZezhang Song 1,2, Junyi Zhao 1,2, Yuanyin Zhang 3*, Dailin Yang 4, Yunlong Wang 4, Li Hu 4,Chenghai Li 4 and Wenlei Liu 41State Key Laboratory of Petroleum Resources and Prospecting, China University of Petroleum, Beijing, China, 2College ofGeosciences, China University of Petroleum, Beijing, China, 3Oil and Gas Survey Center of China Geological Survey, Beijing,China, 4Southwest Oil and Gas Field Company PetroChina, Chengdu, ChinaFluid seepage performance and accumulation in tight sandstone is a critical research topicfor in-depth exploration and development, closely related to the heterogeneity of the porenetwork. The fractal characterization is one of the most compelling and direct ways forquantitative investigation of heterogeneity. However, only one kind of fractal is used in moststudies, and the differences and relations between different fractal dimensions are rarelydiscussed. This paper chose one of the most representative tight sandstone formations inChina, the second member of the Xujiahe Formation, as the research object. First, basedon physical analysis and XRD analysis, we carried out a qualitative investigation on porestructure utilizing thin-section and scanning electron microscopy. Then, detailed porestructure parameters were obtained using high-pressure mercury intrusion (HPMI). Lastly,we combined two-dimensional fractal analysis on thin-section images and three-dimensionalfractalanalysisonHPMIdatatocharacterizetheporenetworkheterogeneity quantitatively. The Xu2 tight sandstone is mainly medium- to fine-grainedlithic feldspathic sandstone or feldspathic lithic sandstone with low porosity andpermeability. Also, the Xujiahe tight sandstone is mainly composed of quartz, feldspar,and clay. The pore types of Xu2 tight sandstones are primarily intergranular pores, micro-fractures, and intra- and intergranular dissolution pores. Moreover, most of the micro-fractures in gas-bearing formation are open-ended, while most are filled by clay minerals inthe dry formation. The r50 (median pore radius) is the most sensitive parameter to seepagecapability (permeability) and gas-bearing status. The 2D fractal dimension (Ds) of gas-bearing samples is significantly larger than that of dry samples, while the 3D fractaldimension (D1, D2) of gas-bearing samples is lower than that of dry samples. There is astrong negative correlation between D2 and gas-bearing status, permeability, quartzcontent, and r50, but a positive correlation between Ds and these parameters. D2represents the heterogeneity of pore space, while the Ds indicates the development ofthe pore network. Tectonic movements that generate micro-fractures and claycementation that blocks the seepage channels are the two main controlling factors onfractal dimensions. Combining 2D and 3D fractal analysis could give a more in-depthinvestigation of pore structure.Keywords: bi-fractal analysis, pore structure, tight sandstone, Xujiahe formation, central Sichuan basinEdited by:Shang Xu,China University of Petroleum (EastChina), ChinaReviewed by:Zhe Cao,SINOPEC Petroleum Exploration andProduction Research Institute, ChinaBenben Ma,China University of GeosciencesWuhan, China*Correspondence:Yuanyin Zhangyuanyinshou@163.comSpecialty section:This article was submitted toEconomic Geology,a section of the journalFrontiers in Earth ScienceReceived: 02 August 2021Accepted: 01 November 2021Published: 30 November 2021Citation:Song Z, Zhao J, Zhang Y, Yang D,Wang Y, Hu L, Li C and Liu W (2021)Bi-Fractal Characterization of the PoreNetwork of Tight Sandstone.Front. Earth Sci. 9:751944.doi: 10.3389/feart.2021.751944Frontiers in Earth Science | www.frontiersin.orgNovember 2021 | Volume 9 | Article 7519441ORIGINAL RESEARCHpublished: 30 November 2021doi: 10.3389/feart.2021.7519441 INTRODUCTIONAfter about 10 years of rapid development of unconventionalhydrocarbon, tight gas plays a vital role in the energy frameworkof China. In 2019, the increases of the world proved gas reservesand gas demand were both driven by China, reaching as high as3 Tcm (trillion cubic meters) and 24 bcm (billion cubic meters),respectively [BP 2019 energy review, (Ersoy et al., 2019)].Compared with shale gas (Xu et al., 2020a; Xu et al., 2020b),tight sandstone gas has a more significant role in China’s naturalgas industry, accounting for 39.2 and 24.6% of China’s total gasreserves and annual natural gas production as early as 2010 (Daiet al., 2012), respectively. To date, tight sandstone gas is one of themost developed unconventional gas resources in the world.A typical tight sandstone gas reservoir is characterized by lowporosity(<10%),ultra-lowpermeability(<0.1 mDunderoverburden pressure, < 1 mD under atmospheric pressure),and strong reservoir heterogeneity (Zou et al., 2012). Thetightnessandsubstantialheterogeneitybringsignificantchallenges to the exploration and development of tight gas:challenging to predict the sweet spots of the reservoir, themain controlling factors for gas-bearing remains unclear(complex gas-water spatial distribution). Comprehensive andin-depth reservoir characterization is the core for successfullysolving the problems mentioned above in tight gas exploration.In recent years, the reservoir characterization of tightsandstoneemphasizedmoreonreservoirheterogeneitycharacterization. Fractal theory, first proposed by Mandelbrotand Wheeler (1983), has been widely utilized in the microscopicheterogeneity characterization for tight sandstone reservoirs.Fractal geometry has a unique advantage in the quantitativecharacterization of irregular objects: it can quantitativelycharacterize the heterogeneity of objects across scales by givingit a numerical value—the fractal dimension (Xie et al., 2001).Unlike Euclidean geometry, fractal dimension is not an integerbut a decimal. Generally, the greater the fractal dimension, themore substantial the roughness or heterogeneity of the porenetwork. Besides, we may further use fractal dimension forpredictingpetrophysicalpropertiessuchaspermeability,tortuosity, and even connectivity (Daigle et al., 2014).In the area of pore structure characterization, based on thevalue range of fractal dimension and experiments, we maycategorize it into two groups—two-dimensional and three-dimensional characterization. Accordingly, the range of fractaldimensions is (1, 2) and (2, 3), respectively, as summarized inTable 1.Two-dimensionalfractalanalysisisbasedonvisualobservation (image analysis), which consists of three steps: (1)obtaining 2D image through direct microscopic observation; (2)extracting and processing on pore structure characteristics; and(3) analyzing the fractal distribution of the pore network in 2Dspace. In addition, the box-counting method is the mostfrequently utilized method. The image for 2D fractal analysiscould be obtained from casting thin section, scanning electronmicroscopy (SEM) and low-field nuclear magnetic resonance(LF-NMR), and others. The casting thin section is the mostconvenientandcost-effectivetechniquetoobtainporestructure by filling the pore network with colorful epoxy resinunder vacuum. Sumantri and Permadi (2018) utilized the 2Dfractal dimension obtained from the casting thin section of theMenggala and Bekasap sandstone of the Miocene age from theCentral Sumatra Basin to predict permeability. It is worth notingthatthispaperhasnotshowntheporenetworkextraction—fractal analysis is directly carried out on thin-section images using the software. Compared with thin-sectionimages, SEM images could represent pore networks on a smallerscale.Forexample,Lianetal.(2004)carriedoutthemorphological decomposition of sandstone pore network onSEM images, then calculated the fractal dimension usingnumber-radius power law. However, the sampling process ofthe two techniques as mentioned above, especially that of theSEM, may significantly alter the pore network. As a moreTABLE 1 | Several examples for fractal characterization of pore structure of (tight) sandstone.DimensionExperimentsMethod/ModelFractaldimensionSamplesSources2DCasting Thin Section ImagesBox-counting(1.4382, 1.8120)Miocene sandstone of SumatrabasinSumantri and Permadi(2018)SEM ImagesNumber-radius power law(1.7640, 1.8261)SandstoneLian et al. (2004)LF-NMR (low-field nuclear magneticresonance) ImagesBox-counting(1.2214, 1.8648)Artificial sandstone coresWang et al. (2012)3DHPMI (high-pressure mercury intrusion)A universal capillary pressuremodel(2.251, 2.4983)The geysers rock (with fractures)Li and Horne (2003;2004)(Li, 2010) model(2.09, 2.44)Bashijiqike tight sandstonesLai and Wang (2015)LF-NMRAnalog of the mass fractalmodel(2.37, 2.41)Berea sandstoneDaigle et al. (2014)(2.6816, 2.9921)Permian tight sandstone of theOrdos basinShao et al. (2017)Spontaneous ImbibitionSpontaneous imbibition ratemodel2.60The Geysers rock (with fractures)Li and Zhao (2012)Frontiers in Earth Science | www.frontiersin.orgNovember 2021 | Volume 9 | Article 7519442Song et al.Bi-Fractal Characterisation of Pore Networknon-destructivetechnique,LF-NMRisalsocapableofrepresenting a pore network. Based on LF-NMR images, Wanget al. (2012) not only calculated the fractal dimension of sandstonebut also discussed the influencing factors (selected box sizes,threshold value, and image resolution) for fractal analysis.Also, it deserves to pay special attention that LF-NMR has itsown weakness brought by the technique itself. The image fromLF-NMR is based on NMR signal intensity from pore fluid. Thus,the LF-NMR image generally is with relatively low resolution andcan only represent pore space with scales between hundreds ofnanometers and micron-scales. Moreover, when it comes to tightsandstones, due to the low content of fluid in the pore network,the signal may be so weak that the LF-NMR image can hardlyreflect the pore space. Hence, by comparison, casting thin sectionis generally the first choice for 2D fractal analysis for tightsandstone owing to low-cost, high-resolution, and sufficientobservation scale for analysis.Three-dimensional fractal analysis is carried out through theinterpretation of fluid flow behavior in three-dimensional porespace. Thus, fluid intrusion techniques, like high-pressuremercuryintrusion(HPMI),LF-NMR,andspontaneousimbibition (SI), are frequently combined with fractal theory toanalyze the fractal characteristics of the pore network. Li andZhao (2012) established a fractal production model based on SIexperiments to estimate sandstone’s fractal dimension to predictproduction rate. Also, in recent years, we noticed more and moreworks applied fractal analysis on LF-NMR data, not only for thepore structure characterization of sandstone (Daigle et al., 2014;Zhao et al., 2017) but also for that of shale (Li et al., 2018; Sunet al., 2019; Yuan and Rezaee, 2019). Compared with the twotechniques mentioned above, the combination of HPMI/capillarypressure and fractal theory has a long history, and numerousmodels have been established for the fractal interpretation ofHPMI data. This is attributed to the convenience and relativelylow cost of the HPMI experiment compared with that of LF-NMRand SI. As early as the 1950s, Corey (1954) proposed a model todemonstrate the relationship between gas and oil relativepermeability, which revealed the power-law relations betweencapillary pressure and effective saturation. It is worth noting thatestablishing a power-law relation between two parameters iscrucial for determining fractal dimension and is regarded asone of the most critical steps for fractal analysis. Later, in theresearch of Thomeer (1960), the pore geometrical factor wasdirectly introduced to the exponent of the power-law betweencapillary pressure and fractional bulk volume occupied bymercury (could be related to saturation). Similarly, Brooks andCorey (1964) introduced a pore structure parameter—pore sizedistribution index, to relate the logarithmic capillary pressure andlogarithmic effective saturation. At the end of the last century,with the explicit proposal of fractal dimension (Mandelbrot andWheeler, 1983; Mandelbrot et al., 1984), capillary pressureanalysis began to combine with fractal dimension. Li andHorne (2003) studied the heterogeneity of Geysers sandstonequantitatively using fractal dimension derived from HPMI data.Then, more and more investigations on the combination offractal theory and HPMI technique further confirmed itseffectiveness in quantitative characterization of pore networkheterogeneity of rocks (Li and Horne, 2004; Li and Horne,2006; Li, 2010; Lai and Wang, 2015; Guo et al., 2019). Also,HPMIcanprovidevariousporestructureparameters,representing pore network connectivity and seepage capacity,which makes it the first choice to combine with fractal theory.It can be seen from the above review that both 2D and 3Dfractal studies have been widely utilized in the heterogeneitycharacterization of pore structure. However, it is rarely seen thatthe differences and connections between two-dimensional andthree-dimensional fractal analysis of pore structure are discussedand compared in one study. In the past 30 years, scholarsworldwidealsopointedoutcoincidentallythatitisinappropriate to depict pore structure heterogeneity merely byone parameter—fractal dimension (Xie et al., 1999; Peng et al.,2018; Hou et al., 2019).Given these research gaps, this study aims to combine 2D and3D fractal analysis for the pore structure characterization of tightsandstone, comparing the advantages and disadvantages of 2Dand 3D fractal characterization and investigating their relations.We selected the tight sandstone of the Xujiahe Formation (one ofthe most important tight sandstone formations in Sichuan Basin)as the research object, integrating 2D fractal analysis on thin-section images and 3D fractal analysis on HPMI data for the porestructure characterization. The related issues are discussed inthis paper.2 GEOLOGIC BACKGROUNDThe Sichuan Basin is a typical superimposed basin with an area ofabout 180,000 km2 and a total sedimentary thickness of about8000–12,000 m (Zecheng et al., 2013). According to the tectonicdivision of the Sichuan Basin, our research area (Figures 1A, 2, 3), thegiant Anyue gas field, is located in the eastern axial region of theLeshan-Longnvsi paleo-uplift in the central Sichuan Basin (Li et al.,2020). The exploration of the target formation, the XujiaheFormation in central Sichuan, began in the 1950s. To date, severalgiant gas fields of Xujiahe Formation, like Guang ‘an, Hechuan,Anyue, and Xinchang, have been discovered successively, withproven gas reserves of more than 0.6 × 1012 m3 (Tao et al., 2014).The upper Triassic Xujiahe Formation is the first continentalstrata since Paleozoic. The overall thickness of the XujiaheFormation in the study area ranges from 532 to 860 m.Vertically, the Xujiahe formation can be divided into sixmembers (Figure 1B). The first, third, and fifth members aremainly black mudstone and shale of lacustrine facies, containingcoal seams and the primary source layers. The second, fourth, andsixth members are mainly grey fine-grained tight sandstoneinterbedded with black mudstone and thin coal lines. Our targetformation is the second member of the Xujiahe Formation (Xu2),mainly braided-river delta front deposits. The lithology of the Xu2member is primarily grey, medium and fine feldspathic quartzsandstone, lithic quartz sandstone, and lithic feldspathic quartzsandstone. The grains are well sorted, sub-angular, point-linecontact, and mainly medium-grained. The cumulative thicknessof the Xu2 member varies from 105 to 543 m, which lays a solidfoundation for a qualified reservoir.Frontiers in Earth Science | www.frontiersin.orgNovember 2021 | Volume 9 | Article 7519443Song et al.Bi-Fractal Characterisation of Pore Network3 SAMPLING AND METHODOLOGYFifteen tight sandstone samples from eight exploration wells wereobtained from the second member of the Xujiahe Formation(Xu2) of the Upper Triassic. The tight sandstone in this study ismainly deposited in the delta-front facies of the braided river. Itslithology is mainly lithic sandstone or lithic quartz sandstone.Also, for comparison’s sake, some of these samples are from thegas-bearing formation when the others are from the dryformation.FIGURE 1 | (A) Geological map, demonstrating: I. the tectonic units of the Sichuan Basin; II. the location of main Xujiahe gas fields; III. the location of our researcharea. (B) The stratigraphic setting of the Xujiahe Formation.FIGURE 2 | (A) The original casting thin section image; (B) Extracted pore structure; (C) Binarised image of pore structure; (D) Cropping the binary image of porestructure into square-shaped for fractal analysis; (E) Fractal analysis on pore structure using the box counting method.Frontiers in Earth Science | www.frontiersin.orgNovember 2021 | Volume 9 | Article 7519444Song et al.Bi-Fractal Characterisation of Pore Network3.1 Sampling and ExperimentsGuided by standard SY/T 5368–2016, thin sections were made byimpregnating with epoxy resin under vacuum. To betterdistinguish the pore network with matrix, we chose to dye theepoxy resin blue. The porosity and permeability were measuredby helium porosimeter in the state key laboratory of petroleumresources and prospecting according to the standard SY/T5336-2006. Also, HPMI was also carried out here. Furthermore, theBeijing Research Institute of Uranium Geology (BRIUG)supported the X-ray diffraction experiment for whole-rockmineral analysis.3.2 2D Fractal Analysis on Casting ThinSection ImagesCastingthinsectionimagescontainsporestructureinformation; however, we should first extract the porenetwork for quantitative image analysis’s sake. Thus, pre-processing of casting thin section images is required. In thisstudy, the pre-processing of casting thin section images before2D fractal analysis contains three steps: (1) Extracting porenetwork (blue part) by filtering out redundant matrixinformation (Figure 2A,B); (2) binarization of the porenetwork to make the calculation simpler (Figure 2C); and(3) cropping the image to a square shape (1024 pixel × 1024pixel) for fractal analysis. After pre-processing, we carried outthe fractal analysis on the pore network using the box-counting method (Figure 2D,E).The box-counting method, also known as the pixel-coveragemethod, is realized by iteratively covering the same set of shapes(square boxes with length r) on a fractal object (image) andcalculating the number of boxes (N(r)) carrying information.When the side length of the box decreases, the number of boxesincreases. Theoretically, the fractal dimension of the fractal objectequals log (N(r))/log (1/r) when the side length of the box tends tozero (Equation 1):D � limr→0log(N(r))log�1r�(1)In this study, the relationship between the calculated boxnumber N(r) and the side length of the box is shown indouble-logarithmic coordinates, then the slope of log(r) andlog (N(r)) is calculated by the least squares method. Thefractal dimension D of the fractal (pore network) equals thenegative slope.3.3 3D Fractal Analysis on HPMI DataThough there are various models established for the fractalanalysis of HPMI data, in this study, we carried out a 3Dfractal analysis on HPMI data using models derived by Li(2010). The detailed derivation of the model could be referredto as the works of Song et al. (2018).4 RESULTS4.1 Petrophysical Properties and MineralCompositionThe overall porosity of the second member of the Xujiahe tightsandstone is relatively low, ranging from 5.45 to 10.29%, withan average value of about 7.86%. Meanwhile, the permeabilityranges from 0.025 to 0.489 mD, with an average value of0.159 mD. The physical parameters further confirm that theXujiahe formation is a typical tight sandstone formation.Further analysis of physical parameters shows that theformation quality of gas-bearing formation is better thanthat of dry formation, with relatively higher porosity andpermeability. The average porosity of gas-bearing samples is8.33%, while that of the dry formation is only 6.93%. Thepermeability difference between gas-bearing formation anddryformationissignificant—morethanoneorderofmagnitude: the permeability of gas-bearing samples rangesfrom 0.130 to 0.489 mD, with an average value of 0.218 mD. Incontrast, the permeability of dry samples ranges from 0.025 to0.058 mD, with an average of 0.041 mD.The Xujiahe tight sandstone is mainly composed of quartz,feldspar, and clay (Figures 4A,B). Quartz is the dominatingmineral, with a relative content from 55.40% to as high as85.60%(averagevalue:75.39%,medianvalue:78.60%).Feldspar (plagioclase and orthoclase) and clay ranked secondand third, accounting for 13.43 and 10.49%, respectively. Wecan also see from the ternary plot that brittle minerals have adominant position in composition. In addition, the content ofplagioclase is slightly higher (4.75%) than that of orthoclase.The content of carbonate minerals such as calcite and dolomiteisrelativelylow(only0.69%onaverage).Moreover,comparative analysis shows that the gas-bearing formation ischaracterized by relatively higher contents of quartz andfeldspar while a relatively lower content of clay minerals.The contents of quartz in the gas-bearing formation are11.18% higher than those in the dry formation, while thecontent of feldspar and clay minerals is 5.12 and 4.34%lower on average, respectively.FIGURE 3 | The scatter plot showing the relationship between porosityand permeability of the Xujiahe Formation.Frontiers in Earth Science | www.frontiersin.orgNovember 2021 | Volume 9 | Article 7519445Song et al.Bi-Fractal Characterisation of Pore Network4.2 Direct Observation of Pore SpaceThis study investigated the microscopic characteristics of 23 Xu2tight sandstone samples based on thin sections. Generally, theXu2 tight sandstone is mainly medium- to fine-grained lithicfeldspathic sandstone or feldspathic lithic sandstone. The particlecontact mode is mainly linear contact (Figure 5), reflecting thatthe tight sandstone of Xu2 has been subjected to medium-strongcompaction. The pore types of Xu2 tight sandstones are primarilyintergranular pores, micro-fractures, and intra- and intergranulardissolution pores (Figure 5). The existence of sharp and angularparticlesdemonstratesalow-mediumstructuralmaturity—relatively short transportation before deposition.Intergranular pores are mostly residual primary intergranularpores and dissolution pores. Due to severe compaction, only asmall fraction of the primary intergranular pores is preserved. Incontrast, the dissolution pores are widely developed due to theacidic environment caused by hydrocarbon generation. Thedissolution occurs at the edge of feldspar and calcite particles,forming large intergranular pores and at some residual feldsparsparticles transformed by kaolinization, forming intragranulardissolution pores. As shown in Figure 5, we can see apparentdifferences between gas-bearing and dry formations: (1) The porespace is more developed in the gas-bearing formation. (2) Micro-fractures are more developed in the gas-bearing formations. Wecan also see some micro-fractures in the dry formations.However, most of them are filled with clay minerals, whilemost of the micro-fractures in the gas-bearing formations areopen-ended. (3) Most of the micro-fractures are developed inquartz particles.To further compare the pore network of gas-bearing and dryformations, this paper studied their pore structure on a smallerscale with the help of SEM (Figure 5). In nanoscale, it is moreevident that though quartz and feldspar dominate in both gas-bearing and dry formations, rigid particles (quartz and feldspar)are highly cracked in the gas-bearing formation than in the dryformation. Also, it is apparent that the majority of intragranularspace and micro-fractures of the dry formation are filled with clayminerals (primarily chlorite).4.3 Quantitative Pore StructureCharacterizationThe experimental results of high-pressure mercury injection ofsamples from the gas-bearing and dry formations show apparentdifferences. First, it is reflected in the curve shape of HPMI. TheHPMI curves of all gas-bearing samples are remarkablyconsistent: the displacement pressure is all lower than 1 MPa,with an average value of 0.62 MPa (Table 2); the injectionbranch has a relatively flat “platform section” (Figure 6); theoverall shape of “platform section” is concave, demonstratingcoarse-grained skewness/relatively large pore-throat diameter;the maximum mercury saturation is basically greater than90%, reaching as high as 96.27%. In sharp contrast with thegas-bearing samples, the HPMI shape of the dry samples isquite different, which is directly reflected in the significantwiderangeinthemaximummercurysaturation(68.18–96.81%). The HPMI shape of all dry samples alsoreflects some similarities: the displacement pressure is allgreater than 1 MPa (average value: 1.50 MPa); comparedwith the gas-bearing formation, the injection branch ismore inclined, showing fine-grained skewness (relativelyfiner pore size).Furthermore, from the perspective of pore size distribution(PSD), the PSD of gas-bearing samples (taking sample A4 as anexample) and dry formation (taking sample A11 as an example)also show distinct characteristics. The PSD (histogram) of gas-bearing samples shows a significant primary peak, which provesthat the pore space is mainly in specific pore size (63–630 nm).FIGURE 4 | Mineralogy comparison (stacked bar chart on the left and ternary plot on the right) between samples from gas-bearing formation and the dry formationbased on XRD data.Frontiers in Earth Science | www.frontiersin.orgNovember 2021 | Volume 9 | Article 7519446Song et al.Bi-Fractal Characterisation of Pore NetworkOn the other hand, from the perspective of contribution topermeability, pores with a diameter from 100 nm to 1 μm play asignificant role in the fluid seepage in pore space. In contrast, thePSD of dry samples does not show a peak state. PSD distributionis flat and gentle, demonstrating that the pore space consists ofpores with a wide range of sizes. The heterogeneity of pore spaceof the dry formation is stronger than that of the gas formationfrom pore size. From the perspective of contribution topermeability, pores with diameters from 63 to 250 nm play asignificant role. On the one hand, it shows that the space of fluidseepage of the dry formation is narrower than that of the gas-bearing formation; on the other hand, it is also explained fromthe point of view of pore structure that the flow of fluid in such aspace is more complicated than that in the gas-bearingformation.Statistics of HPMI parameters from 10 gas-bearing and 5 drysamples help to make a more in-depth quantitative comparisonon the pore structure of gas-bearing and dry formations. P50 (themedian pressure) and r50 (the median pore radius) are the twomost intuitive parameters reflecting the tightness of rock. Theaverage values of P50 of gas-bearing and dry samples are 3.84 and14.04 MPa, respectively. At the same time, r50 (the median poreradius), the representative parameter for pore/throat average size,of gas-bearing and dry samples are 200 and 80 nm, respectively. Arelatively low value of P50 and a high value of r50 correspond tothe coarse skewness of gas-bearing samples. Furthermore, we canFIGURE 5 | Casting thin section images and SEM images, demonstrating the pore structure difference between gas-bearing and dry formations.Frontiers in Earth Science | www.frontiersin.orgNovember 2021 | Volume 9 | Article 7519447Song et al.Bi-Fractal Characterisation of Pore NetworkTABLE 2 | Parameters obtained from HPMI experiment.IDFormationDepth(m)rmax(μm)�r(μm)r50(μm)rmax-PSD(μm)rmax-K(μm)SpDSmax(%)Pd(MPa)P50(MPa)A1Gas2663.221.090.280.190.630.632.317.9196.770.684.05A6Gas2671.891.070.250.190.250.632.158.0896.950.683.98A8Gas2275.541.090.270.200.400.632.157.4797.730.673.78A10Gas2364.191.090.270.210.400.632.268.0397.000.673.74A12Gas2498.491.070.280.170.630.632.357.8794.860.694.45A14Gas2227.561.540.390.201.001.002.796.9195.020.483.90A16Gas2251.991.580.400.260.631.002.325.5597.290.472.84A20Gas2576.441.090.250.180.250.632.399.2896.130.674.24A21Gas2705.881.590.360.220.631.002.637.3993.460.463.59A22Gas2570.651.080.260.200.400.632.288.3797.500.683.87Aver.——1.230.300.200.520.742.367.6996.270.623.84A3Dry2649.440.540.120.090.160.252.0417.5796.811.368.18A7Dry2271.820.540.160.130.250.252.0412.7595.291.365.92A11Dry2411.030.360.110.020.160.252.3631.1368.182.0532.7A13Dry2245.380.540.130.060.160.402.1317.6496.271.3613.47A18Dry2685.960.540.140.080.250.402.2017.0294.251.369.95Aver.——0.500.130.080.200.312.1519.2290.161.5014.04*rmax, maximum pore radius; �r, average pore-throat radius; r50, median pore-throat radius; rmax-PSD, pore distribution peak position; rmax-K, permeability distribution peak position; Sp,sorting coefficient; D, relative sorting coefficient; Smax, maximum mercury saturation; Pd, displacement pressure; P50, saturation median pressure.FIGURE 6 | Typical HPMI curves of samples from gas-bearing (A) and dry (B) formations; and the typical PSDs obtained from HPMI experiments for gas-bearing[(C), sample A4] and dry [(D), sample A11] formations.Frontiers in Earth Science | www.frontiersin.orgNovember 2021 | Volume 9 | Article 7519448Song et al.Bi-Fractal Characterisation of Pore Networkalso see that the average pore radius �r and the maximum poreradius rmax of gas-bearing samples (300 nm and 1.23 μm,respectively) are apparently larger than that of dry samples(130 and 500 nm, respectively). With such a big difference inpore size (distribution), the primary seepage pore space in gas-bearing and dry formations are different: the rmax-K (radiuscorresponds to the peak of permeability distribution curve) ofgas-bearing formation (740 nm) is bigger than that of dryformations(310 nm).Thus,theseepageingas-bearingformation is easier. A better seepage capability can also beverified by Sp (sorting coefficient) and D (relative sortingcoefficient). Gas-bearing samples have a higher Sp and lowerD, indicating that pore size is more concentrated.4.4 2D and 3D Fractal Analysis for PoreStructure CharacterizationWith the help of the “box-counting” method, based on thequalitative characterization of the pore structure of the castingthin section images, we optimized typical casting thin sectionimages that can reflect the pore structure characteristics of thecorresponding sample for two-dimensional fractal analysis.The results (Figure 7) show that: (1) on the bi-logarithmiccoordinates of r and N (r), the data points show a significantlinear distribution, and the fitting goodness of linear regressiongenerally reaches more than 95%, indicating that the two-dimensional analysis is adequate, and the result is reliable. (2)It should be noted that the fitting goodness of gas-bearingsamples is generally better than that of dry samples, which mayindicate a better self-similarity of pore space. (3) For both gas-bearinganddrysamples,thetwo-dimensionalfractaldimension is in the range of (1, 2), demonstrating that thecalculated fractal dimensions are reasonable. (4) The surfaceporosity of gas-bearing samples is significantly better than thatof dry samples. The face rates of two typical gas-bearingsamples are 14.09 and 16.58%, respectively, far better thanthat of dry samples of 1.87 and 1.44%. (5) The two-dimensional fractal dimensionof gas layer samples issignificantly larger than that of dry layer samples. (6) Fromthe point of view of pore network 2D distribution, theextractedporespaceofgas-bearingsamplesismorehomogeneous distributed in 2D space and better connected.Incontrast,the2Dporespaceofdryformationisheterogeneously distributed/concentrated in some areas withlow connectivity.Moreover, statistics (Table 3) show apparent differences intwo-dimensional fractal results between gas-bearing and drysamples. The calculated surface fractal dimension of gas-bearing samples varies from 1.44 to 1.80, with an averagevalue of 1.65. In contrast, the calculated surface fractalFIGURE 7 | Two-dimensional fractal analysis on thin-section images of typical gas-bearing (A16, A6) and dry samples (A11, A3). *Ds: 2D fractal dimension obtainedfrom casting thin section image fractal analysis.Frontiers in Earth Science | www.frontiersin.orgNovember 2021 | Volume 9 | Article 7519449Song et al.Bi-Fractal Characterisation of Pore Networkdimension of dry samples is in the range of (1.12, 1.45), with anaverage value of 1.39. Also, it is more apparent that the surfaceporosity of gas-bearing samples is far better than that of drysamples.Based on the two-dimensional fractal analysis on porestructure with the help of thin-section images, we furthercombined the Li (2010) model and HPMI data to carry outthree-dimensional fractal analysis on pore structure. TheTABLE 3 | Parameters obtained from 2D Fractal analysis.IDFormationDepth(m)D1R2(D1)D2R2(D2)DsR2(Ds)Surfaceporosity(%)A1Gas2663.222.700.5782.120.9281.640.99710.34A6Gas2671.892.930.5732.110.9321.730.99916.58A8Gas2275.542.790.7022.110.9181.660.9968.52A10Gas2364.192.810.5732.110.9311.800.99910.21A12Gas2498.492.620.6082.110.9211.590.99610.61A14Gas2227.562.410.9322.130.9581.640.9968.94A16Gas2251.992.480.7922.090.9141.650.99614.09A20Gas2576.442.890.7432.120.9241.760.99910.05A21Gas2705.882.590.7682.100.9251.580.99514.78A22Gas2570.652.850.5692.120.9581.440.99212.32Average——2.710.6842.110.9311.650.99711.644A3Dry2649.443.590.3362.200.9671.450.9931.44A7Dry2271.823.820.3162.150.9411.260.9784.78A11Dry2411.033.680.2942.230.9691.120.9701.87A13Dry2245.383.530.3842.260.9781.410.9872.14A18Dry2685.963.480.4192.210.9651.290.9842.93Average——3.620.3502.210.9641.310.9822.632FIGURE 8 | Three-dimensional fractal analysis on HPMI data of typical gas-bearing (A,B) and dry samples (C,D).Frontiers in Earth Science | www.frontiersin.orgNovember 2021 | Volume 9 | Article 75194410Song et al.Bi-Fractal Characterisation of Pore Networkresults of three-dimensional fractal analysis on typical gas-bearing and dry samples (Figure 7) show that the data pointsapparently show a two-stage linear distribution on the bi-logarithmic coordinates of Pc and SHg. The piecewise linearregression of data points shows that the goodness of fit ofall samples has reached more than 90%, proving that theappliedfractalmodeliseffectiveinanalyzingHPMIdata. The breakpoint of the two-stage distribution is nearLog (Pc) � 0.8 for both and gas-bearing and dry samples. Thepore/throat radius corresponding to this breakpoint is116.5 nm, indicating that the pore spaces with a poreradius greater than this value and less than this value havedifferent self-similar characteristics (fractal characteristics).In addition, according to Figure 8 and Table 3, D1 of alldry samples is greater than 3, while D2 of all dry layers, D1 andD2 of all gas-bearing samples are within the normal range of3D fractal dimension—(2, 3). Therefore, in the fractal resultsof Li (2010), the relatively smaller pore space in the drysamplescannotbereasonablyexplainedbyfractaldimension, so this part of the pore space is not considered.However, according to the results of Section 3.3, pores lessthan 116.5 nm play an essential role in fluid seepage inthe pore space of dry formations, which should be paidattention to in the analysis.5 DISCUSSIONThe core issue in tight sandstone reservoir characterization is howthe reservoir quality controls the gas-bearing property (Zou et al.,2012; Sakhaee-Pour and Bryant, 2014; Gao and Li, 2015; Schmittet al., 2015; Xi et al., 2016; Zhang et al., 2018). Therefore, in thisstudy, we took the gas-bearing property of the reservoir as aparameter (“Gas-bearing Status” in Figure 9) and made a bi-correlationanalysisonparameters,includinggas-bearingproperty, the mineral composition, reservoir pore structureFIGURE 9 | Heatmap showing the correlation relationships among petrophysical parameters, mineral compositions, pore structure parameters, and gas-bearingstatus. *There are two values for “Gas-bearing status”: True or False, representing gas-bearing and dry formations. The color of the heatmap ranges from dark blue todark red, representing a correlation coefficient of −1 (negative correlation) to 1 (positive correlation), respectively.Frontiers in Earth Science | www.frontiersin.orgNovember 2021 | Volume 9 | Article 75194411Song et al.Bi-Fractal Characterisation of Pore Networkparameters, and fractal dimensions. The results were shown inthe form of a heatmap (Figure 9).5.1 Relationships Between Gas-BearingStatus and Physical Properties, MineralCompositionFirst, the gas-bearing property related issues shall be analyzed.From the perspective of physical properties, there is anapparent positive correlation between permeability and gas-bearing status (the correlation coefficient equals 0.82). Incontrast, the correlation between porosity and gas-bearingstatus is not apparent, indicating that the controlling effect ofreservoir seepage performance on gas-bearing status is moredominant than the porosity. From the perspective of mineralcomposition, quartz and clay minerals, the main minerals ofXu2 formation, have a stronger correlation relationship withgas-bearing status than any minerals except dolomite (thecontent of dolomite is too low to be considered), and thecorrelation coefficients reach 0.52 and −0.50, respectively.Fromthecorrelationdiagrambetweenpermeability,porosity, and mineral components (Figure 10), it can beseen that quartz and clay minerals have a much moresignificantimpactonphysicalpropertiesthanotherminerals. Quartz has a positive correlation with physicalparameters, while clay minerals negatively correlate withphysicalparameters.Whetherquartzorclay,theircorrelation with permeability is significantly higher thanthat with porosity, indicating the existence of quartz andclay minerals primarily controls the seepage capability ofthe reservoir. The positive correlation between quartz androck physical properties and gas-bearing property can beexplained in Figure 5: the rigid particles (mainly quartz) ofgas-bearing samples are intensely cracked to form a goodseepagechannelconducivetonaturalgasmigrationand accumulation. When clay minerals block the formedseepagechannel,theseepageperformancedecreasessharply,resultinginpoorgas-bearingstatusandtheformation of dry formation. In the dry samples shown inFigure 5, we also see many microcracks, but most of themhave been filled with clay minerals. Therefore, there is anegativecorrelationbetweenclay mineral content andseepage performance (permeability) and gas-bearing status.It is also worth noting that even in the dry formation, thoughmost seepage channels (micro-fractures) are filled with clayminerals, micro-fractures developed in rigid particles is stillthe dominant factor affecting their seepage (permeability).Therefore, there is a good positive correlation betweenquartz content in the dry samples and permeability, gascontent.Furthermore, if the gas-bearing status is analyzed from theperspective of diagenesis, most of the diagenetic processes thatcontribute to porosity, like compaction and dissolution, arenot the controlling factors on gas-bearing status. That is whywe see in Figure 3, that even with low porosity, someformation can be gas-bearing. In contrast, clay cementation,which blocks the seepage channel, and tectonic movements,which generate micro-fractures, control gas-bearing status.Correspondingly, we see from Figure 3 that there is a clearboundary between gas-bearing and dry samples, and thepermeability of all gas-bearing samples are higher than thatof dry samples.5.2 Relationships Between Gas-BearingStatus and Pore Structure ParametersTo more specifically characterize the controlling effect of porestructure on gas-bearing status, five pore/throat-size-relatedparameters are selected for comparison in this paper: theaverage pore/throat radius (�r), the maximum pore/throatradius (rmax), the median pore/throat radius (r50), thepermeability distribution peak radius (rmax-PSD), and thepore/throat-size distribution peak radius (rmax-K). We cansee in Figure 9 that almost all pore/throat-size-relatedFIGURE 10 | Correlation relationships between permeability (left), porosity, and mineral composition.Frontiers in Earth Science | www.frontiersin.orgNovember 2021 | Volume 9 | Article 75194412Song et al.Bi-Fractal Characterisation of Pore Networkparameters have a good positive correlation with gas-bearingstatus, seepage performance, porosity, and quartz contentwhile negatively correlating with clay minerals. Furtheranalysis of pore/throat-size-related parameters (Figure 11)with physical properties shows that the correlation betweenpore/throat-size and permeability is generally stronger thanthat with porosity. In addition, among the five pore/throat-size-related parameters, though the average pore/throatradius has the closest correlation with permeability, themedian pore/throat radius (r50) has the most apparentcontrol effect on permeability (a small increase in themedian pore/throat radius can lead to a large increase inpermeability). Thus, r50 is the most sensitive parameter toseepage capability (permeability). In contrast, the maximumpore/throatradiushasarelativelyweakcontrolonpermeability.In contrast to r50, Pd and P50 have a good negativecorrelation with gas-bearing status and permeability, a weaknegative correlation with porosity and quartz content, and aweak positive correlation with clay minerals. When the porestructure tends to coarse skewness (P50 is small), it isconducivetofluidseepageinthereservoir.ThroughFIGURE 11 | Correlation relationships between permeability (A), porosity (B), and five representative pore/throat-size-related parameters.FIGURE 12 | Comparison of gas-bearing status on the correlation between permeability (left), and Pd, P50.Frontiers in Earth Science | www.frontiersin.orgNovember 2021 | Volume 9 | Article 75194413Song et al.Bi-Fractal Characterisation of Pore Networkcomparison of gas-bearing and dry samples (Figure 12), it isfoundthatthePdandP50ofdrysamplesfluctuateconsiderably,reflectingastrongheterogeneityofdrysamples.Incomparison,thegas-bearingsamplesarerelatively uniform: Pd and P50 are small.5.3 Relationships Between FractalDimensions (2D and 3D) and Pore StructureParametersThere is a strong negative correlation between gas-bearingstatus, permeability, and three-dimensional fractal dimension(D1, D2), but a positive correlation with two-dimensionalfractal dimension (Ds) (Figure 9). The reason for this: thethree-dimensional fractal dimension is obtained based on theanalysis of the seepage behavior of the fluid in the pore space.The greater its value, the rougher the surface that the fluidpasses through during seepage, that is, the more substantial theheterogeneityofthethree-dimensionalporespace.Correspondingly, the more complex the fluid seepage in thepore space and the worse the gas-bearing status. In contrast,thetwo-dimensionalfractaldimensionisobtainedbyanalyzingthedistributionofporenetworksintwo-dimensional space using the box-counting method. Themore developed the pore network is in the plane/surface,the better the connectivity is, and the larger the two-dimensional fractal dimension obtained by the box-countingmethod. A more developed pore network is more conducive toFIGURE 13 | Correlation relationship of 2D and 3D fractal dimension with mineral components.FIGURE 14 | Correlation relationship of 2D and 3D fractal dimension with representative pore/throat-size-related parameters.Frontiers in Earth Science | www.frontiersin.orgNovember 2021 | Volume 9 | Article 75194414Song et al.Bi-Fractal Characterisation of Pore Networkthe migration of fluid in pore space. Therefore, the larger theDs, the better the seepage performance and the better the gas-bearing status. From a diagenetic point of view, when micro-fracturesresultingfromtectonicmovementsaremoredeveloped and better connected, the seepage channels forgas flow are more homogeneous; as a result, the D2 issmaller, and the Ds is larger. When the micro-fractures areseverely blocked by clay cementation, the seepage channelsbecome more heterogeneous, and the plane distribution ofconnected pore spaces is limited. Correspondingly, the D2 islarger, and the Ds is smaller. Thus, it seems safe to say that the3D fractal dimension represents pore space heterogeneitywhile the 2D fractal dimension indicates the development ofthe pore network.From the perspective of mineral composition, the analysis ofthe correlation between D2, DS, and mineral components showsthat quartz and clay still have the best correlation with D2 andDS (Figure 13). D2 has a negative correlation with quartzcontent and a positive correlation with clay content—theseepage channel of Xu2 member is mainly micro-fracturesformed by the cracking of rigid particles, which mainly occurin quartz. When the quartz content is relatively high, and theclay mineral content is low, the connectivity of the seepagechannel is good, and the heterogeneity is relatively weak. Withthe increase of clay mineral content, the seepage channels suchasmicro-fracturesareblocked(Figures5A,B),theheterogeneity of seepage channels increases significantly. Incontrast, DS positively correlates with quartz content and anegatively correlates with clay mineral content. As the seepagechannel is mainly the micro-fractures in rigid particles. With theincrease of quartz content, the plane distribution of this seepagechannel network is more expansive, and the connectivity isbetter, so DS is larger. When the content of clay mineralsgradually increases, the seepage channel network is graduallyblocked by clay minerals. As a result, the effective seepagechannels on the plane become less, the connectivity of thepore network becomes worse, even concentrated in somelocal areas, so DS becomes smaller.From the perspective of pore size, the correlation of five pore/throat-size-related parameters with fractal dimensions (D2 andDS) (Figure 14) shows that r50 correlates with fractal dimensions(D2 and DS) best while rmax fluctuates with fractal dimension themost. It is consistent with the analysis result on the correlationbetween permeability and pore/throat-size-related parameters.On the one hand, it shows that using fractal dimension tocharacterize fluid seepage performance is effective; on theother hand, it also shows that r50 is a representative parameterto characterize pore seepage performance.6 CONCLUSIONThis paper combined 2D and 3D fractal analysis for the porestructure characterization to gain insight into the controllingfactors on fluid seepage performance in tight sandstone. Physicalanalysis, mineral composition analysis (XRD), and HPMI werecombined with 2D and 3D fractal analysis for an in-depthinvestigation on the pore structure of tight sandstone. Therelationship between fractal dimensions (2D and 3D) andphysical parameters, mineral composition, and pore structureparameters was discussed. The following conclusions wereachieved.1) The Xu2 tight sandstone is mainly medium- to fine-grainedlithic feldspathic sandstone or feldspathic lithic sandstonewithlowporosityandpermeability.Quartzisthedominating mineral (average value: 75.39%), 11.18%higher in the gas-bearing formation than in the dryformation. The pore types of Xu2 tight sandstones areprimarily intergranular pores, micro-fractures, and intra-and intergranular dissolution pores. Most of the micro-fractures in gas-bearing formation are open-ended whilefilled mainly by clay minerals in dry formation.2) The gas-bearing formation is distinguished from the dry formationby relatively low displacement pressure (<1 MPa), coarseskewness, and apparent peak in PSD (pore size distribution).3) Ds of gas-bearing samples is significantly larger than that of drysamples, while D1 and D2 of gas-bearing samples are lower.4) There is a strong negative correlation between D2 and gas-bearing status, permeability, quartz content, and r50, but apositive correlation between Ds and these parameters.5) D2 represents the heterogeneity of pore space, while the Dsindicates the development of the pore network.DATA AVAILABILITY STATEMENTThe raw data supporting the conclusion of this article will bemade available by the authors, without undue reservation.AUTHOR CONTRIBUTIONSMethodology and writing—original draft preparation, ZS;supervision, YZ and ZS; investigation, JZ; visualisation andformal analysis, JZ; writing—review and editing, ZS and JZ;project administration, LH, CL and WL, sampling andexperiments; DY and YW.FUNDINGThe present research was funded by the National Natural ScienceFoundation of China, grant number 41472118, and the State KeyLaboratory of Petroleum Resources and Prospecting, grantnumber PRP/indep 04-1611 and 2462017YJRC025. All therelevant projects can be found on the website of ZS: https://www.cup.edu.cn/geosci/szdw/fujiaoshou/179957.htm.ACKNOWLEDGMENTSWe thank the PetroChina Southwest Oilfield Company forproviding core samples and necessary data.Frontiers in Earth Science | www.frontiersin.orgNovember 2021 | Volume 9 | Article 75194415Song et al.Bi-Fractal Characterisation of Pore NetworkREFERENCESBrooks, R. H., and Corey, A. T. (1964). Hydraulic Properties of Porous Media.Colorado 26 (1), 352–366. doi:10.13031/2013.40684Corey, A. T. (1954). The Interrelation between Gas and Oil Relative Permeabilities.Producers monthly 19 (1), 38–41.Dai, J., Ni, Y., and Wu, X. (2012). Tight Gas in China and its Significance inExploration and Exploitation. Pet. Exploration Develop. 39 (3), 277–284.doi:10.1016/S1876-3804(12)60043-3Daigle, H., Johnson, A., and Thomas, B. (2014). Determining Fractal Dimensionfrom Nuclear Magnetic Resonance Data in Rocks with Internal Magnetic FieldGradients. Geophysics 79 (6), D425–D431. doi:10.1190/geo2014-0325.1Ersoy, E., Schaffer, M. E., and Ditzen, J. (2019). BP Statistical Review of WorldEnergy 2019.Gao, H., and Li, H. A. (2015). Pore Structure Characterization, PermeabilityEvaluation and Enhanced Gas Recovery Techniques of Tight GasSandstones.J.Nat.GasSci.Eng.28,536–547.doi:10.1016/j.jngse.2015.12.018Guo, X., Huang, Z., Zhao, L., Han, W., Ding, C., Sun, X., et al. (2019). PoreStructure and Multi-Fractal Analysis of Tight sandstone Using MIP, NMR andNMRC Methods: A Case Study from the Kuqa Depression, China. J. Pet. Sci.Eng. 178, 544–558. doi:10.1016/j.petrol.2019.03.069Hou, X., Zhu, Y., Chen, S., Wang, Y., and Liu, Y. (2020). Investigation on PoreStructure and Multifractal of Tight sandstone Reservoirs in Coal Bearing StrataUsing LF-NMR Measurements. J. Pet. Sci. Eng. 187, 106757. doi:10.1016/j.petrol.2019.106757Lai, J., and Wang, G. (2015). Fractal Analysis of Tight Gas Sandstones Using High-Pressure Mercury Intrusion Techniques. J. Nat. Gas Sci. Eng. 24, 185–196.doi:10.1016/j.jngse.2015.03.027LayLian,T.,Radhakrishnan,P.,andDayaSagar,B.S.(2004).MorphologicalDecompositionofsandstonePore-Space:FractalPower-Laws. Chaos, Solitons & Fractals 19 (2), 339–346. doi:10.1016/S0960-0779(03)00046-8Li, A., Ding, W., Jiu, K., Wang, Z., Wang, R., and He, J. (2018). Investigationof the Pore Structures and Fractal Characteristics of marine ShaleReservoirs Using NMR Experiments and Image Analyses: A Case Studyof the Lower Cambrian Niutitang Formation in Northern GuizhouProvince, South China. Mar. Pet. Geology. 89, 530–540. doi:10.1016/j.marpetgeo.2017.10.019Li, K. (2010). Analytical Derivation of Brooks–Corey Type Capillary PressureModels Using Fractal Geometry and Evaluation of Rock Heterogeneity. J. Pet.Sci. Eng. 73 (1-2), 20–26. doi:10.1016/j.petrol.2010.05.002Li, K., and Horne, R. N. (2003). “Fractal Characterization of the Geysers Rock,” inProceedings Of the GRC 2003 Annual Meeting.Li, K., and Horne, R. N. (2006). Fractal Modeling of Capillary Pressure Curves fortheGeysersRocks.Geothermics35(2),198–207.doi:10.1016/j.geothermics.2006.02.001Li, K., and Horne, R. N. (2004). "Universal Capillary Pressure and RelativePermeability Model from Fractal Characterization of Rock", in StanfordGeothermalWorkshoponGeothermalReservoirEngineering,StanfordUniversity, Stanford, California.Li, K., and Zhao, H. (2012). Fractal Prediction Model of SpontaneousImbibition Rate. Transp Porous Med. 91 (2), 363–376. doi:10.1007/s11242-011-9848-0Li, Y., Qin, S., Wang, Y., Holland, G., and Zhou, Z. (2020). Tracing Interactionbetween Hydrocarbon and Groundwater Systems with Isotope SignaturesPreserved in the Anyue Gas Field, central Sichuan Basin, China.GeochimicaetCosmochimicaActa274,261–285.doi:10.1016/j.gca.2020.01.039Mandelbrot, B. B., Passoja, D. E., and Paullay, A. J. (1984). Fractal Character ofFracture Surfaces of Metals. Nature 308 (5961), 721–722. doi:10.1038/308721a0Mandelbrot, B. B., and Wheeler, J. A. (1983). The Fractal Geometry ofNature. Q. Rev. Biol. 147, 468. (Volume 58, Number 3). doi:10.1119/1.13295Peng, J., Han, H., Xia, Q., and Li, B. (2018). Evaluation of the Pore Structure ofTight sandstone Reservoirs Based on Multifractal Analysis: a Case Studyfrom the Kepingtage Formation in the Shuntuoguole Uplift, Tarim Basin,NW China. J. Geophys. Eng. 15 (4), 1122–1136. doi:10.1088/1742-2140/aaab9dSakhaee-Pour, A., and Bryant, S. L. (2014). Effect of Pore Structure on theProducibilityofTight-GasSandstones.Bulletin98(4),663–694.doi:10.1306/08011312078Schmitt, M., Fernandes, C. P., Wolf, F. G., Bellini da Cunha Neto, J. A., Rahner,C. P., and Santiago dos Santos, V. S. (2015). Characterization of BrazilianTight Gas Sandstones Relating Permeability and Angstrom-To Micron-ScalePoreStructures.J.Nat.GasSci.Eng.27,785–807.doi:10.1016/j.jngse.2015.09.027Shao, X., Pang, X., Li, H., and Zhang, X. (2017). Fractal Analysis of Pore Networkin Tight Gas Sandstones Using NMR Method: A Case Study from the OrdosBasin,China.EnergyFuels31(10),10358–10368.doi:10.1021/acs.energyfuels.7b01007Song, Z., Liu, G., Yang, W., Zou, H., Sun, M., and Wang, X. (2018). Multi-fractalDistribution Analysis for Pore Structure Characterization of Tight sandstone-ACase Study of the Upper Paleozoic Tight Formations in the Longdong District,OrdosBasin.Mar.Pet.Geology.92,842–854.doi:10.1016/j.marpetgeo.2017.12.018Sumantri, Y., and Permadi, P. (2018). A Study of Sandstone PermeabilityAnisotropy through Fractal Concept. Am. J. Sci. Eng. Technol. 3 (2), 34–45.doi:10.11648/j.ajset.20180302.12Sun, W., Zuo, Y., Wu, Z., Liu, H., Xi, S., Shui, Y., et al. (2019). Fractal Analysis ofPores and the Pore Structure of the Lower Cambrian Niutitang Shale inNorthern Guizhou Province: Investigations Using NMR, SEM and ImageAnalyses.Mar.Pet.Geology.99,416–428.doi:10.1016/j.marpetgeo.2018.10.042Tao, S., Zou, C., Mi, J., Gao, X., Yang, C., Zhang, X., et al. (2014). GeochemicalComparison between Gas in Fluid Inclusions and Gas Produced from theUpper Triassic Xujiahe Formation, Sichuan Basin, SW China. Org. Geochem.74, 59–65. doi:10.1016/j.orggeochem.2014.05.008Thomeer, J. H. M. (1960). Introduction of a Pore Geometrical Factor Defined bythe Capillary Pressure Curve. J. Pet. Technol. 12 (03), 73–77. doi:10.2118/1324-gWang, H., Liu, Y., Song, Y., Zhao, Y., Zhao, J., and Wang, D. (2012). FractalAnalysis and its Impact Factors on Pore Structure of Artificial Cores Based onthe Images Obtained Using Magnetic Resonance Imaging. J. Appl. Geophys. 86,70–81. doi:10.1016/j.jappgeo.2012.07.015Xi, K., Cao, Y., Haile, B. G., Zhu, R., Jahren, J., Bjørlykke, K., et al. (2016). How Doesthe Pore-Throat Size Control the Reservoir Quality and Oiliness of TightSandstones? the Case of the Lower Cretaceous Quantou Formation in theSouthern Songliao Basin, China. Mar. Pet. Geology. 76, 1–15. doi:10.1016/j.marpetgeo.2016.05.001Xie, H. P., Sun, H. Q., Ju, Y., and Feng, Z. G. (2001). Study on Generation of RockFracture Surfaces by Using Fractal Interpolation. Int. J. Sol. Structures 38 (32-33), 5765–5787. doi:10.1016/s0020-7683(00)00390-5Xie, H., Wang, J.-A., and Kwaśniewski, M. A. (1999). Multifractal Characterizationof Rock Fracture Surfaces. Int. J. Rock Mech. Mining Sci. 36 (1), 19–27.doi:10.1016/s0148-9062(98)00172-7Xu, S., Gou, Q., Hao, F., Zhang, B., Shu, Z., Lu, Y., et al. (2020a). Shale PoreStructure Characteristics of the High and Low Productivity wells, JiaoshibaShale Gas Field, Sichuan Basin, China: Dominated by Lithofacies orPreservation Condition? Mar. Pet. Geology. 114, 104211. doi:10.1016/j.marpetgeo.2019.104211Xu, S., Hao, F., Shu, Z., Zhang, A., and Yang, F. (2020b). Pore Structures ofDifferent Types of Shales and Shale Gas Exploration of the OrdovicianWufeng and Silurian Longmaxi Successions in the Eastern Sichuan Basin,SouthChina.J.AsianEarthSci.193,104271.doi:10.1016/j.jseaes.2020.104271Yuan, Y., and Rezaee, R. (2019). Fractal Analysis of the Pore Structure forclay Bound Water and Potential Gas Storage in Shales Based on NMRand N2 Gas Adsorption. J. Pet. Sci. Eng. 177, 756–765. doi:10.1016/j.petrol.2019.02.082Zecheng, W., Shipeng, H., Deyu, G., Wei, W., and Chong, Y. (2013). GeochemicalCharacteristics of Natural Gases in the Upper Triassic Xujiahe Formation in theSouthern Sichuan Basin, SW China. Int. J. Coal Geology. 120, 15–23.doi:10.1016/j.coal.2013.09.002Frontiers in Earth Science | www.frontiersin.orgNovember 2021 | Volume 9 | Article 75194416Song et al.Bi-Fractal Characterisation of Pore NetworkZhang, Y., Bao, Z., Yang, F., Mao, S., Song, J., and Jiang, L. (20182018). TheControls of Pore-Throat Structure on Fluid Performance in TightClastic Rock Reservoir: A Case from the Upper Triassic of Chang 7Member, Ordos Basin, China. Geofluids 2018, 1–17. doi:10.1155/2018/3403026Zhao, P., Wang, Z., Sun, Z., Cai, J., and Wang, L. (2017). Investigation on thePore Structure and Multifractal Characteristics of Tight Oil ReservoirsUsing NMR Measurements: Permian Lucaogou Formation in JimusaerSag, Junggar Basin. Mar. Pet. Geology. 86, 1067–1081. doi:10.1016/j.marpetgeo.2017.07.011Zou, C., Zhu, R., Liu, K., Su, L., Bai, B., Zhang, X., et al. (2012). Tight Gas sandstoneReservoirs in China: Characteristics and Recognition Criteria. J. Pet. Sci. Eng.88-89 (2), 82–91. doi:10.1016/j.petrol.2012.02.001Conflict of Interest: Authors DY, YW, LH, CL and WL are employed byPetroChina.The remaining authors declare that the research was conducted in the absence ofany commercial or financial relationships that could be construed as a potentialconflict of interest.Publisher’s Note: All claims expressed in this article are solely those of the authorsand do not necessarily represent those of their affiliated organizations, or those ofthe publisher, the editors, and the reviewers. Any product that may be evaluated inthis article, or claim that may be made by its manufacturer, is not guaranteed orendorsed by the publisher.Copyright © 2021 Song, Zhao, Zhang, Yang, Wang, Hu, Li and Liu. This is an open-access article distributed under the terms of the Creative Commons Attribution License(CC BY). The use, distribution or reproduction in other forums is permitted, providedthe original author(s) and the copyright owner(s) are credited and that the originalpublication in this journal is cited, in accordance with accepted academic practice. Nouse, distribution or reproduction is permitted which does not comply with these terms.Frontiers in Earth Science | www.frontiersin.orgNovember 2021 | Volume 9 | Article 75194417Song et al.Bi-Fractal Characterisation of Pore NetworkView publication stats | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/359032698Spousal Violence in Rural Nepal: Prevalence and Risk FactorsArticle · December 2021DOI: 10.3126/jpd.v2i1.43484CITATIONS0READS161 author:Some of the authors of this publication are also working on these related projects:Status of Workplace Sexual Harassment in Nepal View projectManusha PaudelTribhuvan University Patan Multiple Campus7 PUBLICATIONS 2 CITATIONS SEE PROFILEAll content following this page was uploaded by Manusha Paudel on 09 August 2022.The user has requested enhancement of the downloaded file.Journal of Population and Development, June 2021 | Manusha Paudel/Spousal Violence in Rural Nepal: Prevalence116Received Date: January 2021 Revised: March 2021 Accepted: June 2021Spousal Violence in Rural Nepal: Prevalence and Risk Factors Manusha Paudel (M. Phil.)*AbstractSpousal violence is a major public health phenomenon. It is a hidden issue in Nepal especially in rural areas. The objective of this study is to explore the prevalence and associated risk factors of spousal violence in rural Nepal. For this study, the data has been imported from Nepal Demographic and Health Survey (NDHS), 2016. This study is confined in 1510 married women from rural Nepal. Bivariate analysis and logistic regression has been applied to examine the association between variables. Study found that more than half of the respondents were 25 to 34 years aged and the overwhelming majority were (87%) Hindu. Study showed that half of the respondents had no education while 19 percent of respondents’ husbands had no education. More than a fifth (24%), more than a tenth (13%) and about a tenth (8%) had ever experienced physical, emotional and sexual violence respectively. Four in one (28%) women experienced at least one form of violence while 3 percent experienced all three forms of violence. Education of women, education of husbands and consumption of alcohol were highly associated with spousal violence. Women who had no education were 1.7 times more likely to experience violence than higher education. Likewise, other variables for example, age at first marriage, husbands’ education, consumption of alcohol and own financial account were also associating factors of spousal violence in rural Nepal. So that, attention should stand towards rural Nepal’s education for both men and women, excessive consumption of alcohol, upgrade the level of empowerment of women to mitigate intimate partner violence.Key words: Spousal violence, physical violence, sexual violence, emotional violence & married women. IntroductionSpousal violence is a complex social and public health issue and it occurs in all socio-economic, religious and cultural groups. WHO defines that it refers to any behavior within *Ms. Paudel is a Lecturer at the Department of Population Studies, Patan Multiple Campus, TU, Lalitpur, Nepal. Email: manushapaudel@gmail.comJournal of Population and Development, June 2021| Manusha Paudel/Spousal Violence in Rural Nepal: Prevalence117an intimate relationship which causes psychological, sexual and physical harm to those in relationship (Butchart et al., 2010). Culturally in Nepal, the passive role of women and power imbalance between them in husbands’ home create violence (MLJ, 2009; Pun, 2016; Emoy, 2017). On the other hand, because of the sensitivity of the subject, spousal violence is up till a hidden and under estimated social and health problem. It occurs in a pandemic proportion like cancer, HIV/AIDS, traffic accident in the world (Abeya et al., 2011). Historically, during the 1990’s violence against women has emerged as a focus of international attentions and concerns. This is first seen in 1993 when the UN General Assembly passed the declaration on the elimination of violence against women (UN, 1993). Then in 1994 at both the International Conference on Population and Development (ICPD) in Cairo and the 1995 Fourth World Conference on Women in Beijing, Women’s organizations from around the world advocated ending gender violence as a high priority (WHO, 1999). In 1996 the 49th World Health Assembly adopted a resolution declaring domestic violence as a public health priority (WHO, 1997). In 1998, United Nations Development Fund for Women launched regional campaigns in Africa, Asia, pacific and Latin America to draw attention to the issue of violence against women globally. In 1999 the United Nations Population Fund declared violence against women a public health priority (UNFPA, 1999).Research and practice over the last three decades indicates that spousal violence is a formidable social problem, intimate partner violence is violence committed by a spouse, boyfriend, girlfriend or other partner with whom an individual has or previously had on intimate or sexual relationship (Niaz, 2003). Globally, the Demographic and Health Survey (DHS) started to collect the information about intimate partner violence since 1990 in Colombia DHS survey and since 2000, DHS developed a standard module and methodology to collect the data on domestic violence (Hindin, 2008). In Nepal status of women in education and decision making is less than men however over the past decades progress has been made. The levels of maternal mortality, infant mortality have been improved. Regarding the issue of domestic violence, in 2008 Nepal’s parliament passed the domestic violence (crime and punishment) act which criminalized domestic violence for the first time. Then, the government of Nepal developed the National gender based violence plan of action in 2010 and declared this year as the year to combat gender based violence. In Nepal, the area based studies showed that the prevalence of spousal violence is high, though the National level studies on spousal violence has conducted since 2011 in the first time by the Ministry of Health and Population, Nepal Demographic and Health Survey, 2011. This survey showed that 26 percent married women had experience at least Journal of Population and Development, June 2021 | Manusha Paudel/Spousal Violence in Rural Nepal: Prevalence118one form of spousal violence during her married life. This study showed that the level of violence is high in rural area than urban area, such as 29 percent rural women experienced physical or sexual violence whereas only 25 percent urban women experienced such violence (MoHP, 2012). The overall scenario of rural Nepal showed that, it is one step back than urban in every sustainable development indicator, such as empowerment, education, income, awareness and overall development. Intimate Partner Violence is such a variable which hinders to uplift these developmental variables. Therefore, to create the peace and the decent standard of living in rural areas we cannot bypass the study on spousal violence. ObjectiveThe objective of this study is to explore the prevalence and associated risk factors of spousal violence in rural Nepal.MethodologyFor this study the data had come from Nepal Demographic and Health Survey (NDHS) conducted in 2016 a nationally representative sample survey. We used publicly available dataset from the measures DHS website. Analysis was restricted to 1510 currently married women in rural Nepal who had selected for domestic violence module and interviewed. The 2016 NDHS sample was stratified and selected in two stages in rural areas and three stages in urban areas. In rural areas wards were selected as primary sampling units and households were selected from the sample primary sampling unit (PSUs). The study protocol was approved by the Nepal Health Research Council (NHRC) and the ICF Macro Institutional Review Board in Calverton, Maryland, USA. All respondents had provided verbal informed consent prior to data collection. Therefore, independent ethical approval was not required (MoHP, 2017).The measurable outcome of the study was the experience of different forms (psychological, sexual and physical) of spousal violence a dichotomous variable. Experience between explanatory variables and experience of spousal violence was assessed via- bivariate analysis using chi-square tests and binary logistic regression was used to access the net effect of different independent variables on forms of violence. Results are considered at the significant at P<0.05. The Statistical Package for Social Science (SPSS 20.0 for Windows) software was used to analyze the data.Journal of Population and Development, June 2021| Manusha Paudel/Spousal Violence in Rural Nepal: Prevalence119FindingsStudy found that majority of the respondents (nearly two fifth) were aged 25 to 34 years while nearly one fourth were less than 25 years. An overwhelming majority of the women (87%) were Hindus. It is found that the highest percent were janjati (32%) and dalit (14%) were in lowest rank. Majority of the respondents were currently married and 3 percent were widowed whereas divorced and separated were less than one percent. Study found that half of the women had no education and one fifth respondents had primary education but nearly one fourth had secondary. In the higher level the percent had less. Similarly, the data of age at first marriage prevails the typical rural area, it showed that half of the respondents had got marriage between the age of 16 to 19 years while less than one fifth had got marriage 20 years and above. Similarly, wealth index showed than more than half of the respondents were poor.In the characteristics of respondents’ husbands, one fifth had no education; it showed the higher gap in the education of husbands and wives. Furthermore, nearly, 47 percent respondents’ husbands drink alcohol. In addition, nearly two fifth women have their own financial account. Table 1: Background characteristics of respondentsBackground CharacteristicsNumberPercentAge group of womenLess than 25 years36924.425-34 years58839.035 years and above55336.6ReligionHindu132087.4Non-Hindu19012.6EthnicityBrahimin/Chhetri40526.8Janjati49032.4Dalit21514.2Others40126.5Current marital statusMarried145596.3Journal of Population and Development, June 2021 | Manusha Paudel/Spousal Violence in Rural Nepal: Prevalence120Widowed402.7Divorced20.1No longer living together/ Sep-arated140.9Education of womenNo Education75950.3Primary29719.7Secondary35023.2Higher1036.8Age at first marriageLess than 16 years42127.916-19 years76150.420 years and above28018.5Don’t know493.2Wealth indexPoor80653.4Middle35223.3Rich35223.3Education of husbandsNo education28618.9Primary40026.5Secondary43028.5Higher33722.3Don’t know573.8Consumption of alcohol (husband)Yes70346.5No80753.5Has an financial account (women)Yes57137.8No93962.2Total1510100 Source: Nepal Demographic and Health Survey, 2016.Journal of Population and Development, June 2021| Manusha Paudel/Spousal Violence in Rural Nepal: Prevalence121Table 2 shows that a quarter women had experience physical violence. 13 percent women reported that they faced emotional violence while 8 percent women reported that they had faced sexual violence. It is found that more than a quarter (28%) women had faced at least one form of violence (either physical or/ and emotional or/ and sexual) and 3 percent women had experienced all three forms of violence.Table 2: Experience of different forms of violence among rural married women by their husbands in NepalExperience of violenceNumberPercentPhysical violenceYes36224.0No114876.0Sexual violenceYes1177.7No139392.3Emotional violenceYes20213.4No130786.6At least one form of violenceYes41927.7No109172.3All three forms of violenceYes503.3No146096.7Total1510100 Source: Nepal Demographic and Health Survey, 2016.Table 3 shows that nearly one third women had experienced at least one form of spousal violence who were aged 25 to 34 years, while who were aged less than 25 years and 35 years above had 23 percent and 27 percent spousal violence respectively. However p-value indicates that the difference is not significant. Study found that respondents’ ethnicity had significantly associated (p<0.001) with experience of at least one form of violence among rural married women in Nepal. For instance, 28 percent women who had currently married had faced at least one form of violence while all divorced women faced this situation and the current marital status and violence has been significantly Journal of Population and Development, June 2021 | Manusha Paudel/Spousal Violence in Rural Nepal: Prevalence122associated. Similarly, education of women and age of marriage were highly associated at 0.001 levels. For instance, more than one third women who had no education while one tenth women who had higher education faced at least one form of violence. In regards of husbands’ characteristics, those husbands who had no education had experienced more violence (nearly half) than those who had higher education (15%). Similarly, consumption of alcohol and form of violence are significantly associated. Table 3: Background characteristics of rural married women according to experience of violence by their husbands in NepalBackground Characteristics At least one form of violence (%)Numberp-valueAge group of womenLess than 25 years22.836925-34 years31.558835 years and above27.1553ReligionHindu26.71320Non-Hindu35.3190EthnicityBrahimin/Chhetri15.3405<0.001Janjati22.7490Dalit35.8215Others42.1401Current marital statusMarried27.81455<0.001Widowed9.841Divorced1002No longer living together/ Separated71.414Education of womenNo Education35.0760<0.001Primary28.9298Secondary16.6350Journal of Population and Development, June 2021| Manusha Paudel/Spousal Violence in Rural Nepal: Prevalence123Higher9.7103Age at first marriageLess than 16 years33.7421<0.00116-19 years27.776120 years and above17.9280Wealth indexPoor24.3806<0.05Middle32.9353Rich30.7352Education of husbandsNo education44.1286<0.001Primary29.9401Secondary25.4429Higher14.5337Consumption of alcohol (husband)Yes36.6703<0.001No20.1807Has an financial account (women)Yes24.8572<0.01No29.5938Total27.71510 Source: Nepal Demographic and Health Survey, 2016.This study had used binary logistic regression model to show the effects of independent variables on dependent variable. In this model, had included different socio-demographic and economic variables as independent variables and experienced at least one form of spousal violence as a dependent variable. Table 4 contains the odd ratio of logistic regression coefficient and 95 percent confidence interval for odd ratios for each category. From the fitted model shows that women who had aged 25-34 years had more likely (OR=1.259) to experience violence from their husband. On the other hand, women who were Hindu, had less likely (OR=0.700) to experience spousal. Journal of Population and Development, June 2021 | Manusha Paudel/Spousal Violence in Rural Nepal: Prevalence124Likewise, those women who got primary education only had 1.7 times more like to experience spousal violence than who had got higher education. Similarly, those women who got married before 16 years were 1.7 times more likely to experience violence than who got married 20 years and after. Women whose husbands had no education were 2.9 times more likely to experience spousal violence than women whose husbands had higher education. Similarly, women whose husband did not take alcohol had less likely to experience violence than who had taken alcohol. In addition, women who have not own financial account had more likely (OR=1.04) to experience violence by their husbands.Table 4: Odd Ratio (OR) and 95% Confidence Interval (CI) for having experienced of violence from their husbands among rural married women in NepalPredictorsOdd RatioP-Value95% Confidence Interval (CI)Age group of womenLess than 25 years0.9340.7380.62-1.3925-34 years1.2590.1320.93-1.6935 years and above1ReligionHindu0.7000.0630.48-1.02Non-Hindu1EthnicityBrahimin/Chhetri0.3250.0010.21-0.50Janjati0.3590.0010.24-0.52Dalit0.4890.0010.32-0.74Others1Education of womenNo Education1.7070.1820.77-3.74Primary1.4440.3630.65-3.18Secondary1.1400.7380.53-2.45Higher1Age at first marriageLess than 16 years1.7060.2770.63-1.1316-19 years1.4180.1330.46-1.020 years and above1Journal of Population and Development, June 2021| Manusha Paudel/Spousal Violence in Rural Nepal: Prevalence125Wealth indexPoor0.6720.0340.46-0.97Middle0.8080.2580.55-1.17Rich1Education of husbandsNo education2.9000.0011.78-4.71Primary1.8000.0111.14-2.82Secondary1.5130.0550.99-2.30Higher1Consumption of alcohol (husband)No0.3240.0010.24-0.43Yes1Has an financial account (women)No1.0480.7420.79-1.38Yes1 Source: Nepal Demographic and Health Survey, 2016.DiscussionThis study has tried to explore the prevalence and risk factors of spousal violence in rural Nepal. The present study shows that spousal violence is common among rural women and it affects the overall development in rural areas and indicates a need for an effective program on spousal violence issue. In a survey of reproductive-aged women in rural Bangladesh, 47 per cent of the women reported having ever been beaten by their husbands, 19 per cent within the previous year (Schuler, 1996).This study found that physical violence is high (24%), than emotional and sexual violence. In every one in four women has faced physical violence. A study of physical violence in slum areas of Kathmandu valley showed the highest prevalence (65%) during her life time (Paudel, 2019). Similarly a study on violence in Africa found that physical violence is highest and almost half of the countries’ lifetime prevalence is over 40 percent (UN, 2015).Journal of Population and Development, June 2021 | Manusha Paudel/Spousal Violence in Rural Nepal: Prevalence126The bivariate analysis shows some variables such as age, religion, ethnicity, age at first marriage, education, consumption of alcohol and financial account, which are important in explaining spousal violence. Bivariate analysis shows the significant association between dependent and independent variables while multivariate analysis support the findings of bivariate analysis. This study has found that women who were age 25-34 had experienced high IPV than early and later age groups. While, in Europe women in youngest group had high prevalence of spousal violence than older aged (UN, 2015). The WHO multi country study reveals that younger women, especially those age (15-19 years) are at higher risk of experiencing spousal violence (Erulkar, 2013). In Nepal, a study conducted in young couple found that 50 percent young married women falls on sexual violence because of lack of education (Puri, 2008).This study found that women who had no education were 1.7 times more likely to experience violence from their husband. Similar another study which had done at rural Nepal in 2011, found that women’s educational deprivation is the most influential risk factor of spousal violence in Nepal (Lamichhane et al., 2011). In addition a study in India showed that women which have lower level of education had higher risk of spousal violence (Akerson et al., 2008). That type of situation is common in most of developing countries (Niaz, 2003; Hajjar, 2006). Result of this study found that those women who had got married before 16 years were more likely to experience violence than who got married 20 years and above. Similarly a study of UNICEF in nine countries, found that girls who were married before 18 years were more likely to experience domestic violence than peers who married later. Another study in India found that early marriage is associated with low involvement in the decision to marry fewer interactions with one’s spouse and elevated risk of intimate partner violence (Erulkar, 2013). On the other hand, study found that women, who were poor wealth index, had less likely to experience spousal violence than rich. But most of the literatures found that who have poor economic status have higher risk of violence (MoHP, 2012; Puri, 2008). Similarly another study found that, lower-class men often have lack of power and authority in their work environments; they may construct rigid, aggressive models of masculinity in the home (Messerschmidt, 1993) and literature hints that besides level of income, husband’s educational status determines the level of higher and lower class. This study found that those women, whose husbands had no education, had 2.9 times more likely to experience spousal violence than who had higher education. Similarly, a WHO multi country study on Women’s health and domestic violence showed that the highly educated group of husbands had lower odds ratio’s for spousal violence in 10 out of 14 Sites (Abramsky, 2011). This study found that, husbands who did not Journal of Population and Development, June 2021| Manusha Paudel/Spousal Violence in Rural Nepal: Prevalence127consume alcohol were less likely to experience spousal violence. Similarly, a Kenyan DHS report found that, alcohol use by the male partner was associated with women’s ever experiencing either physical or sexual abuse (Kimuna, 2008). In Nepal, a study also reveals that respondents whose husband consumed alcohol were more likely to experience sexual coercion than those whose husband did not consume alcohol (Adhikari, 2009). Another study in Nepal found that nearly 40 per cent women fall on sexual violence whose husband use alcohol (Puri , 2008). Lastly, results showed that, women who had not financial account were more likely to experience violence than who had financial account. Literatures also showed that economic autonomy helps to reduce the violence (Ellsberg, 2015). ConclusionThis study found that spousal violence of married women in rural Nepal is prevalent. Physical violence is high than others. Findings of the study help to know the factors, which are important effect on spousal violence. This study highlights that age at marriage, women’s and husbands’ education, consumption of alcohol and own financial account are important variables to decrease the level of spousal violence in rural areas. Intervention should focus towards upgrade the level of education on both husbands and wives, raise the awareness on early age marriage, consumption of alcohol; and move towards women’s financial inclusion to break the cycle of spousal violence and overall development of rural area in Nepal. ReferencesAbeya, G. S., Afework, F. M., & Yalew, W. A. (2011). Intimate partner violence against women in western Ethiopia: Prevalence, patterns and associated factors. https://www.biomedcentral.com/14/1-24 /58/11/9/3. Abramsky, T., Watts, C. H., Garcia-Moreno, C., Devries, K., Kiss, L., Ellsberg, M., Jansen, H. A., & Heise, L. (2011). What factors are associated with recent intimate partner violence? Findings from the WHO multi-country study on women’s health and domestic violence. BMC Public Health, 11(109).Ackerson, L. K. I., Kawachi, B., & Subramanian, S. V. (2008). Effects of individual and proximate educational context on intimate partner violence. A population based study of women in India. American Journal of Public Health, 98(3), 507-14.Adhikari, R., & Tamang, J. (2009). Sexual coercion of married women in Nepal. BMC Women’s Health 10(31). https://doi.org/10.1186/1472-6874-10-31Journal of Population and Development, June 2021 | Manusha Paudel/Spousal Violence in Rural Nepal: Prevalence128Butchart, A., Garcia-Moreno, C., & Mikton, C. (2010). Preventing intimate partner and sexual violence against women: Taking action and generating evidence. World Health Organization, GenevaEllsberg, M., Arango, D. J., Morton, M., Gennari, F., Kiplesund, S., & Contreras, M. (2015). Prevention of violence against women and girls: What does the evidence say? 385(9977), 1555–66. https://doi.org/10.1016/S0140-6736(14)61703-7 PMID: 25467575Emery, C. R., Thapa, S., & Wu, S. (2017). Power and control in Kathmandu: A comparison of attempted power, actual power, and achieved power. Violence against Women, 23(4), 482–502.Erulkar, A. (2013). Early marriage, marital relations and intimate partner violence in Ethiopia. International Perspectives on Sexual and Reproductive Health. Guttmacher Institute, 39(1), 6-13. https://www.jstor.org/stable/23408822 on 17/03/2014.Hajjar, L. (2006). Religion, state power, and domestic violence in Muslim societies: A framework for comparative analysis. Law and Social Inquiry, 29(1).Hindin, M. J., Sunita, K., & Donna, L. A. (2008). Intimate partner violence among couples in 10 DHS countries: Predictors and health outcomes. DHS Analytical Studies, (18). Calverton, Maryland, USA.Kimuna, S. R., & Djamba, Y. K. (2008). Gender based violence: Correlates of physical and sexual wife abuse in Kenya. Journal of Family Violence, 23 (5), 333-42.Lamichhane, P., Puri, M., Tamang, J., & Dulal, B. (2011). Women’s status and violence against young married women in rural Nepal. BMC Women’s Health, 11(19). https://doi.org/10.1186/1472-6874-11-19.Messerschmidt, J. W. (1993). Masculinities and crime: Critique and reconceptualization of theory: Laham, MD: Rowmah and Littlefield.Ministry of Health and Population (MoHP). (2017). Nepal demographic and health survey 2016. Kathmandu: Ministry of Health and Population New ERA and IFC International, Nepal. Ministry of Health and Population (MoHP). (2012). Nepal demographic and health survey 2011. Kathmandu: Ministry of Health and Population New ERA and IFC International, Nepal. Ministry of Law and Justice (MoLJ). (2009). Domestic violence (offence and punishment) Journal of Population and Development, June 2021| Manusha Paudel/Spousal Violence in Rural Nepal: Prevalence129act, 2009. Kathmandu, Nepal.Niaz, U. (2003). Violence against women in south Asian countries. Women’s Mental Health, 6(3), 173-184.Paudel, M. (2019). Prevalence and factors associated with physical intimate partner violence in slum areas of Kathmandu valley. Electronic Research Journal of Social Sciences and Humanities, 1(4).Pun, K. D., Infanti, J. J., Koju, R., Schei, B., & Darj, E. (2016). Community perceptions on domestic violence against pregnant women in Nepal: A qualitative study. Global Health Action, 9(1). https://doi.org/10.3402/GHA.V9.31964Puri, M. (2008). Exploring the nature and reasons associated with sexual violence within marriage among young couples in Nepal. Annual meeting of PPA. Sheraton New Orleans, USA.Schuler, S. R. (1996). Credit programmes, patriarchy men’s violence against women in rural Bangladesh. Social Science Medicine, 4 (12), 1729-1742.United Nations Population Fund (UNFPA). (1999). Violence against women. New York.United Nations (UN). (1993). Declaration on the elimination of violence against women. General Assembly. United Nations.United Nations (UN). (2015). Violence against women. The World’s women. Worlds Women 2015_chapter6_t.pdf (un.org)World Health Organization (WHO). (1999). Putting women’s safety first: Ethical and safety recommendations for research on domestic violence against women, WHO/EIP/GPE/99.2, Geneva.World Health Organization (WHO). (1997). Violence against women: A priority health issue. WHO/FRH/WHD/97.8, Geneva.View publication stats | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/363563161The Little Fire Ant (Hymenoptera: Formicidae): A Global PerspectiveArticle in Annals of the Entomological Society of America · September 2022DOI: 10.1093/aesa/saac016CITATION1READS1534 authors, including:Some of the authors of this publication are also working on these related projects:Platform for Ecological Restoration Research Infrastructure (PERRI) View projectAustralasian Journal of Environmental Management View projectMichelle MontgomeryUniversity of Canberra7 PUBLICATIONS 28 CITATIONS SEE PROFILECasper VanderwoudeUniversity of Hawaiʻi at Mānoa43 PUBLICATIONS 667 CITATIONS SEE PROFILEA. J. J. LynchACT Government66 PUBLICATIONS 926 CITATIONS SEE PROFILEAll content following this page was uploaded by Casper Vanderwoude on 03 October 2022.The user has requested enhancement of the downloaded file.1© The Author(s) 2022. Published by Oxford University Press on behalf of Entomological Society of America. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.ReviewThe Little Fire Ant (Hymenoptera: Formicidae): A Global PerspectiveM. P. Montgomery,1,2,4, C. Vanderwoude,2 M. Lintermans,1 and A. J. Jasmyn Lynch1,31Institute for Applied Ecology, University of Canberra, ACT 2601, Australia,2Hawai`i Ant Lab, Pacific Cooperative Studies Unit, University of Hawai`i, Hilo, HI 96720, USA,3Conservation and Water Planning Policy Branch, Environment, Planning, and Sustainable Development, ACT Government, Australia, and 4Corresponding author, e-mail: michelle.montgomery@littlefireants.comSubject Editor: Gadi VP ReddyReceived 19 April 2022; Editorial decision 1 August 2022 Abstract Wasmannia auropunctata (Roger) is an invasive tramp ant species that has been transported globally since [at least] the early twentieth century. It is often claimed that despite the negative impacts associated with this species and its listing among the world’s worst invasive species, very little research attention has been paid to W. auropuntata. Although the need for future research exists, there is currently a considerable body of research from around the world and spanning back to the 1920’s on this species. Here we synthesize over 200 peer re-viewed research manuscripts, book chapters, conference presentations, and media reports of new distributions spanning 1929–2022 culminating in a comprehensive literature review on W. auropunctata. This review covers all current knowledge on this species and is intended to serve as a quick reference for future research and pro-vide the reference resources for those seeking more in-depth information on specific topics. Topics included in this review include taxonomic identification, current global distribution and pathways, life history, impacts, de-tection, and control. We discuss where consensus and ambiguity currently lie within the research community, identify contextual considerations for future researchers when interpreting data, and suggest where we believe more research or clarifications are needed.Key words: little fire ant, distribution, biology, impact, controlInvasive alien species (IAS) are species that have been introduced be-yond their native range and negatively impacts those new areas. The negative impacts associated with IAS range from human health risks to reduced species diversity, plant and animal disease transmission, and altered ecosystem processes (Resnik 2018, Liang 2019). Insects are particularly easy to transport inadvertently from one place to another (Meurisse et al. 2019) and eusocial insects are predisposed to be successful invaders wherever they are introduced (Bertelsmeier 2021, Eyer and Vargo 2021) Specifically, invasive Hymenoptera are well known for causing multitiered negative impacts wherever they are introduced.Ants (Hymenoptera: Formicidae) are among the most species-rich taxa in the animal kingdom with over 12,000 described species (Ward 2007). They are important for healthy ecological functioning (Andersen 1988, Abbott 1989, Folgarait 1998, Del Toro et al. 2012) but some cause negative ecological and human health impacts. The worst invasive ant species are also notable ‘tramp’ species due to the ease at which they are transported by anthropogenic means and their ability to thrive in areas disturbed by humans (Wilson and Taylor 1967, McGlynn 1999, Loope and Krushelnycky 2007). Other factors attributed to the success of some invasive ants include a combination of biological and behavioral traits such as polygyny, unicoloniality and low intraspecific aggression, high interspecific aggression, reproduction within the nest, and colony founding via budding instead of nuptial flights (Hölldobler and Wilson 1977, Brandao and Paiva 1994, Passera 1994, Jourdan 1997a, Helms and Vinson 2002, Holway et al. 2002).The Little fire ant, Wasmannia auropunctata (Roger) (Hymenoptera: Formicidae) is an invasive tramp ant listed as one of the world’s worst 100 IAS (Lowe et al. 2000), and has an alarming rate of spread globally with 43% of known new intro-ductions occurring since the year 2000 (Wetterer 2013, Gruber et al. 2016, Espadaler et al. 2018, GBIF.org 2021, Vanderwoude et al. 2021, Chen et al. 2022). Consequently, research interest in this species has increased over the past 20 yr. Previous literature re-views on W. auropunctata to date have either focused on its global Annals of the Entomological Society of America, XX(X), 2022, 1–22https://doi.org/10.1093/aesa/saac016Review Downloaded from https://academic.oup.com/aesa/advance-article/doi/10.1093/aesa/saac016/6697138 by guest on 16 September 20222Annals of the Entomological Society of America, 2022, Vol. XX, No. XXor local distribution, with life history, control, and impacts being minor components (Wetterer and Porter 2003, Herrera and Causon 2008, Wetterer 2013, Vanderwoude et al. 2015) or have focused on the latter three topics while omitting distribution information (Bousseyroux et al. 2019). More recently there has been consider-able research into life history, control, and impacts and reports of new distribution data that has yet to be synthesized.Here, we provide the first comprehensive literature review synthesizing all research on this species before June, 2022. The re-view is timely given the ongoing spread of this serious global pest and the need for research into cost-effective control methods. We review literature spanning the past 90 yr detailing its taxonomy, life history, distribution, range expansion, impacts, detection, and control of W. auropunctata. We provide an overview of its signifi-cance as an invasive species and identify priority areas for further research to fill knowledge gaps. Literature included in the review met two primary criteria: 1) it was published in a peer-reviewed journal, book section, government report or document, university scientific research, or extension report; and 2) W. auropunctata was either the primary focus or one of the primary foci of the study. Exceptions to these criteria were when the information obtained was novel in context and not reported elsewhere, such as new research presented during a conference or media report detailing a first detection for a locality.Taxonomy and SystematicsCommonly referred to as little fire ant, W. auropunctata was first described by Roger in 1863 and originally included as a species of Tetramorium (Smith 1929, Nickerson 1983). After the genus Wasmannia was described by Forel in 1893, some arguments per-sisted around whether the species belonged within Wasmannia or the previously described genus Ochetomyrmex Mayr (Nickerson 1983, Longino and Fernández 2007). However, Longino and Fernández (2007) conducted a taxonomic review of Wasmannia, clearly distinguishing it from Ochetomyrmex, and provided a re-vised key for Wasmannia, including W. auropunctata.Currently, ten Wasmannia species have been described, with W. auropunctata being the most common and widely distributed (Longino and Fernández 2007, Cuezzo et al. 2015). Although 21 Wasmannia species have been described historically, over half of these species were subsequently determined to be synonyms of other species and nine are currently considered to be synonymous with W. auropunctata (Table 1) (Longino and Fernández 2007).Generalized ant taxonomy diagrams are presented in Fig. 1 for those unfamiliar with ant taxonomy or taxonomic terminology. Detailed morphological descriptions of W. auropunctata may be found in Ulloa-Chacon and Cherix (1990), Wetterer and Porter (2003), and Longino and Fernández (2007), but each description includes different morphological characteristics. Summarizing across all three descriptions, W. auropunctata may be identified by the fol-lowing morphological characteristics:Workers (Fig. 2) are tiny (~1.2–1.5 mm), monomorphic, rust-colored to pale yellow-brown myrmicine ants with two pedicle segments (petiole and post petiole) and long propodeal spines. The petiole node is roughly quadrate, approximately as high as it is wide. The hind margin of the node is slightly shorter than the fore margin when viewed in profile and meets the peduncle at nearly a 90-de-gree angle. The head and body are heavily sculptured with transverse Table 1. Past taxonomic synonyms for Wasmannia auropunctata previously described as separate species (Longino and Fernández 2007)Taxonomic synonymsWasmannia atomum (Santschi 1914) W. australisEmery 1894W. glabraSantschi 1894W. laevifronsEmery 1894W. obscuraForel 1912W. panamana(Enzmann 1947)W. pullaSantschi 1931W. nigricansEmery 1906W. rugosa(Forel 1886)Fig. 1. General ant taxonomy diagrams of the whole body as viewed in profile (left) and the head as viewed from the front (right). Diagrams created by Eli Sarnat and used with permission.Downloaded from https://academic.oup.com/aesa/advance-article/doi/10.1093/aesa/saac016/6697138 by guest on 16 September 20223Annals of the Entomological Society of America, 2022, Vol. XX, No. XXrugules and reticulations, and sparse erect setae. The antenna con-sists of 11 segments. Funicular segments of the antenna gradually enlarge with the apical 2 segments distinctly larger than the rest and forming a club. Antennal scrobes are present, although shallow, and distinguishable by the presence of two prominent frontal carinae. Although generally considered ‘monomorphic’, aberrant worker morphs are common in large sample collections.Reproductives (Figs. 3 and 4) are approximately three to four times larger than the workers (queens: 4.5–5.0 mm, males 4.2–4.5 mm). Two sympatric queen morphs have been documented, large- and small-headed queens. The antenna, sculpturing, and pu-bescence of the queens are similar to that of the workers. Queen coloration is uniformly dark brown but younger queens may be light brown and darken with age. Propodeal spines are present but shorter than those of workers. The node of the petiole gradually tapers up-ward but with distinct anterior and posterior angles along the dorsal margin. Wings may or may not be present. When present, the medial and SM1 cells are enclosed and the discoidal cell is absent on the fore wing (Fig. 3c). When not present, the sutures where the wings were detached are clearly visible.Males (Fig. 4) are dark brown with yellowish antenna, legs, and genitalia. Antenna are long, 13-segmented, and without an apical club. Propodeal spines are absent, but the propodium angles sharply downward at a 90-degree angle. The petiolar node is large and dor-sally rounded rather than angular. The parameres of the genital valve are long, curved intero-ventrally, and apically rounded (Fig. 4c.).Life HistoryGeneticsCytogenetics and molecular genetics provide insights on evolu-tionary and biological processes as well as population dynamics and species identification. To date, the mitochondrial genome has been mapped and macrosatellite markers have been identified for W. auropunctata which provided the tools necessary to conduct im-portant foundational research (Fournier et al. 2005b, Souza et al. 2009, de Souza et al. 2011, Duan et al. 2016, Silva et al. 2018). The use of various genetic analyses have allowed researchers to trace the evolutionary history of this species (Chifflet et al. 2016), trace and track historical and current population expansions (Foucaud et al. 2010b, Chifflet et al. 2016, Coulin et al. 2019), distinguish native from exotic populations (Foucaud et al. 2010b), trace the origins of exotic populations (Foucaud et al. 2010b, Coulin et al. 2019), and identify certain biological and behavioral traits linked to inva-sive potential (Fournier et al. 2005b; Foucaud et al. 2006, 2010b; Mikheyev et al. 2009; Souza et al. 2009; Vonshak et al. 2009; Rey et al. 2011; Tindo et al. 2012). Additionally, genetics can help to iden-tify when and where evolutionary adaptations occurred that has led to W. auropunctata being able to invade such a wide range of eco-systems and climates (Rey et al. 2012, Foucaud et al. 2013, Chifflet et al. 2016, Coulin et al. 2019).The field of genetics is rapidly expanding with new technolo-gies for phylogenetics, species identification, detection, and even pest control. Two technologies with especially promising applica-tions are the use of environmental DNA (eDNA) for detection and RNA interference (RNAi) for control of pest ants. Trace amounts of genetic material are now able to be detected. That, with the advent of metabarcoding, has led to the growing practice of using eDNA to detect species presence from water or substrate samples without direct observation or collection of the target species (Kudoh et al. 2020, Uchida et al. 2020). Proof of concept for using eDNA as an ant detection tool has been reported for Linepithema humile Mayr (Yasashimoto et al. 2021). However, eDNA technology requires fur-ther development and testing before it can be deemed as a reliable tool for the detection of other species and under different scenarios. Development of RNAi technology for ant control appears to be promising as a future alternative to conventional pesticides. This is the process by which double-stranded RNA (dsRNA) or DNA (dsDNA) is used to stop the normal functioning of messenger RNA (Allen 2021). A construct of dsRNA or dsDNA may be delivered to the target pest via genetically modified crops, sprays, and bait de-livery systems (Cagliari et al. 2019) and, when integrated into cells, interferes with gene transcription and effectively activates or silences gene expression. However, efficacy varies depending on target spe-cies and delivery system (Allen 2021). Functionality of RNAi on ants has been demonstrated (Allen 2021, List et al. 2022), but consider-ably more work is needed before the technology can be considered a viable control method (Allen 2021, List et al. 2022). Species-specific target genes must be identified and corresponding dsRNA or dsDNA constructs need to be developed. Degradation of dsRNA and dsDNA due to exposure of digestive enzymes, through trophallaxis, and under field conditions for various delivery systems is currently un-known and needs further investigation (Allen 2021). Finally, stand-ardized testing and evaluation procedures need to be agreed upon Fig. 2. Wasmannia auropunctata worker profile (a) and detail or head (b.). Photographs by: E. M. Sarnet, specimen CASENT 0171093. From www.antweb.orgDownloaded from https://academic.oup.com/aesa/advance-article/doi/10.1093/aesa/saac016/6697138 by guest on 16 September 20224Annals of the Entomological Society of America, 2022, Vol. XX, No. XXFig. 3. Wasmannia auropunctata queen profile (a), detail of head (b), and wing diagram (c). Photographs by: A. Nobile, specimen CASENT 0102747. From www.antweb.orgFig. 4. Wasmannia auropunctata male profile (a), detail of head (b), and parameres (c). Profile and head photographs by: A. Nobile, specimen CASENT0102748. Parameres photograph by: A. Nobile, specimen CASENT 0173250. From www.antweb.org.Downloaded from https://academic.oup.com/aesa/advance-article/doi/10.1093/aesa/saac016/6697138 by guest on 16 September 20225Annals of the Entomological Society of America, 2022, Vol. XX, No. XXby the scientific community as it is with conventional pesticides to establish baseline expectancies for field efficacy trials (Allen 2021, List et al. 2022).Reproduction and DevelopmentRudimentary aspects of W. auropunctata reproduction and develop-ment were first described by Ulloa-Chacon and Cherix (1990). Only queens lay eggs and fecundity fluctuates over a queen’s lifespan, sug-gesting that fecundity is influenced by age and colony composition (Ulloa-Chacon and Cherix 1990). While worker brood is produced regularly to maintain nest population, reproductives are produced when high worker:brood ratios occur in the nest or when queen fe-cundity is low (Ulloa-Chacon and Cherix 1990).Wasmannia auropunctata reproduction is unique and com-plex. They are not a typical haplodiploid species as with most so-cial Hymenoptera (Fournier et al. 2005a). In typical haplodiploid reproduction systems, diploid females (queens and workers) are the progeny of sexual reproduction and haploid males are produced through arrhenotokous parthenogenesis, males developing from un-fertilized eggs (Normark 2003). However, W. auropunctata exhibit a complex reproduction system, where the reproductive mode varies between different populations (Foucaud et al. 2007, 2009, 2010a).Reproductive castes are produced in several ways. Haploid males are produced either via arrhenotokous parthenogenesis (arrhen-otoky), males developing from unfertilized eggs, or via androgenesis, males developing from fertilized eggs through the elimination of the entire maternal genome. The latter process is rare in the animal kingdom and results in male clones that are genetically identical to their fathers. Diploid queens are produced through normal sexual reproduction or via automictic thelytokous parthenogenesis with central fusion (Rey et al. 2011). Automictic thelytokous partheno-genesis occurs from the fusion of two meiotic oocytes. Unusually low recombination rates during the meiotic division process result in clonal diploid queen lineages (Rey et al. 2011).The occurrence of one reproductive mode or another is linked to W. auropunctata ecological dominance (Foucaud et al. 2009). Typical haplodiploid reproduction (sexually produced queens and arrhenotokous males) occurs primarily among nondominant varie-gate populations whereas clonal reproduction (automictic par-thenogenesis and androgenesis) occurs primarily among dominant variegate populations (Foucaud et al. 2009, 2010a). Even though there is a trend linking dominance and reproductive mode, this is not a strict rule. Occasionally dominant sexual population, clonal nondominant populations, and rare occurrences of sexual repro-duction within clonal populations have been detected (Foucaud et al. 2006, 2009, 2010b; Tindo et al. 2012). Interestingly, whether males are produced via arrhenotoky or androgenesis is maternally determined (Rey et al. 2013a). Queens from typical haplodiploid nests never produce male clones. They only produce arrhenotokous males whereas clonal queens produce androgen male clones, regard-less of insemination or genetic lineage of the fathers (Foucaud et al. 2010a, Rey et al. 2013a). While it is possible for parthenogenic queens to produce arrhenotokous males, this phenomenon has only been documented during a single laboratory experiment (Tindo et al. 2012) and has not been detected among wild populations. It is possible this occurs at such low rates that research to date has failed to detect it.Speculation around what leads to the expression of one repro-ductive mode over another is ongoing. An early hypothesis was that endosymbiotic bacteria, such as Wolbachia, could be responsible for the shift from typical haplodiploid reproduction to clonal repro-duction (Rey et al. 2013b). Wolbachia is a widespread arthropod endosymbiont that is maternally inherited and influences sex de-termination during reproduction, including through the initiation of thelytokous parthenogenesis (Werren et al. 2008). Wolbachia is a known endosymbiont of W. auropunctata, but Wolbachia is less prevalent in clonal populations than in typical haplodiploid populations suggesting that it was not likely responsible for the reproductive difference (Rey et al. 2013b). The current hypothesis surrounding the expression of one reproductive mode over another involves the influence of ecological factors. Some sources suggest W. auropunctata are typically found amid floodplains (i.e., creek beds) within primary forests and clonal reproduction may have arisen as an evolutionary response to the repeated disturbance ex-perienced in such habitats (Rey et al. 2012, Chifflet et al. 2018). Nests in floodplains would likely be under pressures such as frag-mentation, transportation, and colony founding with every flooding event. Studies on the emergence of clonal reproduction in plants have suggested this reproduction mode may have evolved as an al-ternative lifecycle loop allowing populations to persist despite the absence of the necessaries sustaining the species normal lifecycle, such as a mate (Honnay and Bossuyt 2005). It is possible that the reproductive plasticity of W. auropunctata evolved in a similar way. Repeated flooding events likely increased the frequency of popula-tion fragmentation and a need to establish and persist for a short period of time without males. If this is the case, this phenotypic plasticity or adaptation to ecological pressures of floodplain habi-tats has undoubtedly become a leading factor contributing to their success as an invading species.The unusual reproductive system of Wasmannia auropunctata has been implicated as an important factor contributing to the suc-cessful establishment of small founder colonies (Mikheyev et al. 2009). For newly introduced species, the establishment phase is particularly difficult (Foucaud et al. 2009, Mikheyev et al. 2009). In general, founding populations of an introduced species frequently fail to establish as a consequence of the principal cost of sex; the need to find a mate (Smith 1978). For those that do establish, the genetic diversity of the population is low due to few individ-uals surviving and reproducing. Such a genetic bottleneck leads to inbreeding and loss of heterozygosity over time within sexually re-producing populations and, theoretically, loss of fitness. However, reproductive plasticity in W. auropunctata allows for the preser-vation of heterozygosity when genetic diversity is low and avoids the consequences of inbreeding (Foucaud et al. 2010a, Rey et al. 2013a). It is likely that clonal reproduction via thelytokous par-thenogenesis allows for the persistence of genetic adaptations re-sponsible for W. auropunctata being able to successfully invade a wide variety of human modified habitats. Indeed, introductions of single female and male genotypes can give rise to area-wide infest-ations as seen in New Caledonia (Foucaud et al. 2006), Hawai`i (Mikheyev et al. 2009), Cameroon (Mbenoun Masse et al. 2011) and Israel (Vonshak et al. 2009). Rarely, recombination, muta-tion, and sexually produced queens within clonal populations have been observed which would add small amounts of genetic diversity within otherwise clonal populations (Foucaud et al. 2006, Vonshak et al. 2009, Tindo et al. 2012).Although insemination does not contribute to W. auropunctata genetic diversity and is not strictly required for a queen to lay vi-able eggs, insemination is necessary to maintain egg development and hatching success (Miyakawa and Mikheyev 2015). Despite the possibility of virgin queens producing viable brood, the rate of suc-cessful hatching, pupation, and emergence among uninseminated brood is too low for colony maintenance (Miyakawa and Mikheyev 2015). Sex is therefore an essential part of both reproductive modes.Downloaded from https://academic.oup.com/aesa/advance-article/doi/10.1093/aesa/saac016/6697138 by guest on 16 September 20226Annals of the Entomological Society of America, 2022, Vol. XX, No. XXNesting and Population DynamicsRather than building subterranean nests and mounds like many other ants, W. auropunctata prefer to nest opportunistically in warm, moist, and shaded areas (Ulloa-Chacon and Cherix 1990, Wetterer and Porter 2003). They can exploit ecological or man-made features on the ground, in tree canopies, and in other vegetation (Clark et al. 1982, Wetterer and Porter 2003, Le Breton et al. 2005, Mikissa et al. 2013, Álvarez et al. 2018, Santos et al. 2019, Wisniewski et al. 2019). Ideal nesting locations include leaf litter, under rocks and logs, loose bark, clusters of moss, epiphytes, and plant petioles (Wetterer and Porter 2003), but the species also frequently nests in electrical sockets, vehicles, machinery, pallets, metal pipes, furniture, garbage piles, and anywhere that provides shade, moisture, and refuge.Individual nests are small and often difficult to identify (Clark et al. 1982). Nests are comprised of brood (eggs, larvae, and pupae), sterile workers, multiple queens, and occasionally males. A single nest may be divided into separate aggregations, of which three types have been described; workers + brood + queens, workers + brood, and workers only (Ulloa-Chacon and Cherix 1990). The purpose of separate aggregations and caste segregation has not been identified but could be due to capacity limits at the nest location, protection and defense of brood and queens, or other reasons. Such nesting habits allow W. auropunctata to utilize all available nesting sites in an area and support extremely high population densities (Hölldobler and Wilson 1995, Souza et al. 2008).Typical queen:worker ratios for a W. auropunctata nest have been described as ranging between 1:250 and 1:500 with up to 16 queens per nest (Ulloa-Chacon and Cherix 1990). However, one cannot ignore the question of what defines an individual nest within a supercolony and when a ‘single’ nest may be comprised of separate aggregations. Indeed, many more queens have been observed within proximity to each other amid large infestations (personal observa-tion) but it is uncertain how common this is or what factors influ-ence the number of queens per nest. As with other eusocial insects, older workers forage outside of the nest while younger workers re-main within the confines of the nest and care for queens and brood, a process referred to as temporal polytheism (Robinson et al. 1994, Ortiz-Alvarado et al. 2021). Queens are typically relegated to laying eggs but will forage and care for brood when worker populations are low (Ortiz-Alvarado and Rivera-Marchand 2020). This behavior likely contributes to W. auropunctata surviving the colony founding period when introduced to new locations and during the budding process.Two W. auropunctata ecological variegates are known throughout their native range: nondominant and dominant (Levings and Franks 1982, Tennant 1994, Foucaud et al. 2009, Orivel et al. 2009). The nondominant variegate typically reproduces sexually and is restricted to natural, primary forests throughout its native range. Although common, nests are diffusely dispersed throughout the forests and these W. auropunctata do not display the same level of interspecific aggression as the dominant variegate (Tennant 1994, Salguero Rivera et al. 2011). The dominant variegate typic-ally reproduces clonally and is widely distributed in human modified habitats through its native and introduced ranges (Foucaud et al. 2009, Orivel et al. 2009, Chifflet et al. 2018). The dominant varie-gate displays high levels of interspecific aggression and other be-havioral traits that allow W. auropunctata to successfully dominate other ant species. Interestingly, before the 1980s, the nondominant variegate was unknown (Levings and Franks 1982) despite the current hypothesis that the dominant variegate arose from nearby nondominant populations (Foucaud et al. 2007). The behavioral and physiological plasticity of W. auropunctata and correlation between human disturbance and dominance has led researchers to label W. auropunctata as a ‘disturbance specialist’ (Majer 1999, Solomon and Mikheyev 2005, Foucaud et al. 2009, Orivel et al. 2009, Chifflet et al. 2018, Achury et al. 2020). Despite this designation, it is still unclear whether certain types and intensity of disturbances create unsuitable habitat (Rojas et al. 2021). For example; although W. auropunctata is a well-known agricultural pest, they may not be able to invade annual cropping systems as well as perennial crops and orchards due to annual cropping systems undergoing frequent harvesting, tillage, and replanting (Rojas et al. 2021). Additionally, development of rural areas and increased urbanization dramatic-ally alters landscapes in a way that may reduce suitable habitat for W. auropunctata while becoming more suitable to other ant species (Mbenoun Masse et al. 2021). It is likely the effects of disturbance on W. auropunctata invasion vary from one location to another and this should be looked into further. Low intraspecific aggression al-lows workers to freely move between nest aggregates and share food resources, thus forming three dimensional ‘supercolonies’ (Foucaud et al. 2009). While both the nondominant and dominant varie-gates build supercolonies, those of the nondominant variegates are smaller and multiple, genetically distinct, supercolonies are present throughout a given landscape (Foucaud et al. 2009). Alternately, the dominant variegate is often unicolonial, building a single expansive supercolony with undefined nest boundaries and extending over hun-dreds of kilometers (Hölldobler and Wilson 1977, Le Breton et al. 2004, Errard et al. 2005). This unicolonial social organization differs from most ant species (Hölldobler and Wilson 1990) but is common among invasive ants (Holway et al. 2002). Such cooperative net-works are conducive to exponential population growth, easily sup-porting densities of 20,000 workers and 37–52 queens per square meter (200 million workers and 370,000–520,000 queens per ha) (Ulloa-Chacon and Cherix 1990, Souza et al. 2008). This popula-tion estimate is over five times that of the estimate for polygynous Solenopsis invicta Buren by Macom and Porter (1996) and may be among the highest of all ant species in the world. Hölldobler and Wilson (1990, p. 63) described W. auropunctata as ‘creating a living blanket of ants that kill and eat nearly all other ants in their path’.The underlying cause for the shift towards ecological domin-ance is not well understood. Foucaud et al. (2009) suggested that human disturbance, rather than unicolonality, is likely responsible for triggering ecological dominance. This hypothesis is supported when the history of invasive populations and the impact of W. auropunctata on the local fauna are considered. The alteration of natural ecosystems into agricultural and urban environments results in shifts of biotic and abiotic pressures. Reduced biodiversity in al-tered ecosystems often creates open niches ready for exploitation. Open niches and the inability of local species to successfully defend against W. auropunctata likely allow for the exhibition of dominant behaviors early on in the invasion process and before the forma-tion of large supercolonies. Over time, an invading supercolony is able to expand its boundaries because W. auropunctata is able to dominate, outcompete, and displace other species. The degree of im-pact on local arthropod fauna is not universal since some species are able to successfully defend against W. auropunctata (Le Breton et al. 2007a, b; Mbenoun Masse et al. 2019b, 2021; Perfecto and Vandermeer 2020b). Could the presence of highly competitive local arthropod fauna also play a role in regulating ecological dominance of native and exotic W. auropunctata populations? It is interesting that the global distribution of other well known invasive ant spe-cies overlaps that of W. aurpounctata yet their competitive strength against W. auropunctata seems to vary from one location to another (Kirschenbaum and Grace 2007a, Mbenoun Masse et al. 2019b). Downloaded from https://academic.oup.com/aesa/advance-article/doi/10.1093/aesa/saac016/6697138 by guest on 16 September 20227Annals of the Entomological Society of America, 2022, Vol. XX, No. XXFuture research on the driving factors of ecological dominance, and subsequent ecological impacts should consider the roles different types of human disturbance and multispecies interrelationships have on the degree of ecological dominance (Perfecto and Vandermeer 2020b). Future research should also investigate possible links be-tween W. auropunctata genetics and ecological dominance. It is pos-sible that certain haplogroups or haplotypes are more likely to show greater levels of ecological dominance than others.Social BehaviorsThe complex intra- and inter-specific behaviors of W. auropunctata provide insight into how this species quickly establishes and eventu-ally dominates in some areas while accounting for a mere fraction of the biodiversity elsewhere. Variation in intra- and interspecific behavior between the dominant and nondominant variegates high-lights a behavioral plasticity that is not well understood (Le Breton et al. 2004, 2007a). Since dominant and nondominant variegates exist throughout the native range, it is important to consider behav-ioral studies in the context of this variation rather than simply as native and nonnative populations. However, most studies focusing on intraspecific aggression were conducted before Foucaud (2009) describing the dominant and nondominant variegates and so intra-specific aggression has typically been reported in the context of na-tive versus nonnative populations rather than according to variegate (Le Breton et al. 2004, Errard et al. 2005). The consensus among these studies is that invasive populations display much lower intra-specific aggression than native populations but it is likely that the na-tive populations studied were of the nondominant variegate since the study sites in question were typically within primary forests rather than disturbed habitats. To date, no studies have examined intra-specific aggression in the context of native and nonnative dominant versus nondominant populations.Variation in intraspecific aggression and nest-mate recognition is linked to the chemical composition of cuticular hydrocarbons (CHCs) (Errard et al. 2005, Martin and Drijfhout 2009, Vonshak et al. 2009). Errard et al. (2005) reported a correlation between intraspecific aggression and variation of CHC’s among native and nonnative W. auropunctata populations. Likewise, there is high gen-etic variability among the nondominant variegate and low genetic variability among the dominant variegate due to their different re-productive modes (Foucaud et al. 2007). This supports the hypoth-esis that the native population studied by Errard et al. was of the nondominant variate rather than the dominant variegate; however, research is needed to confirm this. Moreover, CHC production and variability are not only governed by genetic factors but are also in-fluenced by environmental factors such as diet, habitat, and season (Vonshak et al. 2009). Studies have shown that CHC production and intraspecific aggression shift when W. auropunctata are taken from their natural environment and placed into laboratory culture (Vonshak et al. 2009).Wasmannia auropunctata has a reputation for being highly ag-onistic toward other species, with direct aggression being primarily responsible for its success (de la Vega 1994; Kirschenbaum and Grace 2007a, b, 2008; Vonshak et al. 2012). However, although interspecific aggression in W. auropunctata is well documented, many factors contribute to its success as an invader. Wasmannia auropunctata population densities appear to be an underlying factor influencing interspecific aggression. At low densities or when not numerically dominant, W. auropunctata workers are pliant in the presence of other, more dominant species, and act as an insinuator species (Achury et al. 2008, Vonshak et al. 2012, Yitbarek et al. 2017). Its small size may allow W. auropunctata to select what type of interaction is most appropriate and beneficial for a given situation (Tennant 1994, Le Breton et al. 2007a, Achury et al. 2008, Vonshak et al. 2012, Yitbarek et al. 2017). Low population densities occur within populations of the nondominate variegate and early in the invasion process, after colony establishment for the dominant varie-gate. For new introductions, low interspecific aggression may enable W. auropunctata to coexist with other species and exploit resources necessary for colony growth (Vonshak et al. 2012, Yitbarek et al. 2017). Once numerically dominant, a behavioral shift occurs and W. auropunctata workers become highly agonistic toward other species, excluding them from resources and destroying their nests (Vonshak et al. 2012). The low worker:queen ratio, high fecundity rates, and low intraspecific aggression typical of the dominant variegate allow for quick colony growth and numerical dominance (Clark et al. 1982, Ulloa-Chacon and Cherix 1990, Brandao and Paiva 1994, de la Vega 1994, Way and Bolton 1997, Delsinne 2001, Kirschenbaum and Grace 2008).The ability of competitor species to fend off W. auropunctata attacks and defend resources also plays a role in regulating W. auropunctata population dynamics (Le Breton et al. 2007a, Perfecto and Vandermeer 2020a). Controlled behavioral experiments have shown that the presence of W. auropunctata elicits reactive re-sponses from some competitor ant species (e.g., immediate re-cruitment of larger castes able to attack and kill W. auropunctata) but not from other species (Kirschenbaum and Grace 2007a, Le Breton et al. 2007a, Kirschenbaum and Grace 2008, Perfecto and Vandermeer 2020b). It’s not yet known if successful defense against W. auropunctata within its native range is dependent on which eco-logical variegate is encountered and habitat type (i.e., disturbed or natural). Only the most aggressive competitor ant species seem able to fend off W. auropunctata in its native range and these are typic-ally other well-known invasive ant species (Le Breton et al. 2007a, b; Mbenoun Masse et al. 2019b, 2021; Perfecto and Vandermeer 2020b). Interestingly, Pheidole megacephala Fabricus has been docu-mented as a displaced species throughout much of W. auropunctata’s introduced range while also being implicated as a potential cause of a rare invasion contraction event in Cameroon (Mbenoun Masse et al. 2019b, 2021). Additionally, W. auropunctata have been docu-mented nesting near and tolerating other species when not com-peting for food resources (Way and Bolton 1997). This highlights the complexity of interspecific competition and that it is often oversim-plified. It is likely that successful competition and resilience against W. auropunctata rely on both biotic and abiotic factors (Jourdan et al. 2006, Vandermeer and Perfecto 2020).DietWasmannia auropunctata is a true generalist, feeding on whatever is available including nectar (floral and extrafloral) (Schemske 1980, Horvitz 1990, Deyrup 2000, Apple 2001), plant parts (Clemente and Whitehead 2020), other invertebrates (Smith 1942, Feinsinger and Swarm 1978, Clark et al. 1982, Way and Bolton 1997), animal feces (Rosumek 2017), and honeydew-producing phytophagous in-sects (Spencer 1941, Smith 1942, Fabres and Brown 1978, Delabie and Cazorla 1991, Delabie et al. 1994, Naumann 1994, de Souza et al. 1998, Fasi et al. 2013). Few nutrient allocation or dietary studies have been conducted on W. auropunctata, but studies on other ant species indicate that dietary preferences and needs may vary season-ally (Stein et al. 1990) and between arboreal and ground-dwelling conspecific ants (Hahn and Wheeler 2002, Bluthgen et al. 2003). Additionally, W. auropunctata raised under laboratory conditions can display differences in food-lure preferences when compared to wild conspecifics (Montgomery et al. 2020).Downloaded from https://academic.oup.com/aesa/advance-article/doi/10.1093/aesa/saac016/6697138 by guest on 16 September 20228Annals of the Entomological Society of America, 2022, Vol. XX, No. XXIt is common for laboratory-raised insects to behave differently to wild conspecifics and this likely influences observations during laboratory experiments (Herard et al. 1988, Propkopy et al. 1989, Ennis et al. 2015). Since laboratory experiments are valuable com-ponents of the research and experimentation process, differences in foraging behaviors should be accounted for when conducting dietary, palatability, and bait efficacy assessments (Montgomery et al. 2020).Global DistributionThe global distribution and spread of W. auropunctata were first outlined in 2003 and has been tracked since (Wetterer and Porter 2003, Wetterer 2013). Currently, W. auropunctata is more wide-spread globally than S. invicta (Buren), although the latter receives considerably more attention as a research and legislative priority. Wasmannia auropunctata is native to the Neotropics of Central and South America (Wheeler 1929). Its southerly range extends to cen-tral Argentina (Chifflet et al. 2016), east of the Andes, and it was recently postulated that W. auropunctata may be native as far north as the United States—Mexico border (Mikheyev and Mueller 2007, Wetterer 2013).Genetic analysis on specimens collected throughout this region has identified two phylogenetic clades (Clad A and Clad B) with nu-merous haplogroups based on cytochrome c oxidase 1 partial mito-chondrial genome analysis (Chifflet et al. 2016). Even though this research has provided empirical evidence for determining places W. auropunctata is likely native, there is still uncertainty as to how far the native range extends and where this species has been introduced. For example, W. auropunctata is native to northern Argentina, as evidenced by the presence of genetically diverse sexual and clonal populations (Chifflet et al. 2016, 2018). However, the occurrence of only clonal populations of and little genetic variation in central Argentina has been suggested as evidence of range expansion within the past 60 yr and may be indicative of anthropogenic introduc-tions (Chifflet et al. 2016, 2018). Currently, there is a noticeable lack of records from central and western Mexico and no genetic ana-lysis has been done on any W. auropunctata collected in Mexico. To tease out the northern limitations of W. auropuntata’s native range, future research should focus on documenting the ant biodiversity of this area and genetic analysis of W. auropunctata populations throughout Mexico should be conducted. Despite the uncertainty surrounding the northern limitations of its native range and the need for additional work, the Mexican populations are presumed part of W. auropunctata’s native range in this review due to its contiguous distribution where it is known to occur throughout Central America. We acknowledge that this matter requires further investigation be-fore it is fully reconciled.Historically, there has been uncertainty surrounding the status of W. auropunctata in the Caribbean due to its pervasiveness throughout the region with records dating back to the mid-1800s (Wetterer and Porter 2003, Mikheyev and Mueller 2007, Wetterer 2013). However, molecular genetics and analysis of eco-evolutionary pathways iden-tified the presence of the same clonal lineages being present on mul-tiple islands and in parts of South America (Mikheyev and Mueller 2007, Foucaud et al. 2010b). While this does not preclude the idea of the Caribbean being part of W. auropunctata’s native range, it does indicate that numerous introduction events have occurred throughout the region (Mikheyev and Mueller 2007, Foucaud et al. 2010b). The lack of evidence of sexually reproducing populations in the Caribbean is another indication that W. auropunctata is not likely native to the region (Mikheyev and Mueller 2007, Foucaud et al. 2010b). It is possible that the Caribbean distribution contains a mix of native and introduced populations (Wetterer 2013) but, no genetic evidence has been presented that suggests this is the case.The earliest confirmed record of W. auropunctata outside of its presumed native range was from Gabon in 1894 by Emery who rec-ognized it as having the potential to become a serious invasive tramp ant (Wetterer 2013). Indeed, since then, W. auropunctata has been introduced and established in 23 countries and island groups beyond its presumed native range (Wetterer and Porter 2003, Wetterer 2013, Espadaler et al. 2018, Mayron 2019, Vanderwoude et al. 2021). To date, the nondominant variegate has not been documented outside of W. auropunctata’s native range. Since the latest global distri-bution list (Wetterer 2013) there have been five new detections of established W. auropunctata populations in the world. Three new detections were made in Oceania: Wanyaan, Yap, Federated States of Micronesia in 2017 (GBIF.org 2021); Tutuila, American Samoa in 2018 (Gruber et al. 2016); and Suva, Fiji in 2019 (Vanderwoude et al. 2021). A detection in Malaga, Spain in 2018 was confirmed to be the northernmost outdoor population recorded to date (Espadaler et al. 2018). The populations in Israel and Spain represent the only outdoor infestations in Palearctic and Nearctic biogeographical re-gions. All other infestations in these biogeographical regions were documented in greenhouses and indoor nursery settings (Wetterer and Porter 2003). The most recent detection was confirmed in 2022 from Shantou, Guangdong Province in south-eastern China (Chen et al. 2022). This is the first official record of W. auropunctata in the Indomalayan biogeographical region.The global distribution presented in this review is conserva-tive (Fig. 5), with only confirmed, established populations of W. auropunctata as of June of 2021 being included. Not included on the map are W. auropunctata occurrences of unconfirmed detections of possible wild populations, which are included on other distribution lists, specifically, detections from California (USA), southern Texas (USA), Lisca Bianca Island (Italy) (Jucker et al. 2008, Wetterer 2013), and Dhaka, Bangladesh (GBIF.org 2022). Unconfirmed detections are either unsubstantiated reports of establishment or instances in which a single specimen was collected or photographed in the wild and presumably positively identified yet no verification, further record, or knowledge exists. It is possible that established W. auropunctata populations exist in these locales yet no follow-up sampling was ever conducted to verify their existence. Follow-up surveys and sampling should be a priority for all unconfirmed records of W. auropunctata as this information will be invaluable for the understanding of this species presumed native range and potential global distribution. Notably, W. auropunctata is regularly intercepted in California by Department of Agriculture quarantine inspectors, but no wild popu-lations have been confirmed in (G. Arakelian, Los Angeles County Entomologist, personal communication) despitean early claim that it was established in Los Angeles County (Keifer 1937) and repeated citing of this claim in numerous distribution lists. Regulatory and responding agencies around the world should be hyper-vigilant, es-pecially in areas where W. aurounctata is frequently intercepted but not believed to be established. It is important to acknowledge that W. auropunctata are likely established beyond the locations indicated here and have yet to be officially detected due to the ease at which they are distributed and the probability of going unnoticed.Distribution PathwaysTransportation and dispersion of W. auropunctata typically occur in three ways: natural active dispersion, natural passive dispersion, and human-mediated transport. Natural active dispersal occurs primarily via budding (Ulloa-Chacon and Cherix 1990). When a nest becomes Downloaded from https://academic.oup.com/aesa/advance-article/doi/10.1093/aesa/saac016/6697138 by guest on 16 September 20229Annals of the Entomological Society of America, 2022, Vol. XX, No. XXcrowded or the nest is disturbed, a queen will carry a small number of workers to a nearby location and establish a new nest aggregation (Feitosa 2007, Mbenoun Masse et al. 2011). Because of this, outward expansion of an infested area is slow, measured in the tens of meters per year, with dispersal propelled by the ant’s direct effort (Walsh et al. 2004). However, the rate of outward expansion may also be influ-enced by population density and ecological factors (Mikheyev et al. 2008). Wasmannia auropunctata’s unicolonial colony structure, low intra-specific aggression, and generalist nesting preferences allow for population densities far beyond that of multicolonial ant species and species with specialized nesting preferences. Therefore, the rate of outward expansion may increase or decrease depending on site fea-tures such as nesting site availability and physical barriers between the current infested area and the closest suitable habitat. Invasion contraction events (i.e., reduction of previously invaded area) have been documented, but appear to be rare and warrant further inves-tigation (Lester and Gruber 2016, Mbenoun Masse et al. 2019b).Modes of natural passive dispersion occur by rafting downstream on waterways and flood water, landslides, and possibly from high winds and storm events (Lubin 1984, Walker 2006, Vanderwoude et al. 2014). Modes of natural passive dispersion enable dispersion along greater distances than through budding and act as pathways for new introductions with transport via moving water being the most frequently observed (Walker 2006, Vanderwoude et al. 2014).Human-mediated transport is the most common mode of long-distance dispersal (Walsh et al. 2004, Mikheyev et al. 2008, Foucaud et al. 2010b) and both intentional and unintentional intro-ductions have occurred world-wide. Wasmannia auropunctata is frequently transported locally, intra- and inter-nationally through human commerce including, but not limited to, the movement of infested nursery stock and planting media, construction materials, vehicle and machinery, stockyard supplies such as wooden pallets, and by the movement of other types of items held at infested sites to uninfested sites (e.g., furniture, portable toilets, salvaged materials, etc.) (Walsh et al. 2004). Phylogenetic analysis allows tracing of the place of origin for invasive populations and their spread (Mikheyev and Mueller 2007, Silva et al. 2018). Distinct genetic similarities have been found between invasive populations of W. auropunctata in countries with strong trade ties or sharing established shipping routes (Foucaud et al. 2010b).Intentional human introduction and movement of W. auropunctata has occurred for biocontrol purposes (Bruneau de Miré 1969, Wetterer et al. 1999, Ndoutoume-Ndong and Mikissa 2007, Fasi 2009). Although it is well known that the costs associated with W. auropunctata invasion outweigh any potential benefit, it’s pos-sible such intentional introductions will continue due to this species being highlighted repeatedly as a natural enemy and potential bio-control agent for coffee berry borer (Curculionidae: Hypothenemus hampei Ferrari) and Asian citrus psyllid (Liviidae: Diaphorina citri Kuwayama) (Morris and Perfecto 2016, Kondo et al. 2018, Morris et al. 2018, Jiménez-Carmona et al. 2019, Perfecto and Vandermeer 2020a) (Morris and Perfecto 2022).Predicting Future Range ExpansionWasmannia auropunctata has traditionally been considered a trop-ical and subtropical pest species. The outdoor infestations in Spain and Israel show this is a misleading conception and more atten-tion should be given to the species’ current potential range and fu-ture range expansion under climate change. ‘Clade A’ and ‘Clade B’ each display different range potential (Chifflet et al. 2016) with the former distributed primarily in the tropics and the latter having a much wider range into subtropical and Mediterranean climates (Chifflet et al. 2016).Species distribution models (SDMs) are often used to predict the potential range expansion of a target species. However, SDMs are Fig. 5. Global distribution of W. auropunctata as of 2022. Gray shaded areas and points indicate the presumed native range. Black shaded areas and points indicate locations where exotic outdoor populations are established. Black x’s indicate locations where W. auropunctata have been introduced and are documented as indoor greenhouse pests but no outdoor populations have been documented. The current distribution map includes data from J.K. Wetterer’s 2013 distribution map and all subsequent records of confirmed established W. auropunctata populations detected since 2013.Downloaded from https://academic.oup.com/aesa/advance-article/doi/10.1093/aesa/saac016/6697138 by guest on 16 September 202210Annals of the Entomological Society of America, 2022, Vol. XX, No. XXfar from perfect and ecologists are still trying to determine which models and variables best describe limiting factors, potential range predictions, and eco-evolutionary scenarios (Elith et al. 2010, Rey et al. 2012, Federman et al. 2013, Coulin et al. 2019). To date, only two studies have modeled the potential distribution of W. auropunctata (Federman et al. 2013, Coulin et al. 2019). Temperature and pre-cipitation were identified early on as limiting factors for its habitat suitability (Jourdan and Dumas 2004). Since then, minimum and maximum critical thermotolerances of the species have been iden-tified through laboratory experiments and included in SDMs (Rey et al. 2012, Foucaud et al. 2013, Coulin et al. 2019). Some have speculated elevation as a limiting factor due to W. auropunctata not occurring above 700 m in New Caledonia (Jourdan et al. 2006), but infestations have been detected in Hawai`i at over 1,200 m (M. Montgomery personal observation). This suggests that elevation is not likely a limiting factor in and of itself, but as it relates to tempera-ture and moisture thresholds. Additionally, although SDMs using cli-matic factors are undoubtedly useful, the spatial scale at which these models analyze climactic data may not identify localized microcli-mates and the models do not account for human disturbances that influence local microclimates which may aid in triggering natural adaptations within the species (Federman et al. 2013, Foucaud et al. 2013).Because W. auropunctata is considered a disturbance specialist, human disturbances such as irrigation can transform naturally unsuitable habitats into suitable habitats that are unlikely to be detected by correlative SDMs when using naturally occurring pre-cipitation data (Vonshak 2010, Federman et al. 2013). Federman et al. (2013) demonstrated that using a precipitation correction accounting for irrigation estimates increased the precision and ac-curacy of model predictions. Models based solely on data collected from the target species native range are also likely to fail to iden-tify all suitable habitat and global distribution potential of highly adaptable species (Elith et al. 2010). Contrary to correlative SDMs that use climatic and known distribution data, mechanistic SDMs use functional attributes, niches, and spatial data to predict poten-tial range expansion. Models combining mechanistic and correlative SDMs that account for human disturbances and incorporate data from the species’ known global distribution should be investigated further for their potential to refine predictions of habitat suitability and potential global range expansion.ImpactsImpacts and benefits associated with W. auropunctata are multitiered and may be characterized as direct or indirect. Direct impacts are those caused by W. auropunctata while indirect im-pacts are those caused by other organisms influenced by the pres-ence of W. auropunctata. For example, the extirpation of a species due to predation or resource competition by W. auropunctata is a direct impact whereas an increase in plant disease due to higher populations of scale insects farmed by W. auropunctata would be considered an indirect impact. Humans, animals, and entire ecosystems are known to be affected by W. auropunctata. Numerous ecological studies and economic impact assessments have been conducted documenting the multitiered impacts of W. auropunctata and have determined that while some impacts may be considered positive (i. e. biological control of coffee berry borer), the vast majority are negative (Bousseyroux et al. 2019). Careful and thorough cost-benefit considerations should be made before the intentional movement or maintenance of W. auropunctata infestations.Mechanisms for Displacement and Impacts on Natural EcosystemsDisplacement of ants and other invertebrates by W. auropunctata is well documented (Clark et al. 1982; Lubin 1984; Jourdan 1997a,b Armbrecht and Ulloa-Chacon 2003; Le Breton et al. 2003; Wetterer and Porter 2003; Walker 2006; Grangier et al. 2007; Ndoutoume-Ndong and Mikissa 2007; Kirschenbaum and Grace 2008; Fasi 2009; Vonshak et al. 2010; Gasc et al. 2018; Bousseyroux et al. 2019; Mbenoun Masse et al. 2019b). In fact, Silberglied (1972 p. 13) commented that its impact in the Galapagos was ‘the most ser-ious of any introduced animal’. Ecological studies on insect bio-diversity between areas and plots with and without W. auropunctata (Clark et al. 1982; Lubin 1984; Jourdan 1997b; Roque-Albelo et al. 2000; Le Breton et al. 2003, 2005; Walker 2006; Grangier et al. 2007; Ndoutoume-Ndong and Mikissa 2007; Vonshak et al. 2010; Mbenoun Masse et al. 2017, 2019b; Gasc et al. 2018) along with data on resource interference and competition and interspecific ag-gression (Grangier et al. 2007, Vonshak et al. 2012, Yitbarek et al. 2017) provide evidence linking this species to the extirpation of other insects within invaded areas). Such studies have built a founda-tion upon which insect populations and diversity are used as proxies for the impacts of W. auropunctata in forests. Additionally, the cor-relation between biodiversity and W. auropunctata presence or ab-sence has led to W. auropunctata being used as an indicator species for low insect community diversity (Armbrecht and Ulloa-Chacon 2003, Achury et al. 2008, Berman et al. 2013).Studies describing W. auropunctata impacts on insect communi-ties throughout its native and nonnative ranges appear to make the important assumption that its inherent ability to dominate ecosys-tems remains constant and that differences in impact magnitude are due to the resilience of competing species. No consideration is given to whether the populations being studied are of the dominant or nondominant variegate described by Foucaud et al (2009). Since the two variegates differ biologically, physiologically, and behaviorally, results from studies that measure impacts between invasive popu-lations and native populations without consideration of ecological variegate may be inherently flawed (Brandao and Silva 2008, Achury et al. 2012, Rojas and Fragoso 2021). Any future study comparing impacts of native versus nonnative populations should take care to ensure that populations being compared are of the same variegate.Resource and interference competition have been indicated as the mechanisms by which W. auropunctata succeed as invaders and displace other insects. Multiple behaviors have been identified contributing to W. auropunctata’s competitive strength, including acting as an insinuator species when not numerically dominant (Le Breton et al. 2007a, Yitbarek et al. 2017). Wasmannia auropunctata workers frequently take significantly longer to discover and recruit to food resources compared with other ant species (Vonshak et al. 2012, Yitbarek et al. 2017). When first encountering food resources occupied by another species, their passivity toward the other spe-cies likely allows W. auropunctata to avoid aggressive interactions until enough nestmates can be recruited to successfully defend the resource (Vonshak et al. 2012, Yitbarek et al. 2017). Additionally, native ants often appear incapable of forming appropriate responses are unable to defend nesting and food resources creating niche op-portunities that W. auropunctata effectively exploit (Le Breton et al. 2005, 2007a).While the role of resource competition is widely suggested as a mechanism for competitive success, there is some debate as to the magnitude of its contribution. Observations during behavioral la-boratory assays suggest that while W. auropunctata are aggressive toward some species, they are often unsuccessful defenders of food Downloaded from https://academic.oup.com/aesa/advance-article/doi/10.1093/aesa/saac016/6697138 by guest on 16 September 202211Annals of the Entomological Society of America, 2022, Vol. XX, No. XXresources during general interspecific encounters (Kirschenbaum and Grace 2008, Vonshak et al. 2012). However, during one study, al-though W. auropunctata retreated from food resources when con-fronted by a competitor, they also invaded and completely destroyed the competitor’s nests over the course of several days and weeks (Vonshak et al. 2012). Because they are generalist feeders and ac-tive 24 h per day, it’s possible that the cost of aggressive encounters outweighs the benefit of immediate reward and it is better to wait or find an unoccupied resource. Also, laboratory studies don’t often re-flect what happens under natural circumstances. For example: in the wild, the dominant variegate is defined, in part, by its propensity to form expansive supercolonies. This allows for massive recruitment of resources. In contrast, laboratory colonies and experiments are highly structured and controlled and use only a single nest at a time. Field studies have suggested that W. auropunctata may share food resources when fewer than 500 foragers are present (Achury et al. 2008). If this is the case, it is unlikely that a single nest would pro-vide high enough forager recruitment to a resource to successfully defend it during laboratory experiments and behavioral assays. The dichotomy between laboratory and wild conspecifics and the envir-onmental conditions they are exposed to is an important consider-ation with interpreting results from such studies.Although physical aggression, such as biting and stinging, appears to be the primary competitive mechanism by which W. auropunctata directly interferes with other species, chemical defenses may also play a role (Howard et al. 1982, Le Breton 2002, Showalter et al. 2010). Ants primarily communicate chemically and the use of phero-mones is an important mode of communication for foraging, de-fense, and regulation of inner-nest functioning (Howard et al. 1982, Martin and Drijfhout 2009, Showalter et al. 2010, Cha et al. 2019). When distressed, W. auropunctata secrete an alarm pheromone from the mandibular gland which attracts nestmates and aids in mass re-cruitment for defensive action (Howard et al. 1982, Showalter et al. 2010). These mandibular gland secretions may also act as a repellant to competitor species (Howard et al. 1982), although this hypoth-esis has been challenged (Le Breton 2002). It is important to rec-ognize that the two studies investigating repellant properties of the mandibular gland secretions employed very different methodologies and confounding factors may have influenced the results of one or both of the studies. One study tested the effects of extracts from the mandibular gland on the acceptability of mealworms to other ant species (Howard et al. 1982) whereas the other study exposed food lures to W. auropunctata in the field and later presented the exposed food lures to other species at different field locations without W. auropunctata (Le Breton 2002). Given their behavioral plasticity, it is possible that the mandibular gland excretions are secreted only during interspecific encounters, and thus may not have been present on the food lures during foraging when only W. auropunctata was present. Although the mandibular gland excretions are verified alarm pheromones used in mass recruitment and undoubtedly contribute to successful resource defense (Showalter et al. 2010), more research is needed to verify any potential repellant action of W. auropunctata mandibular gland secretions.Wasmannia auropunctata has also been identified as a threat to vertebrates (Jourdan 2001, Walsh et al. 2004, Beavan et al. 2008). Tropical keratopathy (TK) is a condition in which the cornea of the eye becomes clouded over and appears superficially similar to cata-racts. Mild cases of TK often appear as small, individual cloudy cor-neal spots whereas extreme cases appear as clouding over the entire cornea. Wasmannia auropunctata has been directly linked to occur-rences of widespread TK in domestic animals and with occurrences of TK in humans (Theron 2007, Rosselli and Wetterer 2017, Patael et al. 2019). Although no studies have focused on documenting animal TK occurrences in invaded natural ecosystems, trail cam-eras in Gabon captured images of a leopard with severe TK (Walsh et al. 2004) and it is likely this is a problem not yet investigated. Agonistic interactions between W. auropunctata and several lizard (Anolis) species have been documented and it has been speculated that lower herpetofauna diversity may be found throughout infested areas (Jourdan 2001, Jourdan et al. 2001, Wetterer et al. 2007, Perfecto and Vandermeer 2020a). The effects of W. auropunctata on ground-nesting seabirds and forest birds has not been formally as-sessed, but other invasive ant species are known to reduce nesting and hatching success, foraging, and survival rates (Allen et al. 2004, Plentovich et al. 2008, Davis et al. 2009, Kropidlowski 2014). The diversity of suitable habitat for W. auropunctata likely increases the chances of ant-bird interactions, and the potential for negative impacts on avifaunal communities. Research is needed to identify possible impacts and the risk W. auropunctata poses to various avi-faunal communities.In some areas, W. auropunctata may directly and indirectly impact forest plant composition and regeneration through the as-sistance of or interference with seed production, dispersal, and ant-plant mutualisms (Horvitz 1990, Mikissa et al. 2013, Clemente and Whitehead 2020). Wasmannia auropucunctata have been observed harvesting fruits of Piper sanctifelicis Trel (Family: Piperaceae), a common South American shrub (Clemente and Whitehead 2020). Although it is not unusual for ants to harvest seeds of plants typ-ically dispersed by vertebrates, it is the first and only record of W. auropunctata doing this. Additionally, W. auropunctata may aid in fruit set and seed production of some plants through predation on herbivorous insects (Horvitz and Schemske 1984). On the con-trary, W. auropunctata have been observed interfering with ant-plant mutualisms and indirectly hindering forest regeneration by excluding ant mutualists from their obligate host plants (Mikissa et al. 2013). Other ant species, such as yellow crazy ant (Anoplolepis gracilipes [Smith]) have been implicated in ‘invasional meltdowns’ due to interference with natural ecosystem processes (O’Dowd et al. 2003, Green et al. 2011). For example, on Christmas Island in the Indian Ocean, A. gracilipessignificantly reduced populations of en-demic red land crabs (Gecarcoidea natalis Pocock), a keystone spe-cies, and is linked to population explosions of scale insect and sooty mold (O’Dowd et al. 2003, Green et al. 2011). These relationships have altered local litter decomposition rates, plant and insect species diversity, and forest structure, and led to secondary invasions within A. gracilipes invaded areas (O’Dowd et al. 2003, Green et al. 2011). Although W. auropunctata alters ecosystem processes, very little is understood on the ramifications of these relationships on floral com-munities and whether ecological variegation is a factor influencing the magnitude and direction of the impacts. It is possible that such relationships may also lead to ‘invasional meltdowns’, especially within insular island communities, but additional research is needed to assess this.Agricultural SystemsAlthough ecological impacts of W. auropunctata are serious, they are often not enough to trigger concern and human intervention. Agriculture industries are the most affected by W. auropunctata ac-cording to past economic impact assessments (EIAs) (Lee et al. 2015, Angulo et al. 2021) and their multitiered impacts are perhaps most clearly documented within agriculture systems. Agricultural workers are frequently stung while maintaining and harvesting crops. Reports from Tahiti and Florida indicate that farm workers have been unable or unwilling to harvest crops and to perform other duties in heavily Downloaded from https://academic.oup.com/aesa/advance-article/doi/10.1093/aesa/saac016/6697138 by guest on 16 September 202212Annals of the Entomological Society of America, 2022, Vol. XX, No. XXinfested citrus orchards and coffee fields (Spencer 1941, Smith 1965, Fabres and Brown 1978, Delabie 1988, Perfecto and Vandermeer 2020b) resulting, in some cases, in the abandonment of family lands (Vanderwoude et al. 2015). A similar trend has been observed in Hawai`i, where farm workers routinely fail to come to work on days when infested fields are scheduled to be picked (M. Montgomery personal observation).The mutualistic relationship between W. auropunctata and honeydew-producing phytophagous insects results in indirect im-pacts such as population explosions of economically important plant pests (Smith 1965, Fabres and Brown 1978, Delabie 1989, Fasi et al. 2013). In particular, population levels of Homopterans including aphids, scale insects, and mealybugs are positively correlated with W. auropunctata population levels and negatively correlated with plant health and crop productivity (Delabie and Cazorla 1991, de Souza et al. 1998, Fasi et al. 2013). These mutualisms are a conse-quence of the protection that W. auropunctata provides the insects against natural enemies and the ants feeding on the sugary exudates (honeydew) of these insects. Many of these plant pests are known vectors of diseases that reduce plant health leading to crop loss and excess honeydew also acts as a medium for sooty mold growth.Despite the association with economically important plant pests and their direct impact on farm workers, W. auropunctata has been intentionally introduced and moved around as a biocontrol agent against significant crop pests in the Solomon Isands and Gabon (Wetterer et al. 1999, Fasi et al. 2013). Recently, research also con-firmed W. auropunctata is potentially effective at reducing popula-tions of coffee berry borer and Asian citrus psyllid (Kondo et al. 2018, Morris et al. 2018, Jiménez-Carmona et al. 2019, Perfecto and Vandermeer 2020a, Morris and Perfecto 2022). However, re-searchers of those studies noted that the benefits do not outweigh the costs. Conversely, W. auropunctata has been documented interfering with various biocontrol agents (Kondo et al. 2018, Kulikowski 2020, Perfecto and Vandermeer 2020a, Perfecto et al. 2021). Whether in-tentionally or unintentionally introduced, farmers face a variety of issues due to W. auropunctata infestations.Residential and Urban AreasWasmannia auropunctata infestations in urban environments have garnered little attention despite its recognition as a house pest (Smith 1929, Fernald 1947, Delabie 1995, Espadaler et al. 2018, Mbenoun Masse et al. 2019a, Kidon et al. 2022, Mbenoun Masse et al. 2021). The Hawai`i infestation reported by Conan and Hirayama in 2000 represents the first record of W. auropunctata as a major residen-tial pest in any western society (Conant and Hirayama 2000). Their small size, brownish-orange color, and cryptic nesting habits allow this species to go unnoticed for many years and it is common for Hawaiian residents to become impacted by W. auropunctata only after the infestation has spread throughout the entire property and the ants have begun encroaching into homes and other struc-tures (M. Montgomery personal observation). This is not unique to Hawai`i. The recent detection in Malaga, Spain, was due to a resi-dent reporting an infestation of stinging ants in their home to a pest control company (Espadaler et al. 2018). Follow-up surveys by au-thorities revealed a 5.8 ha infestation spanning 50 private residences with some ants nesting in electrical receptacles (Espadaler et al. 2018). It was estimated that W. auropunctata was introduced over five years before detection but residents reported first noticing them only about two to three years prior (Espadaler et al. 2018). Since the original detection in 2018, two other infestations within the Malaga province have been detected at 13 and 17 km away from the original infestation (Espadaler et al. 2020). At this time, these new detections have not yet been completely delimited and their size is unknown (Espadaler et al. 2020).Information gathered through public outreach communica-tions in Hawai`i (M. Montgomery, unpublished data) has revealed that residents are frequently stung in their beds, while using the bathroom, or sitting on certain types of furniture (e.g., sofas, re-cliners). Despite this, individual sentiments toward W. auropunctata vary wildly. Some residents appear determined to manage W. auropunctata infestations and reduce the impacts themselves while others appear despondent and assume their situation is without a solution (M. Montgomery personal observation). For example, W. auropunctata was introduced to the Solomon Islands in the 1960s or early 1970s and continues to spread throughout the archipelago (Fasi et al. 2016). Public attitudes toward W. auropunctata ap-pear to shift with communities having dealt with infestations for longer periods being more tolerant of the ant than communities with more recent invasions (Fasi et al. 2016). Once apathy toward W. auropunctata begins to take hold in a community, management practices likely decline. Similar attitudes are common in Hawai`i. In many Hawaiian communities, some residents are only persuaded to take management action against W. auropunctata when social pres-sure increases and they fear being stigmatized in their community (Niemiec et al. 2018, 2019).Economic ImpactsThere is no doubt that W. auropunctata negatively impacts eco-system functions and quality of life but the reports are generally qualitative, difficult to quantify into monetary units, and rarely used to guide legislation aimed at biosecurity and prevention. Economic impact assessments (EIAs) and cost-benefit analyses (CBAs) attempt to place monetary values on damages incurred across various eco-nomic sectors and costs associated with postarrival management and prevention efforts. These cost factors are used in bioeconomic models to project how economic costs associated with a target spe-cies compound change over time under different management strat-egies. Unfortunately, quantification and reporting of costs associated with W. auropunctata are rare. Because EIAs rely on reported costs that can be quantified monetarily and largely ignore qualitative metrics, EIAs tend to grossly underestimate impacts (Angulo et al. 2021), especially in subsistence economies.Three studies detailing the economic impacts of W. auropunctata have been compiled. Two focus on economic impacts in Hawai`i (Motoki et al. 2013, Lee et al. 2015). The other is a global as-sessment of the economic impact of invasive ants, including W. auropunctata (Angulo et al. 2021). In the global assessment, Wasmannia auropunctata was identified as one of the most econom-ically impactful species in the world, second only to S. invicta with cumulative global costs since 1930 estimated at US$19.91 billion, predominantly incurring over the past ten years (Angulo et al. 2021). The greatest damage and highest costs of management are in the agriculture and public welfare sectors (Lee et al. 2015, Angulo et al. 2021). In Hawai`i, W. auropunctata was projected to cost the state US$6.1 billion over 35 yr if the current management efforts are sus-tained or US$12.9 billion if the management status quo is not main-tained (Motoki et al. 2013, Lee et al. 2015). Conversely, economic costs can be greatly reduced with increased management, rapid re-sponse efforts to newly detected infestations, and enhanced preven-tion measures. Despite higher management cost than the status quo and reduced effort scenarios, overall costs (damages + direct costs of management and prevention efforts) of enhanced action plans were projected to reach only $US51 million over 35 yr (Motoki et al. 2013, Lee et al. 2015).Downloaded from https://academic.oup.com/aesa/advance-article/doi/10.1093/aesa/saac016/6697138 by guest on 16 September 202213Annals of the Entomological Society of America, 2022, Vol. XX, No. XXMost of these assessments focus on costs and damages without considering potential economic benefits of the target species. In Hawai`i, the invasion of W. auropunctata has spurred busi-ness development focusing on ant management in landscapes (M. Montgomery personal observation) and sales of ant baits generate revenues for local businesses. Additionally, the coffee berry borer is a major pest of coffee in Hawai`i with few effective management op-tions. The recent identification of W. auropunctata as a potentially effective biological control (Kondo et al. 2018, Morris et al. 2018, Jiménez-Carmona et al. 2019) may provide a valuable nonchemical pest control option for Hawaiian coffee farmers. The reduction in costs associated with coffee berry borer control could be con-sidered as a cost-saving economic benefit when viewed in isolation. However, W. auropunctata is also a major pest in coffee and creates costs associated with control efforts.Detection and ControlDetectionTraditionally, nontoxic food lures have been the primary tool used for detecting W. auropunctata. Foods high in lipids and/or protein, such as peanut butter, hotdogs, or tuna fish, are typically used as lures during ecological studies to determine W. auropunctata pres-ence with recruitment to the food lures frequently used as a surro-gate for population size or density. Food lures are also used to delimit infestations and monitor the effectiveness of control programs. Few studies have investigated the attractiveness of pure lipids, proteins, and carbohydrates to W. auropunctata (Williams and Whelan 1992, Montgomery et al. 2020), but the consensus is that lipid-rich foods are the most attractive food lures (Williams and Whelan 1992, Meier 1994, Armbrecht and Ulloa-Chacon 2003, Montgomery et al. 2020). This is important because the use of a suboptimal or less attractive, food lure is likely to underestimate distribution, density, and con-found results. Additionally, it is uncertain whether W. auropunctata would defend ideal food resources more than suboptimal resources which could influence the results of competition behavior studies.Given their generalist diet, it is possible that food-lure prefer-ence may vary from one site to another or temporally within the same location depending on the season, available resources, nu-trient deficiencies, and colony needs. For example, in Hawai`i, W. auropunctata recruitment rates to lipid-based food lures have been documented to be 6- —and 11-fold greater than carbohydrate and protein food lures, respectively among wild ants whereas labora-tory colonies showed a preference for carbohydrate food lures (Vanderwoude et al. 2015). It is unknown whether fluctuations in dietary needs of wild populations and available resources influence their attraction to different food lures. To minimize possible con-founding effects of food-lure selection, more information is needed elucidating food preference variability and the relationship between food preference and interspecific competition.Despite detection via food lures being the norm, accuracy of area-wide surveys using food-lure detection protocols varies greatly depending on spacing between lures, foraging distance, and operator experience. Additionally, detection of arboreal colonies is difficult with traditional lure-based surveys and require special consideration (Cox et al. 2020). Early detection and control projects often must seek a compromise between feasibility and accuracy depending on personnel, time, and funding restrictions. Such compromises often manifest as wider spacing between lures and fewer survey efforts, and have resulted in false negative results and premature declar-ations of eradication (Montgomery unpublished data).Food-lure detection methods are inexpensive and easy to use but these benefits come at a cost. Aside from variations in accuracy and precision, they are not species specific to W. auropunctata and ants collected during such surveys must be competently identified. This can be time consuming and further delay time-sensitive control ac-tions. Species-specific detection methods are being developed that hold promise over conventional food-lure detection methods. The use of detector dogs is common for agriculture quarantine inspec-tions and biosecurity programs around the world. Their use in in-vasive species monitoring programs has also grown in recent years (Lin et al. 2011, Wylie et al. 2016, Baker et al. 2017, Poland and Rassati 2019). Similarly, the development and marketing of lateral flow immunoassay rapid tests has revolutionized detection surveys targeting S. invicta by circumventing the need for time-consuming professional or laboratory diagnostics and allowing quick identifi-cation in the field (Valles et al. 2017, 2020). Should a comparable rapid test be developed for W. auropunctata, it would undoubtedly be an invaluable resource for biosecurity, rapid response and control efforts. The use of alarm pheromones for W. auropunctata manage-ment has shown to be ineffective, although they may hold potential as a species-specific detection tool in the near future and the devel-opment of pheromone lures is ongoing (Troyer et al. 2009, Derstine et al. 2012). Remote sensing has been used to identify S. invicta mounds by aerial searches (Vogt et al. 2008). Finally, advances in genetic research will likely lead to improved EDNA technology suit-able for testing terrestrial substrates and may be useful for detecting nascent ant colonies. Although the future holds much promise for improved detection, currently, all but the conventional food lure-based and sniffer dog detection methods are still theoretical or un-available for W. auropunctata.Chemical ControlAnt control is more nuanced than control of most other insect pests. Historically, persistent and nonpersistent general use of con-tact insecticides, mound treatments, insecticidal ant baits (here-after referred to as baits), and physical barriers have been used to control pest ants in various habitats (Osburn 1949, Delabie 1989, Williams 1994, Brooks and Nickerson 2008, Cabral et al. 2011). As the term implies, the general use of contact insecticides (here-after referred to as contact insecticides) kill a wide variety of in-sects on contact through direct spray or when the insect comes into contact with a treated surface. Horizontal transfer, the acquisition of insecticide through physical contact with a treated individual, of active ingredients has been used to control some species of ants in residential and conservation areas (Soeprono and Rust 2004, Choe and Rust 2008, Klotz et al. 2009, Buczkowski 2019, Buczkowski and Wossler 2019, Zhang et al. 2022, Cabral et al. 2011), but few ac-tive ingredients are effective through horizontal transfer. Horizontal transfer, also, has not been tested as a possible control method for W. auropunctata and should be investigated further. Other contact insecticides may be used as a soil drench for potted plants or indi-vidual nests. While use pattern may be useful to treat infested nur-sery stock and potted plants at home, it is impractical to attempt control of W. auropunctata through drenching individual nests given its nesting behavior. Most contact insecticides are ineffective at pro-ducing lasting results for area-wide management because their ef-fects are limited to killing foraging workers only while the rest of the colony within the nest remains unaffected. Baits hold significant advantages over other general use insecticide products in that they are typically lower in toxicity, have fewer nontarget impacts, and minimize insecticide use (Williams 1983, Klotz et al. 2003, Tollerup et al. 2004). Comprised of an attractant, carrier, and small amount Downloaded from https://academic.oup.com/aesa/advance-article/doi/10.1093/aesa/saac016/6697138 by guest on 16 September 202214Annals of the Entomological Society of America, 2022, Vol. XX, No. XXof active ingredient, baits use the ant’s natural foraging behavior to seek out, collect, and share the bait throughout the colony which affects reproductive, worker, and brood castes alike. To be effective, the active ingredient constituent in baits must be nonrepellant, lethal at very low doses and after dilution by trophallaxis, and also main-tain delayed mortality effects (Levy et al. 1973, Williams 1983, Rust et al. 2000, Braness 2002, Tollerup et al. 2004).Baits developed for ‘fire ants’ target lipid-responding ants and have been developed primarily for control of S. invicta. Although W. auropunctata is also a species that is attracted to lipids, there are substantial differences between the two species and the types of habitats they invade which render some baits more effective than others in various situations. For example, S. invicta build mounds in soil and prefer open, dry habitats whereas W. auropunctata nest opportunistically on the ground and in trees, preferring shady, moist habitats (Perfecto and Vandermeer 2020b). Chemical sensitivity dif-ferences between these two species is also apparent which influences efficacy of ‘fire ant’ baits against W. auropunctata (Hara et al. 2014, Montgomery et al. 2015). Site features, chemical sensitivities, and environmental degradation of baits and their active ingredients are the three main issues needing careful consideration when selecting a bait to control W. auropunctata. Management and eradication fail-ures against this species have, in part, been attributed to a failure in addressing one or more of these issues during the eradication planning and implementation processes (Bossin and Padovani 2010).Commercial baits used against W. auropunctata are available as granules, pastes and gels and may be used in bait stations, or broadcast treatments. Outdoor control using bait stations has been reported as inconsistent and less effective than broadcast bait ap-plications (Ulloa-Chacón and Cherix 1994, Souza et al. 2008). Wasmannia auropunctata do not build mounds, therefore, baits are typically broadcast rather than applied to a mound or individual nests. Although broadcast applications are regarded as more ef-fective than bait stations, there may be situations where bait sta-tions are preferred. To date, no studies have attempted to determine foraging distances and effective bait station densities to control W. auropunctata. Such studies often rely on data from mark-capture and mark-release-recapture studies, and marking techniques for this species are under investigation (Montgomery et al. 2019).Granular baits applied to the ground have been found to have little effect on arboreal ants in most situations. In Hawai`i, W. auropunctata nesting in tree canopies, palms, and tall foliage do not always forage on the ground and so some do not encounter the in-secticidal bait (Souza et al. 2008, Taniguchi 2008, Montgomery et al. 2015). However, in Cairns, Australia, arboreal W. auropunctata appear to descend and forage on the ground at times depending on the availability of food resources in the canopy or foliage in which they reside (G. Morton personal communication). Additionally, the distance to the ground and weather are also factors likely to influ-ence this. Therefore, the efficacy of bait applications on the ground may differ from one site to another. In areas where arboreal W. auropunctata do not forage on the ground, bait applications must be three-dimensional and include treatment of tree canopies and vegeta-tion (Souza et al. 2008, Taniguchi 2008, Vanderwoude and Nadeau 2009, Vanderwoude et al. 2010) using a gel or paste bait. Gel bait formulations have the advantage over granules of adhering to vege-tation and vertical surfaces, ensuring arboreal ants’ access to the bait (Vanderwoude and Nadeau 2009, Vanderwoude et al. 2010).Another issue is the weathering effects of sunlight, heat, and mois-ture on bait palatability and degradation. The corn grit carrier of granular baits quickly absorbs moisture from dew and rain making the bait soggy, reducing the oil content, and palatability becomes unpredictable (Banks et al. 1972, Hara et al. 2014, Oi et al. 2022). Oxidation of the soy oil food attractant resulting from heat, light, and humidity exposure also renders baits unpalatable (Markin and Hill 1971, Hara et al. 2014). This can occur from improper or prolonged storage of opened bait containers or after application in the field (M. Montgomery personal observation). Gel and paste baits are also sus-ceptible to weathering. Although dew and light rain may help keep the bait moist and palatable, heavy rains wash the bait away and dry conditions result in desiccation. Additionally, some active ingredients, such as hydramethylnon, undergo rapid aqueous photolysis with a half-life of 42 min when exposed to light and water (Mallipudi 1986). In wet climates, such as windward Hawai`i where average annual rain-fall may exceed 3,200 mm (Giambelluca et al. 2013, accessed 12 Oct 2021), there is a short window of opportunity for W. auropunctata to find, recruit to, feed on, and share the bait. Ensuring maximum at-tractiveness is maintained as long as possible and increasing chances for foragers to encounter the bait following broadcast applications is essential for achieving the highest treatment efficacy.Finally, W. auropunctata appear to be more sensitive to certain active ingredients used in ‘fire ant’ baits. Insect growth regulators (IGRs), such as s-methoprene and pyriproxyfen, appear to be repel-lant to W. auropunctata (Hara et al. 2014, Montgomery et al. 2015). The inclusion of additional phagostimulants or adjuvants has been shown to effectively mask repellent substances, although this is likely only an option for homemade baits as commercial baits are ready-to-use products (Montgomery et al. 2015).When site features, local climate, and chemical sensitivity are factored into the planning process for control operations, W. auropunctata can be effectively controlled using baits (Williams and Whelan 1992, Abedrabbo 1994, Ulloa-Chacón and Cherix 1994, Jourdan and Chazeau 2004, Causton et al. 2005, Vanderwoude et al. 2010, Cabral et al. 2012). Toxic (lethal) baits and IGR baits are equally effective for long-term control and management, but short-term results are drastically different. The use of toxic baits results in the rapid death of ants that ingest a lethal dose of the active in-gredient, whereas IGRs are generally considered nonlethal and af-fect egg production, pupation, and development of reproductive and brood castes (Jourdan and Chazeau 2004; Cabral et al. 2012, 2017). Toxic bait usage results in rapid population knockdown immediately after application followed by a recovery period. Results from IGR bait applications do not show noticeable results immediately after application and instead a gradual population decline is observed over time and with repeated applications. When rapid population knockdown is desired, such as with preharvest treatments to protect harvesters, toxic baits are preferred. However, the nontoxic nature of IGR baits carry lower potential for ecological and nontarget impacts while not compromising long-term efficacy.Nonchemical ControlWhen chemical control is either not effective or not desirable, such as for postharvest treatment of horticultural products including pro-duce, cut flowers, and some nursery stock, nonchemical disinfestation methods are needed. Irradiation and hot water treatments have been tested against W. auropunctata and both are potentially effective phytosanitary procedures (Hara et al. 2011, Calcaterra et al. 2012). However, neither method is 100% effective and thus may not meet phytosanitary requirements (Hara et al. 2011, Calcaterra et al. 2012).Biological ControlThe parasitoid wasp Orasema minutissima is a common parasitoid of W. auropunctata throughout the Caribbean. Rather than seeking Downloaded from https://academic.oup.com/aesa/advance-article/doi/10.1093/aesa/saac016/6697138 by guest on 16 September 202215Annals of the Entomological Society of America, 2022, Vol. XX, No. XXout W. auropunctata, O. minutissima oviposit in plant tissue, and emergent planidia (first instar larvae) are brought into the nest via phoretic attachment to a foraging worker or prey (Heraty 1994, Soto et al. 2010). Once inside the nest, they parasitize brood and produce cuticular hydrocarbons that mimic W. auropunctata brood leading to their acceptance within the nest. Although widespread and common throughout the Caribbean, parts of Central America, and northern South America, it is uncertain to what degree O. minutissima may aid in population control of W. auropunctata (Heraty 1994, Burks et al. 2018). The recent introduction and detection of O. minutissima on Hawai`i Island (Heraty et al. 2021), where W. auropuncata is widespread and a major pest, provides an opportunity to examine O. minutssima impacts on W. auropunctata over time.DiscussionWasmannia auropunctata is a truly fascinating species that has proven adaptable in the most surprising ways. From the unique re-productive systems (Fournier et al. 2005a) to their ability to invade a wide range of habitats and thermotolerance adaptability (Foucaud et al. 2013), W. auropunctataprovides us with a glimpse of evolution in action. For example, clonality may have arisen to ensure successful establishment under the pressures of constant natural disturbances and the success of W. auropunctata in human modified habitats is a consequence of this adaptation. The specific type of clonal repro-duction exhibited by W. auropunctata allows for the preservation of genotypes responsible for adaptations to specific ecological pres-sures, such as the expanded thermotolerance seen in Clad B (Chifflet et al. 2016, 2018; Coulin et al. 2019). What is not clear is whether other genetic adaptations are being preserved that influence the suc-cess of the different clades, haplogroups, or haplotypes.The thermotolerance adaptations seen in ‘Clade B’ has allowed W. auropunctata to successfully invade Mediterranean climates pre-viously believed to be unsuitable for the species (Rey et al. 2012; Chifflet et al. 2016, 2018; Coulin et al. 2019). Although rare, sexual reproduction has been documented in invasive clonal populations (Foucaud et al. 2006, Vonshak et al. 2009, Tindo et al. 2012). Could postinvasion sexual reproduction events lead to additional genetic adaptations? While this may be unlikely with the limited genetic di-versity of invasive clonal populations, repeated introductions of dif-ferent genotypes could lead to admixing and possibly the emergence of new adaptive phenotypes that may enhance its invasive potential. Alternately, genetic diversification might act as a regulating factor for invasive populations over time. The Caribbean invasion is be-lieved to be the result of numerous introductions from Central and South America with evidence of genetic recombination and mutation events (Foucaud et al. 2010b). Given that W. auropunctata has been present in the region for over 100 yr, studying the Caribbean inva-sion could provide insight on the emergence of genetic adaptations of historical invasive W. auropunctata populations over time.Wasmannia auropunctata is a serious pest ant throughout the world and is widely distributed, mostly in tropical and subtropical cli-mates. Its ability to establish with low propagule pressure (Mikheyev et al. 2008) and to spread long-distances via human-mediated trans-port (Walsh et al. 2004, Mikheyev et al. 2008, Foucaud et al. 2010b) will undoubtedly allow this species to continue to spread at an alarming rate in the absence of strong biosecurity and prevention efforts. Once established, W. auropunctata is adept at avoiding con-flict with competitor species while its population is low (Vonshak et al. 2012, Yitbarek et al. 2017). Infestations frequently go unnoticed for several years (Conant et al. 2007) and by the time a new infest-ation is detected, colony fragments have likely been inadvertently transported elsewhere. This is evident in Hawai`i, where new infest-ations on Hawai`i, Maui, Oahu, and Kaua`i islands continue to be detected despite over two decades of state-wide surveys, public out-reach, and eradication attempts (Conant and Hirayama 2000, Null and Gundersen 2007, Vanderwoude et al. 2010, Vanderwoude et al. 2015).Despite several studies investigating the relationship between W. auropunctata genetics, reproduction mode, and behavior (Foucaud et al. 2006, 2007, 2009, 2010a,b; Tindo et al. 2012; Rey et al. 2013b) our understanding is still limited and based on data collected from a select few locations. The biology and physiology associated with the ecological variegates and two clades highlight a plasticity rarely seen in nature. Could genetic adaptations be responsible for this plasticity be linked to specific haplogroups or haplotypes? It is possible that competitive, and thus invasion, potential may vary between clades, haplogroups, or haplotypes. It has been pointed out that exotic populations are unlikely to be traced back to their ancestral native population (Foucaud et al. 2010b). However, this may change with expanded sampling, phylogenetic analysis, and the pooling of gen-etic datasets. Understanding the driving forces, especially the rela-tionships between ecological factors and genetic adaptations, behind biological and physiological shifts in W. auropunctata may elucidate how studying different populations could result in the occurrence of conflicting behavioral observations.Understanding the driving forces behind biological and physio-logical shifts in W. auropunctata may also lead to the development of pesticide-free management tools and methods. For instance, if environmental factors can induce a physiological shift from nondominance to dominance, could the opposite also be true? Could RNAi target genes responsible for the physiological shift between nondominance/dominance? Additionally, rare occurrences of inva-sion contractions have been documented and it’s unclear what led to the contractions (Lester and Gruber 2016; Mbenoun Masse et al. 2019b, 2021). Long term studies looking at expansion and contrac-tion patterns may identify the circumstances necessary for a contrac-tion to occur. If those circumstances were identified, would we be able to artificially induce an invasion contraction?Interestingly, despite being a known invasive tramp ant and pest since the late 1800’s, W. auropunctata garnered little atten-tion from researchers and invasive species specialists until the early 2000s. In fact, the number of research publications focusing on W. auropunctata between 2000 and 2010 was almost double that pub-lished during the previous 80 yr. The importance of W. auropunctata as an invasive tramp ant was likely overshadowed by other pest ant species of importance such as S. invicta and L. humile and interest in it as a research subject was slow to develop. It is also possible that despite knowing the potential impacts of this species, researchers and invasive species specialists have underestimated its distribution potential and adaptability, thus assuming it to be a problem limited to the tropics and not relevant to cooler climates. Regardless, the delay in research focus has slowed the compilation of knowledge and understanding necessary to develop effective prevention and control methods. In this review, we have identified numerous know-ledge gaps spanning distribution and status, life history, impacts, biosecurity, chemical ecology, and management. As our knowledge base on this species grows and with technological advancements, new questions arise. We encourage a greater research focus on this species in line with its global importance as an invasive species. The case of W. auropunctata not only highlights the importance of species-specific understanding for management and control, but also for effective biosecurity and prevention of accidental new species introductions.Downloaded from https://academic.oup.com/aesa/advance-article/doi/10.1093/aesa/saac016/6697138 by guest on 16 September 202216Annals of the Entomological Society of America, 2022, Vol. XX, No. XXAcknowledgmentsFunding for this project was provided by the Hawai`i Invasive Species Council and the Hawai`i Department of Agriculture. This work forms part of the Ph.D. research by the corresponding author at the University of Canberra, Australia, and is supported by funding from the Hawai`i Invasive Species Council and the Hawai`i Department of Agriculture.Data AvailabilityData collected during and used for this project may be available upon reasonable request.Author ContributionsWriting of the original draft was done by Michelle Montgomery and all other authors contributed to revising and editing all subsequent drafts. The final draft has been approved by all authors.References CitedAbbott, I. 1989. The influence of fauna on soil structure, pp. 39–50. In J. D. Majer (ed.), Animals in primary succession – the role of fauna in reclaimed lands. Cambridge University Press, Melbourne.Abedrabbo, S. 1994. Control of the Little Fire Ant, Wasmannia auropunctata, on Santa Fe Island in the Galapagos Islands, pp. 63–72. In D. F. Williams (ed.), Exotic ants. Biology impact and control of introduced species. Westview Press, Boulder Colorado.Achury, R., P. Chacon de Ulloa, and A. Arcila. 2008. Ant composition and competitive interactions with Wasmannia auropunctata in tropical dry forest fragments. Rev. Colomb. Entomol. 34: 209–216.Achury, R., P. Ulloa-Chacon, and A. Arcila. 2012. Effects of the heterogeneity of the landscape and the abundance of wasmannia auropunctata on ground ant assemblages in a Colombian Tropical Dry Forest. Psyche 2012: 12.Achury, R., P. Chacón de Ulloa, Á. Arcila, and A. V. Suarez. 2020. Habitat dis-turbance modifies dominance, coexistence, and competitive interactions in tropical ant communities. Ecol. Entomol. 45: 1247–1262.Allen, M. L. 2021. Prospects for using RNAi as control for ants. Front. Agron. 3: 20.Allen, C. R., D. M. Epperson, and A. S. Garmestani. 2004. Red imported fire ant impacts on wildlife: a decade of research. Am. Midl. Nat. 152: 88–103.Álvarez, O. E. F., J. F. Díaz, and L. Y. G. Padrón. 2018. Apuntes sobre la invasión de Wasmannia auropunctata (Hymenoptera: Formicidae) en tres especies de bromelias de tanque presentes en el Parque Nacional Guanahacabibes, Cuba. Rev. ECOVIDA. 8: 28–38.Andersen, A. N. 1988. Soil of the nest-mound of the seed-dispersing ant, Aphaenogaster longiceps, enhances seedling growth. Aust. J. Zool. 13: 469–471.Angulo, E., B. Hoffmann, L. Ballesteros-Mejia, A. Taheri, P. Balzani, D. Renault, M. Cordonnier, C. Bellard, C. Diagne, and D. Ahmed. 2021. Economic costs of invasive alien ants worldwide. HAL-Open Archive, University of Paris.Apple, J. L., and D. H. Feener Jr. 2001. Ant visitation of extrafloral nectaries of Passiflora: the effects of nectary attributes and ant behavior on patterns in facultative ant-plant mutualisms. Oceologia. 127: 409–416.Armbrecht, I., and P. Ulloa-Chacon. 2003. The Little Fire Ant, Wasmannia auropunctata (Roger) (Hymenoptera: Formicidae) as a diversity indicator of ants in tropical dry forest fragments of colombia. Environ. Entomol. 32: 542–547.Baker, C. M., J. C. Hodgson, E. Tartaglia, and R. H. Clarke. 2017. Modelling tropical fire ant (Solenopsis geminata) dynamics and detection to inform an eradication project. Biol. Invasions. 19: 2959–2970.Banks, W. A., G. P. Markin, J. W. Summerlin, and C. S. Lofgren. 1972. Four Mirex Bait formulations for control of the red imported fire ant. J. Econ. Entomol. 65: 1468–1470.Beavan, A., J. McWilliam, E. van Strydonck, N. Rumboll, and J. Beynon. 2008. Impact of the invasive little red fire ant Wasmannia auropunctata on the Herpetofauna of the West African rainforest, pp. 1–6. James Rennie be-quest Report. University of Edinburgh.Berman, M., A. N. Andersen, and T. Ibanez. 2013. Invasive ants as back-seat drivers of native ant diversity decline in New Caledonia. Biol. Invasions. 15: 23111–2331.Bertelsmeier, C. 2021. Globalization and the anthropogenic spread of invasive social insects. Curr. Opin. Insect Sci. 46: 16–23.Bluthgen, N., G. Gebauer, and K. Fiedler. 2003. Disentangling a rainforest fod web using stable isoptoes: dietery diversity in a species-rich ant commu-nity. Oceologia. 137: 426–435.Bossin, H., and E. Padovani. 2010. Audit des actions menees depuis 2006 en matiere de lutte contre la Petite Fourmi de Feu Wasmannia auropunctata sur lile de Tahiti, pp. 1–27. Institut Lois Malarde.Bousseyroux, A., C. Blanvillain, and J. W. Beardsley Jr. 2019. La petite fourmi de feu (Wasmannia auropunctata): impacts écologiques en zone infestée dans le monde et risques en Polynésie, pp. 344. Bulletin de la Société des Estudes Océaniennes.Brandao, C. R. F., and R. V. S. Paiva (eds.). 1994. The Galapagos ant fauna and the attributes of colonizing ant species. Westview Press, Boulder, CO.Brandao, C. R. F., and R. R. Silva. 2008. Synecology of Wasmannia Auropunctata, an invasive ant species (Hymenoptera: Formicidae), con-tinuous and fragmented areas in the Brazilian Atlantic Forest, pp. 141–151. In T. D. Paine (ed.), Invasive forest insects, introduced forest trees, and altered ecosystems. Springer, Dordrecht, Netherlands.Braness, G. A. Year. Published 2002. Ant bait development: an imidacloprid case study. In S. C. Jones, J. Zhai and W. H. Robinson (eds.), Fourth International Conference on Urban Pests.Brooks, S., and J. C. Nickerson. 2008. Little Fire Ant, Wasmannia auropunctata (Roger) (Insecta: Hymenoptera: Formicidae). University of Florida IFAS Extension. report EENY139, Miami.Bruneau de Miré, P. 1969. Une fourmi utilisée au Cameroun dans la lutte contre les mirides du cacaoyer Wasmannia auropunctata Roger. Café Cacao. 13: 209–212.Buczkowski, G. 2019. Trap–treat–release: horizontal transfer of fipronil in field colonies of black carpenter ants, Camponotus pennsylvanicus. Pest Manag. Sci. 75: 2195–2201.Buczkowski, G., and T. C. Wossler. 2019. Controlling invasive Argentine ants, Linepithema humile, in conservation areas using horizontal insecticide transfer. Sci. Rep. 9: 1–7.Burks, R. A., J. M. Heraty, C. Dominguez, and J. L. Mottern. 2018. Complex diversity in a mainly tropical group of ant parasitoids: revision of the Orasema stramineipes species group (Hymenoptera: Chalcidoidea: Eucharitidae). Zootaxa. 4401: 1–107.Cabral, S. K., A. H. Hara, and R. Niino-DuPonte. 2011. Knock down and residual efficacy of contact insecticides against the Little Fire Ant, Wasmannia auropunctata (Roger) (Hymenoptera: Formicidae), pp. 43–46. In Proceedings of the 2011 Imported Fire Ant Conference.Cabral, S. K., A. H. Hara, and K. L. Aoki. 2012. Efficacy of Hydramethylnon, indoxacarb, and methoprene baits against the Little Fire Ant Wasmannia auropunctata (Roger) (Hymenoptera: Formicidea), pp. 80–84. In Proceedings of the 2012 Imported Fire Ant Conference.Cabral, S. K., A. H. Hara, and R. Niino-DuPonte. 2017. Response of little fire ant (Hymenoptera: Formicidae) colonies to insect growth regulators and hydramethylnon. Proc. Hawaiian Entomol. Soc. 49: 1–10.Cagliari, D., N. P. Dias, D. M. Galdeano, E. Á. Dos Santos, G. Smagghe, and M. J. Zotti. 2019. Management of pest insects and plant diseases by non-transformative RNAi. Front. Plant Sci. 10: 1319.Calcaterra, L. A., C. Coulin, J. A. Briano, and P. A. Follet. 2012. Acute ex-posure to low-dose radiation disrupts reproduction and shortens survival of Wasmannia auropunctata (Hymenoptera: Formicidae) queens. J. Econ. Entomol. 105: 817–822.Causton, C. E., C. R. Sevilla, and S. D. Porter. 2005. Eradication of the Little Fire Ant Wasmannia auropunctata, (Hymeoptera: Formicidae) from Marchena Island, Galapagos: on the edge of success? Fla. Entomol. 88: 159–168.Cha, D. H., D. Skabeikis, R. M. Collignon, M. S. Siderhurst, M. Y. Choi, and R. K. Vander Meer. 2019. Behavioral response of little fire ant, Wasmannia auropunctata (Hymenoptera: Formicidae), to trail chemicals laid on epi-phytic moss. J. Insect Behav. 32: 145–152.Downloaded from https://academic.oup.com/aesa/advance-article/doi/10.1093/aesa/saac016/6697138 by guest on 16 September 202217Annals of the Entomological Society of America, 2022, Vol. XX, No. XXChen, S.-q., Z. Yi, Y.-y. Lu, R. Hao, and Y.-j. Xu. 2022. First record of the little fire ant, Wasmannia auropunctata (Hymenoptera: Formicidae), in Chinese mainland. J. Integr. Agric. 21: 1825–1829.Chifflet, L., M. S. Rodriguero, L. A. Calcaterra, O. Rey, P. A. Dinghi, F. B. Baccaro, J. L. P. Souza, P. Follett, and V. A. Confalonieri. 2016. Evolutionary history of the little fire ant Wasmannia auropunctata before global inva-sion: inferring dispersal patterns, niche requirements and past and present distribution within its native range. J. Evol. Biol. 29: 790–809.Chifflet, L., N. V. Guzmán, O. Rey, V. A. Confalonieri, and L. A. Calcaterra. 2018. Southern expansion of the invasive ant Wasmannia auropunctata within its native range and its relation with clonality and human activity. PLoS One. 13: e0206602.Choe, D.-H., and M. K. Rust. 2008. Horizontal transfer of insecticides in laboratory colonies of the Argentine ant (Hymenoptera: Formicidae). J. Econ. Entomol. 101: 1397–1405.Clark, D. B., C. Guayasamin, O. Pazamino, C. Donoso, and Y. Paez de Villacis. 1982. The tramp ant Wasmannia auropunctata: autoecology and effect on ant diversity and distribution on Santa Cruz Island, Galapagos. Biotropica. 14: 196–207.Clemente, S. R., and S. R. Whitehead. 2020. Ant seed removal in a non-myrmecochorous Neotropical shrub: implications for seed dispersal. Biotropica. 52: 90–100.Conant, P., and C. Hirayama. 2000. Wasmannia auropunctata (Hymenoptera: Formicidae): established on the Island of Hawai`i. Bishop Mus. Occas. Pap. 64: 21–22.Conant, P., R. A. Heu, L. Nakahara, B. Kumashiro, and N. Reimer. 2007. Little fire ant Wasmannia auropunctata (Roger) (Hymenoptera: Formicidae) new pest advisory. Hawai`i Department of Agriculture Plant Pest Control, Honolulu, HI.Coulin, C., J. Gerardo, L. Chifflet, L. A. Calcaterra, and P. E. Schilman. 2019. Linking thermo-tolerances of the highly invasive ant, Wasmannia auropunctata, to its current and potential distribution. Biol. Invasions. 21: 3491–3504.Cox, N. A., G. J. Morton, M. K. McNaught, and R. Wylie. 2020. Novel re-usable canopy trap for sampling arboreal populations of electric ant, Wasmannia auropunctata (Hymenoptera: Formicidae). Austral Entomol. 60: 257–264.Cuezzo, F., L. A. Calcaterra, L. Chifflet, and P. A. Follet. 2015. Wasmannia Forel (Hymenoptera: Formicidae: Myrmicinae) in Argentina: systematics and distribution. Sociobiology. 62: 246–265.[dataset] Global Biodiversity Information Facility (GBIF). iNaturalist con-tributors, iNaturalist (2021). iNaturalist research-grade observations. iNaturalist.org. Occurrence dataset https://doi.org/10.15468/ab3s5x ac-cessed via GBIF.org on 2021-06-16. Available from https://www.gbif.org/occurrence/1640061844. Accessed 16 June 2021.[dataset] Global Biodiversity Information Facility (GBIF). iNaturalist con-tributors, iNaturalist (2022). iNaturalist research-grade observations. iNaturalist.org. Occurrence dataset https://doi.org/10.15468/ab3s5x ac-cessed via GBIF.org on 2022-07-06. Available from https://www.gbif.org/occurrence/3456381581. Accessed 7 July 2022.Davis, N. E., D. J. O’Dowd, R. Mac Nally, and P. T. Green. 2009. Invasive ants disrupt frugivory by endemic island birds. Biol. Lett. Published Online 15 September 2009: 1–4.de la Vega, I. 1994. Food searching behaviour and competition between Wasmannia auropunctata and native ants on Santa Cruz and Isabella, Galapagos Islands. pp. 73–79. In D. F. Williams (ed.), Exotic ants: biology, impact, and control of introduced species. Westview Press, Boulder Colorado.de Souza, A. L. B., J. H. C. Delabie, and H. G. Fowler. 1998. Wasmannia spp. (Hym. Formicidae) and insect damages to cocoa in Brazilian farms. J. Appl. Entomol. 122: 339–341.de Souza, A. L. B., C. S. F. Mariano, J. Delabie, S. G. Pompolo, and J. E. Serrao. 2011. Cytogenetic studies on workers of the neotropical ant Wasmannia auropunctata (Roger 1863) (Hymenoptera: Formicidae: Myrmicinae). Ann. Soc. Entomol. Fr. 47: 510–513.Del Toro, I., R. R. Ribbons, and S. L. Pelini. 2012. The little things that run the world revisited: a review of ant-mediated ecosystem services and disser-vices (Hymenoptera: Formicidae). Myrmecol. News. 17: 133–146.Delabie, J. H. C. 1988. Occurence of Wasmannia auropunctata (Roger, 1863) (Hymenoptera: Formicidae, Myrmicinae) in cacao plantations in Bahia, Brazil. Rev. Theob. 18: 29–37.Delabie, J. H. C. 1989. Preliminary evaluation of an alternative technique for the control of the little fire ant Wasmannia auropunctata in cacao plant-ations. Agrotropica. 1: 75–78.Delabie, J. H. C. 1995. Community structure of house-infesting ants (Hymeoptera: Formicidae) in southern Bahia, Brazil. Fla. Entomol. 78: 264–270.Delabie, J. H. C., and I. M. Cazorla. 1991. Damages caused by Planococcus citri Risso (Homoptera pseudococcidae) to the production of cocoa tree. Agrotropica. 3: 53–57.Delabie, J. H. C., A. M. V. Da Encarnacao, and I. M. Carzorla. 1994. Relationships between the Little Fire Ant, Wasmannia auropunctata and its associated mealybug Planococcus citri in Brazilian cocoa farms, pp. 91–103. In D. F. Williams (ed.), Exotic ants: biology, impact, and control of introduced species. Westview Press, Boulder COlorado.Delsinne, T., H. Jourdan, and J. Chazeau. 2001. Premieres donnees sur la mon-opolisation de reddources par l’envahisseur Wasmannia auropunctata (Roger) au sein d’une myrmecofauna de foret seche Neo-Caledonienne. Actes Coll. Insectes Soc. 14: 1–5.Derstine, N. T., E. J. Troyer, C. N. Suttles, L. A. Siderhurst, E. B. Jang, and M. S. Siderhurst. 2012. Field trapping the little fire ant, Wasmannia auropunctata. J. Insect Sci. 12: 1–13.Deyrup, M., L. Davis, and S. Cover. 2000. Exotic ants in Florida. Trans. Am. Entomol. Soc. 26: 293–326.Duan, X.-Y., X.-Y. Peng, and Z.-Q. Qian. 2016. The complete mitochondrial genomes of two globally invasive ants, the Argentine ant Linepithema humile and the little fire ant Wasmannia auropunctata. Conserv. Genet. Resour. 8: 275–277.Elith, J., M. Kearney, and S. Phillips. 2010. The art of modelling range-shifting species. Methods Ecol. Evol. 1: 330–342.Ennis, D. E., B. J. Mader, K. Burnside, E. Bauce, and E. Despland. 2015. Is feeding behavior on foliage affested by lab rearing on artificial diet? J. Insect Behav. 28: 147–156.Errard, C., J. Delabie, H. Jourdan, and A. Hefetz. 2005. Intercontinental chemical variation in the invasive ant Wasmannia auropunctata (Roger) (Hymenoptera Formicidae): a key to the invasive success of a tramp spe-cies. Naturwissenschaften. 92: 319–323.Espadaler, X., C. Pradera, and J. A. Santana. 2018. The first outdoor-nesting population of Wasmannia auropunctata in continental Europe (Hymenoptera, Formicidae). Iberomyrmex. 10: 1–8.Espadaler, X., C. Pradera, J. A. Santana, and A. R. Reyes. 2020. Dos nuevas poblaciones europeas de la pequeña hormiga de fuego, Wasmannia auropunctata (Roger, 1863) (Hymenoptera: Formicidae) en Andalucía (España). Bol. SAE. 30: 189–192.Eyer, P.-A., and E. L. Vargo. 2021. Breeding structure and invasiveness in social insects. Curr. Opin. Insect Sci. 46: 24–30.Fabres, G., and W. L. Brown jnr. 1978. The recent introduction of the pest ant Wasmannia auropunctata into New Caledonia. J. Aust. Entomol. Soc. 17: 139–142.Fasi, J. 2009. Quantifying the dominance of Little Fire Ant (Wasmannia auropunctata) and its effect on crops in the Solomon Islands. M.Sc. thesis, University of the South Pacific Suva, Fiji.Fasi, J., G. Brodie, and C. Vanderwoude. 2013. Increases in crop pests caused by Wasmannia auropunctata in Solomon Islands subsistence gardens. J. Appl. Entomol. 137: 580–588.Fasi, J., M. J. Furlong, and D. Fisher. 2016. Subsistence farmers management of infestations of the Little Fire Ant in garden plots on Bauro, Makira Province, Solomon Islands. Human Ecol. 44: 765–774.Federman, R., Y. Carmel, and R. Kent. 2013. Irrigation as an important factor in species distribution models. Basic Appl. Ecol. 14: 651–658.Feinsinger, P., and L. A. Swarm. 1978. How common are ant-repellent nectars? Biotropica. 10: 238–239.Feitosa, R. M. 2007. Aggregation and adult transportation in disturbed col-onies of Wasmannia auropunctata Roger (Hymenoptera, Formicidae). Insectes Soc. 54: 200–201.Fernald, H. T. 1947. The Little Fire Ant as a house pest. J. Econ. Entomol. 40: 428–428.Downloaded from https://academic.oup.com/aesa/advance-article/doi/10.1093/aesa/saac016/6697138 by guest on 16 September 202218Annals of the Entomological Society of America, 2022, Vol. XX, No. XXFolgarait, P. J. 1998. Ant biodiversity and its relationship to ecosystem func-tioning: a review. Biodivers. Conserv. 7: 1221–1244.Foucaud, J., H. Jourdan, J. Le Breton, A. Loiseau, D. Konghouleux, and A. Estoup. 2006. Rare sexual reproduction events in the clonal reproduc-tion system of introduced populations of the little fire ant. Evolution. 60: 1646–1657.Foucaud, J., D. Fournier, J. Orivel, J. H. C. Delabie, A. Loiseau, J. Le Breton, G. J. Kergoat, and A. Estoup. 2007. Sex and clonality in the little fire ant. Mol. Biol. Evol. 24: 2465–2473.Foucaud, J., J. Orivel, D. Fournier, J. Delabie, A. Loiseau, J. L. Breton, P. Cerdans, and A. Estoup. 2009. Reproductive system, social organization, human disturbance and ecological dominance in native populations of the little fire ant, Wasmannia auropunctata. Mol. Ecol. 18: 5059–5073.Foucaud, J., A. Estoup, A. Loiseau, O. Rey, and J. Orivel. 2010a. Thelytokous parthenogenesis, male clonality and genetic caste determination in the little fire ant: new evidence and insights from the lab. Heredity. 105: 205–212.Foucaud, J., J. Orivel, A. Loiseau, J. H. Delabie, H. Jourdan, D. Konghouleux, M. Vonshak, M. Tindo, J. L. Mercier, and D. Fresneau. 2010b. Worldwide invasion by the little fire ant: routes of introduction and eco-evolutionary pathways. Evol. Appl. 3: 363–374.Foucaud, J., O. Rey, S. R. L. Crespin, J. Orivel, B. Facon, A. Loiseau, H. Jourdan, M. Kenne, P. S. M. Masse, M. Tindo, M. Vonshak, A. Estoup. 2013. Thermotolerance adaptation to human-modified habitats occurs in the native range of the invasive ant Wasmannia auropunctata before long-distance dispersal. Evol. Appl. 6: 721–734.Fournier, D., A. Estoup, J. Orivel, J. Foucaud, H. Jourdan, J. Le Breton, and L. Keller. 2005a. Clonal reproduction by males and females in the little fire ant. Nature. 435: 1230–1234.Fournier, D., J. Foucaud, A. Loiseau, S. Cros-Arteil, H. Jourdan, J. Orivel, J. Le Breton, J. Chazeau, A. De Jean, L. Keller, et al. 2005b. Characterization and PCR multiplexing of polymorphic microsatellite loci for the invasive ant Wasmannia auropunctata. Mol. Ecol. Notes. 5: 239–242.Gasc, A., J. Anso, J. Sueur, H. Jourdan, and L. Desutter-Grandcolas. 2018. Cricket calling communities as an indicator of the invasive ant Wasmannia auropunctata in an insular biodiversity hotspot. Biol. Invasions. 20: 1099–1111.Giambelluca, T. W., Q. Chen, A. G. Frazier, J. P. Price, Y.-L. Chen, P.-S. Chu, J. K. Eischeid, and D. M. Delparte. 2013. Online rainfall atlas of Hawai`i. Bull. Am. Meteor. Soc. pp. 313–316.Grangier, J., J. L. Breton, A. Dejean, and J. Orivel. 2007. Coexistence be-tween Cyphomyrmex ants and dominant populations of Wasmannia auropunctata. Behav. Process. 74: 93–96.Green, P. T., D. J. O’Dowd, K. L. Abbott, M. Jeffery, K. Retallick, and R. MacNally. 2011. Invasional meltdown: invader–invader mutualism facili-tates a secondary invasion. Ecology. 92: 1758–1768.Gruber, M. A. M., M. Cooling, and A. R. Burne. 2016. PIAT: the pacific in-vasive ant toolkit, Pacific biosecurity. New Zealand Ministry of Foreign Affairs and Trade. Available from https://piat.org.nz/. Accessed 14 April 2021.Hahn, D. A., and D. E. Wheeler. 2002. Seasonal foraging activity and bait preferences of ants on Barro Colorado Island, Panama. Biotropica. 34: 348–356.Hara, A. H., S. K. Cabral, R. Y. Niino-Duponte, C. M. Jacobsen, and K. Onuma. 2011. Bait insecticides and hot water drenches against the little fire ant, Wasmannia auropunctata (Hymenoptera: Formicidae), infesting containerized nursery plants. Fla. Entomol. 94: 517–526.Hara, A. H., K. L. Aoki, S. K. Cabral, and R. Y. Niino-Duponte. 2014. Attractiveness of gel, granular, paste, and solid formulations of ant bait insecticides to the Little Fire Ant, Wasmannia auropunctata (Roger) (Hymenoptera: Formicidae). Proc. Hawaii. Entomol. Soc. 46: 45–54.Helms, K. R., and B. Vinson. 2002. Widespread association of the invasive ant Solenopsis invicta with an invasive mealybug. Ecology. 83: 2425–2438.Herard, F., M. Keller, W. J. Lewis, and J. H. Tumlinson. 1988. Beneficial arthropod behavior mediated by airborn semiochemicals III. Influence of age and experienc on flight chamber responses of Microplitis demolitor Wilkinson. J. Chem. Ecol. 14: 1583–1595.Heraty, J. M. 1994. Biology and importance of two eucharitid parasites of Wasmannia and Solenopsis, pp. 104–120. In D. F. Williams (eds.), Exotic ants: biology, impact and control of introduced species. Westview Press, Boulder, USA.Heraty, J. M., D. V. Rogers, M. T. Johnson, W. D. Perriera, A. J. Baker, E. Bitume, E. Murray, and L. Varone. 2021. New record in the Hawaiian Islands of Orasema minutissima (Hymenoptera: Eucharitidae), an ant-parasitic wasp and a potential biocontrol agent against the Little Fire Ant, Wasmannia auropunctata (Hymenoptera: Formicidae). Bishop Mus. Occas. Pap. 137: 7–18.Herrera, H. W., and C. E. Causon. 2008. Distribution of fire ants Solenopsis geminata and Wasmannia auropunctata (Hymenoptera: Formicidae) in the Galapagos Islands. Galapagos Res. 65: 11–14.Hölldobler, B., and E. O. Wilson. 1977. The number of queens: an important trait in ant evolution. Naturwissenschaften. 64: 8–15.Hölldobler, B., and E. O. Wilson. 1990. The ants. Springer-Verlag, USA.Hölldobler, B., and E. O. Wilson. 1995. Journey to the ants. Harvard University Press, U.S.A.Holway, D. A., L. Lach, A. V. Suarez, N. D. Tsutsui, and T. J. Case. 2002. The causes and consequences of ant invasions. Annu. Rev. Ecol. Syst. 33: 181–233.Honnay, O., and B. Bossuyt. 2005. Prolonged clonal growth: escape route or route to extinction? Oikos. 108: 427–432.Horvitz, C. C., and D. W. Schemske. 1984. Effects of ants and an ant-tended herbivore on seed production of a neotropical herb. Ecology. 65: 1369–1378.Horvitz, C. C., and D. W. Schemske. 1990. Spatiotemporal variation in insect mutualists of a neotropical herb. Ecology. 71: 1085–1097.Howard, D., M. Blum, T. Jones, and M. Tomalski. 1982. Behavioral responses to an alkylpyrazine from the mandibular gland of the ant Wasmannia auropunctata. Insectes Soc. 29: 369–374.Jiménez-Carmona, E., I. Armbrecht, R. Quintero, J. M. Lerma, and L. M. Constantino. 2019. Detección molecular de depredación de hypothenemus hampei (coleoptera: curculionidae) por wasmannia auropunctata (hymen-optera: formicidae). Biotecnol. Sect. Agropecu. Agroind. 17: 82–88.Jourdan, H. 1997a. Threats on Pacific islands: the spread of the tramp ant Wasmannia auropunctata (Hymenoptera: Formicidae). Pac. Conserv. Biol. 3: 61–64.Jourdan, H. 1997b. Are serpentine biota free from biological invasions? An ex-ample of an ant community from southern New Caledonia, pp. 107–108. In The ecology of ultramafic and metalliferous areas.Jourdan, H. 2001. The impact of th little fire ant invasion (Wasmannia auropunctata (Roger)) on New Caledinian herpetofauna: results of a study in scleropihyll forest habitat. Sociobiology. 38: 1–19.Jourdan, H., and J. Chazeau. 2004. Etude comparative de l’efficacité d’appâts toxiques utilisables contre Wasmannia auropunctata.Jourdan, H., and P. Dumas. 2004. Les espèces animales invasives dans le Pacifique: l’apport de la spatialisation dans le cas de la fourmi envahissante Wasmannia auropuctata, pp. 396–408. In Espaces tropicaux et risques: du local au global: actes des 10èmes journées de géographie tropicale des.Jourdan, H., R. A. Sadlier, and A. M. Bauer. 2001. Little fire ant invasion (Wasmannia auropunctata) as a threat to New Caledonian lizards: evi-dence from a sclerophyll forest (Hymenoptera: Formicidae). Sociobiology. 38: 283–299.Jourdan, H., S. Ledoux, Y. Allouche, and J. Foucaud. 2006. Spread of Wasmannia auropunctata in the Pacific. A comparison of interactions with local ant fauna at low altitudes in Tahiti and New Caledonia, IUSSI 2006 Congress, Washington, DC.Jucker, C., F. Rigato, and R. Regalin. 2008. Exotic ant records from Italy (Hymenoptera, Formicidae). Boll. Zool. Agrar. Bachic. Ser. II. 40: 99–107.Keifer, H. H. 1937. The bulletin, pp. 433–435. In C. D. o. Agriculture (ed.).Kidon, M., Y. Klein, and T. Weinberg. 2022. Little fire ant (Wasmannia auropunctata) in Israel – from nuisance to life threatening. Harefuah. 161: 207–209.Kirschenbaum, R., and J. K. Grace. 2007a. Dominant ant species in four habi-tats in Hawai`i (Hymenoptera: Formicidae). Sociobiology. 50: 1069–1073.Kirschenbaum, R., and J. K. Grace. 2007b. Agonistic interactions of four ant species occurring in Hawai`i with Coptotermes formosanus (Isoptera: Rhinotermitidae). Sociobiology. 50: 643–651.Downloaded from https://academic.oup.com/aesa/advance-article/doi/10.1093/aesa/saac016/6697138 by guest on 16 September 202219Annals of the Entomological Society of America, 2022, Vol. XX, No. XXKirschenbaum, R., and J. K. Grace. 2008. Agnostic responses of the Tramp Ants Anoplolepis gracilipes, Pheidole megacephala, Linepithema humile, and Wasmannia auropunctata (Hymenoptera: Formicidae). Sociobiology. 51: 673–683.Klotz, J. H., M. K. Rust, D. Gonzalez, L. Greenberg, H. Costa, P. Phillips, C. Gispert, D. A. Reierson, and K. Kido. 2003. Directed sprays and li-quid baits to manage ants in vineyards and citrus groves. J. Agric. Urban Entomol. 20: 31–40.Klotz, J. H., M. K. Rust, H. C. Field, L. Greenberg, and K. Kupfer. 2009. Low impact directed sprays and liquid baits to control Argentine ants (Hymenoptera: Formicidae). Sociobiology. 54: 101.Kondo, T., A. M. Arcila, L. I. Colorado, Y. Campos-Patiño, and P. Sotelo-Cardona. 2018. Wasmannia auropunctata (Roger)(Hymenoptera: Formicidae), pequeño pero voraz depredador de Diaphorina citri Kuwayama (Hemiptera: Liviidae). Acta Zool. Mex. 34: 1–5.Kropidlowski, S. J. 2014. Investigating the efficacy of commercial baits for the control of Yellow Crazy Ants (Anoplolepis gracilipes) and their im-pacts on Red-tailed Tropicbirds (Phaethon rubricauda). Master of Science, University of Hawai`i Hilo.Kudoh, A., T. Minamoto, and S. Yamamoto. 2020. Detection of herbivory: eDNA detection from feeding marks on leaves. Environ. DNA. 2: 627–634.Kulikowski, A. J. 2020. Ant–scale mutualism increases scale infestation, de-creases folivory, and disrupts biological control in restored tropical forests. Biotropica. 52: 709–716.Le Breton, J., J. Chazeau, and A. Dejean. 2002. Field experiment to assess the use of repellant substances by Wasmannia auropunctata (Formicidae: Myrmicinae) during food exploitation. Sociobiology. 40: 437–442.Le Breton, J., J. Chazeau, and H. Jourdan. 2003. Immediate impacts of inva-sion by Wasmannia auropunctata (Hymenoptera: Formicidae) on native litter ant fauna in a New Caledonian rainforest. Austral Ecol. 28: 204–209.Le Breton, J., J. C. H. Delabie, J. Chazeau, and H. Jourdan. 2004. Experimental evidence of large-scale unicoloniality in the tramp ant Wasmannia auropunctata (Roger). J. Insect Behav. 17: 263–271.Le Breton, J., H. Jourdan, J. Chazeau, J. Orivel, and A. Dejean. 2005. Niche opportunity and ant invasion: the case of Wasmannia auropunctata in a New Caledonian rain forest. J. Trop. Ecol. 21: 93–98.Le Breton, J., J. Orivel, J. Chazeau, and A. Dejean. 2007a. Unadapted Behaviour of native, dominant ant species during the colonization of an aggressive, invasive ant. Ecol. Res. 22: 107–114.Le Breton, J., A. Dejean, G. Snelling, and J. Orivel. 2007b. Specialized preda-tion on Wasmannia auropunctata by the army ant species Neivamyrmex compressinodis. J. Appl. Entomol. 131: 740–743.Lee, D. J., M. Motoki, C. Vanderwoude, S. T. Nakamoto, and P. S. Leung. 2015. Taking the sting out of Little Fire Ant in Hawai`i. Ecol. Econ. 111: 100.Lester, P. J., and M. A. M. Gruber. 2016. Booms, busts and population col-lapses in invasive ants. Biol. Invasions. 01: 1–11.Levings, S. C., and N. R. Franks. 1982. Patterns of nested dispersion in a trop-ical ground ant community. Ecology. 63: 338–344.Levy, R., Y. J. Chiu, and W. A. Banks. 1973. Laboratory evaluation of candi-date bait toxicants against the Red Imported Fire Ant Solenopsis invicta. Fla. Entomol. 56: 141–146.Liang, C. T., C. EShiels, A. B. Sandor, and E. Manette. 2019. Project: the im-pact of non-native predators on pollinators and native plant reproduc-tion in a Hawaiian Dryland Ecosystem SERDP project number: RC-2432 principal investigators. USDA Forest Service, Pacific Southwest Research Station, Institute of Pacific Islands Forestry.Lin, H. M., W. L. Chi, C. C. Lin, Y. C. Tseng, W. T. Chen, Y. L. Kung, Y. Y. Lien, and Y. Y. Chen. 2011. Fire ant-detecting canines: a complementary method in detecting red imported fire ants. J. Econ. Entomol. 104: 225–231.List, F., A. M. Tarone, K. Zhu-Salzman, and E. L. Vargo. 2022. RNA meets toxicology: efficacy indicators from the experimental design of RNAi studies for insect pest management. Pest Manag. Sci. 78: 3215–3225.Longino, J. T., and F. Fernández. 2007. Taxonomic review of the genus Wasmannia, pp. 80. In R. R. Snelling, B. L. Fisher and P. S. Ward (eds.), Advances in ant systematics (Hymenoptera: Formicidae): homage to E. O. Wilson – 50 years of contributions. Memoirs of the American Entomological Institute, pp. 271–289.Loope, L., and P. D. Krushelnycky. 2007. Current and potential ant impacts in the Pacific region. Proc. Hawaii. Entomol. Soc. 39: 69–73.Lowe, S., M. Browne, S. Boudjelas, and M. De Poorter. 2000. 100 of the world’s worst invasive alien species: a selection from the global inva-sive species database. The Invasive Species Specialist Group (ISSG) a specialist group of the Species Survival Commission (SSC) of the World Conservation Union (IUCN), AUckland, New Zealand.Lubin, Y. D. 1984. Changes in the native fauna of the Galapagos Islands fol-lowing invasion by the little red fire ant, Wasmannia auropunctata. Biol. J. Linn. Soc. 21: 229–242.Macom, T. E., and Porter, S. D. 1996. Comparison of polygyne and monogyne red imported fire ant (Hymenoptera: Formicidae) population densities. Ann. Entomol. Soc. America. 89: 535–543.Majer, J. D. a. H. C. D. 1999. Impact of tree isolation on arboreal and ground ant communities in cleared pasture in the Atlantic rain forest region of Bahia, Brazil. Insectes Soc. 46: 281–290.Mallipudi, N. M., S. J. Stout, A.-h. Lee, and E. J. Orloski. 1986. Photolysis of AMDRO fire ant insecticide active ingredient hydromethylnon (AC 217,300) in distilled water. J. Agric. Food Chem. 34: 150–1057.Markin, G. P., and S. O. Hill. 1971. Microencapsulated oil bait for control of the imported fire ant. J. Econ. Entomol. 64: 193–196.Martin, S., and F. Drijfhout. 2009. A review of ant cuticular hydrocarbons. J. Chem. Ecol. 35: 1151–1161.Mayron, S. 2019. Little fire ants alarm, public urged to watch out. Samoa Observer, Apia.Mbenoun Masse, P. S., M. Kenne, R. Mony, A. Dejean, and M. Tindo. 2011. Initial behavior in colony fragments of an introduced population of the invasive ant Wasmannia auropunctata. C.R. Biol. 334: 572–576.Mbenoun Masse, P. S., M. Tindo, M. Kenne, Z. Tadu, R. Mony, and C. Djiéto-Lordon. 2017. Impact of the invasive ant Wasmannia auropunctata (Formicidae: Myrmicinae) on local ant diversity in southern Cameroon. Afr. J. Ecol. 55: 423–432.Mbenoun Masse, P. S., E. T. Giovani, and R. Mony. 2019a. Household and home garden infesting arthropods (Ants and Myriapods) in the city of Yaoundé, Cameroon. Ecosystems. 10: 11.Mbenoun Masse, P. S., M. Tindo, C. Djieto-Lordon, R. Mony, and M. Kenne. 2019b. Diversity of ant assemblages (Hymenoptera: Formicidae) in an urban environment in Cameroon during and after colonization of the area by Wasmannia auropunctata. Eur. J. Entomol. 116: 461–467.Mbenoun Masse, P. S., M. Tindo, C. Djiéto-Lordon, R. Mony, and M. Kenne. 2021. Temporal variation in ant community assemblages along a rural–urban gradient in the Yaoundé metropolis, Cameroon. Urban Ecosyst. 25: 193–204.McGlynn, T. P. 1999. The worldwide transfer of ants: geographical distribu-tion and ecological invasions. J. Biogeogr. 26: 535–548.Meier, R. 1994. Coexisting patterns and foraging behavior of introduced and native ants (Hymenoptera Formicidae) in the Galapagos Islands (Ecuador), pp 44–62. In D. F. Williams (eds.), Exotic ants: biology, im-pact, and control of introduced species. Westview Studies in Insect Biology, Boulder, USA.Meurisse, N., D. Rassati, B. P. Hurley, E. G. Brockerhoff, and R. A. Haack. 2019. Common pathways by which non-native forest insects move inter-nationally and domestically. J. Pest Sci. 92: 13–27.Mikheyev, A. S., and U. G. Mueller. 2007. Genetic relationships between native and introduced populations of the little fire ant Wasmannia auropunctata. Divers. Distrib. 13: 1–7.Mikheyev, A. S., L. Tchingnoumba, A. Henderson, and A. Alonso. 2008. Effect of propagule pressure on the establishment and spread of the little fire ant Wasmannia auropunctata in a Gabonese oilfield. Divers. Distrib. 14: 301–306.Mikheyev, A. S., S. Bresson, and P. Conant. 2009. Single- queen intoductions characterize regional and local invasions by the facultatively clonal little fire ant Wasmannia auropunctata. Mol. Ecol. 18: 2937–2944.Mikissa, J. B., K. Jeffery, D. Fresneau, and J. L. Mercier. 2013. Impact of an invasive alien ant, Wasmannia auropunctata Roger., on a specialised plant-ant mutualism, Barteria fistulosa Mast. and Tetraponera aethiops F. Smith., in a Gabon forest. Ecol. Entomol. 38: 580–584.Downloaded from https://academic.oup.com/aesa/advance-article/doi/10.1093/aesa/saac016/6697138 by guest on 16 September 202220Annals of the Entomological Society of America, 2022, Vol. XX, No. XXMiyakawa, M. O., and A. S. Mikheyev. 2015. Males are here to stay: fertiliza-tion enhances viable egg production by clonal queens of the little fire ant (Wasmannia auropunctata). Sci. Nat. 102: 1–7.Montgomery, M. P., C. Vanderwoude, and A. J. J. Lynch. 2015. Palatability of baits containing (S)-methoprene to Wasmannia auropunctata (Hymenoptera: Formicidae). Fla. Entomol. 98: 451–455.Montgomery, M., J. R. Hagler, C. Vanderwoude, A. J. J. Lynch, and W. A. Robinson. 2019. Laboratory evaluation of egg white and milk external biomarkers for Wasmannia auropunctata (Hymenoptera: Formicidae). J. Insect Sci. 19: 18.Montgomery, M. P., C. Vanderwoude, A. J. J. Lynch, and W. A. Robinson. 2020. The effects of laboratory rearing diet on recruitment behavior of Wasmannia auropunctata (Hymenoptera: Formicidaea). Fla. Entomol. 103: 103–111.Morris, J. R., and I. Perfecto. 2016. Testing the potential for ant predation of immature coffee berry borer (Hypothenemus hampei) life stages. Agric. Ecosyst. Environ. 233: 224–228. doi:10.1002/eap.2653. In press.Morris, J. R., and I. Perfecto. 2022. An aggressive nonconsumptive effect medi-ates pest control and multipredator interactions in a coffee agroecosystem. Ecol. Appl. e2653.Morris, J. R., E. Jimenez-Soto, S. M. Philpott, and I. Perfecto. 2018. Ant-mediated (Hymenoptera: Formicidae) biological control of the coffee berry borer: diversity, ecological complexity, and conservation biocontrol. Myrmecol. News 26: 1–17.Motoki, M., D. J. Lee, C. Vanderwoude, L. J. Nakomoto, and P. S. Leung. 2013. A bioeconomic model of Little Fire Ant Wasmannia auropunctata in Hawai`i, pp. 89. Pacific Cooperative Studies Unit, University of Hawai`i, Honolulu Hawai`i. Report number 186.Naumann, K. 1994. An occurence of two exotic ant (Hymenoptera: Formicidae) species in British Columbia. J. Entomol. Soc. B.C. 91: 69–70.Ndoutoume-Ndong, A., and B. Mikissa. 2007. Influence de la présence de la fourmi Wasmannia auropunctata (Roger 1863) (Hymenoptera: Formicidae) sur les autres espèces de fourmis dans la réserve de la Lopé (centre du Gabon). Ann. Soc. Entomol. Fr. 43: 155–158.Nickerson, J. 1983. The Little Fire Ant, Ochetomyrmex auropunctata (Roger) (Hymenoptera: Formicidae). Florida Department of Agriculture and Consumer Services, Division of Plant Industry. Entomology Circular No. 248.Niemiec, R., N. Ardoin, F. Brewer, S. Kung, and K. Lopez. 2018. Increased neighbor interaction and fear of social sanctions: associations with resi-dent action to control the invasive little fire ant. Soc. Nat. Resour. 31: 1149–1168.Niemiec, R., R. Willer, N. Ardoin, and F. Brewer. 2019. Motivating land-owners to recruit neighbors for private land conservation. Conserv. Biol. 33: 930–941.Normark, B. B. 2003. The evolution of alternative genetic systems in insects. Annu. Rev. Entomol. 48: 397–423.Null, W., and K. Gundersen. 2007. Little Fire Ant Wasmannia auropunctata: its presence on Kauai, pp. 47. Kauai Invasive Species Committee, Hawai`i, USA.O’Dowd, D. J., P. T. Green, and P. S. Lake. 2003. Invasional ‘meltdown’ on an oceanic island. Ecol. Lett. 6: 812–817.Oi, D. H., A. Lucky, and D. M. Liebowitz. 2022. Response of Wasmannia auropunctata (Hymenoptera: Formicidae) to water-soaked imported fire ant baits. Fla. Entomol. 105: 108–114.Orivel, J., J. Grangier, J. Foucaud, J. L. Breton, F. X. Andres, H. Jourdan, J. Delabie, D. Fournier, P. Cerdan, B. Facon, et al. 2009. Ecologically hetero-geneous populations of the invasive ant Wasmannia auropunctata within its native and introduced ranges. Ecol. Entomol. 34: 504–512.Ortiz-Alvarado, Y., and B. Rivera-Marchand. 2020. Worker queens? Behavioral flexibility of queens in the little fire ant Wasmannia auropunctata. Front. Ecol. Evol. 8: 241.Ortiz-Alvarado, Y., R. Fernández-Casas, C. A. Ortiz-Alvarado, E. Diaz-Iglesias, and B. Rivera-Marchand. 2021. Behavioral flexibility in Wasmannia auropunctata (Hymenoptera: Formicidae). J. Insect Sci. 21: 16.Osburn, M. R. 1949. Tests of Parathion for the control of the Little Fire Ant. J. Econ. Entomol. 42: 542–543.Passera, L. 1994. Characteristics of tramp ants, pp. 22–43. In D. F. Williams (ed.), Exotic ants: biology, impact and control of introduced species. Westview Press, Boulder, USA.Patael, Y. Y., O. Segal, and F. Segev. 2019. Fire ant punctate keratopathy: a novel diagnosis based on clinical and anterior segment optical coherence tomography findings. Cornea. 38: 1550–1553.Perfecto, I., and J. Vandermeer. 2020a. Antagonism between Anolis spp. and Wasmannia auropunctata in coffee farms on Puerto Rico: potential com-plications of biological control of the coffee berry borer. Caribb. J. Sci. 50: 43–47.Perfecto, I., and J. Vandermeer. 2020b. The assembly and importance of a novel ecosystem: the ant community of coffee farms in Puerto Rico. Ecol. Evol. 10: 12650–12662.Perfecto, I., Z. Hajian-Forooshani, A. White, and J. Vandermeer. 2021. Ecological complexity and contingency: ants and lizards affect biological control of the coffee leaf miner in Puerto Rico. Agric. Ecosyst. Environ. 305: 107104.Plentovich, S., A. Hebshi, and S. Conant. 2008. Detrimental effects of two in-vasive ant species on weight and survival of colonial nesting seabirds in the Hawaiian Islands. Biol. Invasions. 11: 289–298.Poland, T. M., and D. Rassati. 2019. Improved biosecurity surveillance of non-native forest insects: a review of current methods. J. Pest Sci. 92: 37–49.Propkopy, R. J., M. Aluja, and T. T. Y. Wong. 1989. Foraging behavior of laboratory cultured Mediterranean Fruit Flies on field-caged host trees. Proc. Hawaii. Entomol. Soc. 29: 103–110.Resnik, J. R. 2018. Biodiversity under siege, invasive animals and the National Park Service: a state of the knowledge report. Natural Resource Report NPS/NRSS/BRD/NRR – 2018/1679. National Park Service, Fort Colins, Colorado.Rey, O., A. Loiseau, B. Facon, J. Foucaud, J. Orivel, J. Cornuet, S. Robert, G. Dobigny, J. H. C. Delabie, C. Dos Santos, et al. 2011. Meiotic recombin-ation dramatically decreased in thelytokous queens of the Little Fire Ant and their sexually produced workers. Mol. Biol. Evol. 28: 2591–2601.Rey, O., A. Estoup, M. Vonshak, A. Loiseau, S. Blanchet, L. A. Calcaterra, L. Chifflet, J. P. Rossi, G. J. Kergoat, J. Foucaud, et al. 2012. Where do adap-tive shifts occur during invasion? A multidisciplinary approach to unrav-elling cold adaptation in a tropical ant species invading the Mediterranean area. Ecol. Lett. 15: 1266–1275.Rey, O., B. Facon, J. Foucaud, A. Loiseau, and A. Estoup. 2013a. Androgenesis is a maternal trait in the invasive ant Wasmannia auropunctata. Proc. R. Soc. Lond. 280: 1–17.Rey, O., A. Estoup, B. Facon, A. Loiseau, A. Aebi, O. Duron, F. Vavre, and J. Foucaud. 2013b. Distribution of endosymbiotic reproductive manipula-tors reflects invasion process and not reproductive system polymorphism in the little fire ant Wasmannia auropunctata. PLoS One. 8: e58467.Robinson, G. E., R. E. Page Jr, and Z. Huang. 1994. Temporal polyethism in social insects is a developmental process. Anim. Behav. 48: 467–469.Rojas, P., and C. Fragoso. 2021. A regional approach shows differences among invasive ants Solenopsis geminata and Wasmannia auropunctata (Hymenoptera: Formicidae) within its native range of distribution. J. Insect Sci. 21: 12.Roque-Albelo, L., C. Causton, and A. Mieles. 2000. The ants of Machena Island, twelve years after the introduction of the Little Fire Ant Wasmannia auropunctata. Notic. Galapagos. 61: 17–20.Rosselli, D., and J. K. Wetterer. 2017. Stings of the ant Wasmannia auropunctata (Hymenoptera: Formicidae) as cause of punctate corneal le-sions in humans and other animals. J. Med. Entomol. 20: 1–3.Rosumek, F. B. 2017. Natural history of ants: what we (do not) know about trophic and temporal niches of neotropical species. Sociobiology. 64: 244–255.Rust, M. K., D. A. Reierson, E. Paine, and L. J. Blum. 2000. Seasonal activity and bait preference of the Argentine ant (Hymenoptera: Formicidae). J. Agric. Urban Entomol. 94: 511–515.Salguero Rivera, B., I. Armbrecht, H. Hurtado, and N. Arcila. 2011. Wasmannia auropunctata (Hymenoptera: Formicidae): ¿unicolonial o multicolonial? en el valle geográfico del río Cauca. Rev. Colomb. Entomol. 37: 279–288.Downloaded from https://academic.oup.com/aesa/advance-article/doi/10.1093/aesa/saac016/6697138 by guest on 16 September 202221Annals of the Entomological Society of America, 2022, Vol. XX, No. XXSantos, M. G., G. F. Porto, I. R. Lancellotti, and R. M. Feitosa. 2019. Ant fauna associated with Microgramma squamulosa (Kaulf.) de la Sota (Polypodiaceae) fern galls. Rev. Bras. Entomol. 63: 101–103.Schemske, D. W. 1980. The evolutionary significance of extrafloral nectar pro-duction by Costus woodsoni (Zingibaceae): an experimental analysis of ant protection. J. Ecol. 68: 959–967.Showalter, D. N., E. J. Troyer, M. Aklu, E. B. Jang, and M. S. Siderhurst. 2010. Alkylpyrazines: alarm pheromone components of the little fire ant, Wasmannia auropunctata (Roger) (Hymenoptera, Formicidae). Insectes Soc. 57: 223–232.Silberglied, R. 1972. The little fire ant, Wasmannia auropunctata, a serious pest in the Galapagos Islands. Notic. Galápagos. 19: 13–15.Silva, L. M. R., M. de Oliveira Ramalho, C. Martins, V. G. Martins, and O. C. Bueno. 2018. Diversity of Wasmannia auropunctata (Hymenoptera: Formicidae) and the use of mitochondrial intergenic spacer and leucine tRNA for its identification. Rev. Ciênc. Ambient.12: 81–93.Smith, M. R. 1929. Two introduced ants not previously known to occur in the United States. J. Econ. Entomol. 22: 241–243.Smith, M. R. 1942. The relationship of ants and other organisms to certain scale insects on coffee in Puerto Rico. J. Agric. Univ. P. R. 26: 21–27.Smith, M. R. 1965. House infesting ants of the eastern United States: their recog-nition, biology and economic importance, pp. 105, United States Department of Agriculture, Agriculture Research Services Technical Bulletin 1326.Smith, J. M. (ed.). 1978. The evolution of sex. Cambridge University Press, Cambridge.Soeprono, A. M., and M. K. Rust. 2004. Effect of horizontal transfer of barrier insecticides to control Argentine ants (Hymenoptera: Formicidae). J. Econ. Entomol. 97: 1675–1681.Solomon, S. E., and A. S. Mikheyev. 2005. The ant (Hymenoptera: Formicidae) fauna of Cocos Island, Costa Rica. Fla. Entomol. 88: 417–422.Soto, J., J. Ortiz, E. Murray, B. Morris, L. Varone, and J. Heraty. 2010. Taxonomy and life history of Orasema minutissima (Hymenoptera: Eucharitidae) attacking the little red fire ant in the Caribbean, pp. 339–341. In Entomological Society American Annual Meeting.Souza, E., P. A. Follett, D. K. Price, and E. A. Stacy. 2008. Field suppression of the invasive ant Wasmannia auropunctata (Hymenoptera: Formicidae) in a tropical fruit orchard in Hawai`i. J. Econ. Entomol. 101: 1068–1074.Souza, A. L. B., A. G. Tavares, J. E. Serrao, and J. H. C. Delabie. 2009. Genetic variability of native populations of Wasmannia Forel (Hymenoptera: Formicidae) and their biogeographical implications. Neotrop. Entomol. 38: 376–383.Spencer, H. 1941. The small fire ant Wasmannia in citrus groves – a prelim-inary report. Fla. Entomol. 24: 6–14.Stein, M. B., H. G. Thorvilson, and J. W. Johnson. 1990. Seasonal changes in bait preference by red imported fire ant, Solenopsis invicta (Hymenoptera: Formicidae). Fla. Entomol. 73: 117–123.Taniguchi, G. 2008. Field efficacy studies on Wasmannia auropuncata with ant baits registered for use on tropical fruit crops in Hawai`i. Final Report. University of Hawai`i, Honolulu.Tennant, L. E. 1994. The ecology of Wasmannia auropunctata in primary tropical rainforest in Costa Rica and Panama, pp. 80–90. In D. F. Williams (ed.), Exotic ants: biology, impact, and control of introduced species. Westview Press, Boulder Colorado.Theron, L. 2007. Hypothèse d’une kératopathie liée à Wasmannia auropunctata, le modèle polynésien. Masters, University de Liege Liege, Belgium.Tindo, M., P. S. Mbenoun Masse, M. Kenne, R. Mony, J. Orivel, A. Doumtsop Fotio, A. Fotso Kuate, C. Djieto-Lordon, A. Fomena, A. Estoup, et al. 2012. Current distribution and population dynamics of the little fire ant supercolony in Cameroon. Insectes Soc. 59: 175–182.Tollerup, K. E., M. K. Rust, K. W. Dorschner, P. A. Phillips, and J. H. Klotz. 2004. Low-toxicity baits control ants in citrus orchards and grape vine-yards. Calif. Agric. 58: 213–217.Troyer, E. J., N. T. Derstine, D. N. Showalter, E. J. Jang, and M. S. Ciderhurst. 2009. Field studies of Wasmannia auropunctata alkylpyrazines: towards management applications. Sociobiology. 54: 955–971.Uchida, N., K. Kubota, S. Aita, and S. Kazama. 2020. Aquatic insect commu-nity structure revealed by eDNA metabarcoding derives indices for envir-onmental assessment. PeerJ. 8: e9176.Ulloa-Chacon, P., and D. Cherix. 1990. The Little Fire Ant Wasmannia auropunctata (R.) (Hymenoptera: Formicidae), pp. 281–289. In R. K. Vander meer, K. Jaffe and A. Cedeno (eds.), Applied myrmecology a world perspective. Westview Press, Boulder, USA.Ulloa-Chacón, P., and D. Cherix. 1994. Perspectives on Control of the Little Fire Ant (Wasmannia auropunctata) on the Galapagos Island, pp. 63–72. In D. F. Williams (ed.), Exotic ants: biology, impact and control of intro-duced species. Westview Press, Boulder, USA.Valles, S. M., R. Wylie, C. J. Burwell, M. K. McNaught, and C. Horlock. 2017. Evaluation of a lateral flow immunoassay for field identification of Solenopsis invicta (Hymenoptera: Formicidae) in Australia. Austral Entomol. 56.Valles, S. M., C. A. Strong, R. S. Emmitt, C. T. Culkin, and R. D. Weeks. 2020. Efficacy of the InvictDetectTM Immunostrip® to taxonomically identify the red imported fire ant, Solenopsis invicta, using a single worker ant. Insects. 11: 37. doi:10.3390/insects11010037.Vandermeer, J., and I. Perfecto. 2020. Endogenous spatial pattern formation from two intersecting ecological mechanisms: the dynamic coexistence of two noxious invasive ant species in Puerto Rico. Proc. R. Soc. Lond. B Biol. Sci. 287: 20202214.Vanderwoude, C., and B. Nadeau. 2009. Application methods for paste bait formulations in control of ants in arboreal situations. Proc. Hawaii. Entomol. Soc. 41: 113–119.Vanderwoude, C., K. Onuma, and N. Reimer. 2010. Eradicating Wasmannia auropunctata (Hymenoptera: Formicidae) from Maui, Hawai`i: the use of combination treatments to control an arboreal invasive ant. Proc. Hawaii. Entomol. Soc. 42: 23–31.Vanderwoude, C., D. Haynes, E. Richards, B. Quinquis, D. Moverley, and P. Skelton. 2014. Managing the impacts of the little fire ant (Wasmannia auropunctata) in French Polynesia, pp. 53. Secretariat of the Pacific Regional Environmental Programme, Apia, Samoa.Vanderwoude, C., M. P. Montgomery, H. Forester, E. Hensley, and M. K. Adachi. 2015. The history of little fire ants Wasmannia auropunctata Roger in the Hawaian Islands: spread, control, and local eradication. Proc. Hawaii. Entomol. Soc. 48: 39–50.Vanderwoude, C., S. Boudjelas, M. Gruber, B. Hoffmann, D. Oi, and S. Porter. 2021. Biosecurity plan for invasive ants in the Pacific region, pp. 275–288. In T. Pullaiah and M. R. Ielmini (eds.), Invasive alien species: observations and issues from around the world, vol. 2. John Wiley & Sons, Hoboken, USA.Vogt, J. T., B. Wallet, and S. Coy. 2008. Dynamic thermal structure of imported fire ant mounds. J. Insect Sci. 8: 31.Vonshak, M., T. Dayan, J. Foucaud, A. Estoup, and A. Hefetz. 2009. The inter-play between genetic and environmental effects on colony insularity in the clonal invasive little fire ant Wasmannia auropunctata. Behav. Ecol. 63: 1667–1677.Vonshak, M., T. Dayan, A. Ionescu-Hirsh, A. Freidberg, and A. Hefetz. 2010. The little fire ant Wasmannia auropunctata: a new invasive species in the Middle East and its impact on the local arthropod fauna. Biol. Invasions. 12: 1825–1837.Vonshak, M., T. Dayan, and A. Hefetz. 2012. Interspecific displacement mechanisms by the invasive little fire ant Wasmannia auropunctata. Biol Invasions. 14: 851–861.Walker, K. L. 2006. Impact of the Little Fire Ant, Wasmannia auropunctata, on native forest ants in Gabon. Biotropica. 38: 666–673.Walsh, P. D., P. Henschell, K. A. Abernethy, C. E. G. Tutin, P. Telfer, and S. A. Lahm. 2004. Logging speeds little red fire ant invasion of Africa. Biotropica. 36: 637–641.Ward, P. S. 2007. Phylogeny, classification, and species-level taxonomy of ants (Hymenoptera: Formicidae). Zootaxa. 1668: 549–563.Way, M. J., and B. Bolton. 1997. Competition between ants for coconut palm nesting sites. J. Nat. Hist. 31: 439–455.Werren, J. H., L. Baldo, and M. E. Clark. 2008. Wolbachia: master manipula-tors of invertebrate biology. Nat. Rev. Microbiol. 6: 741–751.Wetterer, J. K. 2013. Worldwide spread of the little fire ant, Wasmannia auropunctata (Hymenoptera: Formicidae). Terr. Arthropod Rev. 6: 173–184.Wetterer, J. K., and S. D. Porter. 2003. The Little Fire Ant, Wasmannia auropunctata: distribution, impact and control. Sociobiology. 41: 1–41.Downloaded from https://academic.oup.com/aesa/advance-article/doi/10.1093/aesa/saac016/6697138 by guest on 16 September 202222Annals of the Entomological Society of America, 2022, Vol. XX, No. XXWetterer, J. K., P. D. Walsh, and L. J. T. White. 1999. Wasmannia auropunctata (Roger) (Hymenoptera: Formicidae), a destructive tramp-ant, in wildlife refuges of Gabon. Afr. Entomol. 7: 1–3.Wetterer, J. K., L. D. Wood, C. J. Johnson, H. Krahe, and S. Pritchett. 2007. Predacious ants, beach replenishment, and nest placement by sea turtles. Environ. Entomol. 36: 1084–1091.Wheeler, W. M. 1929. Two neotropical ants established in the Unites States. Psyche. 36: 89–90.Williams, D. 1983. The development of toxic baits for the control of the im-ported fire ant. Fla. Entomol. 66: 162–172.Williams, D. F. 1994. Control of the introduced pest Solenopsis invicta in the United States, pp. 282–292. In D. F. Williams (ed.), Exotic ants: biology, impact and control of introduced species. Westview Press, Boulder, USA.Williams, D. F., and P. M. Whelan. 1992. Bait attraction of the introduced pest ant Wasmannia auropunctata (Hymeoptera: Formicidae) in the Galapagos Islands. J. Entomol. Sci. 27: 29–34.Wilson, E. O., and R. W. Taylor. 1967. The ants of Polynesia. Pac. Insects Monogr. 14: 1–109.Wisniewski, C., A. J. Bornstein, and D. L. Wood. 2019. Eating out or dining in: insect-plant interactions among several species of Piper in the Rio Abajo forest preserve, Puerto Rico. Selbyana. 33: 1–15.Wylie, R., M. McNaught, J. Oakey, and E. J. Harris. 2016. Eradication of two incursions of the red imported fire ant in Queensland, Australia. Ecol. Manage. Restor. 17: 22–32.Yasashimoto, T., M. K. Sakata, T. Sakita, S. Nakajima, M. Ozaki, and T. Minamoto. 2021. Environmental DNA detection of an invasive ant species (Linepithema humile) from soil samples. Sci. Rep. 11: 1–12.Yitbarek, S., J. H. Vandermeer, and I. Perfecto. 2017. From insinuator to dom-inator: foraging switching by an exotic ant. Divers. Distrib. 23: 820–827.Zhang, L., L. Wang, J. Chen, J. Zhang, Y. He, Y. Lu, J. Cai, X. Chen, X. Wen, and Z. Xu. 2022. Toxicity, horizontal transfer, and physiological and be-havioral effects of cycloxaprid against Solenopsis invicta (Hymenoptera: Formicidae). Pest Manag. Sci. 78: 2228–2239.Downloaded from https://academic.oup.com/aesa/advance-article/doi/10.1093/aesa/saac016/6697138 by guest on 16 September 2022View publication stats | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/264171277Time and Newsweek's Coverage of the Arab Uprisings in 2011: A ContentAnalysis SurveyArticle in The Asian review of World Histories · January 2014DOI: 10.12773/arwh.2014.2.1.081CITATIONS4READS3371 author:Some of the authors of this publication are also working on these related projects:Ahmad Surkitti in Indonesia View projectAhmed AbushoukQatar University18 PUBLICATIONS 88 CITATIONS SEE PROFILEAll content following this page was uploaded by Ahmed Abushouk on 25 June 2015.The user has requested enhancement of the downloaded file.Asian Review of World Histories 2:1 (January 2014), 81-104 © 2014 The Asian Association of World Historians doi: http://dx.doi.org/10.12773/arwh.2014.2.1.081 Time and Newsweek’s Coverage of the Arab Uprisings in 2011: A Content Analysis Survey Ahmed Ibrahim ABUSHOUK Qatar University Doha, Qatar ahmed.abushouk@qu.edu.qa Abstract The popular uprisings that took place in the Arab world, and led to the over-throw of four heads of states, namely Zine El-Abidine Ben Ali (January 14, 2011) of Tunisia, Hosni Mubarak (February 11, 2011) of Egypt, Muammar al-Gaddafi (August 23, 2011) of Libya and Ali Abdullah Saleh (November 23, 2011) of Yemen, have attracted the attention of the world media and policy makers in the West and the Middle East, and triggered their concern for the political future of the region. This article does not offer a comprehensive assessment of these upris-ings, but rather analyzes the coverage of Time and Newsweek of the underly-ing causes of the uprisings and their anticipated consequences. It also investi-gates how the two magazines have highlighted the scenarios that may pose a real challenge to Arab regimes supported by the American administration, and internationally reshape the priorities of American foreign policy in the region. These issues are examined from the two magazines’ perspectives, which under-82 | ASIAN REVIEW OF WORLD HISTORIES 2:1 (JANUARY 2014) line the features of U.S. foreign policy in the region, where the White House is more concerned about the security of the state of Israel, control of the Arab oil and suppression of “Muslim fundamentalism.” Keywords Arab uprisings, Muslim Fundamentalism, Newsweek Magazine, Time Maga-zine, Egypt, Libya, Syria, Tunisia, Yemen TIME AND NEWSWEEK MAGAZINES Time magazine was founded in 1923 by Henry R. Luce (d. 1967) and Briton Hadden (d. 1929) who obtained their university degrees in Journalism and began their careers as editors on the Daily News at Yale. A few years later, they established Time magazine with the objective to deliver world news to the mass public in the United States of America and internationally. During the Vietnam War, Time covered war events from a republican perspective contrasting the Newsweek’s liberal one. This magazine’s conservative leanings seems to have been adopted by its successive thirteen managing editors1, including its current one, Richard Rick Stengel, who has a wide number of publications, including his well-known work enti-tled, Mandela’s Way: Fifteen Lessons on Life, Love, and Courage. Newsweek magazine was founded by Thomas J.C. Martyn, a Time former foreign editor, on February 17, 1933, in New York City. It is currently considered as the second-largest weekly news maga-zine in the United State of America, having trailed behind Time in circulation at national and international levels. It offers a compre-hensive coverage of world events with a global network of corre- 1 James Kelly, “History of Time,” Time, March 31, 2003. Briton Hadden (1923–1929), Henry Luce (1929–1949), and Thomas S. Matthews (1949–1953), whose title was managing edi-tor. From 1949 onward the managing editor became the senior editor of the magazine and his assistants continued as editors. The list of the managing editors who succeeded Thomas S. Mat-thews includes Roy Alexander (1949–1960), Otto Fuerbringer (1960–1968), Henry Grunwald (1968–1977), Ray Cave (1979–1985), Jason McManus (1985–1987), Henry Muller (1987–1993), James R. Gaines (1993–1995), Walter Isaacson (1996–2000), Jim Kelly (2001–2006), and Rich-ard Stengel (2006–present). For further details, see, Wikipedia, s.v. “Time Magazine,” accessed February 20, 2012, http://en.wikipedia.org/wiki/Time _%28magazine%29. ABUSHOUK: “TIME AND NEWSWEEK’S COVERAGE OF THE ARAB UPRISINGS IN 2011” | 83 spondents, reporters and editors covering national and interna-tional affairs, business, science and technology, society, the arts and entertainment.2 Its present managing editor, Daniel Klaidman, is an established journalist. His service as Washington Bureau Chief enabled him to cover the war in Iraq and co-author several stories on Al Qaeda’s resurgence within the U.S. Both Time and Newsweek have given special coverage and de-scriptive analysis of the root causes of the Arab uprisings in 2011, and their anticipated consequences that may change the political landscape in the region. These issues are examined in the context of articles, reports, editorial notes, and communiques published by the two magazines in 2011. A HISTORICAL SURVEY OF THE ARAB UPRISINGS The Arab uprisings started in Tunisia on December 17, 2010, when a policewoman confiscated a vegetable cart of a twenty-six year street vendor, Mohamed Bouazizi, in Sidi Buzid 300 km south of Tunis. As a result, Bouazizi appealed to the provincial headquarters in Sidi Buzid, where he was humiliated by police officers who re-jected his case. An hour later Bouazizi doused himself with flam-mable liquid and set himself on fire. Public outrage immediately erupted in Sidi Buzid where protestors were crushed by police and security forces. Social media sites, such as Facebook and YouTube, featured images of police dispersing youths who had attacked shop windows and damaged cars. At that moment Bouazizi was trans-ferred to a hospital near Tunis, where he was visited by President Zine el Abidine Ben Ali in an attempt to quell the unrest. On Janu- 2 “History of Newsweek,” The Daily Beast, accessed February 24, 2012, http://www .thedailybeast.com/newsweek/2007/10/10/history-of-newsweek.html. 84 | ASIAN REVIEW OF WORLD HISTORIES 2:1 (JANUARY 2014) ary 4, 2011, Bouazizi died of burns in the hospital, and large scale demonstrations broke out in different parts of the country. Security forces brutally tried to suppress the demonstrators, whose prime demands were –“Jobs for all,’’ –“Down with the bribes and favourit-ism,’’ - “Tunisia free’’ and “Ben Ali get lost.’’ 3 In a desperate attempt to quiet down protestors, in his second speech since the outbreak of uprisings in Tunisia, President Ben Ali promised that 300,000 jobs would be created in the next two years, while issuing a deci-sion that closed down schools and universities, and branded pro-testers as “terrorists.’’ These measures did not calm down protes-tors, but rather provoked them to further confrontations with the government, calling for Ben Ali to step down and hand over his power to an elected democratic government. Under this mounting pressure of protests, the President dissolved his government called for parliamentary elections within six months, and promised the protestors that he would step down by the end of his presidential term in 2014. These promises did not improve the political situa-tion on the ground but rather weakened the grip of the govern-ment and forced Ben Ali to leave the country for Saudi Arabia on January 14, 2011, marking the end of his 24 years of authoritarian rule in Tunisia.4 The success of the Tunisian uprisings inspired other Arab countries to revolt against authoritarian leaders in their own coun-tries, who have ruled for many decades, particularly in Egypt and Libya. On January 25, 2011, Egyptian activists called for uprisings against poverty, unemployment and government corruption perpe-trated by President Hosni Mubarak and his closest associates. They urged the President to step down in favour of an elected democrat-ic government that would address the protestors’ demands and ex-pectations. On the next day, the government banned all public gatherings, and security forces used tear gas and concussion gre-nades to disperse the peaceful demonstrations. Protestors were 3 “Timeline of the Tunisian Revolution,” Bulletin of Mass Strike, accessed February 18, 2012, http://www.luxemburgism.lautre.net/spip.php?article127. 4 For further details, see, Vivienne Walt, “Tunisia’s Nervous Neighbors Watch the jas-mine Revolution,” Time, January 31, 2011, 17-21, especially 18. ABUSHOUK: “TIME AND NEWSWEEK’S COVERAGE OF THE ARAB UPRISINGS IN 2011” | 85 dispersed by police and security forces, a curfew was set up, and all forms of communication, including Internet and mobile services, were blocked. The protestors continued their struggle against the authoritarian regime and their revolts spread from the Liberation Square (Midan al-Tahrir) in Cairo to other parts of the country. The first reaction of the President to the demonstrations was that he appointed Omar Sulaiman, Egypt’s intelligence chief, as Vice Presi-dent, and advised him to start dialogue with all political forces. On February 4, 2011, thousands of protesters gathered at Tahrir Square in Cairo and other principal cities of Egypt, calling for the depar-ture of Mubarak without making any concession with the govern-ment. The President was left with no choice but to hand over his power to his newly appointed Vice-President, Omar Sulaiman, and remain in office until the end of his presidential term in 2013. But these proposals did not satisfy the demands of the protestors who maintained their position against the regime as a whole. When the regime reached a dead lock, Mubarak was internally and externally forced to step down on February 11, 2011, leaving the administration of the country to a military council headed by Mohamed Hussein Tantawi and a team of senior military officers who were members of the old guard of Mubarak’s regime. 5 A few days after President Hosni Mubarak’s downfall, anti-government demonstrations began in Benghazi in Libya, when se-curity forces arrested a human rights activist, Fethi Tarbel, who had been fighting to free political prisoners. Diplomats at Libya’s mission to the U.N. sided with the revolt against their country’s leader and urged the Libyan army to help overthrow “the tyrant Muammar Gaddafi.” By the end of February 2011, Gaddafi’s gov- 5 For further details, see, Babak Dehghanpisheh, Christopher Dickey, and Mike Giglio, “Rage Against the Regime,” Newsweek, February 7, 2011, 19-25; Fouad Ajami, “Demise of Dic-tators,” Newsweek, February 14, 2011, 14-21; Christopher Dickey and Babak Dehghanpisheh, “Among the Believers: The Muslim Brotherhood Stands up in Egypt,” Newsweek, February 14, 2011, 25-27. 86 | ASIAN REVIEW OF WORLD HISTORIES 2:1 (JANUARY 2014) ernment had lost control of significant parts of Libya, including major cities such as Misrata, Benghazi, the important harbors at Ra’s Lanul and Brega. The military confrontation between Gaddafi loyalists and anti-Gaddafi forces gradually escalated into a full-scale civil war. In response, the U.N. Security Council and EU gov-ernments intervened by imposing sanctions on Gaddafi and his family, and suspended Libya’s membership in the U.N. On March 17, 2011, the U.N. Security Council passed a resolution imposing a no-fly zone in Libyan airspace and recommended that “all neces-sary measures” be taken to protect civilians against Gaddafi’s forces. Supported by NATO air forces, the Libyan National Council in Benghazi declared itself the legitimate representative of Libya’s people, gaining recognition from Western and Arab countries. The military confrontation continued between the two parties for a couple of months until the forces of the revolutionaries entered Tripoli in the last week of August 2011, and Gaddafi and his forces left the city, taking their final refuge in Bani Walid, Sirte and other cities. After the liberation of Tripoli, fighting between anti- Gaddafi forces and die-hard Gaddafi loyalists continued for about two months until Colonel Gaddafi was captured on October 20, and killed at the city of Sirte. His death marked the end of his 42-year rule, and three days later the Libyan National Council declared the liberation of the country, and started the process of drafting a new constitution and electing a new government. The outcomes of the Tunisian and Egyptian revolutions also inspired opposition party leaders and political activities in Yemen to continue their struggle against the leadership of President Ali Abdullah Saleh who came to power in 1990. To curtail the political situation in Yemen, President Saleh announced that he would nei-ther run for the forthcoming presidential election in 2013 nor hand power over to his son, Ahmad. These promises did not convince opposition party leaders and political activists who organized anti-government rallies on February 3, 2011 under the slogan of a Yeme-ni “Day of Rage,” styled after similarly named protests in Egypt. A series of change squares was established in various cities across Yemen and the protestors’ primary demand was that the President ABUSHOUK: “TIME AND NEWSWEEK’S COVERAGE OF THE ARAB UPRISINGS IN 2011” | 87 should step down for early presidential and parliamentary elections. The spread of protests all over the country prompted President Saleh to fire his entire cabinet, and also encouraged the Yemen ambassador to the U.N. to resign from his post on the grounds that the ruthless treatment of peaceful demonstrators was indefensible. Several top military commanders defected, and Yemen’s ambassa-dor to Syria quit his post and joined the anti-government move-ment that called for Saleh’s resignation. When the situation be-came very complex and out of control in Yemen, the Gulf Coopera-tion Council countries intervened and attempted to mediate be-tween the President and the protestors by calling for a smooth transfer of power, but their proposal was rejected by the govern-ment. On June 7, 2011 President Saleh was more seriously injured in a rocket attack on Yemen’s presidential compound in Sana’a. He was flown to Saudi Arabia, where he received medical treatment and the administration of the country was entrusted to his deputy. While Saleh was receiving medical treatment in Saudi Arabia, pro-testors formed a transition council on August 18, 2011 in order to strengthen their demand for the transfer of power. Under growing internal and external pressure Saleh signed an agreement, brokered by Yemen’s Gulf Arab neighbours, in the Saudi capital, Riyadh, on November 22, 2011, assenting to hand over power to his deputy, Abdrabuh Mansur Hadi, on the concession that he would be given immunity from prosecution. His deputy was then expected to form a national unity government and call for early presidential elec-tions within 90 days. By signing agreement, Saleh ended his 33 years of authoritarian rule on the concession that he will retain his title and certain privileges until the new presidential elections took place in February 2012.6 6 For further details, see, Farhad Khosrokhavar, The New Arab Revolutions that Shook the World (Boulder, CO: Paradigm Publishers, 2012), 28-42, 43-58. 88 | ASIAN REVIEW OF WORLD HISTORIES 2:1 (JANUARY 2014) Apart from these four Arab countries, anti-government demonstrations and demands for regime change in the Arab world had spread to other Arab countries such as Bahrain, Algeria and Syria. The protestors in Bahrain and Algeria were crushed by gov-ernment forces, while their counterparts in Syria have continued their struggle against al-Assad’s regime. The confrontation be-tween the two opposing parties has developed in a form of civil war because the primary demand of the protestors is that President As-sad should resign to a transitional government that would imple-ment constitutional reforms and make the necessary arrangements for parliamentary and presidential elections. Based on this brief overview of the Arab uprisings, the subse-quent sections discuss the underlying causes of these uprisings and their impacts on the political landscape and U.S. foreign policy in the region. The discussion is based on the special coverage given by Time and Newsweek magazines in 2011. ROOT CAUSES OF ARAB UPRISINGS In their weekly issues, both Time and Newsweek attributed the outbreak of the Arab revolts in 2011 to a series of causes associated with political dictatorship and corruption, unemployment, brutal crackdown on protestors, and the significant role of cell phones and social media (i.e. Facebook, YouTube) that mobilized the sup-port of local masses against dictatorial regimes. The most immedi-ate causes of the Arab revolts are addressed in the subsequent par-agraphs that draw their information from articles, reports, editorial notes and communiques published by Time and Newsweek maga-zines from January to December 2011. POLITICAL DICTATORSHIPS AND CORRUPTION In his essay entitled “Demise of Dictators,” Fouad Ajami wrote, “The Arab dictators had taken their people out of politics, they had ABUSHOUK: “TIME AND NEWSWEEK’S COVERAGE OF THE ARAB UPRISINGS IN 2011” | 89 erected and fortified a large Arab prison, reduced men and women to mere spectators of their own destiny.”7 Their post-independence promise, to restore Arab through modernization and social justice, had gone with the wind. Instead Arab regimes became corrupt and authoritarian regimes, losing them the faith of their people. Time and Newsweek both consider the corruption and oppressive poli-cies of the post-independent Arab regimes amongst the factors that led to the eruption of the revolts in Tunisia, Egypt, Libya, Yemen and Syria, and consider them as ingredients for other future revolts in the region. In her essay “Deepening Divide” Aryn Baker inter-viewed Rami Nakhal, a Syrian cyber-activist, working underground in Lebanon, about the main causes that led him to join the anti-Syrian government campaign. In his reply, Nakhal said, “we want what everyone in the region wants: an end to corruption, the abil-ity to choose and dismiss our leaders, freedom of speech, and free-dom of fear.”8 Nakhal’s answer rightly reflects certain features of authoritarian regimes that can be investigated in the following sec-tions. ZINE EL ABIDINE BEN ALI (R. 1987-2011) Habib Bourguiba (1903-2000) was the founder and the first post-independence President of the Republic of Tunisia from 1957 until 1987. He was succeeded by his Prime Minister, Zine El Abidine Ben Ali, based the official medical report that declared Bourguiba as in-competent to rule the country. This medical report gave him the political legitimacy to stage a bloodless coup against his predeces-sor. Shortly after seizing power in 1987, as Barak Dehghanpisheh and Christopher Dickey wrote: 7 Ajami, “Demise of Dictators,” 14-21. 8 Aryn Baker, “Deepening Divide,” Time, June 13, 2011, 26. 90 | ASIAN REVIEW OF WORLD HISTORIES 2:1 (JANUARY 2014) Ben Ali cast himself as a bulwark against militant Islam, and after 9/11 became one of the Bush administration’s most reliable allies in the Global War on Terror. He also played by the economic rules of the In-ternational Monetary Fund, embraced globalization, and was rewarded with money and praise for his progressive policies. Diplomats in Tunis told NEWSWEEK in 2003 that Tunisia was “a country that works”—a relatively benign regime where criticism of the leader might bring tor-ture and jail, but probably not death. It was “a soft dictatorship,” the diplomats said, “more like Singapore or South Korea in the 1980s than like some other Arab countries today.” But all the while, Ben Ali and his wife, along with several members of her family, were living out the starring roles in a real-life gangster movie.9 The corruption of Ben Ali’s family and his in-laws was described in the following words of a Time journalist, Vivienne Walt: The greed and corruption of the First Family were now intolerable. Protesters lambasted Ben Ali’s second wife Leila Trabelsi, a former hairdresser who accumulated vast wealth as First Lady and bestowed lavish gifts on her numerous relatives; she has 10 siblings. About half of Tunisia’s businesses -including a bank, hotels, a property-development firm and the two biggest newspaper companies - are in the names of the extended family. The distributorships of Porsche, Volkswagen, Kia and Alfa Romeo cars all belonged to Ben Ali’s son-in-law, Sakher El Materi. His lavish lifestyle was the subject of a 2009 diplomatic cable, acquired by WikiLeaks, in which the then U.S. ambassador, Robert Godec, warned State Department officials that the ruling family’s ex-cesses could lead to the regime’s collapse. Godec described a sumptu-ous dinner at El Materi’s home, where the young tycoon pressed the ambassador to help him acquire the McDonald’s franchise for Tunisia and where ice cream and frozen yogurt had been flown in from St.-Tropez, France, on his host’s private plane. The household pets includ-ed a caged tiger named Pasha, which reminded Godec of Uday Hus-sein’s caged lion in Baghdad.” 10 9 Barak Dehghanpisheh and Christopher Dickey, “A Dictator Dispatched,” Newsweek, January 31, 2011, 29. 10 Vivienne Walt, “Tunisia’s Nervous Neighbors Watch the Jasmine Revolution,” Time, January 31, 2011, 20. ABUSHOUK: “TIME AND NEWSWEEK’S COVERAGE OF THE ARAB UPRISINGS IN 2011” | 91 The ruling family’s corruption and extravagance were heavily criti-cized by Fouad Ajami who stated that Ben Ali and his in-laws had no “patriotism and love” of their home country. His “ambitious wife, the hairdresser who had come out of nowhere to the pinnacle of power, made a run for it. It had been quite a racket for them, and it was now time to quit the land they had plundered and en-raged.” 11 The public’s perception of Ben Ali’s rampant corruption augmented the anger of protestors, and led them, along with other factors, to revolt against the leadership of Ben and his family, and called them to depart. HOSNI MUBARAK (R. 1981-2011) Hosni Mubarak is also a military officer who took over the leader-ship of the most influential country in the Arab Word, Egypt. He came into power after the assassination of Muhammad Anwar al-Sadat in 1981, and ruled the country with an iron fist for three dec-ades. He suppressed the activities of the Muslim brotherhood and jailed a large number of them and other political opponents. Mu-barak maintained most of his predecessor’s foreign and domestic policies, including the Camp David Treaty and Sadat’s close ties to the United States of America. He also advised King Hussain of Jor-dan and Yasser Arafat of PLO to strike a deal with the state of Isra-el, recognizing Israel’s right to exist in the region. In the last five years of his rule, Mubarak revealed grew increasingly disinterested in sharing power with other political parties, as he paved the way for his son Gamal to succeed him, following the Syrian model. In this respect, a Newsweek writer, Christopher Dickey, wrote: 11 Ajami, “Demise of Dictators,” 20. 92 | ASIAN REVIEW OF WORLD HISTORIES 2:1 (JANUARY 2014) [His wife], Suzanne guided the fortunes of her children and grandchil-dren, looking to establish a political dynasty that might endure for generations. The older son, Alaa, is a businessman who prefers soccer to the game of politics—a fact that has brought him occasional surges of popularity over the years as a big-name, big-mouthed fan of Egypt’s national team. The younger son, the handsome, aloof Gamal, was for years the apparently anointed but undeclared heir to the presidential palace. When writing about his rise, British tabloids never failed to mention the pharaohs’ ancient dynasties. Gamal himself, half-joking with friends and acquaintances even as he ritualistically denied presi-dential aspirations, preferred to speak of the Kennedys, the Bushes, and the Clintons.12 Dickey also described the situation before Hosni Mubarak’s depar-ture in the following words: The night before he finally stepped down as Egypt’s President, the pro-testers in Tahrir Square heard Hosni Mubarak delivers his final address as their head of state. “A speech from a father to his sons and daugh-ters,” he called it, and like many of his orations in the past, it was filled with lies, although he may have believed some of these himself. He would stay as President until September, he promised, because the country needed him for a transition to democracy. This, after three decades of autocracy. The hundreds of thousands gathered in the square wanted to hear him say only one word: “Goodbye.” Amid their screams of fury, one woman could be heard shouting into a phone, “People are sick of the soap opera!” 13 This passage illustrates that the protestors were not ready to accept any concession from the government, except the departure of Hos-ni Mubarak, his family and old guards. Stories of Mubarak’s cor-ruption were widely circulated and many believed that Mubarak and his family had a fortune of $40 to $70 billion; and that he had rigged the last parliamentary elections of 2010. All reports of Newsweek and Time consider this terrific form of corruption 12 Christopher Dickey, “The Tragedy of Mubarak,” Newsweek, February 21, 2011, 28. 13 Ibid., 26. ABUSHOUK: “TIME AND NEWSWEEK’S COVERAGE OF THE ARAB UPRISINGS IN 2011” | 93 among one of the fundamental causes that led to the end of his thirty-year authoritarian regime. MUAMMAR AL -GADDAFI (R. 1969-2011) Colonel Muammar al-Gaddafi staged a military coup in 1969 against the first and last king of independent Libya, Idris al-Sanussi (1951-1969), and served the country as head of state with absolute power. Through his Arab nationalist rhetoric and socialist-style policies, Gaddafi gained the support of the local masses during the early days of his rule, and established himself as the Chairman of the Revolutionary Command Council of Libya, claiming subse-quently to be merely a symbolic figurehead. He involved himself in a series of terrorist activities outside the country, and ruled his own people in a very brutal way, to the extent that the Libyan popula-tion had lost faith in his tyrannical leadership, and eventually re-volted against him in 2011. On the corruption of his regime and in-volvement of his family in the country’s politics, a Newsweek jour-nalist wrote: Like his other six siblings, Saif, 39, was a puppet of his father, who ruled his family like he ruled Libya’s tribes: playing one against the other. The Gaddafi children, for their part, carved up the country’s wealth. Muhammad, the oldest son, controlled telecommunications. Hardliner Mutassim served as -national-security adviser until he lost his father’s favor and was shipped to Egypt. Saadi, perhaps the best-known brother, captained the national soccer team. Aptly named Hannibal [he] earned infamy for beating his model wife to a pulp in a suite at the posh Claridge’s hotel in London. His sister, Aisha, served as a lawyer not only for Hannibal but also for Saddam Hussein, a family friend. The two youngest brothers, Khamis and Saif al-Arab, got lost in the shadows of their older siblings. Of them all, it was Saif al-Islam, the 94 | ASIAN REVIEW OF WORLD HISTORIES 2:1 (JANUARY 2014) self-styled artist, who enjoyed life outside Libya the most. His father used him as a slick ambassador to the West. And Saif loved this role, which allowed him to travel abroad and hang out with his Israeli girl-friend.14 This passage leads us to agree with Fouad Ajami who argued that “The fortunes of the rulers, an Arab businessman once said to me, are the real weapons of mass destruction in the region. The Houses of Assad, of Mubarak, of Gaddafi and of Ben Ali (and of Saddam Hussein before his fall) are rich beyond measure. The line between the wealth of the rulers and the treasure of the realm has been erased.”15 In line with Ajami’s allegations, both Time and Newsweek reported the tyranny of the Arab rulers and corruption of their families to be amongst some of the most important reasons for the uprisings that terminated the leadership of four dictators in the re-gion. ALI ABDULLAH SALEH (R. 1990-2011) Field Marshal Ali Abdullah Saleh was the first President of the Re-public of Yemen after the unification between North Yemen and South Yemen in 1990. The unification of the two Yemens remained as one of his major achievements, enhancing his political image in the eyes of several tribes of Yemen. However, his grip on power slackened when the Arab uprisings inflamed Yemen. Some close associates under his thumb defected, powerful tribes turned against him and the U.S. administration - a longtime friend once happy to back Saleh as a proxy in the war against al-Qaeda - gave him the coldshoulder. His dictatorship, favouritism of his tribes-men and corruption of his regime were singled out among other causes that fueled the revolt in Yemen. Under international and re-gional pressure, Saleh signed an agreement on November 23, 2011, accepting the transfer of power to his Vice President on the con- 14 Eliza Griswold, “In the Name of the Father,” Newsweek, August 28, 2011. 15 Ajami, “Demise of Dictators,” 20. ABUSHOUK: “TIME AND NEWSWEEK’S COVERAGE OF THE ARAB UPRISINGS IN 2011” | 95 cession that he would reserve his title and certain privileges until the new presidential elections take place in February 2012. The whole episode illustrates that the vast majority of the Yemeni pop-ulation had discredited the leadership of Saleh, accusing him of be-ing corrupt, nepotistic and interested in turning the country into a family business, as he groomed his son, Ahmad, to succeed him. YOUTHQUAKE IN THE ARAB WORLD Another contributing factor that led to the Arab Spring is increas-ing number of young people under the age of thirty years.16 As Fouad Ajami wrote, the Arab Human Development Report of 2009 provided a telling portrait of the world of 360 million Arabs. They were overwhelmingly young and about 60% of them lived in urban centres where economic growth and job markets did not meet their expectations, no fewer than 65 million Arabs were living be-low the poverty line of $2 a day and fifty-one million new jobs have to be created by 2020 to accommodate university graduates and others. The case of Mohamed Bouazizi is one of the alarming in-stances of the uprising that shows the frustration of youths in the Arab world, who cannot find jobs to satisfy their ambitions.17 The map below illustrates the distribution of youths who are under 25 years old and the gross domestic product (GDP) per person in each Arab country. 16 Christopher Dickey, “Bye-Bye to Tunisia’s Ben Ali,” Newsweek, January 24, 2011, 7. 17 Ajami, “Demise of Dictators,” 16. 96 | ASIAN REVIEW OF WORLD HISTORIES 2:1 (JANUARY 2014) Figure 1: Distribution of Youths and GDP/person in Arab Countries Source: Cultural Wizard, http://rw-3.com/tag/middle-east/ Based on the above statistics, U.S. Secretary of State, Hillary Clin-ton, alerted the leaders of Arab countries to listen to the demands of their young people since the old foundations of the Arab world “are sinking into the sand.”18 The response to their demands should address issues such as democratization of the political system, cre-ation of new jobs, and suppression of government corruption. CELL PHONES AND SOCIAL MEDIA (FACEBOOK, TWITTER, AND YOUTUBE) Both Time and Newsweek paid special attention to the social media, such as Facebook and Twitter which provided a space for activists to talk, organize rallies against their authoritarian regimes, send out information and challenge any attempt to shut down their public political forums and communications. Their dialogues and discussions on the channels of social media addressed the corrup-tion of the ruling elites, and mobilize the masses, encouraging 18 Walt, “Tunisia’s Nervous Neighbors Watch the Jasmine Revolution,” 18. ABUSHOUK: “TIME AND NEWSWEEK’S COVERAGE OF THE ARAB UPRISINGS IN 2011” | 97 them to organize against their tyrannical leaders and their clients. Social media proved served as a platform for the exchange of steady stream of anonymous text messages, Twitter and Facebook updates. Mobile-phone videos posted documented governments’ reactions against peaceful protestors, including the police beatings and shootings of peaceful protestors who overcame their fear and chal-lenged their autocratic leaders.19 In Egypt, for instance, Newsweek journalists, Babak Dehghanpisheh, Christopher Dickey and Mike Giglio described the role of the social media, saying: One by one, the lines of communication that connected Egypt to the 21st century shut down. Twitter, Facebook, and eventually all Internet access were cut off; text messaging became impossible, and then mil-lions of mobile phones went silent across the country. But the protests and riots continued, as they had for most of the week, with thousands of young Egyptians trying to take down the regime of octogenarian President Hosni Mubarak. They set last Friday for their “day of rage,” drawing in supporters from all over the country, including the out-lawed but powerfully organized Muslim Brotherhood. In the hours leading up to the demonstrations, the government did everything it could to cut them off from each other, and from the rest of the world. 20 Christopher Dickey and Babak Dehghanpisheh also emphasized that the actual instigators of the revolt in Egypt were “a band of young techies who used their mass-communication skills to mobi-lize thousands of people from almost every stratum of Egyptian so-ciety in an uprising against Mubarak’s reign—with the notable ex- 19 Ibid., 20; Christopher Dickey, “Morning the Middle East,” Newsweek, Special Issue, 2012, 9-13. 20 Dehghanpisheh, Dickey and Giglio, “Rage Against the Regime,” 22. 98 | ASIAN REVIEW OF WORLD HISTORIES 2:1 (JANUARY 2014) ception of the Brotherhood, which declined to join the first mas-sive but peaceful demonstrations” 21 on January 25, 2011. In Syria, young activists took advantage of the tools Assad had made available to campaign against him: through the use of cam-eras on their mobile phones protestors were able to record the abuses of the security forces; and through the internet they were able to send these recordings to news agencies around the world. “You can’t quash an uprising if millions of people are acting like their own independent news stations.” 22 All these examples high-light the concern of Time and Newsweek about the role of the so-cial media that facilitated the dissemination of information and the organization of the Arab youths to stand against their corrupt leaders. They also emphasized that economic reforms and political changes were among the primary demands of the protesters who openly criticized the leadership and management of their countries. 23 WHAT WAS THE JUSTIFICATION OF RULING AUTHORITIES? Based on the weekly coverage of Time and Newsweek of the Arab uprisings, the dictators of the Arab countries who had been chal-lenged by protestors tried to attract the support of their Western patrons, claiming that the demonstrations were orchestrated by the loyalists of al-Qaeda in the region. They assumed that this claim would readily arm them with the support of their Western patrons, or at least encourage them to turn a blind eye to any form of brutality that might be used against peaceful protestors. All the challenged dictators presented themselves as key players in the re-gion, leading the fight against extremism and al-Qaeda’s support-ers. A role that was crucial for them to play after 9/11, when they 21 Dickey and Dehghanpisheh, “Among the Believers: The Muslim Brotherhood Stands up in Egypt,” 26. 22 Baker, “Deepening Divide,” 26. 23 For further details on the role of social media in the Arab revolution, see Jean-Pierre Filiu, The Arab Revolution: Ten Lessons from the Democratic Uprising (London: Hurst & Com-pany, 2011), 43-56. ABUSHOUK: “TIME AND NEWSWEEK’S COVERAGE OF THE ARAB UPRISINGS IN 2011” | 99 received substantial support from Western countries and the American administration, in particular, to confront the threat posed by “Muslim extremists” in their home countries. When they were confronted by peaceful protests calling for political and eco-nomic reforms, they raised the issue of Islamic extremism to divert the attention of Western media from their internal problems and corruption. In Syria, for instance, when the army tanks stormed the southern town of Daraa, a military spokesman claimed that the as-sault had targeted “extremist terrorist groups.” President al-Assad himself asserted that the jihadists were behind all the demonstra-tions taking place in Syria.24 In this sense, Arab leaders assumed that if the protestors were branded as the ideological shadows of al-Qaeda, the West would not intervene.25 WHAT WAS THE REACTION OF THE AMERICAN ADMINISTRATION? President Obama was reluctant to intervene in the bloody conflict that took place between the protestors and the die-hard loyalists of the Arab dictators who ruled their people with iron fists and loaded guns. His response to the situation in Libya was that “the best revo-lutions” should be “completely organic,” in the sense, that the American administration should not intervene in the internal af-fairs of the Arab countries. Journalists of Time and Newsweek mag-azines criticized his attitude, particularly Niall Ferguson who listed a number of successful revolutions that were not completely organ-ic, including the American Revolution, but their success was partly based on foreign assistance. Therefore, he suggested that President Obama should be given “a history lesson before thousands of Liby-ans share their fate. It will be tragic indeed if America concludes 24 Nibras Kazimi, “Handing Jihadis Cause,” Newsweek, May 9, 2001, 9-10, especially 9. 25 Ibid., 10. 100 | ASIAN REVIEW OF WORLD HISTORIES 2:1 (JANUARY 2014) from the experiences of overthrowing murderous tyrannies in Af-ghanistan and Iraq that the correct policy is to turn a blind eye to murder in Libya.”26 In line with these remarks, Rosemary Righter wrote: Barack Obama’s public reluctance to point Hosni Mubarak toward the exit door pushed open by millions of disgusted Egyptians was defensi-ble, if not exactly admirable: the U.S. did not wish to be seen peremp-torily dumping so longstanding an ally, although the White House should have seen much earlier that there was no alternative. The pres-ident’s dithering over Libya has been neither defensible nor admirable. His electoral pitch made much of America’s “moral obligation” to in-tervene to prevent atrocities against civilians. Atrocity is a strong word: it does not describe the familiar travails of the subjects of unjust and corrupt rulers. But atrocities are happening in Libya. Obama’s declared principles in foreign affairs face a test he seems loath to recognize. From the first lethal volleys against peaceful demonstrators, it was clear that Muammar Gaddafi would wage war to the death on his own people, a war that his “reformist” son Saif swore to fight “to the last man, the last woman, and the last bullet.” Obama duly declared that Gaddafi had “lost legitimacy” and must go; last week he assured the Libyan people that “we will stand with them in the face of unwarranted violence and the continuing suppression of democratic ideals.” Yet all week, as Gaddafi reduced Zawiyah to rubble and pounded Ras Lanuf, the White House resumed its leisurely consideration of “the full spec-trum of possible responses.” NATO kicked its heels; the EU, as usual, met to decide on what it could not decide. Libyans bled. 27 In his submission, Righter also raised a very provocative question: “Does Obama believe that the era of U.S. leadership should be seen to be over?” She answered the question by saying that “the issue is not the narrow one of the legality or military effectiveness of a no-fly zone. It is what to do about an outlaw.” Since Gaddafi had breached the U.N.’s demand to end violence against his people and meet their “legitimate demands.” The only resort left, from Right- 26 Niall Ferguson, “How to Get Gaddafi: Mr. President, Don’t Send Guns to the Libyans. Send them a Piece of Paper,” Newsweek, March 21, 2011, 13. 27 Rosemary Righter, “Right to Protect: International Law Offers a Robust Doctrine for Intervention in Libya,” Newsweek, March 21, 2011, 14-15, especially 14. ABUSHOUK: “TIME AND NEWSWEEK’S COVERAGE OF THE ARAB UPRISINGS IN 2011” | 101 er’s perspective was to force him to stop fighting and listen to the legitimate demands of the Libyan protestors. On the other hand, a Newsweek journalist, Fareed Zakaria, addressed Obama’s reluc-tance from a different perspective, arguing the President’s concern was that he did not want to repeat the mistakes of his predecessor, George W. Bush, when invading Iraq in 2003. Therefore, his pre-cautions included the establishment of a local opposition move-ment that would be able to wage war against Gaddafi and his loyal-ists; the securing of regional support so that they would not de-nounce foreign intervention in Libya as another example of West-ern imperialism in Muslim lands; and finally, gaining the support of the U.N. would legitimize intervention in Libya against Gaddafi’s regime.28 In the case of Syria, journalists of Time and Newsweek pro-moted a different approach that supported the disinterest of West-ern countries in backing a radical political change of Assad’s re-gime on the grounds that his successors would be the “radical Is-lamists.” In this sense both magazines indirectly advised Western leaders to follow Tel Aviv’s wisdom: “Better the devil you know than the devil you don’t,” 29 meaning, priority should be given to the security of the State of Israel rather than given to the protes-tors’ expectations and demands that call for genuine political and economic reforms in Syria. Their stance on Syria reflects the dou-ble standard in the rhetoric of Time and Newsweek, if we compare their coverage of Libyan events with that of Syria. 28 Fareed Zakaria, “How the Lessons of Iraq paid off in Libya,” Time, September 5, 2011, 18-19, especially 18. 29 Richard Haass, “The Revolution Stops Here,” Time, April 18, 2011, 16. 102 | ASIAN REVIEW OF WORLD HISTORIES 2:1 (JANUARY 2014) PREMATURE CELEBRATIONS Journalists of Time and Newsweek agreed that the Arab uprisings which took place in 2011 did not have charismatic leaders who would put the demands of the ambitious young protestors on the right track and secure their participation in the process of decision making.30 This situation led them to raise concerns about the pos-sible attainment of democratic outcomes, since Islamist opposition parties may take over the leadership of the Arab countries. From their perspectives, this political shift would create three challenges. Firstly, the security of the State of Israel would be vulnerable since the leaders of all Muslim parties have genuine support for the Palestinian cause. Secondly, the Christian monitories in the Arab countries would face some difficulties in dealing with the Islamists, who are in favour of reinstating new forms of Islamic states that may un-dermine their citizenship and human rights. Habib C. Malik de-scribed the fear of Christian Arabs in the following words: At the heart of Christian Arab apprehensions lies a shaken trust in both the staying power and shelf life of Sunni-Muslim moderation. Christians fear that once the dust of the Arab Spring settles, Islamist radicalism could sweep away in its path all the well-meaning, liberal-minded, pro-democracy leaders within the opposition movements. This fear characterizes the present decisive juncture for all Arabs; now is indeed a defining historic moment for moderation in the vast world of Sunni Islam. Will the Arab Sunni moderates, the silent majorities in their respective societies, prove capable of preventing a slide toward intolerance and violence brought on by the fanatical few in their midst? This is the ultimate question of the Arab Spring, but already the Copts appear to have their grim answer: Egypt’s moderate and plural-ist-minded revolutionaries are proving impotent in the face of deter-mined religious extremists and hostile armed forces of the indifferent authorities. 31 30 Bobby Ghosh, “Revolution Delayed,” Time, February 21, 16-22, especially 16. 31 Habib Malik, “Intolerant Arab Spring,” Newsweek, October 24, 2011, 6. ABUSHOUK: “TIME AND NEWSWEEK’S COVERAGE OF THE ARAB UPRISINGS IN 2011” | 103 Thirdly, American and Western interests might be disturbed if the Islamists took over the leadership of the oil rich countries in the near future. These three scenarios lead policy-makers and po-litical activists in the West and the Arab world to be more doubtful of the outcomes of the Arab uprisings, as they neither satisfy the protestors’ expectations nor Western countries’ prospects. There-fore, they consider any forms of celebration as premature since the new governments in Tunisia, Egypt, Libya and Yemen are or will be controlled by Islamists who do not have clear strategic plans to overcome the series of challenges which they are facing. This view is supported by Dr. Schueftan, Director of the National Security Studies Center, the University of Haifa, who argues that “The Arab Spring has led to deep systemic changes -- mostly negative -- in the rules of the game throughout the Middle East. If this trend persists, it will harm the interests of not only the United States and Israel, but also Arab countries, particularly those that have been longstanding U.S. allies.” 32 CONCLUSION This content analysis of Time and Newsweek magazines’ coverage of the Arab uprisings in 2011 shows that the two magazines ad-dressed the underlying causes and anticipated outcomes of the up-risings from a journalistic perspective that was informative but less academic in nature. Newsweek was more comprehensive in its cov-erage and concern by inviting certain scholars in the field, such as Fouad Ajami, to offer their academic insights. However, the two magazines were pessimistic about the outcomes of the uprisings on 32 Dan Schueftan and Michael Singh, “Arab Spring, Arab Storm: Implications for Israel,” Policy Watch no. 1862, Special Forum Report, October 24, 2011, accessed February 24, 2011, http://www.washingtoninstitute.org/templateC05.php?CID=3413. 104 | ASIAN REVIEW OF WORLD HISTORIES 2:1 (JANUARY 2014) the grounds that elections may bring Islamist parties to power, tak-ing into account that these parties are not in favour of American foreign policy in the region. The victory of Islamist parties in Tuni-sia and Egypt had alarmed the two magazines causing them to be less enthusiastic about the interference of Western countries and the U.N. Security Council in supporting the massacre of peaceful protestors in Syria. Their salient rationale is that the success of Is-lamists in Syria may affect the security of the state of Israel and generate a real challenge to U.S. foreign policy in the region, bear-ing in mind that all the Islamist parties have a soft spot towards the Palestinian cause. These anticipated scenarios have led them to suggest that there be a revision of U.S. and Western foreign policy in the region, so as to meet the mandates of the new regimes and their supporters. They also raise a special concern about the future of the long patron-client relationship between the American Ad-ministration and its key clients in the region, such as Saudi Arabia, United Arab Emirates, Qatar and Jordan. View publication stats | RESEARCH GATE |
Journal of Physics: Conference SeriesPAPER • OPEN ACCESSAnalysis of Eucalyptus Supply Chain MethodsBased on Influencing Factors: A ReviewTo cite this article: N Sembiring et al 2023 J. Phys.: Conf. Ser. 2421 012009 View the article online for updates and enhancements.You may also likeSimulation Model of Replanting Eucalyptus: a ReviewNurhayati Sembiring, Humala LodewijkNapitupulu and Meilita Tryana Sembiring-Trends of genetic parameters and standvolume productivity of selected clones ofEucalyptus pellita observed in clonal trialsin Wonogiri, Central JavaD Kartikaningtyas, A Nirsatmanto, SSunarti et al.-Fulfilling Eucalyptus raw materials for pulpand paper production plantsN Sembiring, H L Napitupulu, M TSembiring et al.-This content was downloaded from IP address 168.151.167.23 on 26/01/2023 at 00:56Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distributionof this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.Published under licence by IOP Publishing Ltd2nd TALENTA-International Conference on Science and TechnologyJournal of Physics: Conference Series2421 (2023) 012009IOP Publishingdoi:10.1088/1742-6596/2421/1/0120091 Analysis of Eucalyptus Supply Chain Methods Based on Influencing Factors: A Review N Sembiring*, H L Napitupulu, N N Azmi and R Radot Department of Industrial Engineering, Faculty of Engineering, Universitas Sumatera Utara, Medan, Indonesia *nurhayatipandia68@usu.ac.id, humala_n@yahoo.com, nurulnoviaazmi@gmail.com, rhyvalradot@gmail.com Abstract. The success of a supply chain is an important key in achieving the goals set by producers. The success of a supply chain has multiple objectives that are key to consumer satisfaction. These goals are high quality, low prices, on time delivery, and good customer service. This goal can be achieved if the management can design a good supply chain. In this study, an analysis will be carried out from the results of previous research, regarding the efforts made to increase supply chain productivity for Eucalyptus plants. Every scientific article goes through a rigorous selection process to get articles that are in accordance with the objectives of this research. The results show that the approach using simulation methods is the best way to analyze and evaluate a supply chain. Each simulation approach can provide actual information required for improvement of the supply chain. Improvements are generally carried out on an ongoing basis for each specified period. 1. Introduction Eucalyptus is a very common plant grown in plantations around the world. Currently, the area of plantations that plant Eucalyptus is around 20 million hectares. This plant is more likely to be grown in the countries of South America, Southeast Asia, Africa, Southwestern Europe and Osenia. Eucalyptus plants are very popular with farmers because of their ability to adapt to various types of world climates. In addition, this plant also has a high economic value from its very high quality wood [1]. Eucalyptus has a high level of demand in the global market. Currently the Eucalyptus plant is not only a raw material for making products, but also a raw material for making energy. those derived from high density wood are considered a renewable resource worldwide [2]. Eucalyptus plants have the ability to be used as ethanol fuel. This has led to the expansion of companies producing ethanol using Eucalyptus as raw material. Seeing this condition, we can see how big the global market demand is for the existence of Eucalyptus plants [3]. Supply chain management is an important thing to be applied in producers and managers of Eucalyptus plants in meeting the high level of consumer demand. Supply chain management is carried out to exchange information in the process of material flow from raw materials to finished products. a Supply chain management handled all the flow of information. Supply chain management has a responsibility for the efficiency and effectiveness of the flow of material in a supply chain [4]. This research was conducted to find efforts that can be made to improve the supply chain of Eucalyptus plants. These efforts are obtained from previous research whose quality has been guaranteed. 2nd TALENTA-International Conference on Science and TechnologyJournal of Physics: Conference Series2421 (2023) 012009IOP Publishingdoi:10.1088/1742-6596/2421/1/0120092 These efforts can be in the form of research actions carried out on Eucalyptus plants or other types of plants that have the opportunity to be applied to efforts to increase the supply chain of Eucalyptus plants. 2. Literature Review At the beginning of the concept of sustainable improvement for the forestry sector, the focus of this understanding was still very narrow. The focus of these improvements is only on sustainable wood production to meet consumer demand. The concept of continuous improvement in the forestry sector has developed quite significantly. Currently, continuous improvement considers three dimensions, namely economic, social, and environmental. All of the three dimensions are of concern in the sustainable improvement of the forestry sector, especially in the supply chain [5]. Supply chain management is an important element in increasing efficiency and the flow of materials from raw materials to finished goods until distributed to consumers. Supply chain management is responsible for the flow of information, money and materials from the initial process to the hands of consumers. In addition, supply chain management also pays attention to aspects of customer satisfaction with service, timeliness, and certainty of the product ordered. Therefore, supply chain management is responsible for the flow of raw materials in a society and the exchange of materials and energy with the surrounding environment. Supply chain management is responsible in three dimensions, namely social, economic, and environmental [4]. The supply chain value in the forestry sector is often linked to the production of timber owned. Ideally, a supply chain manager must have the ability to do good supply chain planning and management. This planning is related to operations, maintenance, replanting, harvesting, manufacturing activities, residue utilization, and distribution. Each of these activities must be well planned so that good field execution can be carried out [6]. The concept of sustainable forestry productivity has developed after the existence of sustainable development by the government. So it is often concluded that the practice of continuous improvement is an integration of economic, social, and environmental performance [4]. The research articles that have been selected from the results of rigorous selection will be analyzed. Each selected research article is adjusted to the topic of discussion of this research. The selected previous research articles able to be seen in table 1. Table 1. Previous research articles on Optimization Efforts in Forestry and Plantation Supply Chains (Continue) Publisher Technique Implications Volume and Year Elsevier matrix iterative methods The results showed that an integrated approach to industries with large-scale cases could provide an increase in profits of up to 13% [6]. 259 (2016) Elsevier Analytical assessment models The results of the study provide a representation of the application of a profile to both supply chain management practices that identify employee performance from a sustainable development perspective [4] 76 (2014) Springer Simulation of Ecological Compatibility of Regional Development Scenario analysis allows for a future trajectory and depends on external and internal factors. The external factors are in the form of government subsidies and harvest prices, while the internal factors are innovation and cooperation [7]. 24 (2009) 2nd TALENTA-International Conference on Science and TechnologyJournal of Physics: Conference Series2421 (2023) 012009IOP Publishingdoi:10.1088/1742-6596/2421/1/0120093 Table 1. Previous research articles on Optimization Efforts in Forestry and Plantation Supply Chains (Continue) Publisher Technique Implications Volume and Year Elsevier Review of research articles for the last 10 years with a period of 2000 - 2010 The development of research on supply chain management has evolved each period with a wider variety of approaches and variables [8]. 140 (2012) Crossref Dynamics and Agent-Based Modelling simulation The hybrid simulation model used in the study shows the validation results of the comparison of the simulated NR volume value with the actual conditions. The average of the mean absolute percentage error is 8.6% [9]. 16 (2017) Elsevier composite indicator method The results showed that the use of the suggested tools helped decision makers in determining policies related to the sustainability of rubber supply in North Sumatra province [10]. 241 (2019) Elsevier Evaluation of kinetic parameters using the Coast-Redfren method and multiple linear regression The results showed that there was a chromatogram peak with a complex mixture which was a natural mixture of highly oxygenated content. The complex mixture has non-oxygenated properties such as kenotic, phenolic, aromatic hydrocarbons, amide groups, and others [11]. 8 (2019) Elsevier Coats-Redfern method and multiple linear regression The conclution of the study and review of previous studies (84.6%), found that this study considers two dimensions of sustainability. The two dimensions are economy (31.9%), environment (13.8%), or a combination of the two dimensions (38.8%) [2]. 105 (2019) Elsevier Roundtable on Sustainable Palm Oil (RSPO) Predictions were made that the average AGC value of oil palm trees in the lowlands of Kalimantan had a residual value of 10% to 20% above the average AGC value in the absence of elimination. The results found also showed a positive relationship between plant uniformity and the AGC value. [12]. 248 (2020) Elsevier Agricultural Production Systems simulator (APSIM) next generation. The calibration data and evaluation results of the model give very good results for aboveground biomass (Nash Sutcliffe Efficiency = 0.84 and 0.96 respectively). The results show that the application of the model to a particular site or group will benefit from 469 (2020) 2nd TALENTA-International Conference on Science and TechnologyJournal of Physics: Conference Series2421 (2023) 012009IOP Publishingdoi:10.1088/1742-6596/2421/1/0120094 Table 1. Previous research articles on Optimization Efforts in Forestry and Plantation Supply Chains (Continue) Publisher Technique Implications Volume and Year parameterizing the results to local conditions [1]. Elsevier Shapiro-Wilk (p <.05) and ANOVA Ash and carbon concentrations have lower values for stem fraction at higher planting densities. However, the higher warming rate is not symbolic different from each analysis. The condition of biomass fuel throughout the forest is not affected by management practices [2]. 121 (2018) Elsevier Rosin Adduct-coated Controlled Release Urea Fertilizer (RA-CRUF) The results of the analysis using an optical microscope using the RA-CRUF approach on the cross section resulted in a high homogeneity. The cumulative release values of urea were 41.76% and 47.23% for Rosin Adduct-coated Controlled Release Urea Fertilizer (RA-CRUF) compared to 100% and 82.6% for regular urea. This condition was observed in clay textured soil [13]. 133 (2019) Elsevier Sequential steps to optimize EFB value chain The results showed a very significant economic benefit using the EFB method. EFB utilization in all three cases showed an increase in the profit margin of 47% [14]. 119 (2020) Elsevier Mixed integer linear programming Mixed integer linear models can provide a solution to traditional planning when tested with company data sets [15]. 141 (2017) Elsevier Paper Review Using data in cocoa farming practices can increase productivity [16]. 29 (2013) Elsevier Linear optimization model The results show a regular tendency pattern regarding use for large-scale industrialized transform. The condition of biomass supply will increase by 2030 [3]. 173 (2016) Elsevier Geographical Information Systems (GIS) The results showed an increase in the ratio of the increase in energy balance by 2%, so that high efficiency was created [17]. 157 (2017) Springer Closed-loop supply chain (CLSC) networks Each study result was categorized into numerical and real cases. Green CLSC is focused on industrial cases in the electronics field [18]. 925 (2020) 2nd TALENTA-International Conference on Science and TechnologyJournal of Physics: Conference Series2421 (2023) 012009IOP Publishingdoi:10.1088/1742-6596/2421/1/0120095 3. Methods This research was conducted to (i) find an approach to optimize eucalyptus supply chain, (ii) find factors that affect the productivity of the eucalyptus plant supply chain, (iii) identify improvements that can be carried out on an ongoing basis to achieve the stated goals. set. Each previous literature study has a topic with a unique and different approach. Each difference from the approach given will add to the perspective of the researcher in considering the steps that must be achieved in the eucalyptus supply chain. The researchers conducted an analysis of various types of sources for previous scientific papers on supply chain optimization. Each discussion will be read and examined in detail to obtain the necessary information according to the research hat that has been set. This research is focused on the factors that can be implemented in order to optimize supply chain management of eucalyptus plants. The factor in question may come from research on eucalyptus plants or other types of plants that are deemed feasible to be implemented properly. This research begins by searching for previous scientific articles in accordance with predetermined research topics. Every scientific article is taken from international journal publishers who are competent in their fields. Examples of publishers that have been highlighted by researchers are Elsevier and Springer. After the articles are collected, analysis and sorting will be carried out to get the best articles according to the research topic. After obtaining articles of the expected quality and required quantity, they will be analyzed again. Re-analysis is carried out to determine what implementations and improvements are offered and can be applied for optimization of supply chain management of Eucalyptus plants. The steps of this research can be seen in Figure 1. Stage 1Determination of research objectivesStage 2Search for previous research literatureStage 3Sorting literatures according to the objectives and subject of research Stage 4Research of selected articlesStage 5Evaluate the improvements offered in the articleStage 6The conclusion from the analysis and evaluation results obtained Figure. 1. Research implementation stages 2nd TALENTA-International Conference on Science and TechnologyJournal of Physics: Conference Series2421 (2023) 012009IOP Publishingdoi:10.1088/1742-6596/2421/1/0120096 4. Research Results Efforts made to optimize supply chain management of eucalyptus plants can be carried out with improvements from various dimensions. Repairs made are not temporary. The proposed improvements must be sustainable and also follow the development of technological advances and market demand. Through this research, it can be seen that the classification of improvements offered can be divided into three dimensions, namely: 4.1. Environmental Dimension Optimization of a supply chain management of eucalyptus plants must pay attention to environmental conditions. The environment is one of the factors that must be maintained because it is dynamic and unpredictable. The improvement obtained from the results of this study is that the determination of the excellent rate and region of the eucalyptus plant refine industry is indispensable for designing an environmentally friendly industry followed by an emphasis on production costs. Optimizing a supply chain management system regarding the conversion of forestry land to eucalyptus plantations needs to consider environmental factors and conditions. The level of emission from an environment greatly affects the performance of eucalyptus plants in producing the expected wood quality. If land conversion can be optimized, the value of gas emissions will decrease, which will improve the quality of eucalyptus wood. Quality improvement affects the income of a company [3]. 4.2. Social Dimension Optimization of supply chain management for eucalyptus plants must pay attention to the social dimension. The social dimension is one of the factors that involve the capacity of a company. The social dimension of a plantation such as a eucalyptus plant is usually related to work motivation and also cooperation between farmers. Optimization in the social dimension must be carried out intensively and sustainably to obtain the expected productivity. The high productivity of individual farmers on eucalyptus plantations will support the efficiency and effectiveness of a supply chain. Many researchers use agent-based modelling to get the interaction between social dimensions and the productivity of a company. The designed model will usually help decision makers in determining policies related to social or other dimensions [7]. 4.3. Economic Dimension The economic dimension is the dimension most often discussed from the other two dimensions. The economic dimension usually aims to get the maximum benefit with the smallest amount of capital. This dimension relates to savings and cost cutting for any activities that are deemed unnecessary and have no added value. Optimization of the supply chain of eucalyptus plants is of course aimed at getting big profits. However, the application of this concept must be emphasized further. Achieving high economic dimension values must be done without neglecting the social and environmental dimensions. The integration carried out from these three dimensions aims to optimize sustainable supply chain management. Often the social or environmental dimensions are considered waste and are often ignored. However, upon review, ignoring the other two dimensions will only make the proposed improvements only temporary [6]. One of the studies shows the results of modelling that optimal management of the plantation and forestry sector must be in accordance with environmental, market and social conditions. An integrated approach to optimizing the management of garden and forest products will have an impact on the flow of materials, information and financing in supply chain management. This flow is expected to run efficiently and effectively when these dimensions undergo continuous improvement and evaluation in each period [19]. The implementation of improvements to these three dimensions is expected to increase the productivity of a eucalyptus plantation. Currently, there are 13 million hectares of productive eucalyptus plantations out of 22 million hectares. It is hoped that the amount of productive land will increase if an approach and method is adopted in accordance with the conditions of each country. The eucalyptus plant 2nd TALENTA-International Conference on Science and TechnologyJournal of Physics: Conference Series2421 (2023) 012009IOP Publishingdoi:10.1088/1742-6596/2421/1/0120097 has been cultivated by more than 90 countries in the world. So that the impact of increasing productivity has the opportunity to cover globally [20]. 5. Conclusion The results of this study concluded that an significant factor in improving the optimization of the supply chain of eucalyptus plants is the integration between the economic, social and environmental dimensions which has experienced continuous improvement. Continuous improvement in these three dimensions will encourage an increase in the optimum value of the eucalyptus plant supply chain. This optimization will be integrated without neglecting one dimension. An approach in the form of a simulation model using agent-based modelling can be used to get a picture of the interaction of the three dimensions. This description can be a consideration for decision makers to make continuous improvements. References [1] Philip J. Smethursta, Rafael V. Valadaresb, Neil I. Huthd, Auro C. Almeidaa, Elvis F. Ellie, and Júlio C.L. Nevesb 2020 Generalized model for plantation production of Eucalyptus grandis and hybrids for genotype-site-management applications Forest Ecology and Management 469 https://doi.org/10.1016/j.foreco.2020.118164. [2] Humberto de Jesus Eufrade-Junior, Saulo Philipe Sebasti~ao Guerra, Claudio Angeli Sansígolo, Adriano Wagner Ballarin 2018 Management of Eucalyptus short-rotation coppice and its outcome on fuel quality Renewable Energy 121 https://doi.org/10.1016/j.renene.2018.01.033. [3] J.G.G. Jonker, H.M. Junginger, J.A. Verstegen, T. Lin, L.F. Rodríguez, K.C. Ting, A.P.C. Faaij, F. van der Hilst 2016 Supply chain optimization of sugarcane first generation and eucalyptus second generation ethanol production in Brazil Applied Energy 179 https://doi.org/10.1016/j.apenergy.2016.04.069. [4] Emilie Chardine-Baumann,ValerieBotta-Genoulaz 2014 A frame work for sustainable performance assessmen tof supply chain management practices Computers and Industrial Engineering 76 DOI: 139 10.1016/j.cie.2014.07.029 [5] Andreia Santosa, Ana Carvalho, Ana Paula Barbosa-Póvoa, Alexandra Marques, Pedro Amorim 2019 Assessment and optimization of sustainable forest wood supply chains – A systematic literature review Forest Policy and Economics 105 https://doi.org/10.1016/j.forpol.2019.05.026. [6] M. Bouchard, S. DÕAmours, M. Ronnqvist, R. Azouzi, E. Gunn 2016 Integrated optimization of strategic and tactical planning decisions in forestry European Journal of Operational Research 259 DOI: 10.1016/j.ejor.2016.11.022. [7] Veronika Gaube, Christina Kaiser, Martin Wildenberg, Heidi Adensam, Peter Fleissner, Johannes Kobler, Juliana Lutz, Andreas Schaumberger, Jakob Schaumberger, Barbara Smetschka, Angelika Wolf, Andreas Richter, Helmut Haber 2009 Combining agent-based and stock-flow modeling approaches in a participative analysis of the integrated land system in Reichraming, Austria Landscape Ecol 24 10.1007/s10980-009-9356-6. [8] Elkafi Hassini, ChiragSurti , CorySearcy 2012 A literature review and a case study of sustainable supply chains with a focus on metrics Production Economics 140 https://doi.org/10.1016/j.ijpe.2012.01.042. [9] Chawalit Manisri, Juta Pichitlamken 2017 Hybrid Simulation Model for the Upstream Rubber Supply Chain Industrial Engineering & Management Systems 16 doi.org/10.7232/iems.2017.16.3.330. [10] Muhammad Haikal Sitepu, Alison McKay, and Raymond J. Holt 2019 An approach for the formulation of sustainable replanting policies in the Indonesian natural rubber industry Journal of Cleaner Production 241 doi.org/10.1016/j.jclepro.2019.118357. [11] Mohit Kumar, P.K. Mishra, S.N. Upadhyay 2019 Pyrolysis of Saccharum munja Optimization of process parameters using response surface methodology (RSM) and evaluation of kinetic parameters Bioresource Technology Reports 8 https://doi.org/10.1016/j.biteb.2019.100332. 2nd TALENTA-International Conference on Science and TechnologyJournal of Physics: Conference Series2421 (2023) 012009IOP Publishingdoi:10.1088/1742-6596/2421/1/0120098 [12] Susannah Fleiss, Emily H. Waddell, Bernadus Bala Ola, Lindsay F. Banin, Suzan Benedick, Azlin Bin Sailim, Daniel S. Chapman, Ahmad Jelling, Henry King, Colin J. McClean, Kok Loong Yeong, Jane K. Hill 2020 Conservation set-asides improve carbon storage and support associated plant diversity in certified sustainable oil palm plantations Biological Conservation 248 https://doi.org/10.1016/j.biocon.2020.108631. [13] Irsa Mumtaz, Zahid Majeed, Zainab Ajab, Basharat Ahmad, Kiran Khurshid, Muhammad Mubashir 2019 Optimized tuning of rosin adduct with maleic anhydride for smart applications in controlled and targeted delivery of urea for higher plant’s uptake and growth efficiency Industrial Crops & Products 133 https://doi.org/10.1016/j.indcrop.2019.02.036. [14] Nowilin James Rubinsin, Wan Ramli Wan Daud, Siti Kartom Kamarudin, Mohd Shahbudin Masdar, Masli Irwan Rosli, Sheila Samsatli, John Frederick Tapia, Wan Azlina Wan Ab Karim Ghani, Kean Long Lim 2020 Optimization Of Oil Palm Empty Fruit Bunches Value Chain In Peninsular Malaysia Food and Bioproducts Processing 119 https://doi.org/10.1016/j.fbp.2019.11.006. [15] Celso Herrera-Cáceres, Francisco Pérez-Galarce, Eduardo Álvarez-Miranda, Alfredo Candia-Véjar 2017 Optimization of the harvest planning in the olive oil production: A case study in Chile Computers and Electronics in Agriculture 141 https://doi.org/10.1016/j.compag.2017.07.017. [16] Rolando Saltini, Renzo Akkerman , Stina Frosch 2013 Optimizing chocolate production through traceability: A review of the influence of farming practices on cocoa bean quality Food Control 29 https://doi.org/10.1016/j.foodcont.2012.05.054. [17] S. Sánchez-García, D. Athanassiadis, C. Martínez-Alonso, E. Tolosana, J. Majada, E. Canga 2017 A GIS methodology for optimal location of a wood-fired power plant: Quantification of available woodfuel, supply chain costs and GHG emissions Journal of Cleaner Production 157 DOI: 10.1016/j.jclepro.2017.04.058. [18] Saman Hassanzadeh Amin, Guoqing Zhang, dan Miqdad Nidal Eldali 2020 A review of closed-loop supply chain models Closed-loop supply chain (CLSC) networks 925 https://doi.org/10.1007/s42488-020-00034-y. [19] Vı´ctor Hugo Gutie´rrez, Mauricio Zapata, Carlos Sierra, William Laguado, Alı´ Santacruz 2006 Maximizing the profitability of forestry projects under the Clean Development Mechanism using a forest management optimization model Forest Ecology and Management 226 https://doi.org/10.1016/j.foreco.2006.02.002. [20] N Sembiring, H L Napitupulu, A I Sipahutar, and M T Sembiring 2020 A review of sustainable replanting eucalyptus: higher sustainable productivity Materials Science and Engineering 935 doi:10.1088/1757-899X/935/1/012068. Acknowledgments Funding for this article was supported by TALENTA USU/2020. | RESEARCH GATE |
97Acta Horticulturae et Regiotecturae – Special Issue/2021Slavomír Hološ, Peter ŠurdaDefinition and concept of droughtClimate change has led to an increasing trend of drought in the recent decades, and models predict that the global drought risk will intensify further in the 21st century (Dai, 2013). While mean precipitation will increase, precipitation in the subtropical latitudes tends to decrease, particularly in the Mediterranean. Precipitation changes become statistically significant only when the temperature rises by at least 1.4 °C, and in many regions, the projected changes during the 21st century lie within the range of the late 20th century natural variability (Mahlstein, Portmann, Daniel, Solomon, Knutti. 2012). The research on drought has attracted the attention of scholars, government departments and the public. The development of drought is relatively slow, but its effects are very devastating. Drought is easily reflected in agriculture in reduced yields, but it is also associated with soil degradation and intense erosion. It could also result in the extinction of certain species of animals in the affected areas. In some areas of the Earth, drought can result in malnutrition, hunger and diseases that can lead to reduction of population. Dried soil and vegetation pose a fire hazard (Mouillot, Rambal, Joffre, 2002).Drought is a complex phenomenon and there is no clear physical quantity or definition by which drought can be measured. The lack of precipitation, relative to the climatic average of the area, is the main cause of drought. The increased rate of evapotranspiration, which is increased especially by higher air temperature, low relative humidity, low clouds, more intense sunlight or faster air flow, contributes to a significant intensification of drought. In certain cases, drought may result from the anomaly of other variables, such as temperature or evapotranspiration (Cook, Smerdon, Seager, Coats, 2014; Livneh, Hoerling, 2016; Luo et al., 2017). Moreover, drought may not be a purely natural hazard; human activities such as land use changes and reservoir operation may alter the hydrologic processes and affect drought development (Van Loonet al., 2016a). Overall, the development of drought results from the complicated interactions among the meteorological anomalies, land surface processes and human activities (Mishra, Singh, 2010).Drought types and characterizationTraditionally, drought can be classified into meteorological, agricultural, hydrological and socioeconomic drought, based on both physical and socioeconomic factors (Wilhite, Glantz, 1985). According to Dracup, Lee and Paulson (1980), meteorological drought is determined by the meteorological characteristics such as air temperature, total precipitation and duration of sunshine. Hydrological drought is defined as the lack of water in rivers due to normal flow, lack of groundwater and the lack of water supply in natural or artificial reservoirs. If there is lack of water in soil for animals and plants, we are talking about physiological or agricultural drought. Socio-economic drought is characterised by lack of water for normal social and economic human activities (Pedro-Monzonís, Solera, Ferrer, Estrela, Paredes-Arquiola, 2015).According to Svoboda and Fuchs (2016), it is essential to define drought indices and indicators. Indicators are parameters used to describe drought conditions (e.g. precipitation, temperature, streamflow, groundwater and reservoir levels, soil moisture and snowpack). Indices are Acta Horticulturae et Regiotecturae – Special IssueNitra, Slovaca Universitas Agriculturae Nitriae, 2021, pp. 97–108EVALUATION OF DROUGHT – REVIEW OF DROUGHT INDICES AND THEIR APPLICATION IN THE RECENT STUDIES FROM SLOVAKIASlavomír Hološ1,2*, Peter Šurda21Slovak University of Agriculture in Nitra, Slovakia 2Institute of Hydrology, Slovak Academy of Sciences, Bratislava, SlovakiaDrought has recently become an important topic in Europe but also in Slovakia. Observed results from various studies suggest that this drought phenomenon has a serious impact on hydrology, agriculture and social and economic sectors. The first part of the paper was devoted to the study of literature from the field of existing drought indices, which serve to identify all types of drought such as meteorological, agricultural and socio-economic drought. The second part of the paper dealt with selected scientific studies on drought assessment and the use of drought indices in Central Europe and Slovakia.Keywords: drought, drought index, meteorological drought, agricultural drought, socio-economic droughtDOI: 10.2478/ahr-2021-0015Contact address: Slavomír Hološ, Slovak University of Agriculture, Faculty of Horticulture and Landscape Engineering, Department of Biometeorology, 949 76 Nitra, SlovakiaMaterial and methods98Acta Horticulturae et Regiotecturae – Special Issue/2021Slavomír Hološ, Peter Šurdatypically computed numerical representations of drought severity, assessed using hydrometeorological inputs.More than 100 drought indices have been proposed so far (Heim, 2002; Liu et al., 2018; Vicente-Serrano et al., 2012; Zargar, Sadiq, Naser, Khan, 2011). These indices correspond to different types of drought, including meteorological, agricultural and hydrological drought. With the help of drought indices, we can provide information to the decision-makers in business, government and many stakeholders. These tools can be used to provide an early drought warning system (Lohani, Loganthan, 1997) to calculate the probability of a drought ending (Karl, Quinlan, Ezell, 1987), determine drought relief (Wilhite, Rosenberg, Glantz, 1986), assess the risk of forest fires (Wheaton, 1994), predict crop yield (Sakamoto, 1978; Kumar, Panu, 1997) and examine spatial and temporal characteristics of drought, drought severity and comparison between different regions (Alley, 1985; Soule, 1992; Nkemdirim, Weber, 1999).Meteorological drought The meteorological drought (or precipitation deficit) is generally caused by persistent anomalies in the large-scale atmospheric circulation patterns due to anomalous sea surface temperatures (SSTs) or other remote conditions (Dai, 2011). Globally and also in the conditions of the Slovak Republic, in the recent years, there has been a noticeable increase in the annual average air temperatures, which is closely related to a significant decrease in relative humidity (SHMÚ, 2015). The dry period can only be evaluated on the basis of total precipitation, which also takes into account the duration of precipitation-free days and the time distribution of precipitation. The meteorological drought does not result from a single cause but from a combination of multiple causes (e.g. reduced soil moisture and increased temperature), which may also contribute to the atmospheric anomaly (Dai, 2013; Kam, Sheffield, Wood, 2014). The commonly used meteorological drought indices are listed below.Standardized precipitation indexThe Standardized Precipitation Index (SPI) (Figure 1A a) is a widely used index to characterise the meteorological drought in a range of timescales. McKee, Doesken, Kleist (1995) used the probability of the precipitation occurrence for 3, 6, 12, 24, 48 months, and the output values ranged from -2.0 to +2.0. It was found that the Gamma distribution fits the precipitation time series very well. The Gamma distribution is defined by its frequency or probability density function as: (1) ‐40-4-3-2-10123200420052006200720082009201020112012201320142015201620172018SPI (-)YearSPISPIExtreme droughtSevere droughta050100150200250300350200420052006200720082009201020112012201320142015201620172018PNI (-)YearPNIPNIExtreme droughtSevere droughtbFigure 1A Values of SPI (a), PNI (b) index for the Malanta site, during the period 2004–2018; limit for severe (orange line) and extreme (red line) droughtSource: Šurda, Vitková, Rončák, 202011( )(0)( )xg xxex 99Acta Horticulturae et Regiotecturae – Special Issue/2021Slavomír Hološ, Peter ŠurdaFigure 1B Values of DI (c), MCZI(d), ZSI(e) index for the Malanta site, during the period 2004–2018; limit for severe (orange line) and extreme (red line) droughtSource: Šurda, Vitková, Rončák, 2020 024681012200420052006200720082009201020112012201320142015201620172018DI (-)YearDIDIExtreme droughtSevere droughtc-2.5-2-1.5-1-0.500.511.522.5200420052006200720082009201020112012201320142015201620172018MCZI (-)YearMCZIMCZIExtreme droughtSevere droughtd-3-2-101234200420052006200720082009201020112012201320142015201620172018ZSI (-)YearZSIZSIExtreme droughtSevere droughtewhere:Γ (α) – for gamma functionx – (mm) for precipitation amount (x >0)α – for shape parameter (α >0)β – for scale parameter (β >0)Percent of normal indexIndex PNI (Figure 1A b) was described by Willeke, Hosking, Wallis (1994) as a percentage of normal precipitation. It can be calculated for different time scales (monthly, seasonally and yearly). PNI (Percent of Normal Index) has been found to be rather effective for describing drought for a single region or/and for a single season (Hayes, 2006).PNI is calculated as following: (2)where:Pi – for the precipitation in time increment (mm)P – for the normal precipitation for the study period (mm)PNI100iPP 100Acta Horticulturae et Regiotecturae – Special Issue/2021Slavomír Hološ, Peter ŠurdaDI (deciles)The DI index (Figure 1A c) was defined as a classification of precipitation totals during a time period over the whole monitoring period (Gibbs, Maher, 1967). In particular, monthly precipitation totals data are sorted from the lowest to the highest and are divided into ten equal categories or deciles. Thus, precipitation in a given month can be placed in a historical context by deciles.MCZI (modified z-index)The National Climate Centre in China developed CZI (Figure 1A d) in 1995 as an alternative to the SPI index (Ju, Yang, Chen, 1997). Assuming that the average precipitation totals have a III. Pearson distribution, the CZI is calculated as: (3)where:i – for the observed time span and j is for the current monthCZIij – for the sum of the CZI values in the current month (j) during the period iCsi – for the skew coefficientφtj – for a standardized variationThe MCZI is calculated using the above formula and the median prtecipitation total is replaced by the arithmetic mean value.ZSI (z-sum)The ZSI index (Figure 1A e) is sometimes confused with the SPI index. This drought index is an analogue to the CZI, but does not work with gamma or Pearson‘s distribution of precipitation total data. The ZSI index can be calculated according to the following formula: (4)where:P– – for the average monthly precipitation total (mm)Pi – for the precipitation total in a particular month (mm)SD – for the standard deviation of the precipitation totals over the monitoring time interval (mm)Standardized precipitation-evapotranspiration index (SPEI)The SPEI index is based on the SPI index, but the SPEI index also includes the temperature component. This component allows the index to take into account the effect of temperature on drought. The SPI index is calculated using monthly (or weekly) precipitation as the input data. The SPEI index uses the monthly (or weekly) difference between precipitation and PET. This represents simple climatic water balance (Thornthwaite, 1948) that is calculated on different time scales to obtain SPEI.Reconnaissance drought index (RDI)The RDI index includes potential evapotranspiration and precipitation based on a simplified water balance equation, and the index also contains three outputs: a standardized value, normalized value and an initial value. If the standardized RDI value has a similar character as the SPI index, then it can be directly compared with it. RDI is more representative than SPI, because it uses complete water balance instead of precipitation itself. The parameters that enter the RDI index are: monthly precipitation temperatures and temperatures (Svoboda, Fuchs, 2016).Effective precipitation concept (DEP)Byun and Wilhite (1999) used the term of Effective Precipitation to describe the summed value of daily precipitation with a time-dependent reduction function, representing the daily depletion of water resources. The choice of the best reduction function (equation) remains an unsolved problem, because many parameters, like topography, soil characteristics, ability to keep water in reservoirs, air temperature, humidity, and wind speed, must be considered together precisely to represent the depletion of water resources in nature by runoff and evapotranspiration (Akhtari, Morid, Mahdian, Smakhtinm, 2009; Kalamaras, Michalopoulou, Byun, 2010; Kim, Byun, 2009; Kim, Byun, Choi, 2009; Morid, Smakhtin, Moghaddasi, 2006; Roudier, Mahe, 2010).The EPI index is calculated in a daily time step to overcome the big limitation of other indices – the long-time unit of assessment (most of the current drought indices use a monthly or longer time period as a unit).The EP index is based on the calculation of the effective precipitation during the selected time period. For the purposes of this work, 365-day effective precipitation (EP365) and a linear reduction function were selected, representing uniform loss of water resources throughout the year. (5)where:i – number of days whose total precipitation is included in the EP calculationHzm – total precipitation m-days before the first day included in the EP calculation (mm)For the needs of the complex drought diagnosis, it is necessary to supplement the EP index with other derived values. The first is the value of the long-term EP average for each day of the calendar year (MEP). In this work, the average value of the effective precipitation was calculated from the 14-year series EP365 (period 2004–2018) (Figure 2). Long-term EP (MEPn) for a normal (slightly humid) period was computed for the years 2005–2012 and long-term EP (MEPd) for a long-term dry period for the years 2012–2018, in this work. This time periods were selected according to the DEP index (Figure 3). With respect to the long-term average, the surplus or scarcity of water resources (DEP) for each day of the analysed period can be evaluated, according to the equation: DEP = EP365 – MEP (6)1366CZI126sisiijtjsisiCCCC ZSIiPPSD 11EP/inizmnmHn 101Acta Horticulturae et Regiotecturae – Special Issue/2021Slavomír Hološ, Peter ŠurdaAgricultural droughtAgricultural drought is commonly related to the deficit in soil moisture, which affects plant production and crop yield. This occurs mainly because the soil moisture availability governs the physiological processes in plants, and any paucity of water content in the crop root-zone can impede productivity (Wang, Lettenmaier, Sheeld, 2011; Mannocchi, Todisco, Vergni, 2004). A drought index using soil moisture would be directly related to the crop growth potential and could provide a decision supporting tool. The commonly used agricultural drought indicators according to Ajaz, Taghvaeian, Khand, Gowda, Moorhead (2019), are listed below.Palmer drought severity index (PDSI)This index was developed by Palmer (1965) as one of the first attempts to identify droughts using more than just precipitation data. Monthly precipitation and temperature along with the latitude and the available water capacity of the soil are the input data. PDSI has been used to identify droughts affecting agriculture, and also for identifying and monitoring droughts associated with other types of impacts. It takes into account received moisture (precipitation) as well as moisture stored in the soil, accounting for the potential loss of moisture due to the temperature influences (Svoboda, Fuchs, 2016; Wells, Goddard, Hayes, 2004).Palmer’s Z-IndexPalmer’s Z-Index (Z-Index) is a derivative of PDSI, and the Z values are s part of the PDSI output. It is sometimes referred to as the ‘Moisture Anomaly Index’, and the derived values provide comparable measure of the relative anomalies of a region for both dryness and wetness when compared to the entire record for that location. The moisture loss is multiplied with empirically derived climatic characteristics, and the monthly moisture anomaly index known as Z-Index is estimated (Karl, 1986). Z Index responds to short-term conditions better than PDSI and is typically calculated for much shorter timescales, enabling it to identify rapidly developing drought conditions (Svoboda, Fuchs, 2016).Soil waterdeficit indexThe SWDI was developed by Martínez-Fernández, González-Zamora, Sánchez, Gumuzzio (2015) and was estimated as: (7)where:θ – for the aggregated volumetric water content (VWC) of soil profileθFC – for the VWC at field capacity (FC)Figure 2 Daily values of Hz, MEP and EP365 for the meteorological station of Nitra, for the period 2005–2018Source: Šurda, Rončák, Vitková, Tárnik, 2019 0204060801001201400200400600800100012001400160020052006200720082009201020112012201320142015201620172018Hz (mm)EP365, MEP (-)DateHzEP365MEPFigure 3 Daily values of the DEP index for the meteorological station of Nitra, for the period 2005–2018Source: Šurda, Rončák, Vitková, Tárnik, 2019 -600-400-2000200400600800100020052006200720082009201020112012201320142015201620172018DEP (-)DateDEPDEPSWDI10FCAWC 102Acta Horticulturae et Regiotecturae – Special Issue/2021Slavomír Hološ, Peter ŠurdaθAWC – for the available water content estimated as the difference between VWC at FC and wilting point (WP) (all in m3.m-3)Water deficit indexThis index was developed by Cammalleri, Micale, Vogt (2016): (8)where:n – for an empirical exponent (unitless)θ50 – estimated by averaging VWC between the soil moisture thresholds as described by Cammalleri et al. (2016)θ – aggregated for the soil profile based on depthNormalized soil moistureThe NSM was proposed by Dutra, Viterbo, Miranda (2008) as: (9)where:θm, y – for VWC for the month m and the year y (m3.m-3) – for mean monthly VWC (m3.m-3)σm – for the standard deviation for all studied yearsFinally, remote sensing-based indicators such as the Normalized-Difference Vegetation Index (NDVI) or the fraction of the Absorbed Photosynthetically Active Radiation (fAPAR) are used to monitor drought impacts on the vegetation cover.The normalized difference vegetation index (NDVI) NDVI is defined as: (10)where:αnir and αvis represent surface reflectance averaged over ranges of wavelengths in the visible (λ ~0.6 μm, “red”) and near infrared, IR (λ ~0.8 μm) regions of the spectrum, respectivelyIt is clear from its definition that NDVI (like most other remotely sensed vegetation indices) is not an intrinsic physical quantity, although it is indeed correlated with certain physical properties of the vegetation canopy: leaf area index (LAI), fractional vegetation cover, vegetation condition and biomass.Fraction of absorbed photosynthetically active radiation (fAPAR)The quantity fAPAR is defined as the fraction of incident photosynthetically active radiation (PAR) that is absorbed by a canopy, which usually includes the overstory and sometimes the understory and ground cover (e.g. moss). fAPAR is calculated using: (11)where:PAR↓AC and PAR↑AC – incident (downward) and reflected (upward) PAR above the canopy, respectivelyPAR↓BC and PAR↑BC – the corresponding terms for the below of the canopyHydrological droughtHydrological drought is associated with deficiency in the bulk water supply, which may include water levels in streams, lakes, reservoirs and aquifers. Since it is directly linked to the drought impacts, it is argued that more attention is needed to study the hydrological drought (Cloke, Hannah, 2011; Mishra, Singh, 2010; Pozzi et al., 2013). From the major drought forms, the hydrological drought may be the slowest to develop (Soule, 1992). For example, a shortage of snowfall may not manifest itself as depressed runoff until half a year later. It is possible to minimise the negative impacts of the hydrological drought on the environment and society through the analysis of the minimum flows. These minimum flows are one of the characteristics that can define hydrological drought. The commonly used hydrologic drought indicators include Palmer hydrologic Drought Severity Index (PHDI), Standardized Runoff index (SRI), or reservoir level (Hayes, Svoboda, Wall, Widhalm, 2011).Total water deficit Total water deficit is a traditional assessment of the hydrological drought, synonymous with the drought severity S. This severity is the product of the duration D, during which flows are consistently below a truncation level (e.g., the hydro-climatic mean), and the magnitude M, which is the average departure of streamflow from the truncation level, during the drought period (Dracup et al., 1980). After the drought ends, the total water deficit resets to 0.Cumulative streamflow anomalyA cumulative departure of streamflow from mean conditions can show long-term tendencies in the water availability.Palmer hydrological drought severity indexThe Palmer Hydrological Drought Severity Index (PHDI) is very similar to PDSI, using the identical water balance assessment. Specifically, PDSI considers drought finished when the moisture conditions start an uninterrupted rise that ultimately erases the water deficit, whereas PHDI considers drought ended when the moisture deficit actually vanishes (Heim, 2000, 2002). This retardation is appropriate for the assessment of the hydrological drought, which is slower in developing than the meteorological drought.Surface water supply indexThe Surface Water Supply Index (SWSI) explicitly accounts for snowpack and its delayed runoff (Garen, 1993; Doesken, McKee, Kleist, 1991). SWSI is a suitable measure of the 501d1n ,,NSMm ymm ym ,,Mm ymm ym NDVInirvisnirvis ACACBCBCACPARPARPARPARAPARPARf 103Acta Horticulturae et Regiotecturae – Special Issue/2021Slavomír Hološ, Peter Šurdahydrological drought for regions, such as the mountainous catchments, where snow contributes significantly to the annual streamflow. Computations require measurements for snowpack, precipitation, streamflow and reservoir storage.Standardized runoff index (SRI) Shukla and Wood (2008) applied the concept for SPI in defining a standardized runoff index (SRI) as the standard normal deviation unit associated with the percentile of hydrologic runoff accumulated over specific duration. Different duration (e.g., 1-month, 9-month) and different spatial aggregations of the index can be calculated, depending on the source data resolution and desired application.Socioeconomic droughtThe socioeconomic drought incorporates features or impacts of three other types of drought. Drought impacts span a wide range of societal (e.g., health), economic (e.g., water supply, agricultural production and recreation), and environmental (e.g., forest productivity and wildfires) systems. The subject of the socio-hydrology, firstly conceived by Sivapalan, Savenije, Blöschl (2012), seeks to understand the ’dynamics and co-evolution of coupled human-water systems‘, including the impacts and dynamics of the changing social norms and values, system behaviours such as tipping points and feedback mechanisms, some of which may be emergent (unexpected), caused by non-linear interactions among processes occurring on different spatiotemporal scales. According to Van Loon et al. (2016b), human activities influence water input and output, and storage, and therefore modify the propagation of drought and can even be the cause of drought in the absence of natural drivers of drought. The drought typology based on natural processes should therefore be complemented with drought types based on human processes. According to Pedro-Monzonís et al. (2015), socio-economic drought is associated with the impact of water scarcity on people and the economic activity causing socio-economic, social and environmental impacts. To assess water scarcity, the water resource vulnerability index (Raskin, Gleick, Kirshen, Pontius, Strzepek, 1997), water stress index (Falkenmark, Lundqvist, Widstrand, 1989), critical ratio (Alcamo, Henrichs, Rösch, 2000), the water poverty index (Sullivan, 2002) and water footprint (Hoekstra, 2012) are the most commonly used approaches.Water resource vulnerability index considers scarcity to be the total annual withdrawals, and a percent of the available water resources. It is focused on the assessment of use for being more objective than demanding. Water stress indexCountries may be classified according to the renewable water resources per capita per year. It is easily understood and data are generally available. In contrast, average values may hide scarcity problems on smaller scales. It does not take into consideration the infrastructures that modify the water availability or the variations in demands among different countries.Critical ratio considers scarcity to be the ratio of water withdrawals for human use to the total renewable water resources. The difficulty of distinguishing the amount of water that could be available for human use considering evapotranspiration, return flows, environmental requirements, or the possibility of the society to adapt to water scarcity, belong among its limitations.Water poverty index represents the weighted average of its five dimensions: access to water; water quantity, quality and variability; water use; water management capacity; and environmental aspects. The input data are huge and expert judgments are required.Water footprint is defined as the total volume used to produce goods and services. It can be divided into three types: blue water footprint, green water footprint and grey water footprint. Drought quantification studies for the region of Slovakia over the last decadeDespite some controversy over global drought trends (Dai, 2012; Sheffield, Wood, Roderick, 2012), climate models project increases in mean temperature in the most of the land and ocean regions, hot extremes in the most inhabited regions, heavy precipitation in several regions, and the probability of drought and precipitation deficits in some regions (IPCC, 2018). Between 1990 and 2015, drought in the European Union affected more than 37% of its territory, representing 800.000 km2 and affecting 100 million people. Over a 30-year period (1985–2015), drought has cost the European Union more than € 100 billion (Andreu, Solera, Paredes-Arquiola, Haro-Monteagudo, van Lanen (Eds.), 2015). Therefore, drought has become an important research topic for scientists in the region of Central Europe and also in Slovakia.Meteorological drought studiesJanacova, Labudova, Labuda (2018) assessed the occurrence of the meteorological drought in the regions of the Záhorská, Danubian (Podunajská) and the Eastern Slovakian (Východoslovenská) lowlands and the Southern Slovakian (Juhoslovenská) and the Košická basins in the period 1981–2010. The analysis of the meteorological drought was evaluated on the basis of the monthly data of SPI. Areas which are the most threatened with the meteorological drought in different seasons were identified. There is greater hazard of drought for the Danubian lowland in spring and summer. On the contrary, the eastern part of Slovakia is threatened during the winter season.Nagy, Zelenaková, Kapostasová, Hlavatá, Simonová (2020) evaluated dry and wet periods at six climatic stations in eastern Slovakia using the following indices: standardized precipitation index (SPI), streamflow drought index (SDI), drought reconnaissance index (RDI) and standardized evapotranspiration index (SPEI) in a 12-month step over the period 1960–2015. The evaluation of the results showed an alternation of wet and dry periods and proved that the dry periods also occurred in the north of eastern Slovakia.Vido, Nalevanková, Valach, Šustek, Tadesse (2019) focused on the characterisation of the historical drought occurrences in the Horné Požitavie region, in Slovakia, over Results and discussion104Acta Horticulturae et Regiotecturae – Special Issue/2021Slavomír Hološ, Peter Šurdathe period 1966–2013 using the Standardized Precipitation-Evapotranspiration Index (SPEI). The results showed that drought occurred in the region regularly (recurrent climate feature), while the trend analysis indicated the trend towards more arid climatic conditions. Analyses of the SPEI trends in the individual months showed a decreasing trend of drought occurrences during the cold months of the year (i.e., October to March), while an increasing trend was indicated from April to August.Vido, Nalevanková (2020) analysed drought occurrence and trends using the SPI and the SPEI index in the upper Hron region within the 1984–2014 period. They found that:1. drought incidence decreased with an increasing altitude;2. increasing air temperature increased also the difference in the drought trends between lowlands and mountains during the studied period;3. abrupt changes in the time series of drought indices, which could indicate some signals of the changing atmospheric circulation patterns, were not revealed.Trnka et al. (2016) used the index SPI, PDSI, Z-index and SPEI, in their work. The time series of the drought indices were calculated for 411 climatological stations across Austria (excluding the Alps), the Czech Republic and Slovakia. Up to 45% of the evaluated stations (depending on the index) became significantly drier during the 1961–2014 period. An increase in the evaporative demand of the atmosphere, driven by higher temperatures and global radiation with limited changes in the precipitation totals, was the main driver behind this development. The study by Vido et al. (2015) focused on how drought occurs at higher altitudes of the Tatra National Park, which is a significant biological reserve of the Central European fauna and flora. Authors used the time series of SPI from 1961 to 2010 and standard GIS methods. The results showed that the frequency of drought occurrence has a cyclic pattern with approximately a 30-year period. The spatial analyses showed that the precipitation shadow of mountains influences the risk of drought occurrence. Zelenáková et al. (2018) evaluated the trend analysis applied to the precipitation and temperature monthly data for the period from 1962 to 2014 at sixteen climatic stations in eastern Slovakia. All climatic stations in eastern Slovakia show a positive trend in temperature during the year. Trends in precipitation are also mostly positive during winter and spring. An abrupt shift in precipitation at the highest climatic station, Lomnický peak, began around 1985 (+). Abrupt shifts in temperature began around 1970 (+) at the presented climatic stations. The extremity of climate is confirmed by an analysis of the trends in wet and dry spells. Trends showed increasing tendencies in medium- and long-term wet spells.Zelenáková et al. (2017) analysed the temporal and spatial trends in the annual and seasonal precipitation in Slovakia, in their work, utilising 487 gauging station data collected state-wide in the period from 1981 to 2013. In general, the precipitation data in the studied area have not changed during the last 33 years, and there are no big gaps. However, predominantly increasing trends in the precipitation time series were found at most of the gauging stations in Slovakia. There is also evidence of different rain distribution from the monthly point of view. Decreasing trends were detected in December in the northern part of Slovakia, while the central and southern parts revealed increasing trends. Most of the stations showed increasing summer precipitation trends, especially in July.Nikolová, Nejedlik, Lapin (2016) analysed drought in lowlands of Slovakia on the basis of SPI and SPEI for the period 1961–2011. The results show that temperature has an important role for occurrence of moderate and severe drought at monthly level and precipitation is the main factor for occurrence of extreme drought. There are an increasing number of cases with severe or extreme drought in summer. Future projection of drought shows a general tendency to the increasing frequency of severe dry events in 2001–2050 and 2051–2100, while there will be a little decreasing of the extremely dry months in comparison to 1961–2010.Lapin, Gera, Hrvol, Melo, Tomlain (2009) claimed that regimes of evapotranspiration, soil moisture and runoff have changed mainly in southern Slovakia. A physical model for the estimation of the energy balance equation components has been developed. Input data was gained from 31 meteorological stations in Slovakia since 1951. The 20-year period of 1988–2007 was by 0.9 °C warmer than the normal period mean. Annual precipitation totals have not changed significantly, but the substantial changes have been found in the precipitation regime. The scenarios show significant changes in the hydrological cycle not only at river basins balance but also in case of soil water balance, mainly in southern Slovakia.Hydrological drought studies Fendeková, Fendek (2012) analysed groundwater drought indices, which could be derived for different groundwater parameters, among them for base flow, groundwater head stage, spring yield, or groundwater recharge. Base flow drought assessment methods were proposed in the paper. The base flow drought severity index was applied, calculated as the value of the base flow drought deficit volume divided by the drought duration. After that, the standardized base flow drought severity index was proposed as the ratio of the base flow drought index and the average long-term annual base flow. Proposed methods were applied in the Nitra River basin. Base flow drought occurrence was characterised also from the seasonality point of view.Fendeková et al. (2018) used SPI and SPEI for assessment of the meteorological drought occurrence. The research was established on a discharge time series representing twelve river basins in Slovakia, within the period 1981–2015. Results showed that the drought parameters in the evaluated river basins of Slovakia differed in respective years, most of the basins suffered more by 2003 and 2012 drought than by the 2015 one. Water balance components analysis for the entire period 1931–2016 showed that because of the continuously increasing air temperature and evapotranspiration balance, there is a decrease of runoff in the Slovak territory.Zelenáková et al. (2014) identified statistically significant trends in the stream flow characteristics of the low water content in eastern Slovakia, in their work, which are used in the evaluation of the hydrological drought. This analysis was carried out due to the statistical data from 63 river stations, 105Acta Horticulturae et Regiotecturae – Special Issue/2021Slavomír Hološ, Peter Šurdalying in the eastern part of Slovakia. Mann-Kendall statistical test identifies the frequency of minimal stream flow trends. Obtained results from the statistically significant trends in the stream flows are in a role of a basement for the regionalisation of the eastern Slovakia territory from the point of the hydrological drought risk.Hanel et al. (2014) used global climate models to develop climate change scenarios for four small catchments in the Czech and Slovak Republic. This method applies a nonlinear transformation to precipitation in order to match projected changes in the precipitation variability. Similarly, temperature is transformed considering the changes in mean and variability. The results show an increase in the number of minor droughts and an increase in the most severe droughts. There are clear differences in the changes of drought characteristics related to the dominant runoff regime in a catchment.Blahušiaková et al. (2020) investigated changes in the seasonal runoff and low flows related to the changes in snow and climate variables in the mountainous catchments in Central Europe. The results showed an increase in air temperature, decrease in snowfall fraction and snow depth, and changes in precipitation. Most of the hydrological droughts were connected either to low air temperatures and precipitation during winter or high winter air temperatures, which caused below-average snow storages. Findings show that besides precipitation and air temperature, snow plays an important role in summer streamflow and drought occurrence in the selected mountainous catchments.Agricultural (physiological) drought studiesLabudová, Labuda, Takáč (2016) focused on the assessment of drought intensity impact on crop yields on the Danubian and the Eastern Slovakian Lowland. Limited yield data resulted in the limited length of the assessed period (1996–2013). The standardised yields of ten crops (winter wheat, spring wheat, winter barley, spring barley, rye, maize, potatoes, oilseed rape, sunflower and sugar beet) were correlated with monthly, 2-, and 3-monthly SPI and SPEI. The highest correlation was between maize and the 3-monthly SPEI. Crop yields in the Eastern Slovakian Lowland do not seem to be influenced by wet/dry periods, identified using SPI and SPEI, as their correlation with both indices is quite low and insignificant.Čistý, Jarabicová and Minarič (2016) evaluated a spatial indicator of the threat of droughts, namely the available water capacity of soil. Data from a soil survey and data measured in a laboratory were used for the development of the pedotransfer functions with the help of the Random Forest algorithm. On the basis of the pedotransfer function, the available water capacity was spatially evaluated by geostatistical methods in the investigated area, i.e., in the Záhorská Lowland, in Slovakia. Vido et al. (2016) analysed the physiological response of tree species in central Slovakia to the driest months of 2012. Lukasová et al. (2020) focused on the onset of leaf colouring-LCO-((BBCH)92) of the European beech (Fagus sylvatica, L.). The limiting climate conditions for LCO were defined by the meteorological drought indices: climatic water balance (CWB), standardized precipitation index (SPI), standardized precipitation-evapotranspiration index (SPEI), dry period index (DPI), and heat waves (HW). During 23-year period (1996–2018) of ground-based phenological observations, the timing of LCO was significantly delayed at the middle to high altitudes. Over the last decade, LCO at the middle altitudes started at comparable to low altitudes. This resulted mainly from the significant negative effect of drought prior to this phenological phase. The ongoing warming trend of summer months suggests further intensification of drought spreading from the continual increase of evapotranspiration over the next decades.Bernáth et al. (2020) evaluated the drought impact on the quality parameters of grapes in the locality of the Cultivar Testing Station, Dolné Plachtince. Interanual variability of the drought impact on the grape quality was evaluated according to PDSI. The 1990–2014 period was used as a basis for the evaluation. The PDSI values as well as the sugar and acid contents were correlated to find the strength of relation between them. Short drought periods did not influence the grape quality significantly, while long drought periods caused a decrease of the acid content and an increase of the sugar content.Tuzinsky, Gregor, Tuzinsky, Homolak (2018) analysed the balance of soil water in the spruce stand mountain conditions in the Upper Orava region. The long-term research (1991–2014) shows that the predominant moisture interval in the vegetation period is semi-uvidic soil interval with good or sufficient supply of usable water. Ongoing climatic conditions with a gradual reduction of precipitation and increase of air temperature pose danger, associated with the development of dry periods, to the spruce. Under such conditions, the spruce is threatened with drought, and physiological weakness, reduced evapotranspiration, increased fall of the assimilation organs, reduced increment, degradation of physical and hydrological properties of soil, and reduction of transport of mineral and organic substances are all its responses.Takac, Moravek, Klikusovska, Skalsky (2014) assessed drought severity in the agricultural regions of Slovakia in the years 2011–2013. Standardized index based on the daily available soil water content was used for drought severity classification. The results of the analysis confirmed the occurrence of the meteorological drought in the years 2011 and 2012 and the occurrence of the agronomic drought in the years 2011–2013. Greater areal extension of the impact of drought on crop production was observed only in the years 2012 and 2013.Šiška, Takáč (2009) estimated the climatic index of drought and evapotranspiration deficits for the agricultural regions of the Slovak Republic. Climate change conditions were generated by general circulation model CCCM for emission scenario SIZES B2. Five categories of drought conditions were recognized in the reference period 1961–1990, and additional two very dry categories can be recognized in the agricultural regions of Slovakia, according to both estimated climatic indices.Šustek, Vido, Škvareninová, Škvarenina, Šurda (2017) documented the impact of the 2012 dry season on the decline in the beetle species (Carabids) in the High Tatras. The Standardized Precipitation Evapotranspiration Index was shown, using the cross-correlation of SPEI and number of individuals and species of Carabids as a suitable means 106Acta Horticulturae et Regiotecturae – Special Issue/2021Slavomír Hološ, Peter Šurdato explain and predict such changes for the period of 1–2 years.Brezianská, Vitková, Šurda (2018) analysed the occurrence of drought and reduced soil water storage in the Záhoská Lowland, in 1961–2010. Conclusions1. Drought is a consequence of climate anomalies, as well as of (wrong) human water use practices. This paper has reviewed the literature concerning the existing drought indices. Thus, they serve to identify and quantify all types of drought (meteorological, agricultural, hydrological or socio-economic). The paper has presented a vast number of indices demands by collecting information related to a huge variety of disciplines and representing a complex challenge.2. The second part of paper has been devoted to an overview of the selected scientific studies about the use of various drought indices (and indicators) and drought assessment in the conditions of Slovakia and Central Europe. Most of the studies are focused on meteorological, less on hydrological or agricultural drought and on the impacts of the increased incidence of drought on flora (alternatively fauna).Major conclusions from the reviewed studies: y an increase in air temperature, changes in precipitation patterns, a decrease in snowfall fraction and snow depth y a trend towards more dry (arid) climatic conditions y catchments are becoming drier and runoff is decreasing – an increase in the evaporative demand of the atmosphere, driven by higher temperatures and global radiation with limited changes in precipitation totals are the main drivers behind this development y an important role of snow in summer streamflow and drought occurrence in the mountainous catchments y an increasing number of severe drought events during summer in lowlands y a cyclic pattern of drought events in the High Tatras y delayed phenological phases and lower quality of grapes, caused by drought at some localities.AcknowledgmentsThis contribution was supported by the Scientific Grant Agency VEGA Project No. 2/0150/20.Ajaz, A., Taghvaeian, S., Khand, K., Gowda, P.H., Moorhead, J.E. (2019). Development and evaluation of an agricultural drought index by harnessing soil moisture and weather data. Water, 11, 1375.Akhtari, R., Morid, S., Mahdian, M.H., Smakhtinm, V. (2009). Assessment of areal interpolation methods for spatial analysis of SPI and EDI drought indices. Int. J. Climatol., 29(1), 135–145. https://doi.org/10.1002/joc.1691Alcamo, J., Henrichs, T., Rösch, T. (2000). World water in 2025 – global modeling and scenario analysis for the world commission on water for the 21st century. In: Report A0002, Center for Environmental Systems Research. Kassel, Germany: University of Kassel, Kurt Wolters Strasse 3.Alley, W.M. (1985). The Palmer Drought Severity Index as a measure of hydrologic drought. Water Resour. Bull., 21(1), 105–114.Andreu, J., Solera, A., Paredes-Arquiola, J., Haro-Monteagudo, D., van Lanen, H. (Eds.). (2015). Drought: Research and Science-Policy Interfacing. London: CRC Press. https://doi.org/10.1201/b18077Bernáth, S., Šiška, B., Paulen, O., Zuzulová, V., Pintér, E., Žilinský, M., Tóth, F. (2020). Grape Quality Parameters in Western Carpathian Region under Changing Climatic Conditions as Influenced by Drought. Journal of Ecological Engineering, 21(4), 39-45. https://doi.org/10.12911/22998993/119796Blahušiaková, A., Matoušková, M., Jenicek, M., Ledvinka, O., Kliment, Z., Podolinská, J., Snopková, Z. (2020). Snow and climate trends and their impact on seasonal runoff and hydrological drought types in selected mountain catchments in Central Europe. Hydrological Sciences Journal, 65(12), 2083–2096. https://doi.org/10.1080/02626667.2020.1784900Brezianská, K., Vitková, J., Šurda, P. (2018). Drought analysis and the impact of climate change on soil water supply in the Záhorská lowland. In Aktuálne problémy zóny aerácie pôdy v podmienkach prebiehajúcej klimatickej zmeny, Veda (pp. 307–335) (in Slovak).Byun, H.R. and Wilhite, D.A. (1999). Objective quantification of drought severity and duration. Int. J. Climatol., 12(9), 2747–2756.Cammalleri, C., Micale, F., Vogt, J. (2016). A novel soil moisture-based drought severity index (DSI) combining water deficit magnitude and frequency. Hydrol. Process., 30, 289–301.Čistý, M., Jarabicová, M., Minarič, P. (2016). Spatial Assessment of Soil Water Storage as an Identifier of Areas Threatened by Drought. Procedia Engineering, 161, 1738–1744. https://doi.org/10.1016/j.proeng.2016.08.768Cloke, H.L., Hannah, D.M. (2011). Large-scale hydrology: Advances in understanding processes, dynamics and models from beyond river basin to global scale. Hydrological Processes, 25, 991–995. https://doi.org/10.1002/hyp.8059Cook, B. I., Smerdon, J. E., Seager, R., Coats, S. (2014). Global warming and 21st century drying. Climate Dynamics, 43(9–10), 2607–2627. https://doi.org/10.1007/s00382-014-2075-yDai, A. (2011). Drought under global warming: A review. Wiley Interdisciplinary Reviews: Climate Change, 2, 45–65.Dai, A. (2012). Increasing drought under global warming in observations and models. Nat. Clim. Change, 3, 52–8.Dai, A. (2013). Increasing drought under global warming in observations and models. Nature Clim Change, 3, 52–58. https://doi.org/10.1038/nclimate1633Doesken, N.J., McKee, T.B., Kleist, J. (1991). Development of a Surface Water Supply Index for the western United States. Climatology Rep. 91–93, Colorado Climate Center, Dept. of Atmospheric Science, Colorado State University, Fort Collins, CO (76 p.).Dracup, J.A., Lee, K.S., Paulson, E.G.J. (1980). On the definition of drought. Water Resour. Res., 16(2), 297–302.Dutra, E., Viterbo, P., Miranda, P.M.A. (2008). ERA-40 reanalysis hydrological applications in the characterization of regional drought. Geophys. Res. Lett., 35, 2–6.Falkenmark, M., Lundqvist, J., Widstrand, C. (1989). Macro-scale water scarcity requires micro-scale approaches. Nat. Res. Forum, 13, 258–267.Fendeková, M., Fendek, M. (2012). Groundwater drought in the Nitra River Basin – identification and classification. J. Hydrol. Hydromechanics, 60, 185–193. https://doi.org/10.2478/v10098-012-0016-1Fendeková M., Gauster T., Labudová L., Vrabliková D., Danáčová Z., Fendek M., Pekárová P. (2018). Analysing 21st century meteorological and hydrological drought events in Slovakia. J. Hydrol. Hydromech., 66(4), 393–403.References107Acta Horticulturae et Regiotecturae – Special Issue/2021Slavomír Hološ, Peter ŠurdaGaren, D.C. (1993). Revised Surface-Water Supply Index for western United States. Water Resour. Plan. Manage., 119, 437-454.Gibbs, W., Maher, J. (1967). Rainfall Deciles as Drought Indicators. Melbourne: Bureau of Meteorology (117 p.).Hanel, M., Vizina, A., Martínková, M., Horáček, S., Porubská, D., Fendek, M., Fendeková, M. (2014). Changes of drought characteristics in small Czech and Slovakian catchments projected by the CMIP5 GCM ensemble. In: Hydrology in a changing world: environmental and human dimensions, IAHS (pp. 78–83).Hayes, M.J. (2006). Drought Indices. Van Nostrand’s Scientific Encyclopedia. Hoboken: John Wiley & Sons, Inc.Hayes, M., Svoboda, M., Wall, N., Widhalm, M. (2011). The Lincoln declaration on drought indices: Universal meteorological drought index recommended. Bulletin of the American Meteorological Society, 92, 485–488. https://doi.org/10.1175/2010BAMS3103.1Heim, R.R. (2000). Drought indices: A review. Drought: A Global Assessment, D. A. Wilhite, Ed., Routledge (pp. 159–167).Heim, R.R. (2002). A Review of Twentieth-Century Drought Indices Used in the United States. Bull. Amer. Meteor. Soc., 83, 1149–1166. https://doi.org/10.1175/1520-0477-83.8.1149Hoekstra, A.Y. (2012). Water footprint accounting. In: Godfrey, J.M., Chalmers, K. (Eds.). Water Accounting. International Approaches to Policy and Decisionmaking. Edward Elgar Publishing Limited, Cheltenham, UK – Northampton, MA, USA (pp. 58–75).IPCC (2018). Summary for Policymakers. Global Warming of 1.5 °C. Geneva, Switzerland: World Meteorological Organization (32 p.).Janáčová, T., Labudová, L., Labuda, M. (2018). Meteorological drought in the parts of Slovakia with lowland features in 1981–2010. Geographia Cassoviensis, 12(1), 53–64. Ju, X.S., Yang, X.W., Chen, L.J. (1997). Research on determination of station indexes and division of regional flood/drought grades in China. Quarterly Journal of Applied Meteorology, 8(1), 26–33.Kalamaras, N., Michalopoulou, H., Byun, H.R. (2010). Detection of drought events in Greece using daily precipitation. Hydrol. Res., 41(2), 126–133. https://doi.org/10.2166/nh.2010.001Kam, J., Sheffield, J., Wood, E.F. (2014). A multiscale analysis of drought and pluvial mechanisms for the southeastern United States. Journal of Geophysical Research: Atmospheres, 119, 7348–7367. https://doi.org/10.1002/2014JD021453Karl, T.R. (1986). The sensitivity of the Palmer Drought Severity Index and Palmer’s Z-Index to their calibration coefcients including potential evapotranspiration. J. Clim. Appl. Meteorol., 25, 77–86.Karl, T., Quinlan, F., Ezell, D.S. (1987). Drought termination and amelioration: its climatological probability. J. Climate Appl. Meteor., 26, 1198–1209.Kim, D.W. and Byun, H.R. (2009). Future pattern of Asian drought under global warming scenario. Theor. Appl. Climatol., 98(1–2), 137–150. https://doi.org/10.1007/s00704-008-0100-yKim, D.W., Byun, H.R., Choi, K.S. (2009). Evaluation, modification, and application of the effective drought index to 200 – year drought climatology of Seoul Korea. J. Hydrol., 378(1–2), 1–12. https://doi.org/10.1016/j.jhydrol.2009.08.021Kumar, V., Panu, U. (1997). Predictive assessment of severity of agricultural droughts based on agro-climatic factors. J. Am. Water Resour. Assoc., 33(6), 1255–1264.Labudová, L., Labuda, M., Takáč, J. (2016). Comparison of SPI and SPEI applicability for drought impact assessment on crop production in the Danubian lowland and the east Slovakian lowland. Theor. Appl. Climatol., 128, 491–506.Lapin, M., Gera, M., Hrvol, J., Melo, M., Tomlain, J. (2009). Possible impacts of climate change on hydrologic cycle in Slovakia and results of observations in 1951–2007. Biologia, 64, 454–459.Livneh, B., Hoerling, M. P. (2016). The physics of drought in the U.S. Central Great Plains. Journal of Climate, 29(18), 6783–6804. https://doi.org/10.1175/JCLI-D-15-0697.1Liu, X., Zhu, X., Pan, Y. Bai, J. Li, S. (2018)Performance of different drought indices for agriculture drought in the North China Plain. J. Arid Land, 10, 507–516. https://doi.org/10.1007/s40333-018-0005-2Lohani, V.K., Loganathan, G.V. (1997). An early warning system for drought management using the Palmer Drought Index. J. Am. Water Resour. Assoc., 33(6), 1375–1386. Lukasová, V., Vido, J., Škvareninová, J., Bičárová, S., Hlavatá, H., Borsányi, P., Škvarenina, J. (2020). Autumn phenological response of european beech to summer drought and heat. Water, 12(9), 2610.Luo, L., Apps, D., Arcand, S., Xu, H., Pan, M., Hoerling, M. (2017). Contribution of temperature and precipitation anomalies to the California drought during 2012–2015. Geophysical Research Letters, 44, 3184–3192. https://doi.org/10.1002/2016GL072027Mahlstein, I., Portmann, R.W., Daniel, J.S., Solomon, S.,,Knutti, R. (2012). Perceptible changes in regional precipitation in a future climate. Geophysical Research Letters, 39, L05701. doi:10.1029/2011GL050738Mannocchi, F., Todisco, F., Vergni, L. (2004). Agricultural drought: indices, definition and analysis. Basis Civ. Water Sci., 286, 246–254. http://hydrologie.org/redbooks/a286/iahs_286_0246.pdfMartínez-Fernández, J., González-Zamora, A., Sánchez, N., Gumuzzio, A. (2015). A soil water-based index as a suitable agricultural drought indicator. J. Hydrol., 522, 265–273.McKee, T.B., Doesken, N.J., Kleist, J. (1995). Drought monitoring with Multiple Time scales. Proceeding of the 9th Conference on Applied Climatology. Dallas, TX: American Meteorological Society (pp. 233–236).Mishra, A.K., Singh, V.P. (2010). A review of drought concepts. J. Hydrol., 391(1–2), 202–216. https://doi.org/10.1016/ j.jhydrol.2010.07.012Morid, S., Smakhtin, V., Moghaddasi, M. (2006). Comparison of seven meteorological indices for drought monitoring. Iran. Int. J. Climatol., 26(7), 971–985. https://doi.org/10.1002/joc.1264Mouillot, F., Rambal, S., Joffre, R. (2002). Simulating climate change impacts on fire frequencyand vegetation dynamics in a Mediterranean-type ecosystem. Global Change Biology, 8, 423–437.Nagy, P., Zelenaková, M., Kapostasová, D., Hlavatá, H., Simonová, D. (2020). Identification of dry and wet years in eastern Slovakia using indices. Advances in Environmental Engineering: proceedings, IOP Publishing (pp. 1–6).Nikolová, N., Nejedlík., P., Lapin, M. (2016). Temporal variability and spatial distribution of drought events in the lowlands of Slovakia. Geofizika, 33(2), 119–135.Nkemdirim, L., Weber, L. (1999). Comparison between the droughts of the 1930s and the 1980s in the Southern Prairies of Canada. J. Climate, 12, 2434–2450.Palmer, W.C. (1965). Meteorological Drought. US Weather Bur. Res. Pap. no. 45, 58. https://www.ncdc.noaa.gov/temp-and-precip/drought/docs/palmer.pdfPedro-Monzonís, M., Solera, A., Ferrer, J., Estrela, T., Paredes-Arquiola, J. (2015). A review of water scarcity and drought indexes in water resources planning and management. Journal of Hydrology, 527, 482–493. https://doi.org/10.1016/j.jhydrol.2015.05.003Pozzi, W., Sheffield, J., Stefanski, R., Cripe, D., Pulwarty, R., Vogt, J.V. et al. (2013). Towards global drought early warning capability: Expanding international cooperation for the development of a framework for global drought monitoring and forecasting. Bulletin of the American Meteorological Society, 94(6), 776–785. https://doi.org/10.1175/BAMS-D-11-00176.1Raskin, P., Gleick, P., Kirshen, P., Pontius, G., Strzepek, K. (1997). Water Futures: Assessment of Long-Range Patterns and Prospects. Stockholm, Sweden: Stockholm Environment Institute.108Acta Horticulturae et Regiotecturae – Special Issue/2021Slavomír Hološ, Peter ŠurdaRoudier, P., Mahe, G. (2010). Study of water stress and droughts with indicators using daily data on the Bani River (Niger basin, Mali). Int. J. Climatol., 30(11), 1689–1705. https://doi.org/10.1002/joc.2013Sakamoto, C.M. (1978). The Z-index as a variable for crop yield estimation. Agric. Meteor., 19, 305–313.Sheffield, J., Wood, E.F., Roderick, M.L. (2012). Little change in global drought over the past 60 years. Nature, 491, 435–8.SHMÚ (2015). Manifestations of climate change at the global level. http://www.shmu.sk/sk/?page=1379 (in Slovak).Shukla, S., Wood, A.W. (2008). Use of a standardized runoff index for characterizing hydrologic drought. Geophysical Research Letters, 35, L02405. https://doi.org/10.1029/2007GL032487Šiška, B., Takáč, J. (2009). Drought analyses of agricultural regions as influenced by climatic conditions in the Slovak Republic. Idojárás, 13, 135–143.Sivapalan, M., Savenije, H.H.G., Blöschl, G. (2012). Sociohydrology: A new science of people and water. Hydrol. Process., 26, 1270–1276. doi:10.1002/hyp.8426Soule, P.T. (1992). Spatial patterns of drought frequency and duration in the contiguous USA based on multiple drought event definitions. Int. J. Climatol., 12, 11–24.Sullivan, C.A. (2002). Calculating a water poverty index. World Dev., 30, 1195–1210.Šurda, P., Rončák, P., Vítková, J., Tárnik, A. (2019). Regional drought assessment based on the meteorological indices for locality Nitra. Acta Hydrologica Slovaca, 20(1), 63–73 (in Slovak).Šurda, P., Vitková, J., Rončák, P. (2020). Regional Drought Assessment Based on the Meteorological Indices. Bulletin of the Georgian National Academy of Sciences, 14(2), 69–84.Šustek, Z., Vido, J., Škvareninová, J., Škvarenina J., Šurda, P. (2017). Drought impact on ground beetle assemblages (Coleoptera, Carabidae) in Norway spruce forests with different management after windstorm damage – a case study from Tatra Mts. (Slovakia). J. Hydrol. Hydromech., 65(4), 333–342.Svoboda, M., Fuchs, B. (2016). Integrated Drought Management Programme (IDMP). Handbook of Drought Indicators and Indices. Drought Mitigation Center Faculty Publications (117 p.). http://digitalcommons.unl.edu/droughtfacpub/117Takáč, J., Morávek, A., Klikušovská, Z., Skalský, R. (2014). Drought severity in agricultural land of Slovakia in the years 2011–2013. Mendel and Bioclimatology, 488–506.Thornthwaite, C.W. (1948). An approach toward a rational classification of climate. Geogr. Rev., 38, 55–94.Trnka, M., Balek, J., Štěpánek, P., Zahradníček, P., Možný, M., Eitzinger, J. et al. (2016). Drought trends over part of Central Europe between 1961 and 2014. Clim. Res., 70, 143–160.Tužinský, L., Gregor, J., Tužinský, M., Homolák, M. (2018). Pedohydrological cycles development in the spruce ecosystem under the current climatic conditions of Slovakia. Reports of forestry research, 63(4), 299-310.Van Loon, A.F., Gleeson, T., Clark, J., Van Dijk, A. I. J. M., Stahl, K., Hannaford, J. et al. (2016a). Drought in the Anthropocene. Nature Geoscience, 9(2), 89–91. https://doi.org/10.1038/ngeo2646Van Loon, A.F., Stahl, K., di Baldassarre, G., Clark, J., Rangecroft, S., Wanders, N. et al. (2016b). Drought in a human-modified world: Reframing drought definitions, understanding, and analysis approaches. Hydrology and Earth System Sciences, 20(9), 3631–3650. https://doi.org/10.5194/hess-20-3631-2016Vicente-Serrano, S.M.,Beguería, S., Lorenzo-Lacruz, J., Camarero, J. J., López-Moreno, J. I., Azorin-Molina, C. et al. (2012). Performance of Drought Indices for Ecological, Agricultural, and Hydrological Applications. Earth Interact., 16, 1–27. https://doi.org/10.1175/2012EI000434.1Vido, J., Nalevanková, P., Valach, J., Šustek, Z., Tadesse, T. (2019). Drought Analyses of the Horné Požitavie Region (Slovakia) in the Period 1966–2013. Adv. Meteorol., 1–10.Vido, J., Nalevanková, P. (2020). Drought in the Upper Hron Region (Slovakia) between the Years 1984–2014. Water, 12(10), 2887.Vido, J., Tadesse, T., Šustek, Z., Kandrík, R., Hanzelová, M., Škvarenina, J., Škvareninová, J., Hayes, M. (2015). Drought Occurrence in Central European Mountainous Region (Tatra National Park, Slovakia) within the Period 1961–2010. Adv. Meteorol., 1–8.Vido, J., Střelcová, K., Nalevanková, P., Leštianska, A., Kandrík, R., Pástorová, A., Škvarenina, J., Tadesse, T. (2016). Identifying the relationships of climate and physiological responses of a beech forest using the Standardised Precipitation Index: a case study for Slovakia. J. Hydrol. Hydromech., 64(3), 246–251.Wang, A., Lettenmaier, D.P., Sheeld, J. (2011). Soil moisture drought in China, 1950–2006. J. Clim., 24, 3257–3271.Wells, N., Goddard, S., Hayes, M.J. (2004). A self-calibrating Palmer Drought Severity Index. J. Clim., 17, 2335–2351.Wheaton, E.E. (1994). Impacts of a variable and changing climate on the Canadian prairies’ provinces: A preliminary intergration and annotated bibliography. SRC Publication No. E-2900-7-E-93, Saskatoon, SK: Saskatchewan Research Council.Wilhite, D. A., Glantz, M. H. (1985). Understanding: The drought phenomenon: The role of definitions. Water International, 10(3), 111–120. https://doi.org/10.1080/02508068508686328Wilhite, D.A., Rosenberg, N.J., Glantz, M.H. (1986). Improving federal reponse to drought. J. Climate Appl. Meteor., 25, 332– 342.Willeke, G., Hosking, J. R. M., Wallis, J. R. (1994). The national drought atlas. Institute for Water Resources Report 94-NDS-4. Norfolk, VA: U.S Army Corp of Engineers. Zargar, A., Sadiq, R., Naser, B., Khan, F.I. (2011). A review of drought indices. Environ. Rev., i, 333–49.Zelenáková, M., Purcz, P., Blištan, P., Vranayová, Z., Hlavatá, H., Diaconu, D.C., Portela, M.M. (2018). Trends in Precipitation and Temperatures in Eastern Slovakia (1962–2014). Water, 10, 727.Zelenáková, M., Vido, J., Portela, M.C.A.S., Purcz, P., Blištan, P., Hlavatá, H., Hluštík, P. (2017). Precipitation Trends over Slovakia in the Period 1981–2013. Water, 9, 922.Zelenáková, M., Purcz, P., Solaková, T., Simonová, D., Harbulaková, V.O. (2014). Trends in minimal stream flows at eastern Slovakia. Environmental Engineering 2014, 9th International Conference, selected papers, Vilnius, Lithuania: Gediminas Technical University (pp. 1–7).nnnnnn | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/354582024Effects of mixed and separate gender feeding on the growth performanceindices of Korean native chicken from day 1 to day 35 post-hatchConference Paper · November 2020CITATIONS0READS207 authors, including:Some of the authors of this publication are also working on these related projects:Golden Seed Project View projectRelevance of a Kenyan Feed Database. View projectOketch Elijah OgolaChungnam National University11 PUBLICATIONS 11 CITATIONS SEE PROFILEShan Randima NawarathneChungnam National University24 PUBLICATIONS 32 CITATIONS SEE PROFILEMyunghwan YuChungnam National University14 PUBLICATIONS 18 CITATIONS SEE PROFILEAll content following this page was uploaded by Oketch Elijah Ogola on 14 September 2021.The user has requested enhancement of the downloaded file.Effects of mixed and separate gender feeding on the growth performance indices of Korean nativechicken from day 1 to day 35 post-hatch저자(Authors)O. E. Ogola, H. M. Cho, J. S. Hong, Y. B. Kim, S. R. Nawarathne, M. H. Yu, J. M. Heo출처(Source)한국가금학회 정기총회 및 학술발표회 , 2020.11, 55-55 (1 pages)발행처(Publisher)한국가금학회Korean Society Of Poultry ScienceURLhttp://www.dbpia.co.kr/journal/articleDetail?nodeId=NODE10506677APA StyleO. E. Ogola, H. M. Cho, J. S. Hong, Y. B. Kim, S. R. Nawarathne, M. H. Yu, J. M. Heo (2020). Effects of mixedand separate gender feeding on the growth performance indices of Korean native chicken from day 1 to day 35post-hatch. 한국가금학회 정기총회 및 학술발표회, 55-55.이용정보(Accessed)저작권 안내DBpia에서 제공되는 모든 저작물의 저작권은 원저작자에게 있으며, 누리미디어는 각 저작물의 내용을 보증하거나 책임을 지지 않습니다. 그리고 DBpia에서 제공되는 저작물은 DBpia와 구독계약을 체결한 기관소속 이용자 혹은 해당 저작물의 개별 구매자가 비영리적으로만 이용할 수 있습니다. 그러므로 이에 위반하여 DBpia에서 제공되는 저작물을 복제, 전송 등의 방법으로 무단이용하는 경우 관련 법령에 따라 민, 형사상의 책임을 질 수 있습니다.Copyright InformationCopyright of all literary works provided by DBpia belongs to the copyright holder(s)and Nurimedia does not guarantee contents of the literary work or assume responsibility for thesame. In addition, the literary works provided by DBpia may only be used by the users affiliated to the institutions which executed a subscription agreement with DBpia or theindividual purchasers of the literary work(s)for non-commercial purposes. Therefore, any person who illegally uses the literary works provided by DBpia by means of reproduction ortransmission shall assume civil and criminal responsibility according to applicable laws and regulations.충남대학교168.***.23.1722021/09/14 22:37 (KST)2020 한국가금학회 정기총회 및 학술발표회Effects of mixed and separate gender feeding on the growth performance indices of Korean native chicken from day 1 to day 35 post-hatch55 Effects of mixed and separate gender feeding on the growth performance indices of Korean native chicken from day 1 to day 35 post-hatchO. E. Ogola, H. M. Cho, J. S. Hong, Y. B. Kim, S. R. Nawarathne, M. H. Yu, J. M. HeoDepartment of Animal Science and Biotechnology, Chungnam National University, KoreaBackground & ObjectivesIn order to preserve the important animal genetic resources and meet the increased consumer interest for the flavor, texture and health value of its meat, there has been development and intensified research on the 5 lines of Korean Native Chicken (KNC). However, there is scanty details on the effects of mixed and separate sex feeding on KNC. The present study evaluated the effects of mixed and separate gender feeding on the growth performance of two lines of KNC when compared to a white semi-broiler (WSB) and commercial broiler (CB).Materials & MethodsA total of 576 one-day-old chicks were randomly allotted to 72 cages with 18 replicates per treatment with 8 birds per cage and subjected to the same feeding and environmental conditions over a 5 week period. 288 KNC (K1 and K2), 144 WSB and 144 CB were used. Within the K1 and K2, three groups of males only (MO), females only (FO) and then male+female (FM) were used. Corn-soybean basal diet as per the Korean Feeding Standards for starter (1∼3 wks) and grower (3∼5 wk) diets were provided together with water ad-libitum. Body weights and feed consumed were measured weekly. Using the collected BW and feed consumed data, average daily feed intake (ADFI), average daily gain (ADG) and feed conversion ratios (FCR) were calculated. Data was analyzed using one-way ANOVA technique at P<0.05.ResultsIn an inter-breed comparison, the commercial broiler (CB) had the highest values for growth performance indices measured. It showed higher body weight, ADG, ADFI and lower FCR values (P<0.05) of around 1.54 that shows better feed efficiency. CB was followed by the WSB and finally the KNC. In an intra-breed comparison for KNC, the males showed better growth performance (P<0.05). The mixed group (FM) showed better growth performance than the FO group. No significant differences were recorded for the feed conversion ratio for the K1 line.ConclusionGender and the effects of mixed and separate gender feeding influenced the growth performance, and overall feed and production efficiency for the KNC.▶ Key words: mixed and separate sex feeding, growth performance, native chicken충남대학교 | IP:168.***.23.172 | Accessed 2021/09/14 22:37(KST)View publication stats | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/363662604Vöckler: Politics of Architecture. In: Vöckler (ed): SEE! Urban Transformationin Southeastern Europe. Vienna: Lit 2012, p. 14–25Article · September 2022CITATIONS0READS61 author:Some of the authors of this publication are also working on these related projects:LOEWE research cluster Infrastructure – Design – Society ("project-mo.de") View projectproject mo.de - The Mobility Design Project View projectKai VoecklerHochschule für Gestaltung Offenbach57 PUBLICATIONS 12 CITATIONS SEE PROFILEAll content following this page was uploaded by Kai Voeckler on 19 September 2022.The user has requested enhancement of the downloaded file.1415mark on a post-conflict situation, but the construction of new buildings also has a significant effect. Political power vacuums at the national level, along with the absence of self-monitoring in the civilian population, generate uncontrolled forces, which can seriously damage these cities’ chances for recovery. For this reason, it is necessary to scrutinize the aid and planning strategies we have used, and intensify the search for possible alternatives.1 One fact that often goes unrecognized is that social and political structures are shaped to a great degree by the way a city is developed. This raises the question of how the expertise of architects and planners can be utilized, not only to overcome the consequences of a crisis, but also to avoid more conflict in the future. What would an “architecture of peace” look like?Housing construction is strategically important. Housing is the basis for survival in war or other catastrophic situations, since it guarantees a minimum of safety and protection. And it is the key to urban development. It is not only of great economic significance, but it also helps to constitute new types of social interaction, which have the immediate ability to help institutions begin functioning again. It incorporates economic, social, and political capital. From the way a city is rebuilt, one can see how its inhabitants relate to their future prospects. Still — just like the process of rebuilding urban structures that have been destroyed during armed conflict — the construction of housing, the restoration of architectural legacies, or the expansion of industrial and commercial buildings are not merely questions of architecture or urban planning. Political and economic interdependencies are embedded in these issues. Architects and urban planners have to find their own positions in relation to this complicated association of intertwining forces.In recent years, the importance of the international community in managing local conflicts has increased considerably. Banding together international aid organiza-tions, the international community not only contributes a great deal to the regulation of conflicts, but also to the restoration of political, social, and cultural institutions. Even when the primary task is to provide basic necessities — such as water, electricity, streets, and public transportation — rebuilding a city’s infrastructure means that all of the decisions and steps taken in the process will profoundly influence the further develop-ment of the city. Thus, the urban transfor-mations initiated by global players in very diverse regions around the world often follow the same pattern. However, there are an increasing number of problems: first, it is becoming ever more difficult to discern who is involved, and second, large organiza-tions such as the UN, the OECD, and the World Bank, along with countless NGOs, are unable to coordinate their efforts. In addition, these organizations are frequently in indirect competition with each other for political sway; their financial backing is opaque, there is no legal oversight of their actions, and, last, but not least, there are individuals employed in these organizations who are involved in outright criminal activities, and they often get away scot-free. It has become urgently necessary to subject the methods and procedures of the international aid community to critical scrutiny. 2Global governance and NGO activitiesIn a post-conflict situation, the process of rebuilding a city and restoring acceptable living conditions must be adapted as soon as possible to a specific situation, which is marked by a great deal of ambiguity and insecurity. Most aid and development programs aim to establish new forms of good governance, which often simply imitate the political structures of the donor countries and usually try to apply generally unquestioned notions of a “civil society” to the situation at hand.3 Hence, normaliza-tion strategies around the globe operate with images and concepts — from historical reconstruction to the idea of capacity building. However, despite all of the good intentions, these strategies disregard the realities of these cities. Ideas and concepts, such as “freedom and democracy,” are exported, even though most of the time, they are mere euphemisms, used to cover up the introduction of the same capitalist principles of property and market economies that dominate the donor nations. Even the term “human rights” is regarded as a political norm that should be valid everywhere, and the fact that human rights have a history, as well as a changeable definition, is overlooked. To spell it out, so as not to be misunderstood: of course, human rights are highly desirable and worth defending, but they do not represent a neutral system of values. Rather, they are always permeated by political claims and ought to be negotiat-ed and adapted to suit local conditions. Briefly stated, too little consideration is given to the special circumstances of local political situations, and frequently, there is little knowledge of local culture and society.The result is a kind of “donor speak,” whose goal is to align everything with the political aims of the donor; this language is spoken by all of the members of local initiatives, as well as by those working for international organizations. Unfortunately, this is not a phantasm, but a political reality, and every plan in the crisis situation has to adjust to it. Donors are accountable to their own countries, and have to consider the political situations there, too. Consequently, it is all the more necessary to create a context for planning in crisis situations, and make sure that the goals of the plan reflect the local situation and are suitably appropriate. It is imperative that local political, cultural, and social conditions be considered in the process. Through urban planning, overlapping problems in the development of the society as a whole can be dealt with, and models for solutions can be concocted. However, this means that the “architecture of peace” must have a clear understanding of the specific form of local political governance that arises in a post-conflict situation, and it also has to reflect upon its own position in relation to that.Numerous recent global conferences — for example, the 1992 World Environmental Conference in Rio de Janeiro, the 1996 HABITAT II in Istanbul, the 1999 URBAN 21 in Berlin, and the 2010 World Urban Forum in Rio de Janeiro — have all rehabilitated planning, in terms of influencing and managing social processes. Widely imple-mented by global players, such as the World Bank in the 1980s, the predominant neoliberal ideology and its efforts to minimize state influence has proved crisis-prone and increased the drive to reconfigure state structures and develop new forms of planning. This strategy, christened “good governance,” directly addresses urban agglomeration as the focal point of social transformation, referring to a development that has become increasingly apparent over the last two decades: in conjunction with the significant expansion of cross-border exchange, new transnational structures have also arisen, and they express a new kind of relationship between the local and the global.This is associated with profound alterations in the concept of statehood itself, as well as in planning, for it is no longer the hierarchic and centralistic state that comprises the center of politics and controls development, when it comes to comprehensive planning. Governmental action instead distinguishes itself in various areas. The state now tends to play the roles of moderator and coordinator between the international and local levels. Non-governmental players are increasingly involved in decision-making and implemen-tation processes, this being the aim of the term “governance.” Governance is defined as the sum of all possible options, within which public and private institutions and organizations are constantly regulating their common affairs in order to coordinate their interests and facilitate cooperative action.4 Within this structure, there are active balances of power that are worthy of special attention.Along with the cross-border expansion of communications, transportation, and information systems, the intensification of economic relationships in the global market, and the internationalization of production and labor conditions, new forms of political regulation have arisen on a global scale. First attempts at a kind of global governance have been carried out by NATO, the IMF, the OECD, and the United Nations, more or less successfully. They act on an international level, wherever there is a lack of formal democratic institutions. They are the protagonists in post-conflict situations, since they are the organs of the international community, which is supposed to guarantee reconstruction, as well as integration into the global market and the international community of nations. They frequently replace missing state institutions, or participate to a large degree in their formation (nation building). Here, the international NGOs are important correctives. Evidence of this is the almost dramatic increase in the numbers of international NGOs that have begun operating in post-conflict situations in recent decades. However, their role within the current form of political regulation (policy regime) is also problematic and ought to be critically examined. Are they simply one component of a dominant political and regulatory system, or do they represent the interests of local civil societies?Non-governmental organizations are, according to the general definition, civil society organizations that do not represent the state. They are supposed to be financially and organizationally independent of state apparatuses and private business, and they should not be involved in the pursuit of commercial interests, but the interests of Social and political problems materialize in cities, but since every city’s population is confronted with these problems and must examine how they develop, there is also a chance to influence residents’ view of the future positively, in ways that go beyond the provision of basic necessities. The pivotal meaning of the city as an agent for political and social change becomes particularly obvious in a post-conflict situation. In this case, a city’s development takes place under mostly unsafe, unstable conditions, and it is essentially defenseless against the global mechanisms of political and economic events. This kind of city is scarred by conflicts; it is the expression of the crises in which the city finds itself.Cities that have to regenerate themselves after a conflict always have a recurring, similar problem: the exchange of population segments that occurs when many residents flee a city during periods of armed conflict. Often, many of these people do not return to their homes, for a variety of reasons: they may have better prospects elsewhere, or the political situation in their homeland may have altered so much that they are not interested in returning. Often, the conflicts have not been solved, but have instead become embedded in the city itself — the numerous divided cities around the world — from Belfast to Nicosia, Mostar to Beirut — are evidence of this. In addition, many people migrate to cities from rural areas. The result is that existing communities dissolve, and new neighborhoods are rapidly assembled. The people in them did not know each other previously, and have little to do with the city and its history. And in the case of rural immigrants, they are often not familiar with city life, which requires people to live closely together with many other different types of people. Living conditions are precarious, production has to be set up again, business has to develop, and jobs are scarce. Government institutions have to be rebuilt and do not function well; the political situation is unstable; corruption is every-where. Most crucial though, is the fact that, because of the rural immigrants and the return of refugees after a conflict, the population of a city will grow a great deal in a very short period of time. Cities like Prishtina or Kabul tripled or even quintupled their populations within a few years after military interventions. This results in an unregulated, informal boom in construction, since the demand for living space increases dramatically. Not only does the reconstruction of the city — the rebuilding of ruins — make its Politics of ArchitectureKai Vöckler1 See Sultan Barakat, ed., After the Conflict: Reconstruction and Development in the Aftermath of War, London, New York 2005; Gerd Junne, Willemijn Verkoren, eds., Postconflict Development: Meeting New Challenges, Boulder, London 2005; Kai Vöckler, Prishtina is Everywhere. Turbo Urbanism: the Aftermath of a Crisis, Amsterdam 2008; Jon Calame, Esther Charlesworth, Divided Cities: Belfast, Beirut, Jerusalem, Mostar, and Nicosia, Philadelphia 2009.2 See Regina Bittner, Wilfried Hackenbroich, Kai Vöckler, eds., UN Urbanism. Post-conflict cities Mostar Kabul, Berlin 2010; Linda Polman, War Games: The Story of Aid and War in Modern Times, London, New York 2010.3 See John E. Trent, Modernizing the United Nations System: Civil Society’s Role in Moving from International Relations to Global Governance, Farmington Hills, Opladen, 2007.4 Ulrich Brand, “Stadt als runder Tisch: Zum neuen Leitbild ‘Global’ oder ‘Good Governance,’” HYPER-LINK “http://www.bmgev.de/themen/urban21/brand” www.bmgev.de/themen/urban21/brand. See also Michael Zürn, “Regieren im Zeitalter der Denational-isierung,” in: Claus Leggewie, Richard Münch, eds., Politik im 21. Jahrhundert, Frankfurt am Main 2001.1617the community, instead.5 Ideally, the work of NGOs should conform to the following description:•theyrepresentintereststhathaveno voice in existing political structures (political advocacy)•theyidentifyproblemsandtopics,andinclude them in political negotiations and decision-making processes on both the local and international level (agenda setting)•theymobilizefactualexpertiseandknowledge in order to solve or debate topics in public•theydevelopprojectsthatarenotundertaken by state and supra-state players, or else carry them out when governments or the international community cannot do so for organizational or political reasonsTheir potential lies in their scientific, technical, or even political expertise, as well as in the fact that they are very familiar with the problems in their field and the local structures. This assumes, of course, that an international NGO cooperates with local, civil society organizations, so that they can successfully adapt their strategies to the local political, social, and cultural context. One important component in the success of their work is their ability to mobilize the public on both international and local levels, in order to further the goals of the interests formulated. If they can do this, they can make essential contributions to the processes of reaching the compromises and consen-sus necessary to political decision-making, and thus play an important role as mediators, since they are pursuing neither commercial nor wider political interests.Reality, however, has a somewhat different appearance: very few NGOs can finance themselves through donations, and therefore, they are highly dependent upon the financial support of governments, international governmental organizations (such as the EU), and their semi-public organizations. For financial support, the only other alternative is to turn to private foundations, such as the Aga Kahn’s, or George Soros’, or foundations financed by international corporations.6 Therefore, NGOs are forced to adapt their approaches, to accommodate the goals and programs set forth by their financial support-ers. In addition, professionalizing their work force also means paying their employees, which leads to a factual constraint: they have to compromise their own programs, in order to conform to the donor’s agenda. Over time, NGO “companies” have arisen, which are frequently only guided by self-interest. And the staffs of international NGOs are also part of the elite corps of managers and functionaries whom Mary Kaldor calls the “cosmopolitans” — a community of like-minded, well-educated experts, who work for the United Nations, one of the interna-tional or non-governmental organizations, or for foreign investors (who also count local experts as part of their membership). This, too, is a new form of global domination, which Kaldor describes as the “new divide.” On one side are the generally impecunious local residents and migrants, who are seeking a place in the new urban communi-ties, and on the other side are the global citizens and their allies — the local elite, most of whom have been educated abroad.7 Yet another difficulty is that NGOs specialize in specific topics, and this often prevents them from understanding an overarching complex of problems, which would require a correspondingly comprehensive policy. Last, but not least, is the NGO’s own, non-elected (and therefore undemocratic) position within the political negotiation process. In practice, all of this often results in a condescending, paternalistic way of working.Accordingly, the policies of NGOs operating in post-conflict situations should be critically examined, but they should also be allowed to develop further. An important step toward reinforcing their position enough to resist donors, gain independence from them, and to find their own place within the political regulatory system, is to strengthen their contacts with each other and organize more among themselves, so that, ultimately, they can become institutionalized on an interna-tional level and form a political counterweight. Independent, international collaborations and associations would then also help them to formulate comprehensive political goals. This also means, however, that they have to develop new forms of democratic politics. New, more transparent, and therefore more open ways must be found to select and represent topics and content, as well as to develop decision-making processes. It is always evident that an “intervention” from the outside is only successful when there is collaboration with local groups and support for political self-determination. This means that there must be more effort made to integrate non-governmental organizations and other civil society groups into the international system, and to increase their presence on the scene. The search for new ways must begin.Architects and planners as “agents of change”Architecture and urban planning only have disciplinary perspective when they consider themselves part of this overall political system — that is to say, of “governance.” They must also come up with a new planning strategy that can contextualize itself within various social situations and simultaneously prepare new contexts for participants in society.8 Should one employ the contextual-ization necessary in order to plan (for that, too, can be considered a political project), fundamental principles consequently arise. This is strategic, in the sense that it must react to various social and cultural contexts, while mediating between the special needs of individual social groups and the interna-tional power structure and its mechanisms. Yet it must also be cooperative, in as much as it provides, in its own turn, various participants with new contexts, in order to open up these spaces, whose effects go beyond local (and national) contexts. In this sense, planning must be communicative, because participants can only be mobilized through dialogue. This dialogue is not just local, however; it must also be conducted at an international level. Planners and architects have a seductive perspective of the city from above, which can block the view of how the city should be used, as well as the view of the social relationships and political power relations inscribed in its spaces. Space is not neutral. Accordingly, neither is the position of the architect or planner; there is no such thing as “architecture for architecture’s sake,” nor is there any such thing as a “neutral” plan. Every plan is the result of negotiation and power relations, and these things are expressed through the plan. When architects and planners regard themselves as “agents of change,” and try to intervene positively in city development, then they need to realize that they will need political “patronage.” This support, however, should not be based in opaque relationships and closed-door agreements; rather, it should be transparent and open. Only then would it also be possible to address the programs set forth by other political factions, and to include them in the negotiation process. This requires a delicate tightrope walk in the political arena, which has to be undertaken by all independent groups whenever they attempt to intervene in the urban space. On one hand, one of their essential tasks is to articulate the interests of excluded seg-ments of the population and to scrutinize the process of developing the city by mobilizing the public. On the other hand, they have to position themselves within the negotiation process, and, accordingly, be ready to cooperate or compromise with other political forces. In short, they have to have the ability to criticize the predominant political situation, while being part of it at the same time. This is because critique and public debate alone are not enough, even though they are perquisites for a successful intervention — for if sustainable change is indeed going to come about, then it is always brought about by assertive action; hence, it is a question of power. The regulation of city development cannot be left up to society’s self-regulation and customary law, especially not in post-conflict situations. Accordingly, it is directly linked to the ability of state organs to act assertively, because they have a monopoly on the legitimate use of force. However, no plan, no regulation, no urban development strategy can be carried out if there is no consensus in the society itself about the meaning and purpose of the plan. Yet, even here, limita-tions have to be set: when organs of the state (and the international institutions and organizations that support them) offer only repression and corruption, then they also have no legitimation. Even when they carry out measures through the threat of force, a critical distance must be maintained. But it is naïve to believe that an “architecture of peace” can exist independently of an “architecture of power.”Non-governmental organizations in post- conflict situations must adapt their approach to the local context and thus, only limited generalizations can be made. However, it is possible to propose some fundamental principles for discussion:•reflectuponandcommunicateone’s own position•donotpursueanycommercialinterests•addressdifferentinterestgroups,inorderto incorporate them over the long term•gathervariousintereststogetherforprojects that require groups to cooperate in order to carry them out•balancecollaborations;advantagesandprofits must be shared equally•createtransparency:goalsandproceduresshould be openly discussed in public•createroomfornegotiation,openupperspectives, mobilize knowledge and place it at everyone’s disposal•plananddesigntheprocess;structureitthrough dialogue and keep it openCorrespondingly, the three fundamental pillars of the work are: (1) communication, (2) cooperation, and (3) shaping the process. What does this mean, specifically, for architects and planners who want to intervene in urban development? They should think less about a project, and more about the process of conceiving projects, especially at first. And, as an aside, they should not assume that their project will necessarily wind up in a plan or as a building. As the saying goes, “when you have a hammer, every problem looks like a nail.” From the standpoint of the architect, every problem becomes a construction project. Frequently, however, an urban problem can also be solved by other means: a space can be revitalized through public debate, by using it for something else, or else through an economic strategy or regulation, to name just a few examples. That means, when figuring out projects in a post-conflict situation, it always helps to ask a few basic questions:•Where:whereisthespace?Wherearethere spaces that are not disputed and controversial, but “neutral”?•What:isitpossibletocreateanewspace,which will open up new perspectives?Politics of Architecture 5 See Ulrich Brand, Alex Demirovic, Christoph Görg, Joachim Hirsch, eds., Nichtregierungsorganisationen in der Transformation des Staates, Münster 2001. 6 Archis Interventions’ activities in Southeastern Europe have been mainly supported by ERSTE Foundation. ERSTE Foundation is the main shareholder of ERSTE Group (the savings bank). See HYPERLINK “http://www.erstestiftung.org” www.erstestiftung.org7 Mary Kaldor, “Cosmopolitanism Versus Nationalism: The New Divide?,” in: Richard Caplan, John Feffer, eds., Europe’s New Nationalism: States and Minorities in Conflict, New York, Oxford 1996.8 See Nikolaus Kuhnert, Anh-Linh Ngo, ‘Governance’, Archplus 173 (May 2005) C I T Yin post-conflict situationLocal employees‘Internationals’‘Internationals’WO R LDRefugee employees‘Locals’1819Even though urban development has always been closely connected to the real estate industry, it would be an act of historical hypocrisy to claim that the recent economic crisis is the only reason for bad urban planning. The crisis of modern planning became apparent much earlier and it has been going for decades. Even during the most heroic times of modernism, the crisis was always somewhere around the corner. Maybe the intrinsic failure of planning was hidden; it evaded discussion, because it testified to the crisis of the modern project as such. For modernist governments, architectural plans were always a represen-tational screen, but it became apparent by the end of the 1960s that they were only a blueprint masking the power and hegemony of the elites of modern society (regardless of whether the government’s program was socialist or capitalist). This inability of architects to see the development of their discipline from historical and political perspectives1 is probably the most important cause for present-day confusion about the architect’s role in society and (especially if he or she is practicing urban planning) his or her inability to understand the complex contingencies of today’s urban governance. In that sense, today’s economic crisis is pretty irrelevant.In the post-socialist Balkans, the lack of memory has a long, uninterrupted tradition and is not just particular to architects). After socialism, the collective amnesia and the culture of (re)constructing the past — along with ethnic fundamentalism, which pushed the states of ex-Yugoslavia to go their separate ways — continued in the same fashion as before. If they have not been completely erased from language, written history, or cultural production, traces of the socialist past are constantly under assault. Ironically, now, when it has become obvious that reinventing national identities has failed to produce livable social conditions, the absence of memory has become a cognitive obstacle to understanding today’s social reality.Instead of supporting a constructed and static notion of historical narratives, there should be a different approach. An active historiographical practice, a socially situated activity that is in dialogue and in discursive tension with both past and present, is more appropriate. It is a combination of some kind of forensic activity (forensic, because one has to scrutinize a body that is seemingly dead, i.e., socialist modernism) and observing the social chemistry of elements, which, after being separated, continued to function, albeit in a different way. This is precisely what is missing in today’s urban discourse and practice: an almost technical knowledge that could articulate the direct link between what is known as socialist urban planning and its contempo-rary remains. In making this link, it is inevitable that the technology of urban planning that be primarily dealt with. More than simply analyzing master plans and zoning principles, it would be better to scratch the surface of blueprints to scrutinize the language governing nomenclature and the institutional system. This text should be read as an invitation to acquire this forensic technique for both the recent socialist past and the contemporary, post-socialist present. It is also an oblique attempt to redirect the critical attention of architects to the terrains vagues of urban territory, which in odd ways, unite the failures of urbanism of both the socialist period and its post-socialist stage, opening up a view from the other side of modernity and urban control.Socialist urbanismAt the beginning of the socialist stage, Yugoslav architects and urban planners, unlike their Western colleagues, were not directly tied to market tendencies. Instead, they were highly dependent on political circumstances, because socialist modernism was “happily” married to state ideology. More than being an active agent of political and cultural critique, architecture was a representational screen for the system — an image of a successful, socialist model of a non-aligned Yugoslavia. Political nomenclature transmitted this image toward both sides of the Iron Curtain. Today we have to admit that the lack of professional independence actually had its good side. Until the beginning of the 1970s, the progressive tendencies of political and cultural reforms incubated architectural discourse and practice with fresh ideas based on the concept of social justice and models of governance through self-manage-ment. Therefore, the architecture of social utopia had its ideological foundations (particularly important for urban planning, as the abolition of private ownership and land expropriation enabled modernist urban schemes), while economic prosperity opened up space for outstanding architectural innovations. This is important, because governmental and non-governmental corporate forms that dealt with planning (university departments, governmental departments and institutes, self-managed companies — a socialist version of business enterprise, etc.) developed without historical antecedents (the profession of urban planning barely existed before World War II, so it was practically invented during the socialist era).One should be aware that there has always been a (creative) tension between profes-sional planners and politicians. Professional urban institutes at first operated as parts of municipalities, but during the 1960s and ‘70s they separated, claiming their independence from political decisions on both local and federal levels. Those independent entities provided services to municipalities and federal units, but tried to keep political influence outside of the planning studio. They invested efforts in continuous capacity building (education of employees, publishing books and magazines, and promoting new tendencies in urban planning — including models of civic participation). Unfortunately, the period did not last long enough to yield a coherent discourse and a tradition of independent urban planning. When the deficiencies of the Yugoslavian model started to show, architectural discourse began losing its reformatory capacity. After bureaucratic claims and demands took precedence in the 1980s, the culture of critique and moral authority was even more diminished, and their purposes frustrated to varying degrees. Aside from different forms of state control, market oriented tendencies started to weaken the architectural discourse and social agency as well. This is particularly important to understand, while at the same time connecting the socialist past with contemporary political and economic circumstances. The fact is that the self-man-aged, socialist society in Yugoslavia was a product of continuous changes and adjustments between the models of a state economy and a free market economy. The greatest shift — and probably the crucial phase — occurred after the big constitutional changes in the mid-1960s. Afterward, in the mid-1970s, during the late phase of socialist self-management, the patterns for organizing labor, the inflation of popular culture, the bureaucratization of govern-ment, and a strong shift toward the market, with its attendant development of politically influential techno-elites, created social Margins of ModernityMarko SančaninOn the Institutional Crisis in Post-socialist Urban Planning1 It was Manfredo Tafuri who first rendered this critique with a compelling clarity. In his Towards a Critique of Architectural Ideology (1969), he completely unmasked architectural modernism and literally accused architects of political backwardness and historical naivety.•Why:whoneedsit;whowillprofitfromit?Who are the partners involved in the process? Is it in the interest of the public welfare?•How:whattoolsandstrategiesareneeded,in order to create this space? How can collaborations be established, and how can the process be shaped?Flexible planningIn post-conflict situations, the obvious failure of city government and the international organizations, which are together responsi-ble for city development, can be traced back to a problematic understanding of planning and the role of the expert. Undeniably, plans are needed in order to structure and regulate the city’s development. However, in the traditional top-down approach that is part of hierarchically structured planning — from the strategic master plan, to land zoning, to regulated building plans — there are already a number of troublesome aspects, which are not really taken into consideration. The main problem is the extremely tedious process that often drags out over several years: by the time it comes to an end, reality has already outpaced it. Also, the city is regarded as a coherent form — an object to be shaped and designed. The fact that the city has an active, constantly changing form is ignored. Of course, it is necessary that there be goals in planning, and that planning be a regulated force in urban development, but it is an ongoing process that needs to be flexible whenever circumstances change. This is the paradox of urban planning: on one hand, it establishes and fixes things, but on the other, it also has to be flexible. According-ly, a plan needs to envision the vagaries of the future, and this can only happen when planning itself is regarded as a process. It has to take into consideration the fact that the city is more than just a form to be structured. Rather, it is the product of deeds performed by different actors, and among the protagonists is the city administration. In turn, the actors are guided by norms and values that can be used as orientation points. In this respect, the challenge for planners is to develop ways to cooperate with various actors — ways that are suited to the specific political and social circumstances. In this sense, planning is always political, too. If the traditional model of urban planning is rejected, then new forms of cooperation and processes of negotiation between private parties and governmental institutions need to be developed. Master plans that attempt to treat complex layers of political and economic problems in the same way (comprehensive planning) are of no significance. Instead, it is necessary to develop processual, participatory, and hence, communications-based types of plans (collaborative planning). For this kind of flexible planning, it is crucial that new ways of collaborating be permitted. However, these kinds of cooperative efforts should remain transparent, especially transparent enough to be monitored by the public. Therefore, the forces of civil society should be encouraged to participate; they need to be regarded as an important corrective to the planning process. The key to all of this is communication: public debate had to be stimulated, with the assistance of local media. Public campaigns should also stimulate active participation and provide opportunities for people to get involved. Within the conflict-ridden system of international politics (localized through governmental structures), a globalized market and civil society projects — developed in cooperation with local, independent urban initiatives — could develop new kinds of cooperative and communicative planning, which will help to support the society’s weak civil powers in particular. Thus, positioning within this new framework of governance is of crucial significance, because if an intervention is to succeed, then it has to develop relationships with government institutions and the civil sector alike. The independent, non-governmental, non- commercial initiatives formed by profes-sionals play an important part here, because not only do architects and urban planners bring their expertise to the table, but they frequently also act as both initiator and mediator, employing empowerment strategies to facilitate or influence urban developments. The article was previously published in Volume 26, 2010.2021conditions that comparable to post-Fordist society. Respectively, there was also a comparable influence on urban planning.In essence, there weren’t many differences. Architectural practice became a receptacle to be emptied and filled as party policy or the market dictated. The proof is that artistic and technological experiments appeared in western and ex-Yugoslavian architecture almost simultaneously, as acts of dissent toward reality, and replacing active social engagement. Urban planners were thus reduced to being some kind of highly skilled delivery boy, dealing with the technology of planning. This technology was closely connected to the instrumentals of planning protocols and language. It is possible to trace how the language of planning turned into a kind of meta-language that was applicable irrespective of context. The irony is that it actually became a kind of planning diglossia, which, in spite of its formal purity, had different meanings in different situations, respective of political or financial power.What is left over from socialist modernism are half-realized urban schemes that only mimic successful urban design. The notion that utopian urban schemes would produce utopian social conditions failed to live up to its promise. Our heritage is not modernity, but is only made up of physical objects that are tokens of modernity. In spite of astonishing architectural achievements, the urban environment is the product of modernized artifacts, without structural or institutional modernization. Architects and planners often show regret for unrealized competition entries or unfinished urban schemes, but what they really miss is the fulfillment of the big promise that architecture and planning would evolve in professionally independent and democratic practice. This lack of structural modernization and political culture, combined with market tendencies from the last stage of socialism, was the point of departure for post-socialist urban planning.TransitionTransience is a dynamic quality. To understand it you have to move along and observe the process, not its formal side effects. Even though the transitional 1990s brought about the whole repertoire of new forms of urban planning, economic or cultural and political shifts were far more important. It is also crucial to pay attention to what didn’t change. To some extent the 1990s were shockingly, violently new, but below the deck of the planning “machine,” some parts didn’t change at all. The transitional landscape of planning institutions is a bricolage of rusty parts that still (pretend to) function, and refurbished or new imported parts that (pretend to) function better. They operate in multiple layers and it is not always easy to tell one from another, especially if old nomenclature takes on a totally new agenda, or when old language takes on a different meaning that eludes interpreta-tion outside of a political or financial key.Ljubomir Bratić gave an excellent reading of institutional transformations in post socialism: “We can easily see continutity of institutions. Instutional bonds stayed connected but they changed direction of their sociopolitical, cultural and economic activity. Institutions became instruments that anybody can operate because the old instructions for use are lost and new vocabulary invented.”2 This anomie of professional vocabulary, the free interpre-tation of social norms and the respective diglossia of planning protocols greatly affected urban planning, because previous legal documents, like master plans and physical plans, became obsolete, while building permits were issued arbitrarily from case to case. The master plan stopped being an instrument for long-term public development strategy and became a dynamic map on which political and economic interests were projected during election periods.After going through lists of employees in urban departments and ministries, it is easy to conclude that technical and executive governance never really changed; however, you will also find old names in newly privatized companies. The fact is that highly ranked executives stayed the same and continued operating through a closed network of private alliances. What changed radically is their performance. Not only did ex-socialist urban bureaucrats become zealous advocates of neoliberalism, but they also formed an extra-institutional layer of governance with and within different forms of neoliberal power (investors, banks, media). The ideological arrow of post-socialist planning institutions has turned upside down: instead of social programs for professional support, urbanization, and infrastructural development for the benefit of the community, the tendency is to support short-term, feasible projects that serve particular interests.3 Today’s concept of urban governance operates as an odd symbiosis of old forms of management under the control of state politics and the global financial market. Zagreb is a good example, because in the last ten years big investments were made in the city and city governance was characterized by financial-ization (the city issued short-term bonds, indebting its citizens without giving them the opportunity to vote on the issue) and the corporatization of public services (all city services are united under one corporation which, while trying to run city as a business enterprise, is becoming less dependent on public opinion and more on the shares that international banks have in urban development business). The city is hardly guided by a coherent vision or long-term strategy, under legal scrutiny or civic participa-tion. Therefore, newly planned and built investment schemes do not result in social change, but only add to more of the same.Instead of being surprised, we should finally understand that the nomenclature flipped its agenda so easily, and the language of planning became the instrument of particular interests, because the tendency was already inherent in the concepts of socialist modernism. None of this would have gone as smoothly if there had not been structural weaknesses in planning technology, which were already apparent forty years ago. Somehow it all comes back to the instrumental rationality of modern society and the fact that urban plan was nothing more than an organizational pattern employed by the state or free market capitalism.But what about the terrains vagues of urbaniza- tion that resisted the orthodoxies of urban control and accumulated urban and cultural paradoxes instead? In a time of the general dissolution of the architectural discourse and obvious planning mistakes, it might prove useful at least to pay attention to urban territory that covers a much bigger urban area and is commonly labeled as suburban, marginal, or peripheral. Is there anything to learn from the margins of the modernist project?Margins of ModernityUntil the 1990s, suburbanization of Split was mainly due to industrialization. After the big political changes in the 1990s the city became a migration hub for political and economic refugees from Bosnia. The city perimeter exploded, resulting in an unprecedented encounter between the informal building enterprises of migrants and small building companies. A parallel city (without legal status or basic communal standards) has grown up on land zoned for agriculture. In the absence of public, educational, and cultural facilities — apart from concrete cathedrals, which testified to a renaissance of traditional values — every-thing was improvised: from DIY sewage systems to garages turned into kindergartens. Project Supeprivate did a comparative analysis between the dwelling habits and communal practices of the tenants of socialist multi-family apartment buildings and the new suburban inhabitants.Kozari Bok is one of the biggest and oldest informal settlements situated in Zagreb’s eastern industrial zone. Typical of abrupt industrialization and urbanization of the 1950s and ‘60s, informal settlements inhabited by workers started to cluster in the vicinity of factories. In 2004 Platforma 9,81 – Institute for Research in Architecture started a project titled Superprivate with the aim of investigating the housing and lifestyles of suburban areas in post-socialist cities.Although the Keneta neighborhood (formerly a swamp in the Adriatic coastal city of Durres, Albania) is built on a regular land division, drainage canals were the basis for the subdivision of the land, so a serious lack of the most basic communal standards (water, sewage, and electricity) makes housing conditions barely livable.2 Ljubomir Bratić, “On the Transformation of the Elite at European East” http://transform.eipcp.net/transversal/0208/bratic/sr3 Once I was at a meeting with Zagreb’s chief urban development strategist. I was surprised that he didn’t have a big master plan on the wall, which was pretty common during socialist times. Instead, he had a huge, blank topographic map of the city featuring the fabric of construction. The map was covered in yellow stickers with the telephone numbers of potential investors. Later I realized that, within the logic of short-term development, when a city is struggling to pay off its debts, it is more useful to know about the appetite of potential investors than about urban regulations.2223Margins of modernityFirst of all, the mechanisms of development on the urban periphery changed in a way that differed from that of the representation-al showcase of the modern city center. To be more accurate, it changed slowly and/or under different dynamics. It changed slowly because, since, aside from the interests of the cultural, financial, or political elites, nobody could care less what is happening there. Therefore, the periphery rarely becomes part of any grand scheme of revitalization or improvement (unless improvement means demolishing the periphery to build expansive offices and apartment buildings).When the periphery did change rapidly (illegal building boom and similar phenomena), its economy still did not change. Small investments in private homes and family enterprise were never intended to build according to modern urban standards, just housing. Therefore, the main objection to periphery — that it is not a proper city — mainly is due to the lack of communal amenities and infrastructure like schools, kindergarten, or sewage disposal systems.In comparison with the modern city, the periphery does looks like urban chaos. Very often it is illegal on paper. Its cultural authenticity is based on clichés, which still resonate soundly with the ethnic-religious fundamentalism of the 1990s. All in all, seeing it from the outside, through a modernist lens, it is a form of deep regression. However, one should be cautious about labeling marginality, especially from the Balkan perspective, since the colonial regimes of knowledge within which such entities as “the West,” “the Balkans,” or the Second World are also the product of a modern reading of history. Paraphrasing Chakrabarty on history, modernity remains the sovereign, theoretical subject of all cultures. There is a peculiar way in which all other cultures tend to become variations on a master narrative that could be called “the western modernity.” In this sense, any peripheral modernity is in a position of subalternity. One can only articulate subaltern-peripheral modernity in the name of Modernity itself.4Ironically, maybe the most interesting phenomena of today’s urbanity are far removed from the interests of professional urbanists and their elite institutions. It is at the terrains vagues of urban planning — in neighborhoods of low- and mid-rise, human scale, resilient, and endlessly adaptable semi-legal housing — where the greatest contradictions of any planned city always accumulate. At the margins of Croatian, Serbian, or Albanian cities there are material testaments that urbanization as a planned and controlled activity does have its twin reality.However the freshly improvised vitality of these neighborhoods is interesting primarily for its social potential, not for its formal inventions, which are exhibited all over art galleries. It is the capillary social power and potential for communal micro-policies that make them extremely valuable, especially because there already are examples of grassroots civil initiatives that operate as micro-political governing bodies. Out of a natural solidarity, based on a combination of existential interests, neighborly love, and cultural clichés, inhabitants of those areas gather to build sewage systems and erect garages for community homes. In a time of complete erosion of the political culture in the center, this could potentially mean the rebirth of authentic, direct democracy on the periphery.Maybe what we are witnessing is not a kind of delayed, undeveloped, or unfinished modernity, but rather, the fact that urban margins are becoming examples of another kind of modernity that should be considered an integral part of international modernity per se, because it is most inherent in it. Most cities in Southeastern Europe (SEE) follow the global trends of urban planning, a fact-based, top-down urban development strategy. Using projects done in collabora-tion with Archis Interventions and local planners for Prishtina in Kosovo and Bucharest in Romania as examples, I would like to show an alternative path to urban planning: UWYH - Use What You Have - Urbanism of Opportunities.Today increasing urbanization and changing social practices are addressed in two separate urban discourses. The existing dichotomy between careful, socio-political, urban approaches versus fact-based, economically driven urban developments has widened over the last decade. The urgent need to house millions of people around the world has led to replications of simple urban models and global typologies. Shopping malls, office towers, airports, and repetitive types of housing are the elements of new or expanding cities, regardless of their actual problems and potentials. While urban planners, architects, and artists have rediscovered the city through small- scale interventions and local participation, large-scale urban developments are incapable of incorporating qualities that characterize urban life. Alternative culture spaces, creative spaces, nondescript spaces, flexible and shared spaces are required, in order to generate urban qualities beyond utilitarian needs.UWYH - Use What You Have explores the potential of associating contemporary urban planning strategies with local conditions and participants. Top-down planning strategies need to be complemented by local forces, issues, and activities that will cope with the contemporary urban problems that manifest under our complex social and political conditions. Conversely, bottom-up strategies need to be framed so that they can inform and guide urban planning approaches and provide sustainable improvement to our cities. The question today is not a technical matter of how to “build” a city — it is a cultural, social, and political question of how to structure urban life.The top-down planning approach had a long tradition prior to the founding of C.I.A.M.,1 yet the work of C.I.A.M. laid the foundation for a fact-based, rational approach to urban planning, resulting in a kind of almost Fordist production of cities. Radical neglect and the substitution of different social practices have triggered a counter move by Team 10:2 adopting daily activities as paradigms for urban and architectural developments. Team 10’s practice of urbanism and architecture has followed “as found” design strategies, being more rooted in actual urban and architectural circumstances than Modernist movements are. Many contemporary, bottom-up projects can be traced back to Team 10 and their thinking. UWYH incorpo-rates this bottom-up approach and extrapo-lates the planning process by integrating top-down planning strategies.UWYH’s urban planning discourse consists of different combinations of key strategies employed in the project for Prishtina, Kosovo, and Bucharest, Romania. The following five strategies frame the intentions and goals of the discourse.Strategic Typologies/Zonesconcentrate urban transformation in specific zones or typologies that will have a catalytic effect on the surrounding urban environment.Strategic Planning employs established planning strategies with social, political, or cultural goals, including utilitarian needs, but also going beyond them.Strategic Participation employs local inhabitants as urban produc-ers, critics, and consumers. They are the engines of urbanity and need urban planners, architects, or designers to provide spaces in which to realize opportunities. Strategic Legislation takes advantage of ways to change, amend, or add laws and regulations to achieve social, political, or cultural goals.Strategic Accumulation allows a bottom-up process, where a number of events, activities or spaces amass, either over time or in different locations, and create a critical momentum to make the project successful.These five strategies provide the core of action for “UWYH - Use What You Have.” For any of these strategies to be successful, the project needs a precise urban, social, cultural and political analysis in order to identify problems, potential, and the forces driving the particular urban situation.Legalization of Informal Buildings, Prishtina (Kosovo) 2006 – 20113Production of informal buildings since 1999 has led to excessive urban problems in Prishtina. Today, more than 75% of the buildings in Prishtina have been built without any type of permit. This bottom-up building production has been paralleled by the top-down design of regulatory plans for the main areas of Prishtina. These plans only incorporate legal buildings and were obsolete by the time they were drawn. After the war in 1999 the private sector catered to the housing shortage by constructing informal buildings, while the city of Prishtina has stimulated the development of global typologies through a strategic plan, which incorporates a high-rise zone for offices, hotels, and shopping malls. While this top-down planning strategy was underway, we started our work to legalize the informal buildings in 2005, with the goal of bridging the gap between the city’s top-down urban developments and the inhabitants’ bottom-up construction projects.On our first encounter with the urban system of Prishtina, it became evident that most buildings did not conform to existing building regulations. Under these regulations, most buildings had to be demolished or substantial-ly rebuilt, turning most parts of the city into a blank slate. Surely, destroying capital and burdening the majority of inhabitants could not be the solution. Instead, we decided to use what the city had and explore the opportuni-ties provided by the existing situation.The strategies we employed included strategic legalization as the main incentive; strategic typologies comprising an analytical framework and basis for strategic legalization; strategic participation as a tool for communication and inclusion; and UWYH - Use What You HaveWilfried HackenbroichTowards an Urbanism of Opportunities 4 Dispesh Chakrabarty, Provincializing Europe: Postcolonial Thought and Historical Difference (Princeton: Princeton University Press 2000, p. 27).1 Congrès internationaux d’architecture moderne – C.I.A.M. (International Congresses of Modern Architec-ture). From 1928 to 1959 this group of architects and urban planners conducted conferences around the world to develop and explore the Modernist movement.2 Team 10 was a group of architects responsible for the organization of the 10th th C.I.A.M. congress in 1956. The members of this group had already taken a critical position toward C.I.A.M. at the 9th conference and initiated the dissolution of C.I.A.M. In the following years the group expanded and is known for its alternative urban and architectural approach.3 Team: Archis Interventions / Prishtina (Florina Jerliu, Visar Geci, Vlora Navakazi), Archis Interventions/SEE (Kai Vöckler, Thilo Fuchs), Hackenbroich Architekten (Wilfried Hackenbroich)2425finally, strategic planning, to provide a direction for development in the city of Prishtina and to integrate legalized buildings into the official urban planning process.To understand Prishtina’s urban system, we employed strategic typologies as an analytical tool to identify prototypical problems of the informal buildings and to develop a framework for adequate action. We singled out major typologies, such as the informal suburban housing built accord-ing to building regulations, but without construction permits with a total lack of public infrastructure. Inner city buildings exceeded by far the maximum floor area, had limited fire escape routes, were set too close to neighboring buildings, in many cases, the buildings covered more than ninety percent of the plot. Instead of paring back the size of the buildings, we strove for a regulatory solution that would ensure security and quality of life, with minimal building alterations. Informal roof extensions on socialist housing blocks also exceeded the maximum permissible floor area, but even more cleverly, exceeded the structural capacity of the original building. As far as this type of building was concerned, it was evident that safety and structure had to be re-evaluated. The typological approach enabled us to understand the given situation, as well as the principles required in order to develop initial solutions to the major problems and define characteristics that ought to be preserved.In order to legalize the informal housing stock in Prishtina, we had to include strategic legislation and develop legal guidelines for solving the existing urban and building problems. In our Manual for Legalization we developed simplified building regulations with graphics of the requirements for the building. Legalisation provided benefits and incurred obligations for both the owner and the city, and also made it possible to address urban issues. We developed a simple equation: through legalization, owners were assured of their ownership and had to pay property and land taxes, while the city garnered more tax revenues and had to provide adequate public infrastructure. The manual became the basis for the legalization law in 2010 and is now officially on the books.Implementing the manual meant more than dealing with the legal aspects. To make this transformation successful, a broad swath of stakeholders in the city had to be persuaded to accept it. We used strategic participation to generate two-way communication involving local authorities, professionals, academics, property owners, and the general public. Tools for communi-cation included expanding building capacity through student projects and workshops; professional awareness through an international conference, public awareness through newspapers and other publications; nationwide awareness through a TV show produced exclusively to support the legalization process and increase local participation through public discussions, and workshops with the various owners.Legalizing the informal buildings solved the problems generated by the bottom-up building activities. Nevertheless, these buildings were still disconnected from the top-down urban planning of the city’s strategic plan and regulatory plans. Since the legalization process retained most of the informal buildings, the regulatory plans needed substantial revision, in order to integrate the buildings constructed in the previous dozen years. By making the informal buildings part of the future planning, legalization actually turned informal building activities into a form of participatory urbanism.The regulatory plans were developed separately, and the strategic plan did not provide an urban and economic vision for the city, which made the urban planning of Prishtina questionable. To ensure that the regulatory plans were revised to include effective and sustainable strategies from the bottom up and top down, it became necessary to apply strategic planning and define the direction of city development on economic, social, urban, and cultural levels.Strategic planning was our last project phase and again we applied the logic of UWYH by using what was already planned and built, as well as understanding what was really needed to stimulate the city and make the existing planning relevant.The city of Prishtina has little potential to generate future development. The city is of little value to tourism, and has only a few international businesses, mostly regional trade, and a high unemployment rate. Still, sixty percent of the population in Prishtina is under the age of 25, and combined with the established state university and the increasing number of private universities, there is evidently huge potential for the city’s future. Additionally, as the capital of Kosovo, Prishtina stands for national identity and history, important aspects essential to future development.In order to relate the main issues of youth, history, national identity, education, and economy in a dynamic way, we developed three strategic typologies. These typologies are placed at crucial points in the inner city to form a network of otherwise unrelated places. Programmatically, these typologies facilitate exchanges and relationships among youth, history, national identity, education, and economy. To achieve this kind of exchange, the main incentive must be to support non-academic education and youth entrepreneurship. Typologies included housing for young entrepreneurs and students with minimally equipped apartments and collective functions; the incubator, enabling entrepreneurship by offering small office spaces, information services, collective office functions, and financial support; and finally, the public community center, offering education outside the academic realm via a news and information center, along with a café and a small exhibition and performance space. In terms of space, these typologies are points of attraction and enhance exchanges in the city center, while at the same time forming a bridge to neighboring districts of the city. These three typologies work on social, educational, cultural, and economic levels to bring the city’s youth into a position where they can participate in and contribute to the economic development of the city and stimulate growth of local businesses and investments, independent of global capital. This growth will eventually make top-down planning of global typologies valuable to local business, and therefore it will also become sustainable.Throughout the different stages of the project we synchronised and empowered independent urban and economic develop-ments in the city. By following the dis-course of UWYH, we needed only minimal investments to achieve maximum effects.Magic Blocks – Behind the Concrete CurtainBucharest, Romania 2009–20104The urban fabric of Bucharest is strongly influenced by the socialist housing block, as are the majority of post-socialist cities. But in Bucharest’s case, the housing blocks were not only built on a kind of tabula rasa, they were also built along the existing boulevards. Apartment blocks, eight to twelve stories high, entirely transformed the existing fabric of one- to two-story houses. The slab buildings were like a concrete curtain, framing large areas of low-density housing and giving the boulevards a dense, urban appearance, while protecting the green, suburban setting behind them.Today, socialist collective housing is privatized and low-density housing is intact and highly valued. In the process of privatization, only individual apartments changed ownership, while the land remained city property, leaving a no-man’s-land of public space between the buildings. The city of Bucharest still is incapable of maintaining or developing this no-man’s-land. The devastation of these areas is accelerated by the unrelated juxtaposition of high-rise apartments and low- density housing. Bucharest has focused its attention on planning the city center from the top-down, major infrastructure, and new commercial centers involving global types of structures, such as high-rise offices, shopping malls, and hotels. In 2007, after the global financial crisis, it became evident that this development had tied the city to global forces, leaving the existing local potential behind. When we started our investigation in 2007, in cooperation with Archis Interven-tions and local planners, the city ran a façade renovation program, offering thermal insulation and plastic windows for socialist housing blocks. Minimal improvements had been made along the boulevards, and private property owners had renovated their houses. The public space behind the concrete curtains remained untouched and derelict. A substantial part of Bucharest suffers from this condition, and we decided to develop a prototypical solution for this condition along the Calea Mosilor.Within the context of UWYH we used strategic typologies to identify prototypical situations; we also developed strategic zones to frame strategic legislation to address development problems, strategic planning to reconfigure and reprogram the abandoned spaces, and strategic participation for research, communication, and initiation. Our goal was to go beyond the simple beautification of public space in order to reactivate the abandoned areas behind the slab buildings. Through strategic typologies, we identified particular sets of problems for five prototypi-cal situations. Besides the individual solutions each of them needed, a common factor in all of the areas was that private properties facing the residual space were underdeveloped. Consequently, we devel-oped strategic zones of two kinds: an intervention zone, where planning was needed; and a buffer zone, where private investment was necessary. Private invest-ment was stimulated through strategic legislation, allowing higher building density in the buffer zone. The intervention zone was transformed to improve public spaces for the apartments and to provide high- quality conditions for private investments. Looking at the buffer zone in relation to the entire city, it has the potential to become a huge development area. Its proximity to Bucharest’s main roads ensures good accessibility to the buffer zone from most of the city. Furthermore, an active buffer zone absorbs the existing development pressure in low-density areas, providing them with protection from existing and destructive building activities.In order to resolve the various problems of each prototypical situation, we used strategic planning to reorganize and reprogram the open spaces and usages in the intervention zone. We introduced distinctions between public and semi-public spaces, reorganizing existing activities and injecting new activities. Shortcomings, such as neighborhood parking problems, lack of social infrastructure, limited accessibility, and dilapidated public spaces, were also addressed in the planning, leading to economically interesting possibilities for developments in the buffer zone.In the next step of the project, we used strategic participation to raise awareness in the professional field by producing an exhibition. To address the local public in the neighborhood and to test our strategic planning, we realized a set of small-scale interventions, which explored re-programming and re-organizing abandoned spaces. With minimal means, we were able to engage the inhabitants, received valuable feedback, and paved the way for a prototypical project, so that we could realize our strategies on a larger scale. This step has yet to be taken, but it should trigger a transformation of the concrete curtains throughout Bucharest.The UWYH discourse is not a substitute for existing urban planning strategies; it is a guiding principle for employing strategies and tools as efficiently as possible, offering a much-needed expansion of existing practice. UWYH connects projects to local conditions and solves gridlock problems. It stimulates local activity, so that it can be effective in the local and global contexts alike.While top-down projects demand extensive financial resources and political power to position them in the market, the UWYH strategies are effective with minimal resources. Acceptance by users and markets is far higher, and stimulation of the local market leads to sustainable, lasting develop-ments. The combination of top-down and bottom-up urban strategies maximizes benefits for the city. Global capital, interna-tional business, and international urban practices can be incorporated into the city, while promoting local participants and businesses. When it comes to building urban environments, UWYH represents a crucial discourse.4 Team: Zeppelin (Cosmina Goagea, Constantin Goagea, Ştefan Ghenciulescu), Point 4 (Justin Baroncea, Carmen Popescu), Archis Interventions/ SEE (Kai Vöckler), Hackenbroich Architekten (Wilfried Hackenbroich)UWYH – Use What You HaveLitERSTE Foundation SeriesVolume 2SEE! Urban Transformation in Southeastern EuropeEdited by Kai VöcklerLjubljanaBihačSplitSarajevoTiranaPrishtinaNovi SadSofiaBucharestSkopjeIstanbulNicosiaZagrebMostarKotorBelgradeVlaadičin HanPulaLitERSTE Foundation SeriesVolume 2SEE! Urban Transformation in Southeastern EuropeEdited by Kai VöcklerLjubljanaBihačSplitSarajevoTiranaPrishtinaNovi SadSofiaBucharestSkopjeIstanbulNicosiaZagrebMostarKotorBelgradeVlaadičin HanPulaLitERSTE Foundation SeriesVolume 2SEE! Urban Transformation in Southeastern EuropeEdited by Kai VöcklerLjubljanaBihačSplitSarajevoTiranaPrishtinaNovi SadSofiaBucharestSkopjeIstanbulNicosiaZagrebMostarKotorBelgradeVlaadičin HanPulaLitERSTE Foundation SeriesVolume 2SEE! Urban Transformation in Southeastern EuropeEdited by Kai VöcklerLjubljanaBihačSplitSarajevoTiranaPrishtinaNovi SadSofiaBucharestSkopjeIstanbulNicosiaZagrebMostarKotorBelgradeVlaadičin HanPula | RESEARCH GATE |
1RIScatter: Unifying Backscatter Communicationand Reconfigurable Intelligent SurfaceYang Zhao, Member, IEEE, and Bruno Clerckx, Fellow, IEEEAbstract—Backscatter Communication (BackCom) nodes har-vest energy from and modulate information over an externalelectromagnetic wave. Reconfigurable Intelligent Surface (RIS)adapts its phase shift response to enhance or attenuate channelstrength in specific directions. In this paper, we show howthose two seemingly different technologies (and their derivatives)can be unified to leverage their benefits simultaneously into asingle architecture called RIScatter. RIScatter consists of multipledispersed or co-located scatter nodes, whose reflection statescan be adapted to partially engineer the wireless channel ofthe existing link and partially modulate their own informationonto the scattered wave. This contrasts with BackCom (resp.RIS) where the reflection pattern is exclusively a function of theinformation symbol (resp. Channel State Information (CSI)). Thekey principle in RIScatter is to render the probability distributionof reflection states (i.e., backscatter channel input) as a joint func-tion of the information source, CSI, and Quality of Service (QoS)of the coexisting active primary and passive backscatter links.This enables RIScatter to softly bridge, generalize, and outper-form BackCom and RIS; boil down to either under specific inputdistribution; or evolve in a mixed form for heterogeneous trafficcontrol and universal hardware design. For a single-user multi-node RIScatter network, we characterize the achievable primary-(total-)backscatter rate region by optimizing the input distribu-tion at the nodes, the active beamforming at the Access Point(AP), and the backscatter detection regions at the user. Simulationresults demonstrate RIScatter nodes can exploit the additionalpropagation paths to smoothly transition between backscattermodulation and passive beamforming, and the proposed practicalreceiver accommodates the double fading, double modulation,and symbol period difference of active-passive coexisting net-works without Successive Interference Cancellation (SIC).IndexTerms—Backscattercommunication,reconfigurableintelligentsurface,ambientbackscattercommunication,symbiotic radio.I. INTRODUCTIONA. FundamentalsFUTURE wireless network is envisioned to provide highthroughput, uniform coverage, pervasive connectivity,heterogeneous control, and cognitive intelligence for trillionsof portable devices. As a mature low-power communicationtechnique, Backscatter Communication (BackCom) separates aconventional transmitter into a Radio-Frequency (RF) carrieremitter with power-hungry elements (e.g., synthesizer andamplifier) and an information-bearing node with power-efficientcomponents (e.g., harvester and modulator) [1]. In particular,the node harvests energy from impinging wave and embedsThe authors are with the Department of Electrical and ElectronicEngineering, Imperial College London, London SW7 2AZ, U.K. (e-mail:{yang.zhao18, b.clerckx}@imperial.ac.uk). B. Clerckx is also with SiliconAustria Labs (SAL), Graz A-8010, Austria.This article has been submitted for potential publications.information over scattered signal, and the backscatter readercan be either co-located or separated with the carrier emitter,known as Monostatic BackCom (MBC) in Fig. 1(a) and BistaticBackCom (BBC) in Fig. 1(b). Relevant applications such asRadio-Frequency Identification (RFID) [2], [3] and passivesensor network [4], [5] have been extensively researched,standardized, and commercialized in the era of Internet ofEverything (IoE). However, conventional BackCom nodes onlyrespond when externally inquired by a nearby reader. Totackle this, [6] proposed Ambient Backscatter Communication(AmBC) where battery-free nodes recycle ambient signals(e.g., radio, television and Wi-Fi) to harvest energy andestablish connection in between. As shown in Fig. 1(c), iteliminates the need of dedicated power supply, carrier emitterand frequency spectrum for BackCom, but the decoding issubject to a strong interference from the primary (legacy) link.Cooperative AmBC was then proposed where a co-locatedreceiver decodes both active primary and passive backscatterlinks, and the rate and error performance are evaluated forvarious detection schemes [7]. The concept was further refinedby [8] as Symbiotic Radio (SR) in Fig. 1(d) that cognitivelyincorporates AmBC with existing systems. In a SR system,the active transmitter generates RF wave carrying primaryinformation, the passive node creates additional propagationpath and ride its information over scattered signal, whilethe cooperative receiver jointly, individually, or sequentiallydecodes both links. Those BackCom applications employ scatternodes as pure information sources, and the active primarytransmission (if it exists) can be influenced by the randomnessof backscatter modulation. On the other hand, ReconfigurableIntelligent Surface (RIS) in Fig. 1(e) is a smart planar signalreflector that consists of numerous compact passive elementswith adjustable amplitude and phase responses, whose reflectionpattern over time is deterministic and adaptively optimizedbased on Channel State Information (CSI). RIS smartly reflectsthe surrounding RF waves to customize the wireless propagationenvironment for signal enhancement, interference suppression,scattering enrichment and non-line-of-sight bypassing [9]. Insummary, MBC/BBC (resp. RIS) involves a single passivebackscatter (resp. active primary) link where the passive devicemodulates information (resp. enhances propagation) and theactive device generates carrier wave (resp. modulated signal).On the other hand, AmBC (resp. SR) involves coexisting activeprimary and passive backscatter links in a competitive (resp.collaborative) relationship. In AmBC, the backscatter link onlyneeds to know the basic information (e.g., operating frequency)of the primary link, while the primary link may be completelyunaware of the backscatter link. In SR, relevant CSI is exploitedarXiv:2212.09121v1 [cs.IT] 18 Dec 20222TABLE ICOMPARISON OF SCATTERING APPLICATIONSMBC/BBCAmBCSRRISRIScatterInformation link(s)Passive backscatterActive-passive co-existing (competitive)Active-passive co-existing (collaborative)Active primaryActive-passive co-existing (collaborative)Primary impact onbackscatter detectionCarrierMultiplicativeinterferenceSpreading code or mul-tiplicative interference—Energy uncertaintyBackscatter impacton primary detection—MultiplicativeinterferenceCSI uncertainty or mul-tiplicative interferencePassive beamformingDynamicpassive beamformingCollaborativedevices based on CSI——Co-locatedreceiver (potentiallyprimary transmitter)—Primary transmitter,scatter nodes, andco-located receiverCooperativedecoding strategy—NoUsually SIC fromprimary to backscatter—No SIC, backscatterdetection as part of pri-mary channel trainingPrimary detection—Semi-coherentSemi-coherentCoherentCoherentBackscatter detectionCoherentSemi-coherentCoherent—Semi-coherentReflection patternover time depends onInformation sourceInformation sourceInformation sourceCSIInformationsource, CSI, and QoSReflectionstate (i.e., backscatterinput) distributionEquiprobableEquiprobableEquiprobableor GaussianDegenerateFlexibleLoad-switching speedFastFast or slowVery slowQuasi-staticArbitraryInterrogatorScatterNodeCarrier100(a) MBCCarrierEmitterScatterNodeBackscatterReaderCarrier100(b) BBCPrimaryTransmitterBackscatterReaderScatterNodePrimaryReceiverModulatedDoubly-Modulated(c) AmBCPrimaryTransmitterScatterNodeCo-locatedReceiverModulatedDoubly-Modulated(d) SRTransmitterReflectElementReceiverModulated(e) RISPrimaryTransmitterRIScatterNodeCo-locatedReceiverModulatedFlexible(f) RIScatterFig. 1.Illustration of scattering applications. For each, the blue flow(s)constitute the active primary link while the magenta/green flow denotes thepassive backscatter link.at the co-located receiver for cooperative decoding, and mayalso be exploited at the primary transmitter for cooperativeactive beamforming. A comparison of scattering applicationsis summarized in Table I.B. Related WorksSimilar to Cognitive Radio (CR), active-passive coexistingnetworks can be classified into commensal, parasitic, andcompetitive relationships, whose instantaneous rates, powerschemes, and outage probabilities were acquired in [10],[11]. Joint Maximum-Likelihood (ML) decoding provides thebest error performance but requires prohibitive computationalcomplexity especially for high-order modulations [7]. Indeed,active-passive coexisting networks have three special andimportant properties: 1) primary and backscatter symbols aresuperimposed by double modulation (i.e., multiplication cod-ing); 2) backscatter signal strength is much weaker than primarydue to the double fading effect; 3) backscatter symbol period islonger than primary due to the load switching constraint. Thesecond property motivated [7], [8], [10]–[18] to view SR as amultiplicative Non-Orthogonal Multiple Access (NOMA) andperform SIC from primary to backscatter link. During primarydecoding, the randomness of backscatter modulation can bemodelled as channel uncertainty or multiplicative interferencewhen the backscatter-over-primary symbol period ratio N islarge or small1. Such a SIC receiver is operation-intensive andCSI-sensitive because it involves N re-encoding, precoding,subtraction together with a time-domain Maximal Ratio Com-bining (MRC) to decode each backscatter symbol. When Nis sufficiently large (i.e., commensal SR), the average primaryachievable rate under semi-coherent detection2 asymptoticallyapproaches its coherent counterpart, and both links may bedecoded in an interference-free manner [12]. On the other hand,1For state-of-the-art low-power RFID tags, the load-switching speed usuallyvaries between 100s of kHz to 10s of MHz [19], corresponding to a typicalN between 10 and 1000.2In this paper, semi-coherent detection refers to the primary/backscatterdecoding based on known CSI and unknown backscatter/primary symbols.3efficient backscatter multiple access design remains an open is-sue for BackCom. [15] proposed a NOMA-based SR where theSIC order depends on the backscatter channel strength, but theperformance deteriorates fast as the number of nodes increases.Time-Division Multiple Access (TDMA)-based SR was alsoconsidered in [16] where each node transmits information dur-ing its dedicated slot and harvests energy otherwise. It enablesadaptive transmission time and reflection ratio design but intro-duces extra coordination cost. [20] modifies the load-switchingspeed to shift the scattered signal to the desired frequency band.This enables backscatter Frequency-Division Multiple Access(FDMA) at the cost of larger bandwidth and higher powerconsumption at the scatter nodes. To reduce the coordinationcost, [17] proposed a random code-assisted multiple accessfor SR and evaluated the asymptotic Signal-to-Interference-plus-Noise Ratio (SINR) using random matrix theory. Thiscode-domain solution suffers from the near-far problem andimperfect synchronization, both being severe in BackCom.Conventional RIS design with fixed reflection coefficientsduring each channel block has been extensively studied in com-munication, sensing, and power literatures [21]–[26]. On theother hand, dynamic RIS with independent reflection patternsin different time slots has gained recent attentions in multi-user and multi-purpose wireless networks. The concept wasfirst proposed in [27] to fine-tune the Orthogonal Frequency-Division Multiplexing (OFDM) resource blocks, then extendedto the downlink power and uplink information phases ofWireless Powered Communication Network (WPCN) [28]–[30].It introduces artificial channel diversity and enables flexibleresource allocation, but misses the opportunity to encode itsown message. RIS can also be used as a transmitter whenplaced in the near field of a carrier emitter, and prototypeshave been implemented for Phase Shift Keying (PSK) [31]and Quadrature Amplitude Modulation (QAM) [32]. Froman information-theoretic perspective, [33] reported that usingRIS as a naive passive beamformer to maximize the receiveSignal-to-Noise Ratio (SNR) is generally rate-suboptimal forfinite-input constellations. Instead, the capacity of RIS-aidedchannel is achieved by joint transmitter-RIS encoding, and SIC-based SR can outperform passive beamforming at high SNR.It inspired RIS-empowered BackCom [34]–[43] to combinepassive beamforming and backscatter modulation in the overallreflection pattern. In particular, symbol level precoding mapsthe information symbols to the optimized RIS coefficient sets[34], [35], overlay modulation superposes the informationsymbols over a common auxiliary matrix [36]–[39], spatialmodulation switches between the reflection coefficient setsthat maximize SNR at different receive antennas [40]–[42],and index modulation employs dedicated reflection elements(resp. information elements) for passive beamforming (resp.backscatter modulation) [43]. However, those joint designsincur advanced hardware architecture and high optimizationcomplexity. Relevant literatures also consider either Gaussiancodebook [10]–[16], [38] that is impractical for low-powernodes, or finite equiprobable inputs [7], [8], [17], [18], [34]–[37], [39]–[43] that does not fully exploit the CSI and signalproperties of active-passive coexisting networks.C. ContributionsFirst, we propose RIScatter as a novel scatter protocolthat unifies BackCom and RIS by adaptive reflection state(i.e., backscatter input) distribution design. As shown inFig. 1(f), one or more passive RIScatter nodes ride over anactive transmission and smartly scatters the incident wave tosimultaneously engineer the primary channel and modulateits own information, while a co-located receiver cooperativelydecodes both coexisting links. The reflection pattern over timeis neither fully random like BackCom nor fully deterministiclike RIS, but semi-randomly chosen with the guidance ofreflection state distribution, whose design accommodates theimpact of information source, CSI, and QoS of both links.Such an adaptive backscatter channel coding boils down tothe degenerate distribution of RIS when the primary linkis prioritized, and outperforms the uniform distribution ofline-coded BackCom (by accounting the CSI) when thebackscatter link is prioritized. Besides, when multiple RIScatternodes are co-located, joint encoding can further increase thetotal backscatter rate. RIScatter exploits the transmitter-scatter-receiver path for simultaneous backscatter modulation anddynamic passive beamforming, and can smoothly transitionin between under different QoS. However, such a novelscatter protocol encounters two major challenges, namely thecooperative receiver design and the adaptive backscatter inputdistribution design. This is the first paper to unify BackComand RIS from the perspective of reflection state distribution.Second, we address the first challenge by proposing a practi-cal receiver that accommodates the double fading, double modu-lation, and symbol period difference of active-passive coexistingnetworks. Since joint decoding involves high complexity whileindividual decoding can be inefficient (neither cooperative norexploiting signal properties), most relevant literatures consider asequential primary-backscatter decoding based on SIC. Specif-ically, the decoding of each backscatter symbol requires N re-encoding, precoding, subtraction together with a time-domainMRC, which can be operation-intensive and CSI-sensitive.In contrast, the proposed RIScatter receiver semi-coherentlydecodes the backscatter symbol from the accumulated receiveenergy, re-encodes to recover the exact reflection pattern, andcombine it with relevant CSI to determine the transmitter-scatter-receiver path response. Thanks to the double modulationand symbol period difference, the backscatter symbol canbe modelled within primary channel (similar to RIS) ratherthan being re-precoded and cancelled, and the semi-coherentenergy-based backscatter detection indeed becomes part ofprimary channel training. The strategy is suitable for arbitraryprimary sampling rate and backscatter load-switching speed(i.e., arbitrary N). This is the first paper to consider such a se-quential backscatter-primary decoding and treat semi-coherentbackscatter detection as part of primary channel training.Third, we address the second challenge by considering aspecific scenario where multiple co-located/dispersed RIScatternodes ride over an active point-to-point Multiple-Input Single-Output (MISO) transmission to perform backscatter modulationand dynamic passive beamforming towards a nearby user.For a given QoS, we first express the primary and total4backscatter rates as functions of input distribution at RIScatternodes, active beamforming at the Access Point (AP), andbackscatter decision regions at the user. Then, we characterizethe achievable rate region by optimizing those variables underdifferent QoS. Since the original problem is highly non-convex,we propose a suboptimal Block Coordinate Descent (BCD)algorithm where the Karush-Kuhn-Tucker (KKT) inputdistribution is numerically evaluated by limit of sequences,the active beamforming is iteratively updated by ProjectedGradient Descent (PGD) accelerated by Backtracking LineSearch (BLS), and the decision regions are refined by existingsequential quantization methods for Discrete MemorylessThresholding Channel (DMTC). Uniquely, we account for CSI,QoS, and backscatter constellation in those optimizations, andthe proposed backscatter input distribution design is applicableto more general scenarios. This is also the first paper to revealthe importance of backscatter input distribution and decisionregion designs in active-passive coexisting networks.Fourth, we provide numerical results to demonstrate thebenefits of RIScatter and proposed algorithms. The observationsinclude: 1) adaptive reflection state distribution design canflexibly transition between backscatter modulation and passivebeamforming; 2) when the primary link is prioritized, inputdistribution becomes degenerate and RIScatter nodes coincidewith discrete RIS; 3) when the backscatter link is prioritized,adaptive RIScatter encoding achieves higher backscatter ratethan conventional line-coded BackCom with equiprobableinputs; 4) co-located RIScatter nodes can further leverage totalbackscatter rate by joint encoding; 5) the proposed receivermaintains the passive beamforming benefit and providescomparable backscatter rate to SIC-based SR, with re-encodingcosts reduced to 1/N and no re-precoding/cancellation; 6) italso supports scatter nodes with faster load-switching speedfor potentially higher throughput; 7) PGD active beamformereffectively increases the primary (resp. backscatter) rate byboosting the receive SNR (resp. widening the energy gapunder different reflection states), and can smoothly transition inbetween under different QoS; 8) distribution-aware backscatterdetectors provide higher backscatter rate than the conventionalML detector.Notations: Italic, bold lower-case, and bold upper-case lettersdenote scalars, vectors and matrices, respectively. 0 and 1denote zero and one array of appropriate size, respectively.Ix×y, Rx×y+, and Cx×y denote the real unit, real nonnegative,and complex spaces of dimension x×y, respectively. j denotesthe imaginary unit. diag(·) returns a square matrix with theinput vector on its main diagonal and zeros elsewhere. card(·)returns the cardinality of a set. (·)∗, (·)T, (·)H, |·|, and ∥·∥denote the conjugate, transpose, conjugate transpose, absolutevalue, and Euclidean norm operators, respectively. (·)(r) and(·)⋆ denote the r-th iterated and optimal results, respectively.The distribution of a Circularly Symmetric Complex Gaussian(CSCG) random variable with zero mean and variance σ2 isdenoted by CN(0,σ2), and ∼ means “distributed as”.II. SCATTERING PRINCIPLESRF wave scattering and reflecting can be described bya unified model and realized by variable-load antennas orprogrammable metamaterial [44]. An antenna-based scattererusually consists of an integrated antenna, a load-switchingmodulator, an energy harvester, and on-chip components (e.g.,microcontroller and sensors) [2]. It first receives the impingingsignals, then reradiates some back to the space and dissipatesthe remaining. In comparison, a typical metamaterial-basedscatterer comprises an outer metamaterial layer of numeroussub-wavelength metallic/dielectric patches (with tunablepermittivity/permeability), a middle copper plate layer thatavoids leakage, an inner circuit board layer that adjustspatch responses, and an integrated microcontroller/FPGAfor coordination and control [45]. Ideally, it reflects thewave at the space-metamaterial boundary without receivingthem and mainly applies a phase shift. In practice, bothtypes of scatterers have finite reflection states with non-zeroreflection loss. The scattered signal can be decomposed intoa structural mode component that consistently contributes toenvironment multipath and can be modelled within CSI, andan antenna mode component that depends on the impedancemismatch and can be used for backscatter modulation and/orpassive beamforming [46]. For an antenna-based (resp.metamaterial-based) scatter node with M reflection states, thereflection coefficient at state m∈M≜{1,...,M} isΓm = Zm−Z∗Zm+Z ,(1)where Zm is the antenna load (resp. metamaterial unit)impedance at state m and Z is the antenna input (resp.medium characteristic) impedance. BackCom employ scatternodes as information sources that randomly switching betweendifferent states. For M-ary QAM, constellation point cm mapsto reflection coefficient Γm by [47]Γm =αcmmaxm′|cm′|,(2)where α ∈ I is the common amplitude scattering ratio at thedirection of interest. In contrast, RIS employ reflect elements aschannel adaptors that deterministically choosing the reflectionstate based on CSI. For a RIS element with M available states,phase shift θm maps to reflection coefficient Γm by [21]Γm =βmexp(jθm),(3)where βm ∈I is the overall amplitude scattering ratio at statem.3III. RISCATTERA. ConceptsWe propose RIScatter in Fig. 1(f) as a generalization of Back-Com and RIS, which involves coexisting active primary andpassive backscatter links in a flexible and mutualistic manner.RIScatter nodes leverage CSI- and QoS-based reflection statedistribution design to smoothly transition between backscattermodulation and passive beamforming. Fig. 2 illustrates theblock diagram, equivalent circuit, and scatter model of aRIScatter node. It can be implemented by adding an integratedreceiver [48] and adaptive encoder [49] to off-the-shelf passiveRFID tags. Fig. 3 compares the reflection state distribution andreflection pattern of RIScatter and other applications. Instead of3Most existing RIS literatures assume lossless reflection βm =1, ∀m.5SxRectifierDemodulatorModulatorPower BufferDigitalSection(a) Block DiagramV0XARAI0XmRS,mRP,mXHRHAntennaModulatorHarvester & ChipZAZmZH(b) Equivalent Circuit⃗EI⃗HI⃗Em⃗HmZAm Zm(c) Scatter ModelFig. 2.Block diagram, equivalent circuit, and scatter model of a RIScatter node. The solid and dashed vectors represent signal and energy flows. The scatterantenna behaves as a constant power source, where the voltage V0 and current I0 are introduced by incident electric field ⃗EI and magnetic field ⃗HI [50].00.51BackCom00.51Probability DistributionRIS123400.51Reflection StateRIScatter(a) Reflection State DistributionPBBBChannel BlockBackComRISRIScatterState 1State 2State 3State 4(b) Reflection PatternFig. 3.Reflection state distribution and reflection pattern over time ofBackCom, RIS, and RIScatter. “PB” means primary block and “BB” meansbackscatter block.using fully random or fully deterministic reflection pattern overtime, each RIScatter node semi-randomly chooses the reflectionstate for each backscatter block with guidance of inputprobability P(Γm) at state m. Such an adaptive backscatterchannel coding includes the degenerate distribution of RIS andthe uniform distribution of line-coded BackCom as specialcases. Importantly, RIScatter system design only requires thedirect transmitter-receiver and cascaded transmitter-scatter-receiver CSI. For dispersed RIScatter nodes, relevant CSI canbe estimated by sequential [51]–[53] or parallel approaches [54]originally proposed for BackCom. For co-located RIScatternodes, the estimation can be simplified by group-based [55]and hierarchical [56] trainings originally proposed for RIS.We also propose a novel cooperative receiver that exploits thedouble modulation, double fading and symbol period differenceto reduce the decoding complexity and improve the primary-backscatter tradeoff. As illustrated in Fig. 4, conditioned ondifferent reflection state hypotheses, the accumulated receiveenergy per backscatter block follows Gamma distribution withdifferent scale parameters [57]. Hence, the receiver can semi-coherently (under primary uncertainty) decode RIScatter nodesfrom the accumulated receive energy, re-encode to recover theexact reflection patterns, and combine those with cascadedCSI to determine the multipath response and model within theprimary channel. Such a backscatter-primary decoding enablesFig. 4.PDF of accumulated receive energy per backscatter block conditionedon different reflection state hypotheses. z, t, H and R denote the accumulatedreceive energy, decision threshold, reflection state hypothesis, and decisionregions, respectively.APUser...hHDhHF,khB,kNode k′Node kFig. 5. A single-user multi-node RIScatter system.simultaneous backscatter modulation and dynamic passivebeamforming by only one energy comparison, re-encoding,and precoding during each backscatter (instead of primary)block. It enjoys a three-fold benefit compared to conventionalprimary-backscatter decoding with SIC, namely 1) reduces there-encoding frequencies to 1/N; 2) requires no re-precoding,cancellation, or MRC; 3) preserves backscatter modulationand passive beamforming at smaller N for potentially higherbackscatter throughput.B. System ModelAs shown in Fig. 5, we consider a RIScatter system wherea Q-antenna AP serves a single-antenna user as well as Knearby dispersed/co-located single-antenna M-state RIScatternodes. In the primary point-to-point system, the AP transmitsinformation to the user over the multipath channel enhancedby RIScatter nodes. In the backscatter multiple access system,the AP and user become carrier emitter and backscatter reader,and the RIScatter nodes modulate over the scattered RF signals.For simplicity, we consider a quasi-static block fading modeland focus on a specific block where all CSI remain constant.Without loss of generality, let the backscatter-over-primarysymbol period ratio N be a positive integer. We also omit thesignal reflected by two or more times and ignore the multipathpropagation time difference. Denote the AP-user direct channel6as hHD ∈C1×Q, the AP-node k∈K≜{1,...,K} forward channelas hHF,k ∈ C1×Q, the node k-user backward channel as hB,k,and the cascaded AP-node k-user channel as hHC,k ≜hB,khHF,k ∈C1×Q. Let xk ∈X ≜{c1,...,cM} be the coded backscatter sym-bol of node k and xK ≜(x1,...,xK) be the backscatter symboltuple of all nodes. Due to double modulation, the primaryequivalent channel is a function of backscatter symbol tuple4hHE(xK)≜hHD+�kαkhHC,kxk(4a)=hHD+xHdiag(α)HC,(4b)where αk ∈I is the common amplitude scattering ratio of nodek, α ≜ [α1,...,αK]T ∈ IK, x ≜ [x1,...,xK]H ∈ X K, and HC ≜[hC,1,...,hC,K]H ∈CK×Q. In the following context, we focuson one backscatter block (i.e., N primary blocks). The signalreceived by the user at primary block n∈N ≜{1,...,N} isy[n]=hHE(xK)ws[n]+v[n],(5)where w ∈ CQ is the active beamformer satisfying averagetransmit power constraint ∥w∥2 ≤ P, s ∼ CN(0,1) is theprimary symbol, and v ∼ CN(0,σ2v) is the Additive WhiteGaussian Noise (AWGN) with average power σ2v. Letmk ∈ M ≜ {1,...,M} be the reflection state index of node kand mK ≜(m1,...,mK) be the reflection state index tuple ofall nodes. The backscatter symbol xk (resp. symbol tuple xK)is a random variable (resp. variable tuple) that takes value xmk(resp. value tuple xmK) when state mk (resp. state tuple mK)is employed. Conditioned on mK, the receive signal at eachprimary block follows CSCG distribution CN(0,σ2mK), whereσ2mK =|hHE(xmK)w|2+σ2v(6)is the received variance. Let z =�n��y[n]��2 be the accumulatedreceive energy per backscatter block. Since z is the sum ofN independent and identically distributed (i.i.d.) exponentialvariables, its conditional PDF follows Gamma distributionf(z|HmK)= zN−1exp(−z/σ2mK)σ2NmK(N −1)!,(7)where HmK denotes hypothesis mK. At the receiver, theenergy space is divided into disjoint decision regions associatedwith different hypotheses, as illustrated in Fig. 4.Remark 1. The capacity-achieving decision region designfor DMTC with non-binary inputs in arbitrary distributionremains an open issue. It was proved deterministic detectorcan be rate-optimal, but non-convex decision regions (i.e.,comprise non-adjacent partitions) are generally involved andthe optimal number of thresholds remains unknown [58],[59]. Hence, we limit the backscatter energy detector toconvex deterministic decision regions and consider sequentialthreshold design in the following context.For the ease of notations, we map the state index tuple mKto l∈L≜{1,...,L≜M K}, where σ21,...,σ2L is a non-decreasingsequence. Both notations are used interchangeably in the4(4a) and (4b) are often used in BackCom and RIS literatures, respectively.following context. As such, the decision region of backscattersymbol tuple l can be written asRl ≜[tl−1,tl),0≤tl−1 ≤tl,(8)where tl is the energy decision threshold between hypothesesHl and Hl+1. For a given decision threshold vector t ≜[t0,...,tL]T ∈R(L+1)+, we can formulate a Discrete MemorylessThresholding Multiple Access Channel (DMTMAC) with tran-sition probability from input tuple xmK to output tuple ˆxm′K asP(ˆxm′K|xmK)=�Rm′Kf(z|HmK)dz,(9)over which adaptive backscatter channel coding can be per-formed.C. Achievable RatesDenote the probability of node k choosing reflection statemk as Pk(xmk), and let pk ≜[Pk(c1),...,Pk(cM)]T ∈IM. Fordispersed nodes with independent encoding, the probabilityof backscatter symbol value tuple xmK isPK(xmK)=�k∈KPk(xmk).(10)Following [60], we define the backscatter information functionbetween input value tuple xmK and output variable tuple ˆxK asIB(xmK;ˆxK)≜�m′KP(ˆxm′K|xmK)logP(ˆxm′K|xmK)PK(ˆxm′K),(11)where PK(ˆxm′K) = �mK PK(xmK)P(ˆxm′K|xmK) is theprobability of output value tuple ˆxm′K. We also define thebackscatter marginal information of xmk asIB,k(xmk;ˆxK)≜�mK\{k}PK\{k}(xmK\{k})IB(xmK;ˆxK), (12)where PK\{k}(xmK\{k}) = �q∈K\{k} Pq(xmq). Hence, thebackscatter mutual information can be written asIB(xK;ˆxK)=�mKPK(xmK)IB(xmK;ˆxK).(13)Once nodes are successfully decoded, we can re-encode anddetermine the primary equivalent channel by (4). Therefore,the primary information function conditioned on xmK isIP(s;y|xmK)≜log�1+ |hHE(xmK)w|2σ2v�,(14)the primary marginal information of xmk isIP,k(s;y|xmk)≜�mK\{k}PK\{k}(xmK\{k})IP(s;y|xmK), (15)and the average primary mutual information isIP(s;y|xK)=�mKPK(xmK)IP(s;y|xmK).(16)7With a slight abuse of notation, we define the correspondingweighted sum information function, marginal information, andmutual information asI(xmK)≜ρIP(s;y|xmK)+(1−ρ)IB(xmK;ˆxK),(17)Ik(xmk)≜ρIP,k(s;y|xmk)+(1−ρ)IB,k(xmk;ˆxK),(18)I(xK)≜ρIP(s;y|xK)+(1−ρ)IB(xK;ˆxK),(19)where ρ∈I is the relative QoS of the primary link. We noticethe average primary rate (16) depends on the backscatter inputdistribution and active beamforming, while the total backscatterrate depends on the input distribution and DMTMAC (9) thatrelates to the active beamforming and decision thresholds.IV. RATE-REGION CHARACTERIZATIONTo characterize the achievable primary-(total-)backscatterrate region for the RIScatter system in Fig. 5, we aim to max-imize the weighted sum rate with respect to input distribution{pk}k∈K, active beamforming w, and decision thresholds t bymax{pk}k∈K,w,tI(xK)(20a)s.t.1Tpk =1,∀k,(20b)pk ≥0,∀k,(20c)∥w∥2 ≤P ,(20d)tl−1 ≤tl,∀l,(20e)t≥0.(20f)Problem (20) generalizes conventional BackCom by allowingCSI- and QoS-adaptive input distribution and detection regiondesign. It also generalizes the discrete RIS phase shift selectionby allowing stochastic reflection (i.e., relaxing the feasibledomain from the vertices of M-dimensional probabilitysimplex to the simplex itself). Since problem (20) is highlynon-convex, we propose a BCD algorithm that iterativelyupdates {pk}k∈K, w and t until convergence.A. Input DistributionFor any given w and t, we can construct the equivalentDMTMAC by (9) and simplify (20) tomax{pk}k∈KI(xK)(21a)s.t.(20b),(20c),(21b)which is convex when K = 1 or joint encoding5 over K > 1co-located nodes is available. When the nodes are dispersed,problem (21) involves coupled term �k∈K Pk(xmk) and isnon-convex. Following [60], we first recast the KKT conditionsto their equivalent forms, then propose a numerical methodthat guarantees those conditions on convergence of sequences.Remark 2. As demonstrated in [61], KKT conditions aregenerally necessary but insufficient for total rate maximizationof discrete Multiple Access Channel (MAC). We will showby numerical results that, for a moderate K, the average5Joint encoding formulates an equivalent source of MK codewords, suchthat one can directly design PK(xmK), ∀mK instead of Pk(xmk), ∀k,mk.Algorithm 1: Numerical KKT Input Distribution Evaluationby Limits of SequenceInput: K, N, hHD, HC, α, X, σ2v, ρ, w, t, ϵOutput: {p⋆k}k∈K1: Set hHE(xmK), ∀mK by (4)2:σ2mK, ∀mK by (6)3:f(z|HmK), ∀mK by (7)4:P(ˆxm′K|xmK), ∀mK,m′K by (9)5: Initialize r←06:p(0)k >0, ∀k7: Get P (r)K (xmK), ∀mK by (10)8:I(r)(xmK), ∀mK by (11), (14), (17)9:I(r)k (xmk), ∀k,mk by (12), (15), (18)10:I(r)(xK) by (13), (16), (19)11: Repeat12:Update r←r+113:p(r)k , ∀k by (23)14:Redo step 7–1015: Until I(r)(xK)−I(r−1)(xK)≤ϵachievable rate regions of KKT and global-optimal inputdistributions completely coincide with each other.Proposition 1. The KKT optimality conditions for problem(21) are equivalent to, ∀k,mk,I⋆k(xmk)=I⋆(xK),P ⋆k (xmk)>0,(22a)I⋆k(xmk)≤I⋆(xK),P ⋆k (xmk)=0.(22b)Proof. Please refer to Appendix A.For each node, (22a) suggests each probable state shouldproduce the same marginal information (averaged over allstates of other nodes), while (22b) suggests any state withpotentially less marginal information should not be used.Proposition 2. For any strictly positive initializer {p(0)k }k∈K,the KKT input probability of node k at state mk is given bythe converging point of the sequenceP (r+1)k(xmk)=P (r)k (xmk)exp�ρ1−ρI(r)k (xmk)��m′kP (r)k (xm′k)exp�ρ1−ρI(r)k (xm′k)�, (23)where r is the iteration index.Proof. Please refer to Appendix B.For (23) at iteration r+1, the input distribution of node kis updated over�{p(r+1)q}k−1q=1,{p(r)q }Kq=k�. The KKT inputdistribution design is summarized in Algorithm 1.B. Active BeamformingFor any given {pk}k∈K and t, problem (20) reduces tomaxwI(xK)(24a)s.t.(20d),(24b)8which is still non-convex due to the integration and entropyterms. To tackle this, we rewrite the DMTMAC transitionprobability (9) from input index tuple mK to output indexl as a regularized incomplete Gamma function in the seriesrepresentation [62, Theorem 3]Q�N, tl−1σ2mK, tlσ2mK�=� tl/σ2mKtl−1/σ2mK zN−1exp(−z)dz(N −1)!=exp�− tl−1σ2mK�N−1�n=0� tl−1σ2mK�nn!−exp�− tlσ2mK�N−1�n=0�tlσ2mK�nn!.(25)Its gradient with respect to w∗ can be derived as∇w∗Q�N, tl−1σ2mK, tlσ2mK�= hE(xmK)hHE(xmK)w(σ2mK)2gmK(tl−1,tl),(26)where gmK(tl−1,tl)≜gmK(tl)−gmK(tl−1) andgmK(tl)=tlexp�− tlσ2mK��−1+N−1�n=1�n−tlσ2mK��tlσ2mK�n−1n!�.(27)On top of (25) and (26), we explicitly express the objectivefunction (24a) and its gradient as (28) and (29) at the end ofpage 8, respectively. They allows problem (24) to be solvedby the PGD method, where any unregulated beamformer ¯wcan be projected onto the feasible domain of average transmitpower constraint (20d) byw=√P¯wmax�√P,∥ ¯w∥�.(30)The PGD active beamforming optimization with step size byBLS [63, Section 9.2] is summarized in Algorithm 2.C. Decision ThresholdFor any given {pk}k∈K and w, problem (20) reduces tomaxtI(xK)(31a)s.t.(20e),(20f),(31b)which is still non-convex because variable t appears on thelimits of integration (9). Fortunately, we can further simplifyproblem (31) as a point-to-point rate-optimal quantizer designAlgorithm 2: Iterative Active Beamforming Optimization byPGD with BLSInput: Q, N, hHD, HC, α, X, P, σ2v, ρ, {pk}k∈K, t, α, β, γ, ϵOutput: w⋆1: Set hHE(xmK), ∀mK by (4)2:PK(xmK), ∀mK by (10)3: Initialize r←04:w(0), ∥w(0)∥2 ≤P5: Get (σ(r)mK)2, ∀mK by (6)6:Q(r)�N, tl−1σ2mK ,tlσ2mK�, ∀mK,l by (25)7:I(r)(xK) by (28)8:∇w∗Q(r)�N, tl−1σ2mK ,tlσ2mK�, ∀mK,l by (26)9:∇w∗I(r)(xK) by (29)10: Repeat11:Update r←r+112:γ(r) ←γ13:¯w(r) ←w(r−1)+γ∇w∗I(r−1)(xK)14:w(r) by (30)15:Redo step 5–716:WhileI(r)(xK)<I(r−1)(xK)+αγ∥∇w∗I(r−1)(xK)∥217:Set γ(r) ←βγ(r)18:Redo step 13–1519:End While20:Redo step 8, 921: Until ∥w(r)−w(r−1)∥≤ϵfor a discrete-input continuous-output memoryless channel,thanks to Remark 3 and 4.Remark 3. Backscatter decision design has no impact onthe primary achievable rate, because the primary equivalentchannel (4) at each backscatter block can always bedetermined upon successful backscatter decoding. It suggestsany thresholding scheme that maximize the total backscatterrate (13) is also optimal for problem (31).Remark 4. In terms of total backscatter rate, the potentiallydispersed nodes with known input distribution can be viewedas an equivalent source with symbol tuples as codewords.As such, the DMTMAC (9) becomes a DMTC and problemI(xK)=�mKPK(xmK)�ρlog�1+ |hHE(xmK)w|2σ2v�+(1−ρ)�lQ�N, tl−1σ2mK, tlσ2mK�logQ�N, tl−1σ2mK ,tlσ2mK��m′KPK(xm′K)Q�N, tl−1σ2m′K,tlσ2m′K��(28)∇w∗I(xK)=�mKPK(xmK)�ρhE(xmK)hHE(xmK)wσ2mK+(1−ρ)�l�logQ�N, tl−1σ2mK ,tlσ2mK��m′KPK(xm′K)Q�N, tl−1σ2m′K,tlσ2m′K� +1�×∇w∗Q�N, tl−1σ2mK, tlσ2mK�−Q�N, tl−1σ2mK ,tlσ2mK��m′KPK(xm′K)∇w∗Q�N, tl−1σ2m′K,tlσ2m′K��m′KPK(xm′K)Q�N, tl−1σ2m′K,tlσ2m′K��(29)9t0t1t2t3t4R1R2R3R4zFig. 6. The decision thresholds are selected from fine-grained discrete energylevels instead of continuous space, and each decision region consists of atleast one neighbor energy bins.(31) reduces to the rate-optimal quantization design for adiscrete-input continuous-output memoryless channel.Next, we constrain the feasible domain of problem (31)from continuous space RL+1+to finite candidate set (i.e., fine-grained discrete energy levels) T L+1. As shown in Fig. 6, byintroducing an extra analog-to-digital conversion, we can groupadjacent high-resolution energy bins to construct backscatterdecision regions. Thus, problem (31) can be recast asmaxt∈T L+1IB(xK;ˆxK)(32a)s.t.(20e),(32b)which is solvable using existing rate-optimal sequentialquantizer designs for DMTC. To obtain global optimalsolution, [64] started from the quadrangle inequality andproposed a Dynamic Programming (DP) method accelerated bythe Shor-Moran-Aggarwal-Wilber-Klawe (SMAWK) algorithmwith computational complexity O�L2(card(T )−L)�, while[65] started from the optimality condition for three neighborthresholds and presented a traverse-then-bisect algorithm withcomplexity O�card(T )Llog(card(T )L)�. In Section V, bothschemes will be compared with the ML scheme [66]tMLl=N σ2l−1σ2lσ2l−1−σ2llogσ2l−1σ2l,l∈L\{L},(33)which is generally suboptimal for problem (31) except whenall nodes are with equiprobable inputs.V. SIMULATION RESULTSIn this section, we provide numerical results to evaluatethe proposed input distribution, active beamforming, andbackscatter decision designs for the considered RIScattersystem. We assume the AP-user distance is 10 and at least oneRIScatter nodes are randomly dropped in a disk centered at theuser with radius r. The AP is with an average transmit powerbudget P = 36dBm and all nodes employs M-QAM withcommon amplitude scattering ratio α=0.5. For all channelsinvolved, we consider a distance-dependent path loss modelL(d)=L0�d0d�γ,(34)together with a Rician fading modelH =�κ1+κ¯H+�11+κ˜H,(35)where d is the transmission distance, L0 = −30dB is thereference path loss at d0 =1m, κ is the Rician K-factor, ¯H isthe deterministic line-of-sight component with unit-magnitudeentries, and ˜H is the Rayleigh fading component with standard0204060801001200.1140.1150.116KKT024681012140.120.130.140.15Weighed Sum-RatePGD01234567800.20.40.6Number of IterationsBCDFig. 7.Typical convergence curves at ρ = 0 for Q = 4, K = 8, M = 2,N =20, σ2v =−40dBm and r=2m.i.i.d. CSCG entries. We choose γD = 2.6, γF = 2.4, γB = 2,and κD =κF =κB =5 for direct, forward and backward links.The finite decision threshold domain T is obtained by b-bituniform discretization over the critical interval defined by theconfidence bounds of edge hypotheses (i.e., lower bound ofH1 and upper bound of HL) with confidence 1−ϵ, and wechoose b=9 and ϵ=10−3. All achievable rate points/regionsare averaged over 1000 channel realizations.A. Evaluation of Proposed Algorithms1) Initialization: To characterize each achievable rate region,we progressively obtain all boundary points by successivelyincreasing the primary QoS ρ and solving problem (20). Forρ = 0 where the backscatter link is prioritized, we initializeAlgorithm 1 and 2 by uniform input distribution and MaximumRatio Transmission (MRT) towards the sum cascaded channel�khHC,k, respectively. At the following points, both algorithmsare initialized by the final solutions at the previous point.2) Convergence: In Fig. 7, we plot the weighted sum ofprimary and total backscatter rates at ρ = 0 for KKT, PGDand BCD algorithms on the first call. For K =8 and M =2,Algorithm 1 typically takes around 100 fast iterations by (23)to converge to the KKT input distribution. For Q=4, around10 iterations are required for Algorithm 2 to converge, wherethe gradient is computed by (29) and the step size is refinedby BLS. Overall, the BCD algorithm initially requires at most5 iterations to converge. At the following points (not presentedhere), the convergence of all three algorithms are much fasterthanks to the progressive initialization. Hence, we conclude theproposed algorithms have good convergence performances.B. Comparison of Scattering ApplicationsOn top of the setup in Fig. 5, we consider RIScatter andthe following benchmark applications:• Legacy: Conventional active transmission without antennamode scattering, α=0.• BBC: The primary symbol becomes deterministic s[n]=1and the receive signal at each primary block isyBBC[n]=�hHD+�kαkhHC,kxk�w+v[n],(36)10whichfollowsnon-zeromeancomplexGaussiandistribution CN�(hHD + �k αkhHC,kxmk)w, σ2v�underhypothesis HmK. The corresponding PDF of accumulatedreceive energy over N primary blocks isf BBC(z|HmK)= (z−µBBCmK )N−1exp�−(z−µBBCmK )/σ2v�σ2Nv (N −1)!,(37)where µBBCmK ≜ N���hHD + �k αkhHC,kxmk�w��2. The MLdecision threshold is derived as, ∀l∈L\{L},tBBCl= µBBCl−1 exp�(µBBCl−1 −µBBCl)/σ2v(N −1)�−µBBClexp�(µBBCl−1 −µBBCl)/σ2v(N −1)�−1.(38)• AmBC: The user decodes both links independently andsemi-coherently by treating the other as interference.Hence, the primary achievable rate is approximately6IAmBCP(s;y)≈log�1+|hHDw|2�k|αkhHC,kw|2+σ2v�,(39)while the total backscatter rate follows (13) with uniforminput distribution.• SR: For a sufficiently large N, the average primary rateunder semi-coherent detection asymptotically approaches(16) with uniform input distribution [12]. When s[n] issuccessfully decoded and the direct interference hHDws[n]is perfectly cancelled, the intermediate signal isˆySR[n]=�kαkhHC,kxkws[n]+v[n],(40)which only involves noise uncertainty under hypothesisHmK. During backscatter detection, the primary symbolss[1],...,s[n] can be viewed as a spreading code, and thereceiver employs MRC over N primary blocks. The totalachievable for nodes with equiprobable inputs is [67]IB(xK;ˆySR)=KlogM −ιM K ,(41)where ι≜�mKEˆvlog�m′Kexp(−|xmK −xm′K +ˆv|2/2σ2)and ˆv∼CN(0,σ2v/N).• RIS:SincethebackscattersymboltuplexKisdeterministic, the total backscatter rate is zero and theprimary achievable rate becomes a special case of (16)IRISP(s;y|xK)=IP(s;y|xm⋆K)=log�1+ |hHE(xm⋆K)w|2σ2v�,(42)where m⋆K =argmaxmKIP(s;y|xmK).Fig. 8 compares the typical achievable rate region/pointsof RIScatter and above. First, we observe both BBC and SRalmost ensure noise-free backscatter transmission when N issufficiently large. For BBC with coherent energy detection, theconditional PDF of accumulated receive energy (37) is moreskewed at a large N, such that the equivalent DMTMAC (9)has lower error probabilities. For SIC-based SR, the effectivebackscatter SNR is increased by N times and the penalty term6To provide a preliminary benchmark, we consider the (correlated) scatteredsignal from finite-input backscatter sources as independent interference fromGaussian sources during primary decoding [12].012345600.511.52Primary Rate [bits/s/Hz]BackscatterRate [bits/BB]LegacyBBCAmBCSRRISRIScatterFig. 8. Typical achievable rate region/points of scattering applications forQ=1, K =1, M =4, N =1000, σ2v =−40dBm and r=2m.123400.20.40.60.81Reflection StateProbabilityDistributionρ=0ρ=0.1ρ=0.25ρ=1Fig. 9. Typical RIScatter reflection state distribution at different ρ for Q=1,K =1, M =4, N =20, σ2v =−40dBm and r=2m.ι becomes insignificant. However, the primary rate is onlymaintained at a very large N, which significantly constrainsthe backscatter throughput. Second, the average primary rateslightly increases/decreases in the presence of a AmBC/RISnode, and the multipath benefit of SR is unobvious. This isbecause the cascaded channel can be orders of magnitudeweaker than the direct channel due to the double fading effect.RIS always ensures constructive superposition of direct andscattered components, while SR only creates a quasi-static rich-scattering environment that stochastically enhances the averageprimary rate. When N is moderate, the randomly scatteredsignals should be modelled as primary interference rather thanmultipath components, and the SR point will move towardsthe AmBC point. Third, RIScatter enables a flexible primary-backscatter tradeoff with adaptive input distribution design. Interms of maximum primary achievable rate, RIScatter coincideswith RIS and outperforms the others with passive beamformingprovided by deterministic reflection pattern. On the other hand,for a large N, the maximum backscatter achievable rate ofRIScatter is higher than AmBC but lower than BBC and SR.This is because both RIScatter and AmBC (resp. BBC andSR) employ semi-coherent (resp. coherent) energy detection,while RIScatter with adaptive channel coding can achievehigher backscatter rate than AmBC with equiprobable inputs.Importantly, such a practical RIScatter detection is feasible forarbitrary N, which unleashes the potential of fast-switchingnodes.C. Input Distribution under Different QoSThe objective of this study is to demonstrate RIScatter nodescan leverage CSI- and QoS-adaptive input distribution designto balance backscatter modulation and passive beamforming.For one RIScatter node with M =4, we evaluate the KKT inputdistribution7 at different primary QoS and present the result inFig. 9. At ρ=0 where the backscatter performance is prioritized,7Since problem (21) is convex when K =1, the KKT solution is alwaysglobal optimal in this case.1166.16.26.36.46.500.020.040.060.080.10.120.140.16Primary Rate [bits/s/Hz]Backscatter Rate [bits/BB]CooperationExhaustionKKTEquiprobableMarginalizationDecompositionRandomization6.336.356.370.080.082(a) Input Distribution, Q=101234567800.050.10.150.20.250.30.350.40.450.5Primary Rate [bits/s/Hz]Total Backscatter Rate [bits/BB]PGDE-MRTD-MRT(b) Active Beamforming, Q=401234567800.10.20.30.40.5Primary Rate [bits/s/Hz]Total Backscatter Rate [bits/BB]DPSMAWKBisectionML(c) Decision Threshold, Q=4Fig. 10.Average primary-total-backscatter rate regions by different inputdistribution, active beamforming, and decision threshold schemes for K =2,M =4, N =20, σ2v =−40dBm and r=2m.the optimal input distribution is zero on two states and nearlyuniform on the other two. This is because, due to the weakscattered signal, the conditional energy PDF under differenthypotheses can be closely spaced as illustrated in Fig. 4. In suchcases, the extreme states producing the lowest/highest energyare always assigned with non-zero probability, while the middleones may not provide enough energy difference and thus end upunused. At ρ=1 where the primary performance is prioritized,the optimal input distribution is 1 at the state that maximizes theprimary SNR and 0 at the others. That is, the reflection patternbecomes deterministic and the RIScatter node boils down toa discrete RIS element. Increasing ρ from 0 to 1 providesa smooth transition from backscatter modulation to passivebeamforming, which suggests RIScatter unifies BackCom andRIS from a probabilistic perspective.D. Rate Region by Different Schemes1) Input Distribution: We compare these input distributiondesigns for problem (21):• Cooperation: Joint nodes encoding using a K-dimensionalprobability array PK(xmK) by Algorithm 1;• Exhaustion: Exhaustive search over the M-dimensionalprobability simplex with resolution ∆p=10−2;• KKT: Numerical KKT result evaluation by Algorithm 1;• Equiprobable: Uniform input distribution for all nodes.For dispersed nodes, joint encoding is inapplicable and weconsider these independent input distribution recovery methods(from the joint probability array):• Marginalization: Marginal probability distributions;• Decomposition: Normalized rank-1 Canonical Polyadic(CP) decomposition tensors by Tensor Toolbox [68];• Randomization:Gaussianrandomizationwiththeguidance of correlation matrix [69].Fig. 10(a) shows the average achievable rate regions forthose designs. We observe Cooperation provides the outerbound since joint encoding is always beneficial. The averagerate performance of Exhaustion and KKT completely coincidewith each other, demonstrating KKT input distribution isfavorable for a moderate K as stated in Remark 2. Equiprobableexperiences minor backscatter and major primary rate losseswithout CSI- and QoS-adaptive backscatter encoding. Bothgaps should become larger when M and K increase. For therecovery methods, the simple Marginalization provides a closeresult to KKT, but Randomization and Decomposition fail ourexpectations for most channel realizations. Those observationsemphasize the importance of (joint) adaptive RIScatter encodingand demonstrate the advantages of the proposed KKT inputdistribution design.2) Active Beamforming: We consider three typical activebeamforming schemes for problem (24):• PGD: Iterative PGD optimization by Algorithm 2;• E-MRT: MRT towards the ergodic primary equivalentchannel �mKPK(xmK)hHE(xmK);• D-MRT: MRT towards the direct channel hHD.Fig. 10(b) presents the average achievable rate regionsfor those schemes. In the low-ρ regime, the proposed PGDbeamformer significantly outperforms both MRT schemes interms of total backscatter rate. This is because backscatterdetection relies on the relative difference of accumulated receiveenergy under different symbol tuples. Such an energy diversityis enhanced by PGD that effectively exploits backscatter con-stellation and input distribution knowledge, rather than simplymaximizing the direct/ergodic equivalent SNR. As ρ increases,the primary equivalent SNR outweighs the backscatter energydifference in (28), and PGD beamformer becomes closer to bothMRT schemes. At ρ=1, PGD and E-MRT boil down to MRTtowards the deterministic primary equivalent channel as in RISliteratures [70]. Besides, the difference between E-MRT and D-MRT can be insignificant for dispersed RIScatter nodes. Thoseobservations prove the proposed PGD active beamformingdesign can exploit the CSI, QoS, and backscatter constellationto balance the primary equivalent SNR and backscatter energydifference and enlarge the achievable rate region for RIScatter.3) Decision Threshold: We evaluate the following decisionthreshold strategies for problem (32):• DP: Benchmark DP method for sequential quantizer [64];• SMAWK: DP accelerated by the SMAWK algorithm [64];• Bisection: The traverse-then-bisect algorithm [65];• ML: Maximum likelihood detector (33) [66].Fig. 10(c) reveals the average achievable rate region forthose strategies. The distribution-aware schemes DP, SMAWKand Bisection ensure higher total backscatter rate than the non-adaptive ML. This is because the total backscatter rate (13)is a function of both input distribution and decision regions,and the rate-optimal threshold design heavily depends on inputdistribution. For example, the backscatter symbol tuples withzero input probability should be assigned with empty decisionregions, in order to increase the success detection chances of1201234567800.10.20.30.40.50.60.70.8Primary Rate [bits/s/Hz]Total Backscatter Rate [bits/BB]K =1K =2K =4K =8(a) RIScatter Nodes01234567800.020.040.060.080.10.120.140.16Primary Rate [bits/s/Hz]Backscatter Rate [bits/BB]M =2M =4M =8M =16(b) Input States012345678900.20.40.60.81Primary Rate [bits/s/Hz]Total Backscatter Rate [bits/BB]Q=1Q=2Q=4Q=8(c) Transmit Antennas01234567800.0050.010.0150.020.0250.030.035Primary Rate [bits/s/Hz]Total Backscatter Rate [bits/PB]N =10N =20N =40N =80(d) Symbol Period Ratio02468101200.20.40.60.811.21.41.6Primary Rate [bits/s/Hz]Total Backscatter Rate [bits/BB]σ2n=−20 dBmσ2n=−30 dBmσ2n=−40 dBmσ2n=−50 dBm(e) Average Noise Power012345678900.20.40.60.811.21.4Primary Rate [bits/s/Hz]Total Backscatter Rate [bits/BB]r=0.5r=1r=2r=4(f) Disk RadiusFig. 11.Average primary-total-backscatter rate regions for different systemconfigurations.other hypotheses. It highlights the importance of joint inputdistribution and decision threshold design in rate maximizationproblems.E. Rate Region under Different ConfigurationsIn this study, we choose Q = 4, K = 8, M = 2, N = 20,σ2v =−40dBm and r=2m as a reference.1) Number of Nodes: Fig. 11(a) reveals how the numberRIScatter nodes K influence the primary-backscatter tradeoff.Interestingly, we observe that increasing K has a larger benefiton the total backscatter rate than primary. This is becauseeach RIScatter node not only affects the primary equivalentSNR but also influences the relative energy difference thatother nodes can create. To maximize the total backscatter rate,some nodes closer to the user may need to sacrifice their ownrate and use the state that minimizes the primary equivalentchannel strength, in order to increase the backscatter rate ofother nodes. This accounts for the significant primary ratedecrease in the low-ρ regime. On the other hand, when theprimary link is prioritized, the RIScatter nodes boil down toRIS elements and enjoy a classic passive array gain of N 2.2) Number of States: Fig. 11(b) shows the relationshipbetween available reflection states (i.e., QAM order) M andachievable rate regions. We notice increasing the reflectionstates has a marginal effect on both primary and totalbackscatter rates. This is because once the scope of reflectioncoefficient is determined, using denser constellation points maynot create enough phase resolution and energy diversity forprimary and backscatter links. Due to the maximum amplitudenormalization in (2), the average rate region of 8-QAM (pointson rectangle border) is smaller than that of 4-QAM (points onsquare vertices), and the inner constellation points with smalleramplitudes are less frequently used. On the other hand, 16-QAM provides significant higher backscatter rate benefit thanprimary. It motivates one to use high-resolution metamaterialunits as RIScatter nodes.3) Number of Transmit Antennas: Fig. 11(c) illustrates theimpact of transmit antennas Q on the average performance. AsQ increases, more transmitter-scatter-receiver paths becomeavailable to PGD and their diversity can be further exploitedto enhance the primary equivalent SNR and backscatter energydifference, which leads to a larger achievable rate region. Itemphasizes the importance of multi-antenna RIScatter systemsand demonstrate the effectiveness of the proposed PGD design.4) Symbol Period Ratio: Fig. 11(d) presents how symbolperiod ratio N affects the achievable rate region, where theunit of total backscatter rate is bits per primary block (toindicate throughput). Apparently, using a very large N canseverely constrain the backscatter throughput, since the gain inenergy certainty (by law of large numbers) cannot withstandthe loss in gross rate. As N → ∞, RIScatter nodes boildown to RIS elements with fixed reflection pattern duringwhole channel block and the total backscatter rate approaches0. On the other hand, when N is too small, the equivalentDMTMAC (9) has very high error probabilities and the energydetection can be unreliable. It explains the observation thatN =10 provides lower backscatter achievable rate than N =20,although nodes are encoding at higher gross rate. Therefore,we conclude the design of symbol period ratio N in RIScattersystems should account for multiple factors (e.g., data raterequirements, physical constraints at the nodes, and signalprocessing capability at the user).5) Average Noise Power: Fig. 11(e) depicts the impact ofaverage noise power σ2v on average rate regions. It shows theproposed practical semi-coherent backscatter energy detectionis suitable for a wide range of noise levels. When σ2v relativelyhigh, one can choose a larger N to maintain the backscatterSNR for better detection performance.6) Coverage Disk Radius: Fig. 11(f) shows the relationshipbetween disk radius r and achievable rate region. We observeboth primary and backscatter performance are enhanced whennodes are located closer to the user. This is because the doublefading effect is less severe for near-far setups. In a multi-user RIScatter system with dispersed nodes, each node maybe allocated to the nearest user to guarantee uniformly goodperformance for both links.13VI. CONCLUSIONThis paper introduced RIScatter as a low-power scatter pro-tocol that unifies backscatter modulation and dynamic passivebeamforming by smart input distribution and practical receiverdesign. Starting from scattering principles, we showed howRIScatter nodes include information node of BackCom and re-flect element of RIS as special cases, how they can be built overexisting passive scatter devices, and how they simultaneouslyencode self information and assist legacy transmission. We alsopropose a practical RIScatter receiver that preserves the benefitsof backscatter modulation and dynamic passive beamformingwithout SIC. The achievable primary-total-backscatter rateregion is then studied for a single-user multi-node RIScattersystem, where the input distribution, active beamforming, anddecision thresholds are iteratively updated. Numerical resultsnot only validated the proposed algorithms, but also emphasizedthe importance of adaptive input distribution and cooperativedecoding on both primary and backscatter links.One possible direction is to consider backscatter detectionover the received signal domain rather than energy domain,where multi-antenna and learning-based approaches can bepromising. Another interesting question is how to designRIScatter node and receiver in a multi-user system to fullyexploit the dynamic passive beamforming that naturally originsfrom backscatter modulation.APPENDIXA. Proof of Proposition 1Denote the Lagrange multipliers associated with (20b) and(20c) as {νk}k∈K and {λk,mk}k∈K,mk∈M, respectively. TheLagrangian function of problem (21) is−I(xK)+�kνk��mkPk(xmk)−1�−�k�mkλk,mkPk(xmk)(43)and the KKT conditions are, ∀k,mk,−∇P ⋆k (xmk )I⋆(xK)+ν⋆k −λ⋆k,mk =0,(44a)λ⋆k,mk =0,P ⋆k (xmk)>0,(44b)λ⋆k,mk ≥0,P ⋆k (xmk)=0,(44c)where directional derivative is explicitly written as∇P ⋆k (xmk )I⋆(xK)=I⋆k(xmk)−(1−ρ).(45)Combining (44) and (45), we haveI⋆k(xmk)=ν⋆k +(1−ρ),P ⋆k (xmk)>0,(46a)I⋆k(xmk)≤ν⋆k +(1−ρ),P ⋆k (xmk)=0,(46b)such that�mkP ⋆k (xmk)I⋆k(xmk)=ν⋆k +(1−ρ).(47)On the other hand, by definition (18) we have�mkP ⋆k (xmk)I⋆k(xmk)=I⋆(xK),(48)where the right-hand side is irrelevant to k. (46), (47), and(48) together complete the proof.B. Proof of Proposition 2We first prove sequence (23) is non-decreasing in weightedsum mutual information. Let PK(xmK) = �q∈K Pq(xmq)and P ′K(xmK)=P ′k(xmk) �q∈K\{k} Pq(xmq) be twoprobability distributions with potentially different marginalfor tag k at state mk, and define an intermediate functionJ�PK(xmK), P ′K(xmK)�as (49) at the end of page 14. Itis straightforward to verify J�PK(xmK),PK(xmK)�= I(xK)and J�PK(xmK),P ′K(xmK)�is a concave function for a fixedP ′K(xmK). Setting ∇P ⋆k (xmk )J�PK(xmK),P ′K(xmK)�=0, wehaveS′k(xmk)−S′k(xik)+(1−ρ)log Pk(xik)P ⋆k (xmk) =0,(50)where ik ̸=mk is the reference state andS′k(xmk)≜I′k(xmk)+(1−ρ)�mK\{k}PK\{k}(xmK\{k})×�m′KP(ˆxm′K|xmK)logP ′K(xmK).(51)Evidently, ∀mk ̸=ik, (50) boils down toP ⋆k (xmk)=P ′k(xmk)exp�ρ1−ρI′k(xmk)��m′kP ′k(xm′k)exp�ρ1−ρI′k(xm′k)�.(52)Since Pk(xik) = 1−�mk̸=ik P ⋆k (xmk) has exactly the sameform as (52), the choice of reference state ik does not matterand (52) is indeed optimal ∀mk ∈ M. That is, for a fixedP ′K(xmK), choosing Pk(xmk) by (52) ensuresJ�PK(xmK),P ′K(xmK)�≥I′(xK).(53)On the other hand, we also have∆≜I(xK)−J�PK(xmK),P ′K(xmK)�(54a)=(1−ρ)�mkP ′k(xmk)f ′k(xmk)�m′kP ′k(xm′k)f ′k(xm′k)�m′′KP(ˆxm′′K|xmk)×log�m′kP ′k(xm′k)P(ˆxm′′K|xm′k)f ′k(xmk)�m′kP ′k(xm′k)P(ˆxm′′K|xm′k)f ′k(xm′k)(54b)≥(1−ρ)�mkP ′k(xmk)f ′k(xmk)�m′kP ′k(xm′k)f ′k(xm′k)�m′′KP(ˆxm′′K|xmk)×�1−�m′kP ′k(xm′k)P(ˆxm′′K|xm′k)f ′k(xm′k)�m′kP ′k(xm′k)P(ˆxm′′K|xm′k)f ′k(xmk)�(54c)=0,(54d)where f ′k(xmk) ≜ exp�ρ1−ρI′k(xmk)�and the equality holdsif and only if (52) converges. (53) and (54) together implyI(xK) ≥ I′(xK). Since mutual information is boundedabove, we conclude the sequence (23) is non-decreasing andconvergent in mutual information.Next, we prove any converging point of sequence (23), de-noted as P ⋆k (xmk), fulfills KKT conditions (22). To see this, letD(r)k (xmk)≜ P (r+1)k(xmk)P (r)k (xmk)=f (r)k (xmk)�m′kP (r)k (xm′k)f (r)k (xm′k).(55)14As sequence (23) is convergent, any state with P ⋆k (xmk)>0need to satisfy D⋆k(xmk)≜limr→∞D(r)k (xmk)=1, namelyI⋆k(xmk)= 1−ρρlog�m′kP ⋆k (xm′k)f ⋆k(xm′k),(56)which is reminiscent of (46a) and (22a). That is, givenP (0)k (xmk)>0, any converging point with P ⋆k (xmk)>0 mustsatisfy (22a). On the other hand, we assume P ⋆k (xmk) doesnot satisfy (22b), such that for any state with P ⋆k (xmk)=0,I⋆k(xmk)>I⋆(xK)=�m′kP ⋆k (xm′k)I⋆k(xm′k),(57)wheretheequalityinheritsfrom(19).Sincetheexponential function is monotonically increasing, we havef ⋆k(xmk)>�m′kP ⋆k (xm′k)f ⋆k(xm′k) and D⋆k(xmk)>1. Consid-ering P (0)k (xmk)>0 and P ⋆k (xmk)=0, it contradicts withP (r)k (xmk)=P (0)k (xmk)r�n=1D(n)k (xmk).(58)That is, given P (0)k (xmk) > 0, any converging point withP ⋆k (xmk)=0 must satisfy (22b). The proof is thus completed.REFERENCES[1] C. Boyer and S. Roy, “Backscatter communication and rfid: Coding,energy, and mimo analysis,” IEEE Transactions on Communications,vol. 62, pp. 770–785, Mar 2014.[2] D. Dobkin, The RF in RFID: Passive UHF RFID in Practice.Newnes,Nov 2012.[3] J. Landt, “The history of rfid,” IEEE Potentials, vol. 24, pp. 8–11, Oct2005.[4] G. Vannucci, A. Bletsas, and D. Leigh, “A software-defined radiosystem for backscatter sensor networks,” IEEE Transactions on WirelessCommunications, vol. 7, pp. 2170–2179, Jun 2008.[5] S. D. Assimonis, S. N. Daskalakis, and A. Bletsas, “Sensitive andefficient rf harvesting supply for batteryless backscatter sensor networks,”IEEE Transactions on Microwave Theory and Techniques, vol. 64, pp.1327–1338, Apr 2016.[6] V. Liu, A. Parks, V. Talla, S. Gollakota, D. Wetherall, and J. R. Smith,“Ambient backscatter: Wireless communication out of thin air,” ACM SIG-COMM Computer Communication Review, vol. 43, pp. 39–50, Sep 2013.[7] G. Yang, Q. Zhang, and Y.-C. Liang, “Cooperative ambient backscattercommunications for green internet-of-things,” IEEE Internet of ThingsJournal, vol. 5, pp. 1116–1130, Apr 2018.[8] Y.-C. Liang, Q. Zhang, E. G. Larsson, and G. Y. Li, “Symbiotic radio:Cognitive backscattering communications for future wireless networks,”IEEE Transactions on Cognitive Communications and Networking,vol. 6, pp. 1242–1255, Dec 2020.[9] Q. Wu, S. Zhang, B. Zheng, C. You, and R. Zhang, “Intelligent reflectingsurface-aided wireless communications: A tutorial,” IEEE Transactionson Communications, vol. 69, pp. 3313–3351, May 2021.[10] H. Guo, Y.-C. Liang, R. Long, and Q. Zhang, “Cooperative ambientbackscatter system: A symbiotic radio paradigm for passive iot,” IEEEWireless Communications Letters, vol. 8, pp. 1191–1194, Aug 2019.[11] H. Ding, D. B. da Costa, and J. Ge, “Outage analysis for cooperativeambient backscatter systems,” IEEE Wireless Communications Letters,vol. 9, pp. 601–605, May 2020.[12] R. Long, Y.-C. Liang, H. Guo, G. Yang, and R. Zhang, “Symbioticradio: A new communication paradigm for passive internet of things,”IEEE Internet of Things Journal, vol. 7, pp. 1350–1363, Feb 2020.[13] S. Zhou, W. Xu, K. Wang, C. Pan, M.-S. Alouini, and A. Nallanathan,“Ergodic rate analysis of cooperative ambient backscatter communication,”IEEE Wireless Communications Letters, vol. 8, pp. 1679–1682, Dec 2019.[14] T. Wu, M. Jiang, Q. Zhang, Q. Li, and J. Qin, “Beamforming designin multiple-input-multiple-output symbiotic radio backscatter systems,”IEEE Communications Letters, vol. 25, pp. 1949–1953, Jun 2021.[15] J. Xu, Z. Dai, and Y. Zeng, “Enabling full mutualism for symbioticradio with massive backscatter devices,” arXiv:2106.05789, Jun 2021.[16] Z. Yang and Y. Zhang, “Optimal swipt in ris-aided mimo networks,”IEEE Access, vol. 9, pp. 112 552–112 560, 2021.[17] S. Han, Y.-C. Liang, and G. Sun, “The design and optimization of randomcode assisted multi-bd symbiotic radio system,” IEEE Transactions onWireless Communications, vol. 20, pp. 5159–5170, Aug 2021.[18] Q. Zhang, Y.-C. Liang, H.-C. Yang, and H. V. Poor, “Mutualisticmechanism in symbiotic radios: When can the primary and secondarytransmissions be mutually beneficial?” IEEE Transactions on WirelessCommunications, vol. 1276, pp. 1–1, 2022.[19] R. Torres, R. Correia, N. Carvalho, S. N. Daskalakis, G. Goussetis,Y. Ding, A. Georgiadis, A. Eid, J. Hester, and M. M. Tentzeris,“Backscatter communications,” IEEE Journal of Microwaves, vol. 1, pp.864–878, Oct 2021.[20] G. Vougioukas and A. Bletsas, “Switching frequency techniques foruniversal ambient backscatter networking,” IEEE Journal on SelectedAreas in Communications, vol. 37, pp. 464–477, Feb 2019.[21] Q. Wu and R. Zhang, “Intelligent reflecting surface enhanced wirelessnetwork: Joint active and passive beamforming design,” vol. 18.IEEE,Dec 2018, pp. 1–6.[22] S. Zhang and R. Zhang, “Capacity characterization for intelligentreflecting surface aided mimo communication,” IEEE Journal onSelected Areas in Communications, vol. 38, pp. 1823–1838, Aug 2020.[23] S. Lin, B. Zheng, F. Chen, and R. Zhang, “Intelligent reflectingsurface-aided spectrum sensing for cognitive radio,” IEEE WirelessCommunications Letters, vol. 11, pp. 928–932, May 2022.[24] Y. Liu, Y. Zhang, X. Zhao, S. Geng, P. Qin, and Z. Zhou, “Dynamic-controlledrisassistedmulti-usermisodownlinksystem:Jointbeamforming design,” IEEE Transactions on Green Communicationsand Networking, vol. 6, pp. 1069–1081, Jun 2022.[25] Z. Feng, B. Clerckx, and Y. Zhao, “Waveform and beamformingdesign for intelligent reflecting surface aided wireless power transfer:Single-user and multi-user solutions,” IEEE Transactions on WirelessCommunications, 2022.[26] Y. Zhao, B. Clerckx, and Z. Feng, “Irs-aided swipt: Joint waveform,active and passive beamforming design under nonlinear harvester model,”IEEE Transactions on Communications, vol. 70, pp. 1345–1359, 2022.[27] Y. Yang, S. Zhang, and R. Zhang, “Irs-enhanced ofdma: Joint resourceallocation and passive beamforming optimization,” IEEE WirelessCommunications Letters, vol. 9, pp. 760–764, Jun 2020.[28] Q. Wu, X. Zhou, and R. Schober, “Irs-assisted wireless powered noma:Do we really need different phase shifts in dl and ul?” IEEE WirelessCommunications Letters, vol. 10, pp. 1493–1497, Jul 2021.[29] Q. Wu, X. Zhou, W. Chen, J. Li, and X. Zhang, “Irs-aided wpcns:A new optimization framework for dynamic irs beamforming,” IEEETransactions on Wireless Communications, pp. 1–1, Dec 2021.[30] M. Hua and Q. Wu, “Joint dynamic passive beamforming and resourceallocation for irs-aided full-duplex wpcn,” IEEE Transactions onWireless Communications, pp. 1–1, Dec 2021.[31] W. Tang, J. Y. Dai, M. Chen, X. Li, Q. Cheng, S. Jin, K. Wong, andT. J. Cui, “Programmable metasurface-based rf chain-free 8psk wirelesstransmitter,” Electronics Letters, vol. 55, pp. 417–420, Apr 2019.[32] J. Y. Dai, W. Tang, L. X. Yang, X. Li, M. Z. Chen, J. C. Ke,Q. Cheng, S. Jin, and T. J. Cui, “Realization of multi-modulationschemes for wireless communication by time-domain digital codingmetasurface,” IEEE Transactions on Antennas and Propagation, vol. 68,pp. 1618–1627, Mar 2020.[33] R. Karasik, O. Simeone, M. D. Renzo, and S. S. Shitz, “Beyondmax-snr: Joint encoding for reconfigurable intelligent surfaces,” vol.2020-June.IEEE, Jun 2020, pp. 2965–2970.J�PK(xmK),P ′K(xmK)�≜�mKPK(xmK)�ρlog�1+ |hHE(xmK)w|2σ2v�+(1−ρ)�m′KP(ˆxm′K|xmK)logP(ˆxm′K|xmK)P ′K(xmK)P ′K(ˆxm′K)PK(xmK)�.(49)15[34] R. Liu, H. Li, M. Li, and Q. Liu, “Symbol-level precoding design forintelligent reflecting surface assisted multi-user mimo systems.”IEEE,Oct 2019, pp. 1–6.[35] A. Bereyhi, V. Jamali, R. R. Muller, A. M. Tulino, G. Fischer, andR. Schober, “A single-rf architecture for multiuser massive mimo viareflecting surfaces.”IEEE, May 2020, pp. 8688–8692.[36] X. Xu, Y.-C. Liang, G. Yang, and L. Zhao, “Reconfigurable intelligentsurface empowered symbiotic radio over broadcasting signals,” vol.2020-Janua.IEEE, Dec 2020, pp. 1–6.[37] Q. Zhang, Y.-C. Liang, and H. V. Poor, “Reconfigurable intelligentsurface assisted mimo symbiotic radio networks,” IEEE Transactionson Communications, vol. 69, pp. 4832–4846, Jul 2021.[38] J. Hu, Y. C. Liang, and Y. Pei, “Reconfigurable intelligent surfaceenhanced multi-user miso symbiotic radio system,” IEEE Transactionson Communications, vol. 69, pp. 2359–2371, Apr 2021.[39] M. Hua, Q. Wu, L. Yang, R. Schober, and H. V. Poor, “A novelwireless communication paradigm for intelligent reflecting surface basedsymbiotic radio systems,” IEEE Transactions on Signal Processing,vol. 70, pp. 550–565, Apr 2022.[40] E. Basar, “Reconfigurable intelligent surface-based index modulation:A new beyond mimo paradigm for 6g,” IEEE Transactions onCommunications, vol. 68, pp. 3187–3196, May 2020.[41] T. Ma, Y. Xiao, X. Lei, P. Yang, X. Lei, and O. A. Dobre, “Largeintelligent surface assisted wireless communications with spatialmodulation and antenna selection,” IEEE Journal on Selected Areas inCommunications, vol. 38, pp. 2562–2574, Nov 2020.[42] J. Yuan, M. Wen, Q. Li, E. Basar, G. C. Alexandropoulos, and G. Chen,“Receive quadrature reflecting modulation for ris-empowered wirelesscommunications,” IEEE Transactions on Vehicular Technology, vol. 70,pp. 5121–5125, May 2021.[43] S. Hu, C. Liu, Z. Wei, Y. Cai, D. W. K. Ng, and J. Yuan, “Beamformingdesign for intelligent reflecting surface-enhanced symbiotic radiosystems,” arxiv:2110.10316, Oct 2021.[44] Y. C. Liang, Q. Zhang, J. Wang, R. Long, H. Zhou, and G. Yang,“Backscatter communication assisted by reconfigurable intelligentsurfaces,” Proceedings of the IEEE, 2022.[45] Q. Wu and R. Zhang, “Towards smart and reconfigurable environment:Intelligentreflectingsurfaceaidedwirelessnetwork,”IEEECommunications Magazine, vol. 58, pp. 106–112, 1 2020.[46] R. Hansen, “Relationships between antennas as scatterers and asradiators,” Proceedings of the IEEE, vol. 77, pp. 659–662, May 1989.[47] S. J. Thomas, E. Wheeler, J. Teizer, and M. S. Reynolds, “Quadratureamplitude modulated backscatter in passive and semipassive uhf rfidsystems,” IEEE Transactions on Microwave Theory and Techniques,vol. 60, pp. 1175–1182, Apr 2012.[48] J. Kim and B. Clerckx, “Wireless information and power transfer foriot: Pulse position modulation, integrated receiver, and experimentalvalidation,” IEEE Internet of Things Journal, vol. 9, pp. 12 378–12 394,Jul 2022.[49] X. He, W. Jiang, M. Cheng, X. Zhou, P. Yang, and B. Kurkoski,“Guardrider: Reliable wifi backscatter using reed-solomon codes withqos guarantee.”IEEE, Jun 2020, pp. 1–10.[50] Y. Huang, A. Alieldin, and C. Song, “Equivalent circuits and analysisof a generalized antenna system,” IEEE Antennas and PropagationMagazine, vol. 63, pp. 53–62, Apr 2021.[51] D. Bharadia, K. R. Joshi, M. Kotaru, and S. Katti, “Backfi: Highthroughput wifi backscatter,” vol. 45.ACM, Aug 2015, pp. 283–296.[52] G. Yang, C. K. Ho, and Y. L. Guan, “Multi-antenna wireless energytransfer for backscatter communication systems,” IEEE Journal onSelected Areas in Communications, vol. 33, pp. 2974–2987, Dec 2015.[53] H. Guo, Q. Zhang, S. Xiao, and Y.-C. Liang, “Exploiting multipleantennas for cognitive ambient backscatter communication,” IEEEInternet of Things Journal, vol. 6, pp. 765–775, Feb 2019.[54] M. Jin, Y. He, C. Jiang, and Y. Liu, “Parallel backscatter: Channelestimation and beyond,” IEEE/ACM Transactions on Networking, vol. 29,pp. 1128–1140, Jun 2021.[55] B. Zheng and R. Zhang, “Intelligent reflecting surface-enhancedofdm: Channel estimation and reflection optimization,” IEEE WirelessCommunications Letters, vol. 9, pp. 518–522, Apr 2020.[56] C. You, B. Zheng, and R. Zhang, “Intelligent reflecting surface withdiscrete phase shifts: Channel estimation and passive beamforming.”IEEE, Jun 2020, pp. 1–6.[57] J. Qian, F. Gao, G. Wang, S. Jin, and H. Zhu, “Semi-coherentdetection and performance analysis for ambient backscatter system,” IEEETransactions on Communications, vol. 65, pp. 5266–5279, Dec 2017.[58] T. Nguyen, Y.-J. Chu, and T. Nguyen, “On the capacities of discrete mem-oryless thresholding channels,” vol. 2018-June. IEEE, Jun 2018, pp. 1–5.[59] T. Nguyen and T. Nguyen, “Optimal quantizer structure for maximizingmutualinformationunderconstraints,”IEEETransactionsonCommunications, vol. 69, pp. 7406–7413, Nov 2021.[60] M. Rezaeian and A. Grant, “Computation of total capacity fordiscrete memoryless multiple-access channels,” IEEE Transactions onInformation Theory, vol. 50, pp. 2779–2784, Nov 2004.[61] J. Buhler and G. Wunder, “A note on capacity computation for thediscrete multiple access channel,” IEEE Transactions on InformationTheory, vol. 57, pp. 1906–1910, Apr 2011.[62] G. J. O. Jameson, “The incomplete gamma functions,” The MathematicalGazette, vol. 100, pp. 298–306, Jul 2016.[63] S. Boyd and L. Vandenberghe, Convex Optimization.CambridgeUniversity Press, Mar 2004.[64] X. He, K. Cai, W. Song, and Z. Mei, “Dynamic programming for sequen-tial deterministic quantization of discrete memoryless channels,” IEEETransactions on Communications, vol. 69, pp. 3638–3651, Jun 2021.[65] T. Nguyen and T. Nguyen, “On thresholding quantizer design for mutualinformation maximization: Optimal structures and algorithms,” vol.2020-May.IEEE, May 2020, pp. 1–5.[66] J. Qian, A. N. Parks, J. R. Smith, F. Gao, and S. Jin, “Iot communicationswith m-psk modulated ambient backscatter: Algorithm, analysis, andimplementation,” IEEE Internet of Things Journal, vol. 6, pp. 844–855,Feb 2019.[67] J. Wang, N. Li, W. Shi, Y. Ma, X. Liang, and X. Dong, “Capacity of60 ghz wireless communications based on qam,” Journal of AppliedMathematics, vol. 2014, pp. 1–5, 2014.[68] B. W. Bader and T. G. Kolda, “Tensor toolbox for matlab,”https://www.tensortoolbox.org/, Sep 2022.[69] E. Calvo, D. P. Palomar, J. R. Fonollosa, and J. Vidal, “On thecomputation of the capacity region of the discrete mac,” IEEETransactions on Communications, vol. 58, pp. 3512–3525, Dec 2010.[70] Q. Wu and R. Zhang, “Intelligent reflecting surface enhanced wirelessnetwork via joint active and passive beamforming,” IEEE Transactionson Wireless Communications, vol. 18, pp. 5394–5409, Nov 2019. | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/328254045Twins’ and their Mothers’ Personality: Temperament and Character ClustersPoster · April 2019CITATIONS0READS1177 authors, including:Some of the authors of this publication are also working on these related projects:Dark Character Theory: The Dark Cube and other Person-Centered Methods in the Study of Malevolent Character View projectPerson-Centered Schools: Personality, Well-Being, and Resilience View projectDanilo GarciaLinköping University380 PUBLICATIONS 3,313 CITATIONS SEE PROFILEAli Al NimaRegion Blekinge85 PUBLICATIONS 712 CITATIONS SEE PROFILEThomas NilssonUniversity of Gothenburg83 PUBLICATIONS 1,498 CITATIONS SEE PROFILESebastian LundströmUniversity of Gothenburg174 PUBLICATIONS 3,803 CITATIONS SEE PROFILEAll content following this page was uploaded by Danilo Garcia on 11 December 2019.The user has requested enhancement of the downloaded file. Karlskrona 11th of December 2019 Dear colleague, We are very grateful for your interest in the conference presentation. We are in the process of writing the paper that will be submitted to a scientific journal. Therefore, we are at this moment unable to send you a copy of the conference presentation. If permitted by copyright laws, we will be uploading it as soon as the article is accepted for publication. On behalf of my co-authors Danilo Garcia, PhD Head of Research Blekinge Center of Competence Research and Innovation for the Development of Health and Health Equality Region Blekinge Sweden E-mail: danilo.garcia@icloud.com Associate Professor Department of Psychology University of Gothenburg Sweden View publication stats | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/341756411A general framework of genetic multi-agent routing protocol for improvingthe performance of MANET environmentArticle in International Journal of Artificial Intelligence · June 2020DOI: 10.11591/ijai.v9.i2.pp310-316CITATIONS14READS2138 authors, including:Some of the authors of this publication are also working on these related projects:DDoS attack defense model View projectGa-based feature subset selection in a spam/non-spam detection system View projectMustafa Hamid HassanIraq at Imam Ja'afar Al-sadiq University63 PUBLICATIONS 373 CITATIONS SEE PROFILEMohammed Ahmed JubairImam Ja'afar Al-sadiq University55 PUBLICATIONS 340 CITATIONS SEE PROFILESalama MostafaUniversiti Tun Hussein Onn Malaysia237 PUBLICATIONS 4,326 CITATIONS SEE PROFILEHazalila KamaludinUniversiti Tun Hussein Onn Malaysia13 PUBLICATIONS 95 CITATIONS SEE PROFILEAll content following this page was uploaded by Mustafa Hamid Hassan on 30 May 2020.The user has requested enhancement of the downloaded file.IAES International Journal of Artificial Intelligence (IJ-AI) Vol. 9, No. 2, June 2020, pp. 310~316 ISSN: 2252-8938, DOI: 10.11591/ijai.v9.i2.pp310-316 310 Journal homepage: http://ijai.iaescore.com A general framework of genetic multi-agent routing protocol for improving the performance of MANET environment Mustafa Hamid Hassan, Mohammed Ahmed Jubair, Salama A. Mostafa, Hazalila Kamaludin, Aida Mustapha, Mohd Farhan Md. Fudzee, Hairulnizam Mahdin Faculty of Computer Science and Information Technology, Universiti Tun Hussin Onn Malaysia, Johor, Malaysia Article Info ABSTRACT Article history: Received Jan 30, 2020 Revised Apr 20, 2020 Accepted Apr 29, 2020 These days, the fields of Mobile Ad hoc Network (MANET) have provided increasing prevalence and consequently, MANET is now a subject of considerable significance for the researchers to instigate research activities. MANET is the collaborative commitment of an assemblage of portable (or mobile) hubs (or nodes) without the necessary mediation of any unified (or centralized) gateway (or access point) or existent framework. There exists a growing inclination or course to embrace MANET for business utilization. MANET is a rising domain of research to give different services in communication to end-clients or consumers. However, these communication services of MANET utilize a large amount of transfer speed (or bandwidth) and a huge measure of web speed. Bandwidth optimization is essential in different information interchanges for fruitful acknowledgement and the application of such a technological innovation. This paper integrates the Genetic Algorithm (GA) and the Multi-Agent System (MAS) to improve the QoS requirements. The proposed framework called Genetic Multi-Agent Routing Protocol (GMARP). The aims of the proposed framework are to utilize the benefits of both approaches in order to fulfil QoS such as (delay, bandwidth, and the number of hops) in the different types of routing conventions (or protocols) such as being (proactive and reactive). In this paper is a simulation scenario to demonstrate the ability of the proposed framework to be satisfied with QoS requirements. Keywords: Bandwidth Delay Genetic algorithm Mobile Ad-hoc network Multi-agent system This is an open access article under the CC BY-SA license. Corresponding Author: Salama A. Mostafa Faculty of Computer Science and Information Technology, Universiti Tun Hussein Onn Malaysia (UTHM), 86400, Parit Raja, Batu Pahat, Johor, Malaysia. Email: salama@uthm.edu.my 1. INTRODUCTION The rise of mobile computational gadgets and wireless (or remote) correspondence innovation as of late has brought about the quick development in the quantity of portable terminal clients. A MANET is a multi-hop remote system in which the system segments, for example, PCs, individual computerized associates and remote telephones are portable [1-3]. Generally, the hubs inside the MANET fill in as routers. While relaying with different hubs, these hubs meander unreservedly. The utilization of MANET is increasingly appropriate in instances of inaccessibility of infrastructure and/or additionally costly sending of hub portability. The MANET can be utilized in the development of a cheap and economical system at any place, as it is efficient and needs no specific gateways or hubs. It is viewed as a decent alternative for application in systems in circumstances of crisis or catastrophes. Furthermore, the hubs present in MANETs Int J Artif Intell ISSN: 2252-8938 A general framework of genetic multi-agent routing protocol for... (Mustafa Hamid Hassan) 311 can speak with one another without a manager, as a result of the existence of a disseminated control framework [4-6]. Regardless of the benefits associated with MANET, it additionally has a few constraints which do not exist in fixed systems. For example, the versatility of hubs brings about steady alterations in the topology of the network, in this way resulting in a high transmission of data in the system. Other restrictions of the MANET are the little limit of batteries and constrained data transfer capacity for remote channels. Once more, communication may get unrealizable and the QoS may deteriorate because of information get to that is centred on a solitary point. MANET disseminates important and real-time information to the nodes such as weather information, transit systems, internet access, mobile e-commerce, and other multimedia applications. Most of these applications or systems demand a huge bandwidth capacity and a large measure of internet speed so that users can communicate among themselves. Most of the previous research on ad hoc networking has been done using exist many techniques of bandwidth optimization in MANET. MANETs in recent times have become the choice wireless network due to the numerous advantages it proffers. In a wired system, the accessible data transfer capacity estimation is carried out utilizing a functioning estimation procedure [7-9]. This procedure is inappropriate for MANET since it utilizes test packets when estimating the accessible data transfer capacity present for the source and destination. On the possibility that the quantity of sources to destination pair is sufficiently huge, it will bring about sending more test packets which thus expends a lot of bandwidth. Sarr et al. [10] propose an available bandwidth-based flow affirmation (or admission) control (ABE) calculation for a remote network. Estimation of the accessible transmission capacity is carried out by utilizing the remote channel detecting mechanism. The drawbacks of this technique are when there is an increment in the information traffic load inside a system, the only factor considered is the additional back-off overhead. Other important factors, the collision likelihood is calculated without regarding the hidden and exposed node causing unnecessary delay. An improved available bandwidth (IAB) has been proposed by Zhao et al. [11]. This protocol estimates the available bandwidth of a giving link for QoS support in a wireless ad-hoc network. It considers the synchronization between the source and the destination node by differentiating the busyness caused by the transmitting and receiving node from those caused by the sensing node. Furthermore, the work also improved the accuracy of estimating the overlapping probability of the idle time of two adjacent nodes. The drawback of this technique is also Collision probability is calculated without considering the hidden and exposed node causing unnecessary delay. Cognitive passive estimation of accessible bandwidth (cPEAB) was proposed by [12]. This convention assesses the accessible data transfer capacity of a system in a covered WiFi condition. It considers the extra expense brought about by affirmation frames, which were not considered in both AAC and ABE, therefore assessing the accessible transmission capacity by estimating the extent of waiting and backoff delays, packet collision likelihood, affirmation postponement, and channel inactive time. Nam et al. [13] improved on the work of [7] by enhancing its algorithm to incorporate the retransmission system and back-off cost. The downside of this strategy is that the conflict window expense was not considered with an increase in information traffic load inside the system. However, more time delay and a big amount of packet drop are happening in many schemes like Leaky bucket and Token bucket. In the Token bucket method packet drop is happened less than the Leaky bucket but the time delay is relatively more than the Leaky bucket algorithm. So, in order to improve the performance of mobile ad hoc network (MANET) bandwidth optimization is highly desirable.in this work proposed framework called Genetic Multi-Agent Routing Protocol (GMARP). The aims of the proposed framework are to utilize the benefits of both approaches in order to fulfil QoS such as (delay, bandwidth, and a number of hops) in the different types of routing protocols like (proactive and reactive). 2. MATERIALS AND METHODS 2.1. Genetic algorithm The evolution of Genetic algorithms (GA) was inspired by closely examining and replicating biological evolutions such as reproduction, recombination, and mutation. Over the years the genetic algorithm has become one of the most important methods for obtaining approximate solutions of optimization problems. GA works by letting the competing variables interact with one another to evolve a potential solution naturally. A large number of optimization problems of various types spanning diverse areas of engineering and natural sciences have been solved by the application of GA. The basic principles of GA were introduced, for the first time, by John Henry Holland [14] for solving practical optimization problems and are well documented [15-19]. The point of using a GA is to accomplish better outcomes via selection, hybrid, and transformation (or mutation). The accomplishment of any GA relies upon the blueprint of its ISSN: 2252-8938 Int J Artif Intell, Vol. 9, No. 2, June 2020: 310 – 316 312 search controllers (or operators) including their suitable combination. There is a great deal of writing to enhance the viability of GA operators. In a crossover operator, new strings are created by trading data among strings of mating pools. Numerous hybrid (or crossover) operators are present in the literature of GA. In many crossover operators, two strings are randomly selected from the mating pool and a few bits of the strings are traded between the strings. The two strings taking part in the crossover activity are referred to as parent strings and the subsequent strings are called offspring strings. Practically speaking, for crossover activity with the goal that a portion of the good strings might be protected all parents in the mating pool are not chosen. This is accomplished by choosing a fixed amount of parents from the mating pool. This is called the crossover likelihood. The transformation or mutation is applied to each child individually after crossover. This operator flips or modifies randomly at least one piece values at arbitrarily chosen areas in a chromosome with a small probability. This is not an assurance that global maxima will be achieved. In any case, notwithstanding brute force, GA does not ensure management of non-inconsequential issues. In any case, the chance of being attached to neighbourhood maxima at beginning stages is a problem that one would need to handle for example with a type of reproduced stimulating the rate of decay of mutation. GA takes time before reaching convergence. A nicely estimated populace and various breeds are vital before achieving expected outcomes. With high reproduction/simulation, a resolution will regularly take days to be figured out. Figure 1 shows the main steps of GA. Figure 1. A basic procedure of GA 2.2. Multi-agent system A Multi-Agent framework is portrayed as a lot of insightful specialists or agents that have a connection with one another inside a domain [20]. These agents work together to tackle issues that they can't settle autonomously [21]. Agents have such a large number of attributes which incorporate adjustment, self-governance, adaptability, responsiveness, appropriation and neighbourhood outlook or perspective In view of accomplishing a specific objective, the agents require to collaborate with one another [22-24]. MAS has distinctive application areas, and some of them incorporate aeroplane support or conservation, web agents, observation, ecological check and observations, medical services, military demining, control of spacecraft and industrial control [24-25]. The reason why these agents are used in the kind of systems is to improve: 1. the speed performance and efficiency, 2. the flexibility and scalability of the operations, and 3. the reusability of the system modules. The multi-agents have been used by many researchers to propose different systems based on dynamic routing. For the most part, an agent is an autonomous element that completes single or numerous undertakings in order to achieve a few objectives. In the networking area, an agent works constantly regardless of whether clients disengage from the system. While a few agents run on standardized platforms, some others operate in servers dedicated for such. Figure 2 depicts a simple example of a MAS. Initial Population NEvaluate IndividualSelectionCrossover and mutationNew populationNEndInt J Artif Intell ISSN: 2252-8938 A general framework of genetic multi-agent routing protocol for... (Mustafa Hamid Hassan) 313 Figure 2. A Multi-Agent System 2.3. Genetic algorithm based multi-agent system Generally, MANET is an assortment of hubs that interface between one another without a framework, there are various sorts of routing conventions that are employed to establish the path between nodes in MANET, there are three main kinds of routing protocols; reactive, proactive and hybrid [20]. All of them create the path based on two kinds of messages; Route Request (RREQ) and Route Reply (RREP). The RREQ used to find all hubs/nodes in the network. Then again, all potential paths between source and destination hubs are found via the RREP message. In the reactive routing protocol, the path choice is relay on the minimum delay rather than other parameters like energy, and bandwidth. Otherwise, the proactive routing protocol depends on the shortest path without care to other parameters. These mechanisms achieved a good result; however, it still required more improvement due to other parameters like; bandwidth and energy. This paper employs one of the most popular optimization algorithms that achieved many successes in several domains (like ref), this algorithm is called the Genetic Algorithm (GA). The fundamental GA architecture incorporates chromosome depiction, fitness function, selection, crossover and change operators. GAs initially represents it as an assortment of individually named chromosomes by an encoding scheme. As indicated by the probabilities of crossover and transformation/mutation, these chromosomes are iteratively chosen to produce new chromosomes by crossover and mutation operators. In the interim, all chromosomes compete with one another via a quality measure known as fitness in every alteration cycle. Since crossover and mutation are the operators to create new chromosomes and the probability of the occurrence of a crossover is typically much larger than that of a mutation, the crossover operator constitutes the core operation in GAs. However, GA still suffering from several problems: Global maxima accomplishment is not guaranteed by GA. GA does not ensure management of non-inconsequential issues. The chance of being fixed to nearby maxima at starting stages is a problem that requires handling for example with a type of simulated galvanizing rate of mutation decay. GA takes time before reaching convergence. An adequately estimated populace and various breeds are fundamental preceding accomplishing expected outcomes. With powerful simulation, several days pass before solutions are achieved. To overcome these issues, the Multi-Agent System (MAS) is present to integrate with GA to tackle the aforementioned problems. The MAS is a new technique that applied in different domains and achieves a promising result. In this paper, the GA and MAS (GMARP) will be merged to propose a new framework that used to improve the routing process in MANET. The path selection mechanism will be relay on three QoS parameters such, Number of Hops (NH), Bandwidth (BW), and Delay. The following algorithm depicted the main steps of the GMARP framework. Firstly, determine the source and destination hubs and the source hub communicate a RREQ to locate a particular destination, besides, the destination hub sends a RREP to the source, this procedure finds every conceivable path between the source and destination hubs. At that point, the GMARP saves all paths and assesses the nature or value of them by concentrating on three QoS metrics; NH, BW, and delay, moreover, it categorizes these paths dependent on high value or quality. In the end, it chooses the best path through which data is sent to the destination. In figuring out these metrics, the following formulas: (1-3) are respectively utilized. The paths are categorized by their quality to allow or facilitate the choice of the path; beginning with the path that accomplishes less NH, deferral or delay, and maximum BW. The following equations are used in calculating the best routes: ISSN: 2252-8938 Int J Artif Intell, Vol. 9, No. 2, June 2020: 310 – 316 314 𝑁𝑜. 𝑜𝑓𝐻𝑜𝑝𝑠 = ∑𝑁𝐻𝑆𝐷𝑃𝑖=0 (1) 𝐵𝑊 =∑𝐵𝑊𝑚𝑚𝑗=1ℎ (2) 𝐷𝑒𝑙𝑎𝑦 =∑𝑇𝑟𝑒𝑐𝑒𝑖𝑣𝑒−𝑇𝑠𝑒𝑛𝑡𝑛𝑖=1ℎ∗ 100% (3) where 𝑆, 𝐷 denote the source and destination nodes respectively, 𝑃 refer to the number of paths, 𝑖,𝑛, 𝑗, 𝑚 denote the counter, 𝑇 represents the time and the number of nodes in the path is signified by ℎ. Algorithm: The basic GMARP 1. begin 2. while (t < Max number of iterations) 3. Select the Source and Destination; 4. Source Initialize the route Discovery; 5. Generate initial solutions (routes); 6. 𝛼1evaluate the quality of routes based on NH, BW, Delay; 7. 𝛼2Sorting the routes based on QoS, equations (1-3); 8. 𝛼3Select the optimum routes based on criteria, 𝑓(𝑥) =𝐵𝑎𝑛𝑑𝑤𝑖𝑑𝑡ℎ𝑁𝑜.𝑜𝑓ℎ𝑜𝑝𝑠∗𝐷𝑒𝑙𝑎𝑦; 9. Crossover operator for all routes; 10. Generate new routes and compare with parents; 11. Update the solutions; 12. 𝛼3assess the quality of the new routes; 13. if (new > old) Select best routes 14. else 15. Operate the mutation; 16. Update the solutions (routes); 17. Rank the routes and find the best route; 18. Accept the new solutions; 19. end-if 3. SIMULATION AND RESULTS 3.1. Example scenario Accept a MANET contains 11 hubs/nodes, these nodes are connected with each other with different parameters (QoS). A few paths may be present between the source hub, 𝑆 and the destination hub, 𝐷. The path selection process in MAGA is relay on the values of three parameters (no. of hops, bandwidth, and delay). Figure 3 presents an instance scenario of the system. Figure 3. An example scenario At the underlying stage, a RREQ broadcast is made by the source S node to acquire data about the accessible paths leading to the destination D. In the MAGA routing convention displayed in Figure 3 above, the path with the most transfer speed (or bandwidth), fewer bounces or hops, and deferral (or delay) is given more consideration as shown in Table 1. Int J Artif Intell ISSN: 2252-8938 A general framework of genetic multi-agent routing protocol for... (Mustafa Hamid Hassan) 315 Table 1. Routes details No. Possible Route No. of hop Bandwidth Delay 1 S-1-4-7-D 4 52.5 7 2 S-2-4-7-D 4 55 7 3 S-2-5-8-D 4 57.5 6.7 4 S-2-4-5-7-8-D 6 55 6.5 5 S-2-3-5-9-D 5 64 6.4 6 S-3-6-9-D 4 80 4.5 7 S-3-5-7-8-9-D 6 55 5.5 8 S-3-5-8-D 4 65 6 9 S-3-6-8-D 4 70 6 10 S-1-2-3-6-8-D 6 71.7 6.7 The variation of the NH, BW and Delay for all paths that presented in Table 1 are depicted in Figure 4. Based on Figure 4(a) there are various routes are discovered from the source to the destination nodes to send the data packet, each path has a different number of hops. Otherwise, the bandwidth for each path is shown in Figure 4(b). Finally, the variation of delay is depicted in Figure 4(c). These Figures are used to facilitate the monitoring process. (a) NH (b) BW (c) Delay Figure 4. Variation of QoS parameters 4. CONCLUSION A MANET is a collection of remote/wireless nodes that link together without infrastructure, each node in the network utilize as a sink to collect the data packets or a router to retransmit the data packets. MANET consider an efficient and reliable network for launching communication in a critical circumstance because it has several characteristics. The movement of the nodes are arbitrarily, and the network topology is changed dynamically. Numerous kinds of routing protocols in MANET are proposed, most of them under either reactive or proactive mechanism. However, the routing protocols still suffering from a different problem such as delay, energy consumption, bandwidth, and so on. This paper presented a new framework called Genetic Multi-Agent Routing Protocol (GMARP). The aims of the GMARP framework is to utilize the benefits of both approaches (GA and MAS) in order to fulfil QoS such as (number of hops, bandwidth, and ISSN: 2252-8938 Int J Artif Intell, Vol. 9, No. 2, June 2020: 310 – 316 316 delay) in the different types of routing protocols like (proactive and reactive). The simulation scenario is present to show the ability of the proposed framework to be satisfied with QoS requirements. ACKNOWLEDGEMENTS The authors express appreciation to the Malaysia Ministry of Education (MoE) and Universiti Tun Hussein Onn Malaysia (UTHM). This research is supported by the Fundamental Research Grant Scheme (FRGS) grant vot number K216. REFERENCES [1] P. Joshi, et al., "Comparative Study of Different Routing Protocols for IEEE 802.15. 4-Enabled Mobile Sink Wireless Sensor Network." In Advances in VLSI, Communication, and Signal Processing, pp. 161-170. Springer, Singapore, 2020. [2] K. Prabha, "Performance Assessment and Comparison of Efficient Ad Hoc Reactive and Proactive Network Routing Protocols." SN Computer Science 1, no. 1: 13, 2020. [3] M. H. Hassan, et al., "Integrating African Buffalo Optimization Algorithm in AODV Routing Protocol for improving the QoS of MANET," Journal of Southwest Jiaotong University, 54 , 3 , 2019. [4] M. A Jubair, et al., "Performance Evaluation of Ad-Hoc On-Demand Distance Vector and Optimized Link State Routing Protocols in Mobile Ad-Hoc Networks,". International Journal on Advanced Science, Engineering and Information Technology, 8(4), 1277-1283, 2018. [5] M. A Jubair, et al., "Bat Optimized Link State Routing Protocol for Energy-Aware Mobile Ad-Hoc Networks," Symmetry, 11(11), 1409, 2019. [6] M. A. Jubair et al., "Competitive analysis of single and multi-path routing protocols in mobile Ad-Hoc network," International Journal on Advanced Science, Engineering and Information Technology, Vol 19, No 1, 2020. [7] H. Ziani, et al., "Mobility Condition to Study Performance of MANET Routing Protocols," In Emerging Technologies for Connected Internet of Vehicles and Intelligent Transportation System Networks, pp. 73-82. Springer, Cham, 2020. [8] M. H. Hassan, et al., "Mobile Ad-Hoc Network Routing Protocols of Time-Critical Events for Search and Rescue Missions," Bulletin of Electrical Engineering and Informatics, 2020. In press. [9] S. A. Mostafaet al., "Comparative Analysis to the Performance of Three Mobile Ad-Hoc Network Routing Protocols in Time-Critical Events of Search and Rescue Missions," International Conference on Applied Human Factors and Ergonomics, 2020. In press. [10] C.Sarr, et al., "Bandwidth estimation for IEEE 802.11-based ad hoc networks," IEEE transactions on Mobile Computing, 7, 10, 1228-1241, 2008. [11] H. Zhao, et al., "Accurate available bandwidth estimation in IEEE 802.11-based ad hoc networks," Computer Communications, vol. 32, no. 6, pp. 1050-1057, 2009. [12] S. Tursunova, et al., "Cognitive passive estimation of available bandwidth (cPEAB) in overlapped IEEE 802.11 WiFi WLANs," IEEE Network Operations and Management Symposium-NOMS pp.448-454, IEEE, April, 2010. [13] N. V. Nguye, et al., "Retransmission-based available bandwidth estimation in IEEE 802.11-based multihop wireless networks," In Proceedings of the 14th ACM international conference on Modeling, analysis and simulation of wireless and mobile systems, pp, 377-384, 2011. [14] D. E.Goldberg and J. H. Holland, "Genetic algorithms and machine learning,"1988. [15] M. A Mohammed, et al., "Solving vehicle routing problem by using improved genetic algorithm for optimal solution," Journal of Computational Science, 21, 255-262, 2017. [16] M. A. Mohammed, et al., "Using genetic algorithm in implementing capacitated vehicle routing problem," International conference on computer & information science (ICCIS). IEEE. Vol. 1, pp. 257-262. [17] A. S. Al-Khaleefa, et al., "Optimized authentication for wireless body area network. Journal of Telecommunication," Electronic and Computer Engineering (JTEC), 10(2), 137-142, 2018. [18] M. A. Mohammed, et al., "An anti-spam detection model for emails of multi-natural language," Journal of Southwest Jiaotong University, 54(3), 2019. [19] B. A Khalaf, et al., "Simulation Study of Syn Flood Attack in Cloud Computing Environment," REVISTA AUS 26-1, 2019. [20] S. A Mostafa, et al., "A Multi-Agent Ad Hoc On-Demand Distance Vector for Improving the Quality of Service in MANETs," International Symposium on Agent, Multi-Agent Systems and Robotics (ISAMSR), IEEE, pp. 1-7, 2018. [21] S. N. Mansor, et al., "An emotional agent approach for online customer satisfaction surveys analysis," Int J Eng Tech, 8, 227-233, 2019. [22] M. A. Jubair, et al., " A Survey of Multi-Agent Systems and Case-Based Reasoning Integration," International Symposium on Agent, Multi-Agent Systems and Robotics (ISAMSR), IEEE, pp. 1-6, 2018. [23] M. h. Hassan, et al., "A Statistical Risk Assessment Method of Dynamic Environments: A Case Study of Air Pollution," REVISTA AUS 26-1, 2020. [24] A. B. A. Mosmana et al., "An automated attendance management method for football stadium attendance systems" REVISTA AUS 26-1, 2019. [25] M. H. Hassan et al., "A survey of multi-agent system approach in risk assessment," International Symposium on Agent, Multi-Agent Systems and Robotics (ISAMSR), IEEE, pp. 1-6, 2018. View publication stats | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/331803034PHYSICS AND APPLICATIONS OF OPTICAL LORENTZ FORCEThesis · March 2019CITATIONS0READS1,3991 author:Some of the authors of this publication are also working on these related projects:Electromagnetism & Optics View projectModification of Higher Order Natural Mode in Metamaterial Loaded Patch Antenna View projectMahdy Rahman Chowdhury MahdyNorth South University89 PUBLICATIONS 1,297 CITATIONS SEE PROFILEAll content following this page was uploaded by Mahdy Rahman Chowdhury Mahdy on 16 March 2019.The user has requested enhancement of the downloaded file. PHYSICS AND APPLICATIONS OF OPTICAL LORENTZ FORCE Mahdy Rahman Chowdhury (B.S.C, BUET, EEE, Bangladesh) A THESIS SUBMITTED FOR THE DEGREE OF DOCTOR OF PHILOSOPHY DEPARTMENT OF ELECTRICAL AND COMPUTER ENGINEERING NATIONAL UNIVERSITY OF SINGAPORE April 2017 DECLARATION I hereby declare that the thesis is my original work and it has been written by me in its entirety. I have duly acknowledged all the sources of information which have been used in the thesis. This thesis has also not been submitted for any degree in any university previously. Sign: MAHDY RAHMAN CHOWDHURY December 2016 “He prayeth best, who loveth best All things both great and small; For the dear God who loveth us, He made and loveth all.” - Samuel Taylor Coleridge (The Ancient Mariner, Lyrical Ballads). This PhD thesis is dedicated to my beloved little sister. ‘Love is the master key to solve all the problems’ i Acknowledgements I would like to express my deepest appreciation to my supervisor Prof. Qiu Cheng Wei, who has the attitude and element of a genius. He continuously conveyed an enthusiasm in regard to my full theoretical work. I should also add that: he is not only a good supervisor but also a very kind and helpful person. Without his proper guidance and persistent help this dissertation would not have been possible. I would like to thank my co-supervisor, Prof. Chen Zhi Ning, our collaborators such as Prof. Manuel Nieto Vesperinas and Porf. Weiqiang Ding, who gave me great confidence by showing strong trust and faith in me. They have always encouraged me to work harder and smarter and provided me full freedom to explore my research work while focusing on the final goal. I am grateful to them for helping and guiding me to reach the ultimate final conclusions of this thesis, without which it would not have been possible to finish this thesis. My special thanks to Prof. Qiu Cheng Wei, Prof. Weiqiang Ding, Prof. Golam Dastegir Al Quaderi (Dhaka University, Physics Dept.) and Mr. Arif Shahriar who have spent their valuable time to read my PhD thesis. Without their kind suggestions and timely advice, it would be really tough for me to reach the final version of this thesis. I would like to express my gratitude to all the past and present members in the Microwave Lab of the National University of Singapore and the researchers under the supervision of my supervisor, for their appreciated help and friendship. Special thanks go to Dr. Qasim Mehmood, Mr. Tianhang Zhang, Dr. Mohammaed Danesh, Dr. Zhao Jiajun, Dr. Dongliang Gao, Mr. Amin Kianinejad, Mr. Mei Shengtao, Ms. Srien sithara and Mr. Ashraf Adam for their help throughout the thesis. It has been a real joy working with them, chatting with them and having meals together. I would like to thank all my friends and specially my housemates in Singapore. ii Lastly, I would like to thank my great parents who always gave me their support. I would like to take this opportunity to tell you all that I love you all so much and I will always be in debt to everything you have done for me in my life. I also want to thank my beloved wife, Nabila (Samantha) Alam. You spent much time accompanying me (even not staying with me in Singapore) and waiting for me to complete my research works. Specially, you almost alone handled my little baby for almost two years during my PhD program. I also thank my mother (also my father) and my mother in law (also my father in law) who took care and looked after my little baby. My special gratitude to my father who was a source of inspiration to me. At the beginning of my PhD program, my one and only little sister passed away. I decided not to carry on this PhD program. It was my father who inspired me to earn this PhD and then to go back my Country, Bangladesh. The higher education system of Bangladesh is just developing. My father advised me to finish my PhD with proper learning so that after coming back to Bangladesh, I shall contribute significantly in such a developing education system as a true academician. I also love my country very much and always dream to contribute in the area of theoretical research, especially in physics. These things have always inspired me to work hard during my whole thesis and also made me understood that: ‘Love is the master key to solve all the problems’. Lastly, I am thankful to the creator. If my son, Arham Chowdhury Isa, were not born by the middle of this PhD program, it would never be possible for me to work with mental peace and to finish my thesis so smoothly. iii Abstract In 1918 Albert Einstein wrote in a letter to Walter Dällenbach against his own theory of optical force: “It has long been known that the values I had derived with Laub at the time are wrong; Abraham, in particular, was the one who presented this in a thorough paper. The correct strain tensor has incidentally already been pointed out by Minkowski”. Unfortunately after almost 100 years, it is still not clear why Einstein wrote so regarding the theory of optical/electromagnetic force (the optical stress tensor (ST)) associated with Lorentz force. The electromagnetic force is one of the four fundamental interactions (other three: the strong interaction, the weak interaction and gravitation) in nature. In spite of experiencing much progress from 1970, research on optical/electromagnetic forces has so far revealed only the partial nature of light-matter interaction. For example, inside a material media an appropriate description of photon momenta (i.e. the Abraham-Minkowski dilemma) and an appropriate description of stress tensors along with Lorentz /volumetric forces (known as Minkowski, Abraham, Chu, Einstein-Laub and Ampere/Nelson forces) are still matters of great controversy. In this thesis, physics and applications of Lorentz force along with stress tensors and photon momenta have been investigated in details with special interest in tractor beam effect, plasmonic objects, chiral objects and the objects embedded in a generic material medium. At first, we have considered simpler case: application of Lorentz force for the objects placed in air. Specifically we demonstrate how the mechanism of Lorentz force can be used in a tractor beam-like effect for pulling multiple Rayleigh particles placed outside a dielectric hollow core waveguide and coupler. We also represent the Lorentz force analysis for plasmonic off-axis and on-axis spherical heterodimers and it is shown that the reversal of longitudinal binding force can be easily controlled by forced symmetry breaking. Though it is commonly believed that plasmonic forces iv mostly arise from the surface force and Fano resonance can be a promising way to achieve binding force reversal, our study based on Lorentz force dynamics suggests notably opposite proposals for the case of plasmonic spherical heterodimers. Next, we have also considered objects those are embedded in a material medium. The fundamental results/proposals of this thesis are presented afterwards. We have shown that the well-known Lorentz force leads to inconsistent result (pushing force) instead of the experimentally observed optical pulling force as observed in interfacial tractor beam experiments. As one of the major contributions of this thesis, we identify the ‘exact’ reasons why the well-known distinct Lorentz/volumetric forces fail not only for interfacial tractor beam experiment but also for several other major radiation pressure experiments performed up to date. Later, we demonstrate that it is still possible to establish different equivalent time-averaged Lorentz / volumetric force formulas based on the fulfilment of just two ‘consistency conditions’. Based on those proposed ‘consistency conditions’, finally, we demonstrate that by modifying the Einstein-Laub or Chu formulation; time-averaged STs and volume forces are obtainable, which can overcome the aforementioned inconsistencies. For example- these modified formulations can yield the ‘correct’ time averaged force similar to Minkowski’s force for almost all the previous real experiments. Specially, our proposed modified Einstein-Laub ST can be considered as an efficient mathematical toolkit, an alternative of time and memory consuming volumetric forces, to yield the internal bulk force of a chiral or achiral object embedded in complex material backgrounds (i.e. homogeneous, heterogeneous, bounded etc.). Later, we have also shown an interesting application of modified Lorentz force to control the reversal of optical binding force of plasmonic cubes placed over plasmonic substrate due to strong Fano resonance. Finally, based on our proposals throughout the thesis, at the end of this thesis we propose a new hypothesis named as the ‘existence domain’. ‘Existence domain’ is the region either outside of a scattering body taking only its exterior fields into account, v or in its interior considering only the inside fields. Though almost all the time averaged distinct STs and volumetric force laws are restricted to the idea of ‘existence domain’, we demonstrate that the stress tensor, volumetric force and photon momentum of Minkowski are free from such and other restrictions both in instantaneous and time averaged situations. After almost 100 years of Einstein’s prediction, this thesis, most probably, finally answers why only Minkowski’s theory of optical force and photon momentum remains consistent in all circumstances. Proposals presented in this thesis can be very effective for resolving not only the dilemma of distinct stress tensors and optical Lorentz/ volumetric forces but also for settling the controversy of Abraham-Minkowski photon momenta. This thesis may also open a new window for optical pulling force/tractor beams along with the novel manipulation process of plasmonic dimers. vi Table of Contents Acknowledgements ............................................................................................. i Abstract ............................................................................................................. iii Table of Contents .............................................................................................. iii List of Figures .................................................................................................. xii List of Tables .................................................................................................. xvi List of Acronyms ........................................................................................... xvii List of Publications, Patents and Conferences ................................................ xix Chapter 1 Introduction and Literature Review .............................................................. 1 1.1. Introduction ................................................................................................. 1 1.2. The Simplest Case of Optical Force Calculation: Object placed in air ...... 3 1.3. Optical Force Calculation by Volumetric Force ..................................... 5 1.4. Different Volumetric Force Laws: Problem of Force Distribution ............ 8 1.5. Problem with Optical Force Calculation: Object Embedded in Material Background ........................................................................................................ 9 1.6. Abraham-Minkowski Controversy: Photon Momentum inside Matter .... 13 1.7. Optical Tractor Beams .............................................................................. 14 1.8. Reversal of Near Field Optical Binding Force ................................... …. 17 1.9. Thesis Outline ........................................................................................... 19 vii Chapter 2 Lorentz Force and Its Application on Tractor Beam like Effect .............. 24 2.0. Summary of chapter 2 ............................................................................... 24 2.1. Introduction ............................................................................................... 25 2.2. Single Rayleigh particle outside the waveguide: Force calculation ......... 27 2.3. Detailed analysis of set-up ........................................................................ 30 2.3.1. Electromagnetic fields and modes in a cylindrical waveguide ..... 30 2.3.2. Behavior of external wave-field of a single scatterer placed outside proposed waveguides .............................................................................. 30 2.4. Single or multiple Rayleigh particles outside the waveguides: Lorentz force Dynamics ................................................................................................ 33 2.5. Robustness of pulling force to absorption, wavelength and waveguide length................................................................................................................ 37 2.6. Multiple Rayleigh particles outside the coupler …. ................................. 39 2.7. Conclusion ................................................................................................ 40 Chapter 3 Lorentz Force on Plasmonic Spherical Heterodimers: Reversal of Binding Force ................................................................................................................ 42 3.0. Summary of chapter 3 ............................................................................... 42 3.1. Introduction ............................................................................................... 43 viii 3.2. Optical Force calculation .......................................................................... 48 3.3. Lateral binding force: On-Axis Spherical Heterodimers ......................... 50 3.3.1. Parallel Polarization: No reversal of lateral binding force for Au-Ag, Au-Au and Ag-Ag on-axis heterodimers …… ................................ 50 3.3.2. 3.3.2Perpendicular Polarization: Reversal of lateral binding force for Au-Ag, Au-Au and Ag-Ag on-axis heterodimers ….. ...................... 51 3.4. Longitudinal binding force for Off-Axis Plasmonic Heterodimers: end-fire and near end-fire configuration ........................................................................ 54 3.4.1. Au-Ag off-axis heterodimers: Longitudinal binding force for both polarizations .......................................................................................... 55 3.4.1.1. Au-Ag off-axis heterodimers: Rotating the smaller particle keeping the big one fixed ...................................................................... 56 3.4.1.2 (ii)Au-Ag off-axis heterodimers: Rotating the bigger particle keeping the small one fixed .................................................... 58 3.5. Simplest procedure to reverse the longitudinal binding force for all the off-axis heterodimers ....................................................................................... 62 3.6. Conclusions ............................................................................................... 64 Chapter 4 Problem with Lorentz Force due to Material Background: Interfacial Tractor Beam ................................................................................................. 65 4.0. Summary of Chapter 4 .............................................................................. 65 4.1. Introduction ............................................................................................... 66 ix 4.2. Ray Tracing Method and Minkowski Stress Tensor by Employing Background Fields ........................................................................................... 69 4.3. Explanation of the Observations and the Two Photon Momenta ............. 72 4.4. The Validity of Other Methods ................................................................. 74 4.5. Modified Einstein-Laub Stress Tensor by Employing Interior Fields of the Scatterer ........................................................................................................... 74 4.6. Conclusions ............................................................................................... 75 Chapter 5 Modification of Lorentz Force for embedded chiral and achiral objects ......................................................................................................................... 77 5.0. Summary of Chapter 5 .............................................................................. 77 5.1. Introduction ............................................................................................... 78 5.2. Proposal of Consistency Conditions ......................................................... 80 5.3. The consistent external force with NO GAP METHOD …. .................... 83 5.4. Inconsistency of the GAP METHOD and different other formulations ... 86 5.5. Consistency of the external Minkowski and internal MEL or modified Chu formulations ............................................................................................. 90 5.6. A short discussion on previous tractor beam and lateral force experiments.......................................................................................................................... 93 5.5. Chiral Modified Einstein-Laub and Chu formulations ....................... ….95 5.6. Conclusion .............................................................................................. 101 x Chapter 6 Modified Lorentz Force and Plasmonic Cube Dimers over Substrate: Binding Force Reversal ............................................................................... 104 6.0. Summary of Chapter 6 ............................................................................ 104 6.1. Introduction ............................................................................................. 105 6.2. Optical Force calculation ........................................................................ 107 6.3. Plasmonic cubes over plasmonic substrate: a short discussion on resonance modes and reversal of binding force ...................................... ….. 109 6.4. Plasmonic cubes over different substrates: effect on binding force........................................................................................................................ 113 6.5. Effect of height, size and background material on binding force ........... 118 6.6. Conclusions ............................................................................................. 121 Chapter 7 Consistency of Minkowski’s Theory: Some Final Proposals ................... 123 7.1. Introduction and Summary of Chapter 7 ................................................ 123 7.2. Time averaged Force: Proposal of ‘existence domain’ hypothesis ........ 124 7.3. Instantaneous and Time average Optical Force: Accuracy of Minkowski’s theory ............................................................................................................. 131 7.4. Photon Momentum for instantaneous and Time averaged cases: Accuracy of Minkowski’s theory ................................................................................... 137 xi 7.5. Conclusion ......................................................................................... 139 Chapter 8 Conclusion and Future Works .................................................................... 141 8.1. Summary of the Thesis ........................................................................... 141 8.2. Final Remarks and Future Works ...................................................... 145 Appendices .................................................................................................... 148 Appendix A Detail analysis on GAP METHOD of optical force calculation ...................................................................................................... 148 Appendix B Possible derivation of the internal Modified Einstein-Laub stress tensor .................................................................................................... 154 Appendix C Derivation of the MEL surface force in two fully different ways ............................................................................................................... 159 Appendix D Time averaged external (total force) and internal force (bulk force) for an embedded achiral object ........................................................... 161 Bibliography ................................................................................................. 162 xii List of Figures Figure 1.1: Optical momentum transfers from an incident plane wave of propagating vector ki to a spherical particle .. ................................................... 4 Figure 1.2: Volumetric force distribution for different force laws according to [28] for the Ashkin–Dziedzic experiment.......................................................... 9 Figure 1.3: Though the calculation of time averaged optical force by different volumetric/Lorentz forces lead to same result for the case of air background ............................................................................................................................ 10 Figure 1.4: Simulated radiation pressure on a polystyrene bead (green) placed at the edge of the focal point of a continuous-wave beam (λ0= 532 nm) in a background medium of water (n = 1.33).. ....................................................... 10 Figure 1.5: Different examples of tractor beams.. ........................................... 15 Figure 1.6: Continuous and three dimensional stable manipulations of fully immersed particles by using the optical gradient force generated by superimposing coaxial pair of Bessel beams ................................................... 17 Figure 1.7: An example of binding force reversal due to Fano resonance. .... 17 Figure 1.8: Another example of binding force reversal due to Fano resonance........................................................................................................................... 19 Figure 2.1: Illustration of optical pulling force for a single Silicon particle (refractive index 3.5 and radius of 30 nm) placed at different z-positions outside a 3D cylindrical waveguide. ................................................................ 28 Figure 2.2: Total Poynting vector distribution ................................................. 32 Figure 2.3: Possible Illustration of optical pulling force for multiple Silicon particles (all with refractive index 3.5 and radius 30 nm) placed outside the 3D waveguide with a length of 15 µm; light wavelength 1500 nm. ...................... 33 xiii Figure 2.4: Robustness test .............................................................................. 37 Figure 2.5: Possible Illustration of optical pulling force for multiple Silicon particles placed outside two identical 10 µm sized (length) 3D cylindrical waveguides (a directional coupler). ................................................................. 39 Figure 3.1: Two particles of radii 100 and 50 nm are placed with inter particle distance from surface to surface ‘d’; d= 20 nm throughout this work.. ........... 46 Figure 3.2: Considering perpendicular polarized light for the configuration 52 Figure 3.3: (a) and ‘ȹ’ = 0 degree [on axis Ag-Au]: (a) Extinction co-efficient. (b) The binding force FBind (x) = (FB (x)-FS (x))................................................... 54 Figure 3.4: For both polarizations, electric dipole moment of Ag-Au heterodimers [for configuration of Fig. 3.1 (a)] and Au-Au heterodimers [for configuration of Fig. 3.1 (b)].. ......................................................................... 57 Figure 3.5: SR and BR represent ‘small rotate’ and ‘big rotate’ respectively and ‘ȹ’ = 60, 70 and 90 degree [off axis Ag-Au].. ......................................... 60 Figure 3.6: SR and BR represent ‘small rotate’ and ‘big rotate’ respectively. For off axis Ag-Au and BR [the configuration of Fig.3.1 (d) and ‘ȹ’ = 60 degree]. ............................................................................................................. 61 Figure 3.7: For off axis Ag-Au and by rotating the big particle [the configuration of Fig.3.1 (d) and ‘ȹ’ = 60 degree] ...................................... … 63 Figure 3.8: (a) and (c): extinction coefficient of off-axis [only ȹ =90 degree case shown here] Au-Au hetero-dimers [cf. the configuration of Fig. 3.1(e)]…........................................................................................................................... 63 Figure 4.1: (a) Schematic of photon momentum transfer in an inhomogeneous mixer background. ........................................................................................... 68 xiv Figure 4.2: (a) Changes of momentum per photon in materials with a refractive index n for Minkowski (blue), Abraham (red) and Peierls (black) formulations, respectively. ............................................................................... 71 Figure 4.3: Optical forces derived from different formulas.. ........................... 72 Figure 5.1: Procedure of time averaged optical force calculation by employing stress tensors.. .................................................................................................. 80 Figure 5.2: (a) Optical sorting of a dielectric particle using two obliquely incident plane waves ........................................................................................ 83 Figure 5.3: (a) 3D Illumination geometry for a particle at the interface between two media to model the lateral force set-up. .................................................... 86 Figure 5.4: Illustration of optical force with GAP METHOD. ........................ 87 Figure 5.5: Calculation of time averaged total optical force (NO GAP METHOD) by external Minkowski ST and the time averaged bulk force by internal MEL ST. ............................................................................................. 94 Figure 5.6: Time-averaged forces: Fout at 1.01raa from Minkowski ST and Fin (bulk force) at 0.999raa from the Chiral MEL ST. ................... 96 Figure 5.7: Time-averaged forces: Fout at 1.01raa from Minkowski ST and Fin (bulk force) at 0.999raa from the Chiral MEL ST. ................... 99 Figure 5.8: Time-averaged forces: Fout at 1.01raa from Minkowski ST and Fin (bulk force) at 0.999raa from the Chiral MEL ST. ................. 100 Figure 6.1: (a) Two silver cubes (L=120 nm) are placed over silver substrate and their inter-particle gap, d, is 100 nm (spacer height 5 nm from the bare substrate and the cubes are placed 5 nm away from the spacer). .................. 110 Figure 6.2: (a) Two silver cubes (L=180 nm) are placed over silver substrate and their inter-particle gap,d, is 150 nm. ....................................................... 113 xv Figure 6.3: Two silver cubes (L=120 nm) are placed over different substrates and their inter-particle gap, d, is 150 nm. ...................................................... 114 Figure 6.4: For cube homodimers over silver substrate, plot of Bulk( )Del Fx and Surf (x)Del F. ....................................................................................................... 115 Figure 6.5: For the cases of different substrates, Lorentz force components and the steady state current (Jy) of two 120 nm lengthened silver cubes along with the substrate. .................................................................................................. 116 Figure 6.6: Two silver cubes (L=120 nm) are placed at different heights from the silver substrates (three different columns represent the position of the cubes: 50, 20 and 15 nm away from the substrate) and their inter-particle gap, d, is fixed 100 nm........................................................................................... 119 Figure 7.1: Optical momentum transfers from an incident plane wave of propagating vector ki to a spherical particle .. ............................................... 124 Figure 7.2: The consistency of Minkowski’s and Abraham’s theory everywhere in space for time averaged cases. ............................................... 129 Figure 7.3: Consistency of Minkowski’s and Chu’s theory for instantaneous and Time averaged Force... ............................................................................ 132 Figure 7.4: Consistency of Minkowski photon momentum........................... 138 Figure 1A: When the object is surrounded by a medium other than the free space, a narrow gap may be imagined to exist between the object and its surroundings.. ................................................................................................. 150 Figure 2A: Time-averaged forces: Fout at 1.01raa from Minkowski ST and Fin (bulk force) at 0.999raa from the MEL ST. These forces are always of same trend... ................................................................................... 161 xvi List of Tables Table-1.1: Previous macroscopic tress tensors, electromagnetic momentum densities and force density laws ......................................................................... 7 Table-3.1: An overview on the behavior of binding force for on and off-axis heterodimers ..................................................................................................... 47 Table-7.1: Time averaged macroscopic stress tensors and volume forces: proposal of ‘existence domain’ (not applicable for Minkowski and Abraham formulations) .................................................................................................. 129 Table-7.2: Optical force laws: difference between instantaneous forces and their corresponding time averaged forces ...................................................... 134 xvii List of Acronyms (in Alphabetic Order) 3D three dimensional APM Abraham photon momentum BR big rotate CP circularly polarized DD dipole dipole DQ dipole quadrupole DWTB dielectric waveguide tractor beam EL Einstein-Laub FDTD Finite difference time domain ITB interfacial tractor beam LP linearly polarized LSP localized surface plasmon MEL modified Einstein-Laub MPM Minkowski photon momentum NOF negative optical force OM optical manipulation xviii QD quadrupole dipole QQ quadrupole quadrupole ST stress tensor SR small rotate SPP surface plasmon polariton xix List of Publications, Patents and Conferences Publications in Peer-reviewed International Journals †=Equal Contribution, *=Corresponding author. 1. Qiu, C.-W. †*, Ding, W. †, Mahdy, M. R. C., Gao, D., Zhang, T., Cheong, F. C., Dogariu, A., Wang, Z. & Lim, C. T. Photon Momentum Transfer in Inhomogeneous Dielectric Mixtures and Induced Tractor Beam. Light: Science and Applications 4, e278 (2015). [Nature Publishing Group] 2. Rahman, M. M., Sayem, A. A., Mahdy, M. R. C. *, Haque, M. E., Islam, R., Chowdhury, S. T. R. & Matin, M. A., Tractor beam for fully immersed multiple objects: Long distance pulling, trapping, and rotation with a single optical set-up, Annalen der Physik 527, 777 (2015). [Cover page and cover story of December 2015 issue of Annalen Der Physik] 3. Zhu, T. †, Mahdy, M. R. C. †, Cao, Y., Lv, H., Sun, F., Jiang, Z. & Ding, W. *, Optical pulling using evanescent mode in sub-wavelength channels. Opt. Express. 24, 18436 (2016). 4. M.R.C. Mahdy†*, M. Q. Mehmood† , Weiqiang Ding, Tianhang Zhang, Zhi Ning Chen, Lorentz Force and the Optical Pulling of Multiple Rayleigh Particles outside the Dielectric Cylindrical Waveguides , Annalen der Physik 529, 1600213 (2017). 5. M.R.C. Mahdy†*, Tianhang Zhang†, Weiqiang Ding*, Amin Kianinejad, Manuel Nieto-Vesperinas, Consistency of time averaged optical force laws for embedded chiral and achiral objects. arXiv preprint arXiv:1704.00334 (2017). [Submitted version in journal: M.R.C. Mahdy†*, Tianhang Zhang†, Weiqiang Ding*, Amin Kianinejad, Manuel Nieto-Vesperinas, A solution to the problem: time averaged total optical force.] 6. Dongliang Gao, Weiqiang Ding, Manuel Nieto-Vesperinas, Xumin Ding, Mahdy Rahman, Tianhang Zhang, Chwee Teck Lim and Cheng-Wei Qiu* , Optical Manipulation from Microscale to Nanoscale: Fundamentals, Advances, and Prospects. Light: Science and Applications 6, e17039 (2017). [Nature Publishing Group] xx 7. Tianhang Zhang, M.R.C. Mahdy, Yongmin Liu, Jinghua Teng, Lim Chwee Teck, Zheng Wang and Cheng-Wei Qiu*, All-Optical Chirality-sensitive Sorting via Reversible Lateral Forces in Interference Fields. ACS Nano (2017). 8. M.R.C. Mahdy, Md. Danesh, Tianhang Zhang, Weiqiang Ding, Cheng-Wei Qiu*, Lorentz Force and Plasmonic on-axis and off-axis Spherical Heterodimers: Reversal of Optical Binding Force as the Effect of Symmetry Breaking. (will be submitted in journal soon) 9. M.R.C. Mahdy, Tianhang Zhang, Md. Danesh, Weiqiang Ding, Cheng-Wei Qiu*, Plasmonic Cube Dimers over Substrate: A Flexible Approach of Binding Force Reversal Induced by Fano Resonance. (will be submitted in journal soon) 10. M.R.C. Mahdy, GD Al-Quaderi, Tianhang Zhang, Weiqiang Ding, Manuel Nieto-Vesperinas, Cheng-Wei Qiu* , An investigation on the consistency of optical forces and photon momenta for instantaneous and time averaged scenario (under preparation) [a small part of the article already discussed in: Mahdy, M. R. C., Dongliang Gao, Weiqiang Ding, M. Q. Mehmood, Manuel Nieto-Vesperinas, and Cheng-Wei Qiu, arXiv:1509.06971 (2015)] Chapter 1 Introduction and Literature Review 1 Chapter 1 Introduction and Literature Review 1.1 Introduction The electromagnetic force is one of the four fundamental interactions (commonly called forces) in nature. The other three fundamental interactions are the weak interaction, the strong interaction, and gravitation. By utilizing the optical/electromagnetic force, optical manipulation (OM) [such as optical cooling [1], trapping [2], binding [3-5], sorting and transporting [6-8]] have experienced an intensive development in the past 40 years [9-11]. OM is now considered as one of the most important tools in many scientific areas, including optics [9], atomic physics [12,13], biological science [14] , and chemistry [15]. Traditionally, the configuration of a small dielectric sphere in the focus is usually investigated by considering the scattering of the fundamental mode of a Gaussian beam [1,2]. According to the theory of electrostatics state, small Rayleigh particles develop an electric dipole moment in response to the electric field of the light and the induced dipole is then drawn by the field intensity gradients; this phenomenon results in a force called Optical Gradient Force [16]. Again, due to the transfer of momentum from the photons to an object, another optical force namely Optical Scattering Force, comes into action. Optical Scattering Force is the combination of radiation pressure and spin curl component [16]. Recently, OM has gone beyond this scenario, and more complex configurations have been investigated. For example, OM resolution has been increased beyond the Rayleigh limit by extending it to the near-field, thus taking advantage of non-radiative forces [17,18]. The manipulated objects can be generic Mie or bigger sized objects [rather than a Chapter 1 Introduction and Literature Review 2 single and small dielectric medium] such as magnetodielectric [16], chiral [19-21], plasmonic [22,23] or multilayered coating [24]. For such cases, the common way to yield the total optical force is to apply the well-known Maxwell stress tensor (ST) method [25] instead of the approximate force calculation method of dipolar force [16] (i.e. gradient, scattering and spin curl force). In contrast, the Lorentz/volumetric force is applied to compute the force density throughout the particles, which demonstrates regions of compression and tension within the medium [26]. Most importantly, Lorentz force also describes how the volumetric force on a generic object is distributed [27], which is described as Lorentz volumetric force distribution. This information cannot be retrieved from the optical stress tensor based [25] force calculation. The process of yielding the total force become more complex when the embedding background is considered material media [28] instead of air or vacuum. The aim of this chapter is to present an introductory picture of OM (i.e. literature review); from fundamental photon momentum transfer, dilemma, existing problems and advances toward future prospects on nanoscale and microscale objects and structures, which have recently attracted considerable attention [17, 28, 29]. At first, in the next section, we have reviewed very shortly how to calculate the optical force on an object placed in air. After that, the Lorentz/volumetric force and its connection with stress tensor have been reviewed for simple case: object placed in air. However, when an object is placed in material media instead of the air background, optical force calculations require more appropriate and accurate description of photon momenta and forces. This ‘appropriate and accurate description’ is still quite controversial [28]. As a result, dilemma on distinct photon momenta and electromagnetic stress tensors [27], which are connected with different Lorentz/volumetric force densities inside matter, are reviewed in the later part in this chapter. Almost half part of this thesis has tried to cover and solve such controversial but fundamental issues on optical stress tensors, Lorentz /volumetric forces and the Chapter 1 Introduction and Literature Review 3 photon momenta inside material medium. The other half part of this thesis mainly focuses on the novel applications of the Lorentz and proposed modified Lorentz force on plasmonic dimers and counter-intuitive tractor beam effect. Though previously OM was restricted mainly on Gaussian beams or simple backgrounds, recently the incident light is also constituted by special wave fields such as non-diffraction Bessel beams, multiple beam interference, or complex backgrounds. A representative example is the recently proposed optical tractor beams [30-36], which can exert a negative non-conservative force on a body pulling it opposite to its propagation direction [37]. In addition, when the objects are resonant with the incident light (i.e. plasmonic object [38]), the optical force may be greatly enhanced. Recently force on plasmonic objects has achieved considerable attention due to some unusual behavior of plasmonic objects connected with Fano resonance [39], super scattering [40], localized surface plasmon [41] and surface plasmon propagation [42]. As a result, just before the end of this chapter, two very special applications of optical force such as counter-intuitive optical pulling force with tractor beams and the reversal of near field optical binding force due to plasmonic objects have also been discussed. In fact, the applications of Lorentz force dynamics is still not investigated in details for such exciting cases of optical manipulations: tractor beams and plasmonic dimers. Almost half part of this thesis has tried to cover such issues based on Lorentz and proposed modified Lorentz force. Finally, at the end of this chapter, a detail outline of the thesis has been presented in the last section. 1.2 The Simplest Case of Optical Force Calculation: Object placed in air Chapter 1 Introduction and Literature Review 4 Fig. 1.1 Optical momentum transfers from an incident plane wave of propagating vector ki to a spherical particle with permittivity p and permeability p . The total force calculated using the Stress Tensor (ST) is: Fout when the time-averaged ST is evaluated from fields outside the particle, through a surface of radius (blue lines), and Fin when the time-averaged ST is evaluated from the inside of the particle through a surface of radius (red lines). We specify that throughout this thesis we refer to 'exterior' or 'outside' forces as those evaluated outside the volume of the macroscopic particles, while ‘interior’ or 'inside' refer to those quantities inside this object volume. To determine the optical force experienced by a particle, we refer to Fig. 1.1. If we set a boundary very close to the particle from outside, i.e., radius r= a+, (a is the particle radius, assuming it spherical (in three dimensions) or cylindrical (in two dimensions)), and calculate the time-averaged force using only exterior fields, this will be considered as the outside force. On the other hand, if we set a boundary very close to the particle from inside, i.e. at r= a-, and calculate the time averaged force using only interior fields, this will be the inside force. Usually the ‘outside optical force’ [35, 43] is calculated by the integration of time averaged Minkowski [35, 43] stress tensor at r= 1.01a = a+ (It means we are selecting a spherical surface with a radius slightly larger than the object, on which the 1.001raa0.999raaChapter 1 Introduction and Literature Review 5 integration of stress tensor is calculated for the total optical force) employing the background fields of the scatterer of radius a: . (1.1) Where ‘out’ represents the exterior total field (i.e. incident plus the scattered field) of the scatterer; E, D, H and B are the electric field, displacement vector, magnetic field, and induction vectors, respectively, while represents the time average and the unity tensor. On the other hand, if Eq (1.1) is applied at r= 0.999a = a- [43, 44] by employing the internal field of the scatterer placed in air, it leads to zero force [44] for lossless scatterers. As a result, the usual way is to determine the total force is the ‘outside force’ based on Eq (1.1). But a question arises: what is the connection of Eq (1.1) with optical volumetric/Lorentz force? Answer of this last question is discussed in details in the next section. 1.3 Optical Force Calculation by Volumetric Force Answer of the last question, connection of Lorentz force with stress tensor, comes from the linear momentum conservation equation [17]. According to linear momentum conservation equation, the total transferred momentum to an object should be written as [17]: (1.2a) is the electromagnetic momentum density and is the volumetric force density. Eq (1.2a) is considered as one of the main pillars of classical TOutoutTotalout****outoutoutoutoutoutoutout1 Re12(][)2dDEBHEDHBFTsTII·t fTGGfChapter 1 Introduction and Literature Review 6 electrodynamics along with Maxwell’s equations. Eq (1.2a) can also be described in an alternative approach [17]: TotalMech.Non-Mech.ppp Here p represents momentum. Ultimately in the time average form, Eq (1.2a) turns into the force equation: (1.2b) It should be noted that term vanishes due to time average. We have already discussed regarding the left side of Eq (1.2b), the surface integral of optical stress tensor, in the previous section. The right side of Eq (1.2b) is volume integral. Now a problem arises. Though the background is air, the Lorentz /volumetric force, , is connected with the idea of optical force distribution inside the object. As a result, though the left side of Eq (1.2a) [i.e. total momentum] is a single conserved quantity, in the right side of Eq (1.2) the form of or is not unique. In Eq (1.2a), the ∂G/∂t term is connected with photon momentum, ; where and is the electromagnetic momentum density. Inside matter, two different definition of photon momenta are considered, which is known as the Abraham-Minkowski controversy [27]. This controversy will be discussed shortly in a forthcoming sub-section. However, the usual common/generic ‘compact’ form of the time averaged Lorentz volumetric force is written as [45]: (1.3) is the total volumetric charge, contains the densities of free as well as bound currents arising from polarization and magnetization densities and is the magnetic induction [45]. Surprisingly, there are five different ‘elaborate’ forms of volumetric force [27,28] and their corresponding stress tensors as shown in Table-1.1. ddv TsFft GfGfNon-Mech.pNon-Mech.dV pGGtotaltotalf =E + JBtotaltotalJBChapter 1 Introduction and Literature Review 7 It is still a matter of debate which volumetric Lorentz force is more appropriate than the other one [27,28,46]. Interestingly, when the background of an embedded object is air or vacuum, all the Lorentz volumetric forces shown in Table-1.1 leads to exactly same time averaged total force according to Eq (1.2b). So, when the background of an object is air, for the calculation of time averaged total force, there is no problem. But the problem arises from the idea of volumetric force distribution process. This matter is discussed in the next section. Table-1.1: Previous macroscopic tress tensors, electromagnetic momentum densities and force density laws Stress tensor, , and force density, f Electromagnetic momentum density, G Minkowski 22121122 TDEBHB HD E IfEH DB Abraham 22221211;221Antc AffTDEEDBHHBB H + D E IfEHEH 2cEH Einstein-Laub 001 ()2(()·)·oott T = DEBHH HE E IPMfPEMHHE 2cEH Chu 00000··1 ()2()()ott EEHHH HE E IEHHETPfP 2cEH Ampere/ Nelson 1122000012()()t TEEBBB +EIPfP EBM+μB 0EB Polarization, P ; magnetization, M. TChapter 1 Introduction and Literature Review 8 1.4 Different Volumetric Force Laws: Problem of Force Distribution One remarkable example on the inconsistency of optical volumetric force distribution is the famous Ashkin–Dziedzic experiment in ref. [47]. A two-dimensional fluid dynamic simulations of the Ashkin–Dziedzic experiment has been conducted in ref. [28] as shown in Fig. 1.2. In Fig. 1.2, the lower fluid (blue) represents water (n = 1.33) and the upper fluid (white) represents air (n = 1). In [28], the air–water interface is excited by a 60 ns-long and 530-nm-wavelength pulse at normal incidence. The time-averaged force density distributions exerted by the pulse are calculated using the Minkowski, Abraham, Einstein–Laub, Chu, and Ampere formulations, which are shown next to the resulting velocity field (blue arrows) of the deformed interface [28] in Fig. 1.2. The Abraham, Minkowski, and Einstein–Laub formulations predict an upward bulge for TE and TM polarizations for both illumination directions [consistent with experimental observations [47]]. The Ampere and Chu formulations predict an upward and downward bulge for TM and TE polarizations, respectively [inconsistent with experimental observations [47]]. Chapter 1 Introduction and Literature Review 9 Fig. 1.2: Volumetric force distribution for different force laws according to [28] for the Ashkin–Dziedzic experiment [47]. The Abraham, Minkowski, and Einstein–Laub formulations predict an upward bulge for TE and TM polarizations for both illumination directions [consistent with experimental observations [47]]. The Ampere and Chu formulations predict an upward and downward bulge for TM and TE polarizations, respectively [inconsistent with experimental observations [47]].Figures adapted and reproduced with permission from ref. [28], © 2015, Institute of Physics (IOP) publishing. 1.5 Problem with Optical Force Calculation: Object Embedded in Material Background Chapter 1 Introduction and Literature Review 10 Fig. 1.3: Though the calculation of time averaged optical force by different volumetric/Lorentz forces lead to same result for the case of air background, this fact does not remain valid for the case of material background. (a) The prototype of the interfacial tractor beam experiment reported in [35]. So far two different independent works ([28] and [43]) have investigated that same experiment reported in [35]. Different volumetric force laws lead to distinct results; especially Chu and Ampere forces lead to inconsistent pushing force. (b) Another famous example of such disagreements [53]: the 1954 Jones–Richards experiment [55] and the 1978 Jones–Leslie experiment [56]. It is observed that only Minkowski and Abraham formulations always predict the time averaged results which are consistent to the real experimental observations in [55] and [56] according to [28]. Figures adapted and reproduced with permission from (a) ref. [35], © 2013, Nature publishing group and (b) ref. [53], © 2011, Optical Society of America (OSA). Fig. 1.4: Simulated radiation pressure on a polystyrene bead (green) [28] placed at the edge of the focal point of a continuous-wave beam (λ0= 532 nm) in a background medium of water (n = 1.33). The polystyrene bead (n = 1.58) has a diameter of 820 nm and the force acting on it has been determined in [28] under the rigid body assumption by integrating the force density distribution within the bead calculated using the Minkowski, Abraham, Einstein–Laub (EL), Chu, and Ampere/Nelson formulations. These formulations predict distinct magnitudes of total force in [28] and lead to inconsistent result (direction of total force) with Chu and Ampere formulations for TM polarization. Figures adapted and reproduced with permission from ref. [28], © 2015, Institute of Physics (IOP) publishing. Chapter 1 Introduction and Literature Review 11 The calculation process of total optical force on a body embedded in a medium has different versions [27,28] [cf. Table- 1.1]. But this force is usually derived by integrating the momentum flux over any closed surface surrounding this object [cf. Fig. 1.1]. For example- initially, Minkowski [48] suggested that a field energy-momentum tensor should preserve its form in all Lorentz frames. According to his suggestion, the stress tensor should be: M12TDEBHB HD E I (1.4) The volumetric force density of Minkowski’s stress tensor is known as Helmholtz force density: 22H1122 fEH. (1.5) Major radiation pressure experiments have always supported Minkowski’s theory [27,28, 35, 46, 49,50]. Alternatively, Abraham suggested that the symmetric electro-magnetic stress tensor and force density in matter as [51,52]: (1.6) 22A1122AffEH (1.7) where is the extra “Abraham force” density [46,50]. Experiments in favour of Abraham formulation will be found in detailed in Refs. [28,46,49,50]. However, it should be remarked that in most configurations, dealing with time-averaged forces of an object embedded in any material medium, both theories yield the same time-averaged optical force [28, 46, 49, 50]. In fact, the difference between Minkowski’s 1 [()],2ADEEDBH + HBB HD E ITAfChapter 1 Introduction and Literature Review 12 force and Abraham’s force arises only in the instantaneous force [46, 50] mainly due to the extra “Abraham force”. It is remarkable that, like the Minkowski’s formulation, Abraham’s theory does not separate the momentum into the sum of field and material components. It describes D and B as a single quantity so that the polarization P and magnetization M are not separately formulated [cf. Table-1.1]. There are also three major stress tensors and volume forces other than that of Abraham or Minkowski formulations [27, 28, 46]. Those stress tensors and volume forces are: Einstein-Laub, Chu and Nelson/ Ampere stress tensor as shown in Table- 1.1 along with their corresponding Lorentz volumetric force density. But when the background is considered as a material medium rather than air or vacuum, those different volumetric force formulations given in Table-1.1 predict distinct magnitudes and directions of the time-averaged force on an embedded scatterer in several experiments [28,46] [cf. Fig. 1.3 and 1.4]. The notable fact is that the time averaged total force based on Eq (1.4) or (1.6) is usually considered as ‘outside total force’ [35,43] [which is calculated at r=1.01a = a+ employing the exterior background fields of the embedded scatterer]. On the other hand, whether Einstein-Laub, Chu and Nelson/ Ampere formulations should be applied as ‘outside force’ or ‘inside force’ (which is calculated at r= 0.999a=a- employing the interior fields of the embedded scatterer) is still not properly understood. The time averaged total forces significantly differ for different volumetric formulations according to [28] and [46] for real experiments conducted so far [2,35,43,53-56] considering background as a material medium and few other cases [cf. Fig. 1.3 and 1.4]. Hence it is still a quite ambiguous task to determine which volumetric force and its associated ST are more appropriate when the background is material media instead of air or vacuum. The controversies in the area of optical force calculation would be resolved if the corresponding ranges of validity of the optical force formulations are clearly identified. This issue has been investigated in some major parts of this thesis. Chapter 1 Introduction and Literature Review 13 However, it should be noted that: As Minkowski’s (or Abraham’s) time averaged formulations have always led to the correct time averaged total force for almost all major radiation pressure experiments, the formulations of Minkowski (or Abraham) is yet considered the most reliable approach [50] for the time-averaged force on objects of any size or shape. 1.6 Abraham-Minkowski Controversy: Photon Momentum inside Matter Generally speaking, the force is a consequence of the change of momentum carried by photons [1,2,35,43,47,53]. Interestingly, there is a long-standing debate over the determination of the momentum of photons in media, which may now be synthesized as the Minkowski-Abraham controversy [27, 57]. A century has now passed pertaining to the correct form of optical momentum in media. Both Experiments and theories have been applied at classical and quantum levels in attempt to resolve this problem. The ensuing debate on photon momentum, p, in matter is essentially known as the Abraham-Minkowski (AM) or related controversies [58] in electrodynamics, i.e., whether one should write: or (1.8) where is the reduced Planck’s constant, represents the angular frequency of light, c is its speed in vacuum, and denotes the refractive index of the medium. In terms of G inside a medium, Eqs in (1.8) are written as [58]: and . (1.9) ,Minnpc ,Abrpnc nAbr2dVcEHpMindVpDBChapter 1 Introduction and Literature Review 14 The Abraham MD is , whereas the Minkowski momentum density is characterized by . Though there are insightful arguments in favour of both momentum densities, the real issue is which one is measurable in a particular event, rather than the fact of which one is correct. Although it is believed that the idea of Abraham photon momentum as kinetic momentum of photon and Minkowski photon momentum as canonical momentum of photon resolves the conceptual complexity regarding these two rivalry momenta [58], their connections with the rivalry stress tensors and Lorentz/volumetric force formulas are still not clear. This issue has been investigated in some parts of this thesis. 1.7 Optical Tractor Beams Manipulation of objects with laser beams has always been a hot topic in scientific fields and science fiction. In the latest decade, "tractor beams" which drag trapped objects towards the light source, have attracted much attention [25,30-36,59-64]. 2()AbrcGEH /.Mink GDBChapter 1 Introduction and Literature Review 15 Fig.1.5 Different examples of tractor beams. (a) Distribution of optical force along the wave axis on a silver nanoparticle at the wavelength of 400 nm. The inset shows the region of negative optical force (NOF). (b) the optical force on a polystyrene sphere by a Bessel beam. It shows that optical pulling force is possible for different radii of polystyrene particles. (c) the longitudinal forces Fz and radial forces as a function of radial distance between the particle and beam axis. The shaded region represents stable pulling effect due to and . Objects near the axis are confined by gradient force and transported in propagation direction by scattering force. (d) polystyrene particles with radii of 800 nm and 1000 nm are separated by s-polarized beam. Figure reproduced with permission from: (a) ref. [32]; © 2010, Optical Society of America (OSA), (b) ref. [59]; © 2011, Nature publishing group, (c), ref.[25]; © 2011, American Physical Society (APS), (d) ref.[36]; © 2013, Nature publishing group. A tractor beam is a traveling wave which exerts a negative scattering force to a scatterer and pulls it opposite to the propagation direction of light. Most beams of light do not act as tractor beams because radiation pressure tends to drive illuminated objects down-stream. A tractor beam pulls and transports an object over a long rF0zF 0rF Chapter 1 Introduction and Literature Review 16 distance, whereas an optical tweezers cannot be termed as a tractor beam because of its inherently limited range and due to the use of gradient force. Previous works regarding optical pulling using tractor beams can be divided into well-defined procedures. For example: (a) It has been demonstrated [32] that there is a region of negative optical force (NOF) for a class of designed optical beams [Fig. 1.5(a)]. However, the regions of NOF are small compared to the incident wavelength. Thus a nondiffraction beam constituted by 24 plane waves with the same kz component was used in Ref. [31], (b) generating a negative force by exciting multipoles inside a scatterer [59], (c) the use of a single non-paraxial Bessel beam [Fig.1.5(c)] [25, 30], (d) A tractor beam generated by the interference of a single Gaussian beam and its reflection by a dielectric mirror also demonstrated the pulling effect [36]. This kind of tractor beam can sort different sizes of particles by switching between s- and p-polarized beams [Fig.1.5 (d)]. (e) Building a standing wave named the optical conveyor-belt [60, 61], (f) employing gain [62] or metamaterial [63] background media, (g) placing the particle between two different media to make linearly polarized light [35,43] to act as a tractor beam and (h) introducing the optical conveyor [33] as a travelling wave with two super-imposed zeroth ordered Bessel beams and later the all-rounder beam with two super-imposed higher order Bessel beams with frequency difference between them [34] as shown in Fig. 1.6 (a) and (b). However, most of the times optical pulling effects have been discussed from the point of view of quasi static analysis based on the external dipole force methods [30,34,62]; without considering much analysis on full electrodynamic calculation or the Lorentz force [27,28,65]. What is the physical connection of optical pulling force with Lorentz volumetric force? Answers of these questions are investigated later in this thesis. Chapter 1 Introduction and Literature Review 17 Fig.1.6: Continuous and three dimensional stable manipulations of fully immersed particles by using the optical gradient force generated by superimposing [34]: (a) coaxial pair of Bessel beams of reversed orders and different frequencies and (b) Non coaxial setup for a tractor beam with two Bessel beams of different helical nature. Only the particles located within a certain spatial range can be levitated using this latter beam set-up. Figures adapted and reproduced with permission from ref. [34], © 2015, WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim. 1.8 Reversal of Near Field Optical Binding Force Fig.1.7: An example of binding force reversal due to Fano resonance. (a) The geometry of the system and the incoming light configuration. The plasmonic dimer consists of two nanorods with L1 = 100 nm, L2 = 280 nm, and gap g = 10 nm [22]. The diameter of the nanorods is d = 40 nm. (b) Optical features associated with the Dipole-Quadrupole Fano resonance. Optical extinction cross sections and optical scattering cross sections [22]. (c) Optical forces between the two nanorods and along the k-direction [22]. Figures adapted and reproduced with permission from ref. [22], © 2013, Optical Society of America (OSA). Chapter 1 Introduction and Literature Review 18 In addition to the optical forces directly from light-matter interaction, optical binding forces can emerge from the momentum exchange between multiple particles. Light-mediated optical binding offers a new way for self-assembly, organization, and cell sorting [3,4, 66-68]. The spheroidal nanoparticles could arrange themselves into clusters, chains, photonic lattices and linear lines by optical binding [69,70]. Alternately, computer-generated holographic tweezers could also be used to organize particles into such nanostructures in three dimensions [71,72]. Especially optical binding forces of metallic nanoparticles have been found to be over 20 times larger than the gradient force [73] and they decayed slower than the gradient force, [74] which enables extremely stable nanometer manipulation. Recently the reversal of near field optical binding force has achieved much attention [22,23,38,75,76]. However, such reversal of near field binding force is mostly connected with the plasmonic dimers [22, 23, 38, 75]. Based on Fano resonance, the reversal of near field optical binding force has been reported in [22] and [23] for nanorod structures [22] and for disk along with a ring structure [23] as shown in Figs. 1.7 and 1.8. ‘Whether such Fano resonance (raised from heterodimer interaction) is a universal process of achieving binding force reversal?’- Answer of this question is still unknown. In addition, though Lorentz force analysis has been applied previously in [65,77-80] to understand the mechanism of chirality induced force, Luneburg lenses, mechanical interaction between light and graded index media, cloaking effect, background effect on radiation pressure; such analysis has never been applied to understand the plasmonic effects and plasmonic binding force. How the optical force distributes itself and which part of the total force is mainly responsible for plasmonic binding force reversals will be discussed based on the Lorentz force dynamics in this thesis. Chapter 1 Introduction and Literature Review 19 Fig. 1.8: Another example of binding force reversal due to Fano resonance. (a) disk-ring structure and the incident light [23]. (b) Scattering cross section of the disk-ring structure (black-solid line) and the individual nanodisk (black-dashed line) [23]. The local field enhancement of the individual nanoring (blue dash-dotted line) is also plotted. (c) Optical binding force spectrum of the disk-ring structure. The reversal points are marked with A, B, and C with respect to the vertical dashed lines, and the other reversal points are marked with A’, B’, and C’ according to the vertical dotted lines [23]. Figures adapted and reproduced with permission from ref. [23], © 2013, Optical Society of America (OSA). 1.9 Thesis Outline One of the main goals of this thesis is to investigate the problems/dilemmas of optical volumetric/Lorentz forces specially observed in real radiation pressure experiments performed up to date [cf. sections 1.4 to 1.6]. Moreover, Lorentz force based analysis is so far restricted to simple cases. In literature, Lorentz force analysis for optical pulling force (tractor beams) and the reversal of near field optical binding force (i.e. for plasmonic objects) have not been investigated in details [cf. sections 1.7 and 1.8]. As a result, in this thesis, physics and applications of Lorentz force have Chapter 1 Introduction and Literature Review 20 been investigated in details with special interest in tractor beam effect, plasmonic objects, chiral objects and the objects embedded in a generic material medium. At the beginning part of this thesis, in chapter 2 and 3, we have considered comparatively simple cases such as objects placed in air. Time averaged total force calculated by well-known Lorentz force and the external stress tensor method are found in full agreement for those two chapters. More specifically, chapter 2 demonstrates the Lorentz force analysis for tractor beam like effect. At the end of section 1.7 of this chapter, we have already discussed shortly that the stimulating connection of optical Lorentz force with counterintuitive tractor beam effect has not been investigated in details so far. Chapter 2 attempts to do so. It is shown based on Lorentz force analysis that the bound surface charges of Rayleigh scatterer experience backward force, which overcomes the positive bulk force and ultimately results in the net pulling of the scatterer for several spatial regions outside the two dielectric hollow core cylindrical waveguide. Later, this idea of ‘dielectric waveguide tractor beam’ has been extended for dielectric coupler set-ups. At the end of section 1.8 of this chapter, we have discussed shortly that the connection of optical Lorentz force dynamics with plasmonics and specially the reversal of optical binding force have not been investigated in details so far. Chapter 3 demonstrates the Lorentz force analysis for plasmonic off-axis and on-axis spherical heterodimers. It is shown that the reversal of longitudinal binding force can be easily controlled due to forced symmetry breaking. Though it is commonly believed that plasmonic forces mostly arise from the surface force and Fano resonance can be a promising way to achieve binding force reversal, our study based on Lorentz force dynamics suggests notably opposite proposals for the case of plasmonic spherical heterodimers. In contrast, in chapters 4, 5 and 6, we have considered a little bit complicated cases such as objects those are embedded in material medium. Specially chapters 4 Chapter 1 Introduction and Literature Review 21 and 5 deal with very fundamental topics of optical force: consistency of stress tensors, volumetric forces and photon momenta inside matter. The controversies and the problems related with such topics have already been discussed shortly in sections 1.4, 1.5 and 1.6 of this chapter. In chapter 4 it is shown that the well-known Lorentz force leads to inconsistent result (pushing force) instead of the experimentally observed optical pulling force for interfacial tractor beam experiment. Our suggested interpretation in chapter 4 supports the Minkowski approach only for the purpose of optical momentum transfer to the embedded scatterer for the interfacial tractor beam experiment [35] rejecting Peierls’ and Abraham’s approach, although the momentum of photon in the continuous background medium should be considered as the type of Abraham for the calculation of the bulk part of Lorentz volumetric force distribution. Later chapter 5 discusses in details why well-known Lorentz forces fail not only in interfacial tractor beam experiment but also for several other real experiments. It is demonstrated that at the boundary of an object, the difference of the external Minkowski ST and internal ST of Chu (and Einstein-Laub) is found in agreement with the surface force yielded by Chu (and Einstein-Laub) volumetric force only when the background is air rather than a material. We identify this as one of the main reasons of the disagreements observed in the major radiation pressure experiments which include material medium as background. Some other notable reasons of such disagreements have also been identified and discussed in that chapter. We also demonstrate that it is still possible to establish different equivalent time-averaged Lorentz/volumetric force formulas based on the fulfilment of just two ‘consistency conditions’. Based on those proposed ‘consistency conditions’, finally, we show that by modifying the Einstein-Laub or Chu formulation; time-averaged STs and volume forces are obtainable, which can overcome the aforementioned inconsistencies and yield the ‘correct’ force similar to Minkowski’s force for almost all the previous real experiments. Specially, our proposed modified Einstein-Laub ST can be considered Chapter 1 Introduction and Literature Review 22 as an efficient mathematical toolkit, an alternative of time and memory consuming volumetric forces, to yield the internal bulk force of a chiral or achiral object embedded in complex material backgrounds (i.e. homogeneous, heterogeneous, bounded etc.). The behavior of Fano resonance and the reversal of near field optical binding force of dimers over different substrates have not been studied so far. This matter is investigated in details in chapter 6. In chapter 5, it has been discussed in details why Lorentz force should be modified when the embedding background is material one instead of air. In chapter 6, we have shown an interesting application of modified Lorentz force [i.e. modified Chu force] to control the reversal of optical binding force of plasmonic cubes placed over plasmonic substrate due to strong Fano resonance. Last but not least, based on our proposals in chapter 5, at the end of this thesis (chapter 7) we propose a new hypothesis named as ‘existence domain’. ‘Existence domain’ is the region either outside a scattering body taking only its exterior fields into account, or in its interior considering only the inside fields. Though almost all the time averaged distinct STs and volumetric force laws are restricted to the idea of ‘existence domain’, we demonstrate that the time averaged stress tensor and volumetric force law of Minkowski are free from such restrictions. In addition, we have also discussed in details the differences between time averaged and instantaneous force laws for different formulations. Notably only Minkowski’s theory remains valid for all circumstances without any modification. After almost 100 years of Einstein’s prediction, this thesis, (most probably) finally answers why Minkowski’s theory remains valid everywhere in space (both in instantaneous and time averaged scenario) and can be considered as the base of a unified theory of optical force and photon momentum. Proposals presented in this thesis can be very effective to resolve not only the dilemma of distinct stress tensors and optical Lorentz forces but also the controversy Chapter 1 Introduction and Literature Review 23 of Abraham-Minkowski photon momenta. Our proposals may also open a new window of optical pulling force/tractor beams due to the exclusion of conventional structured tractor beams along with the artificial exotic matters. Last but not least, the proposed ideas for the reversal of near field optical binding force may also be useful for the future applications of plasmonic dimers in the areas of improved plasmonic sensors, particle clustering and aggregation. Chapter 2 Lorentz Force and Its Application on Tractor Beam like Effect 24 Chapter 2 Lorentz Force and Its Application on Tractor Beam like Effect 2.0 Summary of chapter 2 The stimulating connection between the counter-intuitive optical pulling effects and the Lorentz force has not been investigated in literature. This chapter demonstrates that multiple absorbing or non-absorbing dielectric Rayleigh objects can be pulled locally with gradientless travelling waves outside a finite-sized cylindrical nano or micro waveguide, if it is made up of a hollow core along with the cladding of at least two different dielectrics of appropriate refractive indices. Lorentz force analysis reveals that the bound surface charges of Rayleigh scatterer experience backward force, which overcomes the positive bulk force and ultimately results in the net pulling of the scatterer for several spatial regions outside the waveguide. Finally, in order to control the pulling of multiple Rayleigh particles based on scattering force and binding force, we have proposed a possible cylindrical coupler set-up. The work presented in this chapter may open a new window of optical pulling force due to the exclusion of conventional structured tractor beams along with the artificial exotic matters. Chapter 2 Lorentz Force and Its Application on Tractor Beam like Effect 25 2.1 Introduction Controlling the optical force at nanoscale has important applications in emerging technological developments [81, 82]. In order to control the optical force at nanoscale, nano-opto mechanical systems are of great interest nowadays for different novel optical manipulations [83-89]. For example, a way to create optical pulling effect for Rayleigh particles is an evanescent wave [83, 84], which is conventionally applied in nano-opto mechanical systems for particle manipulation. A slot waveguide is reported in [85], which proposes a way of possible long distance transportation by pushing the particles along the slot and also trapping the particles in a two-dimension way. Later, the trapping and transport of polystyrene nano particles and DNA molecules have been demonstrated using such slotted silicon waveguides [86]. Nano-fiber can also be used for trapping or pushing atoms [87]. Instead of optical pulling, the light driven transport of nano particle against the fluidic flow has been experimentally demonstrated in presence of micro-meter sized fiber [88]. Using two oppositely directed beams from two ends of a 10-μm dielectric waveguide, trapping and pulling of nano particles have also been predicted [89]. However, a conventional dielectric micro-waveguide (cf. Ref. [89]) with the aid of only forward travelling fields has not been used to pull non-resonating conventional dielectric nano particles outside the fiber or waveguide, which can be considered as the tractor beam like effect. In contrast to the conventional beams with pushing force, a tractor beam is a customized travelling light beam which pulls any scatterer opposite to the propagation direction of light [25, 30, 33,34,35, 36, 62, 90, 91], which are distinct from the aforementioned opto-mechanical systems. In this chapter, we demonstrate a possible scheme of three dimensional (3D) nano-opto-mechanical system to create local tractor beam effect for multiple non-Chapter 2 Lorentz Force and Its Application on Tractor Beam like Effect 26 resonating Rayleigh particles (Silicon) outside the proposed finite-sized nano or micro waveguides with non-structured light and dielectric materials. The proposed pulling method does not require sophisticated control of structured light or exotic materials or the complicated background [25, 30, 33, 34, 35, 36, 62, 90, 91]. Only a waveguide or a coupler with a hollow air core along with the cladding of two different conventional dielectrics with appropriate permittivity is needed, which can be excited by simple non-structured source of light. We name our setup as “Dielectric Waveguide Tractor Beam” (DWTB), in which a forward propagating wave can drive multiple Rayleigh particles locally in a backward direction outside the finite-sized waveguide in absence of any backward propagating wave [91]. Based on the interaction between the scattering force [92] and optical binding force [3], we also propose a local pulling mechanism created by a cylindrical waveguide coupler. Due to the presence of additional controllable parameters, probably the manipulation of small nano particles [93, 94] (i.e. Rayleigh particles) will be flexible outside the proposed coupler setups. However, most of the times optical pulling effects have been discussed from the point of view of quasi static analysis based on the external dipole force methods [30, 34, 62, 92]; without considering much analysis on full electrodynamic calculation or the Lorentz force [44,65,77,78,95,96]. Though Lorentz force analysis has been applied previously in [65, 77, 78, 96] to understand the mechanism of chirality induced force, Luneburg lenses, mechanical interaction between light and graded index media, cloaking effect, background effect on radiation pressure; such analysis has never been applied to understand the tractor beam like effect. Our verifications on the consistency between the full electrodynamic analysis based external stress tensor and volume/Lorentz force method in three dimensional (3D) environment may open a new window to understand the physical mechanism [65, 77, 96] of optical force not only for optical pulling but also for more complicated set-ups reported in [38, 97-99] (where dipolar force [92] does not lead to the consistent result). Though in this work Chapter 2 Lorentz Force and Its Application on Tractor Beam like Effect 27 we have limited our discussion on non-resonating dielectric Rayleigh particles (both non-absorbing and absorbing), we believe that our proposed DWTB based pulling can also be applicable for resonating and plasmonic Rayleigh objects [38] due to the dominance of the identical part of the Lorentz force. 2.2 Single Rayleigh particle outside the waveguide: Force calculation We specify that throughout this thesis we refer to 'exterior' or 'outside' forces as those evaluated outside the volume of the macroscopic particles, while ‘interior’ or 'inside' refer to those quantities inside this object volume. In order to consider the realistic effects, in this chapter we have done all the numerical calculations /full wave simulations in 3D structures [100]. The proposed simplest setup is illustrated in Fig. 2.1(a). The Silicon scatterer (with refractive index of ns=3.5) is placed at the outside of the finite-sized nano or micro waveguide, which is made up of a hollow core and two different dielectric layers of cladding [i.e. Glass and Potassium Niobet (KNbO3); where the refractive index of the second clad should be higher than the first one]. The ‘outside optical force’ [35, 43] is calculated by the integration of time averaged Minkowski [35, 43] stress tensor at r= 1.01a = a+ [It means we are selecting a spherical surface with a radius slightly larger Chapter 2 Lorentz Force and Its Application on Tractor Beam like Effect 28 5001000150020002500-0.2-0.10.00.1(b)Fz (pN) z (nm) Fig. 2.1: Illustration of optical pulling force for a single Silicon particle (refractive index 3.5 and radius of 30 nm) placed at different z-positions outside a 3D cylindrical waveguide. Dotted (….) lines represent the possible time averaged force on the particle at different z-positions based on the actual force on the particle indicated by ‘.’ marked points. (a) A 10 µm sized (this specific length has been considered according to Ref. [89]) three dimensional (3D) waveguide is excited by a plane wave source of wavelength 1500 nm. The radius are r0=500 nm, r1=600 nm, and r2=700 nm. The refractive indexes are n0=1.0, n1=1.5 and n2=2.25. A single silicon particle of radius 30 nm and refractive index of 3.5 is placed at different z-distances outside the waveguide. The radial-directed distance, d, is fixed at 200 nm for different z-distances/positions of the particle. (b) Fz (in pico Newton) on the single Rayleigh scatterer. Figures adapted and reproduced with permission from the work: Mahdy et al., Annalen der Physik, doi: 10.1002/andp.201600213 © 2016, WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim. than the object, on which the integration of Stress tensor is calculated for the total optical force (for details cf. Fig. 1.1 in chapter 1)] employing the background fields of the scatterer of radius a: OutoutTotalout****outoutoutoutoutoutoutout1 Re12(][)2dDEBHEDHBFTsTI. (2.1) Where ‘out’ represents the exterior total field [i.e. incident plus the scattered field] of the scatterer; E , D , H and B are the electric field, electric displacement field, magnetic field, and magnetic induction field, respectively, while represents the time average and I the unity tensor. Chapter 2 Lorentz Force and Its Application on Tractor Beam like Effect 29 On the other hand, based on the Lorentz force, the total force including the surface force and the bulk force [65,77,78]) can be written as: TotalVolumeBulkSurfBulkSurfacedvdsFFFFff (2.2) Where outinoutinouti**Surfaceat**0atatnoutin[]ˆˆ ()(),22oeavgmavgr ar ar afEHEEHHEEnHHn(2.3) Surfacef is the surface force density (the force which is felt by the bound electric and magnetic surface charges of a scatterer), which is calculated just at the boundary of a scatterer [65,77,78]. ‘in’ represents the interior fields of the scatterer; ‘avg’ represents the mathematical average of the fields just at the boundary [i.e. exactly at r=a].e and m are the bound electric and magnetic surface charge densities of the scatterer, respectively. The unit vector ˆn is an outward pointing normal to the surface. On the other hand, the time averaged bulk force density can be expressed as [65,77,78]: ****Bulk0inin0inin0inin0inin11Re[ ()()]Re[ (){}(){}2]2ssii EHfEEBDH (2.4) Bulkfis the bulk force density, which is calculated from the interior of the scatterer by employing the inside field [65,77,78]. s is the permittivity and s the permeability of the scatterer. It is important to note that the quasi static analysis based on ‘external dipolar force’ [38,92], which has also been described as the Lorentz force in [38] is quite different from the full electrodynamic analysis based on the Lorentz force defined in Eqs (2.2) - (2.4). Chapter 2 Lorentz Force and Its Application on Tractor Beam like Effect 30 2.3 Detailed analysis of set-up 2.3.1 Electromagnetic fields and modes in a cylindrical waveguide The primary source is a simple x-polarized plane wave 0i zxEE e and 0E has been set to 1 V/m in the simulation model. This source [which is placed at around z=0 (the starting length of the waveguide) in the full wave simulation model; cf. Fig.2.1 (a)] is used to excite the conventional/modified Bessel fields inside/outside our finite-sized cylindrical waveguide [101, 102]. This intrinsic wave-fields can be calculated by a transfer matrix method [101, 102]. Throughout this work three different sized (length) 3D cylindrical waveguides have been considered: 10 µm (this specific size has been considered according to Ref. [89]), 15 µm and 20 µm. In each uniform cylindrical cladding of our waveguides, the four field components (Ez, Hφ, Hz, Eφ) can be expressed as a linear combination of any two types of Bessel functions [102]. Our waveguide has two dielectric claddings, where the guiding modes are leaky as the confinement is not perfect (i.e. the higher refractive index of the second clad). Just outside the waveguide (i.e., in the air background), there will be no incoming waves for the idealistic case [102]. It should be noted that the fields (Ez, Hφ, Hz, Eφ) calculated in [102] are the intrinsic fields in absence of a scatterer. Due to the finite structure of our waveguide and the presence of the scatterer, the intrinsic fields in [102] should be modified near a scatterer; and this has been managed by the full wave simulation based techniques applied in this work. 2.3.2 Behavior of external wave-field of a single scatterer placed outside proposed waveguides In Fig.2.1 (b), it is shown that when a single scatterer is placed around our proposed waveguide at different z-distances, the scatterer experiences a net optical Chapter 2 Lorentz Force and Its Application on Tractor Beam like Effect 31 pulling force towards the source in several spatial regions. Calculations have been done based on the outside stress tensor and also volume force density formula, respectively. They are in quite good agreement. Though in [103] the optical pulling outside a dielectric waveguide slab has been explained based on the modal conversion and resonant coupling of the scatterer, optical pulling force in our set-ups is not due to resonant coupling. We also find that a 3D cylindrical waveguide with hollow core along with a single dielectric clad is not capable of pulling the Rayleigh scatterer in continuous spatial position (which is in agreement with refs. in [104] where quasi static analysis has been emphasized to explain the time averaged total force). In case of optical pulling with our two-layered hollow core waveguide, the scatterer (which is a generic Rayleigh particle) is not in resonance at all [also cf. the introduction of ref. [97], the trapping of non-resonant particles]. Ultimately, the additional second clad along with the presence of a weak standing wave in it assists the pulling effect and alters the spin-orbit interaction [105] in presence of the particle. The effect of altered electrodynamic interaction (i.e. the pulling force of non-resonating particle), due to the addition of the suitable second clad for our setups, cannot be included in a simple manner in the quasi static force formula: such as external dipolar force in ref. [92]. As a result, a full electrodynamic analysis will be discussed in the next section. In order to investigate the connection of our observed pulling force with the Poynting vector distribution [92], the total Poynting vector distribution of the waveguide is examined via full wave simulation [cf. Fig. 2.2 (a), (b) and (c)]. It is interesting to note that the real part of Ez is found in opposite phase after each 500-nm z-distance. Another interesting fact is that the Poynting vector direction at z=1000 nm plane is negative for the air core, which is discussed in detailed in [106]. However, it is not the generic reason of optical pulling force outside the waveguide [104]. For example- at z=1500 nm, the force is still negative but the total Poynting vector is Chapter 2 Lorentz Force and Its Application on Tractor Beam like Effect 32 positive everywhere just like at z=500 nm where the force is positive. So, the conclusion is that: negative Poynting vector [106] is not responsible for such optical pulling force [104] in general. Chapter 2 Lorentz Force and Its Application on Tractor Beam like Effect 33 Fig. 2.2: Total Poynting vector distribution at the planes of z=500 nm (a), 1000 nm (b) and 1500 nm (c). The insets show the responding Re(Ez). It is observed that at z=1000 nm plane when optical pulling force occurs [cf. Fig. 2.1(b)] outside the cylindrical waveguide of Fig.2.1 (a), negative Poynting vector [cf. Fig. 2.2(b)] takes place. At the same time, inset of Fig. 2.2(b) represents opposite phase shift of the Real (Ez) with respect to Fig. 2.2(a). However, neither the negative Poynting vector nor the phase shift of Real (Ez) is the generic reason of the pulling effect. For example, Fig. 2.2(c) and in its inset: the total Poynting vector is positive and the Real (Ez) does not experience any opposite phase in comparison with the Real (Ez) given in the inset of Fig. 2.2(a). In Fig. 2.1(b) it is observed that the particle experiences local pulling force at z=1500 nm but it experiences pushing force at z=500 nm. Figures adapted and reproduced with permission from the work: Mahdy et al., Annalen der Physik, doi: 10.1002/andp.201600213 © 2016, WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim. 2.4 Single or multiple Rayleigh particles outside the waveguides: Lorentz force Dynamics Chapter 2 Lorentz Force and Its Application on Tractor Beam like Effect 34 05001000150020002500-0.10.00.1(b) Fz (pN)z (nm) FVolume FOut FSurf FBulkx=200nm Fig. 2.3: (a) Possible Illustration of optical pulling force for multiple Silicon particles (all with refractive index 3.5 and radius 30 nm) placed outside the 3D waveguide with a length of 15 µm; light wavelength 1500 nm. For clarity, it does not represent the actual number of particles. In actual simulation set-up, due to time and computer memory restriction, we have placed total 46 particles (two different rows; each contains 23 particles) covering up to 2.5 µm of the total 15 µm sized waveguide. The z-direction center-to-center separations between the particles is 50 nm, and the separations of them to the waveguide are d=70 nm and d=200 nm, respectively. (b) Dotted (….) lines represent the possible time averaged force on the particle at different z-positions based on the actual force on different particles indicated by ‘.’ marked points.zF for particles at x=200 nm. outF is obtained using the time-averaged ST, whereas the force VolumeF is calculated by Eq (2.2) from fields inside the object, and SurfF just from the surface. It is observed that the negative force is mainly the ultimate result of negative surface force for each scatterer, which overcomes the positive bulk force. In fact the time averaged total force follows the trend of surface force almost in all spatial positions. Figures adapted and reproduced with permission from the work: Mahdy et al., Annalen der Physik, doi: 10.1002/andp.201600213 © 2016, WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim. When several dielectric Rayleigh objects are placed outside the waveguide in Fig. 2.3(a), we have shown that optical pulling force remains valid as shown in Fig. 2.3(b). However, the magnitudes of the forces change in comparison with the single particle case due to the presence of binding force [3]. The explanation of the observed pulling force does not remain simple for this set-up because the idea of conventional conservative forces like gradient, scattering and curl force or the total external dipolar force cannot be applied even for object radius a ≪ λ Chapter 2 Lorentz Force and Its Application on Tractor Beam like Effect 35 [38, 97-99]; where: (i) the modification of modes due to interaction with the Rayleigh scatterers significantly alters [98] the properties of the force of an intrinsic dipole [92,107] and (ii) multipoles [38] are excited along with the retardation effect. For such cases, Lorentz force analysis [44,65,77,78,95,96] (which is a full electrodynamic analysis) can be an alternative, accurate and physically insightful way to explain complex setups like ours, though so far no attempt has been taken based on the Lorentz force to explain the tractor beam effects [25, 30, 33,34,35, 36, 62, 90, 91] or complex situations reported in refs [38, 97-99,104,107]. From the numerically calculated results in Fig. 2.3(b), it can be explained that the bound charges in the surface of the scatterer experience backward surface force SurfF and it overcomes the positive value of the bulk force BulkF caused by the interaction between magnetic induction and induced polarization current (i.e. bound current [78]). Hence, the resultant force ultimately becomes negative and the dielectric Rayleigh particle experiences local backward pulling in some continuous spatial positions outside our proposed waveguide structure. In presence of external electric field, bound charges can oscillate around a mean position [108]. In this work, we are also observing a case where optical pulling force is arising mainly due to dominant negative surface force, while previous tractor beam works [25, 30, 33, 34, 35, 36, 62, 90, 91] have not investigated based on the idea of surface force and bulk force. Although such dominance of surface /polarization charges (that should lead to the dominance of surface force) is common for plasmonic structures [109] due to local field enhancement, the dominance of surface force for dielectric scatterers with positive permittivity is a little bit unusual and highly dependent on the shape of the scatterer (cf. the force on crescent shaped object in [78]). Although it is sometimes considered that the bound charges and currents are merely computational tools for calculating electric and magnetic fields in matter, yet these Chapter 2 Lorentz Force and Its Application on Tractor Beam like Effect 36 charges and currents are no less real than free charges and currents and can be measured experimentally [110,111]. According to ref. [110] [cf. Eq (10) given in [110]], within a linear, homogeneous, isotropic dielectric, in regions where there are no (free) volumetrically distributed charges, there are also no bound volumetric charges. This leads to an interesting conclusion that: in case of optical force, electrons that are more or less free to roam around the lattice may be said to act as free electrons and can be modelled as bound charges [95]. For the force exerted by the E-field is −(�·P)E [95] or 0() EE, as shown in our Eq (2.4). Here, P is the polarization vector. Though this force due to volume charge density vanishes inside an isotropic and homogeneous object [65,77,78], the net effect is that surface bound charges contribute to the total force as shown in Eq (2.3). We have examined several other setups with placing multiple particles outside the proposed cylindrical waveguides. For all the set-ups with finite-sized cylindrical waveguides or couplers, the common reason of optical pulling effect is observed from the dominance of the negative surface force due to bound surface charges of the Rayleigh particles. I.e. in Fig. 2.1(b), nearly at z=1000 nm, the magnitude of optical pulling force on a single scatterer reaches local minima (negative maxima) due to strong negative surface force on the bound charges of the scatterer. This conclusion on surface force remains valid even for the surface force on a particle placed with multiple other particles in Fig. 2.3(b) for the 15-µm waveguide. However, in Fig. 2.3(b), due to multiple scattering and binding force [3], the bulk force changes significantly in comparison with the case of a single scatterer. As a result, though the total force on a single scatterer reaches the local minima (negative maxima) value nearly at z=1000 nm in Fig. 2.1 (b) along with strong surface force, total force on the scatterer in presence of multiple scatterers do not reach the local minima (negative maxima) value at z=1000 nm even with the presence of strong surface force, as shown in Fig. 2.3(b). Chapter 2 Lorentz Force and Its Application on Tractor Beam like Effect 37 2.5 Robustness of pulling force to absorption, wavelength and waveguide length Fig. 2.4: Robustness test. (a) Same configuration of Fig. 2.3(a). But the length of the waveguide is 10 µm. (b) Dotted (….) lines represent the possible time averaged force on the particles at different z-positions based on the actual force on different particles (total 46 particles; 23 in each row) indicated by ‘.’ marked points [also cf. Fig. 2.3]. Moderate loss (n= 3.5+0.5i) is introduced to all the particles placed outside the 10 µm waveguide. The magnitude of the pulling force is decreased very slightly and still pulling force is quite possible in presence of moderate absorption. It proves the robustness of our scheme. (c) The effect of the change of incident wavelength is shown for lossless particles at x=200 (total 46 particles; 23 in each row). If wavelength of the light source is changed for the set-up of Fig. (a) [i.e. 1000, 1250, 1750 and 2000 nm] but the particle size remains fixed; the force magnitudes have been dramatically varied for different wavelengths of the source. In lower wavelengths (i.e. 1000 nm wavelength) pulling force is greatly reduced, which indicates that the proposed idea of DWTB is applicable only for Rayleigh particles. Figures adapted and reproduced with permission from the work: Mahdy et al., Annalen der Physik, doi: 10.1002/andp.201600213 © 2016, WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim. Chapter 2 Lorentz Force and Its Application on Tractor Beam like Effect 38 In order to examine the robustness of our proposed set-ups, we have varied different key parameters. For example- at first we have introduced moderate loss (ns=3.5+0.5i) to all the particles placed outside the waveguide [cf. Fig. 2.4(a) and (b)]. For absorbing scatterers, Eq (2.4) gets two different parts known as the bound force and conduction force [44]: BulkbcFFF (2.5a) Currents arise in materials when there is a non-uniform distribution of charge. In dielectric materials, there is a current density corresponding to the net movement of electric dipole moments per unit volume, i.e. the polarization P. When the polarization density changes with time, the time-dependent bound-charge density creates polarization current density (bound current [78], [110]). This bound current interacting with magnetic induction causes the bound force part of total bulk force. Here, the bound force bF can be written as [44]: **0inin0inin1 Re[(Re[]){}(Re[])}2]{bbssdviiFfEBDH (2.5b) On the other hand, the time averaged conduction force cF (which is an additional interior force that arises due to the free currents inside the scatterer) takes the value [44]: **inininin1 Re2ccIIdvdvFfEBHD (2.5c) Here I and I are the imaginary parts of the permittivity and permeability of a scatterer, respectively. It is observed that even if multiple absorbing scatterers are placed outside our waveguide, still optical pulling force occurs very similar to non-absorbing case as shown in Fig. 2.4 (b). Chapter 2 Lorentz Force and Its Application on Tractor Beam like Effect 39 Next, we have varied the wavelength of light source in Fig. 2.4(c). We conclude that our proposed pulling effects work only for the Rayleigh limit of the particles (for both absorbing and non-absorbing). If the primary source wavelength is decreased too much (or by increasing the particle size; which is connected with the electrical size = ksa; where ks is the wave number ‘inside’ the material of scatterer), only few particles at random positions experience optical pulling force [cf. Fig. 2.4(c)]. However, as long as the Rayleigh limit (electrical size) is maintained, local pulling effect is observed for different wavelengths of the primary source. Finally, the overall length of our waveguide structures has also been varied, from 10 μm to 20 μm, and we have found that all the proposals introduced so far still remain valid. 2.6 Multiple Rayleigh particles outside the coupler 10002000300040005000-0.3-0.2-0.10.00.10.20.3(b)x=200nmCylindrical z (nm) Fx (pN) Fz (pN) Fig. 2.5: Possible Illustration of optical pulling force for multiple Silicon particles placed outside two identical 10 µm sized (length) 3D cylindrical waveguides (a directional coupler). (a) The ‘possible’ set-up of 3D particles placed together at different z-distances at d=200 nm apart from each waveguide. Material and the configuration of the waveguides and particles are the same as those used in Fig. 2. 3. (b) Fig. (a) does not represent the actual number of particles. In actual simulation set-up due to time and computer memory restriction, we have placed total 48 particles in one row covering up to 5 µm of the total 10 µm sized waveguides. The forces (zF andxF ) on an individual Rayleigh scatterer placed at different z-distances outside the waveguide (at fixed d=200 nm). Dotted (….) lines represent the possible time averaged force on the particle at different z-positions based on the actual force on different particles indicated by ‘.’ marked points. The two identical waveguides are excited by a plane wave source of wavelength 1500 nm. We verified that the pulling Chapter 2 Lorentz Force and Its Application on Tractor Beam like Effect 40 force due to cylindrical waveguide in Fig. 2.5(b) is also originated from negative surface forceSurfF(not shown). Figures adapted and reproduced with permission from the work: Mahdy et al., Annalen der Physik, doi: 10.1002/andp.201600213 © 2016, WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim. Based on the interaction between the scattering force [92] and optical binding force [3], we have also proposed a local pulling mechanism created by a cylindrical waveguide coupler as shown in Fig. 2.5. Importantly, (i) by changing the phase of the guided modes in the two waveguides [76] or (ii) by differing the intensity (input power) of primary light source [89] or (iii) by increasing the intensity (input power) of primary light source [88], it may be possible to achieve a flexible way of multiple particle manipulation based on the model system of coupler set-up proposed in Fig. 2.5. In Fig. 2.5, it is also observed that optical pulling force on multiple particle towards -z direction exists with moderate magnitude not only at the starting length but also at the later parts of the waveguide. However, the surface force part of the Lorentz force,SurfF, (not shown) on the scatterer has also been observed negative and dominant for that set-up. It should be noted that the coupler set-up is just a possible model system [also cf. ref. [112]]. If we consider the lines of multiple Rayleigh particles from different distances from the two waveguides [i.e. different fixed x-positions, particle aligned in z-direction]; particles, which are almost at the same distance from both the waveguides, should experience smaller lateral gradient force in comparison with the single waveguide system. 2.7 Conclusion In summary, we have demonstrated a possible way of achieving the local pulling force for multiple non-resonating Rayleigh objects considering both the exterior and Chapter 2 Lorentz Force and Its Application on Tractor Beam like Effect 41 interior dynamics behaviors of optical force. The robustness of our proposed scheme, local pulling force of multiple particles in presence of a structured cylindrical waveguide or a coupler, has also been tested by varying several key parameters. Probably the space dependent local optical forces can be adjusted (positive or negative) for multiple nano particles outside the finite-sized nano or micro waveguides/couplers based on the ideas presented here. The consistency between the Lorentz force method and external stress tensor (i.e. full electrodynamic calculation [113] instead of the approximate quasi static analysis) has been demonstrated in details for complex 3D set-ups. The work presented in this chapter has provided not only a novel way to realize opto-mechanical tractor beam effect with the simple light source and material but also a physical insight on the observed effects considering the Lorentz force dynamics of both absorbing and non-absorbing scatterers. Though we have restricted our discussion for non-resonating dielectric Rayleigh particles (both non-absorbing and absorbing), we believe that our proposed DWTB based pulling can also be applicable for resonating and plasmonic Rayleigh objects [38,114] due to the very high dominance of surface force [114] part of the Lorentz force. For very recent trend of metamaterial cladded hollow core cylindrical waveguides [115, 116], this work may also open a new way of future investigations: i.e. an alternative mechanism of optical pulling force on the particles placed inside or outside the core of such cylindrical [115, 116] waveguides. Chapter 3 Lorentz Force on Plasmonic Spher. Heterodim.:Rev.of Bind. Force 42 Chapter 3 Lorentz Force on Plasmonic Spherical Heterodimers: Reversal of Binding Force 3.0 Summary of chapter 3 In previous chapter the connection between Lorentz force and tractor beam effect has been investigated. In this chapter, the stimulating connection between the reversal of plasmonic binding force and Lorentz force dynamics has been investigated. As both bonding and anti-bonding modes are present in the visible spectra of well-known spherical plasmonic heterodimer sets, binding force reversal is commonly believed to occur for all such heterodimers. But this chapter suggests a very different proposal. We demonstrate that for the symmetry broken heterodimer configurations: reversal of lateral (for on-axis heterodimers) and longitudinal (for off-axis heterodimers: end-fire and nearly end-fire configurations) binding force follow fully distinct mechanisms. Interestingly, the reversal of longitudinal binding force can be easily controlled just by changing the direction of light propagation or just their relative orientation. Though it is commonly believed that plasmonic forces mostly arise from the surface force, and Fano resonance can be a promising way to achieve binding force reversal, our study based on Lorentz force dynamics suggests notably opposite proposals for the case of plasmonic spherical heterodimers. Chapter 3 Lorentz Force on Plasmonic Spher. Heterodim.:Rev.of Bind. Force 43 3.1 INTRODUCTION Fano resonances, super-scattering and plasmonic hybridization in nanostructures [39-41], [117] have received substantial attention in the area of plasmonics. The promising applications of plasmonic hybridization, super-scattering and Fano resonances have been investigated in improved sensitivity of the resonance [118], bio sensing [119], surface-enhanced Raman scattering [120,121], photonic propagation and wave guiding [122,123], plasmon-induced transparency [124] to super scattering [125] and many others [126,127]. In contrast, less attention is dedicated on near field optical force due to Fano resonance and plasmonic hybridization; especially for plasmonic dimers. When two metal nanoparticles are placed very closely to each other, the properties of their surface plasmons are dramatically modified. This configuration of nanoparticles is known as a “dimer”. Different behaviors (other than optical force) of such dimers have been studied in refs. [39,41,117,128-132]. Among the dimers, specially heterodimers show remarkable properties: Fano resonances [41,130], avoided crossing behavior [41] and optical nanodiode effect [41]. But the behavior of near field optical force for such heterodimers have not been studied in details. So far only two works [38, 133], as per our knowledge, have studied the behavior of binding force for on-axis spherical heterodimers. Though the behavior and reversal of near field optical binding force for spherical plasmonic on-axis homodimers [134- 136] (due to bonding and anti-bonding modes without plasmonic substrate) have been studied in details, such detail studies lack for the on-axis [38, 133] and off-axis spherical heterodimers. Here off-axis means end-fire [38] and nearly end-fire configuration [cf. Fig. 3.1 when the rotation angle, φ, of Chapter 3 Lorentz Force on Plasmonic Spher. Heterodim.:Rev.of Bind. Force 44 the particle is between 60 to 120 degrees]. Considering the heterodimer cases, based on Fano resonance, the reversal of near field optical binding force has been reported in [22] and [23] for nanorod structures [22] and for disk along with a ring structure [23]. ‘Whether such Fano resonance (raised from heterodimer interaction) is a universal process of achieving binding force reversal?’- Answer of this question is still unknown. A plasmonic spherical “heterodimer” set-up supports both bonding and antibonding plasmon modes at the same time due to its broken symmetry [117]. Hence it is expected that binding force reversal should occur almost for all the spherical heterodimer structures. But our work in this chapter suggests that reversal of lateral (for on-axis heterodimers) and longitudinal (for off-axis heterodimers) binding force of symmetry broken heterodimers follow fully different key parameters/mechanisms; i.e. later one depends on relative orientation and constructive dipole-quadrupole resonance (though not super-scattering mode [40]) but the former one on light polarization and the induced electric resonance. Most importantly, the reversal of longitudinal binding force can be easily controlled due to forced symmetry breaking just by changing the direction of wave propagation for a specific set-up of off-axis heterodimers or by changing their relative orientation. Interestingly, though reversal of optical binding force occurs for nano rods or other shapes due to Fano resonance [22, 23], we have demonstrated that Fano resonance [39, 41] does not contribute to binding force reversal for spherical heterodimers. These observations are quite different than the homo-dimer cases reported in [113,134-137]. Though Lorentz force analysis has been applied previously in [65,77,79,96] to understand the mechanism of chirality induced force, Luneburg lenses, mechanical interaction between light and graded index media, cloaking effect, background effect on radiation pressure; such analysis has never been applied in details to understand Chapter 3 Lorentz Force on Plasmonic Spher. Heterodim.:Rev.of Bind. Force 45 the plasmonic effects and plasmonic binding force. Although it is commonly believed that plasmonic forces mostly arise from the surface force/polarization induced charges [109,114], our study suggests a notably different proposal especially for the off-axis heterodimers. In addition, how the optical force distributes itself and which part of the total force is mainly responsible for plasmonic binding force reversals have also been discussed based on the Lorentz force dynamics without applying any approximate method. Table-3.1 of this chapter (given below) represents a very short overview of our overall investigation throughout this chapter. Observations of this work should be very useful for the future plasmonic applications of the heterodimers in the areas of improved sensors [39, 118,119], particle clustering and aggregation [113,134,135]. Chapter 3 Lorentz Force on Plasmonic Spher. Heterodim.:Rev.of Bind. Force 46 Fig. 3.1: Two particles of radii 100 and 50 nm are placed with inter particle distance from surface to surface ‘d’; d= 20 nm throughout this chapter. One particle centered at 0,0,0 and the other centered at cos ,sin ,0RR with 12=170 nmRdrrthe center-to-center distance of the two object. The angular displacement is ‘φ’ which is considered 0 degree when the dimers are on-axis in x direction. And the angular displacement is considered +90 degree [end fire configuration] when the dimers are on-axis in –y direction. This ‘φ’ should not be mixed up with the azimuthal angle of spherical co-ordinate system. Two different polarized light sources are applied propagating towards –y direction [in order to break the symmetry; light is shined from a specific side]: (i) For parallel polarization: x-polarized plane wave 0i yxEE e (ii) For perpendicular polarization: z-polarized plane wave0i yzEE e. Yellow color represents Au and Silver color represents Ag: (a) Au-Ag (b) Au-Au (c) Ag-Ag (d) Ag-Au (e) Au-Au and (f) Ag-Ag. Chapter 3 Lorentz Force on Plasmonic Spher. Heterodim.:Rev.of Bind. Force 47 Table-3.1: An overview on the behavior of binding force for on and off-axis heterodimers Heterodimer Set Number in Fig. 3.1 On- axis [ȹ=0] Off- Axis [ȹ=60 to 120 deg.] Pol. Binding Force reversal: (i)Lateral (ii)Longitudinal Comment (a) Inter-particle edge to edge gap, d, is always fixed 20 nm. (b) Heterodimer radii are fixed: 50 nm and 100 nm. Ag-Au (a) [=(d)] Yes Ʇ (i)Yes Lateral binding force reverses only for perpendicular polarization. For higher wavelength region: Reversal of force occurs due to zero surface and bulk Lorentz force at a specific wavelength near bonding (attractive force) resonance. For lower wavelength region: such force reversal can be recognized from the reversal of electric dipole moment of the smaller object. In fact, such reversals (repulsive to attractive) occur due to induced electric resonance near the bonding resonance mode. Au-Au (b) [=(e)] Yes Ʇ (i)Yes Ag-Ag (c) [=(f)] Yes Ʇ (i)Yes Ag-Au (a) [=(d)] Yes ‖ (i)No Au-Au (b) [=(e)] Yes ‖ (i)NoAg-Ag (c) [=(f)] Yes ‖ (i)No Ag-Au (a) Yes Ʇ and ‖ (ii) NoLongitudinal binding force reverses [for only Ag-Au and Ag-Ag case] only when the bigger particle rotates and the light is perturbed by the fixed smaller object at first. This reversal occurs due to the constructive dipole-quadrupole resonance and due to the dominance of the bulk Lorentz force. However, for all heterodimer sets attractive and repulsive force can be very easily controlled by changing the light propagation direction or changing the relative orientation of the dimers. Such control is not possible with the spherical homo-dimers. Au-Au (b) Yes Ʇ and ‖ (ii) NoAg-Ag (c) Yes Ʇ and ‖ (ii) NoAg-Au (d) Yes Ʇ and ‖ (ii) YesAu-Au (e) Yes Ʇ and ‖ (ii) NoAg-Ag (f) Yes Ʇ and ‖ (ii) Yes Chapter 3 Lorentz Force on Plasmonic Spher. Heterodim.:Rev.of Bind. Force 48 3.1 Optical Force calculation We again specify that throughout this thesis we refer to 'exterior' or 'outside' forces as those evaluated outside the volume of the macroscopic particles, while ‘interior’ or 'inside' refer to those quantities inside this object volume. In order to consider the realistic effects, in this chapter we have done all the numerical calculations /full wave simulations [100] in three dimensional (3D) structures. The proposed simplest set-up is illustrated in Fig. 3.1. The Gold (Silver) and Silver (Gold) particles are placed near to each other. The real and imaginary part of the permittivity of Gold is taken from the standard CRC data [100, 138]; but for Silver it is taken from standard Palik data [100,139] (the criteria is the best agreement with the FDTD fitting model in [100] for full wave simulation). Inter particle distance is ‘d’. Suppose the source is a simple x-polarized plane wave 0i yxEE epropagating towards –y direction. This set-up is a forced symmetry broken system which later plays a vital role for the reversal of binding force. If the heterodimer set-up is shined from the top, such ‘forced symmetry breaking’ is not possible. The ‘outside optical force’ [43, 84] is calculated by the integration of time averaged Minkowski [43,84] stress tensor at r=a+ employing the background fields of the scatterer of radius a: OutoutTotalout****outoutoutoutoutoutoutout1 Re12(][)2dDEBHEDHBFTsTI. (3.1) Chapter 3 Lorentz Force on Plasmonic Spher. Heterodim.:Rev.of Bind. Force 49 Where ‘out’ represents the exterior total field of the scatterer; ,,and are the electric field, displacement vector, magnetic field and induction vectors respectively, represents the time average and is the unity tensor. On the other hand, based on the Lorentz force, the total force (surface force and the bulk force [65,77,79,96]) can be written as: TotalVolumeBulkSurfBulkSurfacedvdsFFFFff (3.2) Where outinoutinoutinouti**Surface*n*0ˆˆ ()(),22eavgmavogfEHEEHHEEnHHn (3.3) ****Bulk0inin0in0inin0inin11Re[()()]Re[(){}(){}22]ssii EHfEEBDH (3.4) is the surface force density (the force which is felt by the bound electric and magnetic surface charges of a scatterer), which is calculated just at the boundary of a scatterer [65,77,79,96]. is the bulk force density, which is calculated from the interior of the scatterer by employing the inside field [65,77,79,96]. ‘in’ represents the interior fields of the scatterer; ‘avg’ represents the average of the field. and are the bound electric and magnetic surface charge densities of the scatterer respectively. is permittivity and is permeability of the scatterer. The unit vector is an outward pointing normal to the surface. As per we know, the Lorentz force dynamics of plasmonic particles and especially heterodimers have not been discussed previously. It is notable that the ‘external dipolar force’ [43, 84] (which has also been described as Lorentz force in [38]) is quite different than the Lorentz force defined in E DHBISurfacefBulkfemssˆnChapter 3 Lorentz Force on Plasmonic Spher. Heterodim.:Rev.of Bind. Force 50 our Eqs (3.2) - (3.4). Even if the quasi static analysis (i.e. dipolar force [40,41]) leads to wrong conclusion (for example- in refs [38, 97-99]); the agreement of Lorentz volume force [65,77,79,96] and external ST method based on full electrodynamic analysis, which is considered for all the force calculations in this work, should lead to the consistent result for realistic experiments. 3.3 Lateral binding force: On-Axis Spherical Heterodimers Behavior of optical binding force for on-axis spherical heterodimers has been studied in [38] considering the inter particle edge to edge gap of only 2nm. In addition, the size of the spherical objects has been considered only 10 nm and another one maximum 40 nm in [38]. However, we have observed that if the inter particle gap is increased (i.e. 20 nm instead of 2 nm), the reversal of optical binding force dies out for both polarizations of light. However, a more generic way of the reversal of binding force has been investigated and demonstrated in the next sub-sections. Such an investigation is also important to find the answer of the question: ‘Whether Fano resonance (raised from heterodimer interaction) is a universal process of achieving binding force reversal?’. 3.3.1 Parallel Polarization: No reversal of lateral binding force for Au-Ag, Au-Au and Ag-Ag on-axis heterodimers It is observed in full wave simulation that for Ag-Au, Au-Au and Ag-Ag heterodimer configurations, the reversal of lateral binding force [FBind (x) = (FB (x)-FS (x))] does not occur for the light polarized parallel to the dimer axis. Here FB(x) and FS(x) are the +x directed time averaged force on big and small particle respectively [cf. Figs in 3.1]. Chapter 3 Lorentz Force on Plasmonic Spher. Heterodim.:Rev.of Bind. Force 51 According to our several full wave simulation results, the important conclusion is that: although reversal of optical binding force occurs for nano rods or other shapes due to Fano resonance [22, 23], Fano resonance is in general not the reason of the reversal of optical binding force. 3.3.2 Perpendicular Polarization: Reversal of lateral binding force for Au-Ag, Au-Au and Ag-Ag on-axis heterodimers Fig.3.2: Considering perpendicular polarized light for the configuration of Fig.3.1 (a) and ‘ȹ’ = 0 degree [on axis Ag-Au]: (a) Extinction co-efficient. (b) The binding force FBind (x) = (FB (x)-FS (x)). (c) Difference of bulk Lorentz force. (d) Difference of surface Chapter 3 Lorentz Force on Plasmonic Spher. Heterodim.:Rev.of Bind. Force 52 force. Considering same polarization of light for the configuration of Fig.3.1 (b) and ‘ȹ’ = 0 degree [on axis Au-Au]: (e) Extinction co-efficient (f) The binding force FBind (x) = (FB (x)-FS (x)) (g) Difference of Lorentz bulk force for x components. (h) Difference of Lorentz surface force for x components. We consider two on-axis Ag-Au, Au-Au and Ag-Ag particles of 100 and 50 nm with inter particle distance (edge to edge distance) of 20 nm [cf. Figs. 3.1 (a), (b) and (c)] and perpendicular polarized light. (1) The difference of the bulk part of the total Lorentz force on a plasmonic object [also cf. Eq (3.4)] should describe the relative bulk force experienced by the optical molecule: BulkmBulk ()()Bulk ( )( )FF[]Del Del []BBSSdvdv ff (3.5) Here; (B) and (S) represent: bigger and smaller object respectively. At the same time the difference of the surface part [which originates from induced surface charges; cf. Eq (3.3)] of the total Lorentz force on a plasmonic object should describe the relative surface force experienced by the optical molecule: SurfeSurface ()()Surface ( )( )[]Del []FDel FBBSSdsds ff (3.6) It should be noted that: ( )( )( )Bulk( )Surf) ( x Del DFFel xBindxBxSxFFF. Subscript (x) represents the force in +x direction. It is observed that near the bonding resonance mode reversal of the optical binding force (negative to positive) occurs at wavelength 646 nm [cf. Fig. 3.2(a), (b) for Ag-Au and Fig. 3.2 (e), (f) for Au-Au]. Reversal of optical binding force occurs at that specific wavelength mainly due to the individual zero surface (Surf (x)Del F0) and bulk (Chapter 3 Lorentz Force on Plasmonic Spher. Heterodim.:Rev.of Bind. Force 53 Bulk( )FD l 0ex ) Lorentz force [cf. Fig. 3.2 (c), (d) for Ag-Au and Fig. 3.2(g), (h) for Au-Au]. (2) We now consider a different idea: the electric dipole moment p of the objects to explain the reversal of binding force based on same and opposite electric charges. This idea of electric dipole moment p should be a more generalized idea than the electric polarizability (discussed in ref. [38]), as the overall size of the dimer set-up is higher than the dipolar limit in this work. The real part of electric dipole moment of an object is defined as: 0inRe [()]spidv E (3.7) It is also demonstrated that whenever the 2nd reversal (positive to negative) of the lateral binding force occurs for transverse/perpendicular polarization near the wavelength 500 nm [cf. Fig. 3.2(a), (b) for Ag-Au and Fig. 3.2 (e), (f)], the real part of the induced electric dipole moment reverses its sign near the resonance of the smaller object in Fig. 3.3 (a) and (c).This does not occur for parallel polarized case as shown in Fig. 3.3(b) and (d). So, the reversal of lateral binding force near this specific wavelength can better be explained based on the idea of induced same or opposite electric charges similar to the idea (reversal of the electric polarizability near resonance) proposed in ref. [38]. Results of on-axis Ag-Ag heterodimers are similar to Ag-Au and Au-Au cases; which have been observed (but not shown here) in our full wave simulations. Chapter 3 Lorentz Force on Plasmonic Spher. Heterodim.:Rev.of Bind. Force 54 Fig. 3.3: For both polarizations, electric dipole moment of Ag-Au heterodimers [for configuration of Fig. 3.1 (a)] and Au-Au heterodimers [for configuration of Fig. 3.1 (b)]. 3.4 Longitudinal binding force for Off-Axis Plasmonic Heterodimers: end-fire and near end-fire configuration In this section we mainly focus on the Ag-Au heterodimers to explain the behaviour of the off-axis heterodimers. Reversal of the optical longitudinal binding force has been observed for only Ag-Au and Au-Au off-axis heterodimers (only when the smaller object perturbs the propagating light at first) and this issue is connected with the presence of constructive interference of dipole-quadrupole mode. Another notable point is that: mutual attraction and repulsion of all the off-axis heterodimers can be very easily controlled by changing the direction of propagating light or by changing Chapter 3 Lorentz Force on Plasmonic Spher. Heterodim.:Rev.of Bind. Force 55 the relative orientation of the particles. All the conclusions of the forthcoming sections have been noted very shortly in Table-3.1. 3.4.1 Au-Ag off-axis heterodimers: Longitudinal binding force for both polarizations Now, we consider Au-Ag particles of 100 and 50 nm respectively with inter particle distance of 20 nm [cf. Fig.3.1 (a) and (d)] but considering rotation angle, φ, of the particle is between 60 to 120 degrees [i.e. end fire or nearly end fire configuration [38]]. The light source is same. We start to create angular displacement from the x- axis considering two cases: (A) Rotating the smaller object keeping the bigger one fixed [cf. Fig. 3.1(a)] and (B) Rotating the bigger object keeping the smaller one fixed[cf. Fig. 3.1(d)]. Now the question arises: ‘Should there be any difference on longitudinal optical binding force for these two cases- (A) and (B)?’. The notable observation of this work: the behavior of longitudinal binding forces are quite different for these two cases due to the forced breaking of symmetry by placing the light source at one side of the dimer configuration instead of at the top of the set-up. If the light source were placed at the top of the set-up, such difference should not arise. According to our forthcoming observations, forced symmetry breaking is detected as one of the key ways to control the inter-particle attraction and repulsion. Some previous symmetry broken set-ups have been discussed in [140, 141] (but not for optical force), which are different than our case. However, for both cases- (A) and (B), the extinction cross sections reveal that bonding mode resonance blue shifts for increasing angular displacement in Fig. 3. 4(a), (c), (e), (g) for aforementioned both the cases. We have observed that this is also true for Au-Au heterodimers. It appears that an ‘angular ruler’ may also be possible similar to previously defined ‘inter-particle gap ruler’ in ref. [142]. Chapter 3 Lorentz Force on Plasmonic Spher. Heterodim.:Rev.of Bind. Force 56 For the off-axis heterodimers, the attractive force can be defined as the positive value of the optical binding force FBind (y) (SR)= (FS (y)-FB (y)) and FBind (y) (BR)= (FB (y)-FS (y)) [here SR means small rotating and BR means big rotating], considering two important facts: (a) the angular displacement angles should be much higher and φ should be as close as 90 degree [i.e. 60< φ <120] and (b) x directed lateral force F(x) is at least ten times smaller than y directed force F(y) (which is usually satisfied, as the y directed/longitudinal scattering force is usually much higher than the x directed lateral force for plasmonic spherical heterodimers). It should also be noted that the scattering force of the bigger object is always pushing force [negative value of FB (y)], which is one of the key issues to control the reversal of the y directed binding force (this will be explained next). 3.4.1.1 Au-Ag off-axis heterodimers: Rotating the smaller particle keeping the big one fixed At first we consider the rotation of the smaller object [case (A); cf. Fig. 3.1(a)] for both perpendicular and parallel polarizations of light. For φ =60 to 90 degrees, it is observed that only the scattering force of the smaller object experiences the reversal at bonding resonance region. On the other hand, scattering force of the bigger object (FB (y)) is always pushing force. But the most important fact is that( ) ( ) SyByFF; always. As a result, FBind (y) (SR) = (FS (y)-FB (y)) is always positive [attractive force as shown in Fig. 3.4(b) and (f)]. Importantly, the real part of electric dipole moment of the smaller object reverses its sign near the bonding resonance [not shown] but FBind (y) (SR) always remains attractive with no reversal of sign. In fact, the difference of the particle radius of both the particles plays a vital role. When one of the particles in the heterodimer is much larger than the other one and the propagating light is perturbed by the bigger object at first, the scattered field from the large particle becomes much larger compared to the incident field. When the field enhancement is quite high at the inter particle gap position, this enhanced field forces the dipole on Chapter 3 Lorentz Force on Plasmonic Spher. Heterodim.:Rev.of Bind. Force 57 the small particle to oscillate in phase. Accordingly, with larger radius of bigger particle, the optical force between the particles becomes always attractive. Fig.3.4: SR and BR represent ‘small rotate’ and ‘big rotate’ respectively and ‘ȹ’ = 60, 70 and 90 degree [off axis Ag-Au]. Considering perpendicular polarized light- for the configuration of Fig.3.1 (a): (a) Extinction co-efficient (SR) (b) The binding force FBind (y) (SR); and for the configuration of Fig.3.1 (d): (c) Extinction co-efficient (BR) (d) the binding force FBind (y) (BR). Considering parallel polarized light: for the configuration of Fig.3.1 (a): (e) Extinction co-efficient (SR) (f) The binding force FBind (y) (SR); and for the configuration of Fig.3.1 (d): (g) Extinction co-efficient (BR) (h) The binding force FBind (y) (BR). Chapter 3 Lorentz Force on Plasmonic Spher. Heterodim.:Rev.of Bind. Force 58 3.4.1.2 Au-Ag off-axis heterodimers: Rotating the bigger particle keeping the small one fixed Now, we shall consider the alternate orientation [case (B); cf. Fig. 3.1 (d)] by rotating the bigger object and keeping the smaller object fixed. If the bigger object is rotated remaining the smaller one fixed, FBind(y) (BR) reverses during the antibonding type resonance mode and near spectral dip position. This is explained next. When the smaller object was rotated and the propagating light was perturbed by the bigger object at first, the scattering force on the bigger object (always pushing) was always higher than the smaller one. But when the bigger object is rotating and the propagating light is perturbed by the smaller object at first, there are some chances to find some wavelength regions when the longitudinal/scattering force on the smaller object becomes higher than the bigger object. In this way the binding force FBind (y) (BR)= (FB (y)-FS (y)) can be observed attractive in those wavelength regions. This is what exactly happens during the anti-bonding type resonance modes as shown in Fig. 3.3 (c), (d) and (g), (h); which is quite different than the conventional idea of optical binding force with homodimers [113]. For homodimers, according to the quasi-static approximation limit [113]: the bonding modes and antibonding modes have positive and negative definite slopes, respectively. As a result, consequently they must, respectively, induce attraction and repulsion. But we clearly observe the opposite scenario for the longitudinal binding force of heterodimer set (at a fixed edge to edge distance of 20 nm) when the light is perturbed by the smaller object at first. Then the question rises why this kind of opposite behavior is observed for such symmetry broken heterodimer sets. Its answer lies in the electrodynamic calculations and force distribution analysis instead of the quasi-static analysis; mainly due to the generation of multipoles. Based on the results demonstrated in Fig. 3.5 (a) - (h) we shall discuss the detail dynamics considering a specific case: φ =60 degree. Chapter 3 Lorentz Force on Plasmonic Spher. Heterodim.:Rev.of Bind. Force 59 In Fig. 3.5 (d) and (h) we have plotted the difference of the bulk Lorentz force, which clearly suggests that the total binding force is dominated by the bulk part of Lorentz force [which is in contrast with the commonly observed dominance of surface [114]/ polarization charge induced force [109] for plasmonic objects]. This force can be considered as the scattering force part [114, 143] of the total force, which is physically originating from the multiple scattering between the smaller and the bigger object. Fig. 3.5 (c) and (g) suggest that during the anti-bonding resonance mode, the directive forward scattering of the bigger object is much higher than the smaller object. Surface charge plots in Fig. 3.6 suggest that for the parallel polarized illumination, during the wavelength spectrum around 350 nm to 470 nm, constructive interference occurs due to dipole quadrupole resonance. Though this is not super-scattering [40], it is recognized that the forward scattering of the bigger object increases during this spectra [cf. the extinction spectra in Fig. 3.4(g) where the magnitude of extinction co-efficient increases for higher rotation angles during this specific spectrum regime]. On the other hand, exactly opposite scenario takes place for the bonding mode resonance. For example- at higher wavelength regime, during bonding mode resonance, the smaller object even experiences optical pulling force [cf. Fig. 3.5(c) and (g)] because of: (i) very strong effective forward scattering along with (ii) more reflected light from the bigger object. It is observed that the reversal wavelength of the optical binding type force FBind (y) (BR)= (FB (y)-FS (y)) remains almost fixed along with the spectral dip position [cf. Fig. 3.4(c), (d) and (g), (h)], though the bonding mode resonance blue shifts gradually with the rotation of the bigger object. Moreover, very similar to ref [22], the reversal of the phase of the steady state current takes place near the spectral dip in our heterodimer set-ups as shown in Fig. 3.6 (m)-(r), though it is constructive dipole quadrupole resonance instead of destructive one. Chapter 3 Lorentz Force on Plasmonic Spher. Heterodim.:Rev.of Bind. Force 60 Fig. 3.5: SR and BR represent ‘small rotate’ and ‘big rotate’ respectively. For off axis Ag-Au and BR [the configuration of Fig.3.1 (d) and ‘ȹ’ = 60 degree]: Considering perpendicular polarized light (a) Extinction co-efficient (b) The binding force FBind (y) (BR) (c) Time averaged force on each particle. (d) Difference of bulk Lorentz force. Considering parallel polarized light for same configuration (e) Extinction co-efficient (f) The binding force FBind (y) (BR) (g) Time averaged force on each particle. (h) Difference of bulk Lorentz force for y components. Chapter 3 Lorentz Force on Plasmonic Spher. Heterodim.:Rev.of Bind. Force 61 Fig. 3.6: For off axis Ag-Au and by rotating the big particle [the configuration of Fig.3.1 (d) and ‘ȹ’ = 60 degree]: Considering parallel polarized light, from left first two columns represent surface charges [(a)-(l)] and the third column represents steady state current [(m)-(r)]. We have chosen six wavelengths for six different rows (from top to bottom): 338, 354, 400, 457, 485 and 612 nm. Charge distributions: (a) QQ (b) DQ (c) DQ (d) DQ (e) QQ (f) DD; where Q and D mean quadrupole and dipole respectively. Chapter 3 Lorentz Force on Plasmonic Spher. Heterodim.:Rev.of Bind. Force 62 3.5 Simplest procedure to reverse the longitudinal binding force for all the off-axis heterodimers So far we have shown that: when the bigger object is rotated and the propagating light is perturbed by the smaller object at first, for only Ag-Au and Ag-Ag off-axis heterodimers reversal of longitudinal binding force occurs [i.e. the dynamics of Ag-Ag and Au-Au heterodimers has been shown in Fig. 3.7 and 3.8 respectively]. Especially after the anti-bonding resonance mode, the longitudinal binding force is observed always repulsive for such heterodimers. On the other hand, for Au-Au heterodimers this force is always repulsive for such configuration for the whole visible wavelength spectrum. In contrast, when the smaller object is rotated and the propagating light is perturbed by the bigger object at first, for all the spherical heterodimers no reversal of longitudinal binding force occurs. Binding force is always attractive for such configuration. So, (1) if we consider the higher wavelength regions and change the direction of propagating light manually by bringing the light source from one side of the dimers to another side, it will be easily possible to observe the mutual repulsion and attraction of all the heterodimer sets just due to the automatic change of the relative dimer position of smaller and bigger objects. Or (2) simply by changing the relative orientation of the heterodimers manually (not light propagation direction), it is also possible to observe such reversal at higher wavelength regions. Such simplest controls of force reversal due to ‘forced broken symmetry’ are certainly impossible with the plasmonic homo-dimers. Chapter 3 Lorentz Force on Plasmonic Spher. Heterodim.:Rev.of Bind. Force 63 Fig 3.7: Reversal of longitudinal binding force for off-axis Ag-Ag heterodimers [cf. Fig. 3.1 (f)] for both polarizations. Fig. 3.8: (a) and (c): extinction coefficient of off-axis [only ȹ =90 degree case shown here] Au-Au hetero-dimers [cf. the configuration of Fig. 3.1(e)]. (b) and (d): Longitudinal binding force on off-axis [only ȹ =90 degree case shown here] Au-Au hetero-dimers. Chapter 3 Lorentz Force on Plasmonic Spher. Heterodim.:Rev.of Bind. Force 64 3.6 CONCLUSIONS In order to identify the conclusions of this chapter at a glance, we have listed our key observations in Table-3.1. It is expected that binding force reversal should occur almost for all the spherical heterodimer structures due to the presence of bonding and anti-bonding mode in the visible spectra. But our work suggests that reversal of lateral (for on-axis heterodimers) and longitudinal (for off-axis heterodimers) binding force of symmetry broken heterodimers follow fully different key parameters/mechanisms; i.e. later one depends on relative orientation and constructive dipole-quadrupole resonance but the former one on light polarization and the induced electric resonance. Most importantly, the reversal of longitudinal binding force can be easily controlled due to forced symmetry breaking just by changing the direction of wave propagation for a specific set-up of off-axis heterodimers or by changing their relative orientation. In addition, though it is commonly believed that plasmonic forces mostly arise from the surface force and Fano resonance can be a promising way to achieve binding force reversal, our study based on Lorentz force dynamics suggests notably opposite proposals for the case of plasmonic spherical heterodimers. Notably, the longitudinal binding force of spherical heterodimers originates almost fully from the difference of the bulk part of Lorentz force, which strongly suggests the connection of bulk Lorentz force with the scattering force. As much study has not been done on the connection of Lorentz force with gradient and scattering forces, this work may open a new window for such investigations. Chapter 4 Problem with Lorentz Force: Interfacial Tractor Beam 65 Chapter 4 Problem with Lorentz Force due to Material Background: Interfacial Tractor Beam 4.0 Summary of Chapter 4 In previous two chapters applications of Lorentz force have been demonstrated considering dielectric and plasmonic particles placed in air, where the well-known Lorentz force has led to accurate and consistent result. In this chapter, we consider a particle sub-merged into an inhomogeneous background composed of two different dielectric materials instead of air. Interestingly, the well-known Lorentz force leads to time averaged pushing force instead of the experimentally observed pulling force for the half immersed objects. As a result, two alternative approaches of force calculation other than commonly used external Minkowski stress tensor method have been implemented. For example- Ray tracing method is adopted to capture the direct process of momentum transfer from the complex background media, which is validated by the proposed modified Einstein-Laub method only employing interior fields of the particle in calculation. Our suggested interpretation supports the Minkowski approach only for the purpose of optical momentum transfer to the embedded scatterer rejecting Peierls’ and Abraham’s approach, although the momentum of photon in the continuous background medium should be considered as the type of Abraham for the calculation of the bulk part of Lorentz volumetric force distribution. Chapter 4 Problem with Lorentz Force: Interfacial Tractor Beam 66 4.1 Introduction Following the pioneering work of Marston in acoustics [144], optical “tractor beams” have attracted much interest due to their unusual mechanism in micromanipulation [16,25,30-33,36,59,62,90,145,146]. Generally speaking, a tractor beam is a customized light beam which exerts a negative scattering force (NSF) to a scatterer and pulling it opposite to the propagation direction of light, in contrast with to the conventional pushing forces [147]. Optical pulling force opens up a new way for gradientless optical manipulation technique distinct from optical tweezers [1,2,9], optical conveyor [33, 60,148], or nano-opto-mechanical systems [76, 149]. Recently, different types of tractor beam have been experimentally demonstrated using a Gaussian beam with an optical mirror (interfering incident and reflecting lights in certain limited regions) [36] and with dodecane droplets sitting on the dielectric interface [35]. However, in the presence of high power laser, the hydrodynamic effects (uneven heat dissipation, particle absorption, temperature gradient, liquid convection, surface energy well, etc.) may contribute too. Moreover, the stability criteria of the tractor beams have also not been investigated so far, which is very important for practical application purposes. Though the mechanical effect has been demonstrated [35] as an overall consequence of all possible contributing factors, the mechanism of optical momentum transfer from a mixed bi-medium background (air and water) to a partially sub-merged particle is still unclear, particularly in the issue of what proper stress tensor and volumetric force law (i.e. Lorentz force) should be adopted for non-vacuum backgrounds. This motivates us to investigate the fundamental physics governing the momentum transfer from light to particle within an inhomogeneous (mixed) background matter, which leads to stimulating debates of Abraham-Minkowski controversy [27, 49,50, 58, 150-153]. In general, both Minkowskian [27, 50,153] and Chapter 4 Problem with Lorentz Force: Interfacial Tractor Beam 67 non-Minkowskian [49, 90] formulations have been supported for the purpose of measuring the optical momentum transfer from homogenous background to an embedded particle. After a century of debate, it is not yet known hitherto which stress tensor (ST), volumetric force law (i.e. Lorentz force) and photon momentum scheme in matter should be the most appropriate one (e.g., Table 1 given in Ref. [27] and Table-1.1 in the introduction chapter of this thesis). It appears that since the identification of adequate 'particle momenta' is quite ambiguous and challenging [154], a recent resolution of the Abraham-Minkowski dilemma [58] still cannot address adequately the appropriate ST and volumetric Lorentz force law to be employed in complex (mixed) backgrounds [154, 155]. In this context, it is valuable to pursue a simplified optical configuration without those limitations or implications discussed above, in which we can discuss the mechanical light-matter interaction isolate. For this purpose, we investigate the light momentum transfer and related optical force on a transparent scatter floating on a dielectric liquid-gas interface [33-35], as shown in Fig. 4.1(a). Although Webb et al. [90, 155] have supported Abraham photon momentum (APM) as an appropriate one for the purpose of optical momentum transfer from homogenous background to an embedded particle, we have shown that optical momentum transfer for our case is more appropriate to consider as the transfer of Minkowski photon momentum (MPM). Interestingly, non-Minkowskian formulations such as external Chu stress tensor and the well-known Lorentz force show pushing force, which contradicts the experimental observation in [35]. As a result, two alternative approaches of force calculation other than external Minkowski stress tensor method have been implemented in this chapter. Detailed calculations by ray tracing method and modified Einstein-Laub equations (using the interior fields only for the latter one) show that negative pulling force and optical tractor beam is natural in our scheme. Chapter 4 Problem with Lorentz Force: Interfacial Tractor Beam 68 However, our suggested interpretation supports the Minkowski approach only for the purpose of optical momentum transfer (or emission) to (from) the embedded scatterer (near the scatterer and the dielectric interface) rejecting Peierls’ and Abraham’s approach although the momentum of photon in the continuous background medium should be considered as the type of Abraham for the calculation of the bulk part of Lorentz volumetric force distribution. Since this scheme can be extended to any gas-liquid interface, we name our tractor beam “interfacial tractor beam” (ITB). The importance of this scheme is not only for that the tractor beam per se becomes practically realizable but also for that there is a clear-cut insight specially based on photon momentum for practical implementation [43]. Fig. 4.1: (a) Schematic of photon momentum transfer in an inhomogeneous mixer background. 1 ,1 ,1irtA and 2 ,2 ,2irtA denote the propagation direction of the incident, reflection, and transmission at point 1 and 2, respectively [43] (lengths of the arrows indicating mn , the refractive index of the associated medium). 1,2 describe the angles of incident and transmission rays with x axis. (b) Ray tracing patterns of the system with an 120 with 0.75yxrr (rx and ry are the semi axes of the scatter along the x and y directions, respectively), 11n , 21.33n , and 31.45n . The black (solid) and red (dashed) lines show the incident and refractions rays, Chapter 4 Problem with Lorentz Force: Interfacial Tractor Beam 69 respectively. (c) The same as (b) except for 120 . Figures adapted and reproduced with permission from ref. [43], © 2015, Nature publishing group. 4.2 Ray Tracing Method and Minkowski Stress Tensor by Employing Background Fields The proposed background mixture is illustrated in Fig. 4.1(a). The scatterer (with refractive index of 3n ) is suspended at the interface of a liquid (e.g., water, with a refractive index of 2n ) and a gas (e.g., air, 1n). The incident and the scattered beams may lie in different mediums (also shown in the ray tracing patterns of Fig. 4.1(b) & (c)). Therefore, not only the direction but also the amplitude of the momentum of the light is changed. Without loss of generality, we restrict the calculation mostly in two dimensional structures for clarity. In the ray tracing method [43], the momentum of a photon in each medium should be clearly defined first of all. Generally, the momentums 1,2p carried by a photon in medium-1 and -2 respectively are 11012202,ppppcc (4.1) Here , and c are reduced Planck constant, angular frequency, and light velocity in vacuum, respectively. 0pc is the momentum of the photon in vacuum. For the material related constants 1,2, however, different values are proposed and there is also a long standing controversy on this point [24, 27, 29], such as n for Chapter 4 Problem with Lorentz Force: Interfacial Tractor Beam 70 Minkowski formula [27, 50,152], 1n for Abraham formula [27, 50,152], and 244710nnnn for Peierls formula [158, 159]. In order to calculate the momentum transfer and corresponding forces quantitatively, we set the mediums of gas, liquid and scatterer to be air, water and silicon sphere with 11.0n , 21.33n , and 31.45n , respectively. Using the ray tracing method [43], all the rays illuminated on the scatter are traced (as illustrated in Fig. 4.1(b) and (c)), and momentum changes and in turn the optical force are calculated correspondingly. Fig. 4.2(a) shows that the Minkowski and Abraham momentums are the upper and lower boundaries, respectively; while the Peierls momentum is about the average value of them. Using the above momentum formulations (Minkowski, Abraham and Pierels), the forces of xF can be calculated by using the ray tracing method, and the results are shown in Fig. 4.2(b). It is seen that only the Minkowski momentum results in a negative pulling force in a wide range of incident angles while those of Abraham and Peierls momentums result in positive pushing forces regardless of the incident angles. Fig. 4.2: (a) Changes of momentum per photon in materials with a refractive index n for Minkowski (blue), Abraham (red) and Peierls (black) formulations, respectively. (b) Optical forces (s-polarization) calculated by ray tracing simulations when Minkowski (blue), Abraham (red), and Peierls (black) momentums are used. Figures adapted and reproduced with permission from ref. [43], © 2015, Nature publishing group. Chapter 4 Problem with Lorentz Force: Interfacial Tractor Beam 71 Fig. 4.3(a) shows the changes of xF with the shape of the scatter at the case of 10. When 0.425/0.775yxrr, both p and s polarizations can produce negative force. Fig. 4.3(d) shows the force changes with incident angle on a circular scatter with xyrr, and NSF are achieved within a broad incident direction. Fig. 4.3: Optical forces derived from different formulas. The parameters are 11.0n , 21.33n , 31.45n , and 0 , unless otherwise specified. (a) Optical forces change with the shape /yxrr at 10 . The forces are calculated using the ray tracing method, in which the Minkowski momentum approach of photon is adopted. (b) Optical forces changing with incident direction 1 for the p polarization calculated by the internal modified Einstein-Laub formula (MEL, circle line) proposed in this chapter, volume integration of Lorentz force density (Lorentz, triangle line), external Chu ST (Chu, asterisk line), as well as external Minkowski ST (Minkowski, square line). The size of the scatter is 2.0xr μm and 0.4yr μm (as shown in (c)). (c) Field pattern for the Hz at the case of 130 . (d) Optical forces changing with the incident direction 1 for a circular scatter of 2.0xyrrμm. The forces are calculated by ray tracing method. (e, f) The same as (b, c) respectively, except force 2.0yr μm. Figures adapted and reproduced with permission from ref. [43], © 2015, Nature publishing group. Chapter 4 Problem with Lorentz Force: Interfacial Tractor Beam 72 These results agree with the analysis above. In order to calculate the force using ST’s, we calculate the scattering using the finite difference in time domain (FDTD) method with a plane wave incidence for a circular scatter (with the semi axis along x direction of rx, and y direction of ry of 2.0xyrr μm) and an ellipse scatter (with 2xr μm and 0.4yr μm). Then, the optical force experienced by the particles are calculated by the integration of time averaged external Minkowski stress tensor on a closed contour [see the dashed rectangles in Fig. 4.3© and 4.3(f)] surrounding the scatterer, ****1 ()](41 Re[2.2)2ITDEBHEDHB The calculated forces for the elliptical and circular scatters are shown in Fig. 4.3(b) and 4.3(e) (square lines), respectively, and good agreements with ray tracing method are observed. Here, s-polarization results are not shown, since they are similar to those of p-polarization case. In the calculation, local permittivity and permeability are used along the integration paths (as shown in Fig. 4.3(c) and (f)). Fig. 4.3(c) and (f) show the field pattern of zH scatterer at the case of 130 for the circular and elliptical scatterers, respectively. The arrows show the element force vector on the closed contour, along which the total net force xF is calculated by integration. One can find that most part of the pulling force is exerted when the photons left the scatterer, which agrees with our analysis. 4.3 Explanation of the Observations and the Two Photon Momenta Chapter 4 Problem with Lorentz Force: Interfacial Tractor Beam 73 It is important to note that a previous analysis and the experimental observation in [35] cannot tell that the travelling momentum of the photon inside the continuous second background (water) is the Minkowski one. Previously for one same experiment (Jones and Leslie experiment [154, 155]), the optical force density formula based approaches have supported the transfer of both APM [155] and MPM [154]. Such ambiguities for the transfer of photon momentum from the background can only be overcome by direct photon momentum approach [27] such as the ray tracing method, which is the main purpose of our previous section. But surprisingly, for this ITB experiment [35], the force felt by the embedding water medium can also be calculated using Einstein-Laub formula employing the field inside water, which indeed supports APM (Cf. Eq (21) given in [155] and also ref. [161]). If the force of a lossless continuous medium is calculated via Minkowski ST (associated with MPM) employing interior fields, it gives zero experienced force. In the ITB experiment, at the interface of the scatterer and the water [35], MPM should arise due to the reduced impedance mismatch [162]. In fact, MPM generates a translation of the electromagnetic field [151]. This translation, relative to the host, is exactly the quantity required to represent the displacement of an embedded object [58]. As a result, MPM appears in almost all the major radiation pressure experiments that measure the displacement of an embedded object [58, 151]. In this discussion, it is important to mention the experimental observations of Ashkin and Dziedzic regarding the force density distribution in water [47], which support the Einstein–Laub force– density expression associated with the APM, according to Mansuripur et al. [161]. So, the connection of APM is related with the force distribution inside water medium, which suggests the role of APM as the pure electromagnetic part of photon momentum in a continuous medium. Therefore, the experimental observation of a pulling force on the scatterer in Ref. [35] indeed cannot prove that the MPM is the only correct photon momentum. In fact, both the MPM and the APM are probably correct, but their functionalities are quite different. Chapter 4 Problem with Lorentz Force: Interfacial Tractor Beam 74 4.4 The Validity of Other Methods Hereto, the transfer of optical momentum has been calculated based on Minkowski photon momentum approach, which is also in good agreement with the experimental observations reported in [35]. Surprisingly, any other photon momentum (Abraham or Peierls) does not predict optical pulling force (see Fig. 4.2(a) and (b)). In addition, force calculation results based on Lorentz vlumetric formula and external Chu ST shown in Fig. 4.3(b) and (e) do not guarantee negative forces, which are calculated by the force density integration of Lorentz force density (from the interior of the embedded particle) and external Chu’s ST (from exterior of the embedded particle) [27] (external Nelson [163] ST leads to the same result). Although external Chu ST employing exterior fields shows optical pulling force for the 2D circle case (Fig. 4.3(e)), it shows pushing force for the 2D ellipse (Fig. 4.3(b)). 4.5 Modified Einstein-Laub Stress Tensor by Employing Interior Fields of the Scatterer Another method to calculate the optical force unambiguously is using the force density inside the scatterer. Surprisingly, the time averaged Lorentz force [44] of the embedded particle also fails to provide the accurate result of total mechanical force, as shown in Fig. 4.3 (b) and (e). It appears that the Lorentz force reported in [44] (actually Chu force for nonmagnetic media [27]) is applicable only when the background media is air. If the background media is not air, Lorentz force [44] for the interior of the embedded particle should be modified (which will be discussed in details in the next chapter of this thesis). According to our point of view, the interior Chapter 4 Problem with Lorentz Force: Interfacial Tractor Beam 75 mechanical force should be calculated via the proposed time averaged modified Einstein-Laub (MEL) stress tensor [164]: MTotalEL1 Re[(in)]2dSFT (4.3) Where 2MEL**2(in)1 ()2inininiininnbbD EETB HHI , b and b are permeability and permittivity of the background medium of a particle, and I is the unit tensor. In our setup, they are the local permeability and permittivity of air and water defined respectively in the specific 2D configuration of full wave simulation [Fig. 4.3(b) and (e)]. The quantities Ein, Hin , Din & Bin are at the interior of the silica particle. The calculations via MEL stress tensor from the interior are fully in agreement with ray tracing method and Minkowski formulations from the exterior (see Fig.4.3(b), (e)). Detail discussion on MEL formulations [164] will be presented in next chapter, chapter 5. 4.6 Conclusions In summary, our suggested interpretation supports the Minkowski approach only for the purpose of optical momentum transfer (or emission) to (from) the embedded scatterer (near the scatterer and the dielectric interface) rejecting Peierls’ and Abraham’s approach, although the momentum of photon in the continuous water background medium should be considered as the type of Abraham for the purpose of calculating the bulk part of Lorentz volumetric force. Interestingly, the well-known Lorentz force leads to time averaged pushing force instead of the experimentally observed pulling force for the half immersed objects. As a result, two alternative approaches of force calculation other than external Minkowski stress tensor method have been implemented. We believe that our proposal is very effective in order to Chapter 4 Problem with Lorentz Force: Interfacial Tractor Beam 76 resolve not only the Abraham-Minkowski photon momentum dilemma but also the dilemma associated stress tensors and optical force including interfacial tractor beams. Chapter 5 Modification of Lorentz Force for embed. chiral and achiral obj. 77 Chapter 5 Modification of Lorentz Force for embedded chiral and achiral objects 5.0 Summary of Chapter 5 Although it is commonly believed that all the Lorentz volumetric force laws lead to the same total optical force for both chiral and achiral objects, this idea has been invalidated in the recent review by M.Bethune-Waddell and K. Chau [cf. Rep. Prog. Phys. 78, 122401 (2015)] discussing several previous experiments. In our previous chapter we have also shown one example of such cases. To identify the exact reason of such significant disagreements, we inspect in this chapter two tractor beam and one lateral force experiments on using distinct stress tensors (STs). In this chapter we have identified the exact reasons of such disagreements. We propose that it is possible to make different time-averaged force formulas equivalent based on the fulfilment of just two ‘consistency conditions’. Finally, we demonstrate that by modifying the Einstein-Laub or Chu formulation, time-averaged STs and volume forces are obtainable those can overcome the aforementioned inconsistency for both chiral and achiral Mie objects embedded in even complex material backgrounds. Our work presented in this chapter also explains why such modification should be adopted for Chu or Einstein-Laub formulations in order to obtain an accurate and reliable time-averaged total force, both in magnitude and direction, on both chiral and achiral objects embedded in a generic material background. Chapter 5 Modification of Lorentz Force for embed. chiral and achiral obj. 78 5.1 Introduction A tractor beam is a customized light beam which exerts negative force to a scatterer [24, 25, 32-36, 43, 59, 62,165] and pulling it opposite to the propagation direction of light, in contrast to the conventional pushing forces. Not only the counter intuitive idea of optical pulling (i.e. tractor beam effect) but also the idea of optical lateral force [166-169] is also growing famous in recent literature. Tractor beam and lateral force experiments, which involve the material background [35, 36, 24, 166], can also be investigated in details to understand the persistently debated roles of different STs, Lorentz/ volumetric forces and photon momenta such as Abraham-Minkowski controversy [27, 28, 57, 58]; as the employment of inappropriate approach may lead to pushing force (or inconsistent lateral force) instead of the experimentally observed pulling one (or consistent lateral force). Throughout this thesis we refer to 'exterior' or 'outside' magnitudes as those evaluated outside the volume of the embedded macroscopic object, while by ‘interior’ or 'intside' we shall refer to those quantities inside this object volume [cf. the force calculation process in Fig.5.1 (a) - (c)]. Although it is commonly believed all the volume force laws [27, 170] (i.e. Minkowski, Abraham, Chu, Einstein-Laub and Ampere/Nelson) should lead to the same time averaged total optical force, this idea has been invalidated in recent works [43, 28] considering several experimental set-ups [28] involving material background. In fact, if one involves the material background; several complexities arise: (i) In order to calculate the total force on an embedded object, normally two conventions, i.e., GAP METHOD AND NO GAP METHOD, are applied [26, 78, 170-173] but which one is more accurate than the other is still not properly understood: (a) GAP METHOD [78,170,172,173]: By introducing a very small gap Chapter 5 Modification of Lorentz Force for embed. chiral and achiral obj. 79 between the scatterer and the background to yield the force as shown in Fig. 5.1(a). (b) NO GAP METHOD [26, 28, 35, 36, 43, 166, 171]: Without introducing any small gap outside the embedded scatterer as shown in Fig. 5.1(b) and (c). Between these two distinct methods, which one is more appropriate? (ii) In addition, the involvement of chirality [19,20,77,174-181] can be a key factor to judge the consistency of the optical force calculation methods which are considered applicable for embedded achiral objects. Previous analyses of optical force in chiral objects are still restricted in these factors: (a) the dipolar limit of object [167,175,176], (b) air or vacuum background [77,174-181] and (c) internal force distribution inside a slab embedded in air [77,181]. But the stress tensor and force distribution inside any arbitrary generic embedded chiral object [i.e. more complex than Mie object which are used in real experiments [19, 20]] are still not investigated in details. This chapter attempts to solve the aforementioned complexities. We demonstrate that even without introducing any artificial gap between the background and the embedded object, it is possible to make different time-averaged force formulas equivalent based on the fulfilment of just two ‘consistency conditions’ [denoted as C(I) and C(II) with details shown below]. This may also solve the reported problems [28, 35, 43] of total force calculation for almost all the real experiments conducted so far involving a material background. In addition, this work introduces the less time and memory consuming internal Modified Einstein-Laub (MEL) stress tensor method, (rather than the longer time consuming bulk volumetric force method [26, 183]) for both achiral [35,36,43,28] and chiral objects [19, 20] embedded in a generic background. Chapter 5 Modification of Lorentz Force for embed. chiral and achiral obj. 80 5.2 Proposal of Consistency Conditions Fig. 5.1: Procedure of time averaged optical force calculation by employing stress tensors. (a) GAP METHOD: a small gap between the scatterer and the background should be considered. Time averaged total force on the scatterer should be calculated using the time-averaged ST GAPoutF evaluated from fields strictly outside the object [i.e. gap field and background field], putting the integration boundary in the gap [170] (black dashed circle). However, the volume force calculation method is a little bit different, which is discussed in details in [78]. (b) and (c) NO GAP METHOD: In both examples the total force obtained by using the time-averaged ST is outF evaluated from fields strictly outside the object considering no gap, at 1.001raa, (black circles); whereas this force is inF or bulk force when the ST is determined from fields strictly inside the object considering no gap at 0.999raa(white circles). In (b), a sphere or cylinder is immersed in an unbounded and heterogeneous background medium. In (c) a core-shell sphere or cylinder (i.e., the core is embedded in a bounded background). One of the fundamental proposals of this chapter is: if inT [applied at r=a- employing only the interior field of a scatterer, where a is the particle radius; cf. Fig. 5.1 (b) and (c)] is a valid internal stress tensor and f (in) is its corresponding volume force; and (out)T is a valid external stress tensor which is applied at r=a+ using only exterior fields of an object; then the following ‘validity condition’ should be fulfilled: “On the boundary r= a of any object, this equation: Surfaceˆ[ (out)in ]TTnf should hold ( ˆn being the local unit outward normal of the object surface). Also, the same surface force Surfacef should independently be found from the volume force Chapter 5 Modification of Lorentz Force for embed. chiral and achiral obj. 81 density f (in) by applying the appropriate boundary conditions at r= a. These two aforementioned conditions must be satisfied simultaneously.” For example- in [182], the process of obtaining Surfacef has been shown only from a volumetric force density stemming from the well-known Chu and Einstein-Laub force, and by considering only air as the background. At the beginning of Appendix A (at the end of this thesis) it is shown that when the background is air, the difference of the external ST (notice that then all STs are same) and the internal ST of Chu (and of Einstein-Laub) at the object boundary is in complete agreement with the fully independently calculated surface force given by the volumetric formulation of Chu and Einstein-Laub. But when material background is involved: the first question should be: Q(I) which (out)T should be appropriate? Then the second question is: Q(II) If BulkBulk ininin0dvdFfTs ; which inT, along with its corresponding f, will satisfy these two following ‘consistency conditions’ simultaneously: C(I) the aforementioned ‘validity condition’ and C(II) the consistent time-averaged force equation: TotalBulkSurface outConsistentind TsF= FF? In this chapter we show that when the background is a material medium instead of air, (out)T and inT cannot be arbitrary STs which satisfy C(I) and C(II) simultaneously. Concerning Q(I), in order to sort out the appropriate ST (out)T, we investigate two major tractor beam experiments [35, 36] along with the recent set-up of a lateral force experiment [166]. We identify that the vacuum ST [170] of the GAP method leads to inconsistent results for tractor beams and lateral force experiments, especially when the symmetry is broken [35,166]. Both external Minkowski and Chapter 5 Modification of Lorentz Force for embed. chiral and achiral obj. 82 Abraham STs of a NO GAP METHOD lead to a consistent time-averaged total force for all those experiments, (and also for all previous experiments [28,46]). However, this does not mean external STs other than Minkowski (or Abraham) are incorrect. As a result, even to sort out appropriate (out)T, it is required to satisfy aforementioned consistency conditions C(I) and C(II) simultaneously. Now, we shall consider Q(II). As always BulkBulk ininin0dvdFfTs for internal Minkowski ST for the transparent/non-absorbing objects, to sort out the appropriate inT [and also (out)Tof Q(I)] we examine several cases for an object embedded in material background [19,20, 28,35,36,46, 84,97, 98, 183,184,185, 186]. For example- condition C(I) has been violated when: (i) (out)T is considered Mink.(out)T and inT is well known internal Chu or Einstein-Laub ST; and (ii) (out)T and inT both are considered ELTrespectively. More details will be discussed later. In contrast, C(II) is seriously violated when: (iii) (out)T and inT both are considered ChuTrespectively. More details will be discussed later. (iv) Already we have discussed very shortly that external vacuum ST of GAP METHOD violates the left hand side equation of C(II). These are identified as the main reasons of disagreements observed in [28] and [46] for different volumetric force laws. Finally we demonstrate that whenBulk in0F; to satisfy C(I) and C(II) simultaneously for a chiral or achiral Mie scatterer embedded in generic material backgrounds [i.e. homogeneous, heterogeneous, bounded or unbounded], (out)Tshould be considered as Minkowski ST and inT should be the modified version of Einstein-Laub (or Chu) formulation along with modified f(in). So, in order to satisfy both C(I) and C(II) simultaneously, (out)T and inT cannot have the same form Chapter 5 Modification of Lorentz Force for embed. chiral and achiral obj. 83 for embedded transparent/non-absorbing scatterers. To explain the last observation for realistic situations [19,20, 28,35,36,46, 84,97, 98, 183- 186], we conclude at the end: though both the external Minkowski (and Abraham) ST and proposed internal MEL (and Modified Chu) methods lead to the same consistent time-averaged total force [i.e. they are mathematically equivalent], they should better be considered two fully different operations/process from the physical point of view. 5.3 The consistent external force with NO GAP METHOD Fig. 5.2: (a) Optical sorting of a dielectric particle using two obliquely incident plane waves reported in [36]. The external force will be calculated based on NO GAP METHOD. Two TE-polarized plane waves ( = 532 nm) incident at varying angle onto a polystyrene cylinder (n = 1.58) immersed in water (n = 1.33), shown: (b) For r = 320 nm, at steady state when the plane waves exert a pushing force in the +x direction. Force on the bead calculated by external Minkowski, Einstein–Laub, Chu, and Ampere/Nelson ST for TE polarization as a function of incident angle . (c) For r = 410 nm, at steady state when the plane waves exert a pulling force in the −x direction. Force on the bead calculated by external Minkowski, Einstein–Laub, Chu, and Ampere/Nelson ST for TE polarization as a function of incident angle . Pulling forces are only achieved for TE polarization within a short range of angles experimentally observed in [36]. In this section we investigate the consistency of the total force calculation for one tractor beam experiment [36] and one lateral force experiment [166] by the major Chapter 5 Modification of Lorentz Force for embed. chiral and achiral obj. 84 stress tensors namely: external Minkowski, Chu, Ampere and Einstein-Laub with NO GAP METHOD [cf. Fig.5.1(b) and (c)]. All the 3D simulations throughout this chapter are conducted using incident power of 0.57mW/µm2. For example- the external time averaged total outside force by commonly applied Minkowski ST is written as [35,36,43,166]: outTotal Minkout,dFTs (5.1a) out****Minkoutoutoutoutoutoutoutout11Re22TDEBHBHDEI. (5.1b) ‘out’ stands for fields outside the object, [e.g., on r=a+, if it is a sphere or cylinder of radius a, cf. Fig. 5.1(b)] and I is the unity tensor. The electromagnetic vectors in Eq (5.1b) correspond to the total field, namely, incident plus field scattered by the body. For the force calculations with other stress tensors (i.e. external Chu, Ampere and Einstein-Laub), we shall also use this ‘total’ outside fields. Without introducing any small gap (named as NO GAP METHOD in the introduction), total force has been calculated by different external stress tensors for two beam tractor configuration reported in [36] in Fig. 5.2(a) - (c) and also for lateral force experiment [166] configuration in Fig. 5.3(a)-(c). Though Minkowski’s (or Abraham’s) formulation leads to the most accurate time averaged force for some other experiments [2,35, 36, 47, 55,56] reported in [28]; for two tractor beam [36] and lateral force experiments [166] it may not be possible to recognize which one is the most consistent external ST. Due to the very big size of the object [166] considered in the real lateral force experiment [special technique has been applied in the first ref. of [166] due to the big size of particle: 4500 nm], we are modeling that same set-up with comparatively small sized object [cf. the second reference in [166], where a small object has been considered]. However, we are considering two cases: spherical and elliptical object to check the consistency of the lateral force considering the arbitrary Chapter 5 Modification of Lorentz Force for embed. chiral and achiral obj. 85 shape of the testing object. Previously for interfacial tractor beam experiment [35,43], only external Minkowski ST lead to the accurate prediction for arbitrary shaped objects [35,43] modelled as spherical and elliptical shaped objects. Though there is no difference among the signs of the total forces for the experiments reported in [36] and [166], their magnitudes are observed quite different in Fig. 5.2(b), (c) [two beam tractor] and in Fig. 5.3 (b), (c) [lateral force]. But the important fact is that: For the NO GAP METHOD: (A) The size based sorting of embedded particles by two beam method [cf. supplement of ref. [36]] is consistently predicted by all external STs. (B) The sign change of the force for two different handedness of polarizations observed for lateral force experiment [166] has also been correctly predicted by all external STs for spherical and elliptical objects respectively. In Fig. 3(a) given in the second ref. of [166], for 1500 nm sized object, a single direction of lateral force has been observed for a single handedness of circularly polarized light (consistent with our observation). These aforementioned observations will be imperative for our next investigation: the consistency of GAP METHOD for the tractor beam experiments [35,36] and the lateral force experiment [166]. Chapter 5 Modification of Lorentz Force for embed. chiral and achiral obj. 86 Fig. 5.3: (a) 3D Illumination geometry for a particle at the interface between two media to model the lateral force set-up in [166]. Either a right-handed or a left-handed circularly polarized (CP) beam of wavelength 1064 nm is incident along the x–z plane at angle 1 . k1 and k2 are wave vectors in media 1 and 2, respectively. The external force has been calculated based on the NO GAP METHOD. (b) Transversal force (negative for a left- handed CP and positive for right-handed CP) as a function of angle of incidence for a 500 nm (radius) spherical TiO2 particle located at the water–air interface calculated by external Minkowski, Einstein–Laub, Chu, and Ampere/Nelson ST. The Minkowski and Einstein-Laub STs predict a smaller lateral force in comparison with the time-averaged force yield by the external Chu and Ampere ST. (c) Transversal force, (negative for left-handed CP and positive for right-handed CP) as a function of the angle of incidence for an elliptical polystyrene (PS) particle [rx=800 nm and ry=rz=(800/3) nm] located at the water–air interface, calculated by external Minkowski, Einstein–Laub, Chu, and Ampere STs. The Minkowski and Einstein-Laub STs predict lower lateral force in comparison with the time-averaged force yielded by the external Chu and Ampere STs. 5.4 Inconsistency of the GAP METHOD and different other formulations Chapter 5 Modification of Lorentz Force for embed. chiral and achiral obj. 87 Fig. 5.4: Illustration of optical force with GAP METHOD. (a) Optical force for a 2D spherical shaped scatterer placed in air-water interface (cf. Fig. 2(d) and Fig. 2(e) given in [43]). The parameters are n1=1.0, n2=1.33, n3=1.45. The variation in the optical forces with the incident angle [43] for the p-polarization case, as calculated via the external vacuum stress tensor considering a small gap of 2 nm [the possible smallest gap with 2D full wave simulation set-up; gap size << incident wavelength] between the scatterer and the water background. The size of the scatterer is defined by rx = ry =2.0 μm [3]. Instead of optical pulling [43], optical pushing is achieved based on GAP METHOD. We have also examined our results by varying the size of the artificial gap, (i.e. 6nm, 10 nm and 20 nm). The results are almost same for all those gaps. (b) Optical force (by vacuum ST) of a 3D dielectric particle given in Fig. 5.2(c) using two obliquely incident plane waves but considering a small gap of 10 nm [possible smallest gap with 3D full wave simulation set-up; gap size << incident wavelength] between water background and the embedded scatterer. Both GAP and NO GAP METHOD lead to consistent result. (c) For the 3D set-up of Fig. 5.3(b), the time averaged lateral force has been calculated by vacuum ST considering a small gap (10nm) between the scatterer and the water background for right hand CP wave. (d) For the 3D set-up of Fig. 5.3 (c), the time averaged lateral force has been calculated by vacuum ST considering a small gap (10 nm) between the scatterer and the water background for left hand CP wave. The sign of the lateral force is in complete disagreement with the time averaged force yield by NO GAP METHOD in Fig. 5.3(c). One convention of calculating optical force is that if the background is material medium; the force should be calculated considering an extremely small gap [cf. Fig. 5.1(a)] between the embedded object and the background as discussed in ref. [78,170,172,173]. The main goal of this GAP METHOD approach is the idea that all Chapter 5 Modification of Lorentz Force for embed. chiral and achiral obj. 88 the optical force formulations should lead to the same time averaged total force [27, 170], as the total time averaged force is always the same for all the volume force formulations (when an object is placed in air). As a result, in this section we shall investigate the consistency of this GAP METHOD for the scatterers embedded in material media such as in [35], [36] and [166]. According to [170], if a small gap is introduced [cf. Fig. 5.1(a)], the external vacuum stress tensor (Chu type stress [170]) to yield the total outside force should be written in terms of gap fields (gE ,gH ) as: outTotal Vacuumout,dFTs (5.2a) out****Vacuum0gg0gg0gg0gg11Re22TE EH HHHEEI. (5.2b) In Fig. 5.3(a) we have calculated the external force considering such an extremely small gap [cf. Fig. 5.1(a)] between the half immersed scatterer and the water background and by employing the stress tensor of Eq (5.2b). Instead of the experimentally observed pulling force [35, 43], we obtain pushing force due to such an extremely small gap as shown in Fig. 5.4(a). In contrast, both GAP METHOD and NO GAP METHOD lead to consistent pulling force for two beam tractor experiment [36] as shown in Fig. 5.4(b).Then we have considered the lateral force experiment reported in [166]. Though for the spherical sized object the total force seems in good agreement (cf. Fig. 5.4(c)) with the previously calculated forces with NO GAP METHOD (cf. Fig. 5.3 (b)), for elliptical shaped object the sign significantly alters (cf. Fig. 5.4(d)) in comparison with NO GAP METHOD (cf. Fig. 5.3(c)). So, the problem of GAP METHOD mainly arises when the background is inhomogeneous/ symmetry broken case [35,43,166] (or in general heterogeneous type as shown in Fig. 5.1(b)) [serious violation of C(II)]. Another detail analysis has been done in favor of GAP METHOD in ref. [78] (and also in refs. [172] and [173]). Instead of the external ST Eq (5.2b), it is suggested in Chapter 5 Modification of Lorentz Force for embed. chiral and achiral obj. 89 [78] to yield the total force by an appropriate volume force for some specific cases (described as method II in [78]). However, in Appendix A (at the end of this thesis) we have discussed in details why method II reported in [78] may not be a general way to yield the total force and to explain the so far reported experiments, especially for the experiments with inhomogeneous (or heterogeneous) background due to the violation of C(I). In fact, even considering no such gap, the well-known Einstein-Laub volume force [28] (but not Chu and Ampere/Nelson force) predicts consistent time averaged total force for almost all previous experiments as shown in details in [28]. However, the magnitude of the total force by EL law is not in full/exact agreement with the total force calculated by external Minkowski ST or Helmholtz force [28,46] for all those experiments. But still we cannot comment regarding the inconsistency of EL force for material background case. One needs to examine C(I) and C(II) at first. This issue is discussed next. At the beginning of Appendix A (at the end of this thesis) it is shown that exactly at the object boundary, the difference of well-established external Minkowski ST and internal Chu ST (also similarly applicable for Einstein-Laub ST) is found in agreement with the surface force of Chu (or Einstein-Laub) when the background is air. But if the background is any material medium instead of air or vacuum, this conclusion does not remain true. For example- condition C(I) has been violated when: (i) (out)T is considered Mink.(out)T and inTis well known internal Chu or Einstein-Laub ST; and (ii) (out)T and inT both are considered ELT respectively. It should also be noted that EL force leads to inconsistent result for few experiments [violation of C(II)]: Hakim-Higham experiment [46], Rasetti experiment [50], few cases of Jones’ experiments according to [28] [cf. Fig. 9 (a)-(d) given in [28]] and some other notable cases reported in [188]. Chapter 5 Modification of Lorentz Force for embed. chiral and achiral obj. 90 In contrast, C(II) is seriously violated when: (iii) (out)T and inT both are considered ChuT respectively. For example- when TotalBulkSurface inFFF is calculated by employing internal volumetric force of Chu [or by Chu ST: Bulk Chuin(in)dFTs] for the several real experiments reported in [28], it does not lead to the correct time averaged total force, Total ConsistentF [Violation of right hand side of the equation given in C(II)]. In addition, in [43] it is shown that left hand side of equation of C(II) has been violated for Chu(out)Tbased on NO GAP METHOD. These are identified as the main reasons of disagreements observed in [28] (and also in [46]) for different volumetric forces. In the next section we shall demonstrate that not only the surface force but also the bulk force of the well-known Einstein-Laub (or Chu) force law is responsible for such disagreements those reported in [28] and [46] for the real experiments. It would be possible to overcome such inconsistencies if and only if the GAP METHOD were applicable where the external Minkowski ST turns into Vacuum ST [cf. Eq (5.2b)]. But in this section we have already demonstrated the problem/inconsistency of the GAP METHOD. Hence the possible solution of such problems will be addressed in the next three sections. 5.5 Consistency of the external Minkowski and internal MEL or modified Chu formulations Though so far Minkowski's ST has led to consistent time-averaged result for real experiments [28,46] by employing the exterior field of an embedded scatterer Chapter 5 Modification of Lorentz Force for embed. chiral and achiral obj. 91 [fulfilment of the left side of the equation in C(II)], more complex configurations may arise where Minkowski's ST and the Helmholtz force may fail [i.e. cf. refs. [189-191]]. As a result, we shall now examine whether external Minkowski leads to the fulfilment of both C(I) and C(II) together with any other internal ST or not. To handle future situations, and to sort out new optical force laws, if they proceed, the validity conditions C(I) and C(II) together and the analysis to follow below should be very effective. A modification on the well-known EL volume force (and also Chu force [26]) has been predicted in [171]. Interestingly that modified version of the EL force led to exactly the same time-averaged force predicted by the external Minkowski ST in [171] for a dielectric object embedded in another dielectric. Commercial software [183] applies this modified version of the Einstein-Laub [171] or the Chu volume force density [26] to yield the total volumetric force on dielectric objects embedded in another dielectric. In order to explain the interfacial tractor beam experiment [35], the achiral internal MEL ST [43] has been applied previously without any derivation and consistent description. So far no explanation is given in [43] and [171] why both Minkowski ST and the MEL volume force [171] (and also modified Chu force [26,183]) and the MEL ST [43] lead to same/consistent time averaged result for dielectric cases. This section (and the final conclusion of this chapter) explains the reason, even considering more general cases such as magnetodielectric objects embedded in a generic magnetodielectric background. Achiral MEL ST (which should yield the bulk force of an embedded scatterer) inside an object embedded in a generic heterogeneous background should be written as: b ( )( )Bulk****MEL( )ininininsinininins1in()().2jb jjssTD EB HHHEEI (5.3) Chapter 5 Modification of Lorentz Force for embed. chiral and achiral obj. 92 Here j=1,2,3,...., N represents the number of background regions sharing interface with the object, [cf. our Fig. 5.1(b), (c) and forthcoming Fig. 5.7(a)]. b and b are fixed background permittivity and permeability and s and s are fixed permittivity and permeability of the scatterer. A possible derivation of Eq (5.3) is given in Appendix B where we have shown that the achiral MEL ST should be written in the time averaged form along with its time averaged volume force density (i.e. MEL force law for magnetodielectric object embedded in a generic magnetodielectric heterogeneous background): Bulk****MEL( )Eff( )inEff( )inEff( )inEff( )in1inRe.2jjjjjiifPEMHPBMD (5.4) Consistency of Eq (5.3) and (5.4) has been shown for surface force calculation in Appendix C, which satisfies the ‘validity condition’ described in the introduction [fulfilment of C(I)]. In Eq (5.4), the effective polarization and magnetization are defined as [26,171]: EffbinSPE and EffbinSMH . The total time averaged Bulk force on the embedded object should be: BulkBulkBulkMEL MEL( )( )MEL( )( )inininjjjjjjddvF=Tsf. (5.5) Consistency of Eq (5.5) can be determined analytically considering the internal force on a magnetodielectric slab embedded in another magnetodielectric. Both the internal ST and volumetric force lead to the same result [164]. So, internal MEL method and external Minkowski method fulfil the condition C(II) analytically for this simple case. However, internal Chu, EL and Ampere method do not lead to that same force.In addition, the external force calculated by Minkowi’s force also lead to the exactly same result [164]. Chapter 5 Modification of Lorentz Force for embed. chiral and achiral obj. 93 Now, we consider the consistency condition C(I). By applying the proper boundary conditions [cf. Appendix C], the surface force of Modified Einstein-Laub method can be written from two different ways: (i) By the volume force method of Eq (5.4) [as shown in [182] only for volumetric force] and (ii) from the difference of external Minkowski ST and internal MEL ST in Eq (5.3) just at the boundary. These two different ways lead to exactly same result: outinoutinouSurfaceoutMELMinkMEL( )( )( )atattinoutinˆ[(out)in ]ˆˆ ()()22jr ab jb jr ar afTTnEEHHEEnHHn (5.6) Eq (5.6) clearly explains why in [171], the time averaged result of total force predicted by Minkowski ST is in exact agreement with the MEL volume force reported in [171]. According to Eq (5.6), the total time averaged force calculation by MEL force is equivalent with the total force calculation based on Minkowski ST or Helmholtz force for any generic case. So, the total time averaged force on an embedded achiral object according to MEL method should finally be written as: outBulkSurfaceMinkMEL MEL ind Ts = FF (5.7) Hence at least analytically we have arrived at the fulfillment of both C(I) and C(II). In the next part of this work, the validity of Eq (5.7) [our proposed C(II)] will be investigated mainly based on numerically [i.e. full wave simulations] calculated results. 5.6 A short discussion on previous tractor beam and lateral force experiments Chapter 5 Modification of Lorentz Force for embed. chiral and achiral obj. 94 Fig. 5.5: Calculation of time averaged total optical force (NO GAP METHOD) by external Minkowski ST and the time averaged bulk force by internal MEL ST. These forces are always of same trend. By adding the surface force of achiral MEL with bulk force [cf. Eqs (5.6) and (5.7)], the magnitude exactly matches with external time averaged total force by Minkowski ST. (a) For the two beam tractor set-up in Fig. 5.2(b) with 320 nm object. (b) For the two beam tractor set-up in Fig. 5.2(c) with 410 nm object. (c) For the lateral force set-up in Fig. 5.3(b) [only left hand CP wave incident case] with spherical object. (d) For the lateral force set-up in Fig. 5.3(c) [only right hand CP case] with elliptical object. For all the cases the trend of the time-averaged bulk force, obtained by employing the internal field only, is very similar to the total outside force calculated by the external Minkowski ST, using fields exterior to the scatterer. The bulk force by modified Chu volume force or stress tensor does not follow the trend of the total force. Moreover, the internal force calculation by the MEL ST is very less-time and memory consuming in comparison with the modified volumetric force calculation method [cf. the detailed discussion in [26] and [183]]. These are the main computational advantages of MEL ST method. The most probable reason of the success of Minkowski’s theory lies on the fact that, in contrast with other force formulations, Minkowski’s formulations, i.e. external Minkowski ST, Helmholtz force [35, 43, 28, 46, 192] and Minkowski photon momentum based ray tracing method [2,3] correctly account for the linear increase of transferred/emitted photon momentum at the boundary between the embedded object Chapter 5 Modification of Lorentz Force for embed. chiral and achiral obj. 95 and the background [35, 43]. Interestingly, when ( )2[1()]b jsis very small for a dielectric object embedded in another dielectric (for example- most of the real experiments [35,36,166]), the surface force of MEL in Eq (5.6) almost vanishes, as ( )Surface2MEL[1()]b jsf. Only for such special casesTotalBulkMink. MEL outinFF[as shown in [43] without any explanation]. In [43], the time averaged external force by Minkowski ST and the internal time averaged bulk force by MEL ST match well because the surface force almost vanishes. However, there must be a surface force part of MEL force. In [43], the time averaged external force by Minkowski ST and the internal time averaged bulk force by MEL ST match well because the surface force almost vanishes. But in general, for the exact total force formulation, that surface force should be added with the time averaged bulk force calculated by internal MEL ST [cf. Eq (5.7)] so that TotalBulkSurfaceMink MEL MEL outin F= FF. If the internal force for the tractor beam experiment in [35, 36] and the lateral force in [166] is calculated, our aforementioned conclusions still remain valid. It is shown in Fig.5.5 (a) and (b), total external time averaged force by Minkowski ST and the total internal bulk force by MEL ST are in almost full agreement for two beam tractor beam experiment in [36] due to very small value of ( )2[1()]b js. However, the difference between the bulk force of MEL ST and the total force of external Minkowski ST is clearly observable for the lateral force experiment when ( )2[1()]b js is not very small [i.e. by considering a TiO2 object embedded in air-water interface in Fig. 5.5(c)]. Also there is an effect of the size/shape change of the object on the bulk force of MEL ST as shown in Fig. 5.5(d). However, in Appendix D the bulk force calculation by MEL ST is shown for a Mie or more complex objects embedded in Chapter 5 Modification of Lorentz Force for embed. chiral and achiral obj. 96 homogeneous, heterogeneous and bounded background. As per we know, previously force calculation for heterogeneous medium has not been discussed in literature. 5.7 Chiral Modified Einstein-Laub and Chu formulations Fig. 5.6: Time-averaged forces: Fout at 1.01raa from Minkowski ST and Fin (bulk force) at 0.999raa from the Chiral MEL ST. These forces are always of same trend. By adding the surface force of Chiral MEL with bulk force, the magnitude exactly matches with external time averaged total force. (a) Calculation procedure of force on a dielectric sphere with a=500 nm (i.e., a Mie object) at 1064nm. The unbounded homogeneous dielectric background parameter:04b. (b) Force on that chiral dielectric sphere (chirality parameter,0.4 ) by varying the permittivity of the sphere, illuminated by a linearly polarized plane wave0i kztxEE e. Notice that the internal forces (bulk forces) calculated by all other STs (i.e. EL, Chu, Nelson and Minkowski) are not so close to the total time averaged force. (c) Force on the same embedded chiral dielectric sphere by varying the permittivity of the sphere, illuminated by a linear polarized and a circularly polarized (xyEiE:0i kztxyEE eE) wave. (d) Force on the embedded chiral dielectric (04s) sphere by varying the chirality parameter of the sphere, illuminated by a linear polarized and a circularly polarized wave. Chapter 5 Modification of Lorentz Force for embed. chiral and achiral obj. 97 In this section we shall show the consistency of the proposed internal MEL method with the external Minkowski ST method [174]. Especially for chiral objects, no built in software [183] technique is available to yield the total volumetric force. The primary goal of this section is to set an efficient computational way for embedded chiral objects. The constitutive relations inside a chiral Mie object can be written as [77]: chiralinsinin(/ ); jcDEH (5.8a) chiralinsinin(/ ).jcBHE (5.8b) If a magneto-dielectric chiral object is embedded in a material background instead of air, the internal stress tensor that may yield the total force of the scatterer is the chiral MEL stress tensor for chiral object: b ( )( )chiralchiral*chiral***MEL( )ininininsinininins1in()().2jb jjssTDEBHHHEEI (5.9) Where j=1,2,3,...., N represents the number of background regions sharing interface with the chiral object. The total time averaged bulk force on the embedded object should be: chiralBulkMEL( )( )ininjjjd F=Ts. A possible derivation of Eq (5.9) can be yielded very similar to our achiral MEL ST shown in Appendix B. The Bulk force of chiral MEL method can also be written as: Chiral****MEL( )Chiral( )inChiral( )inChiral( )inChiral( )in1BulkRe.2jjjjjiifPEMHPBMD (5.10) In Eq (5.10), the effective polarization and magnetization are defined as: Chapter 5 Modification of Lorentz Force for embed. chiral and achiral obj. 98 chiralbinin ,,( / )eceScjcPPMPEMH (5.11a) chiralnnbinin ,,( / )cScjcMMPMHPE . (5.11b) However, chiralinD and chiralinBin Eq (5.10) should be written from Eq (5.8a) and (5.8b) respectively. The surface force part has also been derived similar to achiral case discussed in previous section. The total time averaged force should be the surface force plus the bulk force. Consistency of chiral MEL ST for the unbounded homogeneous background is shown in Fig. 5.6 (a)-(d). In Fig. 5.7 (a)-(d), consistency of the chiral MEL ST has been shown considering a 2D infinite cylinder embedded in a heterogeneous background. Finally, the result shown in Fig. 5.8 [the bounded background case] bears some important physical insight. We have considered b=800 nm and a=600 nm. If b is made even much smaller and very close value of a (but b > a+), still the conclusion presented in Fig. 5.8 (a)-(d) remains valid. Hence we can conclude that the boundary between the scatterer and background play a vital role to yield the total time averaged force, which may not be properly explained by GAP METHOD of force calculation. The internal MEL ST (internal bulk force) leads to almost time averaged total force for several situations (or at least follows the trend of total time averaged force), which can be very useful from computational point of view for both embedded chiral and achiral objects. Finally, considering all the cases discussed in this work, probably we can assure the fulfilment of ‘consistency conditions’ C(I) and C(II) simultaneously by external Minkowski ST and the internal modified EL or Chu formulation both analytically and numerically. Chapter 5 Modification of Lorentz Force for embed. chiral and achiral obj. 99 Fig. 5.7: Time-averaged forces: Fout at 1.01raa from Minkowski ST and Fin (bulk force) at 0.999raa from the Chiral MEL ST. These forces are always of same trend. By adding the surface force of Chiral MEL with bulk force, the magnitude exactly matches with external time averaged total force. (a) Calculation procedure of Force on a magneto-dielectric infinite chiral (chirality parameter,0.5 ) cylinder of 00(,)(5, 4)ss and radius 2000 nm embedded in heterogeneous unbounded background of four different magneto-dielectric layers: (,bb ) = (03 ,02 ); (04 ,03 );(05 ,04 ); (06 ,05 ) at 1064nm. (b) Force on that cylinder when the plane wave 0i kztxEE eilluminates at varying angles. Notice that the internal force calculated by all other STs (i.e. EL, Chu, Nelson and Minkowski) are not in the same trend of the total external force. (c) Force on the same embedded cylinder when the illuminating circularly polarized wave:xyEiE:0i kztxyEE eEincidents at varying angles. (d) Force on the cylinder when the illuminating plane waves at angle 45 degree at varying chiral parameter . Chapter 5 Modification of Lorentz Force for embed. chiral and achiral obj. 100 Fig. 5.8: Time-averaged forces: Fout at 1.01raa from Minkowski ST and Fin (bulk force) at 0.999raa from the Chiral MEL ST. These forces are always of same trend. By adding the surface force of Chiral MEL with bulk force, the magnitude exactly matches with external time averaged total force. (a) Calculation procedure of a 3D magneto-dielectric core where the whole core-shell sphere is embedded in air. Core radius, a=600 nm,08s, 04s Bounded local immediate background (i.e. the shell) parameters: radius, b=800 nm and 04s;02s. This entire core-shell is illuminated at wavelength 1070nm. (b) For plane wave illumination, 0i kztxEE e: CoreoutFat different chirality parameters obtained from Minkowski ST at r=a+ using the fields in the shell. Force CoreinF based on the Chiral MEL ST at r=a- using core fields. The bulk force on the core by other STs do not follow the trend of the total external force. (c) For circularly polarized wave illumination (xyEiE:0i kztxyEE eE): still the bulk force by Chiral MEL ST is of the same trend of the external time averaged total force for the variation of chirality parameter of the core. (d) For linear and circularly polarized wave illumination: again our conclusions remain valid for the variation of chirality parameter of the core. Chapter 5 Modification of Lorentz Force for embed. chiral and achiral obj. 101 5.8 Conclusion In this chapter we have shown that: the ‘validity condition’ of inT and f (in), defined at the beginning, plays a key role for the severe disagreement reported in [28, 46] of different force laws employed to describe real experimental results. We have proposed a solution to this problem based on the MEL ST, the MEL volume force [171] or alternatively by the modified Chu volume force [26, 183]. However, If Bulk inin0dFTs; (out)T and inT cannot have the same form to satisfy both consistency conditions C(I) and C(II) simultaneously, when the background is a material medium rather than air. What is the physical reason behind that? A possible answer of this important question is discussed next. Let us consider the total momentum conservation equation [17]: .dsdvdvtTf G and TotalMech.Non-Mech.ppp where p represents momentum. Here G is the electromagnetic momentum density. Though the total momentum Totalpis always a conserved quantity, whenever one calculates/measures the photon momentum transfer from the background [58,151], it leads to Minkowski photon momentum where instead of simple mechanical and non-mechanical momentum part, the linear momentum equation can better be described as [58]: medTotalCano.Minkppp where MinkMinkdv pG with Mink GDB . On the other hand, Abraham photon momentum [PAbr (in)] is considered as the travelling momentum of photon [151], which [PAbr (in)] can also be considered as the remaining electromagnetic part of the photon momentum after delivering the mechanical momentum of photon inside an object setting that object in motion [hence kinetic momentum [58]:medTotalKinAbrppp , where AbrAbrdv pG with 2Abrc EHG]. So, the role of Minkowski and Abraham photon momenta are fully different [58, 151] and Chapter 5 Modification of Lorentz Force for embed. chiral and achiral obj. 102 hence in the aforementioned conservation equation, G should also be different when we describe two different process: transfer of momentum from background (i.e. transfer ofMinkpfrom background to an embedded object due to Doppler effect [58]) and delivery of momentum inside an object (i.e.Abrpinside the Einstein-Balaz’s box [58]). This suggests that: as total momentum is a conserved quantity, f can also be different due to different G and they should bear fully different physical meanings. Hence the proposed explanation is that: time averaged Minkowski’s ‘total’ force describes how much mechanical momentum has already been transferred to an embedded object from the background medium, which can be calculated just by employing the external fields of the embedded object. In contrast, time averaged MEL ‘bulk’ force describes how much force (due to local fields) has been felt by the induced dipoles of an embedded object by which (dipoles) that object is made. Now, most importantly, though by adding the surface force of MEL with the time averaged bulk force we get exactly the same value of time averaged total force predicted by Minkowski’s external force [ TotalBulkSurfaceMink MEL MEL outin F= FF], they [left side and right side of aforementioned equation] are two fully different things from physical point of view. In fact, this issue has already been pointed regarding only photon momenta in refs. [58, 151]. But such dissimilarity may, in general, also be true for the role of distinct time averaged stress tensors, which has been overlooked so far. So, we suggest the distinct physical meanings of the left and right hand sides of this equation: TotalBulkSurfaceMink MEL MEL outin F= FF, which (distinct physical meanings) is also applicable for the modified Chu formulation [26]: TotalBulkSurfaceMink Chu Chu outin F= FF. Last but not least, our work explains not only the reason of the inconsistency of different time averaged volumetric forces reported in [28], [46], but also provides an efficient alternative solution to calculate the time-averaged bulk force with the Chapter 5 Modification of Lorentz Force for embed. chiral and achiral obj. 103 modified Einstein-Laub (MEL) stress tensor method, saving much calculation time and memory with respect to the time-consuming bulk volumetric force method [183] for both embedded achiral [19,20, 28,35,36,46, 84,97, 98, 183- 186] and chiral objects [19,20]. Similar time-averaged modified formulations are also possible on employing modified Chu methods [26,183]. In fact, the main goal of this chapter is not to show superiority of one formula over another. Rather the main targets of this work were to make fully distinct formulas mathematically equivalent to yield accurate and consistent time-averaged total optical force, as well as explaining the exact reason of discrepancies of previous theories for real experimental observations. Chapter 6 Mod. Lorentz Force and Plasmonic Cube Dimers over Substrate 104 Chapter 6 Modified Lorentz Force and Plasmonic Cube Dimers over Substrate: Binding Force Reversal 6.0 Summary of Chapter 6 In previous chapter it has been discussed in details why Lorentz force should be modified when the embedding background is material one instead of air. In this chapter we apply the modified Lorentz force (modified Chu force) to study the behavior of binding force for two cubes over different substrates fully immersed in water. The behavior of Fano resonance and the reversal of near field optical binding force of dimers over different substrates have not been studied so far. In this work, we observe that if the closely located plasmonic cube homodimers (over glass or high permittivity dielectric substrate) are illuminated with plane wave polarized parallel to dimer axis, no reversal of optical binding force occurs. But if we apply the same set-up over a plasmonic substrate, stable Fano resonance occurs along with the reversal of binding force. It is observed that during such Fano resonance stronger coupling occurs between the dimers and plasmonic substrate along with the strong enhancement of the substrate current even in presence of dielectric spacer. Interestingly, for both the strong quadrupole-quadrupole and dipole-dipole resonances of the dimers, the binding force is dominated by Coulombic surface charge induced modified surface Lorentz force ; whereas during the Fano resonance binding force reversal is dominated by polarization current induced modified bulk Lorentz force. Chapter 6 Mod. Lorentz Force and Plasmonic Cube Dimers over Substrate 105 6.1 Introduction Fano resonance and the Fano line shape [193] cannot be described by the Lorentz resonance formula [194]. In plasmonic nanostructures, Fano resonance can happen due to the resonant destructive interference between a super-radiant (bright) mode and subradiant (dark) modes [195]. The promising applications of plasmonic hybridization and Fano resonances [39,41] have been investigated in improved sensitivity of the resonance [196], bio sensing [119], surface-enhanced Raman scattering [197], photonic propagation and wave guiding [198], plasmon-induced transparency [124] and many others [127]. In contrast, much less attention is dedicated on near field optical force due to Fano resonance; especially for plasmonic dimers [39,41] as discussed only in [22,23]. Specially, plasmonic particle over substrate show remarkable properties: Fano resonances [199], Boradband tenability in plasmonic resonance [200], modification of energies of the plasmon modes [201], changing the localized density of states [202], radiative enhancement [203], frequency shift of an electric dipole resonance [204] and so on. But the behavior of near field optical force for plasmonic dimers over different types of substrate and the effect of substrate on the reversal of near filed optical binding force have not been studied in literature. In fact, only few works have been reported on the reversal of near field optical binding force due to Fano resonance, i.e. for nanobar structures [22], for disk along with a ring structure [23]. But reversal of near field binding force are highly dependent on inter-particle distance and particle size for those set-ups. According to ref. [205]: “The inter particle separations are usually comparable to the wavelength of the illuminating laser beam and, therefore, mainly long-range interactions mediated by the far-field scattered field are considered while the near-Chapter 6 Mod. Lorentz Force and Plasmonic Cube Dimers over Substrate 106 field coupling is usually omitted in the related studies.” But with the recent technology, the inter particle distance between the nano particles can be precisely controlled [206]. More importantly, the size and shape of the gap between the nanoparticle and film can also be controlled to sub nanometer precision bottom-up fabrication approaches [206]. The film-coupled nanoparticle geometry has recently been applied to enhancing optical fields, accessing the quantum regime of plasmonics [207], and the design of surfaces with controlled reflectance [208]. Still no investigation has been carried out regarding the behavior of both Fano resonance and the reversal of near field optical binding force with respect to the inter particle distance of two dimers over different substrates. In this work, it is observed that if two plasmonic cube homodimers are placed closely without any substrate or above glass or high refractive index substrate, no reversal of optical binding force occurs. But if we apply the same set-up over a plasmonic substrate, stable and stronger Fano resonance occurs along with the reversal of binding force. It is well known that usually Fano resonance is ultra-sensitive and previously reported reversal of optical binding force dies out very quickly if the inter particle gap of the heterodimers [22] increase even a little bit. Moreover, this [22] previously reported Fano resonance induced reversal of binding force is highly dependent on particle size, which has not been achieved in [22] and [23] by homodimers. In general, reversal of near field lateral optical binding force for the same polarization of light is quite uncommon with the homodimers placed without substrates [113, 134, 135, 209]. Even if the homodimers are placed over substrates, reversal of near field lateral optical binding force has not been observed in refs. [136, 210,211] for spherical shaped and rod shaped plasmonic homodimers. As a result, we have investigated a fully alternative way to achieve strong and stable reversal of optical binding force based on cube homodimer and substrate system, where stable and much stronger Fano resonance can be achieved due to the Chapter 6 Mod. Lorentz Force and Plasmonic Cube Dimers over Substrate 107 substrate coupling with larger area of the lower portions of cube dimers. Though nanoparticles of sphere is easy to synthesize in experiment, the substrate-induced hybridization of the plasmon modes can be much larger [199] for a planar metallic nanoparticle (i.e. a cube) deposited on a plasmonic substrate [206,208] than for a spherical particle [201]. This happens because the plasmon-induced steady state currents and surface charges will be located closer to the strong surface currents and plsmonic screening charges induced on the surface of the substrate over a large area. We have demonstrated that from the interplay between localized surface plasmon and propagating surface plasmon polariton along with the strong coupling between the two particles with the plasmonic substrate, the reversal of the optical binding force occurs during the strong Fano resonance. The physical mechanisms of binding force reversal have been explained based on modified Lorentz force dynamics. This lateral binding force reversal for the homodimers may be more flexible than the previously reported cases in [22, 23] where the reversal of near field binding force of the heterodimers dies out with the change of: (i) inter particle distance and (ii) particle size. Our configuration provides much relaxation of those parameters and hence can be verified experimentally with simpler experimental set-ups. 6.2 Optical Force calculation We again specify that throughout this thesis we refer to 'exterior' or 'outside' forces as those evaluated outside the volume of the macroscopic particles, while ‘interior’ or 'inside' refer to those quantities inside this object volume. In order to consider the realistic effects, we have done all the numerical calculations using full wave simulations [100] in three dimensional (3D) structures. Chapter 6 Mod. Lorentz Force and Plasmonic Cube Dimers over Substrate 108 The proposed simple set-up is illustrated in Fig. 6.1(a). The Silver particles are placed near to each other above a silver substrate (whole set-up is embedded in water). Spacer of height 5 nm is always considered above all the substrates in this chapter. The real and imaginary part of the permittivity of silver is taken from the standard Palik data [100, 139]. Inter particle distance is ‘d’. The source is a simple x-polarized plane wave 0i zxEE e propagating towards –z direction. This set-up is a symmetry broken system which later plays a vital role for the force reversal. The ‘outside optical force’ [33] is calculated by the integration of time averaged Minkowski [43, 84] stress tensor at r=a+ employing the background fields of the scatterer of radius a: OutoutTotalout****outoutoutoutoutoutoutout1 Re12(][)2dDEBHEDHBFTsTI. (6.1) Where ‘out’ represents the exterior total field of the scatterer; ,,and are the electric field, displacement vector, magnetic field and induction vectors respectively, represents the time average and is the unity tensor. On the other hand, based on the modified Lorentz force (i.e. modified Chu force), the total force (surface force and the bulk force [26,171,183] part of modified Lorentz force) can be written as: TotalVolumeBulkSurfBulkSurfacedvdsFFFFff (6.2) Where outinoutinoutinout**Surfaceatnai**att[]ˆˆ ()(),22eavgmavgrabbrarafEHEEHHEEnHHn (6.3) EDHBIChapter 6 Mod. Lorentz Force and Plasmonic Cube Dimers over Substrate 109 is the modified surface force density (the force which is felt by the induced bound electric and magnetic surface charges of a scatterer), which is calculated just at the boundary of a scatterer [26,171,183]. ‘in’ represents the interior fields of the scatterer; ‘avg’ represents the average of the field. and are the bound electric and magnetic surface charge densities of the scatterer respectively. The unit vector is an outward pointing normal to the surface. b is permittivity and b is permeability of the background (i.e. water). ****Bulk0inin0inininininin11Re[ ()()]Re[ (){}](){}22sbsbii EHfEEBDH (6.4) is the modified bulk force density, which is calculated from the interior of the scatterer by employing the inside field [26,171,183]. is permittivity and is permeability of the scatterer. As per we know, the Lorentz force dynamics of plasmonic particles and specially dimers have not been discussed previously. It is notable that the ‘external dipolar force’ [34, 38] (which has also been described as Lorentz force in [38]) is quite different than the Lorentz force defined in our Eqs (2) - (4). 6.3 Plasmonic cubes over plasmonic substrate: a short discussion on resonance modes and reversal of binding force SurfacefemˆnBulkfssChapter 6 Mod. Lorentz Force and Plasmonic Cube Dimers over Substrate 110 Fig. 6.1: (a) Two silver cubes (L=120 nm) are placed over silver substrate and their inter-particle gap, d, is 100 nm (spacer height 5 nm from the bare substrate and the cubes are placed 5 nm away from the spacer). (b) Comparison of the extinction coefficients of two silver cubes [set-up of Fig. (a)] and a single silver cube placed over silver substrate. (c) The binding force of set-up (a). (d)-(f): charge distribution at wavelengths: 966 (DD), 754 (Fano) and 627 (QQ) nm respectively. D and Q represent dipole and quadrupole respectively. (g)-(i): Electric field distribution for those same wavelengths respectively. In Fig. 6.1(a) we have considered two silver cubes of length 120 nm each (100 nm apart). These cubes are placed in water above a silver substrate. A spacer of 5 nm height is considered between the cubes and the substrate and the cubes are placed 5 nm above the spacer. This set-up is shined by a plane wave described in previous section. The extinction spectra in Fig. 6.1(b) reveals that Fano dip takes place at around 754 nm and the strength of extinction coefficient increases two times than a single cube placed over silver substrate. The quadropolar resonance (at wavelength 627 nm), dipolar resonance (at 966 nm) and Fano dip (at around 754 nm) do not experience shifting due to the addition of the identical cube. It is demonstrated in Fig. 6.1(c) that reversal of optical binding force [i.e. positive value of Bind) (xFmeans Chapter 6 Mod. Lorentz Force and Plasmonic Cube Dimers over Substrate 111 attractive force and negative value means repulsive force] takes place near the Fano dip at 754 nm wavelength. It should also be noted that if we decrease the inter particle distance of the cubes, the magnitude of the lateral binding force increases where lateral binding force,Bind 1( )( )( 2 ) xxxFFF. Here subscript (x), (1) and (2) represent: +x direction, left cube and right cube respectively. It is also observed in our full wave simulations that the reversal of binding force occurs at the Fano dip position (around the wavelength 1000 nm) even if we change the length of the cube from 120 to 180 nm [cf. Fig. 6.2 (a)-(c)]. Such reversal of binding force is quite generic for plasmonic cubes of other sizes placed over plasmonic substrates. The only countable issue to achieve such force reversal is to achieve Fano type dip at first with the single cube. However, the force reversal wavelength and the Fano resonance both red shift with the increase of the cube dimension. The reason of Fano resonance for the dimers can be very easily explained based on the destructive interference of quadropolar and dipolar mode of the dimer as previously discussed considering high refractive index substrate in [199]. The distinct charge distributions of the plasmonic cubes have been shown in Figs 6.1(d) –6.2(f) for different resonance modes [i.e. dipole-dipole (DD), Fano and quadrupole-quadrupole (QQ) resonance respectively]. In addition, in Figs 6.1(g)-(i) the electric field magnitudes of different modes [i.e. DD, Fano and QQ resonance respectively] have been demonstrated. The field distributions are similar (but not identical) to ref. [199] where the substrate is considered as high refractive index dielectric. The strength of electric field magnitude increases due to the use of plasmonic substrate, which is a necessary condition for stronger coupling and resonance that leads to binding force reversal. However, the mixed nature of electric field at the Fano dip clearly supports the mixing of dipolar and quadrupolar resonance [199, 200]. But the question is whether Fano resonance is the only factor to achieve such force reversal. The answer is: strong Fano resonance induced by strong plasmonic coupling Chapter 6 Mod. Lorentz Force and Plasmonic Cube Dimers over Substrate 112 is the key factor of such force reversal but it is not the only reason. For example- in the next section we shall demonstrate that Fano resonance also occurs for Silicon substrate with cube dimers but no reversal of binding force takes place for such case. In addition, more details on the physical mechanism of such optical force reversal will be discussed in the next sections. At the DD resonance mode (i.e. 966 nm wavelength), field lines form closed loops around the separation between the particle and the metallic film [212]. Maximum value of the repulsive binding force occurs at this wavelength as shown in Fig. 6.1(c). This resonance mode corresponds to the localized surface plasmon (LSP) mode associated to the silver particle. According to another point of view [206], this resonance mode can be explained as a magnetic cavity mode that roughly satisfies the half wavelength criteria. In particular, the metal (silver)−insulator (spacer)−metal (silver) region supports a transmission line type of mode modified by the plasmonic dispersion of the metal. A large effective impedance mismatch occurs at the edges of the cube and most of the energy is reflected back under the cube [cf. the electric field distribution in Fig. 6.1(g)]. On the other hand, a resonance peak of extinction spectra occurs at around 627 nm as shown in Fig. 6.1 (c), which corresponds to QQ resonance. Local maxima of the attractive binding force occurs during this resonance. It corresponds to the surface plasmon polariton (SPP) mode propagating on the silver/air interface of the metallic film, which is excited near this wavelength [cf. ref. [212]]. The corresponding enhancement of the particle excitation field increases the intensity of field scattered at infinity [cf. the electric field distribution in Fig. 6.1(i)]. From another point of view [206], this higher order mode is an example of waveguide modes and occurs at shorter wavelengths such that the cavity is longer than a half wavelength of the mode. Chapter 6 Mod. Lorentz Force and Plasmonic Cube Dimers over Substrate 113 Fig. 6.2: (a) Two silver cubes (L=180 nm) are placed over silver substrate and their inter-particle gap,d, is 150 nm. The spacer (refractive index 1.4) height is 5nm for all the cases discussed in this work. Cubes are placed 5nm (h= 5nm) away from the spacer. x-polarized plane wave is propagating towards –z-direction. (b) Extinction co-efficient (c) Optical binding force. 6.4 Plasmonic cubes over different substrates: effect on binding force So far we have discussed regarding single or double plasmonic cubes over plasmonic substrates. Now we shall compare the behavior of optical binding force considering two plasmonic cubes without substrate and over the glass or Silicon (high refractive index) or silver (plasmonic substrate) substrate fully immersed in water [ cf. Fig. 6.3 (a) –(c)]. No reversal of optical binding force is observed in Fig. 6.3 (c). This same conclusion is true for the 2nd case [cf. Figs. 6.3(d)-(f)] when no reversal of optical binding force is observed. However, in Fig. 6.3(g)-(i), the substrate is Silicon (modelled as Palik data [100, 139]). Ultimately Fano dip spectral region is observed in Fig. 6.3(h) due to Fano resonance. But no reversal of optical binding force is observed in Fig. 6.3 (i). This is an important observation. This observation suggests that Fano resonance is not the only criteria to achieve the reversal of optical binding force [which occurs for silver slab in Fig. 6.3 (l)] rather some other factors (will be discussed next) significantly influence the origin of the reversal of optical binding force. Now we shall discuss the behavior of binding force distribution for these three resonances: (i) DD (ii) QQ and (iii) Fano resonance, when the substrate is silver one. Chapter 6 Mod. Lorentz Force and Plasmonic Cube Dimers over Substrate 114 Fig.6.3: Two silver cubes (L=120 nm) are placed over different substrates and their inter-particle gap, d, is 150 nm. The spacer height is always 5nm. Cubes are placed 5nm (h= 5nm) away from the spacer. x-polarized plane wave is propagating towards –z-direction. (a)-(c): No substrate is placed; the extinction coefficient and binding force for that configuration. (d)-(f): Glass substrate (refractive index 1.5) is placed; the extinction coefficient and binding force for that configuration. (g)-(i): Silicon substrate is placed; the extinction coefficient and binding force for that configuration. (j)-(l): Silver substrate is placed; the extinction coefficient and binding force for that configuration. For the DD resonance mode, the resonance process with the plasmonic substrate is quite different in comparison with the other substrates. During the DD resonance mode, the accumulated charges are strongly coupled with the plasmonic substrate and the total power is mostly concentrated between the cube and the plasmonic substrate. The charge distribution essentially results in a current loop between the nano cube and metal film inducing a strongly enhanced magnetic field. This resonance is then mainly of magnetic nature [206, 208, 212]. But the most interesting part is that: though the resonance is of magnetic nature, the difference of the scattering part Chapter 6 Mod. Lorentz Force and Plasmonic Cube Dimers over Substrate 115 (which originates from magnetic induction: inJB ; where J is the polarization induced current) of the total Lorentz force vanishes during the peak of LSP or DD resonance mode [cf. Fig. 6.4 (a), (c) for both 120 and 180 nm lengthened cubes]: Bulk( )Bulk (1)(1)Bulk (2)(2)Del F][] [xdvdvff (6.5) So, the total binding force at LSP or DD resonance peak is fully due to the surface charges and the difference of surface force can be expressed as: Surf (x)Surface ()(1)Surface ( )(2)[]Del []FBSdsdsff (6.6) It should be noted that: ( )( )Bin( )Bulk( )Sud 1 rf2(x) Del Del FFxxxxFFF. In Fig 6.5(i) it can be seen that the flipped Surface (1)(1)[]dsf [and alsoSurf (x)Del F; cf. Fig 6.4 (b) and (d)] reaches its maximum negative value near the DD resonance for plasmonic silver substrate [also cf. Fig 6.3(l)]. Fig. 6.4: For cube homodimers over silver substrate, plot of Bulk( )Del Fx and Surf (x)Del F[cf. Eq (6.5) and (6.6)]. Spacer height is always 5 nm and height of the cubes from the spacer is fixed 5nm from the spacer. For first column: Length of each cube is 120 nm and they are placed 100 nm apart. For second column: Length of each cube is 180 nm and they are placed 150 nm apart. Chapter 6 Mod. Lorentz Force and Plasmonic Cube Dimers over Substrate 116 Fig. 6.5: For the cases of different substrates, Lorentz force components and the steady state current (Jy) of two 120 lengthened silver cubes along with the substrate. First row represents the Lorentz force components: surface force, bulk force and total force only on cube-1 (left cube, denoted as P1). ‘°’ represents the chosen wavelengths for which the steady state current distributions are plotted later. Second, third and fourth row represent the steady state current (Jy) from front view [in xz plane; setting the window very close to the cube surfaces from front view] for three different wavelengths (marked as ‘°’) respectively: (a)-(d): For glass substrate where wavelengths are chosen: 754, 622 and 550 nm respectively. (e)-(h): For Silicon substrate where resonance wavelengths are: 816 (DD), 679 (Fano) and 578 nm (QQ) respectively. (i)-(l): For Silver substrate where resonance wavelengths are: 966 (DD) ,754 (Fano) and 627 nm (QQ) respectively. From the first row of Fig. 6.5, it can also be seen that such reversal of Surface (1)(1)[]dsf does not happen for other substrates. This can be explained based on the electric field coupling between the cubes and also between cubes with plasmonic substrate. Electric field enhancement becomes much stronger for the case of plasmonic substrate in comparison with other substrates especially at this particular Chapter 6 Mod. Lorentz Force and Plasmonic Cube Dimers over Substrate 117 resonance. The surface force in Eq (6.3) depends on the electric fields of the cube boundary/interface, which becomes much stronger during this DD resonance. Though the maximum value of repulsive binding force for DD or LSP mode is just due to surface charges, the local maximum value of attractive binding force for QQ or SPP resonance at wavelength 627 nm [cf. Fig. 6.3(l)] is the combined effect of static (opposite surface charges as shown in Fig. 6.1 (f)) and dynamic (propagation of SPP) process. If we give a close look at Lorentz force in Fig. 6.4, it is clearly observable that at this QQ resonance: though the resultant binding force is due to the dominance of Surf (x)Del F, Bulk( )Del Fx also exists. This fact can be verified from Fig. 6.5(i), where the dominance of Surface ()(1)[]Bdsf on cube-1 is clearly observable. The opposite charges of the cubes create the attractive force between the cubes placed over any substrate such as plasmonic or glass or silicon substrate. At the same time, the propagating surface plasmon of the plasmonic substrate also create force on the cubes in the opposite direction of its propagation, which ultimately causes the local maximum value of attractive force at this QQ or SPP resonance [cf. Fig. 6.3(l)]. Now we shall consider the behavior of optical force for Fano resonance when the reversal of lateral binding force takes place [cf. Fig. 6.3(l)]. If we compare the extinction efficiency of the four cases [cf. the second column of Fig. 6.3], the magnitude is almost 2.5 times stronger for the plasmonic (silver) substrate case. In addition, if we consider the strength of induced current at resonant frequencies (i.e. Fano resonance and previously discussed DD resonance mode), the surface current for the case of silver slab is quite stronger in comparison with all other cases: no substrate, glass substrate and silicon substrate case. Now if we give a closer look to the individual cube (i.e. cube-1): it is clearly observable that the bulk force part turns into negative force only near the Fano resonance regime for the case of plasmonic substrate. However, this bulk force (which is mainly connected with the scattering Chapter 6 Mod. Lorentz Force and Plasmonic Cube Dimers over Substrate 118 force of plasmonic objects [114, 143]) is never dominant, when the substrate is glass or silicon. The total force of cube-1 for glass and silicon substrate case is always dominated by surface force. In contrast, the scenario is fully reversed for the case of silver substrate especially during the Fano resonance. In Fig. 6.5(i) it can be observed that the reversal of optical binding force occurs due to the reversal of bulk Lorentz force. Our ultimate conclusions are: (i) the reversal starts to occur from strong multiple scattering. (ii) Though it is commonly believed that internal wave-field does not contribute much for optical force on plasmonic objects (cf. the dominance of surface force in [109, 114]), during the Fano resonance the scenario is quite different. 6.5 Effect of height, size and background material on binding force In previous section we have discussed the effect of different substrates on lateral binding force. In this section we shall focus our discussion on lateral binding force due to the change of dimer heights from the plasmonic substrate. In Fig. 6.6(a)-(f), it is shown that when the cubic silver dimers are far away (45 nm away from the spacer) from the plasmonic substrate, Fano resonance does not occur. For these dimers, placed 45 nm away from the spacer, the dominant resonance in Fig. 6.6(a) is the DD resonance [but for the isolated dimer case in Fig. 6.3(a) the dominant resonance is not DD mode [199]]. It should be noted that the binding force at DD resonance mode behaves quite differently for these two cases: (i) Far from the substrate- strong mutual attractive force occurs and (ii) very close [i.e. 5nm away from spacer as shown in Fig. 6.3 (j)-(l)] to the substrate- strong mutual repulsive force occurs [the reversal mainly occurs before this dipole resonance at the strong Fano resonance, when stronger coupling with the plasmonic substrate starts to occur]. Chapter 6 Mod. Lorentz Force and Plasmonic Cube Dimers over Substrate 119 Fig. 6.6: Two silver cubes (L=120 nm) are placed at different heights from the silver substrates (three different columns represent the position of the cubes: 50, 20 and 15 nm away from the substrate) and their inter-particle gap, d, is fixed 100 nm. The spacer height is 5nm. x-polarized plane wave is propagating towards –z-direction. For different heights from the substrate: First row: extinction co-efficient; second row: binding force; ‘°’ represents the chosen wavelengths for which the steady state current distributions are plotted later. Third row: Lorentz force components along with the total force only on cube-1 (left cube, denoted as P1); fourth, fifth and sixth row: steady state current (Jy) from front view [in xz plane; setting the window very close to the cube surfaces from front view] for two silver cubes along with the substrate in three different wavelengths (marked as ‘°’). For first column [50 nm height, (d)-(f)] wavelengths are chosen as: 855, 553 and 503 nm respectively. For second column [20 nm height, (j)-(l)] resonance wavelengths are: 872 (DD), 631 (Fano) and 567 nm (QQ) respectively. For third column [15 nm height, (p)-(r)] resonance wavelengths are: 890 (DD), 674 (Fano) and 590 nm (QQ) respectively. Chapter 6 Mod. Lorentz Force and Plasmonic Cube Dimers over Substrate 120 Now, in Fig. 6.6(g)-(l), when the cubic silver dimers are closely placed (15 nm away from the spacer) from the plasmonic (silver) substrate, Fano resonance does not die out. However, such Fano resonance is not strong enough and optical binding force does not reverse. This configuration can better be comparable with the case: plasmonic homodimers (cube) over the previously discussed Si substrate. The current, electric fields and magnetic fields of such set-ups are also similar (but not identical) to that previously discussed case of dimers over Si substrate. Now, in Fig. 6.6 (m)- (r), when the cubic silver dimers are closely placed (10 nm away from the spacer) from the plasmonic substrate, strong Fano resonance just starts to occur along with strong coupling with the substrate. During this resonance, the bulk part of Lorentz force reverses and just starts to dominate the total force on each dimer [cf. Fig. 6.6(o)]. Ultimately these lead to the reversal of optical binding force as shown in Fig. 6.6 (n). The mechanism of binding force for this case is very similar to the case [5 nm away from the substrate] discussed previously in details in previous section. The waveguide description [213] of resonance mode [206, 208] also aids in understanding the observed dependence of the resonance shift on both the size of the nano cube [as can be seen from the comparison of Fig. 6.1(b), (c) and Fig. 6.2 (b), (c)] and the gap between the silver cube and the silver substrate [as shown in Figs. 6.6]. (1) For larger cube, the length of the cavity increases and will therefore support a resonance at a longer wavelength. As a result, the resonance redshifts. (2) Similarly, when the gap between the substrate and the cube gets smaller, the effective refractive index of the cavity mode increases. This effectively lengthens the cavity [206] and results in a resonance condition at a longer wavelength. For case (1) and (2), when the second cube is placed closely to the first cube, the induced stable Fano resonance also Chapter 6 Mod. Lorentz Force and Plasmonic Cube Dimers over Substrate 121 red shifts following the rapid red shift of DD resonance and hence the binding force reversal wavelength also red shifts following the Fano dip in figures of Fig. 6.6 and Fig. 6.1 (b), (c); Fig. 6.2 (b), (c). Throughout this work, the background material has been considered water. The simplest way to blue shift all the resonances along with the wavelength of binding force reversal is to decrease the background refractive index. Last but not least, with the increase of the inter-particle gap, the binding force magnitude decreases slowly but still this overall set-up provides a very relaxed mechanism to verify the reversal of optical binding force. 6.6 Conclusions In summary, we have investigated a very simple possible configuration to demonstrate the reversal of lateral optical binding force with plasmonic homodimers based on strong Fano resonance. Among all the substrates (i.e. glass, Si, Ag, Au etc.), the closely placed plasmonic particles should remain very close only to the plasmonic substrate so that the bulk part of the total Lorentz force dominates the total lateral force of each cube during the substrate mediated Fano resonance. The surface current (along with the strong electric and magnetic coupling) increases significantly only for the close presence of the plasmonic substrate, which bears significant influence on the reversal of binding force of the dimers. Though we have shown only the homodimer cases, our proposed idea should also work for the cube heterodimers providing more flexibility on particle size. We believe that our proposals can be verified experimentally due to the simplicity of the proposed set-ups; and thus the attractive Chapter 6 Mod. Lorentz Force and Plasmonic Cube Dimers over Substrate 122 and repulsive forces between two plasmonic objects can be robustly adjusted based on the idea proposed in this chapter. Chapter 7 Consistency of Minkowski’s Theory: Some Final Proposals 123 Chapter 7 Consistency of Minkowski’s Theory: Some Final Proposals 7.1 Introduction and Summary of Chapter 7 In chapter 5, we have suggested the modification of Lorentz force for the embedded chiral and achiral objects. Later, in chapter 6, we have explained the dynamics of binding force reversal of plasmonic cubes based on the idea of modified Lorentz force discussed in chapter 5. In contrast, based on the proposed ‘consistency conditions’ of chapter 5, in this chapter we have proposed a new idea/hypothesis named as ‘existence domain’. ‘Existence domain’ is the region either outside a scattering body taking only its exterior fields into account, or in its interior considering only the inside fields. Though almost all the time averaged distinct STs and volumetric force laws are restricted to the idea of ‘existence domain’, we demonstrate that the time averaged stress tensor and volumetric force law of Minkowski are free from such restrictions. In addition, we have discussed in details the differences between time averaged and instantaneous force laws for different formulations in this chapter. Notably only Minkowski’s theory remains valid for all circumstances without any modification. As a result, considering ‘all’ the proposals in this chapter, it appears that Minkowski formulations are most probably more accurate description of physical optical force along with photon momentum. Chapter 7 Consistency of Minkowski’s Theory: Some Final Proposals 124 7.2 Time averaged Force: Proposal of ‘existence domain’ hypothesis Fig. 7.1 Optical momentum transfers from an incident plane wave of propagating vector ki to a spherical particle with permittivity and permeability . The total force calculated using the Stress Tensor (ST) is: Fout when the time-averaged ST is evaluated from fields outside the particle, through a surface of radius (blue lines), and Fin when the time-averaged ST is evaluated from the inside of the particle through a surface of radius (red lines). According to chapter 5, though the proposed modified Einstein-Laub (MEL) ST (and the bulk volume force of MEL or modified version [26] of the well-known Chu force [214]) leads to a consistent time-averaged ‘bulk’ force for embedded generic objects, it is applicable at r=a- employing only the interior fields of an embedded object [cf. Fig. 7.1]. On the other hand, Minkowski (and Abraham) ST leads to consistent time-averaged ‘total’ force when they are applied at r=a+ [cf. Fig. 7.1] employing only the exterior fields of an embedded object (and zero force when applied at r=a- on employing interior fields). We have also explained the reason behind that in chapter 5, and suggested the distinct physical meanings of the left and right-hand sides of the equation: TotalBulkSurfaceMink MEL MEL outin F= FF (also alternatively applicable for Chu formulations). pp 1.001raa0.999raaChapter 7 Consistency of Minkowski’s Theory: Some Final Proposals 125 Ultimately above proposal leads to a very surprising but significant conclusion for the well-known distinct time averaged stress tensors. We henceforth characterize the idea of ‘existence domain’ for the time averaged optical stress tensors (and volumetric forces) as the region either outside a scattering body taking only its exterior fields into account, or in its interior considering only the inside fields. The fundamental proposals [named as two consistency conditions] in favour of the idea of ‘existence domain’ proposed in chapter 5. In chapter 5, we have considered a special condition for non-absorbing objects: BulkBulk ininin0dvdFfTs. But if we consider a more generic case where the existence of time averaged bulk force is not any big deal for a non-absorbing object, it is possible to arrive an optical force law (Minkowski force) which is not restricted to the idea of ‘existence domain’. To clarify the aforementioned proposal by a very simple example (more complex cases already discussed in chapter 5), let us consider a magnetodielectric slab embedded in another magnetodielectric. The surface force part of well-known Chu and Einstein-Laub force vanishes for the slab case. But the bulk part plays the main role to decide the time averaged total force [215]. The lossless embedded slab illuminated at normal incidence by a linearly polarized plane wave propagating along the z-direction with electric vector:0i kztxEE e. The ‘outside force’ can be calculated considering the transfer of Mink (out) pk from the background to the embedded slab (cf. Eq (60) given in [215]). At the same time, the interior force on the embedded slab is calculated by internal bulk force of modified Einstein-Laub or modified Chu. All of them are in full agreement with each other: Chapter 7 Consistency of Minkowski’s Theory: Some Final Proposals 126 2Transfer Process220From Background1121b SS biirrttsbssbssEN kN kN kRT (7.1) Here R and T denote the reflection and transmission coefficients of the slab, while a and b are constants determined from the boundary conditions. The quantity /NSdenotes the photon flux and S is the time-averaged Poynting vector. ‘i’, ‘r’ and ‘t’ mean incident, reflected and transmitted, respectively. Now, the important conclusions are: (1) Thus the left side of Eq (7.1) describes the transfer of momentum to the embedded slab employing the incident, reflected, and transmitted Mink (out) pk , respectively. The left part of Eq (7.1) can be derived from the Minkowski’s formulation as shown in [216]. In chapter 5, we have already discussed in details that except the external Minkowski ST, other formulations do not support the accurate calculation of time averaged total ‘external force’. (2) In contrast, the right side of Eq (7.1) represents the total internal MEL or Modified Chu force felt by the slab, obtained by employing only its internal field. Without modifying the well-known EL or Chu force, it is not possible to obtain the equivalent results [more details discussed in chapter 5]. Though, the time averaged internal Minkowski ST leads to zero bulk force, the surface force part (obtained by Helmholtz force) also leads to the total force value equal to the one shown in Eq (7.1) [for details on Helmholtz force based calculations cf. the examples given in Ref. [215]]. So, above simple example clearly points the fact that: in time averaged situation, the Minkowski force remains valid at least at (i) the surface of the object [i.e. Helmholtz force] and (ii) outside the object [i.e. the background medium]. (iii) For the last part: Chapter 7 Consistency of Minkowski’s Theory: Some Final Proposals 127 the bulk force part of Minkowski’s force, one needs to consider absorbing slab. Such an example has already been shown in the last section of Ref. [215]. All the above mentioned proposals can be verified in an implicit way: based on the ‘validity condition’ [denoted as C(I) in chapter 5] proposed in chapter 5. As a result, now, we shall show that in instantaneous and in time average case, Minkowski force remains valid everywhere in space. The detail calculations are given below. Even from the physical point of view, time averaged Minkowski ST can be applied both inside and outside the object [44] and hence free from the restriction of ‘existence domain’. In fact, the calculations given below also suggest this same fact. The surface force of Minkowski / Helmholtz volumetric formulation: Surface22Helmholtz1122r a fEH (7.2) Now, from the non-diagonal (ND) components of Minkowski ST we get: outoutininm outoutm ininq oMIXMIXMiutoutq ininoutoutininmnkMinkoutoutm ininˆˆ[][]ˆ[(out)in ]ˆ[]ˆˆ[][]bsbsr absr absbsEEEEn +EEEEmTTnEEEEqHHHHn +HHHHq outoutq inin ˆ[]bsr amHHHHq (7.3) ‘MIX’ represents the mixed diagonal and non-diagonal elements of the stress tensor, which are not connected with the identity tensor, I . b and b are fixed background permittivity and permeability, and s and s are fixed permittivity and permeability of the scatterer. ‘out’ represents the total fields (incident plus scattered field) outside a scatterer. ‘in’ represents the fields inside a scatterer. Electric field at the object and background boundary is defined as:mqˆˆˆn E = E nE mE q where ˆn , ˆm and ˆq are mutually orthogonal arbitrary unit Chapter 7 Consistency of Minkowski’s Theory: Some Final Proposals 128 vectors, which are applicable for different co-ordinate systems such as Cartesian or Spherical or Cylindrical. ˆn is the local unit normal of the object surface, which is considered aligned towards the direction of wave vector direction (for simplicity). E and E are the parallel and perpendicular components of electric fields at the background and object boundary. In a very similar way, the magnetic field has also been defined. Now, by employing the electromagnetic boundary conditions in above Eq (7.4): outoutininoutouMIXMIXMtininnnikMi kˆˆˆ[(out)in ][][]r absbsTTnEEEEnHHHHn (7.4) In contrast, from the pure diagonal (D) components we get [after employing the electromagnetic boundary conditions]: outoutininm outm outminminq outq outq inq inoutoutininmDDMinkMi kountˆ[(out)in ]ˆˆ]1 2ˆˆ1 2r absbsr absbsb TTnEEEEn + EEEEm +EEEEqHHHHn+ Hm outminminq outq outq inq inˆˆsr absHHHm + HHHHq (7.5) Now, by adding Eq (7.4) and (7.5) and after doing some calculations, we get exactly Eq (7.2) at the object boundary. Our final conclusions on time averaged forces are shown at a glance in Fig. 7.2. To get a clear view of the operations physically involved with the different time averaged STs and volumetric force laws, we list them in Table-7.1. Chapter 7 Consistency of Minkowski’s Theory: Some Final Proposals 129 Fig. 7.2: The consistency of Minkowski’s and Abraham’s theory everywhere in space for time averaged cases. Our discussions in section 7.2 and later in Table-7.1 suggest that: for time averaged scenario, almost all stress tensors and volumetric force laws are restricted to the idea of ‘existence domain’ except Minkowski and Abraham STs and their corresponding volumetric force laws. Table- 7.1: Time averaged macroscopic stress tensors and volume forces: proposal of ‘existence domain’ (not applicable for Minkowski and Abraham formulations) Time averaged force Comment Minkowski MinkMinkMinkoutoutoutoutoutoutoutoutMinkMinkMinkinininininininin1(out)Re(out) .2(out)121(in)Re(in) .2(in)12at r aat r adsds********FTTDEBHBHDEIFTTD EB HBHDEI The time averaged external Minkowski ST characterizes the transfer of momentum transported by the field from any background medium to an embedded scatterer. The time averaged external Minkowski ST yields the correct (so far for almost all real experiments) total ‘outside force’ only from fields outside the embedded object with the integration at r=a+. The correct total ‘outside force’ can also be obtained directly from calculations employing PMink (i.e. by ray tracing method). Though the time averaged internal Minkowski ST yields a zero mechanical force (‘felt force’) at the interior of a lossless object, it leads to the correct internal force (bulk part of Helmholtz force) felt by the free charges for an absorbing objects. Exactly at the boundary of a generic object, the difference between the external and internal Minkowski stress tensor leads to the Helmholtz surface force. Hence, time averaged Minkowski ST (and Helmholtz Chapter 7 Consistency of Minkowski’s Theory: Some Final Proposals 130 force) should be considered valid as both external and internal ST (and force), which is fully opposite case of time averaged EL, Chu or Ampere/Nelson STs and force laws. Abraham AbrAbrAbroutoutoutoutoutoutoutoutoutoutoutoutAbrAbrAbrinininininininininininin1(out)Re(out) .21(out)[2]1(in)Re(in) .21(in)[2at raat radsds***********FTTDEEDBHHBBH+ DEIFTTD EE DB HH BBH+ DE ]*I The time averaged Abraham ST yields the same force as Minkowski ST. But the correct total ‘outside force’ cannot be directly obtained from the PAbr (i.e. by ray tracing method). The difference between Minkowski’s force and Abraham’s force arises only in instantaneous force. However, Abraham formulations are not relativistically invariant. Einstein-Laub ELELELinininininin0inin1(in)Re.212oat rads****HEFTTD EB HHEI The time averaged EL ST yields the ‘bulk force’ felt by the induced dipoles or ‘inside force’ from fields just inside the object embedded in air or vacuum. This ST does not characterize the flow of momentum transported by the field from the background medium to an embedded scatterer. Hence it is not applicable to get the ‘outside force’. For an object embedded in air or vacuum, the difference of external vacuum ST and the internal EL ST at the object boundary leads to a surface force Surfacef which exactly matches with the surface force fully independently calculated by the EL volumetric force [cf. Appendix A and C]. But when an object is embedded in a material medium, the difference of the external EL ST and internal EL ST does not lead to the surface force of EL. As a result, MEL ST should be applied to yield the internal time averaged total force and to satisfy the two consistency conditions [C(I) and C(II)] presented in chapter 5. Chu ChuChuChu0inininininin0inin1(in)Re(in) .2(in)1 2ooat r ads****EHHEFTTEHHEI Time averaged Chu ST yields the ‘bulk force’ felt by the induced charges and currents or ‘inside force’ from fields just inside the object embedded in air or vacuum. This ST does not characterize the flow of momentum transported by the field from the background medium to an embedded scatterer. Hence it is not applicable to get the ‘outside force’. For an object embedded in any material medium, the difference of external vacuum ST and the internal Chu ST at the object boundary leads to a Surfacef which exactly matches with the surface force fully independently calculated by the Chu volume Chapter 7 Consistency of Minkowski’s Theory: Some Final Proposals 131 force [cf. next section, section 7.3, of this conclusion and Appendix A]. But both the internal and external Chu ST do not lead to the time averaged consistent force for an embedded object. Only by modifying the internal Chu formulations along with its surface force, it is possible to yield the time averaged total force on an embedded object. Ampere/ Nelson AmpAmp0inin0inin220i10nin11(in)Re(in) .2(in)1 2at r ads**FTTE E+μ BE+B BI The Nelson ST and force law behave exactly same as the Chu formulation for a dielectric object placed in air. For magneto-dielectric objects, it behaves differently. This ST does not characterize the flow of momentum transported by the field from the background medium to an embedded scatterer. Hence it is not applicable to get the ‘outside force’ very similar to EL and Chu ST. Modified Einstein-Laub MELMELMELinininininininin1(in)Re(in) .2(in)1 2at rabbds****HEFTTD EB HHEI MEL ST is similar to time averaged EL ST but when the background is material medium instead of air. For Air background, the MEL ST turns into EL ST. The time averaged MEL ST yields the ‘bulk force’ felt by the induced dipoles or ‘inside force’ from fields just inside the object embedded in material medium. This ST is also not applicable to get the ‘outside force’. For an object embedded in material medium, the difference of external Minkowski ST and the internal MEL ST at the boundary of that embedded object leads to a version of Surfacef which exactly matches with the surface force fully independently calculated by the MEL volumetric force [cf. Appendix C]. Also note that: similar arguments are applicable for internal modified version of Chu ST for objects embedded in material medium [cf. our discussion in chapter 5]. 7.3 Instantaneous and Time averaged Optical Force: Accuracy of Minkowski’s theory Chapter 7 Consistency of Minkowski’s Theory: Some Final Proposals 132 Fig. 7.3: Consistency of Minkowski’s and Chu’s theory for instantaneous and Time averaged force. Our discussions in section 7.3 and later in Table-7.2 suggest that: for instantaneous scenario, the consistency of only Minkowski’s and Chu’s theory remain everywhere in space based on: (i) consistent surface force calculation [described as ‘validity condition’ C(I) in chapter 5] and (ii) Relativistic invariance. In previous section we have discussed that except Minkowski’s and Abraham’s stress tensors and volumetric force laws, all other force formulations are restricted with the idea of ‘existence domain’ for the case of time averaged force calculation. In addition, to yield the time averaged total force on an embedded object, we require to modify the Einstein-Laub, Chu and Ampere formulations. However, though the time averaged form of the Einstein-Laub and Chu formulations require modifications inside any embedded object, for the instantaneous case, well-known form of Chu force remains consistent everywhere (i.e. inside the embedded object, at the surface of the object and also at the background) among these two forces. When we consider an object embedded in a material medium, the surface force of Chu can be determined from the well-known volumetric force of Chu by appropriate boundary conditions [182]: outinoutiSurfacenoutinouChu0tin0ˆˆ()()22r ar aEEHHfEEnHHn (7.6) Chapter 7 Consistency of Minkowski’s Theory: Some Final Proposals 133 Now, from the non-diagonal (ND) components of Chu ST we get: outoutininm outouMIXMIXVacuumChu00000000tm ininq outoutq ininoutoutinin0m outˆ[(out)in ]ˆˆ[][] ˆ[]ˆ[][r ar aTTnEEEEn +EEEEmEEEEqHHHHn +HHoutm ininq outoutq inin000ˆ]ˆ[]r aHHmHHHHq (7.7) Now, by applying the electromagnetic boundary conditions, we finally arrive for non-diagonal components: outoutininm outm inoutinq outq inoutMIXMIXVacuumChu00000inˆ[(out)in ]ˆˆ[][]2 ˆ[]2[ r ar aTTnEEEEn +EEEEmEEEEqHoutoutininm outm inoutinq outq 00inoutin0ˆˆ][]2ˆ[]2r aHHHn + HHHHm HHHHq(7.8) Now, we can also write for diagonal components: outoutDDVacuumChu0000ininoutoutinin1ˆˆˆ[(out)in ]2r ar a TTnEEEEn+HHHHn (7.9) ‘D’ represents pure diagonal elements of the stress tensor. Now, by adding Eqs (7.8) and (7.9), after some calculations we get exactly Eq (7.6), the surface force of well-known Chu force. But this type of matching does not happen for the Einstein-Laub volumetric force considering embedded object. It is also notable that both Chu and Minkowski stress tensor and their volumetric force laws are relativistically invariant as shown in [217, 218]. On the other hand, Chapter 7 Consistency of Minkowski’s Theory: Some Final Proposals 134 Abraham’s and Einstein-Laub’s force laws are not relativistically invariant [217-219]. Now, our final conclusions on instantaneous forces are shown at a glance in Fig. 7.3. In the Table-7.2, we have discussed in details the differences between time averaged forces and instantaneous force formulations. Table -7.2: Optical force laws: difference between instantaneous forces and their corresponding time averaged forces Stress tensor T and Force density, f Electromagnetic momentum density, G Comment Minkowski 22121122 TDEBHB HD E IfEH DB (1) Let us consider at first that a lossless glass object is embedded in water. Helmholtz/Minkowski’s volumetric force distribution is applicable at each point of space. But in continuous medium (if lossless), it vanishes even in instantaneous scenario and appears only at the boundary of the object. But this does not mean Minkowski force is not applicable everywhere. For example- Now, we consider the glass object is absorbing. Then a local/bulk force part exists for Minkowski force along with the Helmholtz’s surface force. This bulk force describes how much momentum is transferring ‘only’ to the free carries due to local bulk fields of that glass object, which can also be obtained from internal Minkowski ST at r=a-. Now, the interesting point is that: the time averaged external Minkowski ST at r=a+ yields the time averaged total force [Helmholtz surface force plus the bulk force on free carriers] on that embedded absorbing glass object just by employing the external fields of the scatterer (fields from water background). In addition, exactly at the object boundary, the difference of internal Minkowski ST and the external Minkowski ST leads to that surface force of Helmholtz. (2) When absorption takes place in background, a local/bulk force arises in Minkowski’s force connected with the local fields in the background; Chapter 7 Consistency of Minkowski’s Theory: Some Final Proposals 135 which represents how much momentum is transferring ‘only’ to the free carriers of the background (but not to the embedded glass object). - Hence Minkowski force has different physical operations in different spatial regions such as inside the absorbing embedded object; at the interface of the background and embedded object; and at the absorbing water background. But still it is applicable everywhere in space without any modification and also continuous for the case of instantaneous force (also relativistically form invariant). Hence Minkowski ST/ force is consistent everywhere in space for both cases. Abraham 22221211;221AAntc DEEDTBHHBB H + D E IfEHfEHf 2cEH The difference between Minkowski’s force and Abraham’s force arises only for instantaneous force due to Af term in Abraham’s force. Though in time averaged case Abraham force law is consistent everywhere in space, it is not relativistically invariant. Hence for instantaneous force, it cannot be considered consistent. Einstein-Laub 001 ()2()(··)oottT = DEBHH HE E IfPEMHPMHE 2cEH (1) If Einstein-Laub force is applied to a lossless glass object (embedded in water background) employing the local field of the object, the EL force inside the object does not represent the ‘correct’ magnitude of bulk force felt by the object [cf. the discussion after Eq (7.1) in section 7.2 of this conclusion]. It is the modified Einstein-Laub volumetric force [or MEL ST; cf. chapter 5] which should be applied to yield the time averaged bulk force of an embedded object. But it should be noted that the MEL version of volumetric force may not the correct form of the force distribution for instantaneous force, as it creates some important problems with Einstein-Balazs thought experiment [220]. (2) For an object embedded in air or vacuum, the difference of external vacuum ST and the internal EL ST at the object boundary leads to a Surfacef which exactly matches with the surface force fully independently calculated by the EL volumetric force. But the difference of external and internal EL ST does not lead to the consistent Surfacef for an object embedded in material medium. For such an object, the difference of external Minkowski ST and the internal MEL ST at the boundary of that embedded object leads to a version of Surfacef which exactly matches with the surface force fully independently calculated by the MEL volumetric Chapter 7 Consistency of Minkowski’s Theory: Some Final Proposals 136 force [cf. Appendix A and C]. (3) On the other hand, the well-known Einstein-Laub force in the water background employing the local field of the water represents the bulk force felt by the water background itself. It does not represent how much momentum has been transferred from the water background to the embedded object. - So, based on (1)-(3), we conclude that: in time averaged scenario: the MEL force should be applied inside the embedded scatterer [cf. chapter 5]. On the other hand, the well-known time averaged EL force should be applied in the background medium to determine the force felt by the background water medium. For the case of instantaneous force: EL force is not relativistically invariant. Hence EL ST/force is not consistent everywhere in space for both cases. Chu 00000··1 ()2 ()()ott EEHHH HE E IEH HETfPP 2cEH The conclusions for Chu force are very similar to the Einstein-Laub force except one difference: the instantaneous force behavior. Instantaneous Chu force is more physical than the instantaneous EL force due to its nature of relativistic invariance. So, from the physical point of view, it can be suggested that the instantaneous volumetric force distribution based on the bound charges and currents of Chu formulation is more accurate than the induced dipole based EL formulation. In chapter 5 we have highlighted the importance of both the time averaged modified EL and modified Chu force laws. Though instantaneous Chu force remains consistent everywhere, in time averaged scenario it does not remain consistent everywhere. For example- though the time averaged force felt by the water background should be calculated by well-known Chu force, the time averaged force on the embedded scatterer should be calculated by its modified version (very similar behavior of time averaged scenario of EL/MEL force). Ampere/ Nelson 0020112012()()t TEEBB B +EIPfP EB M+μB 0EB The Nelson/Ampere ST (or its modified version) and force law behaves exactly same as the Chu formulation (or its modified version) for a dielectric scatterer. Chapter 7 Consistency of Minkowski’s Theory: Some Final Proposals 137 7.4 Photon Momentum for instantaneous and Time averaged cases: Accuracy of Minkowski’s theory Based on the discussions in the previous sections [sections 7.2 and 7.3], finally, a question arises: can we model the momentum of photon as the one of Minkowski everywhere in the space? It is so far understood that the travelling momentum of photon (i.e. at the background) can better be modelled as Abraham/kinetic photon momentum [151] and the transferred momentum from the background to the embedded object should be considered as Minkowski/canonical photon momentum [43,151]. Classically Minkowski photon momentum arises just at the interface of the background and the embedded object [162]. Also the well-known Snell’s law at the interface of two different mediums can be derived only from the conservation of the tangential component of Minkowski photon momentum. So far almost all the real experiments [43,58] have strongly supported the Minkowski photon momentum as the physical momentum of photon. Our analysis throughout the thesis suggests that either in instantaneous or in time average case, Minkowskian formulations remain valid everywhere in space and so far all the major radiation pressure experiments have unambiguously supported Minkowski’s force formulations [28,46] and photon momentum [43,58]. One of the main arguments in favor of Abraham photon momentum mainly arises from the validity of Einstein-Laub’s volumetric force distribution, which is connected with Abraham photon momentum, in case of Ashkin’s experiment on water surface [161] (in other words: from the volumetric force calculations those support Abraham momentum density [28,151]). But Einstein-Laub formulation may not be appropriate to describe the physical reality considering both instantaneous [218] and time averaged analysis [cf. our discussion in previous two sections and chapter 5]. At the same time, such experiments [28,161] can also be explained better based on the Minkowskian force and Minkowski photon momentum Chapter 7 Consistency of Minkowski’s Theory: Some Final Proposals 138 [28,221]. On the other hand, the well-known Chu force (that also supports Abraham photon momentum), which remains valid everywhere in instantaneous scenario [but not in time averaged case], does not lead to the correct force distribution for Ashkin’s experiment on water surface [28]. One recent study has also invalided the idea of Abraham photon momentum even in instantaneous scenario [222]. Our final conclusions on photon momentum are shown at a glance in Fig. 7.4. Fig. 7.4: Consistency of Minkowski photon momentum. Based on our discussions in this conclusion chapter [specially cf. Fig. 7.2 and 7.3], it can be understood that Minkowski’s force formulations remain consistent everywhere in space. This fact implicitly suggests that: for both time averaged and instantaneous cases, Minkowski photon momentum can also be considered as the consistent photon momentum everywhere in space. Ray tracing method, Snell’s law, almost all real experiments on radiation pressure directly support only the Minkowski photon momentum as the observable physical momentum of photon. On the other hand, even if it can be considered that Abraham photon momentum is the travelling momentum of photon in a continuous medium, most probably there is no way to reject the idea that: Minkowski photon momentum arises at the boundary of the embedded objects [cf. chapter 4 of this thesis and in chapter 5 it is also shown that the GAP METHOD, which is modelled to support Abraham’s photon momentum everywhere in space, leads to inconsistent result of time averaged total force]. One recent study has invalided the idea of Abraham photon momentum even in instantaneous scenario [222]. As a result, for a unified consistent theory of optical force and photon momentum, either in instantaneous or in time averaged case, most probably Minkowski photon momentum is the strongest candidate for appropriate photon momentum as shown in above figure. Chapter 7 Consistency of Minkowski’s Theory: Some Final Proposals 139 7.5 Conclusion This chapter and Chapter 5 of this thesis partially answer the question after almost 100 years of Einstein’s prediction [223] why Minkowski’s theory is more accurate than Einstein-Laub theory to obtain the time averaged total force [due to: (i) ‘existence domain’ and (ii) modification issues]. Based on the proposals in chapter 5, we have proposed the idea/hypothesis of ‘existence domain’ in this chapter [section 7.2]. Though almost all the time averaged distinct STs and volumetric force laws are restricted to the idea of ‘existence domain’, we demonstrate that the time averaged stress tensor and volumetric force law of Minkowski are free from such restrictions. In addition, we have discussed in details the differences between time averaged and instantaneous force laws for different formulations in this chapter [section 7.3]. Notably only Minkowski’s theory remains valid for all circumstances without any modification. So, after almost 100 years of Einstein’s prediction [223], this chapter, most probably finally answers why Minkowski’s theory is more accurate for all circumstances. Considering ‘all’ the proposals in this chapter, it appears that Minkowski formulations are most probably more accurate description of physical optical force along with photon momentum. So, according to our proposals in this chapter, it may also be possible to consider the presence of Minkowski photon momentum everywhere in space along with the interface of the two different medium. However, if it would be possible to experimentally measure the presence of bulk force distribution of any ideal non-absorbing medium by any future experiment, it would certainly prove our proposals in this chapter physically a little bit inaccurate (as no bulk force should take place for Minkowski formulations in case of non-absorbing medium). However, whatever the experimental observation in future, from physical point of view: Minkowski’s force most probably always describes the transfer of Chapter 7 Consistency of Minkowski’s Theory: Some Final Proposals 140 optical momentum from background to an embedded object (external Minkowski ST) or to the embedded free charges (internal Minkowski ST). Chapter 8 Conclusion and Future Works 141 Chapter 8 Conclusion and Future Works 8.1 Summary of the Thesis Radiation pressure, in conjunction with photon momentum, has always constituted an intriguing phenomenon in physics. An accurate prediction of the electromagnetic force is important for complex biological systems, stable optical manipulation, tractor beams, MEMs and nano-opto-mechanical systems, among many applications. Most experiments in those areas need determinations of these optical forces on objects immersed in a non-vacuum environment. However, inside matter different definitions and descriptions of the macroscopic optical force and the photon momentum have been put forward, among others by Minkowski, Abraham, Ampere/Nelson, Einstein-Laub, Chu or Peierls. Even after extensive debates spanning over a century, there are still several unsolved problems regarding exactitudes of stress tensors or force laws, and on their individual validities or limitations inside matter. One of the main goals of this thesis was to investigate the problems/dilemmas of optical volumetric/Lorentz forces specially observed in real radiation pressure experiments performed up to date. Moreover, Lorentz force dynamics was previously restricted to simple cases. In literature, Lorentz force analysis for optical pulling force (tractor beams) and the reversal of near field optical binding force (i.e. for plasmonic objects) have not been investigated in details so far. As a result, in this thesis, physics and applications of Lorentz force have been investigated in details with special Chapter 8 Conclusion and Future Works 142 interest in tractor beam effect, plasmonic objects, chiral objects and the objects embedded in a generic material medium. At the beginning part of this thesis, in chapters 2 and 3, we have considered the Lorentz force dynamics for comparatively simple cases such as objects placed in air. Time averaged total force calculated by well-known Lorentz force and the external stress tensor method are found in full agreement for those two chapters without any modification in Lorentz force. More specifically, as the stimulating connection between the Lorentz force and optical pulling force has not been investigated in details previously, chapter 2 demonstrates the Lorentz force analysis and its applications for tractor beam like effect [214]. It is shown based on Lorentz force analysis that the bound surface charges of Rayleigh scatterer experience backward force, which overcomes the positive bulk force and ultimately results in the net pulling of the scatterer for several spatial regions outside two dielectric hollow core cylindrical waveguide. Later, this idea of dielectric waveguide tractor beam has been extended for dielectric coupler set-ups for stable transverse equilibrium and long distance operation. On the other hand, Chapter 3 demonstrates the Lorentz force analysis for plasmonic off-axis and on-axis spherical heterodimers. It is shown based on Lorentz force analysis that the reversal of longitudinal binding force can be easily controlled through forced symmetry breaking: (i) just by changing the direction of light/wave propagation for a specific set-up of off-axis heterodimers or (ii) just by changing their relative orientation while keeping the light/wave propagation in a fixed direction. Notably, the longitudinal binding force of spherical heterodimers originates almost fully from the difference of the bulk part of Lorentz force, which strongly suggests the connection of bulk Lorentz force with the scattering force. As much study has not been performed on the connection of Lorentz force with gradient and scattering forces, the work presented in chapter 3 may open a new window for such investigations. Chapter 8 Conclusion and Future Works 143 In contrast, in chapters 4, 5, 6 and 7, we have considered much complicated cases such as chiral and achiral objects embedded in material medium. Specially chapters 4 and 5 of this thesis deal with very fundamental topics of optical force: consistency of stress tensors, volumetric forces and photon momenta inside matter. In chapter 4 it is shown that the well-known Lorentz force leads to inconsistent result (pushing force) instead of the experimentally observed optical pulling force for interfacial tractor beam experiment [43]. Our suggested interpretation in chapter 4 supports the Minkowski approach only for the purpose of optical momentum transfer to the embedded scatterer for the interfacial tractor beam experiment [35], although the momentum of photon in the continuous water background can be considered as the type of Abraham for the calculation of the bulk part of Lorentz volumetric force distribution. Later chapter 5 discusses in detail why the well-known Lorentz/volumetric forces fail not only in interfacial tractor beam experiment but also for several other real experiments. Exactly at the boundary of an object, the difference of the external Minkowski ST and internal ST of Chu (and Einstein-Laub) is found in agreement with the surface force yielded by Chu (and Einstein-Laub) force only when the background is air or vacuum rather than a material. We identify this as one of the main reasons of the disagreements observed in the major radiation pressure experiments which include material medium as background. Some other notable reasons of such disagreements have also been identified and discussed in that same chapter. We also demonstrate that it is still possible to establish different equivalent time-averaged Lorentz / volumetric force formulas based on the fulfilment of just two ‘consistency conditions’. Based on those proposed ‘consistency conditions’, finally, we propose that by modifying the Einstein-Laub (EL) or Chu formulation, time-averaged STs and volume forces are obtained those can overcome the aforementioned inconsistency for almost all previous experiments. Specially, our proposed modified Einstein-Laub ST can be considered as an efficient mathematical toolkit, an Chapter 8 Conclusion and Future Works 144 alternative of time and memory consuming volumetric forces, to yield the internal bulk force of a chiral or achiral object embedded in complex material backgrounds (i.e. homogeneous, heterogeneous, bounded etc.). In chapter 6, we have shown an interesting application of modified Lorentz force, i.e. modified Chu force. We have observed that if the closely located plasmonic cube homodimers (over glass or high permittivity dielectric substrate) are illuminated with plane wave polarized parallel to dimer axis, no reversal of optical binding force occurs. But if we apply the same set-up over a plasmonic substrate, stable and stronger Fano resonance occurs along with the reversal of binding force. Interestingly, for both the strong quadrupole-quadrupole and dipole-dipole resonances of the dimers, the binding force is dominated by Coulombic surface charge induced force; whereas during the Fano resonance binding force reversal is dominated by polarization current induced modified bulk Lorentz force (which is unconventional in comparison with the usual idea of optical force on plasmonic objects). Finally in chapter 7 (the chapter before conclusion), we have proposed the hypothesis of ‘existence domain’ inspired by the idea presented in chapter 5. Though almost all the time averaged distinct STs and volumetric force laws are restricted to the idea of ‘existence domain’, we demonstrate that the time averaged stress tensor and volumetric force law of Minkowski are free from such restrictions. In addition, we have discussed in details the differences between time averaged and instantaneous force laws for different formulations in chapter 7. Notably only Minkowski’s theory (stress tensor, force law and photon momentum) remains valid for all circumstances without any modification. The main contributions of this thesis are the following: Firstly, to identify the exact reasons of the failure of distinct time averaged optical force formulas in real radiation pressure experiments performed up to date and to settle them. Secondly, the consistency conditions [C(I) and C(II) in chapter 5] of optical forces , which lead to Chapter 8 Conclusion and Future Works 145 the equivalent formulations of several distinct time averaged optical forces applicable for objects embedded in generic backgrounds [i.e. homogeneous, heterogeneous, bounded and so on]. Thirdly, the hypothesis of ‘existence domain’ in chapter 7, which along with aforementioned consistency conditions lead to sort out the most probable consistent theory of optical force and photon momentum applicable everywhere in space (i.e. Minkowski’s theory) both in instantaneous and time averaged scenario. Fourthly, the applications of Lorentz force and modified Lorentz force dynamics for tractor beam effect and plasmonic dimers, which lead so several unconventional conclusions not presented before in literature. 8.2 Final Remarks and Future Works In 1918 Albert Einstein wrote against his own theory of optical force to Walter Dällenbach [223]: “It has long been known that the values I had derived with Laub at the time are wrong; Abraham, in particular, was the one who presented this in a thorough paper. The correct strain tensor has incidentally already been pointed out by Minkowski”. Unfortunately, even after almost 100 years, it is not clear why Einstein predicted so regarding the theory of optical force generally known as optical stress tensor associated with optical Lorentz force. This thesis (especially chapters 5 and 7), most probably, finally answers why Minkowski’s theory is scientifically more accurate for all circumstances. Hence, it appears that Minkowski formulations are most probably more appropriate description of physical optical force along with photon momentum. It may also be possible to consider the presence of Minkowski photon momentum everywhere in space along with the interface of the two different medium according to previous chapter (chapter 7). However, if it would be possible to experimentally measure the presence of bulk Chapter 8 Conclusion and Future Works 146 force distribution of any ideal non-absorbing medium by any future experiment, it would certainly prove our proposals in favor of Minkowski’s theory a little bit inaccurate physically (as no bulk force should take place for Minkowski formulations in case of non-absorbing medium). However, whatever the experimental observation in future, from physical point of view: Minkowski’s force always describes the transfer of optical momentum from background to an embedded object (external Minkowski ST) or to the embedded free charges (internal Minkowski ST). According to the conclusion of chapter 5, if time averaged bulk force is found not zero for a non-absorbing medium (i.e. in any future experiment), then it should better be considered that: (i) both time averaged external Minkowski force and time averaged internal modified Chu (or modified Einstein-Laub) force are physical entities but (ii) their physical operations should be considered quite distinct from each other [for details cf. the conclusion of chapter 5]. And hence the forms of the optical force law can be fully different to describe these fully different physical process/physical operations. In addition, if bulk force distribution for non-absorbing medium will be experimentally observed in real time analysis, then the idea of travelling momentum of photon as the one of Abraham [cf. our discussion in chapter 4] may not be neglected. As a result, such a future experiment on bulk force (especially in real time analysis), if properly proceeds, will probably resolve many misconceptions and dilemmas in the area of distinct optical forces along with photon momenta. This may be possible by introducing much modifications in the recent experiments of ‘Ashkin-Dziedzic type’ [47] as discussed in ref. [221, 224, 225] and by measuring the state of bulk part of water (or any ideal non-absorbing liquid) instead of the water surface. Proposals presented in this thesis can be very effective to resolve not only the dilemma of distinct stress tensors and optical Lorentz forces but also the controversy of Abraham-Minkowski photon momenta. In addition, so far the idea of tractor beams are mostly restricted with highly sophisticated structured beams or exotic artificial Chapter 8 Conclusion and Future Works 147 materials [226]. This thesis may also open a new window of optical pulling force/tractor beams due to the exclusion of conventional structured tractor beams along with the artificial exotic matters. Last but not least, the proposed ideas for the reversal of near field optical binding force may also be useful for the future applications of plasmonic dimers in the areas of improved plasmonic sensors, particle clustering and aggregation. Appendices 148 APPENDICES Appendix A Detail analysis on GAP METHOD of optical force calculation Before discussing the volume force method of GAP METHOD (defined in the introduction of chapter 5 and also cf. Fig. 1A below), we highlight the surface force determination of Chu force. It is already shown in [182] that when we consider an object placed in air, the surface force of Chu can be determined from the appropriate boundary conditions [78, 173, 182] of Chu volumetric force: SurfaceChu00aoutit r=aanoutinoutinoutritn=aˆˆ()()2A21EEHHfEEnHHn However, in [182] the surface force has not been calculated by employing the stress tensors. We shall show the process of surface force calculation by employing the stress tensor method in this section. The internal Chu stress tensor (ST) can be written as: in****Chu0inin0inin0inin0inin12TE EH HHHEEI. Here ‘in’ represents the fields inside a scatterer. Now, we shall show that the surface force of Chu in Eq (A1) can be yielded from the difference of external vacuum stress tensor Vacuum(out)T and Appendices 149 the internal Chu ST [cf. Eq (5.2b) in chapter 5 for vacuum ST; for the case of air background: the gap field (gE ) itself is the outside field (outE) in Fig. 1A given below]: outoutininm outoutm ininqMIXMIXVacuumChu0000000000outoutq ininoutoutininm outoutm iniu0nq o tˆˆˆ[(out)in ][][]ˆ[]ˆˆ[][][r aTTnEEEEn +EEEEmEEEEqHHHHn +HHHHmHoutq inin0ˆ]HHHq ‘MIX’ represents the mixed diagonal and non-diagonal elements of the stress tensor, which are not connected with the identity tensor, I . ‘out’ represents the total fields (incident plus scattered field) outside a scatterer. Electric field at the object and background boundary is defined as:mqˆˆˆ E = E nE mE q where ˆn , ˆmand ˆq are mutually orthogonal arbitrary unit vectors, which are applicable for different co-ordinate systems such as Cartesian or Spherical or Cylindrical. ˆn is the unit normal of the object surface. E and E are the parallel and perpendicular components of electric fields at the background and object boundary. In a very similar way, the magnetic field has also been defined. Now, by applying the electromagnetic boundary conditions, we finally arrive for mix components: outoutininm outm inoutinq outq inoutinoutoutininm outm 000MIXMIXVacuumChu00o0nut0iˆˆ[][]2ˆ[(out)in ]ˆ[]2ˆ[][2r ar aEEEE n +EEEE mTTnEEEE qHHHH n + HHHinq outq inoutin0ˆ]ˆ[]2r aH m HHHH q (A2) Now, we can also write: Appendices 150 outoutDDVacuumChu0ininoutoutinin000ˆˆˆ[(out)in ][][]r ar aTTnEEEEn+HHHHn (A3) ‘D’ represents pure diagonal elements of the stress tensor. Now, by adding Eqs (A2) and (A3), after some calculations we get exactly Eq (A1), the surface force of Chu: outinoutVacuumChu00ainoutinouttinr=aat r=aˆˆˆ[(out)in ][()()]22r aEEHHTTnEEnHHn (A4) So, when the object is placed in air: (i) internal Chu stress tensor yields the time averaged bulk force of Chu volume force and (ii) the external vacuum ST leads to the total time averaged force (bulk force plus the additional surface force). This is also true for the Einstein-Laub volumetric force. We have found analytically that the difference of external vacuum ST and the internal Einstein-Laub ST leads to a surface force of Einstein-Laub [173,182], which exactly matches with the independently calculated surface force from the volumetric force of Einstein-Laub [182]. Fig. 1A: When the object is surrounded by a medium other than the free space, a narrow gap may be imagined to exist between the object and its surroundings according to refs. [170, 78,172,173]. The integration boundary is then placed within the gap. gE represents the electric field in that narrow gap. Now, we can go back to our discussion on the topic of GAP METHOD [170, 78,172,173] as illustrated in Fig. 1A. In [78] considering the consistency of the Appendices 151 volume force, finally it has been concluded that: When such a gap is given between the scatterer and the surrounding background, the total force of the embedded scatterer should be calculated based on (described as method II in [78] and cf. Fig. 1A): TotalBulkSurfaceGap Method Gap Method Gap Method in F= FF (A5) *SurfaceoutinoutinGap Method0gin0ginat r=aat r=a1ˆˆRe[()()]222EEHHfEEnHHn (A6) Bulk**Gap Methodinin1inRe.2iifPBMD (A7) Here 0inSPE , 0inSMH , ininSBHand iinnSDE. But the bulk and surface force calculation based on the above method may not be a general procedure due to these reasons: (1) Though above Eq (A5) is successful for small object only for few specific cases [78], it may lead to inconsistent result for embedded Mie or more complex objects according to the same ref. [78]. (2) The surface force of Chu force calculated in [182] from the direct volume force method. It is also shown in Eqs (A1-A4) that the same surface force can also be calculated for a dielectric object (placed in air) from the difference of the external ST, Vacuum(out)T, and the internal ST of Chu, Chu inT,just at the boundary of the scatterer. So, the total force of a scatterer placed in air calculated by volume force method is nothing but the force calculated by the external ST. Appendices 152 Now, we again consider the work of Barnett et al. [182] where the procedure of surface force calculation has been illustrated by properly applying the boundary conditions. If we apply those boundary conditions and at the same time consider that there is a small gap between the scatterer and the surrounding background, the consistent equation leads to the force equation where the average field at the boundary of the scatterer should be calculated with gap fields (gE ,gH ) instead of the real fields (outE,outH) of the background medium (described as method III in [78] and cf. Fig. 1A): *ginginSurfaceGap Method0gin0ginat r=aat r=a1ˆˆRe[()()]222EEHHfEEnHHn (A7) It is also interesting that if we consider a vacuum stress tensor at the gap [170] shown between the scatterer and the background, the difference between the external vacuum stress tensor and the internal Chu stress tensor leads to the same surface force of Eq (A7): ggGAPVacuumChu0g0gat r=aat r=aininininˆˆˆ[(out)in ][()()]22r aEEHHTTnEEnHHn (A8) But that equation (method III in [78]) does not lead to the consistent result according to that same ref. [78]. In addition, we have already shown in chapter 5 that the total force calculated by external ST GAPVacuum(out)Tin Eq (A8) [method III in [78] should also lead to the same force of Eq (5.2a) given in chapter 5] leads to wrong result for experiments which include inhomogeneous background (or in general heterogeneous background). Appendices 153 So, Eq (A5) above is an independent equation (described as method II in [78] which is a more accurate formulation than method I and III according to ref. [78]) which cannot be derived from the volume force of Chu [182] by employing the proper boundary conditions to the static force part shown in [182]. At the same time, there is no such external (and also internal) ST which leads to the surface force yielded by Eq (A6) above from the difference of external ST and the internal Chu ST just at the object boundary. (3) Last but not least, the difference of any external ST and the internal Chu ST does not lead to the surface force given in Eq (A6). So, Eq (A5) above (described as method II in [78]) is a fully independent equation to yield the time averaged total force, which may not be a generic way to yield the total time averaged force of the embedded objects. However, an alternative way to yield the time averaged total force (by volume force method) for the embedded objects can be well-known Einstein-Laub force [28] as shown for several previous experiments [28] but without considering any small gap between the scatterer and the background. However, the problems of EL force for real experiments, those involve material background, have been discussed shortly in chapter 5. Appendices 154 Appendix B Possible derivation of the internal Modified Einstein-Laub stress tensor Previously the Eisntein-Laub (EL) stress tensor was derived in [227] from the well-known EL force density method. In this section we shall derive the modified EL ST from the modified EL (MEL) force in the same way as described in [227] but in the time averaged form. For an embedded achiral dielectric object embedded in another homogeneous dielectric, the EL force density takes a modified form [171]: **MELEffinEffin1 Re.2ifPEPB (B1) In Eq (B1), it is interesting to note that the polarization have been written as effective polarization as [26,171,183]: inEffbinEffEffbin; ,SDDEPPE (B2) In the same way, the effective magnetization can also be written as [26]: inEffbinEffEffbin; .SBBHMMH (B3) It should be noted that b and b are fixed constant parameters [For example- they do not vary with respect to time]. In [183] it is argued in favor of Eq (B2): “Note that the equation for net force without rescaling with the background permittivity will give the total force on the volume including force on the background material which Appendices 155 does not result in motion of the particle.” The idea of effective induced electric (and magnetic) dipole moment has been also applied previously to yield the ‘outside force’ of dipolar objects in [16, 97, 186] in a very similar way of the idea of effective polarization and magnetization given in Eq (B2) and (B3). For example- the induced effective dipole moment has been derived in Eq (22) given in ref. [97] from the effective polarization model of Eq (B2) but with the exterior field of an embedded scatterer for external dipolar force. However, in [26], it is shown that for a magneto-dielectric object embedded in another magneto-dielectric, the external force by Minkowski ST matches with the time averaged force via Modified Chu (MCHU) volume force expressed in terms of effective polarization and magnetization as: MCHUMCHUMCHUTotalSurfaceBulkfff (B4) **MCHUSurfaceat r=aatoutinoutinoutinoutinr=a1ˆˆRe[()()]222bbEEHHfEEnHHn (B5) MCHU**BulkEffinEffin1Re[].2iifPBMD (B6) For several cases [26,183, 228, 229], this volume force method has been applied to yield the total force. For example- the effect of background permittivity has been described in [229] for the cases of binding forces. Considering the consistency of Eqs (B4-B6) [26], [183], the modified EL force law in [171] can be written in a generic form for a magneto-dielectric object embedded in another magneto-dielectric in accordance with [161] as: Bulk*****MELEffinEffinEffinEffEffinEff1(in)Re.2iifPEMHPBMMDP Appendices 156 (B7) In Eq (B7), the *EffEffiPM terms cancel out finally. As it is shown in [188] that the application of EL formulation may lead to misleading results for very simple situations, the final time averaged form of MEL force should be written in terms of B instead of 0 H according to [161] (also cf. reference [230] where the difference between the time varying and the time averaged force has been argued): Bulk****MELEffinEffinEffinEffin1(in)Re.2iifPEMHPBMD (B8) Now, in order to yield the final form of the MEL stress tensor, the intermediate step of the MEL force [not the final time averaged form given in Eq (B8)] can be written from Eq (B7) as: BulkMELEffinEffinEffbinEffbin1(in)Re[()()]2iifPEMHPHME (B9) Eq (B9) ,which represents only an intermediate step of final MEL force in Eq (B8), turns into the EL force equation if the background is air. It should be noted that the Maxwell equations remain unique everywhere in terms of D and B only. However, as the internal induced polarization and magnetization are affected/influenced by the background permittivity and permeability specially due to the effect of the boundary/interface, the division of D and B can be expressed based on Eq (B2) and (B3).As a result, considering PEff and MEff in Eq (B2) and (B3) as a consistent mathematical description, the Maxwell-Chu [155] equations can be written for the embedded object as: ininbinEff() , ii EB+ MH (B10) ininbinEff (),ii PHDE (B11) inbinEff )0 EDP (B12) Appendices 157 inbinEff )0 B+ MH (B13) In order to determine the final form of the stress tensor, let us consider the **bbininbbininRe[()()]ii EHHE term which is zero. By adding this **bbininbbininRe[()()]ii EHHEterm with previous Eq (B9), we finally obtain: Bulk****MELEffinEffinEffbinEffbin**bbininbbinin1(in)Re[[()()]2[()() ]·]iitii TfGPEMHPHMEEHHE (B14) In above equation t G is zero where G is a non-mechanical momentum density. After some calculations, we get from Eq (B14): Bulk****MELEffinEffinbininbinin1(in2·)Re[] TfPE+ MHEEHH (B15) By applying Eqs (B2) and (B3), we can write from Eq (B15): Bulk****MELininininbininbinin111(in)Re[ ()(.)()]222· TfD EB HEEH.H (B16) Hence the time averaged MEL ST inside an object embedded in a homogeneous background can be written as: Bulk****bbMELininininsinininins11inRe[()()].22ssTD EB HHHEEI (B17) Eq (B17) turns into EL Stress tensor [227] when the embedding background is air instead of material medium. However, still one important issue remains unsolved: Appendices 158 Are Eqs (B1) [171], (B8) and (B17) consistent equations, which lead to the same surface force from two different ways of calculations (as shown at the beginning of Appendix A): (i) from direct volume force [183] and (ii) from the difference of external and internal ST. This issue is answered in the next appendix. Appendices 159 Appendix C Derivation of the MEL surface force in two fully different ways At first we consider the surface force calculation process shown in [183] (which applied in [183] to yield the surface force: (i) only from volumetric force and (ii) only for air background) and apply that process for the static part of Eq (B8) at the boundary of the embedded object (at r=a) [considering there is no artificial gap between the scatterer and the background]: SurfaceMELat r=aaoutinoutinoutinoutit r=anˆˆ[()()]22bbEEHHfEEnHHn (C1) At the object boundary, the average polarization requires to be considered as Eff / 2Pto arrive at Eq (C1) and; Eq (C1) turns into the surface force of well-known Einstein-Laub force when background is considered air. Now considering the difference between the external Minkowski ST and the internal MEL ST just at the boundary of an embedded scatterer: MIXMIXMinkMoutoutininoutouEtininLˆˆˆ[(out)in ][][]r absbsr aTTnEEEEn +HHHHn (C2) outoutininoutoDDMinkMutinnELi1ˆˆ[(out)in ][[] []]2r abbbbr aTTnEEEE+HHHHn (C3) Now, adding Eq (C2) and (C3) together, after some calculations we get: Appendices 160 outinoutoutMininoutinoukMELat r=aat r=atinˆˆˆ[(out)in ][()()]22r abbEEHHTTnEEnHHn (C4) -Which exactly matches with Eq (C1). Appendix D Appendices 161 Time averaged external (total force) and internal force (bulk force) for an embedded achiral object: All the 3D simulations throughout chapter 5 and this appendix are conducted using incident power of 0.57mW/µm2. Fig. 1D: Time-averaged forces: Fout at 1.01raa from Minkowski ST [231,232] and Fin (bulk force) at 0.999raa from the modified Einstein-Laub (MEL) ST [232]. These forces are always of same trend. By adding the surface force of MEL with the bulk force, the magnitude exactly matches with external time averaged total force. (a) Homogeneous background: Force on a non-absorbing dielectric sphere by varying its permittivity (with a=500 nm, illuminated by a linearly polarized plane wave 0i kztxEE e at 1064nm). The unbounded homogeneous dielectric background parameter is: 05b. (b) Heterogeneous background: Force on a magneto-dielectric infinite cylinder of (,ss ) = (05,04 ) and radius 2000 nm embedded in an heterogeneous unbounded background of four different magneto-dielectric layers: (,bb ) = (03 ,02 ); (04 ,03 );(05,04 ); (06 ,05 ). The illuminating plane wave with 1064nm indices at varying angles. (c) A 3D core-shell magneto-dielectric sphere embedded in air is illuminated by a linearly polarized plane wave with varying wavelengths. Core radius, a=500 nm, 08s, 03s. Bounded local immediate background (i.e. the shell) parameters: radius, b=600 nm and 04b; 02s. CoreoutoutFF, at different illumination wavelengths λ obtained from Minkowski ST at r=a+ using the fields in the shell. CoreininFF based on MEL ST at r=a- using core fields. Bibliography 162 Bibliography [1] A. Ashkin, J. M. Dziedzic, J. E. Bjorkholm, and S. Chu, Observation of a single-beam gradient force optical trap for dielectric particles. Opt. Lett. 11, 288 (1986). [2] A. Ashkin, Acceleration and trapping of particles by radiation pressure. Phys. Rev. Lett. 24, 156 (1970). [3] K. Dholakia and P. Zemánek, Colloquium: Gripped by light: Optical binding. Reviews of Modern Physics 82, 1767 (2010). [4] M. M. Burns, J.-M. Fournier, and J. A. Golovchenko, Optical binding. Phys. Rev. Lett. 63, 1233 (1989). [5] P. C. Chaumet and M. Nieto-Vesperinas, Optical binding of particles with or without the presence of a flat dielectric surface. Phys. Rev. B 64, 035422 (2001). [6] E. Almaas and I. Brevik, Possible sorting mechanism for microparticles in an evanescent field. Phys. Rev. A 87, 063826 (2013). [7] A. Jonáš and P. Zemánek, Light at work: The use of optical forces for particle manipulation. sorting, and analysis. Electrophoresis 29, 4813 (2008). [8] M. M. Wang, E. Tu, D. E. Raymond, J. M. Yang, H. Zhang, N. Hagen, B. Dees, E. M. Mercer, A. H. Forster, I. Kariv, P. J. Marchand, and W. F. Butler, Microfluidic sorting of mammalian cells by optical force switching. Nat. Biotechnol. 23, 83 (2004). Bibliography 163 [9] J. R. Moffitt, Y. R. Chemla, S. B. Smith, and C. Bustamante, Recent advances in optical tweezers. Annu. Rev. Biochem. 77, 205 (2008). [10] D. G. Grier, A revolution in optical manipulation. Nature 424, 810 (2003). [11] K. G. Dholakia and T. Čižmár, Shaping the future of manipulation. Nat. Photonics 5, 335 (2011). [12] D. Chang, J. Thompson, H. Park, V. Vuletić, A. Zibrov, P. Zoller, and M. Lukin, Trapping and manipulation of isolated atoms using nanoscale plasmonic structures. Phys. Rev. Lett. 103, 123004 (2009). [13] A. Reiserer, C. Nölleke, S. Ritter, and G. Rempe, Ground-state cooling of a single atom at the center of an optical cavity. Phys. Rev. Lett. 110, 223003 (2013). [14] F. M. Fazal and S. M. Block, Optical tweezers study life under tension. Nature Photon. 5, 318 (2011). [15] H. Li, D. Zhou, H. Browne, and D. Klenerman, Evidence for resonance optical trapping of individual fluorophore-labeled antibodies using single molecule fluorescence spectroscopy. J. Am. Chem. Soc. 128, 5711 (2006). [16] M. Nieto-Vesperinas, J. J. Sáenz, R. Gómez-Medina, and L. Chantada, Optical forces on small magnetodielectric particles. Opt. Express 18, 11428 (2010). [17] M. Nieto-Vesperinas, P. C. Chaumet, and A. Rahmani, Near-field photonic forces. Philos. Trans. A Math Phys Eng. Sci 362, 719 (2004). [18] M. L. Juan, M. Righini, and R. Quidant, Plasmon nano-optical tweezer. Nat. Photonics 5, 349 (2011). Bibliography 164 [19] G. Tkachenko and E. Brasselet, Optofluidic sorting of material chirality by chiral light. Nat.comm. 5, 3577 (2014). [20] G. Tkachenko and E. Brasselet, Helicity-dependent three-dimensional optical trapping of chiral microparticles. Nat.comm. 5, 4491 (2014). [21] A. Hayat, J. P. B. Mueller, and F. Capasso, Lateral chirality-sorting optical forces. Proc. Nat. Acad. Sci. 112, 13190 (2015). [22] Q. Zhang, J. J. Xiao, X. M. Zhang, Y. Yao, and H. Liu, Reversal of optical binding force by Fano resonance in plasmonic nanorod heterodimer. Opt. Express 21, 6601 (2013). [23] Q. Zhang and J. J. Xiao, Multiple reversals of optical binding force in plasmonic disk-ring nanostructures with dipole-multipole Fano resonances. Opt. letters 38, 4240 (2013). [24] V. Shvedov, A. Davoyan, C.R. Hnatovsky, N. Engheta, and W.A Krolikowski, Long-range polarization-controlled optical tractor beam. Nat. Photonics 8, 846 (2014). [25] A. Novitsky, C.-W. Qiu, and H. Wang, Single gradientless light beam drags particles as tractor beams. Phys. Rev. Lett. 107, 203601 (2011). [26] B. A. Kemp, T. M. Grzegorczyk, and J. A. Kong, Lorentz Force on Dielectric and Magnetic Particles. Journal of Electromag. Wave Appl. 20, 827 (2006). [27] B. A. Kemp, Resolution of the Abraham-Minkowski debate: Implications for the electromagnetic wave theory of light in matter. J. Appl. Phys. 109, 111101 (2011). [28] M. Bethune-Waddell and K. J. Chau, Simulations of radiation pressure experiments narrow down the energy and momentum of light in matter. Bibliography 165 Reports on Progress in Physics 78, 122401 (2015). [29] K. Dholakia, P. Reece, and M. Gu, Optical micromanipulation. Chem. Soc. Rev. 37, 42 (2008). [30] A. Novitsky, C.-W. Qiu, and A. Lavrinenko, Material-independent and size-independent tractor beams for dipole objects. Phys. Rev. Lett. 109, 023902 (2012). [31] S. Sukhov and A. Dogariu, Negative nonconservative forces: Optical “tractor beams” for arbitrary objects. Phys. Rev. Lett. 107, 203602 (2011). [32] S. Sukhov and A. Dogariu, On the concept of "tractor beams". Opt. Lett. 35, 3847 (2010). [33] D. B. Ruffner and D. G. Grier, Optical conveyors: A class of active tractor beams. Phys. Rev. Lett. 109, 163903 (2012). [34] M. M. Rahman, A. A. Sayem, M. R. C. Mahdy, M. E. Haque, R. Islam, S. T. R. Chowdhury, and M. A. Matin, Tractor beam for fully immersed multiple objects: Long distance pulling, trapping, and rotation with a single optical set-up. Annalen der Physik 527, 777 (2015). [35] V. Kajorndejnukul, W. Q. Ding, S. Sukhov, C. W. Qiu, and A. Dogariu, Linear momentum increase and negative optical forces at dielectric interface. Nat. Photonics 7, 787 (2013). [36] O. Brzobohatý, V. Karásek, M. Šiler, L. Chvátal, T. Čižmár, and P. Zemánek, Experimental demonstration of optical transport, sorting and self-arrangement using a ‘tractor beam’. Nat. Photonics 7, 123 (2013). [37] A. Dogariu, S. Sukhov, and J. Sáenz, Optically induced 'negative forces'. Nat. Photonics 7, 24 (2013). Bibliography 166 [38] V. D. Miljkovic, T. Pakizeh, B. Sepulveda, P. Johansson, and M. Kall, Optical Forces in Plasmonic Nanoparticle Dimers. The Journal of Physical Chemistry C 114, 7472 (2010). [39] O. Peña-Rodríguez, U. Pal, M. Campoy-Quiles, L. Rodríguez-Fernández, M. Garriga, and M. I. Alonso, Enhanced Fano resonance in asymmetrical Au: Ag heterodimers. The Journal of Physical Chemistry C 115, 6410 (2011). [40] W. Wan, W. Zheng, Y. Chen, and Z. Liu, From Fano-like interference to superscattering with a single metallic nanodisk. Nanoscale 6, 9093 (2014). [41] L. V. Brown, H. Sobhani, J. B. Lassiter, P. Nordlander, and N. J. Halas, Heterodimers: plasmonic properties of mismatched nanoparticle pairs. Acs Nano 4, 819 (2010). [42] M. I. Petrov, S. V. Sukhov, A. A. Bogdanov, A. S. Shalin, and A. Dogariu, Surface plasmon polariton assisted optical pulling force. Laser and Photonics Reviews 10, 116 (2016). [43] C.-W. Qiu, W. Ding, M.R.C. Mahdy, D. Gao, T. Zhang, F.C. Cheong, A. Dogariu, Z. Wang, and C.T. Lim, Photon momentum transfer in inhomogeneous dielectric mixtures and induced tractor beams. Light: Science and Applications 4, e278 (2015). [44] B. A. Kemp, T. M. Grzegorczyk, and J. A. Kong, Optical momentum transfer to absorbing Mie particles. Phys. Rev. Lett. 97, 133902 (2006). [45] M. Mansuripur, Resolution of the Abraham–Minkowski controversy. Optics Communications 283, 1997 (2010). [46] I. Brevik, Experiments in phenomenological electrodynamics and the electromagnetic energy-momentum tensor. Phys Rep. 52, 133 (1979). Bibliography 167 [47] A. Ashkin and J. M. Dziedzic, Radiation pressure on a free liquid surface. Phys. Rev. Lett. 30 139 (1973). [48] H. Minkowski, Die grundgleichungen für die elektromagnetischen vorgänge in bewegten körpern. Math. Ann. 68, 472 (1910) and H. Minkowski, The basic equations for electromagnetic processes in moving bodies. Nachr. Ges. Wiss. Gottingen, Math. Phys. Kl. 53 (1908). [49] R. Pfeifer, T. Nieminen, N. Heckenberg, and H. Rubinsztein-Dunlop, Colloquium: Momentum of an electromagnetic wave in dielectric media. Rev. Mod. Phys. 79, 1197 (2007). [50] P. W. Milonni and R. W. Boyd, Momentum of light in a dielectric medium. Adv. Opt. Photonics 2, 519 (2010). [51] M. Abraham and R. C. Circ., Mat. Palermo 28, 1 (1909). [52] M. Abraham and R. C. Circ., Mat. Palermo 30, 33 (1910). [53] B. A. Kemp and T. M. Grzegorczyk, The observable pressure of light in dielectric fluids. Optics letters 36, 493 (2011). [54] S. S. Hakim and J. B. Higham, Proc. Roy. Soc. 80, 190 (1962). [55] R. V. Jones and J. C. S. Richards, The pressure of radiation in a refracting medium. Proc. R. Soc. Lond. 221, 480 (1954). [56] R. V. Jones and B. Leslie, The measurement of optical radiation pressure in dispersive media. Proc. R. Soc. Lond. A 360, 347 (1978). [57] K. T. McDonald, Bibliography on the Abraham-Minkowski Debate: http://www.physics.princeton.edu/~mcdonald/examples/ambib.pdf (2015). Bibliography 168 [58] S.M. Barnett, Resolution of the Abraham-Minkowski Dilemma. Phys. Rev. Lett. 104, 070401 (2010). [59] J. Chen, J. Ng, Z. Lin, and C. T. Chan, Optical pulling force. Nat. Photon. 5, 531 (2011). [60] CizmarT et al. Appl. Phys. Lett. 86, 174101 (2005). [61] CizmarT, Siler M, Zemánek. Appl. Phys. B 84, 197 (2006). [62] Mizrahi A and Fainman, Opt. Lett. 35, 3405 (2010). [63] B.A. Kemp, J.A. Kong, and T.M. Grzegorczyk, Reversal of wave momentum in isotropic left-handed media. Phys. Rev. A 75, 053810 (2007). [64] C. Wang, Z. Gong, Y. L. Pan, and G. Videen, Laser pushing or pulling of absorbing airborne particles. Applied Physics Letters 109, 011905 (2016). [65] H. Chen, B. Zhang, Y. Luo, B. A. Kemp, J. Zhang, L. Ran, and B.I. Wu, Lorentz force and radiation pressure on a spherical cloak. Phys. Rev. A 80, 011808 (2009). [66] T. Čižmár, L. C. D. Romero, K. Dholakia, and D. L. Andrews, Multiple optical trapping and binding: New routes to self-assembly. J. Phys. B: At., Mol. Opt. Phys. 43, 102001 (2010). [67] S. K. Mohanty, J. T. Andrews, and P. K. Gupta, Optical binding between dielectric particles. Opt. Express 12, 2746 (2004). [68] B. N. Slama-Eliau, and G. Raithel, Three-dimensional arrays of submicron particles generated by a four-beam optical lattice. Physical Review E 83, 051406 (2011). Bibliography 169 [69] Z. Yan, R. A. Shah, G. Chado, S. K. Gray, M. Pelton, and N. F. Scherer, Guiding spatial arrangements of silver nanoparticles by optical binding interactions in shaped light fields. ACS Nano 7, 1790 (2013). [70] Z. Yan, U. Manna, W. Qin, A. Camire, P. Guyot-Sionnest, and N. F. Scherer, Hierarchical photonic synthesis of hybrid nanoparticle assemblies. The Journal of Physical Chemistry Letters 4, 2630 (2013). [71] F. C. Cheong, B. J. Krishnatreya, and D. G. Grier, Strategies for three-dimensional particle tracking with holographic video microscopy. Opt. Express 18, 13563 (2010). [72] M. Polin, Y. Roichman, and D. G. Grier, Autocalibrated colloidal interaction measurements with extended optical traps. Physical Review E 77, 051401 (2008). [73] V. Demergis and E. L. Florin, Ultrastrong optical binding of metallic nanoparticles. Nano Lett. 12, 5756 (2012). [74] D. S. Bradshaw and D. L. Andrews, Optically induced forces and torques: Interactions between nanoparticles in a laser beam. Phys. Rev. A 72, 033816 (2005). [75] R. Zhao, P. Tassin, T. Koschny, and C. M. Soukoulis, Optical forces in nanowire pairs and metamaterials. Opt. Express 18, 25665 (2010). [76] M. Li, W. H. P. Pernice, and H. X. Tang, Tunable bipolar optical interactions between guided lightwaves. Nat. Photonics 3, 464 (2009). [77] M. Li. H. Wang, D. Gao, L. Gao, J. Xu, and C. W. Qiu, Radiation pressure of active dispersive chiral slabs. Opt. Express. 23, 16546 (2015). Bibliography 170 [78] A. R. Zakharian, P. Polynkin, M. Mansuripur, and J. V. Moloney, Single-beam trapping of micro-beads in polarized light: Numerical simulations. Opt. express 14, 3660 (2006). [79] A. Akbarzadeh, M. Danesh, C. W. Qiu, and A. J. Danner, Tracing optical force fields within graded-index media. New Journal of Physics 16, 053035 (2014). [80] M. Mansuripur, Optical manipulation: Momentum exchange effect. Nat. Photonics 7, 765 (2013). [81] A. Ashkin and J. M. Dziedzic, Optical trapping and manipulation of viruses and bacteria. Science 235, 1517 (1987). [82] H. Yin , M. D. Wang, K. Svoboda, R. Landick, S. M. Block, and J. Gelles, Transcripting against an applied force. Science 270, 1653 (1995). [83] L. N. Ng, B. J. Luff, M. N. Zervas, and J. S. Wilkinson, Forces on a Rayleigh Particle in the Cover Region of a Planar Waveguide. J. Lightwave Technol. 18, 388 (2000). [84] T. Zhu, M. R. C. Mahdy, Y. Cao, H. Lv, F. Sun, Z. Jiang, and W. Ding, Optical pulling using evanescent mode in sub-wavelength channels. Opt. Express. 24, 18436 (2016). [85] V. R. Almeida, Q. C. Xu, A. Barrios, and M. Lipson, Guiding and confining light in void nanostructure. Opt. Lett. 29, 1209 (2004). [86] A. H. J. Yang, S. D. Moore, B. S. Schmidt, M. Klug, M. Lipson, and D. Erickson, Optical manipulation of nano particles and biomolecules in sub-wavelength slot waveguides. Nature 457, 71 (2009). Bibliography 171 [87] L. Tong, R. R. Gattass, J. B. Ashcom, S. He, J. Lou, M. Shen, I. Maxwell, and E. Mazur, Sub-wavelength diameter silica wires for low-loss optical wave guiding. Nature 426, 816 (2003). [88] C. Xu, H. Lei, Y. Zhang, and B. Li, Backward transport of nano particles in fluidic flow. Opt. Express 20, 3 (2012). [89] L. Zhao, Y. Li, J. Qi, J. Xu, and Q. Sun, Quasi three-dimensional optical trapping by two counter-propagating beams in nano-fiber. Opt. Express. 18, 6 (2010). [90] K. J. Webb and Shivanand, Negative electromagnetic plane-wave force in gain media. Phys. Rev. E 84, 057602 (2011). [91] A. V. Maslov, Resonant Pulling of a Microparticle Using a Backward Surface Wave. Phys. Rev. Lett. 112, 113903 (2014). [92] D. Gao, A. Novitsky, T. Zhang, F. C. Cheong, L. Gao, C. T. Lim, B. Luk'yanchuk, and C.-W. Qiu, Unveiling the correlation between non-diffracting tractor beam and its singularity in Poynting vector. Laser Photon. Rev 9, 75 (2014). [93] Y. Pang and R. Gordon, Optical Trapping of 12 nm Dielectric Spheres Using Double-Nanoholes in a Gold Film. Nano Lett. 11, 3763 (2011). [94] Wikipedia link: http://en.wikipedia.org/wiki/100_nanometres and S. M. Grist, S. A. Schmidt, J. Flueckiger, V. Donzella, W. Shi, S. T. Fard, J.T. Kirk, D. M. Ratner, K. C. Cheung, and L. Chrostowski, Silicon photonic micro-disk resonators for label-free biosensing. Opt. Express 21, 7994 (2013). [95] M. Mansuripur and A. R. Zakharian, Maxwell’s macroscopic equations, the energy-momentum postulates, and the Lorentz law of force. Phys. Rev. E 79, Bibliography 172 026608 (2009). [96] A. Akbarzadeh, J. A. Crosse, M. Danesh, C. W. Qiu, A. J. Danner, and C. M. Soukoulis, Interplay of optical force and ray-optic behavior between Luneburg lenses. ACS Photonics 2, 1384 (2015). [97] M. L. Juan, R. Gordon, Y. Pang, F. Eftekhari, and R. Quidant, Self-induced back-action optical trapping of dielectric nano particles. Nat. Phys. 5, 915 (2009). [98] J. T. Rubin, and L. Deych, On optical forces in spherical whispering gallery mode resonators. Opt. Express 19, 22337 (2011). [99] J. T. Rubin and L. I. Deych, Optical forces due to spherical microresonators and their manifestation in optically induced orbital motion of nanoparticles. Phys. Rev. A 84, 023844 (2011). [100] Lumerical FDTD: https://www.lumerical.com/tcad-products/fdtd/ [101] P. Yeh, A. Yariv, and E. Marom, Theory of Bragg fiber. J. Opt. Soc. Am. 68, 1196 (1978) and H. Munir, D.K. Soumitra, M.R.C. Mahdy, H. Dihan, and Md. Matin, Robust Optical Fiber Using Single Negative Metamaterial, IEEE Phot. Tech. Letter 25, 1043 (2013). [102] S. Gou, S. Albin, and R.S. Rogowski, Comparative analysis of Bragg fibers. Opt. Express 12, 198 (2004). [103] V. Intaraprasonk and S. Fan, Optical pulling force and conveyor belt effect in resonator–waveguide system. Opt. letters 38, 3264 (2013). [104] H. Ren, A. Salandrino, G. A. Siviloglou, and D. N. Christodoulides, Anomalous Optical Force Fields around High-Contrast Subwavelength Bibliography 173 Nanowaveguides. Quantum Electronics and Laser Science Conference, 1-2, (2010) and F. L. Kien, and A. Rauschenbeutel, Negative azimuthal force of nanofiber-guided light on a particle. Phys. Rev. A 88, 063845 (2013). [105] D. Haefner, S. Sukhov, and A. Dogariu, Spin hall effect of light in spherical geometry. Phys.rev.lett. 102, 123903 (2009) and S. Sukhov, V. Kajorndejnukul, and A. Dogariu, Dynamic Consequences of Optical Spin-Orbit Interaction. Nat. Photonics 9, 809 (2015). [106] S. Mokhov, R. El-Ganainy, and D. N. Christodoulides, Power circulation via negative energy-flux wormholes in optical nano waveguides. Opt. Express 14, 3255 (2006). [107] F. J. Rodríguez-Fortuño, G. Marino, P. Ginzburg, D. O’Connor, A. Martínez, G. A. Wurtz, and A. V. Zayats, Near-field interference for the unidirectional excitation of electromagnetic guided modes. Science 340, 328 (2013). [108] J. D. Jackson, Classical electrodynamics, 3rd edition, New York: Wiley (1999). [109] T. V. Raziman and O. J. Martin, Internal optical forces in plasmonic nanostructures. Opt. Express 23, 20143 (2015). [110] A. Herczyński, Bound charges and currents. Am. J. Phys. 81, 202 (2013). [111] D. J. Griffiths, Introduction to Electrodynamics, 3rd edition Upper Saddle River, NJ: prentice Hall (1999). [112] T. M. Grzegorczyk, B. A. Kemp, and J. A. Kong, Stable Optical Trapping Based on Optical Binding Forces. Phys. Rev. Lett. 96, 113903 (2006). [113] J. Ng, R. Tang, and C. T. Chan, Electrodynamics study of plasmonic bonding and anti-bonding forces in a bisphere. Phys. Revi. B 77, 195407(2008). Bibliography 174 [114] C. Min, Z. Shen, J. Shen, Y. Zhang, H. Fang, G. Yuan, L. Du, S. Zhu, T. Lei, and X. Yuan, Focused plasmonic trapping of metallic particles. Nat. comm. 4, 2891 (2013). [115] M. Zeisberger, A. Tuniz, and M. A. Schmidt, Analytic model for the complex effective index dispersion of metamaterial-cladding large-area hollow core fibers. Opt. express 24, 20515 (2016). [116] M. J. Shawon, G. A. Mahdiraji, M. M. Hasan, B. H. Shakibaei, S. Y. Gang, M. R. C. Mahdy, and F. R. M. Adikan, Single Negative Metamaterial-Based Hollow-Core Bandgap Fiber With Multilayer Cladding. IEEE Photonics J. 7, 4600812 (2015). [117] S. Sheikholeslami, Y. W. Jun, P. K. Jain, and A. P. Alivisatos, Coupling of optical resonances in a compositionally asymmetric plasmonic nanoparticle dimer. Nano Lett. 10, 2655 (2010). [118] P. K. Jain and M. A. El-Sayed, Noble metal nanoparticle pairs: effect of medium for enhanced nanosensing. Nano Lett. 8, 4347 (2008). [119] N. Verellen, P. Van Dorpe, C. J. Huang, K. Lodewijks, G. A. E. Vandenbosch, L. Lagae, and V. V. Moshchalkov, Plasmon Line Shaping Using Nanocrosses for High Sensitivity Localized Surface Plasmon Resonance Sensing. Nano Lett. 11, 391 (2011). [120] P. L. Stiles, J. A. Dieringer, N. C. Shah, and R. R. Van Duyne, Annu. Rev. Anal. Chem. 1, 601 (2008). [121] A. M. Michaels, J. Jiang, and L. Brus, Ag nano-crystal junctions as the site for surface-enhanced Raman scattering of single rhodamine 6G molecules. The Journal of Physical Chemistry B 104, 11965 (2000). Bibliography 175 [122] K. Li, M. I. Stockman, and D. J. Bergman, Self-similar chain of metal nanospheres as an efficient nanolens. Phys. Rev. Lett. 91, 227402 (2003). [123] S. A. Maier, P. G. Kik, and H. A. Atwater, Optical pulse propagation in metal nanoparticle chain waveguides. Phys. Rev. B 67, 205402 (2003). [124] S. Zhang, D. A. Genov, Y. Wang, M. Liu, and X. Zhang, Plasmon-Induced Transparency in Metamaterials. Phys. Rev. Lett. 101, 047401(2008). [125] L. Verslegers, Z. Yu, Z. Ruan, P. B. Catrysse, and S. Fan, From electromagnetically induced transparency to superscattering with a single structure: a coupled-mode theory for doubly resonant structures. Phys. Rev. Lett. 108, 083902 (2012). [126] A. J. Pasquale, B. M. Reinhard, and L. Dal Negro, Engineering Photonic-Plasmonic Coupling in Metal Nanoparticle Necklaces. ACS Nano 5, 6578 (2011). [127] T. Shegai, S. Chen, V. D. Miljkovi_c, G. Zengin, P. Johansson, Käll, and M., A Bimetallic Nanoantenna for Directional Colour Routing. Nat. Comm. 2, 481(2011). [128] P. Nordlander, C. Oubre, E. Prodan, K. Li, and M. I. Stockman, Plasmon hybridization in nanoparticle dimers. Nano Lett. 4, 899 (2004). [129] Z. J. Yang, Z. S. Zhang, L. H. Zhang, Q. Q. Li, Z. H. Hao, and Q. Q. Wang, Fano resonances in dipole-quadrupole plasmon coupling nanorod dimers. Optics letters 36, 1542 (2011). [130] J. N. Li, T. Z. Liu, H. R. Zheng, F. Gao, J. Dong, Z. L. Zhang, and Z. Y. Zhang, Plasmon resonances and strong electric field enhancements in side-by-side tangent nanospheroid homodimers. Opt. Express. 21, 17176 (2013). Bibliography 176 [131] L. Shao, C. Fang, H. Chen, Y. C. Man, J. Wang, and H. Q. Lin, Distinct plasmonic manifestation on gold nanorods induced by the spatial perturbation of small gold nanospheres. Nano Lett. 12, 1424 (2012). [132] S. Marhaba, G. Bachelier, C. Bonnet, M. Broyer, E. Cottancin, N. Grillet, and M. Pellarin, Surface plasmon resonance of single gold nanodimers near the conductive contact limit. The Journal of Physical Chemistry C 113, 4349 (2009). [133] P. Chu and D. L. Mills, Electromagnetic response of nanosphere pairs: Collective plasmon resonances, enhanced fields, and laser-induced forces. Phys. Rev. B 77, 045416 (2008). [134] A. J. Hallock, P. L. Redmond, and L. E. Brus, Optical forces between metallic particles. Proc Natl Acad Sci U S A 102, 1280 (2005). [135] P. Chu and D. L. Mills, Laser-induced forces in metallic nanosystems: The role of plasmon resonances. Phys. Rev. Lett. 99, 127401(2007). [136] H. Liu, J. Ng, S. B. Wang, Z. H. Hang, C. T. Chan, and S. N. Zhu, Strong plasmon coupling between two gold nanospheres on a gold slab. New Journal of Physics 13, 073040 (2011). [137] C. Van Vlack, P. Yao, and S. Hughes, Optical forces between coupled plasmonic nanoparticles near metal surfaces and negative index material waveguides. Phys. Rev. B 83, 245404 (2011). [138] R. C. Weast, M. J. Astle, and W. H. Beyer, CRC handbook of chemistry and physics (Vol. 69). Boca Raton, FL: CRC press (1988). [139] Edward D. Palik, Handbook of optical constants of solids. Vol. 3. Academic press, 1998. Bibliography 177 [140] L. S. Slaughter, Y. Wu, B. A. Willingham, P. Nordlander, and S. Link, Effects of symmetry breaking and conductive contact on the plasmon coupling in gold nanorod dimers. Acs Nano 4, 4657 (2010). [141] C. Tabor, D. Van Haute, and M. A. El-Sayed, Effect of orientation on plasmonic coupling between gold nanorods. ACS Nano 3, 3670 (2009). [142] P. K. Jain, W. Huang, and M. A. El-Sayed, On the universal scaling behavior of the distance decay of plasmon coupling in metal nanoparticle pairs. Nano Lett. 7, 2080 (2007). [143] R. Quidant and C. Girard, Surface-plasmon-based optical manipulation. Laser Photonics Rev. 2, 47 (2008). [144] P. L.Marston, Axial radiation force of a bessel beam on a sphere and direction reversal of the force. J. Acoust. Soc. Am. 120, 3518 (2006). [145] S.-H. Lee, Y. Roichman, and D. G. Grier, Optical solenoid beams. Opt. Express 18, 6988 (2010). [146] A. Salandrino and D. N. Christodoulides, Reverse optical forces in negative index dielectric waveguide arrays. Opt. Lett. 36, 3103 (2011). [147] G. A. Swartzlander, Jr, T. J. Peterson, A. B. Artusio-Glimpse, and A. D. Raisanen, Stable optical lift. Nat. Photon. 5, 48 (2011). [148] P. D. Miller, N. N. Akhmediev, and A. Ankiewicz, Optical conveyor belts: a new scheme for fiber communications. Opt. Lett. 21, 1132 (1996). [149] G. S. Wiederhecker, L. Chen, A. Gondarenko, and M. Lipson, Controlling photonic structures using optical forces. Nature 462, 633 (2009). Bibliography 178 [150] M. Mansuripur, Trouble with the Lorentz law of force: Incompatibility with special relativity and momentum conservation. Phys. Rev. Lett. 108, 193901 (2012). [151] S. M. Barnett and R. Loudon, The enigma of optical momentum in a medium. Phil. Trans. R. Soc. A 368, 927 (2010). [152] C. Baxter and R. Loudon, Radiation pressure and photon momentum in dielectrics. J. Mod. Opt. 57, 830 (2010). [153] R. Loudon, Radiation pressure and momentum transfer in dielectrics: The photon drag effect. Phys. Rev. A 71, 063802 (2005). [154] D. J. Griffiths, Resource Letter EM-1: Electromagnetic Momentum. Am. J. Phys. 80, 7 (2012). [155] K. J. Webb, Dependence of the radiation pressure on the background refractive index. Phys. Rev. Lett. 111, 043602 (2013). [156] E. Brasselet, R. Wunenburger, and J.-P. Delville, Liquid optical fibers with a multistable core actuated by light radiation pressure. Phys. Rev. Lett. 101, 014501 (2008). [157] O. Emile and J. Emile, Low-power laser deformation of an air-liquid interface. Phys. Rev. Lett. 106, 183904 (2011). [158] R. Peierls, The momentum of light in a refracting medium. Proc. R. Soc. London Ser. A 347, 475 (1976). [159] R. Peierls, The momentum of light in a refracting medium. II. Generalization. Application to oblique reflexion. Proc. R. Soc. London Ser. A 355, 141 (1977). Bibliography 179 [160] E. R. Tkaczyk, Vectorial laws of refraction and reflection using the cross product and dot product. Opt. Lett. 37, 972 (2012). [161] M. Mansuripur, A. R. Zakharian, and E. M. Wright, Electromagnetic force distribution inside matter. Phys. Rev. A 88, 023826 (2013). [162] M. Mansuripur and A. R. Zakharian, Whence the Minkowski momentum? Opt. Commun. 283, 3557 (2010). [163] R. Loudon, L. Allen, and D. F. Nelson, Propagation of electromagnetic energy and momentum through an absorbing dielectric. Phys. Rev. E 55, 1071 (1997). [164] Mahdy, M. R. C., Dongliang Gao, Weiqiang Ding, M. Q. Mehmood, Manuel Nieto-Vesperinas, and Cheng-Wei Qiu, A unified theory correcting Einstein-Laub electrodynamics solves dilemmas in the photon momenta and electromagnetic stress tensors. arXiv:1509.06971 (2015). [165] A. Novitsky and C.-W. Qiu, Pulling extremely anisotropic lossy particles using light without intensity gradient. Phys. Rev. A 90, 053815 (2014). [166] S. Sukhov, V. Kajorndejnukul, R. R. Naraghi, and A. Dogariu, Dynamic Consequences of Optical Spin-Orbit Interaction. Nat. Photonics 9, 809 (2015) and S. Sukhov, V. Kajorndejnukul, and A. Dogariu, Dynamic Consequences of Optical Spin-Orbit Interaction. ArXiv:1504.01766 (2015). [167] S. B. Wang and C. T. Chan, Lateral optical force on chiral particles near a surface. Nat. comm. 5, 3307 (2014). [168] R. Fortuño, J. Francisco, N. Engheta, A. Martínez, and A. V. Zayats, Lateral forces on circularly polarizable particles near a surface. Nat. comm. 6, 8799 (2015). Bibliography 180 [169] C. Huajin, Y. Jiang, N. Wang, W. Lu, S. Liu, and Z. Lin, Lateral optical force on paired chiral nanoparticles in linearly polarized plane waves. Opt. Lett. 40, 5530 (2015). [170] B. A. Kemp, Macroscopic Theory of Optical Momentum. Prog. in Opt. 60, chapter 5 (2015). [171] M. Dienerowitz, M. Mazilu, and K. Dholakia, Optical manipulation of nanoparticles: a review. J. Nanophoton. 2, 021875 (2008). [172] M. Mansuripur, A. R. Zakharian, and J. V. Moloney, Radiation pressure on a dielectric wedge. Opt. Express 13, 2064 (2005). [173] M. Mansuripur, A. R. Zakharian, and J. V. Moloney, Equivalence of total force (and torque) for two formulations of the Lorentz law. Proc. of the SPIE Conf., 6326 (2006). [174] Q. C. Shang, Z. S. Wu, T. Qu, Z.J. Li, L. Bai, and L. Gong, Analysis of the radiation force and torque exerted on a chiral sphere by a Gaussian beam. Opt. Express 21, 8677 (2013). [175] A. Canaguier-Durand, J. A. Hutchison, C. Genet, and T. W. Ebbesen, Mechanical separation of chiral dipoles by chiral light. New J. of Phys. 15, 123037 (2013). [176] K. Ding , J. Ng, L. Zhou, and C. T.Chan, Realization of optical pulling forces using chirality. Phys. Rev. A 89, 063825 (2014). [177] G. Tkachenko and E. Brasselet, Spin controlled optical radiation pressure. Phys. Rev. Lett. 111, 033605 (2013). Bibliography 181 [178] R. J. Hernandez, A. Mazzulla, A. Pane, K. V. Sepúlveda, and G. Cipparrone , Attractive-repulsive dynamics on light-responsive chiral microparticles induced by polarized tweezers. Lab on a chip 13, 459 (2013). [179] M. G. Donato, J. Hernandez, A. Mazzulla, C. Provenzano, R. Saija, R. Sayed, S. Vasi, A. Magazzù, P. Pagliusi, R. Bartolino, PG.Gucciardi, O. M. Maragò, and G. Cipparrone, Polarization-dependent optomechanics mediated by chiral microresonators. Nat. Comm. 5, 3656 (2014). [180] D. E. Fernandes and M. G. Silveirinha, Optical tractor beam with chiral light. Phys. Rev. A 9, 061801(R) (2015). [181] G. Li, M. Wang, H. Li, M. Yu, Y. Dong, and J. Xu, Wave propagation and Lorentz force density in gain chiral structures. Opt. Mat. Express 6, 388 (2016). [182] S. M. Barnett and R. Loudon, On the electromagnetic force on a dielectric medium. J. Phys. B: At. Mol. Opt. Phys. 39, S671 (2006). [183] Volumetric Technique of Lumerical Software: https://kb.lumerical.com/en/nanophotonic_applications_optical_tweezers_volumetric_technique.html and https://kb.lumerical.com/en/index.html?nanophotonic_applications_optical_tweezers_volumetric_technique.html [184] A. Rohrbach, Stiffness of optical traps: Quantitative agreement between experiment and electromagnetic theory. Phys. Rev. Lett. 95, 168102 (2005). [185] J. Guck, R. Ananthakrishnan, T. J. Moon, C. C. Cunningham, and J. Kas, Optical Deformability of Soft Biological Dielectrics. Phys. Rev. Lett. 84, Bibliography 182 5451 (2000). [186] M. Hoeb, J.O. Radler, S. Klein, M. Stutzmann, and M.S. Brandt, Light-induced dielectrophoretic manipulation of DNA. Biophys J. 93, 1032 (2007). [187] F. Rasetti, Deflection of mesons in magnetized iron, Phys. Rev. 66, 1 (1944). [188] K. T. Mcdonald, Biot-Savart vs. Einstein-Laub Force Law. Joseph Henry Laboratories, Princeton University, Princeton, NJ 08544, (2013), http://www.physics.princeton.edu/~mcdonald/examples/laub.pdf. [189] J. Hu, S. Lin, L. fC. Kimerling, and K. Crozier, Optical trapping of dielectric nanoparticles in resonant cavities. Phys. Rev. A 82, 053819 (2010). [190] V. V. Datsyuk and O. R. Pavlyniuk, Maxwell stress on a small dielectric sphere in a dielectric. Phys. Rev. A 91, 023826 (2015). [191] R. Saija, P.Denti, F.Borghese, O. M. Marago, and M.A. Iati, Optical trapping calculations for metal nanoparticles. Comparison with experimental data for Au and Ag spheres. Opt. Express 17, 10231 (2009). [192] I. Brevik and S. A. Ellingsen, Comment on “Casimir force acting on magnetodielectric bodies embedded in media”. Phys. Rev. A 79, 027801 (2009). Bibliography 183 [193] U. Fano, Effects of Configuration Interaction on Intensities and Phase Shifts. Phys. Rev. 124, 1866 (1961). [194] Y. S. Joe, A. M. Satanin, and C. S. Kim, Classical analogy of Fano resonances. Phys. Scr. 74, 259 (2006). [195] A. E. Miroshnichenko, S. Flach, and Y. S. Kivshar, Fano resonances in nanoscale structures. Rev. Mod. Phys. 82, 2257 (2010). [196] S. S. Acimovic, M. P. Kreuzer, M. U. Gonzalez, R. Quidant, Plasmon near-field coupling in metal dimers as a step toward single-molecule sensing. ACS Nano 3, 1231(2009). [197] A. M. Michaels, J. Jiang, L. Brus, Ag nanocrystal junctions as the site for surface-enhanced Raman scattering of single rhodamine 6G molecules. J. Phys. Chem. B 104, 11965 (2000). [198] K. R. Li, M. I. Stockman, D. J. Bergman, Self-similar chain of metal nanospheres as an efficient nanolens. Phys. Rev. Lett. 91, 227402 (2003). [199] S. Zhang, K. Bao, N. J. Halas, H. Xu, and P. Nordlander, Substrate-induced Fano resonances of a plasmonic nanocube: a route to increased-sensitivity localized surface plasmon resonance sensors revealed. Nano Letters 11, 1657 (2011). [200] D. Sikdar, W. Zhu, W. Cheng, and M. Premaratne, Substrate-Mediated Broadband Tunability in Plasmonic Resonances of Metal Nanoantennas on Finite High-Permittivity Dielectric Substrate. Plasmonics 10, 1663(2015). [201] M. W. Knight, Y. Wu, J. B. Lassiter, P. Nordlander, and N. J. Halas, Substrates matter: influence of an adjacent dielectric on an individual plasmonic nanoparticle. Nano Letters 9, 2188 (2009). Bibliography 184 [202] K. Joulain, R. Carminati, J. P. Mulet, and J. J. Greffet, Definition and measurement of the local density of electromagnetic states close to an interface. Phys. Rev. B 68, 245405 (2003). [203] Z. Wu, and Y. Zheng, Radiative Enhancement of Plasmonic Nanopatch Antennas. Plasmonics 11, 213 (2016). [204] W. R. Holland, and D. G. Hall, Frequency shifts of an electric-dipole resonance near a conducting surface. Phys. Rev. Lett. 52, 1041(1984). [205] O. Brzobohatý, T. Čižmár, V. Karásek, M. Šiler, K. Dholakia, and P. Zemánek, Experimental and theoretical determination of optical binding forces. Opt. Express 18, 25389 (2010). [206] J. B. Lassiter, F. McGuire, J. J. Mock, C. Ciracì, R. T. Hill, B. J. Wiley, and D. R. Smith, Plasmonic waveguide modes of film-coupled metallic nanocubes. Nano Letters 13, 5866 (2013). [207] C. Manolatou and F. Rana, Subwavelengthnanopatch cavities for semiconductor plasmon lasers. IEEE Journal of Quantum Electronics 44, 435(2008). [208] A. Moreau, C. Ciracì, J. J. Mock, R. T. Hill, Q. Wang, B. J. Wiley, and D. R. Smith, Controlled-reflectance surfaces with film-coupled colloidal nanoantennas. Nature 492, 86 (2012). [209] Z. Xiao-Ming, X. Jun-Jun, and Z. Qiang, Optical binding forces between plasmonicnanocubes: A numerical study based on discrete-dipole approximation. Chinese Phys. B 23, 017302 (2013). Bibliography 185 [210] Q. Zhang, J. J. Xiao, X. M. Zhang, and Y. Yao, Optical binding force of gold nanorod dimers coupled to a metallic slab. Optics Communications 301, 121 (2013). [211] A. S. Zelenina, R. Quidant, and M. Nieto-Vesperinas, Enhanced optical forces between coupled resonant metal nanoparticles. Opt. Lett. 32, 1156 (2007). [212] G. L. Úvŕque and O. J. Martin, Optical interactions in a plasmonic particle coupled to a metallic film. Opt. Express 14, 9971(2006). [213] P. T. Bowen and D. R. Smith, Coupled-mode theory for film-coupled plasmonic nanocubes. Phys. Rev. B 90, 195402 (2014). [214] M. R. C. Mahdy, M. Q. Mehmood, Weiqiang Ding, Tianhang Zhang, ………and Zhi Ning Chen, Lorentz force and the optical pulling of multiple ………rayleigh particles outside the dielectric cylindrical waveguides. Annalen der ………Physik 529 (2017). [215] W. Frias and A. I. Smolyakov, Electromagnetic forces and internal stresses in dielectric media. Phys. Rev. E 85, 046606 (2012). [216] B. Kemp, T. Grzegorczyk, and J. Kong, Ab initio study of the radiation pressure on dielectric and magnetic media. Opt. Express 13, 9280 (2005). [217] V. B. Veselago, Energy, linear momentum and mass transfer by an electromagnetic wave in a negative-refraction medium. Phys. Usp. 52, 649 (2009). [218] C. J. Sheppard, and B. A. Kemp, Kinetic-energy-momentum tensor in electrodynamics. Phys. Rev. A 93, 013855 (2016). Bibliography 186 [219] V. B. Veselago and V. V. Shchavlev, On the relativistic invariance of the Minkowski and Abraham energy-momentum tensors. Phys. Usp. 53, 317 (2010). [220] K. J. Chau and H. J. Lezec, Revisiting the Balazs thought experiment in the presence of loss: electromagnetic-pulse-induced displacement of a positive-index slab having arbitrary complex permittivity and permeability. Applied Phys. A 105, 267 (2011). [221] O. A. Capeloto V. S. Zanuto, L. C. Malacarne, M. L. Baesso, G. V. B. Lukasievicz, S. E. Bialkowski, and N. G. C. Astrath, Quantitative assessment of radiation force effect at the dielectric air-liquid interface. Scientific reports 6, 20515 (2016). [222] C. Wang, Can the Abraham Light Momentum and Energy in a Medium Constitute a Lorentz Four-Vector? Journal of Modern Physics 4, 1123 (2013). [223] A. Einstein, The Collected Papers of Albert Einstein. Vol. 8, Princeton University Press, Princeton, NJ (1998). [224] L. Zhang, W. She, N. Peng, and U. Leonhardt, Experimental evidence for Abraham pressure of light. New Journal of Physics 17, 053035 (2015). [225] G. Verma and K. P. Singh, Universal Long-Range Nanometric Bending of Water by Light. Phys. Rev. Lett. 115, 143902 (2015) and N. G. Astrath, L. C. Malacarne, M. L. Baesso, G. V. Lukasievicz, and S. E. Bialkowski, Unravelling the effects of radiation forces in water. Nat. Comm. 5, 4363 (2014). Bibliography 187 [226] Dongliang Gao, Weiqiang Ding, Manuel Nieto-Vesperinas, Xumin Ding, Mahdy Rahman, Tianhang Zhang, Chwee Teck Lim and Cheng-Wei Qiu, Optical Manipulation from Microscale to Nanoscale: Fundamentals, Advances, and Prospects. Light: Science and Applications 6, e17039 (2017). [227] K. J. Chau and H. J. Lezec, Revisiting the Balazs thought experiment in the case of a left-handed material: electromagnetic-pulse-induced displacement of a dispersive, dissipative negative-index slab. Opt. Express 20, 10138 (2012). [228] T. M. Grzegorczyk and J. A. Kong, Analytical expression of the force due to multiple TM plane-wave incidences on an infinite lossless dielectric circular cylinder of arbitrary size. JOSA B 24, 644 (2007). [229] T. M. Grzegorczyk, B. A. Kemp, and J. A. Kong, Trapping and binding of an arbitrary number of cylindrical particles in an in-plane electromagnetic field. JOSA A 23, 2324 (2006). [230] B. A. Kemp, Comment on “Revisiting the Balazs thought experiment in the presence of loss: electromagnetic-pulse-induced displacement of a positive-index slab having arbitrary complex permittivity and permeability”. Applied Phys. A 110, 517 (2013). [231] Tianhang Zhang, Mahdy Rahman Chowdhury Mahdy, Yongmin Liu, Jing hua Teng, Lim Chwee Teck, Zheng Wang and Cheng-Wei Qiu, All-Optical Chirality-sensitive Sorting via Reversible Lateral Forces in Interference Fields. ACS Nano (2017). [232] M.R.C. Mahdy, Tianhang Zhang, Weiqiang Ding, Amin Kianinejad, Manuel Nieto-Vesperinas, Consistency of time averaged optical force laws for embedded chiral and achiral objects. arXiv:1704.00334 (2017). View publication stats | RESEARCH GATE |
sensorsArticleA Standard-Based Internet of Things Platform and Data FlowModeling for Smart Environmental MonitoringTércio Filho 1,*, Luiz Fernando 2, Marcos Rabelo 2, Sérgio Silva 1, Carlos Santos 3, Maria Ribeiro 4,Ian A. Grout 5, Waldir Moreira 6and Antonio Oliveira-Jr 6,7�����������������Citation: Filho, T.; Fernando, L.;Rabelo, M.; Silva, S.; Santos, C.;Ribeiro, M.; Grout, I.A.; Moreira, W.;Oliveira-Jr, A. A Standard-BasedInternet of Things Platform and DataFlow Modeling for SmartEnvironmental Monitoring. Sensors2021, 21, 4228. https://doi.org/10.3390/s21124228Academic Editor: Antonio GuerrieriReceived: 19 May 2021Accepted: 16 June 2021Published: 20 June 2021Publisher’s Note: MDPI stays neutralwith regard to jurisdictional claims inpublished maps and institutional affil-iations.Copyright: © 2021 by the authors.Licensee MDPI, Basel, Switzerland.This article is an open access articledistributedunderthetermsandconditions of the Creative CommonsAttribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).1Institute of Biotechnology (IBiotec), Federal University of Catalão (UFCAT), Catalão 75705-220, Brazil;sergio@ufg.br2Graduate Program in Science and Technology (PPGCET), Federal University of Catalão (UFCAT),Catalão 75705-220, Brazil; luizfernandoelias@gmail.com (L.F.); marcosrabelo@ufg.br (M.R.)3Campus Palmas, Federal Institute of Education, Science and Technology of Tocantins (IFTO),Palmas 77021-090, Brazil; carlosedu@ifto.edu.br4Institute for Systems and Computer Engineering, Technology and Science (INESC-TEC),4200-465 Porto, Portugal; maria.r.ribeiro@inesctec.pt5Department of Electronic and Computer Engineering, Faculty of Science and Engineering,University of Limerick, V94 T9PX Limerick, Ireland; Ian.Grout@ul.ie6Fraunhofer Portugal AICOS, 4200-135 Porto, Portugal; waldir.junior@fraunhofer.pt (W.M.);antoniojr@ufg.br (A.O.-J.)7Institute of Informatics (INF), Federal University of Goiás (UFG), Goiânia 74690-900, Brazil*Correspondence: tercioas@ufg.brAbstract: The environment consists of the interaction between the physical, biotic, and anthropicmeans. As this interaction is dynamic, environmental characteristics tend to change naturally overtime, requiring continuous monitoring. In this scenario, the internet of things (IoT), together withtraditional sensor networks, allows for the monitoring of various environmental aspects such as air,water, atmospheric, and soil conditions, and sending data to different users and remote applications.This paper proposes a Standard-based Internet of Things Platform and Data Flow Modeling for SmartEnvironmental Monitoring. The platform consists of an IoT network based on the IEEE 1451 standardwhich has the network capable application processor (NCAP) node (coordinator) and multiple wirelesstransducers interface module (WTIM) nodes. A WTIM node consists of one or more transducers, adata transfer interface and a processing unit. Thus, with the developed network, it is possible tocollect environmental data at different points within a city landscape, to perform analysis of thecommunication distance between the WTIM nodes, and monitor the number of bytes transferredaccording to each network node. In addition, a dynamic model of data flow is proposed where theperformance of the NCAP and WTIM nodes are described through state variables, relating directlyto the information exchange dynamics between the communicating nodes in the mesh network. Themodeling results showed stability in the network. Such stability means that the network has capacityof preserve its flow of information, for a long period of time, without loss frames or packets dueto congestion.Keywords: IoT platform; IEEE 1451 standard; smart environmental monitoring; data flow modeling1. IntroductionThe use of technology for environmental monitoring has become an important toolfor the management of public health and the inspection of flora. In addition, such a toolallows studies focusing on climate and micro-climate analyses, and weather forecasting.These studies can provide essential knowledge for a better management of resources andplanning in a specific region. Possible areas that can benefit of such knowledge includeagriculture, livestock, fish farming, apiculture, among others [1].Sensors 2021, 21, 4228. https://doi.org/10.3390/s21124228https://www.mdpi.com/journal/sensorsSensors 2021, 21, 42282 of 22Atmospheric processes and factors have been studied by meteorologists and clima-tologists, respectively, with the focus on weather forecasting and characterization of theclimate in macro- and micro-regions [2,3]. The factors of the weather in a region, also calledmeteorological variables, such as humidity, temperature, precipitation, and wind directionthat have been analyzed over decades allow characterization of the climate according tothe seasons of the year. This knowledge can be used in the management of resources andfor planning in activities such as civil construction and crop growth. This knowledge alsobecomes important for the analysis of air pollution, its dispersion, concentration and possi-ble interactions of the environment with living beings present in a given place. However, itis difficult to monitor the weather of a specific region for a long period of time to conduct aclimate analysis without fully automating the data acquisition process. To narrow this gap,wireless sensors networks have been developed.A wireless sensor network (WSN) can be defined as a network of devices, capable ofmeasuring aspects of the environment and propagate the information collected through thenetwork. The data are transmitted through the network nodes, possibly via multiple jumpsuntil a master node, using a wireless interfaces. The data collected by a sensor networkcan be heterogeneous and distributed. In addition, the data are collected for a long timeand are stored on a server automatically, which can form large data repositories. In a WSNproject, specific factors must be taken into account: fault tolerance, scalability, low cost,low power consumption, programmability, and security. In this case, security regardingaspects such as access control, message integrity, confidentiality and protection againstreproduction [4,5].For the integration of networked devices, the internet of things (IoT) emerged as a newparadigm making objects connected via the Internet. The concept becomes increasinglypossible due to the miniaturization of devices and transducers, in addition to the increaseddata transmission rate over the network. Such features result in a significant increase in thenumber of objects connected in a network, making them indispensable in people’s lives. Themost used technologies for the development of networks based on IoT are radio frequencyidentification (RFID), WSN, middleware, cloud computing, and IoT application software [6].IoT-based systems are applied in several areas: smart homes, smart cities, health care,energy management, environmental monitoring, and agriculture. Each application can usea specific protocol for IoT systems, such as hypertext transfer protocol (HTTP), transmissioncontrol protocol/user datagram protocol (TCP/UDP), extensible messaging and presenceprotocol (XMPP), message queuing telemetry transport (MQTT), and message queuingtelemetry transport-sensor network (MQTT-SN). A standard that has been adapted to theIoT is the IEEE 1451 standard.The IEEE 1451 standard provides functionalities, protocols, communication interfaces,and plug-and-play features through transducers electronic data sheet (TEDS). Basically, theIEEE 1451 standard specifies three entities: network capable application processor (NCAP),transducer interface module (TIM), and Client. NCAP is a network gateway that receivesdata from the external network, processes the data, and forwards the processed data to theinternal network based on one of the subcommittees of the IEEE 1451 standard. TIM, is anetwork module that contains a communication interface with NCAP, sensors/actuators,processing unit, signal conditioning circuit and TEDS. TEDS are the TIM module infor-mation stored in a non-volatile memory. The client is the network user or an end device.With the advent of transducer networks, the client entity is defined according to: IEEE1451.1.1 (TCP/UDP), IEEE 1451.1.2 (HTTP), IEEE 1451.1.3 (web interface), IEEE 1451.1.4(XML extensible markup language (XMPP)), IEEE 1451.1.5 (simple network managementprotocol (SMNP)), IEEE 1451.1.6 (MQTT), and IEEE 1451.1.7 (MQTT-SN) [7]. In this context,it is possible to apply the IEEE 1451 standard and the IoT concept to smart transducernetworks, increasing the flexibility of devices to access data, performance, insertion of newdevices accessing data and decreasing costs [8]. Figure 1 shows the model define by IEEE1451 for development of transducers network.Sensors 2021, 21, 42283 of 22Figure 1. Network topology developed.This paper proposes a standard-based IoT platform and data flow modeling for smartenvironmental monitoring employing ZigBee interfaces with the MQTT protocol (IEEE1451.1.6). Thus, an NCAP/Broker was developed for the integration of the two networkmodels and the collection of data from the WTIM modules over the Internet. The NCAPnetwork node converts the data to the Internet using the publish/subscribe standard. Thenetwork was applied in the field for the tests of communication and reading of the data attwo weather stations developed with ZigBee transmitters. Thus, it was possible to carryout the network analysis in practice through data collection using the MQTT protocol andthe IEEE 1451 standard according to Figure 1.To validate the proposed network model, it is possible to use theoretical/computationalmodels that analyze the degree of parity between the theoretical model and the experi-mental model. One of the central issues is the control of network congestion and the studyof techniques that are able to prevent its emergence. In general, the control mechanismstake into account techniques that manipulate the data flow between the nodes of the [9]network. One way to model the flow is to consider the network as a dynamic system [10].Among the various topics that can be addressed, the control of flow congestion has arousedinterest. This is due to the fact that congestion control strategies are built through dynamicmodels whose complexity is not yet fully understood [10–12]. Thus, concepts such as phaseportrait, stability, fixed point, and bifurcations, widely used in classical dynamic systems,can be applied to analyze the behavior of the data flow in the network [12]. In addition tothe development of the NCAP/Broker network node, the contributions of this work areas follows:•development of a mesh sensor network based on the IEEE 1451 standard for IoT;•data flow analysis of the internal network;•analysis of network communication based on the RSSI signal;•analysis of packet loss between communication networks; and•modeling the network data flow for congestion analysis.The remainder of this paper is organized as follows. Section 2 discusses related worksand highlights the contributions presented in this research. Section 3 shows the networkdeveloped, the NCAP, wireless transducer interface module (WTIM), and commands.Section 4 details the wireless sensor network implemented in the field, and Section 5presents the network data flow modeling for analysis. The last section, Section 6, providesconclusions to the paper.2. Related WorkThe application of IoT networks [13] is increasingly becoming part of the peoplelives, carrying out the control of complex systems [14], environmental monitoring [15],precision agriculture [16,17], digital farming [18], digital health [19], and smart homes.Several researches have employed IoT networks to collect big data for analysis and complexdecision making. In [20], the authors use a multidimensional spatial scan technique toSensors 2021, 21, 42284 of 22discover household movement patterns, in order to help in the development and planningof public policies. The research in [21] reviews studies of IoT applied to smart cities andsustainable cities, including research aimed at conceptual, analytical, and overarchinglevels, as well as research on specific technologies and applications. In [22], the big dataof of supply chains in manufacturing industries and artificial neural networks are usedto identify predictors of sustainable business in emerging economies. In [23], a social-ecological model is developed to understand how ecosystems services drive urban growth.The research in [24] links IoT to Industry 4.0, deriving an empirical formulation to establisha mapping between Industry 4.0 attributes and sustainability goals. For a more detailedoverview of smart environmental monitoring IoT-based solutions and platforms, we directthe reader to the work in [25].In addition to the importance of IoT in providing high-value applications in variousareas of knowledge, it is essential that the IoT infrastructure uses open standards in away that is easily replicable, maintainable, and scalable. Currently, an open standardthat complies with these characteristics is IEEE 1451. Several studies have been carriedout using the IEEE 1451 standard and IoT systems, as, for example, in [26], where theauthors presented a TIM using two communication interfaces. The highway addressableremote transducer (HART)-based module and TIM modules are based on the IEEE 1451standard. The HART is a hybrid analog/digital industrial automation protocol. The HART-based module was coded using very high speed integrated circuit (VHSIC) hardwaredescription language (VHDL) and implemented using the field programmable gate array(FPGA) to communicate with devices through the use of the universal asynchronousreceiver/transmitter (UART) protocol.In [8], the combination of IoT concepts and the active subcommittees of the IEEE 1451standard are considered. They also presented an integration of the IEEE 1451 standardusing IoT concepts and a new proposal for how the transducer electronic data sheet (TEDS)is referred to as the health electronic data sheet (HEDS). HEDS is an adaptation of TEDSfor a health area, making it possible to distribute health assessment data at all levels of asystem hierarchy.The authors of [27] presented the IoT architecture in intelligent environments and acomparison between the technologies used for environmental monitoring. The authorspresented technologies and sensors, such as the Arduino UNO, DHT11 sensor, and theESP8266 wireless transmitter. The study carried out by the authors on intelligent monitoringusing IoT facilitates the development of future work by other researchers.In another work [28], the authors presented a new concept for services critical to IoTdefined as critical IoT (CIoT). In more critical services, IoT protocols become vulnerablehaving to work with a communication prioritization scheme. Thus, a new criticality level(CL) and a security level (SL) are proposed to categorize and prioritize various types ofrelevant transmission. The authors also describe the use of the NCAP-IEEE 1451 standard,WTIM and TEDS for a mobile cloud computing (MCC) connection solution. In this way, itwas possible to present a prioritization scheme involving the indices of CL and SL proposedto solve the problem related to latency tolerant.In [14,29,30], the authors developed a wireless sensor network based on the Zigbeestandard to different environments and performed analyses of performance. For example,the authors of [29] introduced an open-source wireless mesh network (WMN) module thatintegrated the functions of network discovery, automatic routing control, and transmissionscheduling. In [14], a smart grid architecture-based system was developed that used alow-power Zigbee mesh network in response to adaptive traffic on the road. The sensornetwork was implemented and tested in a real environment. The system demonstrated thatis capable of energy savings depending on the variations in daylight hours between summerand winter. In [30], the authors proposed a linear sensor network to deployment at oil andgas pipeline using a custom sensor board accompanied with algorithms to solve networkcreation, leak interrupt detection, and routing of high-priority messages with reliability whileSensors 2021, 21, 42285 of 22keeping network active. The authors consider the development of the network to be lowenergy consumption, latency, access to sensors/actuators, and data reliability.Projects have been developed in the area of IoT using the IEEE 1451 standard, carryingout the monitoring of the sequence of painting and washing tubs [31]. In [32], TIMs weredeveloped using an MSP430F5529 board and a Raspberry Pi as the NCAP. Another RaspberryPi was used as an MQTT broker based on IEEE 1451.1.6. The author’s analysis was realizedbased on three transducers: temperature sensor, voltage sensor, and light-emitting diode(LED). The work also presents the description of TEDS and the common characteristics:identification, representation, communication, life cycle, function, and interoperability.In [33], the development of a semantic sensor web (SSW) as a combination of sensornetworks, web services, database, and semantic web technologies for monitoring envi-ronmental conditions is shown. The data were collected in the environment, stored in adatabase, and made available to the software through an application programming inter-face (API). In order to provide the data obtained by the API, a semantic web was used whichcreated the data interoperability in a system such as resource description framework (RDF),a language used to process metadata, the SPARQL protocol, and RDF query language(SPARQL) which is an integration class between RDF and hypertext preprocessor (PHP).Several models have been presented in the literature to deal with network congestion.In [10], results on flow stability are presented through the design of congestion controllers,using models of dynamic systems to describe the network. Due to the number of parametersinvolved in network models, it is common for bifurcations to appear at equilibrium points.In [34], Hopf bifurcation problems are studied in the context of congestion controller design.In [10], a general multi-link/multi-source model for TCP/random early detection(RED) transfer protocols was proposed and it was demonstrated that the dynamic behaviorof the packet flow is strongly influenced by the stability of the TCP/RED protocol. Ensuringa relationship between the stability of the system and the communication delay timebetween processes, allowing us to conclude that the increase in the delay time generatesinstability in the protocol.In the aforementioned presented works, different types of wireless networks wereidentified and studies were carried out. However, they did not define a new descriptionmodel for the external network based on the IEEE 1451 standard. Our work has usedtwo models—the HTTP and MQTT protocols—according to Figure 1. In addition to theexternal network, the developed network presents the signal level between the nodes andthe request and response time for each command, storing the data so that the networkanalysis can be performed. Our proposal also presents the formulation and analysis of thedata flow of the developed network, differentiating it from the state of the art.3. Our Proposal: IEEE 1451 Standard-Based IoT PlatformThe IEEE 1451 standard defines two network modules, called NCAP and WTIM,and a set of subcommittees for the definition of the communication interface, protocols,and TEDS [35,36]. The IEEE P1451.1.6 subcommittee specifies the characteristics of theNCAP and the communication with the external network using the MQTT communicationprotocol. The IEEE 1451.0-2007 standard describes common functions, communication pro-tocols and TEDS, that is, descriptions of the transducers stored in a non-volatile memory inWTIM. The IEEE 1451.5-2007 standard specifies wireless network communication, whetherWiFi, ZigBee, or Bluetooth. In this project, the ZigBee interface was used to communicatebetween NCAP and WTIM for internal network, and to external network MQTT-IEEE1451.1.6. Figure 2 shows the network model developed for the collection of environmentaldata and analysis.Sensors 2021, 21, 42286 of 22Figure 2. Network model developed.3.1. Network Capable Application Processor—NCAP (IEEE 1451.1.6)The NCAP is a node with local processing capacity, capable of receiving data froman external network and converting the data to the internal network based on the IEEE1451 standard, thus performing the reading/control of the transducers connected to TIM.For this, the NCAP node has the characteristic of identifying the type of network to whichit is connected, thus giving the concept of interoperability, in addition to abstractinginformation of the type of connected transducer, facilitating the introduction of the plug-and-play operation mode [37].NCAP was developed with the objective of creating an autonomous and expandablenetwork that can carry out data collection for a long period of time. It has the functionalityof discovering WTIM nodes in a plug-and-play way, reading sensors through WTIM nodesat predefined time intervals, storing data, and transmitting to an external network usingthe MQTT protocol.The WTIM’s have a logical functionality in which the objective is to receive framesfrom the NCAP network node, verify that the frame corresponds to the WTIM and, ifso, the frame is decoded and the command is carried out. If the command is not for thegiven WTIM, the frame is forwarded to the next WTIM. For the external network, theMQTT protocol was defined and the Mosquitto server was installed and configured on theNCAP network node. The NCAP protocol manager was developed in Python, makingthe requests for WTIM’s modules on the internal wireless ZigBee network and forwardingit to the MQTT broker. Thus, in turn, it makes a publication on the external network forforwarding to customers. Thus, the development of NCAP was divided into two levels:hardware and software. Section 3.1.1 presents a hardware and interface configuration, andSection 3.1.2 describes the software.3.1.1. Hardware and InterfacesThe NCAP was developed using the Raspberry Pi B+ made by the Raspberry PiFoundation that has a Broadcom BCM2837 64bit ARMv8 Cortex-A53 Quad-Core 1.4 GHz,1 GB of random access memory (RAM), and the following interfaces: Ethernet, Bluetooth,high-definition multimedia interface (HDMI) output, 4 universal serial bus (USB) ports, agraphics processing unit (GPU) and 40 pins, configured as 28 general purpose input/output(GPIO) pins (shared with serial peripheral interface (SPI), inter-integrated circuit (I2C).The transceiver module used was the XBee Pro S2C, manufactured by Digi International,having 2.4 GHz of frequency band, transmission rate of 250 Kb/s and can reaches a rangeup to 3200 m of communication distance according to the proprietary specification. TheRaspberry and XBee Pro S2C communicate through connecting directly using the receive(Rx), transmission (Tx) from UART, and to turn on the X Bee Pro S2C, were used the pins3.3 V and GND from the Raspberry Pi [38]. In this work, the security of the transmitted datais performed through the ZigBee standard using the XBee Pro S2C module. ZigBee securityin the media access control (MAC) layer uses Encrypted Communication using SymmetricKeys, Frame Integrity through CRC Redundancy Checks and Frame Sequentiality. In addi-tion to MAC layer security, ZigBee has Cryptographic Key Establishment, Key Transport,Frame Protection and Device Management. ZigBee uses 128-bit keys working with 3 typesSensors 2021, 21, 42287 of 22of keys, Network Key (used by all nodes in the network), Link Key (secret session keys,between connected devices) and Master Key (used to generate the key Link) [39].The XBee Pro S2C modules were configured using the X-CTU software defined as coor-dinator and router. For the coordinator, the following parameters were changed: “PAN-ID”(definition of PAN ID); “CE”—Coordinator Enable (activated); “DL”—Destination AddressLow (0 × FFFF) and “NI”—Node Indetification (Coordinator). For the router: “PAN-ID”(same as the coordinator); “CE”—Coordinator Enable (disabled); “DL” Destination AddressLow (zero-(0)) and “NI”—Node-Indetification (WTIM). Thus, the commands are sent to thenetwork nodes with the destination MAC, if the MAC is not from the node that received thecommand, it forwards to the next node. When the command has the same address as thenetwork node, then is executed and returns with the response command to the coordinatorusing the coordinator node address or default address: “0 × 0000000000000000”.3.1.2. NCAP/MQTT SoftwareFor the development of the logical part of the NCAP, the following tools were installed:Maria DB, Mosquitto, and Apache servers. The NCAP manager was developed in thePython language, using a digi.xbee.models package library to make the logical part of thecommunication between the Raspberry and the X Bee Pro S2C module. For sending datausing the MQTT protocol, the paho.mqtt library was used.When the NCAP module is started, it sends a command on the network to add theWTIM modules. Upon receiving, each the WTIM modules returns the X Bee Pro S2Cmodule ID, storing it in the database. New WTIM modules can be added at any time(plug and play) and, when the NCAP command occurs again, the new WTIMs will alsohave their IDs added to the database. If the ID is stored in the database, the NCAP sendsa TEDS request command using the X Bee Pro S2C module ID; thus, it is possible toidentify the WTIM module and perform the recognition of the transducers on the network.Figure 3 shows the flowchart of the NCAP network node and the recognition part of theWTIMs modules.When the recognition of the WTIMs modules is complete, the NCAP automaticallyenters the data reading mode according to a predetermined time configured on the networkcoordinator. When the time is up, the NCAP sends a command to the first WTIM torequest the data for each sensor. After that, the next module is verified in the databaseand forwards the request command. Upon receiving data from the internal network bythe broker, the publication is made to customers. In the external network, the sender andthe receiver are independent and can receive and send data simultaneously. The brokerreceives the data according to the predetermined time by the NCAP protocol manager andforwards it via MQTT using the publish according to Figure 2. An important characteristicof the broker is the filtering of messages, which is possible in 3 ways: topic, content, ortype. In filtering by topic, the messages are sent in a hierarchical structure using the 8-bitunicode transformation format (UTF8) string in general. In content filtering, the messagesare based on a filter language, and by type, messages are sent according to the type or classof the message. In this project, messages were sent using topics according to the followingstructure: Station _1/ Sensor/Temperature/Minimum, being divided into 4 levels, thusit was possible to define the sensors of each station. Figure 3 shows the flowchart of theNCAP protocol manager.Sensors 2021, 21, 42288 of 22Figure 3. Flowchart of the logical part of the NCAP.3.2. Wireless Transducer Interface Module—WTIMThe WTIM has characteristics similar to the NCAP, Raspbian operating system, andcommunication with the X Bee Pro S2C module is through the UART interface. However,the WTIM has sensors, signal conditioning circuits, and TEDS stored in files. Storage ofTEDS files was specified by the name of each TEDS, such as Meta_TEDS.txt, Transducer-Channel_TEDS.txt, UserTransducer_TEDS.txt, and Phy_TEDS.txt. The sensor module usedwas the BME280 that measures pressure, temperature and humidity. The characteristics ofthe BMP280 sensor are supply voltage of 1.8 V and 5V DC, I2C interface (up to 3.4 MHz),SPI (up to 10 MHz). In addition to the BMP280, the wind direction, wind speed, and rainsensor were used. Since the Raspberry Pi B + does not have an internal analog-to-digitalconverter (ADC), an external 16-bit ADC was used that communicates with the moduleusing the I2C protocol. Figure 4 shows the WTIM node and Figure 5 presents the schematicof the developed electronic circuit.Figure 4. WTIM developed in the laboratory.The WTIM software is illustrated by the flowchart in Figure 6. The WTIM was developedusing the Python language, using the same library for the development of NCAP, andcommunication with X Bee Pro S2C: digi.xbee.models package. When the WTIM is initialized,the UART interface, the interrupts and communications with sensor using the I2C areinitialized. Then, the WTIM waits for an NCAP request. When it receives it, the frame isdecapsulated and, then, is checked the command class based on the IEEE 1451 standard. Ifthe class is 0 × 02, the command is for the manipulation of TEDS, if the class is 0 × 03, itchecks the function of the command of access to transducers. The TEDS or sensor data isplaced into the reply command based on IEEE 1451 standard, encapsulated in a ZigBee frameand devolved to the NCAP.Sensors 2021, 21, 42289 of 22Figure 5. Signal conditioning circuit diagram.Figure 6. Flowchart of the logical part of WTIM.3.3. Request and Response CommandsThe ZigBee and IEEE 1451.5 standards establish rules for communication betweenthe devices envolved. In the ZigBee application layer, more specifically, in the applicationframework, the data transfer unit, called the API frame, is implemented.The API frame is the basis of the communication between the devices of the ZigBeemesh network, once all devices communicate through frames. In the IEEE 1451.5 standard,communication between devices is done through IEEE 1451.5 frames. Thus, request andreply frames were implemented, referring to the sensor readings and the TEDS.For the development of the mesh network using the ZigBee standard and IEEE 1451.5,the frames of both standards were joined. Thus, frames based on the IEEE 1451 standardare encapsulated in ZigBee frames and transmitted over the network. The API framecan contain several structures defined through the Type field. Each type of API framecontains a unique identifier, called the API Identifier ID (API). Through this identifier thatthe network and application layers recognize the characteristics of a frame sent or received.Two types of API frames were used to management and communication between thedevices of the developed mesh network: AT command (0 × 08) and ZigBee transmissionrequest (0 × 10). The “AT Command” frame was used to recognize the MAC ID of the XBee Pro S2C connected to the Raspberry. The command is implemented on the main nodeof the network, that is, on the NCAP/Coordinator ZigBee node as shown in the Figure 7.The frame is used, generally, to configure or “read” parameters related to the XBee S2Cmodule connected locally or remotely in the ZigBee mesh network. The generic structureof an “AT Command” frame is illustrated in Figure 7.Sensors 2021, 21, 422810 of 22Figure 7. Generic protocol with AT command.The frame request performs the exchange of messages between the devices of theZigBee mesh network. The objective is to forward the IEEE 1451.5 frames through the dataencapsulation regarding the acquisition and response in its data payload.In the transmission request frame shown in Figure 8, the byte that indicates the typeof frame will be equal to 0 × 10 (request). Thus, according to type 0 × 10, bytes 6–13 areassigned to the MAC ID of the target device. Bytes 14–15 represent the short 16-bit addressof the target device. Byte 16 corresponds to the broadcast radius of the message. Byte17 is the transmission option field. Bytes 18–n corresponds to the data load of the frame.This is the field where the IEEE 1451.5 request frame will be stored. Figure 8 illustratesthe encapsulation of a response frame. The difference in the structure of the responseframe in relation to the allocation of a request frame is in the data payload field (18–n),where the IEEE 1451.5 response frame is stored. The IEEE 1451 command for accessingthe transducers has the following structure: ID Channel (2 Bytes), Class (1 Byte), Function(1 Byte), Length (2 Bytes), and Data (N Bytes). The ID Channel defines the transducerchannel; the Command Class specifies whether the command function is for access tothe transducers, TEDS or, other configuration parameters; the Function specifies the taskto be performed and, Length, the field size of data. Class 0 × 03 was used for access totransducers and 0 × 01 for access to TEDS. For Function, parameter 0 × 01 was used toread the transducers and 0 × 02 to read the TEDS. The TEDS tables are specified in thedata field, being: 0 × 01-Meta TEDS, 0 × 03-Transducer TEDS, 0 × 0C-User TEDS and 0 ×0D-PHY TEDS.The WTIM response is made by entering the coordinator’s MAC ID or the standardframe return value, being: “x00 × 00 × 00 × 00 × 00 × 00 × 00 × 00”. The response com-mand defined by the IEEE 1451.0 standard is encapsulated in the ZigBee frame according toFigure 9 and forwarded to NCAP. The first Byte identifies whether the command was success-ful/failed, where different 0 × 00 means success. The next two Bytes, represents the size ofthe data field and the remaining Bytes, represents the data field with the command response.To read the TEDS, the command follows the same model as the request and responsefor reading the sensors, inserting the IEEE 1451.0 command to read the TEDS in the datafield of the ZigBee frame.Figure 8. Request command.Sensors 2021, 21, 422811 of 22Figure 9. Response command.4. Mesh Network and System DescriptionThe WTIMs nodes were distributed in different points of the city. Figure 10 shows themap with the distribution of this sensor network nodes and the distance between them.For WTIM_1, an intermediate router was inserted to communicate with NCAP due to theelevation of the soil and the woods. On WTIM_2, the communication was made directlywith the NCAP coordinator, however, to improve the communication a 10 meter tower wasbuid and an X Bee S2C module installed at the top.Figure 10. Network topology applied in the city.The coordinator makes the data available to the external network in two ways: usingthe HTTP protocol based on the IEEE 1451.1.2 standard and the MQTT, IEEE 1451.1.6standard. The NCAP node sends reading requests once every 30 minutes through theZigBee mesh network, receives the readings from WTIM nodes, and stores it in the database.Then, it sends to the client through the MQTT protocol. Thus, the data are visualized at themoment of reading through an MQTT client or through the web pages provided by theApache server at the coordinator. Figure 11 shows the data received by coordinator to eachsensor through to MQTT EXPLORER software.Sensors 2021, 21, 422812 of 22Figure 11. Reading of data via MQTT.The reading of the data can be achieved through a web page, in this case, the informa-tion is consulted in the database. In addition to the sensor data, received signal strengthindication (RSSI) was collected before each request was sent to read the sensor and the time(Request/Response) to read data in the WTIM module. Table 1 shows the results obtained.To obtain the signal level, an AT command was used according to Figure 7. In the datafield, the parameter “DB” was assigned.Table 1. Reading table of sensors and network data.DateTimeTemp. ºCLowRSSIdbmReq/RespSecondsPress.LowRSSIdbmReq/RespSecondsWindm/sRSSIdbmReq/RespSeconds14 November00:3019.29−740.2313915.591−740.228880.2194−740.2392714 November01:3019.33−740.27177915.081−740.213780.2277−730.4221614 November02:3019.52−740.45759915.086−740.373620.5611−750.2371314 November03:3019.81−750.49296914.829−740.477350−740.0858914 November04:3019.5−740.2077915.038−740.20251.3111−740.2113314 November05:3019.44−740.2277915.346−7514 November0.5777−750.2246514 November06:3019.59−720.47874915.639−750.435660.4888−740.4940914 November07:3020.33−740.43999916.259−750.477010.6305−740.458415. Validation: Network Data Flow ModelingThis section presents an abstract modeling of the developed network. The justificationthat underlies this methodology lies in the fact that such models are efficient for measuringthe flow of data between the components of the communication network. When approachingsuch a methodology, a central issue is the time delay that arises between us through the com-munication links. To deal with delay in the data flow, whether it is bit/s or packets/s, severalstrategies have been reported in the literature, among them, we mention control techniques,routing being one of the most used [9,40]. For these reasons, a better understanding ofthe nature and mechanisms that are responsible for the delay and how these mechanismsrelate to the parameters of the network is important for a safe analysis of the stability ofthe network.To proceed with the modeling, we consider the communication links as flow innetworks where the bits are transmitted with a certain rate per second. This number iscalled the link’s transmission capacity. Generally, the link capacity is related to physicalparameters of the communication channels, as well as to the interfaces related to it, that is,the data rate that the interface accepts to transmit. Thus, the management of the data flowin the network communication links has relevant effects on packet delay [9].Sensors 2021, 21, 422813 of 22One of the established ways in the literature [9], to model data flow in the network isthrough transmission in a single row (multiplexing) with a first-come/first-served policy.In the language of queuing systems, customers arrive, with a random time, to be servedby the processor. In the present work, NCAP is the service processor, while WTIMs arecustomers who arrive in line to be served. The probability distribution of time betweensuccessive arrivals and the probability of time in task processing are known.In the context of the data network, clients represent the data packets transmittedthrough the communication links by WTIM’S while the server is represented by the datatransmission carried out by NCAP. The service time corresponds to the packet transmissiontime made by NCAP is equal to LC, where L is the length of the packet to be transmittedand C represents the communication capacity of the channel, presented in Figure 12.The analysis of networks by queuing theory has the disadvantage of knowing inadvance the probability distribution associated with the stochastic process that representsthe dynamics of the random variables that describe the system. Another approach is toview the flow in the network as a dynamic system. In this direction, the network nodescan be seen as varying states in time that interact with each other through communicationlinks [12,40]. The advantage of using the dynamic systems methodology is that all theconcepts, tools and properties of the theory can be applied in the context of network flow.In this sense, a topic of extreme importance is the concept of stability that allows measuringthe characteristic of the system in maintaining its operability, that is, maintaining the flowof communication between network nodes even with disturbances in the input signals.In the literature, the contributions of dynamic systems in the analysis of stability andbifurcations have been shown to be quite intense. In what follows some contributions fromrecent research will be shown in order to contextualize the present work. Challenges innetwork research of underwater wireless sensors, reliability problems in transport flow innetworks, routing protocols in networks, analysis of the dynamics of underwater sensornetwork congestion control models can be found at [12,41–44]. Themes involving stabilityanalysis and bifurcation control can be found at [11,34,45–47]. The influence of the delayin network flow stability is found in [12,45,48]. Stability in impulse systems can be foundat [49].As previously mentioned, bifurcation stability issues are important issues to be con-sidered in the data flow in the network, as branches of bifurcations can generate unstableequilibrium points and are responsible for the emergence of limit cycles and overflow, [50].On the other hand, the presence of delay times can be a source of instability, [12,45,48,51].To analyze the influence of the delay on the network stability, generally two approachesare considered, namely, the delay time can be known or not. Systems where the delay timeis known are very rare to happen. Thus, to determine the delay, analytical methods areapplied or sampling based on numerical algorithms based on models is performed. In thepresent work, the method of numerical algorithms will be applied, which will be done inthe next subsection.5.1. Model DescriptionIn this section, we modeled the network, Figure 10, in an abstract way based onthe Figure 12. In this model, WTIM’s and NCAP are represented, respectively, by si andi = 1, 2, 3; m1, n1 represent, respectively, the density of information flow given in bits/m3or packets/m3; the communication links with the transfer rate ri, i = 1, 2 represent thetransfer rate given in bits/s or packets/s; K, L represent the communication capacity ofthe links li, i = 1, 2; the bit loss rate due to proximity penalty between si, i = 1, 2, sensorsare denoted by α, β, respectively; θi, i = 1, 2, represent the bit rate sent by the si sensors,i = 1, 2; from the control point of view, the parameters m, n, µ are considered, representingrespectively the loss of bits due to the stack overflow between nodes s1 and s3, the lossof bits between nodes s2 and s3 due to stack overflow, µ represents the bit loss rate dueSensors 2021, 21, 422814 of 22to stack overflow at node s3. The control of network congestion is given by the followingsystem of differential equations.˙x1(t) = r1x1(1 − x1K ) − αx1x2 − mx1x3˙x2(t) = r2x2(1 − x2L ) − βx1x2 − nx2x3˙x3(t) = (m1x1 + n1x2)x3 − µx3 − θ1 f (x1(t − τ))x3 − θ2g(x2(t − τ))x3.(1)It is important to highlight that in the differential equation system, Equation (1),α, β, µ, m, n, θ1 and θ2 together with the functions f, g are control parameters being, in thecase of f, g used to manage the bit distribution of the s1 and s2 sensors.Figure 12. Architecture of the sensor networks.5.2. Stability Analysis and Hopf BifurcationsImportant properties related to the solutions of Equation (1) can be obtained throughthe stability analysis of the linearized equations around the equilibrium points. Thus,proceeding to the linearization of Equation (1) we have˙x = Ax + Bx(t − τ),(2)where τ represents the delay time in sending packages from both s1 and s2 to s3; in thiswork, it is assumed that τ is common for both nodes si, i = 1, 2; A, B ∈ R3×3 are arraysgiven byA=−r1x∗1K−αx∗1−mx∗1−βx∗2− r2x∗2L−nx∗2m1x∗3+n1x∗30.,(3)B=000000−θ1 f ′(x∗1)x∗3−θ2g′(x∗2)x∗30,(4)where the vector x∗ = (x∗1, x∗2, x∗3) are the equilibrium points satisfying the followingcondition:r1K x1 + αx2 + mx3 =r1�nr1K − mβ�x1 + (nα − mr2L )x2 =nr1 − mr2m1x1 + n1x2 − θ1 f (x1) − θ2g(x2) =µ(5)Sensors 2021, 21, 422815 of 22Equation (5) can be rewritten as follows:x2 = L(γ1x1 + γ2)(6)m1x1 + n1L(γ1x1 + γ2) − µ − θ1 f (x1) − θ2g(L(γ1x1 + γ2)) = 0(7)x3 = r1m�1 − x1K�− αm L(γ1x1 + γ2)(8)where γ1 and γ2 are constants given byγ1=(nr1 − Kmβ)K(mr2 − nαL)γ2=(mr2 − nr1)(mr2 − nαL).In Equations (6)–(8), just study the solutions to the Equation (7), as the equilibriumpoint x∗ = (x∗1, x∗2, x∗3) is parameterized in x1. The characteristic of Equation (3) is given byp(λ, τ)=λ3 + b1λ2 + b2λ + b3 + (b4λ + b5)e−λτ(9)whereb1=r1x∗1k+ r2x∗2lb2=r1r2x∗1x∗2kl− αβx∗1x∗2 + mm1x∗1x∗3 + nn1x∗2x∗3b3=mm1r2x∗1x∗2x∗3L+ nn1r1x∗1x∗2x∗3K− αm1nx∗1x∗2x∗3 − βmn1x∗1x∗2x∗3(10)b4=−mθ1 f ′(x∗1)x∗1x∗3 − nθ2g′(x∗2)x∗2x∗3b5=− nr1θ2g′(x∗2)x∗1x∗2x∗3k− mr2θ1 f ′(x∗1)x∗1x∗2x∗3l+mβθ2g′(x∗2)x∗1x∗2x∗3 + nαθ1 f ′(x∗1)x∗1x∗2x∗3.In the presented work, the characteristic equation method will be used which, accord-ing to the work in [52], consists of the study of the stability of Equation (1) investigating theroot distribution region of p(λ, τ) = 0 in the complex plane. Note that the introduction ofthe delay parameter in the Equation (1) makes the stability analysis more complex. Whilein systems without delay, the number of roots of p(λ, τ)��τ=0is finite, for the case with τ ̸= 0,the roots of p(λ, τ) = 0, are infinite in number.However, according to the work in [52], the roots of p(λ, τ) = 0 with a positive realpart, that is, Reλ(τ) such that p(λ(τ), τ) = 0 are upper bound and isolated. Furthermore,the roots of the characteristic equation are isolated on the complex plane. We can then saythat the quasi-polynomial p(λ, τ) given in Equation (9) has a finite number of roots with apositive real part. In particular, the roots of p(λ, τ) in any set of the form{λ = x + iy, a ≤ x ≤ b},with a, b ∈ R arbitrary real numbers are finite.In what follows, we will use the Pontryakin theorem to obtain estimates of the networkparameters that guarantee the asymptotic stability of Equation (1). For this we consider thefollowing polynomials:R(ω, τ)=Rep(iω, τ),S(ω, τ)=Imp(iω, τ),(11)Sensors 2021, 21, 422816 of 22what according to Equation (9) gives usR(ω, τ)=−b1ω2 − b4ω sin(ωτ) + b5 cos(ωτ),S(ω, τ)=−ω3 + b2ω + b4ω cos(ωτ) − b5 sin(ωτ).(12)According to the Pontryakin criterion the system, Equation (1), is asymptotically stable,if R(ω) and S(ω) have real, simple, intertwined roots, and the following condition is truefor every real ω:T(ω, τ) = R(ω, τ) dSdω (ω, τ) − dRdω (ω, τ)S(ω, τ) > 0,(13)for more details on the concepts covered in the Pontryakin criterion see in [52]. FromEquation (13) there are some highlighted points:•the function T(ω, τ) establishes a relationship between the characteristic roots of thequasi-polynomial, Equation (9), and the system parameters, given by the coefficientsbi, i = 1, . . . , 5;•from the inequality T(ω, τ) > 0 it is possible to define regions of stability accordingto the equilibrium points, the system parameters and the delay time;•the equilibrium points, and consequently the functional relationship T(ω, τ) dependon the sampling functions f, g;Table 2 shows the equilibrium point convergence for the parameters, m1, n1, µ, θi, ri,K, l, α, β, m, n, i = 1, 2 and sample distribution functions f = sin(x), g = cos(x).Table 2. Variation of the balance points according to the network parameters.m1n1µθ1θ2r1r2Klαβmnx∗1x∗2x∗30.0600.0150.0800.0200.0500.50.510100.00800.00800.1300.1302.2982.4036.1210.0650.0200.0850.0250.0551.01.015150.00850.00850.1350.1342.1982.4829.3760.0700.0250.0900.0300.0601.51.520200.00900.00900.1400.1382.0572.58912.4880.0750.0300.0950.0350.0652.02.025250.00950.00950.1450.1421.8732.71515.4440.0800.0350.1000.0400.0702.52.530300.01000.01000.1500.1461.6432.85618.2520.0850.0400.1050.0450.0753.03.035350.01050.01050.1550.1501.3663.00720.9200.0900.0450.1100.0500.0803.53.540400.01100.01100.1600.1541.0433.16523.4590.0950.0500.1150.0550.0854.04.045450.01150.01150.1650.1580.6733.32825.8780.1000.0550.1200.0600.0904.54.550500.01200.01200.1700.1620.2583.49428.187The numerical values of the reference parameters contained in the Table 2 weretaken from [40]. The negative equilibrium values shows the system instability, meaningan amount of data above the processing capacity of the devices, in the case of WTIM’sand NCAP. Continuing with the analysis, the next step is determine the parameters bi,i = 1, . . . , 5. This is done by replacing the parameters given in the Table 2 in Equation (11),to obtain the Table 3.Table 3. Value of bi parameters.b1b2b3b4b50.3130.1870.0220.2140.0090.3510.3120.0440.3940.0150.3710.4500.0710.6460.0230.3820.5960.0980.9980.0290.3850.7420.1231.4810.0340.3820.8780.1402.1300.0370.3740.9950.1412.9830.0390.3601.0810.1174.0740.0380.3411.1220.0565.4060.025Sensors 2021, 21, 422817 of 22Figure 13 represents the profile from T(ω, τ) to ω ∈ [−10, 10] and τ ∈ [0, 5] and theparameter vector bi, i = 1, . . . , 5 given by the first line of the Table 3. As can be seen, T(ω, τ)has positive values which, according to the Pontryagin criterion, assure the stability of theequilibria. Based on this analysis, the Figure 14 shows the performance of nodes s1 and s2(WTIM nodes), with respect to the number of bits per sample unit. The profiles presentedshow a stabilization of the number of bits per sample unit around 2.2 and 2.3 bits/sampleunit for nodes s1 and s2, respectively, while Figure 15 shows a stabilization of the node s3(NCAP node) around 15.2 and 15.3 bits/sample unit. Figure 16 shows the performanceof the three nodes with respect to the phase plane, in which presents a stable behaviorconverging to an equilibrium point, compatible with the values found, referring to theequilibrium points shown in the Table 2.Figure 13. Quasi-characteristic polynomial.Figure 14. Performance graph of S1 and S2 nodes.Sensors 2021, 21, 422818 of 22Figure 15. S3 node performance graph.Figure 16. Phase picture of the sensor network.6. ConclusionsIn this paper, we presented a mesh sensor network based on the IEEE 1451 standard,where the NCAP node in a plug-and-play manner detects the WTIM nodes, send requeststo read the WTIM’s sensors, and routes the sensor data back in the network to the NCAPnode that then stores the data. The nodes developed can communicate over long distances,without packet losses. Based on these characteristics, a mesh network can be easily expandedto perform a data acquisition in large areas without modifications in the network. The testsperformed proved that the network is feasible for data collection purposes. In a futuredevelopment, the plan is to insert a range of meteorological sensors in the WTIM’s, analyzepower consumption, and to analyze the influence of meteorological conditions and externalsign interference on communication. There are similar studies in the literature relatedto our proposal. However, in this paper the NCAP was implemented using a structuredSensors 2021, 21, 422819 of 22programming in a Raspberry pi B+ to be autonomous, extensible, in which the mesh networkwas evaluated in a real scenario, analyzing aspects of communication between the nodes.The work developed based on the IEEE 1451.1.2 and IEEE 1451.1.6 standard demonstratedthe increased accessibility to the system from the external network side, being possible to beaccessed through a web based system and IoT with a Bocker on the NCAP network node.A dynamic model was also presented to describe the flow of communication betweenthe NCAP and WTIMs. The numerical results confirm the existence of an equilibriumconfiguration with respect to the shared data flow between the NCAP and the WTIM.Author Contributions: The research was conducted with different professionals specialized in differ-ent areas: (T.F.), professor coordinating the research project, fundraising, methodology, supervision,data collection and sensor network developer. (L.F.), doctoral student, contextualization of the textand part of the development of the sensor network. (A.O.-J.), contextualization of the text, support inthe development of the network, data collection and analysis of the sensor network. (W.M.), concep-tualization, review the entire manuscript and funding acquisition. (S.S.), development of the network,assistance in the implementation of the network in the field, analysis of the data collected, writingand revision of the text. (M.R. (Marcos Rabelo)), Data curation, formal analysis and revision of thetext. (M.R. (Maria Ribeiro)) and (C.S.), responsible for the contextualization of the mathematical partof the text, equating and generating the graphics through the captured data. (I.A.G.), contextualizedthe text and revised it. All authors have read and agreed to the published version of the manuscript.Funding: This research was partially funded by National Council for Scientific and TechnologicalDevelopment-CNPq, process: 431552/2016-9 and partially funded by Fraunhofer Portugal AICOS.M.R. (Maria Ribeiro) acknowledges Fundação para a Ciência e a Tecnologia (FCT) under scholarshipSFRH/BD/138302/2018.Institutional Review Board Statement: Not applicable.Informed Consent Statement: Not applicable.Data Availability Statement: Not applicable.Acknowledgments: The authors are grateful to the Digital Systems Laboratory at the Federal Univer-sity of Catalão—UFCAT, Biotechnology Institute—IBiotec, Computer Science Course and NationalCouncil for Scientific and Technological Development—CNPq.Conflicts of Interest: The authors declare no conflict of interest.AbbreviationsThe following abbreviations are used in this manuscript:APIApplication Programming InterfaceCIoTCritical IoTCLCriticality LevelFPGAField Programmable Gate ArrayGPIOGeneral Purpose Input/OutputGPUGraphics Processing UnitHARTHighway Addressable Remote TransducerHDMIHigh-Definition Multimedia InterfaceHTTPHypertext Transfer ProtocolI2CInter-Integrated CircuitIEEEInstitute of Electrical and Electronic EngineersIoTInternet of ThingsLEDLight Emitting DiodeMACMedia Access ControlMCCMobile Cloud ComputingMQTTMessage Queuing Telemetry TransportMQTT-SNMessage Queuing Telemetry Transport-Sensor NetworkNCAPNetwork Capable Application ProcessorPHPHypertext PreprocessorSensors 2021, 21, 422820 of 22RAMRandom Access MemoryRDFResource Description FrameworkREDRandom Early DetectionRFIDRadio-Frequency IdentificationRSSIReceived Signal Strength IndicationSLSecurity LevelSMNPSimple Network Management ProtocolSPARQLSPARQL Protocol and RDF Query LanguageSPISerial Peripheral InterfaceSSWSemantic Sensor WebTCP/UDPTransmission Control Protocol/User Datagram ProtocolTEDSTransducers Electronic Data SheetUARTUniversal Asynchronous Receiver/TransmitterUSBUniversal Serial BusUTF88-bit Unicode Transformation FormatVHDLVHSIC Hardware Description LanguageVHSICVery-High-Speed Integrated CircuitWMNWireless Mesh NetworkWSNWireless Sensor NetworkWTIMWireless Transducer Interface ModuleXMLExtensible Markup LanguageXMPPExtensible Messaging and Presence ProtocolReferences1.Moi¸s, G.D.; Sanislav, T.; Folea, S.C.; Zeadally, S. Performance Evaluation of Energy-Autonomous Sensors Using Power-HarvestingBeacons for Environmental Monitoring in Internet of Things (IoT). Sensors 2018, 18, 1709. [CrossRef] [PubMed]2.Gupta, G.S.; Quan, V.M. Multi-sensor integrated system for wireless monitoring of greenhouse environment. In Proceedings ofthe 2018 IEEE Sensors Applications Symposium (SAS), Seoul, Korea, 12–14 March 2018; pp. 1–6. [CrossRef]3.Gore, R.W.; Deshpande, D.S. An approach for classification of health risks based on air quality levels. In Proceedings of the 20171st International Conference on Intelligent Systems and Information Management (ICISIM), Aurangabad, India, 5–6 October 2017;pp. 58–61. [CrossRef]4.Essa, A.A.; Zhang, X.; Wu, P.; Abuzneid, A. ZigBee network using low power techniques and modified LEACH protocol. InProceedings of the 2017 IEEE Long Island Systems, Applications and Technology Conference (LISAT), New York, NY, USA,5 May 2017; pp. 1–5. [CrossRef]5.Hsueh, C.; Wen, C.; Ouyang, Y. A Secure Scheme Against Power Exhausting Attacks in Hierarchical Wireless Sensor Networks.IEEE Sens. J. 2015, 15, 3590–3602. [CrossRef]6.Lee, I.; Lee, K. The Internet of Things (IoT): Applications, investments, and challenges for enterprises. Bus. Horiz. 2015, 58, 431–440.[CrossRef]7.Velez, J.; Trafford, R.; Pierce, M.; Thomson, B.; Jastrzebski, E.; Lau, B. IEEE 1451-1-6: Providing common network services overMQTT. In Proceedings of the 2018 IEEE Sensors Applications Symposium (SAS), Seoul, Korea, 12–14 March 2018; pp. 1–6.8.Trafford, R.; Shin, S.; Schmalzel, J.L. Provisioning IEEE Smart Transducer Standards (P21451.1) to Include Health Metrics viaHEDS. In Proceedings of the 2020 IEEE Sensors Applications Symposium (SAS), Kuala Lumpur, Malaysia, 9–11 March 2020;pp. 1–5. [CrossRef]9.Bertsekas, D.; Gallager, R. Data Networks; Prentice-Hall International, Ed.; Prentice-Hall International: Hoboken, NJ, USA, 1992.10.Low, S.H.; Paganini, F.; Wang, J.; Doyle, J.C. Linear stability of TCP/RED and a scalable control. Comput. Netw. 2003, 43, 633–647.[CrossRef]11.Wang, X.F. Controlling bifurcation and chaos in internet congestion control system. In Proceedings of the IEEE 4th WorldCongress on Intelligent Control and Automation (Cat. No. 02EX527), Shanghai, China, 10–14 June 2002; Volume 1, pp. 573–576.12.Dong, T.; Liao, X.; Huang, T. Dynamics of a congestion control model in a wireless access network. Nonlinear Anal. Real WorldAppl. 2013, 14, 671–683. [CrossRef]13.Oliveira-Jr, A.; Cardoso, K.; Sousa, F.; Moreira, W. A Lightweight Slice-Based Quality of Service Manager for IoT. IoT 2020, 1,49–75. [CrossRef]14.Shahzad, G.; Yang, H.; Ahmad, A.W.; Lee, C. Energy-Efficient Intelligent Street Lighting System Using Traffic-Adaptive Control.IEEE Sens. J. 2016, 16, 5397–5405. [CrossRef]15.Mois, G.; Folea, S.; Sanislav, T. Analysis of Three IoT-Based Wireless Sensors for Environmental Monitoring. IEEE Trans. Instrum.Meas. 2017, 66, 2056–2064. [CrossRef]16.García, L.; Parra, L.; Jimenez, J.M.; Lloret, J.; Lorenz, P. IoT-Based Smart Irrigation Systems: An Overview on the Recent Trends onSensors and IoT Systems for Irrigation in Precision Agriculture. Sensors 2020, 20, 1042. [CrossRef]Sensors 2021, 21, 422821 of 2217.Deepika, G.; Rajapirian, P. Wireless sensor network in precision agriculture: A survey. In Proceedings of the 2016 InternationalConference on Emerging Trends in Engineering, Technology and Science (ICETETS), Pudukkottai, India, 24–26 February 2016;pp. 1–4.18.Oliveira-Jr, A.; Resende, C.; Pereira, A.; Madureira, P.; Gonçalves, J.; Moutinho, R.; Soares, F.; Moreira, W. IoT Sensing Platform asa Driver for Digital Farming in Rural Africa. Sensors 2020, 20, 3511. [CrossRef]19.Sali, S.; Parvathi, C.S. Integrated wireless instrument for heart rate and body temperature measurement. In Proceedings of the2nd International Conference for Convergence in Technology (I2CT), Mumbai, India, 7–9 April 2017; pp. 457–463.20.Pan, H.; Chen, S.; Gao, Y.; Deal, B.; Liu, J. An urban informatics approach to understanding residential mobility in Metro Chicago.Environ. Plan. B Urban Anal. City Sci. 2020, 47, 1456–1473. [CrossRef]21.Bibri, S.E. The IoT for smart sustainable cities of the future: An analytical framework for sensor-based big data applications forenvironmental sustainability. Sustain. Cities Soc. 2018, 38, 230–253. [CrossRef]22.Raut, R.D.; Mangla, S.K.; Narwane, V.S.; Gardas, B.B.; Priyadarshinee, P.; Narkhede, B.E. Linking big data analytics andoperational sustainability practices for sustainable business management. J. Clean. Prod. 2019, 224, 10–24. [CrossRef]23.Pan, H.; Page, J.; Cong, C.; Barthel, S.; Kalantari, Z. How ecosystems services drive urban growth: Integrating nature-basedsolutions. Anthropocene 2021, 35, 100297. [CrossRef]24.Tiwari, K.; Khan, M.S. Sustainability accounting and reporting in the industry 4.0. J. Clean. Prod. 2020, 258, 120783. [CrossRef]25.Srivastava, M.; Kumar, R. Smart Environmental Monitoring Based on IoT: Architecture, Issues, and Challenges. In Advancesin Computational Intelligence and Communication Technology; Gao, X.Z., Tiwari, S., Trivedi, M.C., Mishra, K.K., Eds.; Springer:Singapore, 2021; pp. 349–358.26.Ga¸sparesc, G.; Papazian, P.; Babata, M. IEEE 1451 compliant smart HART modem. In Proceedings of the 38th InternationalConference on Telecommunications and Signal Processing (TSP), Prague, Czech Republic, 9–11 July 2015; pp. 294–298.27.Singh, A.K.; Raj, M.; Sharma, V. Architecture, Issues and Challenges in Monitoring based on IoT for Smarter Environment. InProceedings of the 2020 Fourth International Conference on Computing Methodologies and Communication (ICCMC), Erode,India, 11–13 March 2020; pp. 142–146. [CrossRef]28.Wu, C.K.; Tsang, K.F.; Liu, Y.; Zhu, H.; Wang, H.; Wei, Y. Critical Internet of Things: An Interworking Solution to Improve ServiceReliability. IEEE Commun. Mag. 2020, 58, 74–79. [CrossRef]29.Lee, H.; Lin, H. Design and Evaluation of an Open-Source Wireless Mesh Networking Module for Environmental Monitoring.IEEE Sens. J. 2016, 16, 2162–2171. [CrossRef]30.Ali, S.; Ashraf, A.; Qaisar, S.B.; Afridi, M.K.; Saeed, H.; Rashid, S.; Felemban, E.A.; Sheikh, A.A. SimpliMote: A Wireless SensorNetwork Monitoring Platform for Oil and Gas Pipelines. IEEE Syst. J. 2018, 12, 778–789. [CrossRef]31.Pinto, R.; Pereira, J.; da Rocha, H.; Martin, R.I.; Espírito-Santo, A. A Discussion about the Implementation of a WSN to Industry4.0 based on the IEEE 1451 Standard. In Proceedings of the 2019 IEEE 17th International Conference on Industrial Informatics(INDIN), Helsinki, Finland, 22–25 July 2019; Volume 1, pp. 1573–1578.32.Pereira, J.L.D.; da Rocha, H.; Santo, A.E. A Platform for IEEE 1451 Standard’s Education, Development and Validation forIndustry 4.0. In Proceedings of the 2020 IEEE International Instrumentation and Measurement Technology Conference (I2MTC),Dubrovnik, Croatia, 25–28 May 2020; pp. 1–6.33.Rasyid, M.U.H.A.; Sayfudin, A.; Basofi, A.; Sudarsono, A. Development of semantic sensor web for monitoring environmentconditions. In Proceedings of the 2016 International Seminar on Intelligent Technology and Its Applications (ISITIA), Lombok,Indonesia, 28–30 July 2016; pp. 607–612. [CrossRef]34.Ding, D.; Qin, X.; Wu, T.; Wang, N.; Liang, D. Hopf bifurcation control of congestion control model in a wireless access network.Neurocomputing 2014, 144, 159–168. [CrossRef]35.Kumar, A.; Hancke, G. Energy Efficient Environment Monitoring System Based on the IEEE 802.15.4 Standard for Low CostRequirements. IEEE Sens. J. 2014, 14, 2557–2566. [CrossRef]36.Kumar, A.; Srivastava, V.; Singh, M.K.; Hancke, G.P. Current Status of the IEEE 1451 Standard-Based Sensor Applications. IEEESens. J. 2015, 15, 2505–2513. [CrossRef]37.IEEE Standards Association. IEEE Standard for a Smart Transducer Interface for Sensors and Actuators—Network Capable ApplicationProcessor Information Model; IEEE: Piscataway, NJ, USA, 2000; pp. 1–480. [CrossRef]38.Li, Y.; Cheng, J.; Wang, X. An Optophone Based on Raspberry Pi and Android Wireless Communication. In Proceedings of the2020 IEEE International Conference on Advances in Electrical Engineering and Computer Applications (AEECA), Dalian, China,25–27 August 2020; pp. 952–956. [CrossRef]39.Romashchenko, V.; Brutscheck, M.; Chmielewski, I. Investigation and Implementation of Robust Wireless Zigbee Based SecuritySystem. In Proceedings of the 2018 29th Irish Signals and Systems Conference (ISSC), Belfast, UK, 21–22 June 2018; pp. 1–6.[CrossRef]40.Dong, T.; Hu, W.; Liao, X. Dynamics of the congestion control model in underwater wireless sensor networks with time delay.Chaos Solitons Fractals 2016, 92, 130–136. [CrossRef]41.Akyildiz, I.F.; Pompili, D.; Melodia, T. Underwater acoustic sensor networks: Research challenges. Ad Hoc Netw. 2005, 3, 257–279.[CrossRef]42.Akan, O.; Akyildizi, I. Event-to-sink reliable transport in wireless sensor networks. IEEE/ACM Trans. Netw. 2005, 13, 1003–1016.[CrossRef]Sensors 2021, 21, 422822 of 2243.Akkaya, K.; Younis, M. A survey on routing protocols for wireless sensor networks. Ad Hoc Netw. 2005, 3, 325–349. [CrossRef]44.Jiang, J.; Han, G.; Guo, H.; Shu, L.; Rodrigues, J.J. Geographic multipath routing based on geospatial division in duty-cycledunderwater wireless sensor networks. J. Netw. Comput. Appl. 2016, 59, 4–13. [CrossRef]45.Rezaie, B.; Motlagh, M.R.J.; Analoui, M.; Khorsandi, S. Stabilizing fixed points of time-delay systems close to the Hopf bifurcationusing a dynamic delayed feedback control method. J. Phys. A Math. Theor. 2009, 42, 395102. [CrossRef]46.Hong, K.S. Hopf bifurcation control via a dynamic state-feedback control. Phys. Lett. A 2012, 376, 442–446. [CrossRef]47.Liu, F.; Wang, H.O.; Guan, Z.H. Stability analysis and control of bifurcation in a TCP fluid flow model of wireless networks. InProceedings of the 10th WCICA, Beijing, China, 6–8 July 2012; pp. 1026–1030.48.Zhang, S.; Xu, J.; Chung, K.W. On the stability and multi-stability of a TCP/RED congestion control model with state-dependentdelay and discontinuous marking function. Commun. Nonlinear Sci. Numer. Simul. 2015, 22, 269–284. [CrossRef]49.Ruan, S.; Wei, J. On the zeros of transcendental functions with applications to stability of delay differential equations with twodelays. Dyn. Contin. Discret. Impuls. Syst. Ser. A 2003, 10, 863–874.50.Diniz, P.; da Silva, E.; Netto, S. Digital Signal Processing: System Analysis and Design; Cambridge University Press: Cambridge,UK, 2010.51.Huang, C.; Cao, J.; Alsaedi, M.X.A.; Alsaadi, F.E. Controlling bifurcation in a delayed fractional predator-prey system withincommensurate orders. Appl. Math. Comput. 2017, 293, 293–310. [CrossRef]52.Hu, H.; Wu, Z. Stability and Hopf bifurcation of four-wheel-steering vehicles involving driver’s delay. Nonlinear Dyn. 2000,22, 361–374. [CrossRef] | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/367073040JORNALISMO E ENGAJAMENTO CLIMÁTICOTechnical Report · January 2023CITATIONS0READS109 authors, including:Some of the authors of this publication are also working on these related projects:Os riscos climáticos no circuito da notícia local: percepção, comunicação e governança View projectObservatório de Jornalismo Ambiental View projectEloisa Beling LooseUniversidade Federal do Rio Grande do Sul89 PUBLICATIONS 170 CITATIONS SEE PROFILEI.M.T. GirardiUniversidade Federal do Rio Grande do Sul27 PUBLICATIONS 82 CITATIONS SEE PROFILEDébora Gallas SteiglederUniversidade Federal do Rio Grande do Sul13 PUBLICATIONS 23 CITATIONS SEE PROFILEEliege Maria FanteUniversidade Federal do Rio Grande do Sul32 PUBLICATIONS 30 CITATIONS SEE PROFILEAll content following this page was uploaded by Eloisa Beling Loose on 12 January 2023.The user has requested enhancement of the downloaded file.JORNALISMO E ENGAJAMENTO CLIMÁTICOPESQUISA 2022© 2022. Modefica Esta obra está licenciada com uma Licença Creative Commons 4.0 International (CC BY-NC-SA 4.0)JORNALISMO E ENGAJAMENTO CLIMÁTICOINSTITUTO MODEFICAGRUPO DE PESQUISA JORNALISMO AMBIENTAL (GPJA)COMO CITAR ESTE DOCUMENTO PALAVRAS CHAVESDIRETORA PRESIDENTE Marina ColeratoCOORDENAÇÃO DO PROJETO Marina ColeratoPROJETO GRÁFICO Alain SchranerCOORDENAÇÃO DE PEQUISA Eloisa Beling Loose e Ilza GirardiAUTORAS E AUTORES Bibiana DavilaCaroline Maldaner JacobiDébora Gallas SteiglederEliege FanteEloisa Beling LooseEvilene PaixãoLásaro ThiesenMODEFICA. Jornalismo e Engajamento Climático. São Paulo, 2022. 1. Jornalismo Ambiental 2. Mudanças Climáticas 3. Estudo de RecepçãoO GPJA, registrado no CNPq em 2008, é voltado para os estudos das interfaces entre jornalismo e meio ambiente. Coordenado pela professora Dra. Ilza Maria Tourinho Girardi, no âmbito da Universidade Federal do Rio Grande do Sul (UFRGS), tem como objetivo contribuir com a constituição de um referencial teórico na área e a ampliação do debate ambiental no meio jornalístico. As principais atividades do GPJA hoje envolvem as investigações científicas, coletivas e individuais, a organização dos Encontros Nacionais de Pesquisadores em Jornalismo Ambiental (ENPJA) e a produção semanal de análises críticas para o Observatório de Jornalis-mo Ambiental. Acesse: jornalismoemeioambiente.comO Modefica é uma organização de mídia, pesquisa e educação que trabalha para justiça ambiental e climática com uma perspectiva ecofeminista. Nós procuramos fazer do jornalismo, da pesquisa e da educação ferramentas de transformação, colaborando com causas sociais e ajudando a construir uma narrativa de responsa-bilização e autoridades privadas. Nossa missão é aumentar a consciência sistêmi-ca de problemas ambientais para promover relações de equidade e justiça entre humanos, não humanos e a Natureza. Acesse: modefica.com.br GRUPO DE PESQUISA JORNALISMO AMBIENTAL (GPJA)INSTITUTO MODEFICA09153975538892SUMÁRIOAPRESENTAÇÃO PONTO DE PARTIDA METODOLOGIACONSIDERAÇÕES FINAISGRUPOS FOCAISREFERÊNCIAS BIBLIOGRÁFICAS APÊNDICES" Para engajar precisamos fazer algo diferente, através da construção de narrativas que legitimam as experiências locais das pessoas e povos, que desvelam os diferentes papéis dos grupos sociais. É importante evidenciar as interconexões dos temas abordados, de modo que aproximem as causas e consequências das mudanças climáticas com a vida das pessoas, bem como com as possibilidades de transformação positiva na relação sociedade e clima. (GF SUDESTE, 18.06.2022)PESQUISA 20229APRESENTAÇÃOConforme nosso mundo fica mais quente e governos e corporações ficam cada vez mais distantes de cumprir metas e acordos globais no âmbito climático, cresce a atenção para os potenciais - e os limites - do jornalismo para um engajamento cívico pró-clima. De acordo com os cientistas, o jornalismo desempenha um importante papel na amplificação da discussão, definição de sentidos, apresentação de argumentos, valores e visões de mundo a respeito das mudanças climáticas. Sobretudo o jornalismo ambiental é entendido como uma das perspectivas emergentes que podem colaborar com o engaja-mento climático na medida em que seus pressupostos reforçam o compromisso com a sustentabilidade da vida, e assumem a crítica ao capitalismo, à colonialidade e à lógica do Norte Global. Ao se apre-sentar como uma prática engajada, o jornalismo ambiental assume como função política a sensibilização dos públicos para formas mais efetivas de enfrentamento à crise climática, entendendo que a in-formação qualificada auxilia em tomadas de decisão mais acertadas e alinhadas ao exercício da cidadania planetária. Contudo, será que essa é a mesma percepção da audiência? Para além de entender a percepção do público sobre a atual co-bertura climática na mídia, a presente pesquisa buscou levantar insumos para responder a questão: como os jornalistas promo-vem ou podem promover engajamento das pessoas nas ações pró-clima no seu fazer jornalístico? Gunster (2017) afirma que o jornalismo pode contribuir com o engajamento a partir de quatro pontos, sendo o primeiro deles a priorização de públicos pro-pensos a se envolver com o tema (em oposição à destinação de mensagens genéricas). Dessa forma, centramos nossa pesquisa num público previamente engajado com temas sociais (aqueles que se percebem como ativistas) e, portanto, mais predisposto ao engajamento cívico, entre 18 e 35 anos. O recorte de idade foi orientado por pesquisas de comportamento que mostram a maior preocupação dos jovens da geração Y e Z com questões ambien-tais. Ainda sobre o público, destacamos que entendemos esses sujeitos como qualificados e acreditamos que os insumos gerados Por Eloisa Loose e Marina Colerato 10JORNALISMO E ENGAJAMENTO CLIMÁTICOPESQUISA 2022APRESENTAÇÃO11por esse grupo podem colaborar de forma mais significativa com os objetivos desta investigação, gerando assim insights relevantes para profissionais da comunicação. Considerando as relações entre crise climática, desmatamento e Amazônia no Brasil, a presente pesquisa buscou entender a per-cepção do público sobre tais conexões e sobre a clareza das infor-mações transmitidas pela mídia sobre as mesmas. Até que ponto tal relação faz sentido para os brasileiros das diferentes regiões do país? As diferenças associadas às realidades locais é outro ponto que atravessa o horizonte do estudo.Desde a concepção da proposta, nosso objetivo foi coletar percep-ções e entendimentos de jovens ativistas de todo o Brasil sobre engajamento e jornalismo climático e, ao mesmo tempo, permitir que o espaço dos grupos de discussão fosse de reflexão sobre a interseccionalidade da pauta climática, de modo a integrá-la às demais causas. Apesar das restrições de tempo, avaliamos, a partir das falas dos participantes, que a dinâmica se mostrou positiva para além da recolha de dados.Nossos esforços nesta pesquisa tentam aproximar prática e ciên-cia para aprimorar a comunicação climática, pois acreditamos que conhecer melhor os entendimentos dos públicos pode reorientar estratégias e ampliar o debate em um sistema de múltiplas informa-ções, que concorre com a desinformação. Para o Instituto Modefica, o presente estudo busca também contribuir com o financiamento de pesquisas científicas e com a aproximação entre academia e sociedade. Para o Grupo de Pesquisa Jornalismo Ambiental (CNPq/UFRGS), é uma oportunidade para trabalhar com estudos de recep-ção e se debruçar, ainda mais, sobre as interfaces entre jornalismo e engajamento. Por fim, registramos nossos agradecimentos às pessoas que responderam ao questionário online da primeira etapa e, especialmente, aquelas e aqueles que aceitaram voluntariamente integrar os grupos focais, que consistiram na segunda etapa. A pesquisa "Jornalismo e Engajamento Climático" aponta que os ativistas participantes dos grupos focais das cinco regiões brasi-leiras concordam uns com os outros muito mais do que discor-daram (em raras ocasiões houve confronto de opiniões). Embora nem todos façam clara distinção entre o que é jornalismo e outras formas de comunicação/divulgação, as percepções associadas ao engajamento climático estavam mais na ordem do que deveria ser (no sentido prescritivo) do que, de fato, eles acompanhavam. Na concepção majoritária dos participantes, o jornalismo tradicional/mainstream sobre clima (identificado nos vídeos apresentados) possui limites para mobilizar, já que são revestidos de uma pers-pectiva predominantemente científica, com termos técnicos, e não se aproximam do dia a dia de seus públicos.Ainda que reconheçam o papel informativo do que é apresenta-do, sugerem que é preciso ter posicionamentos mais categóricos, denunciando os responsáveis e convocando a população para se unir em ações concretas, de forma explícita. Formas especializadas e segmentadas de jornalismo foram citadas como mais capazes de envolver o público, porém entende-se que a partir da imprensa massiva muitos terão contato com o tema.Apesar das especificidades regionais, relacionadas ao acesso de informações jornalísticas e à ausência de produção conectada à re-alidade local, pode-se destacar sete pontos-síntese resultados do estudo de recepção empreendido com um público qualificado sob o foco no jornalismo e engajamento climático: ǿ Na visão dos respondentes, para engajar, o jornalismo precisa ouvir os mais afetados; ǿ O jornalismo precisa pensar nas soluções/respostas para além das ações individuais, reforçando a orientação para um enga-jamento político (desconstrução da ideologia neoliberal, que enfatiza a responsabilização do sujeito);RESUMO EXECUTIVO12JORNALISMO E ENGAJAMENTO CLIMÁTICOPESQUISA 2022APRESENTAÇÃO13 ǿ As causas e os responsáveis pela crise climática precisam ser mais expostos; ǿ Sobre o efeito de choque, verifica-se que, mesmo entre os ativistas, há percepções diferentes o quanto que a ênfase no medo pode provocar reação ou apatia; ǿ Em conformidade com o que já é apontado pela literatura cien-tífica, os ativistas destacam para o uso de uma linguagem clara e simples (crítica aos termos técnicos), a aproximação do tema com a realidade próxima do leitor (crítica à perspectiva nacio-nal, oriunda do eixo Rio-São Paulo), “humanização” do debate (reforçada pela necessidade de dar protagonismo aos defenso-res da floresta e outras populações que atuam localmente), a credibilidade ou confiabilidade associadas ao comunicador ou meio de comunicação (críticas aos grandes grupos que se po-sicionam a favor do agronegócio, por exemplo), a necessidade de evidenciar as diferentes implicações do clima para setores e aspectos sociais (aproximar clima com alimentação, saúde, economia, desigualdades sociais, etc.), o trabalho em rede (em parceria com as comunidades), o conhecimento prévio dos pú-blicos e suas demandas (reconhecendo a necessidade de uma produção segmentada); ǿ Os ativistas indicam que o jornalismo mainstream deve incor-porar aspectos do jornalismo não hegemônico (especialmente o comunitário, o engajado ou de advocacy, e o ambiental), ao mesmo tempo em que reconhecem os obstáculos estruturais dessa demanda; ǿ O público pesquisado possui um entendimento amplo do que é referência jornalística, mesclando aspectos do jornalismo tradicional e alternativo, em simultâneo com outras produções derivadas de organizações ambientais.VISUALIZAR AS CONSEQUÊNCIAS E AS PESSOAS AFETADAS ↑ Famílias Impactadas por Enchentes no Brasil. Foto: © Isis Medeiros / Greenpeace " Conectar as pessoas com possíveis instrumentos de transformação, através de exemplos a serem seguidos, como iniciativas locais, sejam coletivas e individuais, projetos em andamento nas comunidades. Não relativizar as informações, trazendo a realidade com transparência, verdade e pé no chão. Também contemplar a racialização, denunciando a urgência da justiça climática e cultivando a esperança. (GF SUL, 22.06.2022)PESQUISA 202215PONTO DE PARTIDANo Brasil, as investigações de recepção orientadas para materiais jor-nalísticos são escassas e não há um estudo representativo com foco na questão climática. No entanto, para avançarmos em estratégias de comunicação mais assertivas, buscar saber como os grupos enten-dem e fazem uso dessas informações é fundamental. Esta pesquisa busca compreender como se dá o engajamento pró-clima por meio do jornalismo, sobretudo para ampliar o entendimento a respeito das relações entre desmatamento da Amazônia e crise climática.Para tanto, realizou-se um levantamento bibliográfico sobre estudos de recepção que envolvam cobertura jornalística de clima, o que nos confirmou que ainda são poucas as pesquisas que tratem da relação entre o consumo jornalístico e o engajamento climático. Na sequência, discutiu-se qual seria o perfil da recepção mais interes-sante para trabalhar com engajamento a partir de achados de outras pesquisas, tentativa de ouvir grupos das diferentes regiões e enten-dendo que há perfis que poderiam contribuir mais com a discussão do engajamento, como o de jovens ativistas. Contudo, a intenção não foi limitar a pesquisa aos ambientalistas, pressupondo que tais informantes teriam mais conhecimento sobre o clima; o objetivo da pesquisa foi trazer a questão climática para outros movimentos e pautas, sublinhando sua transversalidade.Para coleta de dados, foram mobilizados grupos focais (GATTI, 2005) nos meses de junho e julho de 2022, de forma virtual, com o intuito de ouvir percepções sobre o tema e também observar reações a partir da exibição de dois produtos audiovisuais com características tradicio-nais do jornalismo. Os grupos focais foram gravados e, posteriormen-te, as discussões foram transcritas. Para sua análise, houve a categori-zação das falas, a partir da Análise de Conteúdo (BARDIN, 1979). Os respondentes participaram de forma voluntária sob o resguardo do anonimato, e os convites foram realizados a partir de uma pri-meira etapa de mapeamento, por meio de um questionário virtual disponibilizado nas redes sociais do Modefica, e, posteriormente, segundo indicações, aplicando a técnica conhecida como “bola de neve” (snowball sampling).PONTO DE PARTIDA16JORNALISMO E ENGAJAMENTO CLIMÁTICOPESQUISA 202217As mudanças climáticas têm representado um dos grandes desafios da humanidade há várias décadas (APPELGREN; JÖNSSON, 2021). O último relatório produzido pelo Painel Intergovernamental sobre Mudanças Climáticas (IPCC, na sigla em inglês) afirma que precisamos fazer mudanças drásticas para evitar os piores cenários, evidenciados am-plamente pelos cientistas. Além de reduzir as emissões de gases de efeito estufa (GEE), é preciso lembrar que as emissões passadas se-guirão intensificando a mudança do clima. Desde a era pré-industrial, a humanidade emitiu 2,4 trilhões de toneladas de CO2, sendo que 58% desse total foi lançado entre 1850 e 1989, e 42% entre 1990 e 2019 (IPCC, 2022), o que denota uma relação entre a globalização neoliberal e o aumento das emissões de gases de efeito. Soma-se à identificação da aceleração das emissões de GEE os estudos que mostram como as políticas públicas adotadas até 2020 são insufi-cientes: elas promoveriam um aquecimento de 3,2 graus, mais que o dobro do que o limite estabelecido no Acordo de Paris (2015).Outro destaque desse último relatório (IPCC, 2022) diz respeito à (in)justiça climática, ou seja, o fato daqueles que mais sofrem com os impactos desse processo serem justamente os que menos con-tribuem para o problema. Há uma distribuição bastante desigual de riscos, efeitos e oportunidades associadas à crise climática, que precisa ser socialmente reconhecida e inserida com mais frequência no debate público. A ação humana, responsável pela potencialização do fenômeno climático, não é homogênea. Um levantamento da agência de classificação de risco S&P Global (S&P GLOBAL, 2022) em 135 países a respeito dos efeitos derivados da crise climática, como eventos extremos, apontou para uma perda de 4% do PIB mundial por ano até 2050, sendo que, nos piores cenários, os países mais vulneráveis aos riscos climáticos, como os do sul da Ásia, podem acumular perdas anuais de até 18% do PIB - o triplo do estimado na América do Norte. As emissões de GEE estão vinculadas aos modos de vida moder-nos, orientados por um modelo de desenvolvimento que não incor-PANORAMA DAS PESQUISAS NA ÁREA DE COMUNICAÇÃO CLIMÁTICApora os limites ecossistêmicos. Isso revela que a questão climática possui dimensões sociais, culturais, políticas e econômicas, extra-polando em muito a perspectiva reducionista de que meio ambien-te seria apenas fauna e flora. Obviamente há uma conexão forte com a perda da biodiversidade, mas até mesmo esta última precisa ser observada de forma integrada, a partir das relações humanas com o ambiente.No Brasil, a principal causa das emis-sões de GEE é a mudança de uso da terra, decorrente do desmatamento e, muitas vezes, associada à expansão da fronteira agropecuária. De acordo com o último relatório do Sistema de Estimativas de Emissões e Remoções de Gases de Efeito Estufa - SEEG (2021), 46% do total bruto (998 milhões de toneladas de CO2) foram emitidas em 2020 em razão disso, seguido de 27% oriundas da atividade agropecuária e 18% relacionadas ao setor de energia. Todavia, a fiscalização e as ações de combate a essas práticas destrutivas seguem sendo pífias. Segundo o Monitor da Fiscalização (MAPBIOMAS ALERTA, 2022), 98% dos alertas de desma-tamento registrados no Brasil, desde janeiro de 2019, não foram autorizados ou foram alvos de fiscalização por órgãos do governo federal, e é na Amazônia que se concentram três quartos desses alertas (149.631). Isso significa que há um índice altíssimo de im-punidade, que favorece a prática de desmatamento ilegal no país.Contudo, o problema não se concentra apenas na Amazônia, a despeito de sua relevância para o contexto planetário. Existe um afrouxamento das políticas públicas ambientais já há alguns anos em curso, que ganhou força a partir da gestão do governo Jair Bolsonaro em todo o País (ARAGÃO DOS SANTOS ET AL., 2021; FEARNSIDE, 2019). No caso específico do enfrentamento climático, para além das tentativas de se negar a existência do problema, o relatório do SEEG (2021) conclui que, encerrado o ciclo inicial da PNMC [Política Nacional sobre Mudança do Clima], em 2020, o Brasil não possui plano para cumprir a meta do Acordo de Paris, sendo que o assunto tem sido negligenciada pelo governo federal.Logo, as iniciativas brasileiras que estão fazendo frente à crise climática hoje emergem, em grande medida, da articulação da sociedade civil, que ocupou o vácuo deixado pela fragilização e No Brasil, a principal causa das emissões de GEE é a mudança de uso da terra, decorrente do desmatamento e, muitas vezes, associada à expansão da fronteira agropecuária.PONTO DE PARTIDA18JORNALISMO E ENGAJAMENTO CLIMÁTICOPESQUISA 202219inoperância das instituições públicas. Nesse sentido, é importante conhecer como a pauta climática é compreendida por esses grupos comprometidos em transformar a realidade e observar como a comunicação jornalística circula nesse âmbito - e de qual forma contribui para o debate e participação social.Estudos de opinião já revelaram que os latino-americanos (WORLD RISK POLL, 2019) possuem uma preocupação grande com a questão ambiental, concordando que as ações humanas tinham um peso sig-nificativo na intensificação das mudanças climáticas (a maioria com-preende que não é um fenômeno apenas natural). Dentre 26 países investigados, o Brasil ficou na oitava posição dos que percebem as mudanças climáticas como uma grande ameaça (PEW RESEARCH, 2019).Pesquisas recentes realizadas pelo ITS-Rio e Yale University sobre a percepção em relação às mudanças climáticas em 2021 e 2022 reforçam que os brasileiros dão muita importância à questão do aquecimento global: em 2020, 78% declararam que consideram o tema muito importante e, em 2021, essa pro-porção foi de 81%. Embora o grau de preocupação seja alto, o conhecimento a respeito da problemática ainda é insuficiente: apenas 21% consideram saber muito sobre o tema, enquanto a maior parte dos respondentes (46%) avaliam que sabem mais ou menos sobre a questão climática.Perguntas específicas sobre a preservação da Amazônia e as queimadas, derivadas dos acontecimentos de 2019 e 2020, foram feitas pensando no contexto nacional. De forma geral, os resulta-dos demonstram a preocupação dos brasileiros com a região, entendendo que sua destruição impacta na qualidade de vida e que seu cuidado pode ser um ativo econômico e político nas relações internacionais. Além de madeireiros, agricultores e pecua-ristas e criadores de animais serem apontados como os responsáveis pelas queimadas, políticos também foram citados (ITS-RIO/YALE, 2022).Segundo essas pesquisas, há vários fatores em jogo, como o fato de a maior escolaridade estar relacionada à maior preocupação com o tema. Outros aspectos, como gênero e posicionamento político, também foram observados: mulheres e sujeitos situados à esquer-da, em termos políticos, se declaram mais preocupados com o meio ambiente do que homens e aqueles posicionados ao centro ou à di-reita. Os jovens afirmam ter mais preocupação em relação às demais faixas etárias. O acesso à informação é outro fator bastante inves-tigado: na última pesquisa do ITS-Rio (2022) a proporção dos que disseram saber muito sobre a questão foi maior entre os usuários de internet em comparação aos não usuários. Cabe destacar que o tipo de informação acessada não foi questionado neste estudo, existindo várias lacunas a serem preenchidas no que tange o consumo e recepção das informações ambientais.Nesta pesquisa, debruça-se sobre a produção jornalística, enten-dendo que esse tipo de informação segue orientações éticas e de-ontológicas, que conduzem à verificação dos fatos, diferentemente da produção de outros conteúdos. Além disso, a credibilidade e o alcance da imprensa são elementos que justificam os estudos de como tais produtos estão sendo consumidos. Diferentes autores si-tuam essa área como uma arena-chave para amplificar a discus-são, definir sentidos, apresentar argumentos, valores e visões de mundo a respeito das mudanças climáticas (LOOSE; CARVALHO, 2017; HULME, 2009). O jornalismo é um tipo de produção específica den-tro da área de comunicação climática, que, de acordo com Moser (2010), engloba diferentes níveis: o de informar e educar as pes-soas sobre a problemática climática, o de buscar envolvê-las para que exerçam sua cidadania, e o de promoção para que haja mudanças individuais (algo além das pressões políticas), que se estenderia para a transformação de normas, valores e ideias Ativistas do Greenpeace colocam uma faixa ao redor de uma castanheira "Castanheira" derrubada. Foto: ©Greenpeace / Daniel Beltrá VISUALIZAR AS CAUSAS E OS PRINCIPAIS RESPONSÁVEIS ↓PONTO DE PARTIDA20JORNALISMO E ENGAJAMENTO CLIMÁTICOPESQUISA 202221do próprio contexto cultural. Importante dizer ainda que, para que os fatos ambientais sejam transformados em preocupação pública (e, quiçá, em ações para evitá- los ou manejá-los), a mediação da imprensa se faz fundamental (HANNIGAN, 1995).Embora a questão climática tenha, gradativamente, recebido mais atenção midiática nas últimas décadas, os estudos de interface com o campo da Comunicação, em geral, e com o subcampo do Jornalis-mo, especificamente, ainda apresentam uma série de lacunas, so-bretudo nos países do Sul Global. Historicamente, as pesquisas so-bre essa interface partem de países ditos desenvolvidos, sobretudo os europeus e os Estados Unidos (SCHÄFER; SCHLICHTING, 2014), com foco em análise de representações sobre o tema em jornais hegemô-nicos. Conforme Moser (2010), desde que o debate climático com ênfase na ação antropogênica ganhou espaço, na década de 1980, há esforços em direção em como melhor comunicar essa temática. Entretanto, os estudos de recepção ou voltados para como se dá o consumo das informações climáticas permanecem sendo pontuais, o que dificulta a construção de estratégias mais assertivas, que poderiam, de fato, ter o potencial de mobilizar os públicos.Além da falta de pesquisas na área, é preciso considerar que o aspecto comunicacional é um, dentre outros fatores, que podem ser associados à mudança de atitudes e comportamen-tos e colaborar com o enfrentamento climático. Ainda que o engajamento seja cada vez mais acionado neste debate, seu entendimento é bastan-te difuso, sendo que há autores que defendem, no contexto climático, a necessidade de um engajamento polí-tico, que extrapole as ações individuais (CARVALHO; VAN WASSEL; MAESEELE, 2016). De acordo com esses autores, o engajamento político encoraja a oposição às políticas dominantes e coloca em relevo a agência dos cidadãos, sendo que o engajamento público está mais associado à implementação de políticas governamentais de cima para baixo e concebe os cidadãos como um público a ser mobilizado.A maioria dos estudos de comunicação climática, em diferentes países, têm identificado que os meios de comunicação utilizam poucos recursos visuais e de animação com fins informativos e ex-plicativos, além de dar mais destaque para as consequências/efei-tos do que as causas (PARRATT FERNÁNDEZ; MERA FERNÁNDEZ; ABEJÓN MENDONZA, ... é preciso considerar que o aspecto comunicacional é um, dentre outros fatores, que podem ser associados à mudança de atitudes e comportamentos e colaborar com o enfrentamento climático.2021; LOOSE, 2021). Também já foi assinalada a pouca atenção dada à cobertura das ações de mitigação e adaptação, que prejudica a tomada de decisão participativa em relação ao futuro (MOSER, 2017). Soma-se a isso o fato das notícias climáticas serem, em grande parte, derivadas de agências internacionais (VÉLEZ ET AL., 2017), que abarcam predominantemente o tema a partir da mitigação e como uma oportunidade de negócio (TAKAHASHI; MARTÍNEZ, 2017).Há estudos que revelam que os mais afetados nem sempre são ouvidos nas reportagens sobre clima, como é o caso dos povos indígenas (LOOSE, 2021; STODDART; SMITH, 2016), sinalizando para a neces-sidade de enquadramentos a partir da justiça climática. Ademais, já são bem conhecidos os gatilhos que desencadeiam a cobertura do clima, sendo as COPs (Conferências das Partes da Convenção--Quadro das Nações Unidas sobre Mudança do Clima), a divulga-ção dos estudos científicos, sobretudo os do IPCC, e os eventos extremos climáticos acontecimentos que alavancam a discussão sobre a crise climática (LOOSE, 2021).Assim como na discussão de outros temas ambientais, as respostas recorrentemente visibilizadas como soluções/alternativas estão cen-tradas nos indivíduos (enquanto consumidores), esvaziando o debate sobre o papel das estruturas sociais ou da cidadania. Federovisky (2018) destaca que a ideologia neoliberal responsabiliza os sujeitos enquanto desvia a atenção das verdadeiras causas da deterioração ambiental/climática. Além disso, a despolitização ocorre quando o tema é transformado em um discurso calcado em modelos e núme-ros, onde questões muito distantes dos cidadãos tornam-se centrais, como ocorre quando o tema é reduzido ao funcionamento do merca-do de carbono, por exemplo (CARVALHO; VAN WASSEL; MAESEELE, 2016).Soma-se a isso o fato de que o debate da responsabilização pode ser representado de modo a afastar a questão do dia a dia dos sujeitos. Quando há ênfase sobre o aspecto global, países ditos desenvolvidos costumam ser alvo, o que não deve eximir os demais países de agi-rem em prol da mitigação. A complexidade da questão demanda que se entenda o problema em todas as escalas e setores, exigindo ações de cima para baixo e de baixo para cima em simultâneo (GIDDENS, 2010).Recorda-se que a percepção do que é ou não um problema e de como se pode estar envolvido ou não nas suas causas não depende apenas de informação bem apurada e contextualizada. Há uma sé-rie de fatores culturais, econômicos, sociais e até psicológicos que precisam ser considerados a fim de que as estratégias utilizadas facilitem a mobilização (LOOSE, 2020).PONTO DE PARTIDA22JORNALISMO E ENGAJAMENTO CLIMÁTICOPESQUISA 202223No Brasil, há alguns (poucos) estudos que buscam entender as percepções/entendimentos dos públicos em relação à questão climática, porém baseados em públicos bem específicos, já fami-liarizados com o tema, como o desenvolvido por Cavalcanti (2019) que entrevistou informantes que atuavam em ONGs e trabalhavam diretamente com mudanças climáticas, ou o de Loose (2020), que buscou leitores que tivessem interesse associados à sustentabili-dade e/ou meio ambiente. Mesmo assim, os resultados apontam para a dificuldade de explicar o que são mudanças climáticas em razão de sua complexidade e ressaltam o distanciamento do tema com a realidade cotidiana dos sujeitos. O fato de as mudanças climáticas serem um fenômeno global, com múltiplos e diversos impactos, faz com que a explicação do fenômeno represente um desafio por si só.Cavalcanti (2019) ainda sublinha que a concorrência com outros te-mas, percebidos ou sentidos como mais urgentes, faz com que a crise climática não entre na “cota finita de preocupações” que os indiví-duos têm – a percepção do risco climático tende a se tornar muito pequena ou até mesmo inexistente quando comparada ao risco do desemprego, da fome, da doença ou da violência, por exemplo. Essa outra questão remete ao contexto dos públicos com os quais quere-mos falar, revelando que é preciso conhecer a situação da recepção.A forma como a mudança climática é representada nas comunica-ções também influencia na maneira como os sujeitos reagem a ela. Nesse sentido, estudos afirmam que o viés catastrofista pode de-sencadear reações que, ao invés de mobilizar a população, podem gerar apatia ou paralisia. Embora chamadas e imagens negativas possam despertar uma atenção sobre a questão, o medo decorren-te dos piores cenários pode se mostrar bastante contraproducente para envolver as pessoas (O’NEILL; NICHOLSON-COLE, 2009).O negacionismo climático pode ser decorrente da falta de informa-ções corretas ou do acesso contínuo à desinformação, de uma mes-cla entre tais aspectos e aqueles associados à tentativa de sobrevi-ver sem medo do futuro (reforçado pelas mensagens apocalípticas) e da concorrência existente entre os riscos que serão enfrentados a cada dia. Interesses outros, como oportunidades econômicas, também podem ser consideradas.Todos esses desafios confirmam que tratar da comunicação climática não é tarefa simples e que, mesmo quando feita com bons resultados, não significa, automaticamente, que os públicos transformarão aquela mensagem processada em atitude capaz de auxiliar no enfrentamento do problema. E os limites entre preocu-pação, conhecimento e engajamento (seja ele individual ou político) nem sempre são nítidos. Bayes, Bolsen e Druckman (2020) afirmam que, apesar das mensagens sobre consenso científico fortalecerem a opinião pública sobre o clima, outros aspectos devem ser conside-rados, como o que motiva as pessoas quando elas formam crenças sobre as mudanças climáticas, por exemplo.Formas mais efetivas de comunicar as mudanças climáticas ten-dem a apontar para uma linguagem clara e simples, a aproximação do tema com a realidade próxima do leitor e a “humanização” do debate, a credibilidade ou confiabilidade associadas ao comuni-cador ou meio de comunicação, a necessidade de se associarem temas outros (evidenciando as diferentes implicações do clima para setores e aspectos sociais), as soluções, o trabalho em rede, o conhecimento prévio dos públicos e as suas demandas (o que de-veria fazer parte do planejamento de qualquer estratégia de comu-nicação) (LOOSE, 2021; CAVALCANTI, 2019).Gunster (2017) afirma que o jornalismo pode contribuir ainda mais com o engajamento a partir de quatro pontos: 1) priorizando públicos propensos a se envolver com o tema (em oposição à des-tinação de mensagens genéricas); 2) ampliando o enquadramen-to da justiça social, de modo que as dimensões éticas, políticas e normativas das mudanças climáticas recebam mais atenção; 3) abordando valores culturais que poderiam promover atitu-des, crenças e comportamentos; e 4) focando em experiências e emoções de pessoas e comunidades que estão colaborando para enfrentar as mudanças climáticas.Na revisão de estudos da área feita por Moser (2016), destaca-se a lacuna existente em estudos direcionados para efeitos de médio e longo prazo, já que a grande maioria dos estudos de recepção e con-sumo conformam-se em estudos sem continuidade. Quando falamos sobre engajamento precisamos lidar com fatores que vão além de uma exposição pontual. Outro ponto fundamental constatado pela autora é a distante conexão entre os pesquisado-res de comunicação do clima e aqueles que se dedicam à sua produção. Se por um lado não conseguimos ainda pensar em um desenho de pesquisa longitudi-nal, que depende de recursos no longo prazo, nossos esforços nesta pesquisa tentam aproximar prática e ciência para aprimorar a comunicação climática.... nossos esforços nesta pesquisa tentam aproximar prática e ciência para aprimorar a comunicação climática.PONTO DE PARTIDA24JORNALISMO E ENGAJAMENTO CLIMÁTICOPESQUISA 202225Para contextualizar a discussão de engajamento no jornalismo é preciso entender a estrutura com a qual ela busca romper, baseada no ideal da objetividade. Conforme Genro Filho (1987), a objetividade é herança das correntes funcionalistas estadunidenses voltadas aos estudos da comunicação social entre o final do século XIX e o início do século XX. A partir desta perspectiva, buscava-se aplicar a lógica empiricista, com base em modelos matemáticos, à produção da notícia. Desta maneira, fazer jornalismo significaria seguir uma fórmula no processo de seleção, hierarquização e relato dos acontecimentos.Traquina (2004), por sua vez, atenta para a importância de não enxergar a objetividade jornalística a partir de uma dicotomia com a subjetividade. Isto porque a prática jornalística, segundo o autor, é um híbrido entre o comprometimento com o interesse público e com os valores democráticos, e o entendimento do jornalismo enquan-to negócio, sujeito a pressões de mercado e de seus financiadores. Desta forma, o ideal da objetividade não consistiria em um problema quando significa buscar a informação o mais apurada possível para corresponder à responsabilidade com o interesse público – como um ritual da rotina do repórter (TUCHMAN,1993). O ponto de atenção, contudo, reside na associação da objetividade com valores inalcançá-veis pelo jornalismo enquanto construção social da realidade, como imparcialidade e neutralidade.É neste sentido que abordagens críticas emergem. No caso do Brasil, Figaro et al. (2021) descrevem o fortalecimento de arranjos alternati-vos e independentes na prática jornalística a partir do fim da primeira década do século XXI. São experiências alternativas à mídia hegemôni-ca, construídas a partir de relações horizontais, com “engajamento aos temas democráticos e populares” (FIGARO ET AL., 2021, P.436). Apesar das limitações estruturais e financeiras, tais experiências estão comprome-tidas com a democratização dos meios de comunicação, segundo os autores, dando espaço para perspectivas pouco exploradas pela grande mídia, especialmente de grupos sub-representados no debate público, como mulheres periféricas, população negra e povos indígenas.JORNALISMO E ENGAJAMENTOSob esse olhar, várias vertentes de práticas jornalísticas podem ser identificadas. Pepermans e Maeseele (2017) assinalam que for-mas jornalísticas que questionam o status quo, direcionadas para uma produção comprometida e reflexiva podem ser encontradas no advocacy journalism, constructive journalism e slow journalism, perspectivas que contestam a suposta neutralidade do jornalismo e que estão associadas à modalidade alternativa. Também Oliveira (2017, p.209), ao tratar do jornalismo emancipatório, a partir das ideias de Paulo Freire, afirma que jornalistas precisam tomar o parti-do dos segmentos sociais oprimidos para “[...] procurar entender os fenômenos sociais dentro da perspectiva de que há essa relação de opressão e que é necessário superá-la”. É nesse mesmo sentido que atua também o jornalismo ambiental, como se verá mais adiante.Com o questionamento às ideias de neutralidade no campo jornalístico, toma forma o entendimento do jornalismo enquanto motivador de mudanças sociais. Neste aspecto, Appelgren e Jöns-son (2021) mencionam o engajamento, que significa se importar e estar disposto e apto a agir diante de uma questão de interesse público. Assim, o engajamento leva ao interesse na participação política com vistas a colaborar com possíveis soluções para as problemáticas identificadas.Moraes (2019; 2021) defende que o ativismo é um elemento per-tinente para o jornalismo, especialmente no atual contexto político brasileiro, em que o governo de ultradireita acentua discursos de ódio e intolerância contra pessoas e grupos. O jornalismo de sub-jetividade proposto pela autora não nega a objetividade enquanto procedimento que envolve observação, apuração e pesquisa, por exemplo, os quais garantem a acurácia da informação; a proposta é dar atenção às posições de classe, de gênero, geográficas, raciais e questionar as estruturas sociais. O jornalismo das grandes redes e conglomerados seria considerado a norma e, portanto, um jorna-lismo “universal”. Todas as práticas desviantes, neste sentido, são classificadas como ideológicas, e o ativismo ou engajamento são associados de forma negativa a uma cobertura parcial. Conforme os argumentos da autora, entende-se que uma cobertura baseada em estereótipos e representações exotificantes do outro, que descon-sidera vieses de raça, gênero, classe social e orientação sexual, por exemplo, não é ela própria isenta ou imparcial.A partir daqui, aprofundamos mais o potencial de engajamento do jornalismo diante de questões ambientais e da crise climática. Ga-nha destaque o papel do jornalismo ambiental, perspectiva histori-camente relacionada à abordagem comprometida com a sustenta-PONTO DE PARTIDA26JORNALISMO E ENGAJAMENTO CLIMÁTICOPESQUISA 202227bilidade, de acordo com Bueno (2007), Frome (2008) e Girardi et al. (2012), e catalisadora do pensamento crítico. Este difere do jornalis-mo sobre meio ambiente, no qual a questão ambiental aparece ape-nas como temática, e não como transversal à abordagem, conforme destacam Loose e Girardi (2021). O jornalismo ambiental, portan-to, prevê a conexão das problemáticas ambientais com elemen-tos de política, economia, cultura e outras áreas do conhecimen-to, contemplando uma diversidade de saberes e olhares.Historicamente, a postura engajada inerente ao jornalismo ambien-tal encontra resistência nas redações brasileiras, principalmente nos grandes veículos, com proprietários identificados com a ideologia pró-modernização. Este é um obstáculo relevante, pois “o jornalista ambiental trabalha comprometido com a promoção da qualidade de vida planetária” (BELMONTE, 2017, P. 119). Outro desafio importante da prática do jornalismo ambiental reside em não apenas mencionar a necessidade de preservação ambiental, mas, principalmente, como a população pode contribuir para tal compromisso, explorando as possíveis soluções coletivas, o que aproxima o jornalismo do inte-resse público. Quanto à prática atual do jornalismo ambiental, “o seu engajamento, que pretende mobilizar a sociedade para os desafios ambientais do século XXI, pode e deve ser potencializado com as técnicas contemporâneas de apuração, como as do jornalismo guia-do por dados” (BELMONTE, 2017, P. 122).A fim de entender melhor a prática, o Grupo de Pesquisa em Jornalis-mo Ambiental UFRGS/CNPq mapeou o entendimento de jornalismo ambiental assumido pelos próprios jornalistas identificados com o campo na América Latina, Caribe, Portugal, Espanha e países africanos de língua portuguesa. Girardi, Loose e Silva (2018, p.58) sintetizam os achados de questionário aplicado com 60 profissionais, sublinhando que 55% dos respondentes afirmaram que o jornalismo ambiental “[...] tem papel educativo e transformador, contribuindo para a mobili-zação social, para mudanças de atitudes e pensamentos". Os jornalistas percebem o jornalismo ambiental como uma prática especializada, que tem o engajamento como princípio, e não é perme-ada por imperativos éticos de isenção ou imparcialidade. No entanto, alguns profissionais têm ressalvas com este termo, na medida em que pode dar a entender de que se trata de um jornalismo mais alinhado à causa do que ao compromisso de atender o interesse público, ressal-tam as autoras. Alguns afirmaram ser necessário estar engajado com a prática jornalística, de forma ampla, sendo comprometido e respon-sável, independentemente da pauta; há uma tentativa de evitar dizer que se é engajado (GIRARDI, LOOSE, SILVA, 2018).Alguns estudos se debruçam sobre as possibilidades de aborda-gem do engajamento ambiental no jornalismo profissional. Miguel (2019) analisa os cruzamentos entre o trabalho de organizações ambientalistas em prol da educação ambiental e a adaptação das normativas jornalísticas à lógica das mídias sociais digitais. A cultura participativa e a inteligência coletiva viabilizam o consumo de narrativas transmídias que promovem experiência sensorial e imersiva, e, portanto, próxima aos públicos a quem se destina. Tal estratégia fortalece um jornalismo engajado e militante, sensível às problemáticas sociais, por permitir a exploração de diversos forma-tos e maneiras de apresentar o conteúdo.O jornalismo ambiental cumpre três funções específicas – a infor-mativa, a pedagógica e a política (BUENO, 2007), pois está comprome-tido com a sustentabilidade da vida e com a emancipação social. Desta forma, os pressupostos do jornalismo ambiental alinham-se à crítica à colonialidade, ao capitalismo e a seus instrumentos de dominação e de manutenção do status quo (LOOSE E GIRARDI, 2021). O jornalismo ambiental considera outras concepções de mundo. Dessa maneira, pode e deve ter papel ativo na emergência das perspectivas do Sul Global. VISUALIZAR AS SOLUÇÕES ↑Turbinas eólicas em Fortaleza (CE), Brasil. Foto: ©Greenpeace / Flavio CannalongaPONTO DE PARTIDA28JORNALISMO E ENGAJAMENTO CLIMÁTICOPESQUISA 202229Frente ao desafio imposto pelas mudanças climáticas, novas es-tratégias são assumidas pelos veículos jornalísticos para tornar a questão mais palpável e próxima do cotidiano. O jornalismo climáti-co, ou climate journalism, sustenta que as fontes sobre tal questão deixam de ser apenas autoridades ou cientistas, tendo em vista o crescimento de organizações não-governamentais que trabalham com o tema, por exemplo (SCHÄFER; PAINTER, 2021). Com as mudanças velozes nos ecossistemas midiáticos e a emergência de veículos especializados no tema, esta é uma área de estudos que demanda cada vez mais atenção dos pesquisadores.Alguns direcionamentos dessa cobertura especializada precisam ser definidos. Por exemplo, como o jornalismo garante o conhecimento e o envolvimento dos cidadãos sem que a incerteza sobre os cenários da emergência climática gere medo paralisante nos atores sociais? Segundo Balbé e Loose (2020), com base em Lorenzoni, Nicholson--Cole e Whitmarsh (2013), o enfrentamento diante das mudanças climáticas envolve três níveis que se articulam entre si: a emoção, o conhecimento e o comportamento. Tais aspectos interagem de diferentes formas de acordo com o contexto e as vivências de cada sujeito e são potencializados quando articulados de forma coleti-va. Portanto, não há como padronizar respostas sobre a mobiliza-ção dos atores diante das mudanças climáticas. E, embora o medo seja um elemento presente nas abordagens jornalísticas, inerente aos critérios de noticiabilidade como intensidade, raridade e dra-maticidade, não há consenso sobre como lidar com essa emoção na comunicação dos riscos climáticos.O jornalismo pode, contudo, criar capacidade de agir nos cidadãos, provocando o questionamento das estruturas de poder e estimu-lando escolhas políticas alinhadas com a sustentabilidade. Con-forme Balbé e Loose (2020), a percepção do risco relacionado ao volume das emissões GEE deve ser acompanhada da percepção de que também existem alternativas e experiências já em curso que resgatam a conexão com a natureza. Assim, a cobertura deve trazer informações de fácil compreensão e ser propositiva, pois somente alarmar os indivíduos pode resultar em passividade diante do inevitável. Segundo as autoras, “o enfrentamento das ACs [alterações climáticas] requer estratégias diversas, asso-ciadas a cada perfil de público, de modo a evitar generalizações que acarretem mais apatia e desesperança do que respostas ao problema” (BALBÉ; LOOSE, 2020, P.50).Ao estudarem o engajamento sobre mudanças climáticas a partir de experiências de jornalismo de dados na Suécia, Appelgren e Jönsson (2021) destacam que, em termos de enquadramentos selecionados, o jornalismo guiado por dados não difere do jorna-lismo tradicional que traz pautas sobre meio ambiente de modo fortuito e pouco contextualizado. Assim, a interação possibilitada por tecnologias digitais não necessariamente leva o público à ação. Para isso acontecer, pode ser necessário trabalhar de forma mais direcionada com audiências segmentadas. Também é im-portante rever o posicionamento do jornalismo de dados como uma prática “o mais objetiva possível”, o que pode prejudicar a conexão emocional com a pauta e, por consequência, o engaja-mento dos cidadãos.Bolsen e Shapiro (2017) destacam algumas estratégias para mobilizar o público: selecionar enquadramentos que envolvam emocionalmente os indivíduos, priorizando mensagens positivas e de esperança, de modo que estes se sintam mobilizados a parti-cipar da solução dos problemas. Os autores afirmam que reforçar a existência de consenso científico sobre contribuição das ações humanas para as mudanças climáticas pode aumentar a percepção das pessoas sobre a urgência da questão, embora a estratégia seja menos eficaz em contextos de grande influência política, como em grupos de republicanos. Por fim, acentuar os impactos das mudan-ças climáticas em nível local pode favorecer para que as pessoas daquela região fiquem propensas a buscar mais informações.A partir do construtivismo social, Laws e Chojnicka (2020) alegam que a separação de caráter positivista, firmada a partir do sécu-lo XIX, entre jornalismo “reportador” – baseado na confiança na objetividade – e jornalismo orientado para “advocacy” – que toma partido e defende uma causa – não se sustenta mais. Os autores partem do pressuposto de que é impossível haver uma verdade universal e que, portanto, a neutralidade é um valor que pode ser almejado, mas na prática não pode ser alcançado pelo jornalismo. Enquanto o primeiro tenta disfarçar seu posicionamento através da linguagem, o segundo é honesto sobre as perspectivas que guiam seu trabalho. Ao reconhecer a importância dos contextos históri-cos, sociais e políticos na construção do conhecimento, o estudo propõe, então, outra diferenciação: entre o jornalismo partidário de mudanças sociais – “progressive (or regressive) advocacy journalism” – e o jornalismo partidário do status quo – “status quo advocacy journalism” (LAWS; CHOJNICKA, 2020, P.2).Laws e Chojnicka (2020) assinalam que o jornalismo partidário das mudanças sociais que trabalha na perspectiva da sustentabi-lidade deveria ser baseado em uma abordagem integrada, ou seja, PONTO DE PARTIDA30JORNALISMO E ENGAJAMENTO CLIMÁTICOPESQUISA 202231que leve em consideração os pilares da ecologia, da economia e da sociedade. Os autores aplicam tal pressuposto ao estudo da cobertura jornalística sobre a campanha de Bernie Sanders como pré-candidato do Partido Democrata à presidência dos Estados Unidos em 2016. A partir do conhecido posicionamento do sena-dor em prol de projetos de combate às mudanças climáticas e à degradação ambiental, o fato de sua campanha ter obtido menos cobertura em relação aos pleitos dos também candidatos Hillary Clinton e Donald Trump teria relação, segundo os autores, com o desconforto da mídia corporativa em dar espaço às críticas de San-ders ao sistema econômico do qual tais conglomerados são atores proeminentes. Tal viés pode ter, inclusive, influenciado a decisão do Partido Democrata em nomear Hillary Clinton como sua repre-sentante no processo eleitoral. A pesquisa revela, portanto, que o jornalismo partidário das mudanças sociais ainda está restrito, sobretudo, a meios alternativos e eletrônicos, o que limita seu alcance e impacto social; e que discursos de mídia aparente-mente progressistas, como os que destacaram os posicionamen-tos de Clinton como progressistas em comparação aos de Trump – embora Sanders, com sua ampla atuação parlamentar diante das questões sociais e ambientais, estivesse no páreo –, podem não ser totalmente transformadores se não desafiam a ideologia neoliberal vigente.A partir da revisão de literatura sobre jornalismo e engajamen-to, é possível realizar alguns apontamentos. De acordo com os autores consultados, entendemos que o engajamento passa pela crítica à noção de objetividade jornalística e pela aceitação do ativismo em prol de uma causa como fundamento para uma prática jornalística honesta, responsável e comprometida com o interesse público. Esta abordagem jornalística identificada com uma causa, frequentemente referenciada na literatura através do termo advocacy, não se refere à defesa de um interesse pessoal do jornalista ou corporativo do veículo em que atua. Trata-se de priorizar o interesse da coletividade, em prol da sustentabilidade da vida no planeta.Sendo assim, o jornalismo ambiental é uma das perspectivas emergentes que podem colaborar com o engajamento climático na medida em que seus pressupostos reforçam a crítica ao capi-talismo, à colonialidade e à lógica do Norte Global. Ao se assu-mir como uma perspectiva e uma prática engajada, o jornalismo ambiental compromete-se com a sensibilização dos públicos para formas mais efetivas de enfrentamento à crise climática. VISUALIZAR AS CONSEQUÊNCIAS DAS MUDANÇAS CLIMÁTICAS ↑ Em Bailique, erosão, falta de luz e salinização do Amazonas se tornaram desafios permanentes. Foto: Ianca Moreira PONTO DE PARTIDA32JORNALISMO E ENGAJAMENTO CLIMÁTICOPESQUISA 202233A fim de analisar como os participantes de movimentos sociais se relacionam com o jornalismo e com a pauta das mudanças climáti-cas, torna-se necessária uma aproximação com suas experiências de consumo midiático e sentidos produzidos a partir das mensa-gens comunicacionais, uma dimensão que é bem contemplada pelos estudos de recepção e de consumo midiático.Toaldo e Jacks (2013) afirmam que os estudos de consumo midiáti-co têm como centro de sua investigação a relação dos consumidores com os meios de comunicação e produtos midiáticos, em termos de experiência e contexto de consumo, assim como de apropriações, compreendendo o que, como e por que consomem, enquanto os es-tudos de recepção têm como foco a produção de sentidos a respeito das mensagens. Na prática, esses estudos tendem a se entrecruzar, pois são complementares. A investigação de um auxilia na compre-ensão do outro e vice-versa. Assim, a maioria dos estudos tende a circular por ambas as esferas e essa separação tem como objetivo principal auxiliar os pesquisadores a compreender qual o foco de estudo (TOALDO; JACKS, 2013). Escosteguy e Jacks (2005) relatam que há muitos autores que incluem todos esses trabalhos no mesmo âmbi-to, considerando estudos de recepção todos aqueles que analisam o polo oposto da emissão, os receptores das mensagens e as relações com os meios.Ao realizar pesquisa bibliográfica sobre esse tema, verificou-se, como esperado, que a maioria dos resul-tados investigavam as mensagens de produtos midiáticos a partir de análises de conteúdo e de discurso ou a produção jornalística a partir de entrevistas com profissionais da área, o que demonstra a escassez das pes-quisas de recepção e justificam sua importância.CONSUMO MIDIÁTICO E RECEPÇÃO DE PRODUTOS JORNALÍSTICOSSENTIDOS CIENTÍFICOS DISTANTES E, POR VEZES, NEGACIONISTASOs artigos de Taks (2019) e Casanova-Pérez, Dávila e Alonso (2019) apresentam trabalhos que analisam a recepção de mensagens so-bre as mudanças climáticas e sua tradução em relação ao cotidiano desses públicos específicos. A maior contribuição dos trabalhos re-side na aproximação por meio de métodos de pesquisa qualitativos (em geral entrevistas) com públicos específicos, análise e compara-ção com os textos midiáticos, e apontamentos de possíveis ade-quações ou lacunas da linguagem/abordagem midiática.Taks (2019) realizou sua pesquisa no Uruguai a partir da análise de notícias, mesas de discussão com especialistas e de uma observa-ção participante de seis a oito meses em três locais: um laboratório de climatologistas; uma zona de pesca artesanal e um território marítimo dos capitães e trabalhadores de uma boia de petróleo. A análise de notícias feita por Taks (2019) demonstra que o governo e a mídia forneceram um enquadramento institucional, tecnocien-tífico e catastrófico das mudanças climáticas, ao aproveitar uma oportunidade de inserção do tema a partir de eventos climáti-cos extremos, como tornados e inundações, e COPs. No entanto, essas abordagens geram um problema de tradução (LATOUR,2012), entre conhecimento científico, político e o âmbito cotidiano dos receptores. A fala de um climatologista ilustra um dos problemas da tradução. Ele questiona o motivo de associar as mudanças climáticas às catástrofes, visto que se essas não ocorrem ou se se torna possível achar outra razão para sua ocorrência, a crença nas mudanças climáticas pode se desfazer. Como exemplo, Taks (2019) observou que petroleiros e pescadores percebem mudan-ças nos níveis e condições de estabilidade das águas que afetam consideravelmente o seu cotidiano, no entanto, muitos acabaram abraçando o discurso de que essas mudanças se dão em virtude de ciclos climáticos mais curtos, de 30 a 40 anos, que eles dizem já conhecer, um argumento que condiz com uma fala negacionista de um ex-ministro da agricultura. Por outro lado, esses profissionais também acabam achando soluções ou melhorias tecnológicas para os problemas causados pelas más condições das águas, o que os afasta da urgência de ação em relação às MCs.Já a pesquisa de Casanova-Pérez, Dávila e Alonso (2019) utiliza-se de dados secundários, de revisão documental de matérias jornalís-ticas (produto midiático), de pesquisas quantitativas e entrevistas em profundidade com agricultores da região subúmida do México para demonstrar a importância da adaptação da linguagem cientí-fica à linguagem simples, específica e contexto local dos agriculto-... a maioria dos resultados investigavam as mensagens de produtos midiáticos a partir de análises de conteúdo e de discurso ou a produção jornalística a partir de entrevistas com profissionais da área ...PONTO DE PARTIDA34JORNALISMO E ENGAJAMENTO CLIMÁTICOPESQUISA 202235res. Como exemplo, os pesquisadores apontam diferenças entre a semântica científica distante como no uso do termo chuvas torren-ciais, em espanhol “lluvias torrenciales”, e os termos usados em espanhol pelos agricultores, “aguacerones” ou “torrentales”.Ambos os trabalhos demonstram a necessidade de mensagens midiáticas que falem a públicos específicos com linguagens e abordagens próximas sobre os efeitos gerais das mudanças cli-máticas às suas condições atuais e futuras de vida e produção, fornecendo informações sobre medidas de adaptação e ação política, sem reter o foco à explicação tecnocientífica da crise climática ou em catástrofes.RELAÇÕES ENTRE CONSUMO MIDIÁTICO, CRENÇA NAS MUDANÇAS CLIMÁTICAS E APOIO A POLÍTICAS PÚBLICAS O estudo de Nathania, Zunaidah, Yanuarita e Reinaldo (2021), conduzido durante a pandemia com estudantes de um campus “verde” na Indonésia, realizou um questionário aberto com 26 estudantes millenials (nascidos em 1982-2005) e entrevistas por Zoom com cinco desses estudantes. Os resultados indicam para um consumo de notícias sobre meio ambiente que não ocorre por meio de uma busca ativa, mas no momento em que elas apa-recem em seu feed de notícias, principalmente no Instagram, a partir de perfis sobre o tema que os estudantes seguem. Alguns estudantes destacaram que o conteúdo audiovisual é essencial para eles, preferindo esse formato a longos textos. Eles alegam que esse consumo os inspira a aderir a hábitos mais ecológicos e, em alguns casos, a conversar e sugerir ações a amigos e familia-res. O fato de fazerem parte do campus “verde” foi reconhecido pelos estudantes, apesar do campus não explorar muito a comu-nicação e a promoção de ações “verdes” em todo seu potencial (NATHANIA ET AL., 2021).Já o estudo de Hu, Zeng e Yi (2019) é feito a partir de aplicação de questionário fechado com amostra significativa da população chinesa. O trabalho investigou a relação entre o consumo midiático e a satisfação com a ação governamental em relação ao meio am-biente. Os resultados mostram que, na China, o consumo de mídias tradicionais está diretamente associado com a satisfação em rela-ção às políticas públicas ambientais, enquanto o consumo midiá-tico digital está associado à insatisfação com as políticas públicas chinesas de meio ambiente. Os pesquisadores sinalizam para a influência do governo nas mídias tradicionais, o que incentiva a sa-tisfação pelos consumidores desse meio, e o acesso a informações mais variadas e aprofundadas na internet, o que poderia embasar a insatisfação dos internautas.Em estudo do tipo survey realizado com amostra significativa da população da Índia, país "em desenvolvimento" e com 25% da população vivendo abaixo da linha da pobreza, Thaker, Zhao e Leiserowitz (2017) também encontraram relações entre o con-sumo midiático e o apoio a políticas públicas, porém com foco no aquecimento global (o artigo parece ter preferência por este termo ao invés de MCs). Os resultados revelam uma tendência geral da crença no aquecimento global ser positivamente asso-ciada a um aumento do consumo de mídias, à percepção e ao apoio a políticas públicas, porém com especificidades para cada tipo de consumo midiático. A televisão foi o meio comunicacio-nal mais associado à crença no aquecimento global e ao apoio a políticas públicas, porém sem estar ligada ao ganho de conhe-cimentos científicos sobre o tema por parte dos consumidores. Já o consumo de jornais não estava necessariamente associado à crença, à percepção de risco e ao apoio a políticas públicas e, também, o consumo de internet (que, segundo os autores, é restrito a um público muito pequeno e não implica necessaria-mente em consumo de informações sobre o aquecimento global), estava negativamente associado à percepção de risco e ao apoio a políticas públicas.Retomando os estudos, observamos que Nathania et al. (2021) investigam um público jovem que já faz parte de um contexto de política pública relacionada às mudanças climáticas, um campus “verde”, que demonstra sua vontade de consumir notícias sobre o assunto e sua recepção positiva das mensagens quando essas aparecem em seu feed virtual. O estudo de Hu, Zeng e Yi (2019), realizado na China, aponta para uma ligação entre consumo midi-ático de mídias tradicionais e o apoio a políticas públicas, decor-rente do contexto governamental específico chinês. E o trabalho de Thaker, Zhao e Leiserowitz (2017), realizado na Índia, constata, assim como outros estudos internacionais, que um aumento no consumo midiático tende a contribuir com a conscientização am-biental, ou seja, esses dados podem ainda melhorar consideravel-mente a partir da continuidade do trabalho jornalístico e aumento de acesso à informação em um país em que o consumo midiático ainda é muito restrito.PONTO DE PARTIDA36JORNALISMO E ENGAJAMENTO CLIMÁTICOPESQUISA 202237fico e sobre a falta de ações em relação às mudanças climáticas, o escasso planejamento urbano e o papel da defesa civil. A pesquisa demonstra a possibilidade de novas ferramentas de comunicação para a expressão de receptores, os quais respondiam a tweets de jornalistas e autoridades.Estas duas pesquisas destacam o papel ativo do receptor, sendo que no primeiro caso a colaboração ocorre pela sugestão de perguntas e possível presença nos programas. Já no segundo trabalho, os tweets são tanto o relato da produção de sentidos do receptor em relação à sua apropriação de notícias sobre as chuvas do RJ quanto sua pró-pria produção de conteúdo como emissores na rede Twitter.Os resultados indicam um alto nível de confiança nas informações científicas apresentadas nos programas e os receptores demonstram ter aprendido informações específicas sobre o seu contexto em relação às mudanças climáticas, o que representa alguns impactos positivos da experiência em termos de educação ambiental e jornalismo colaborativo, apesar da participação dos receptores ter sido limitada.Os artigos de Nettlefold e Pecl (2022) e Andrade, Barreto e Hen-riques (2020) são caracterizados pelo papel ativo dos receptores. Em estudo realizado na Austrália, no estado insular da Tasmânia conhecido por ser inclinado a políticas e conscientização “verde”, Nettlefold e Pecl (2022) relatam a ação e pesquisa sobre o projeto Curious Climate da ABC, rede de televisão australiana, que convi-dou sua audiência a enviar perguntas para que cientistas e jornalis-tas do clima respondessem durante programas com a participação da audiência. Após os programas, os participantes responderam a um questionário. Os resultados indicam um alto nível de confiança nas infor-mações científicas apresentadas nos programas e os receptores demons-tram ter aprendido informações específicas sobre o seu contexto em relação às mudanças climáticas, o que representa alguns impactos positivos da experiência em termos de educa-ção ambiental e jornalismo colaborati-vo, apesar da participação dos recep-tores ter sido limitada.Já o estudo de Andrade, Barreto e Henriques (2020) analisou 53 tweets que remetiam às mudanças climáti-cas, dentre os 357 mil tweets sobre o evento climático das chuvas fortes que atingiram o Rio de Janeiro entre 07 e 10 de abril de 2019. A metodolo-gia empregada foi a análise temática, que permite categorizar padrões exis-tentes e selecionar aqueles que dialogam com o tema de pesquisa (BRAUN; CLARK; 2006). Os tweets mapeados foram analisados em três te-mas - bairros, influenciadores e mudanças climáticas - e os tweets sobre mudanças climáticas foram organizados em seis categorias: negacionismo climático, adaptação, governança, vulnerabilidade, críticas ao ceticismo climático, educação e informação. A maioria dos tweets eram referentes mais aos temas governança e críticas ao ceticismo climático, o que exemplifica a utilização da rede para cobrar as autoridades sobre medidas contra o negacionismo cientí-RECEPTORES ATIVOS E PARTICIPATIVOSPilhas de eucalipto no Macapá. Foto: © Daniel Beltrá / GreenpeaceVISUALIZAR AS CAUSAS E OS PRINCIPAIS RESPONSÁVEIS ↓" O jornalismo precisa ter consciência do seu poder de engajar, gerando identificação com o problema e responsabilizando os atores. É necessário mostrar a conexão com a realidade local, a partir das evidências científicas, e a pluralidade de vozes, contemplando os temas indígenas e incluindo as soluções possíveis. (GF CENTRO-OESTE, 02.07.2022)PESQUISA 202239METODOLOGIAA fim de debater os achados do estudo com a bibliografia mais recente sobre jornalismo e engajamento climático, realizou-se, na primeira etapa, uma pesquisa sistemática no SciELO e no Google Scholar no período dos últimos cinco anos (2017-2021), privile-giando os artigos revisados por pares. Detivemo-nos em três eixos teóricos: a comunicação climática (e suas aproximações com o con-texto brasileiro); as conexões entre jornalismo e engajamento; e os estudos de recepção e consumo midiático orientados para o tema.Vale ressaltar ainda que as buscas no Google Scholar se mostra-ram mais profícuas. As palavras-chave acionadas, em português e inglês, variaram de acordo com o tema central de cada eixo. Uma série de combinações foi realizada na pesquisa bibliográfica, am-pliando a questão climática para a ambiental e os aspectos jorna-lísticos para os comunicacionais, de modo a rastrear a literatura mais próxima aos interesses desta pesquisa. Os resultados desse esforço podem ser observados ao longo de todo o trabalho, sobre-tudo na seção anterior.METODOLOGIA40JORNALISMO E ENGAJAMENTO CLIMÁTICOPESQUISA 202241A técnica para recolha de dados adotada, o grupo focal, possibili-ta uma construção plural, diversa e aprofundada de conhecimen-tos. O grupo focal promove um encontro, entre pessoas inte-ressadas em um tema comum. Os diálogos facilitados por uma pessoa na função de moderador, orientado por um roteiro com perguntas-chave, possibilita a irrupção de informações e dados, cuja análise posterior vai trazer resultados como novos saberes, soluções e questões sobre o tema. Em síntese, pesquisadores destacam que, não raro, ouvem de participantes que, naquela ocasião, tiveram a primeira oportunidade de parar para refle-tir a respeito e com acompanhamento qualificado. Assinalam, também, a satisfação manifestada pelos participantes devido ao enriquecimento através dos diálogos ou por contribuírem com a produção de conhecimento.A técnica do grupo focal pode ser usada de diferentes formas: como espaço de entrevistas, debates e diálogos. A finalidade vai da busca por consensos e diferenças, ao aprofundamento de temas fortes e complexos, até o esclarecimento de dúvidas ou obtenção de avalia-ções (NASCIMENTO; ABREU; FONSECA, 2020; SOUZA, 2020; GATTI, 2005).Neste estudo, os grupos focais foram realizados de forma virtual, pela plataforma Zoom, sendo, ao menos, dois grupos focais por região do país, a fim de contemplar a diversidade regional brasi-leira, de extensão continental. Para realizar os convites, além de demarcarmos um perfil, fizemos uso de um questionário online para identificação dos jovens ativistas interessados em contribuir com um debate sobre jornalismo e engajamento climático.Segundo os dados obtidos pelo questionário, a maioria dos respon-dentes são da região Sudeste (91; 42,1%), depois Sul (59; 27,3%), Centro-Oeste (27; 12,5%), Nordeste (24; 11,1%) e Norte (15; 6,9%). Mesmo com os esforços específicos de circulação da pesqui-sa nas regiões Norte e Nordeste, o quadro não se alterou. O gráfico abaixo apresenta a distribuição por região das respostas obtidas.Apesar do questionário proporcionar dados interessantes, percebeu--se que os contatos obtidos não estavam se mostrando suficientes para a composição dos grupos focais, respeitando a representativi-dade das cinco regiões brasileiras. Assim, foram realizados convites para integrar a segunda etapa da pesquisa (também chamados de grupos de discussão virtuais). Essa fase consistiu em busca ativa nas redes, organizações e movimentos que pudessem indicar ativistas jovens (com até 35 anos) e, depois, indicações daqueles que já haviam participado. Com isso, chegamos a um total de 60 participantes:COLETA DE DADOSORIGEM GEOGRÁFICA DOS RESPONDENTES DA PESQUISAGRÁFICO 1 Fonte: Os autores (2022).CENTRO-OESTE 12,5%SUL27,3%NORTE 6,9%NORDESTE 11,1%SUDESTE 42,1%METODOLOGIA42JORNALISMO E ENGAJAMENTO CLIMÁTICOPESQUISA 202243DISTRIBUIÇÃO DE PARTICIPANTES DOS GRUPOS FOCAIS POR REGIÃOQUADRO 1 Fonte: Os autores (2022).1211111313NÚMERO DE PARTICIPANTESSulSudesteCentro-OesteNordesteNorteREGIÃOÉ importante mencionar que, em todos os casos, foram convidados (e confirmados) mais ativistas do que os números apresentados (tivemos desistências, após a confirmação, em todos os grupos). Realizamos também um esforço de mobilização para que os convidados ficassem durante todo o período da discussão (que teve duração de duas ho-ras), mas em alguns casos tivemos entradas no meio da dinâmica ou saídas antecipadas. Tais situações foram esporádicas e não geraram, na perspectiva dos pesquisadores, prejuízos aos dados obtidos.Ressalta-se ainda que o convite sempre reforçou a faixa etária pesquisada e a necessidade de cada participante se sentir engaja-do com alguma causa ou se ver como ativista (no sentido amplo do termo, não precisando estar formalmente ligado às ONGs ou movi-mentos sociais, por exemplo). E, verificou-se em duas respostas do questionário que participantes dos grupos focais se colocaram como não engajados. Isso pode ter ocorrido em razão da interpretação pessoal do que é, de fato, ser engajado. Durante os grupos focais, percebe-se que havia um entendimento forte de que o ativismo de-veria ser constante e alguns apontavam que não se sentiam fazendo o bastante para se “enquadrar” nessa definição, como apresentado por um dos ativistas da Região Sul:A maioria dos respondentes do questionário, 161 (75,2%), se con-siderou engajado em alguma causa ou movimento social, enquanto 53 (24,8%) afirmaram que não eram. Esse é um ponto interessante, afinal, mesmo que a pesquisa tenha sido direcionada a ativistas, verificamos que quase um quarto dos que participaram da primeira fase responderam não se identificar com esse perfil, o que sugere que há um interesse pelo tema da pesquisa por acadêmicos e pro-fissionais que não entendem seu envolvimento como ativismo.Recorda-se que, nos grupos focais, antes da aplicação do roteiro, foi informado aos participantes o que estávamos entendendo por engajamento, já que no pré-teste, organizado com ativistas con-vidados pelo Modefica, o termo foi, muitas vezes, reduzido a uma ação nas redes sociais. Esclareceu-se que:PANORAMA DAS CIDADES DOS RESPONDENTES DA PESQUISAGRÁFICO 2Fonte: Os autores (2022).NorteNordesteCentro-OesteSudesteSul""Falo muito do meu pessoal: eu já fui muito mais engajada. Hoje estou parada, vejo notícia e não engajo muito, nada mais me abala muito ao ponto de fazer algo. Tenho certeza de que é por rotina [...]. Fico mal de fazer tão pouco, mas é por rotina, não tenho tempo livre pra pensar em outra coisa e também por ter o sentimento horrível de que está tão terrível, que não tem mais solução pra nada.A pesquisa trata de engajamento, no sentido de se importar e estar disposto e apto a agir diante de uma questão de interesse público (nesse caso a crise climática). Assim, o engajamento que buscamos promover é aquele que leva à participação política e públicatão terrível, que não tem mais solução pra nada.SÃO PAULOMANAUSCURITIBAFREDERICO WESTPHALENFLORIANÓPOLISBARBACENACAMPO GRANDERIBEIRÃO PRETOPORTO ALEGRESANTO ANDRÉSÃO LUÍSILHABELARIO DE JANEIROBARRA DO GARÇASFORTALEZASÃO JOSÉBRASÍLIATAGUATINGABELÉMMACAPÁSALVADORCAMPINASJUIZ DE FORASUZANOPALMAS VILA VELHAFOZ DO IGUAÇUMETODOLOGIA44JORNALISMO E ENGAJAMENTO CLIMÁTICOPESQUISA 202245Embora esta pesquisa, com os recursos previstos, não tivesse sido desenvolvida se não fosse virtualmente, destacamos as limitações derivadas da aplicação do método. Por conta do tempo em frente à tela e da própria exaustão das atividades online, reflexo da pandemia de covid-19, as discussões tiveram de ser limitadas, configurando, em alguns momentos, uma entrevista coletiva na qual os responden-tes concordavam ou discordavam das falas anteriores.Foram realizados dez grupos de discussão com jovens ativistas, dois em cada região brasileira, entre 22 de junho e 20 de julho de 2022. Registra-se ainda que antes dessa etapa foi realizado um grupo focal pré-teste, com jovens identificados pelo Instituto Modefica, que per-mitiu o aprimoramento e ajuste do roteiro e organização da dinâmica, que consistiu, na sua versão final, em uma breve apresentação da pesquisa, dos pesquisadores e dos participantes, divisão das salas para coleta de dados com duas partes:1) perguntas gerais sobre engajamento e jornalismo climático; 2) percepções sobre duas peças jornalísticas audiovisuais sobre clima (a primeira com foco em causas, que tratava do desmatamento da Amazônia, produzida pela TV Cultura; e a segunda que traz também algumas considerações sobre soluções, produzida pelo Brasil de Fato), e, ao final, uma atividade coletiva de construção de síntese sobre como deveria ser o jornalismo para engajar no enfrentamento às mudanças climáticas.Cada sessão contou com um pesquisador-moderador para conduzir a discussão e controlar o tempo, baseado em um roteiro comum, mais um relator, que registrava, em tempo real, o que estava sendo ma-nifestado em cada sala. Além disso, por segurança, todos os grupos foram gravados para recuperação posterior de dados (transcrição das falas). Os grupos focais foram realizados na condição de anonimato, por isso, faremos a citação dos participantes somente levando em conta sua região.Os dados coletados nos grupos focais, após transcritos, foram cate-gorizados a partir da revisão de literatura já feita e das observações dos pesquisadores durante as dinâmicas, considerando sempre os objetivos da pesquisa: compreender como se dá o engajamento pró-clima por meio do jornalismo, sobretudo para ampliar o enten-dimento sobre as relações entre desmatamento da Amazônia e crise climática. As análises foram separadas por região e depois discuti-das coletivamente, visando entender pontos comuns (para o cenário nacional) e específicos (por região). VISUALIZAR AS CAUSAS E OS PRINCIPAIS RESPONSÁVEIS ↑ Monitoramento de Desmatamento na Região da AMACRO, Rondônia. Foto: © Bruno Kelly / GreenpeaceMETODOLOGIA46JORNALISMO E ENGAJAMENTO CLIMÁTICOPESQUISA 202247ANÁLISE DE DADOSPara investigar as relações entre o consumo de jornalismo e o enga-jamento de ativistas com a pauta climática a partir dos dados quali-tativos, coletados nos grupos focais, adotamos a Análise de Conteú-do, que segundo Bardin (1979), consiste em um conjunto de técnicas que permitem a análise sistemática de conteúdo verbal e audiovisual das mensagens comunicacionais. Isso ocorre pois, apesar da natu-reza interpretativa das palavras, gestos e imagens, essa metodologia mapeia o conteúdo manifesto, como as aparições de palavras ou imagens, de forma que uma pesquisa possa ser replicada por diver-sos pesquisadores gerando resultados similares (BARDIN, 1979).Conforme Bardin (1979, p.37), a Análise de Conteúdo é “um con-junto de técnicas de análise das comunicações visando obter, por procedimentos sistemáticos e objetivos de descrição do conteúdo das mensagens, indicadores (quantitativos ou não) que permitam a inferência de conhecimentos relativos às condições de produção/recepção (variáveis inferidas) destas mensagens”.Além disso, através da apresentação de ambiguidades dos textos, embates e direcionamentos, o pesquisador que realiza a análise de conteúdo consegue captar e explorar outros aspectos, como as ideologias e o conteúdo implícito na mensagem, indo além da percepção inicial, superficial, de um texto. Para Bardin (1979), a função essencial desta análise é o desvendar crítico.Segundo Franco (2008) um dos procedimentos mais relevantes da aplicação da análise de conteúdo consiste no desenvolvimento de categorias que sustentam a classificação dos significados manifes-tos das mensagens, agrupando os dados segundo critérios semân-ticos (temas), sintáticos, léxicos ou expressivos. A criação das ca-tegorias é um processo desafiante e que, em geral, pode ocorrer de duas formas: definidas a priori, a partir do problema e objetivos de pesquisa, ou não a priori, emergindo da "fala" ou demais elemen-tos dos dados coletados, o que implica em repetidos cruzamentos entre material coletado e teoria.A presente pesquisa desenvolve sua categorização não a priori. A partir de uma análise inicial de temas emergentes nas falas dos participantes do grupo focal pré-teste, foram delineadas catego-rias, as quais também respondem aos objetivos de pesquisa. Foi mantida a possibilidade de inserção de novas categorias ou subca-tegorias a partir da análise dos dados coletados nos demais grupos focais – o que ocorreu durante o período de categorização.Para que não existam categorias amplas demais, que impeçam a observação de nuances de significados, e nem um número muito grande de categorias, o que poderia fragmentar a análise a ponto de se perder pontos de convergência, Franco (2008) indica a orga-nização das categorias em molares (mais amplas) com indicadores moleculares que reorganizam os dados. A autora também relata a importância de criar as categorias de acordo com três princípios: a exclusão mútua (as categorias não devem se sobrepor, cada uma abarca um tipo de conteúdo que exclui outro); a pertinência (a cate-goria deve ser relevante para a pesquisa); a objetividade e fidedig-nidade (erros de codificação acontecem quando as categorias não estão claras, por isso a importância de serem objetivas).Assim, foram elaboradas as seguintes categorias e subcategorias:VISUALIZAR AS CONSEQUÊNCIAS E AS PESSOAS MAIS IMPACTADAS ↑Seca no Amazonas. Foto: © Greenpeace / Daniel BeltráMETODOLOGIA48JORNALISMO E ENGAJAMENTO CLIMÁTICOPESQUISA 202249Fonte: Os autores (2022).ENGAJAMENTOJORNALISMO CLIMÁTICOENGAJAMENTO POR MEIO DO JORNALISMOENGAJAMENTO INDIVIDUALENQUADRAMENTO DO ASSUNTOLINGUAGEM JORNALÍSTICAPERSPECTIVAS DAS FONTESENGAJAMENTO COLETIVO/POLÍTICOReúne aspectos (constatações e críticas) sobre mobilização e incentivo à ação individual, refle-xões sobre "o que eu posso fazer". O que me move enquanto ativista? Quais os limites da ação individual?Abarca a maneira como o assunto é predominantemente enfatizado pela imprensa. ǿ Foco nas causas – concentra percepções sobre como as causas que levaram à intensifi-cação das mudanças climáticas são tratadas ou invisibilizadas (discussão sobre que tipo de desenvolvimento nos trouxe até aqui). De que maneira as causas das mudanças climáticas são apresentadas pelo jornalismo? ǿ Foco nas consequências ou efeitos – traz as percepções sobre como os efeitos climáticos (eventos climáticos extremos, aumento do nível do mar, mudança nos regimes de chuva, etc.) são representados pela imprensa. De que maneira os efeitos das mudanças climáticas são apresentados pelo jornalismo? ǿ Foco nas soluções – congrega as percepções sobre como o jornalismo apresenta (ou não) so-luções ou alternativas para enfrentar as mudanças climáticas (iniciativas ligadas à mitigaçãoe adaptação), e quais são as soluções propagadas. De que maneira as soluções para enfrentar as mudanças climáticas são apresentadas pelo jornalismo?Reúne comentários e observações a respeito de como os fatos são relatados nas matérias jorna-lísticas e como isso está relacionado ao engajamento ǿ Termos técnicos - trata das percepções de como o jornalismo lida com a ciência das mu-danças climáticas e de sua complexidade. Qual a linguagem das reportagens sobre mudanças climáticas? De que maneira a linguagem técnica interfere no engajamento? ǿ Escolha do público e especificidades dos formatos - reúne observações que consideram as especificidades de cada público, dos formatos e linhas editoriais e (que alterarão o tipo de linguagem) Para quem eu estou comunicando? Quem é o público deste veículo? O que consigo apresentar neste espaço de tempo?Abarca comentários sobre quem são as fontes que estão sendo consultadas ou deveriam ser pelos jornalistas e como tais enfoques direcionam as notícias. ǿ Perspectivas científicas - trata das percepções sobre a inclusão dos cientistas/pesquisado-res nas reportagens/notícias assim como outras lógicas científicas. Como a ciência colabora para produção de engajamento via jornalismo? ǿ Autoridades políticas e empresariais - trata das percepções sobre empresas, políticos e demais tomadores de decisão nas reportagens/notícias. Como as autoridades políticas e em-presariais são percebidas no jornalismo? Como isso colabora para o engajamento? ǿ Pessoas afetadas – aquelas que já sofrem em decorrência das mudanças climáticas. Os afetados estão sendo ouvidos? Inclui a perspectiva da justiça climática.Trata de como é feito o ativismo em grupos, movimentos sociais, ações relacionadas aos esforços coletivos, em comunidade. O que gera mobilização coletiva? Quais os desafios da mobilização social?Reúne percepções de como o jornalismo aborda as mudanças climáticas.Engloba diferentes percepções de como ocorre o engajamento, sem distinção de ações individuais ou coletivas.Reúne percepções explícitas sobre o papel do jornalismo para o engajamento climático. Aqui coletamos impressões mais gerais sobre confiança, credibilidade e outros aspectos amplos do jornalismo e do engajamento que extrapolam as subcategorias abaixo.METODOLOGIA50JORNALISMO E ENGAJAMENTO CLIMÁTICOPESQUISA 202251EFEITO DE CHOQUEEngloba apelos de medo e de amplificação dos efeitos negativos (pode-se dizer sensacionalis-mo) para despertar as pessoas para ação (há divergências sobre a eficácia de tal estratégia). Como o jornalismo demonstra a urgência e gravidade da crise climática?RELAÇÃO DO PROBLEMA COM O BOLSOCHAMADO EXPLÍCITODESNATURALIZAÇÃOTrata da vinculação econômica à crise climática; forma de trazer determinados públicos para pensar e agir pelo clima. Como o jornalismo trata do impacto econômico da crise?Percepções de que o jornalismo poderia ser mais engajado e ultrapassar a fronteira de apenas “informar”. O que o jornalismo faz ou deveria fazer para engajar?Reúne percepções e entendimentos de que o jornalismo deveria romper com a “naturalização” ou “normalização” de certas visões de mundo. Qual o papel do jornalismo na naturalização ou desnaturalização dos acontecimentos?CONEXÕES COM O CLIMACOMUNICAÇÃO COMUNITÁRIA E/OU ALTERNATIVAINFLUENCERS DIGITAISMANIFESTAÇÕES ARTÍSTICASPOVOS DA FLORESTA/AMAZÔNIDASPOLÍTICA AMBIENTAL E QUESTÕES ECONÔMICASTrata da compreensão entre clima, desmatamento e outros aspectos derivados dessa relação, como crise hídrica. Qual a relação entre os assuntos?Envolve jornais populares, rádios comunitárias, atividades autônomas independentes, conver-sas presenciais.Manifestações sobre o papel de influenciadores digitais e formadores de opinião que se enga-jam de forma a chamar para a ação, porém não no papel de fontes jornalísticas ou de profissio-nais de jornalismo.Diz respeito ao uso da arte e cultura, em suas mais diversas formas, como teatro, música, rap, bata-lhas de slam, grafite, filmes de ficção, fotografia, festas culturais, etc. para engajar sobre o clima.Reúne percepções sobre a floresta e seus habitantes invisibilizados, indígenas, quilombolas, além de seus conflitos. Quem são as pessoas envolvidas nessa relação? E de que forma?Reúne aspectos das políticas públicas, da gestão governamental, as discussões internacionais e de cunho econômico que envolvem desmatamento na Amazônia. Por que esse tema é de interes-se público? A quem interessa?Engloba manifestações sobre as ligações entre desmatamento, Amazônia e mudanças do clima.Abarca manifestações que extrapolam a lógica do subcampo jornalístico.DESMATAMENTO NA AMAZÔNIACOMUNICAÇÃO PARA ALÉM DO JORNALISMOPROXIMIDADE GEOGRÁFICATrata de comentários e manifestações sobre a necessidade de trazer situações próximas a vida das pessoas (perspectiva local x global). Como o jornalismo aproxima a população com a questão climática?Fonte: Os autores (2022)." É necessário aproximar o tema, com afeto, das realidades vividas, através de conexões com os grupos mais vulnerabilizados pela crise do clima, de forma transversal e interseccional com a sociedade, observando questões como raça, gênero, classe e idade. A partir da representatividade, da democratização da linguagem, da materialidade e da apresentação de soluções, é possível trazer um jornalismo climático inclusivo, acessível e efetivo. (GF NORDESTE, 13.07.2022)PESQUISA 202253GRUPOS FOCAISO jornalismo é considerado por muitos autores como um propulsor importante de mudanças socioambientais em razão de sua credi-bilidade e alcance na esfera pública (BUENO, 2007, GIRARDI ET AL., 2012; NEELIMA; REDDY, 2014, LOOSE, 2021) especialmente quando direcionado a públicos específicos (APPELGREN; JÖNSSON, 2021) e utilizam enquadra-mentos que envolvam emocionalmente os indivíduos, de modo a orientá-los para ação (BOLSEN; SHAPIRO, 2017). Ao escolher consumi-dores de notícias que se identificavam como ativistas, parte-se do olhar daqueles que já se tornaram engajados para recolher pistas de como a imprensa poderia mobilizar mais. Nesse primeiro momento, a perspectiva individual (do que é preciso para eu me envolver) foi citada, predominantemente, sendo acionados sentimentos associados à empatia, à conexão, ao pertencimento e à proximidade com as causas. Também a indignação, a injustiça e fatos que impactam foram citados como possíveis desencadeadores de uma preocupação que pode vir a se tornar engajamento.PERCEPÇÕES RECORRENTES ENTRE OS ATIVISTAS DE TODAS AS REGIÕES BRASILEIRASENGAJAMENTOMas, afinal, o que gera engajamento?"Você tem que sentir familiaridade, coisas que te impactam e te induzem a lutar por aquilo, lutar por aquilo que você não pode aceitar. (ATIVISTA DO NORDESTE)GRUPOS FOCAIS54JORNALISMO E ENGAJAMENTO CLIMÁTICOPESQUISA 202255Foram menos frequentes as respostas ligadas ao engajamento co-letivo. Geralmente a questão coletiva aparecia após o envolvimento individual.Nesse sentido, a informação (por meio do jornalismo) é, às vezes, citada para mostrar os impactos que nem sempre podem ser sen-tidos na pele por todos, assim como permitir conhecer o que está acontecendo (partindo da ideia de que só cuidamos daquilo que conhecemos).""""Eu acho que essa aproximação das questões ao seu dia a dia e àquilo que você conhece [...] eu acho que, isso leva a agir. Quando as pessoas percebem que o problema tá no hoje, [...] quando ela começa a perceber que esses problemas não estão lá longe [...] aí eu acho que isso leva, consegue engajar mais e levar à ação. (ATIVISTA DO NORDESTE)Tinha pensado na identificação primeiro, mas acho que também a emoção, pra mim, o que começou a me mover foi esse sentimento de revolta, mas depois passou a ser algo que eu precisava acreditar que era possível. (ATIVISTA DO SUL)E também é muito inspirador ver que outras pessoas estão fazendo coisas, tipo se outras pessoas estão fazendo e eu não, é quase um desrespeito em relação ao trabalho dessas pessoas. (ATIVISTA DO SUL)[...] implicar o indivíduo, sentir que tem um papel nisso, como isso me afeta, e como que os modos de vida e os hábitos, enfim, estão todos conectados. Então acho que é um pouco isso: trazer a responsabilidade individual e coletiva, não no sentido só daquela coisa mais simples, de economizar água, reciclar... existem processos muito maiores, como você participa de uma cadeia e apoia práticas melhores ou piores. Eu acho que um pouco dessa mobilização de indivíduos e de organizações. (ATIVISTA DO CENTRO-OESTE) JORNALISMO CLIMÁTICO E ENGAJAMENTO POR MEIO DO JORNALISMOReunimos nesta seção as duas categorias que abordam especifica-mente o jornalismo por compreender uma série de intersecções en-tre as subcategorias. Por exemplo: os enfoques da cobertura orien-tados para os efeitos podem ser articulados às falas associadas ao efeito de choque, assim como as com foco nas soluções costumam ser atreladas à proximidade geográfica, aos exemplos positivos que vêm sendo sustentados pelas comunidades.De acordo com os cientistas, o jor-nalismo desempenha um importante papel na amplificação da discussão, definição de sentidos, apresentação de argumentos, valores e visões de mundo a respeito das mudanças climáticas (LOOSE; CARVALHO, 2017; HULME, 2009). Contudo, será que essa é a per-cepção do público?Quando se pergunta de forma explíci-ta como o jornalismo pode contribuir com esse engajamento, há, em um primeiro momento, uma série de falas que redundam no próprio fazer jor-nalístico, como, por exemplo, o papel de levar a informação para um público amplo. Essa função permi-tiria que as pessoas ficassem cientes dos problemas ambientais e, quiçá, fizessem algo a respeito. Todavia, no desenrolar das falas, aparecem aspectos mais específicos, que tornariam os discursos mais efetivos no que tange à mobilização, como a aproximação das pautas climáticas com a vida das pessoas (algo bastante recorrente na literatura sobre comunicação climática, já que envolve um fenômeno global que afeta diferentes escalas), dando ênfase à necessidade de se ouvir quem já está sendo afetado e/ou luta nessa causa, como quilombolas, ribeirinhos, povos indí-genas, populações rurais, e classes sociais vulnerabilizadas. Um dos ativistas do Centro-Oeste destaca que já há jornalismos nessa direção, mas que eles não possuem o alcance que a imprensa hegemônica possui:Corroborando com esta ideia, lembra-se que o Jornalismo Am-biental abarca a cobertura dos riscos climáticos (LOOSE; GIRARDI, 2017) apresentando uma série de pressupostos orientados para a De acordo com os cientistas, o jornalismo desempenha um importante papel na amplificação da discussão, definição de sentidos, apresentação de argumentos, valores e visões de mundo a respeito das mudanças climáticas. Contudo, será que essa é a percepção do público?"Acho que o jornalismo hegemônico tem que aprender com o jornalismo ativista. Porque já existe um jornalismo que pauta todas essas questões e já existe um jornalismo que se engaja. Eu acredito que o jornalismo hegemônico, aquele de grandes empresas, [...] acho que ele precisa aprender com esses jornalistas que já tratam desses temas. (ATIVISTA DO CENTRO- OESTE)GRUPOS FOCAIS56JORNALISMO E ENGAJAMENTO CLIMÁTICOPESQUISA 202257pluralidade de vozes, maior contextualização, responsabilidade com a mudança de pensamento, cobertura sistêmica e próxima à realidade do leitor, assimilação do saber ambiental e comprometimento com a quali-ficação ambiental, sendo que a discussão das soluções e a incorporação do princípio da precaução também devem ser considerados. Tal prática é realizada, com diversas ênfases, pelos veículos especializados, sobre-tudo os alternativos, frequentemente citados pelos participantes, razão pela qual a ênfase parece ser na distribuição desses conteúdos - e não necessariamente na falta deles.Destaca-se ainda o papel pedagógico, alinhado ao dever de simplificar os termos técnicos e tornar mais palpável as questões complexas que envolvem a crise ambiental/climática. De forma geral, colocam-se exem-plos de como aproximar uma crise tão ampla, de caráter global, com o cotidiano das pessoas:Em acordo com Moser (2010), ativistas entendem que o jornalismo pode-ria buscar envolver os públicos tanto para o exercício da cidadania como para uma transformação no âmbito dos valores e das ideias vigentes, embora relatem não observar isso com frequência:""""[...] acho que a nossa casa tem muita coisa que está conectada com esses processos sistêmicos, então, se eu tenho um computador, tem a ver isso; se eu como comida, tem a ver isso. Então eu acho que muita coisa começa em casa, no sentido de: ‘vamos trazer conexões para o dia a dia das pessoas’. (ATIVISTA DO NORDESTE)Nunca vi algo na TV que me despertasse sentimento de causa que, tá... agora preciso tomar uma ação. É sempre uma coisa assim que... já passou ou tem alguém fazendo alguma coisa lá, então tá tudo certo, ou não é tão grave. Mostram o problema, mas não como realmente é, a profundidade da coisa, sabe? Então, eu acho que não comove o suficiente as pessoas para tomarem uma ação. (ATIVISTA DO SUL) Considerando que a pauta climática é interseccional, complexa e bastante dependente do campo científico, saber traduzir, tornar compreensível e acessível uma notícia sobre o tema, é mais um papel relevante do jornalismo. A tradução ou simplificação dos vocábulos e conceitos foram recorrentemente citados por todos os grupos. Pontua- se aqui que a perspectiva científica é criticada por ser difícil, abstrata e distante da população, porém reconhece-se que não dá para suprimi- la: “A gente precisa da ciência e do que a ciência produz, então a lógica é desmistificar esses espaços e essa linguagem e transformar”. (ATIVISTA DO SUL) Então acho que o exemplo é uma ótima forma de engajar, de fazer as pessoas se moverem mais em prol de alguma coisa e também mostrar o que tá sendo feito mais próximo da pessoa pra ela ver que próximo a ela tem uma iniciativa sendo feita pra melhorar a situação. Então, ela pode se sentir motivada a participar também. (ATIVISTA DO SUL)Tal posição coincide com a discussão acadêmica sobre quem seriam as fontes de informação do jornalismo climático. Schäfer e Painter (2021) assinalam que as autoridades e cientistas deixam de ser as únicas fontes, para compartilharem espaço junto com as ONGs que trabalham com o tema, de forma a tornar a questão mais palpável e próxima do cotidiano das pessoas.Após assistirem aos vídeos já citados, os participantes, em sua maioria, dizem não perceber como esse tipo de jornalismo pode engajar, embora façam ressalvas de que há aspectos positivos (como apresentação de soluções e menção aos rios voadores que surgem da Amazônia, por exemplo) e que tudo depende do público de tais reportagens. Muitos afirmam que, por já estarem sensibili-zados com a causa, o vídeo interessa e pode servir de subsídio para outras atividades, porém, ao pen-sar nas pessoas distantes do tema, apontam uma série de problemas: uso de jargão técnico (mitigação, COP, gases de efeito estufa, precipitação, etc.), gráficos de difícil compreensão, uso da língua inglesa apenas com legenda1, fontes apenas do campo científico/ ausência de fontes locais/dos afetados, silenciamento sobre os responsáveis pelo problema, etc.Nessa etapa verificou-se que há ativis-tas que relativizam as especificidades das linhas editoriais e dos públicos, assim como os limites de uma reporta-gem televisiva, e outros que realizam 1Não é usual na imprensa brasileira fazer uso de legendas, mas o vídeo escolhido, de um meio não hegemônico, optou por não dublar ou reformular a fala do secretário- geral da ONU.[...] há ativistas que relativizam as especificidades das linhas editoriais e dos públicos, assim como os limites de uma reportagem televisiva, e outros que realizam críticas que, mais uma vez, extravasam o campo de atuação, como, por exemplo, a questão dos monopólios de comunicação, que reduzem a diversidade dos conteúdos que chegam às casas da maioria dos brasileiros.GRUPOS FOCAIS58JORNALISMO E ENGAJAMENTO CLIMÁTICOPESQUISA 202259críticas que, mais uma vez, extravasam o campo de atuação, como, por exemplo, a questão dos monopólios de comunicação, que redu-zem a diversidade dos conteúdos que chegam às casas da maioria dos brasileiros. A conexão da Rede Globo com o agronegócio foi citada em diferentes momentos, demonstrando uma ciência dos ativistas das es-truturas econômicas e políticas que afetam o jornalismo mainstream:Registra-se ainda que, apesar do debate posterior à passagem dos vídeos sinalizar para muitas faltas (não têm fontes locais/afetadas, não responsabilizar os culpados, não é didático, não mobiliza, etc.), alguns manifestaram o entendimento de que esses e mais vídeos deveriam ser mais frequentes:Nas duas partes dos grupos focais (antes e depois da apresentação dos vídeos), muitos ativistas mencionaram a necessidade de mostrar o que podemos fazer, muitas vezes ponderando o real efeito das ações individuais (como redução do consumo de carne ou economia de energia) em comparação às ações de grandes setores e governos:Ainda que mencionem a necessidade de se tratar dos impactos/consequências/efeitos, reforçam que é preciso ir além, mostrando o que podemos fazer diante de tal cenário. Essa posição foi majo-ritária nos GFs e coincidente com o que pesquisadores dessa área manifestam: “o enfrentamento das ACs [alterações climáticas] requer estratégias diversas, associadas a cada perfil de público, de modo a evitar generalizações que acarretem mais apatia e deses-perança do que respostas ao problema” (BALBÉ; LOOSE, 2020, P.50).""""""[...] como fica uma comunicação para passar nos meios de maior circulação, que são financiados por essas indústrias, acho que é um desafio e tanto, ainda bem que eu sou bióloga e não jornalista”. (ATIVISTA DO SUDESTE)Uma forma de poder contribuir ainda mais é se descentralizando mesmo, quanto mais plural e quanto mais vozes puderem existir, sobretudo pra causa das mudanças climáticas, mais rico vai ficar o debate. (ATIVISTA DO NORDESTE)[...] o que que nós enquanto cidadãs e cidadãos podemos fazer e aí pensando também naquilo que a colega falou, a ação do indivíduo, levando também para algo que gere uma mobilização mais coletiva e o engajamento político nesse tema. (ATIVISTA DO SUL)Acho que não tem um investimento dos problemas climáticos estarem nas grandes mídias, nos grandes veículos que vão chegar em todo mundo, na massa. Eu vejo esse ponto da forma como a informação é passada, ser muito frágil nessas questões que são ambientais, porque são impessoais às vezes, e daí a pessoa não consegue enxergar onde é que aquilo ali está afetando ela, tipo: o que eu tenho a ver com as calotas de gelo derretendo lá no extremo do globo? (ATIVISTA DO NORDESTE)[É preciso] ir conectando e fazer um acúmulo, né? Porque eu sei que sempre tá chegando leitor novo e tal, para explicar que combustível fóssil vai... a causa do aquecimento do planeta. Mas a gente tem que... acho que contar com um acúmulo que cada reportagem vai somando, que puxa a outra, seja no escrito, somando com um vídeo, somando com a televisão, mas tem que incentivar esse acúmulo, de fazer esse resgate do que já foi feito, já foi falado antes. (ATIVISTA DO NORTE)Além do reconhecimento do problema, [...] teve que ter uma conversa sobre o que que dá para fazer. E aí construir esperança dessas pessoas. Tipo assim: "ó, a gente é limitado, a gente consegue fazer x, mas se a gente tiver junto a gente consegue fazer tanto". E aí, conversar [...] sobre solução, falar sobre possibilidades [...] vamos construir junto essas possibilidades, acho que isso gerou mais engajamento, tanto para trazer pessoas para o coletivo quanto pra a gente conseguir atingir os territórios que a gente consegue atingir. [...] essa possibilidade da pessoa tá envolvida também na solução do problema acho que é uma coisa que engaja muito. (ATIVISTA DO NORTE)DESMATAMENTO NA AMAZÔNIANa última questão, que provoca os ativistas a pensar sobre o por-quê do desmatamento da Amazônia aparecer em um dos vídeos, surgem afirmações que relacionam adequadamente a destruição da floresta com a maior emissão de gases de efeito estufa e, conse-quentemente, com a intensificação das mudanças climáticas, mas há também leituras associadas à visibilidade nacional e internacio-nal do tema, articulando à política federal deste governo: “porque o desmatamento é um projeto de governo que está sendo instaurado GRUPOS FOCAIS60JORNALISMO E ENGAJAMENTO CLIMÁTICOPESQUISA 202261agora” (ativista do Sul). Nesse sentido, muitas críticas foram feitas sobre o desmonte das políticas ambientais brasileiras e o descaso com os ambientalistas e povos da floresta, principalmente devido à grande repercussão midiática que teve os então recentes assas-sinatos do jornalista britânico Dom Phillips e do indigenista Bruno Pereira Araújo na Amazônia.Embora, de uma forma geral, tiveram muitos relatos associados às subcategorias “proximidade geográfica” e “perspectivas das pessoas afetadas”, no qual os ativistas reivindicavam protago-nismo para aqueles que já estavam atuando na causa ambiental, questões envolvendo diretamente os povos da floresta/amazôni-das foram pontuais:Complementar a essa observação, verificou-se que ainda é muito distante falar da Amazônia, tendo em vista as dimensões continen-tais e a diversidade de biomas do Brasil.""A gente vê também o desmatamento sendo muito falado porque quem está defendendo a floresta tá recebendo ameaças de morte; nosso Brasil é um dos países que mais tem matado, acho, assassinado ambientalistas, pessoas que defendem a natureza. (ATIVISTA DO NORDESTE)Então, eu acredito que a mensagem mesmo é a gente se compreender enquanto povo amazônida que vive aqui nessa região e daí partir essa mensagem de verdade [...], e do que está acontecendo agora, não é do que vai acontecer, é do que já está acontecendo, então, eu acredito dessa forma. (ATIVISTA DO NORTE)COMUNICAÇÃO PARA ALÉM DO JORNALISMOApós o GF pré-teste, sentiu-se a necessidade de categorizar insi-ghts e outras discussões que não fazem parte do escopo jornalís-tico, mas são outras formas de comunicar e engajar para o enfren-tamento das mudanças climáticas. Durante a realização dos GFs regionais essa categoria foi menos acionada do que o imaginado, já que ao se referir ao jornalismo, os ativistas consideraram modos de fazer ampliados, misturando gêneros comunicacionais, o que observamos em interface com o conceito de hibridismo midiático. Entretanto, vale ressaltar que houve uma ênfase na prática da co-municação comunitária, que pode ser entendida pela mobilização dos próprios habitantes dos territórios, visando utilizar dispositivos de comunicação e produzir conteúdos que atendam aos interesses das suas comunidades (PERUZZO, 2006). Ela decorre da sistemática exclusão feita pela comunicação hegemônica e da dificuldade financeira de custear o acesso aos meios, buscando fazer circular as informações que necessitam para sua integração e organização social. Um exemplo são as rádios comunitárias, cuja programação é montada pela própria comunidade.O jornalismo comunitário, por sua vez, busca essa aproximação das comunidades. Ao difundir suas rei-vindicações, trabalha para reduzir a sub-representação desses cidadãos nas mídias, em especial na grande imprensa, cuja ênfase às comuni-dades mais distantes das sedes dos poderes está sobre acontecimentos extremos, aspectos negativos, rati-ficando estereótipos e preconceitos, em vez de aprofundar os porquês da injustiça social e das desi-gualdades (GALLI, 2021).""Não é só a gente direcionar olhando de cima, mas, às vezes, que o comunicador surja desse público, que a gente quer falar. [...] Então, cativar outras pessoas para se entenderem como comunicadores das suas comunidades, para poder adaptar aquela notícia na linguagem, no formato, que melhor é atendida para a comunidade que faz parte, acho isso muito bacana, uma rede. (ATIVISTA DO SUDESTE)Eu trabalho muito com mídia de favela, mídias comunitárias e tal, e eu acho que essa galera tem feito um trabalho muito, muito incrível assim. Aqui no Rio e em São Paulo. Eu acompanho muito as mídias comunitárias de SP e eles têm feito um trabalho muito incrível, porque eles têm trazido essas discussões que ficavam muito restritas a um público diferente. Tem trazido essas discussões de uma forma mais simples, mais acessível. Se você abrir esses veículos eles estão discutindo racismo ambiental, estão discutindo ilhas de calor, estão discutindo o que é justiça climática, de uma forma acessível. (ATIVISTA DO SUDESTE)O jornalismo comunitário, [...]. Ao difundir suas reivindicações, trabalha para reduzir a sub-representação desses cidadãos nas mídias, em especial na grande imprensa [...]GRUPOS FOCAIS62JORNALISMO E ENGAJAMENTO CLIMÁTICOPESQUISA 202263Em relação a outras abordagens que excedem o jornalismo, um participante da região Sul citou uma iniciativa de um cartunis-ta como uma forma de traduzir o conhecimento científico para mais públicos:Tal relato coincide com a necessidade de recursos visuais e de animação com finalidade didática e informativa nas notícias, con-siderada escassa em pesquisas antecedentes (PARRATT FERNÁNDEZ; MERA FERNÁNDEZ; ABEJÓN MENDONZA, 2021; LOOSE, 2021).Outro ativista do Sudeste mencionou uma carência de autori-dades que tenham a "confiança" do público, o que entendemos como referências que gerem empatia ou que sejam carismáticas:Lembra-se que alguns jornalistas acabam assumindo um pouco dessa função nas redes sociais, como ocorre com jornalistas bas-tante reconhecidos na área ambiental, como Paulina Chamorro e André Trigueiro. Entretanto, identificou-se que tais referências não se limitam ao campo jornalístico."Quando o Dráuzio Varela vai falar alguma coisa, eu penso: ‘é sério isso aqui’. Eu, pelo menos. Eu sinto falta de autoridade. [...]. Eu sinto falta de uma pessoa que seja daqui e que nos oriente, que a gente confie. Porque assim, quando fica muito impessoal...tem dados, tem imagens, tem muitas coisas, mas não tem alguém, uma figura aqui, uma pessoa que esteja ali falando e talvez, que tem uma trajetória que a gente vai acompanhando. (ATIVISTA DO SUDESTE)"Tem um perfil chamado “Árvore Ser Tecnológico”. É um cartunista, um ilustrador que faz uns desenhos incríveis e ele traduz as coisas numa linguagem muito bacana, que é essa linguagem do desenho. Então você pega um público que vai começar a entender as coisas, porque você usa um outro tipo de linguagem, que é o desenho. Então, acho que pensar nisso é sempre importante na hora de desmistificar o linguajar científico e traduzir isso para outras coisas, porque, muitas vezes, é complexo. (ATIVISTA DO SUL)VISUALIZAR AS SOLUÇÕES →Agrofloresta em Santa Catarina. Foto: © Peter Caton / GreenpeaceGRUPOS FOCAIS64JORNALISMO E ENGAJAMENTO CLIMÁTICOPESQUISA 202265Os grupos focais do Sul contemplaram pouca diversidade de raça ou cor e de gênero – um dos grupos foi composto apenas por mu-lheres, o que não se repetiu em nenhuma outra região, e apenas um participante não se autodeclarou branco. Em estados fortemente dependentes do agronegócio, a escassez de chuvas, que impacta a produção agrícola, foi associada ao clima e ao desmatamento da Amazônia. Também as foram citadas, como forma de aproximar as mudanças climáticas com a realidade das pessoas:Mais do que em outras regiões, o apelo econômico, que mexe no bolso do consumidor, foi citado: “Acho que mostrar como afeta diretamente, às vezes apresentar a escassez, muita chuva [...] daí as plantações não deram certo, isso aumenta o preço [da alimenta-ção]. O que afeta a pessoa no bolso infelizmente é o que as pessoas ficam mais de olho”.Mesmo inseridos em uma discussão de que apenas ações individu-ais não são suficientes, as conexões das mudanças do clima com o aumento do preço dos produtos foram citadas por ser uma forma de mostrar o impacto diretamente na vida das pessoas. Isso pode ESPECIFICIDADES POR REGIÃOREGIÃO SULter ocorrido em função das representações reiteradas dos meios locais e mesmo da mídia hegemônica que tem seu olhar a partir do eixo Rio- São Paulo, e tentam aproximar a destruição da flo-resta com os prejuízos econômicos das lavouras. Sublinha-se que isso não é um problema em si, mas reduz a questão e invisibiliza outras questões. Por isso, a cobertura sistemática, com diferen-tes abordagens, se faz necessária, especialmente em um país de tamanho continental como o Brasil, que possui distintas realidades no mesmo território.Outro aspecto que chamou atenção foi a de um jornalismo que se coloca como ativista (advocacy journalism), fazendo um chamado explícito a seus públicos, pois apenas cobrindo o tema dentro da lógica jornalística tradicional seria muito difícil engajar. O chamado explícito se repetiu nos comentários como uma forma de mostrar que as pessoas também têm responsabilidades:Os participantes entendem que o papel de informar está sendo cumprido, mas não é suficiente, inclusive pode gerar banalização “[...] a informação está circulando, só que, muitas vezes, isso gera quase um esvaziamento da pauta”. Há estudos que mostram que a cobertura climática mundial (AKAS, 2022) e da América Latina (RA-DAR CLIMÁTICO, 2022) representam menos de 2% das notícias gerais publicadas pela imprensa, mas o fato é que há uma populariza-ção do tema que deriva em uma falsa sensação de já sabermos muito sobre ele, como foi verificado pela pesquisa do ITS-Rio/Yale (2022) no Brasil quando compara preocupação e conheci-mento sobre o tema.Endossaram ainda a necessidade de estar com as bases: "[...] com as comunidades indígenas, quilombolas, enfim, com os bair-ros, com as pessoas que estão vivendo mesmo as coisas na pele, porque, às vezes, a gente se deixa levar por coisas grandes, inter-nacionais, por grandes eventos e grandes acontecimentos e todos os dias têm coisas cotidianas que acontecem e não são divulgadas, "Acho que trazer o impacto para o quintal, né? E outra coisa mesmo quando você fala de trazer os impactos para o Sul Global, você também tem muitas diferenças aqui né, então não adianta nada eu dar exemplos da Amazônia para Santa Catarina, por exemplo, as pessoas não entendem, elas não vão entender. Eu preciso falar sobre enchente, sabe, que é o que afeta anualmente a população aqui nas cidades ou no campo ou crise hídrica, né, escassez hídrica, que são temas que estão afetando as pessoas aqui com o conhecimento da Amazônia lá. (ATIVISTA DO SUL)"[...] Para o que você tá fazendo agora, escuta isso, a gente precisa fazer alguma coisa juntos e juntas. É urgente. É pra já e você tem que fazer algo no seu dia a dia né. [...]. Cada pessoa que está falando, você tá escutando a pessoa, tá falando com a pessoa diretamente, como se estivesse olhando no olho, se referindo a cada indivíduo, não tanto como um contexto geral informativo, de que o mundo está passando por isso, não, mas que você tem essa responsabilidade. [...] (ATIVISTA DO SUL)GRUPOS FOCAIS66JORNALISMO E ENGAJAMENTO CLIMÁTICOPESQUISA 202267não tem nem espaço para falar sobre elas, então eu acho que isso é muito importante”. A questão da proximidade geográfica abarca tanto a pluralidade de vozes quanto a articulação de causas--consequências-soluções, indicada na literatura como forma do jornalismo se aproximar dos públicos.Os participantes do Sudeste trazem exemplos relacionados aos efeitos das chuvas intensas para ilustrar a importância de abor-dagens jornalísticas que considerem a responsabilidade do poder público nas causas e consequências das mudanças climáticas. Infere-se, portanto, que esta é uma questão próxima à realidade dos participantes.Os ativistas desta região falaram bastante sobre a necessidade de se cobrir a atuação das instituições, sobretudo o Estado, com a im-plementação de políticas públicas que auxiliem no enfrentamento das mudanças climáticas. Para eles, é importar mostrar os respon-sáveis: “[...] o vídeo não informa, porque ele não diz quem é o vilão”.Em termos de linguagem, citaram a necessidade de construir notícias que afetem o outro, em oposição a textos carregados de números, que não são apreendidos pela maioria dos públicos. "Quando você desgasta muito uma coisa da mesma forma, como é o caso da Amazônia, por exemplo, [...] 'Já queimou 50 ha, se quei-mar 52 não vai fazer diferença'. Gera essa banalização”. A crítica aos enquadramentos repetidos e ao uso de números que não permitem a dimensão real do fato ocorreu em outras regiões pesquisadas.Com mais destaque do que em outros grupos, ativistas associa-vam a cobertura do clima à própria lógica neoliberal que man-tém nosso sistema capitalista. Houve apontamento de tal sistema como grande responsável pela crise ambiental, trazendo à tona as falsas soluções divulgadas a partir da individualização: desligar a torneira enquanto milhares de litros estão sendo consumidos para produção de carne, por exemplo.Quando trataram da proximidade geográfica e da necessidade de se conectar com as vivências, trouxeram aspectos associados ao REGIÃO SUDESTEREGIÃO CENTRO-OESTEpouco conhecimento e, consequentemente, pertencimento que os brasileiros possuem em relação aos nossos biomas:Ao tratar do desmatamento da Amazônia, os ativistas do Sudeste mencionaram as populações indígenas, a fauna e a flora, reve-lando a importância do território de forma ampla (não apenas da crise climática ou associada a questões econômicas e políticas):Os participantes dessa região trouxeram o debate para a realidade do bioma Pantanal, conectando o desmatamento da Amazônia com a mudança nos regimes de chuva e consequentemente com as quei-madas, associadas à estiagem prolongada. A questão da proximida-de geográfica foi pontuada de diferentes maneiras, sendo relevante, na percepção dos ativistas, dar voz àqueles que são impactados e exemplificar os efeitos a partir da realidade local: """Entender a realidade que te cerca tem a ver, porque a gente não tem muito pertencimento. E eu nem ia tão longe nas calotas polares, a gente não tem pertencimento da própria Mata Atlântica, da floresta amazônica... quer uma coisa que a gente não tem conhecimento nenhum? É a Caatinga, a Caatinga não é meio ambiente para as pessoas, ‘claro que não, porque é meio ambiente o que é verde’... Então, trazer essa realidade de que isso interfere diretamente, interfere no seu poder de compra, interfere na sua saúde, interfere na água que você consome ou deixa de consumir. (ATIVISTA DO SUDESTE)[...] ontem teve a veiculação no Instagram de uma fala indígena que foi vinculada acho que até pela Célia Xakriabá, uma ativista né, falando assim 'Vocês têm que começar a olhar pra Amazônia a partir de quem está dentro da Amazônia, das pessoas que preservam a floresta, porque não é ficar olhando para a Amazônia através de satélites. (ATIVISTA DO SUDESTE)[...] acho que é trazer mais para a realidade das pessoas, as pessoas estão vendo que a água está diminuindo, [estão] com mais dificuldade de encontrar peixe, você tem lá a sua agricultura familiar, a sua horta, sua produção de subsistência que GRUPOS FOCAIS68JORNALISMO E ENGAJAMENTO CLIMÁTICOPESQUISA 202269está mais difícil, no ano passado não deu aqui perto... É lá no Mato Grosso, na verdade, tem um quilombo que faz todo ano a festa com a banana, há décadas já, [ano passado] não teve banana, porque não tinha água. Então todo mundo consegue entender o que está acontecendo, se trazer de uma maneira um pouquinho mais concreta. (ATIVISTA DO CENTRO-OESTE)Entretanto, não é o suficiente ouvir as fontes locais, mas entender com certa profundidade o contexto. Assim, surge a crítica das coberturas sobre a região feitas por profissionais do Sudeste desconhecedores das dinâmicas dos territórios; segundo os ativistas, tal produção perderia força, pois é um "olhar de fora": Sobre o papel do jornalismo, a função política, trazida por Bueno (2007) quando trata do Jornalismo Ambiental ficou evidente: “[...] o jornalismo podia funcionar de uma forma de conscientizar as pessoas, no sentido de que elas podem ajudar a pressionar tomadores de decisão, diante dessas questões, sabe?". Para tanto, é posto que ele não pode ser isento, mas engajado, e atuar de forma interseccional: “[...] tem um costu-me muito forte do jornalismo de não falar de forma transversal sobre esses assuntos, [como] gênero, mudanças climáticas, população indígena. Então começar a trazer essas questões de forma transversal, falar de economia para falar de mudanças climáticas, falar de política para falar de mu-danças climáticas, trazer isso em várias editorias”. Somado a isso é assinalado que o jornalismo precisa prestar atenção, e “tornar o familiar estranho”, no sentido de questionar algumas práticas e discursos que são naturalizados, como “sempre pegou fogo desse jeito”.A região Nordeste registrou que há temas locais, como a deser-tificação, que são pouco tratados pela imprensa, impedindo que as pessoas entendam as razões de certos fenômenos próximos estarem acontecendo. Também abordaram que há uma suposta naturalização de que desmatamento só ocorre na Amazônia, como se outras regiões estivessem livres das atividades destru-tivas, como mineração e grilagem.Além disso, mais do que em outras regiões, as desigualdades socioeconômicas que tornam as pessoas mais vulnerabilizadas frente aos riscos climáticos (e outros) foram mencionadas, as-sim como a necessidade da imprensa cobrar mais governantes (por conta das políticas públicas) e empresários:Observa-se uma necessidade de maior denúncia frente à pa-ralisia de quem poderia tomar alguma medida preventiva: “[...] mostrar que as pessoas que estão nos lugares de poder não estão interessadas em fazer as mudanças que a gente precisa”. Assim como Andrade, Barreto e Henriques (2020), observa-se a crítica à falta de ações efetivas para o enfrentamento às mudanças climá-ticas. Nessa direção, há uma ideia de que o jornalismo deveria ser mais combativo, enfrentando os grandes setores que afetam a sustentabilidade do planeta (ao invés de se unir a eles, como faz a mídia hegemônica com o agronegócio).Finalmente, o próprio acesso à informação jornalística se mostrou algo bastante recorrente, pela própria questão de infraestrutura de internet ou TV. Seja por não ter acesso, sobretudo em regiões do interior, seja por falta de recursos para custear o acesso às notícias, as populações mais vulnerabilizadas ficam mais suscetíveis ao universo da desinformação. Diante de tantas carências, a cobertura da crise climática não se mostra conectada com tais situações: “[...] "Um jornalista aqui do Sudeste, que estudou em universidades do Sudeste, fez intercâmbio, voltou e falou "não, vocês têm que parar com a mudança climática", os caras olham e não dão credibilidade, né? Falam "ah, esse cara não entende, nunca viveu no mato, não sabe o que ele tá falando". É diferente de uma pessoa que tá ali, morando há 40 anos ali, viveu a vida inteira ali, e está sentindo os efeitos dessas mudanças. (ATIVISTA DO CENTRO-OESTE)REGIÃO NORDESTE[...] o jornalismo precisa prestar atenção, e “tornar o familiar estranho”, no sentido de questionar algumas práticas e discursos que são naturalizados, como “sempre pegou fogo desse jeito”."[...] muito essa visibilidade no que diz respeito à fiscalização, de exposição mesmo de situações, de exposição de grandes empresas, sei o quanto isso pode ser prejudicial pra essas empresas e pessoas que trabalham nelas, mas sinto falta dessa exposição, do quanto a gente vê empresas que dizem que fazem pelo meio ambiente, mas tem alguém pronto pra expor isso? (ATIVISTA DO NORDESTE)GRUPOS FOCAIS70JORNALISMO E ENGAJAMENTO CLIMÁTICOPESQUISA 202271existe situação complicada de falta de internet, de falta de recursos básicos... eu, particularmente, sinto aqui que parece que essa pau-ta do meio ambiente parece muito elitista, não é, mas parece, pelo menos para as pessoas das camadas mais baixas da sociedade”. Em tais contextos, formas específicas de jornalismo são indicadas: “[...] talvez um jornalismo mais comunitário, no sentido de estar mais próximo das camadas mais baixas, para trazer esse engaja-mento de baixo para cima”.Os ativistas acreditam que para engajar os públicos caberia ao jornalismo utilizar recursos de influenciadores através do infotenimento (DEJAVITE, 2006) e aproxi-mar a sua agenda midiática àquela dos comunicadores populares, indígenas, mulheres negras, e outros atores das regiões. Tal concepção de jornalismo ali-nha-se, em certa medida, à definição de Charron e Bonville (2016). Neveu (2006, p.164) expressa quem seria o profissio-nal nesse jornalismo de comunicação: "[...] age como vulgarizador, conselheiro, até como um amigo, mantendo uma re-lação de familiaridade com seu público, divertindo-o. Ele se despoja de sua postura de autoridade, de autor de uma relação cívica”. Assim como a Região Nordeste, há dificuldade de acesso às infor-mações jornalísticas específicas da área climática (há relato de que são poucos os jornalistas da região dedicados ao tema). Um dos respondentes do questionário inclusive apontou essa dependência a partir do seu estado:"No Tocantins existe uma falta de acesso e análise de dados, além do baixo interesse em cobrir determinados temas. A falta de investimento no jornalismo regional faz com que ele seja alimentado principalmente por press releases, assim, muitas vezes, a reportagem e investigação é deixada de lado e a narrativa dominante é advinda das assessorias de comunicação. (ATIVISTA DO NORTE)REGIÃO NORTEA falta de jornalismo local/regional sobre clima se reflete na queixa sobre pouca representatividade que os amazônidas possuem nas notícias, assim como seus reais problemas – ela é retratada quando afeta outras regiões ou interesses de ordem nacional/internacio-nal. Para os participantes, é preciso dar mais protagonismo para as populações locais que já estão há anos lutando pela proteção da floresta, pluralizando as fontes e entendendo a riqueza do bioma e de seus povos:A regionalização da cobertura cli-mática no Norte é um caminho para engajar por meio do jornalismo, evitando apenas as perspectivas científicas e trazendo soluções que façam sentido para a população. Por isso, a crítica à diminuição do consu-mo de carne foi emblemática, pois a dieta predominante dos habitantes é baseada em peixe: "Não dá pra você chegar aqui [Amazônia] e dizer que vai impactar na carne, sendo que aqui todo mundo só come peixe, vive de comer peixe. [...] Então é aproximar essa realidade do que está acontecendo hoje, do que vai impactar na realidade do território". Outra crítica realizada no campo das soluções foi a redução ou sim-plificação de uma única medida, como plantar árvores:""[dar protagonismo a] [...] essas vozes, pelas próprias pessoas que vivem e vivenciam, que estão na linha de frente do que as mudanças climáticas causam, do que a crise climática causa. Então, não tem como saber o que acontece se a gente não conhecer essas pessoas, não conhecer essas realidades e se a mídia não der espaço para que essas vivências possam ser visibilizadas dentro do debate público né. (ATIVISTA DO NORDESTE)Uma coisa que eu lembro muito era de um... um artigo de uma galera que falou que se plantasse um trilhão de árvores e aí saiu um milhão de jornais do mundo todo falando que se plantasse um trilhão de árvores ia tudo certo, e as pessoas simplesmente esqueceram que tem que parar de emitir carbono de combustível fóssil e outras coisas relacionadas, né? Então, às vezes do jeito que [se] trata as Os ativistas acreditam que para engajar os públicos caberia ao jornalismo utilizar recursos de influencers através do infotenimento e aproximar a sua agenda midiática àquela dos comunicadores populares, indígenas, mulheres negras, e outros atores das regiões.A regionalização da cobertura climática no Norte é um caminho para engajar por meio do jornalismo, evitando apenas as perspectivas científicas e trazendo soluções que façam sentido para a população.GRUPOS FOCAIS72JORNALISMO E ENGAJAMENTO CLIMÁTICOPESQUISA 202273Esta fala remete à necessidade de se discutir a complexidade das causas e a necessidade de um rol de ações. Repercutir com exaustão medidas paliativas pode gerar a falsa impressão de que a crise climática já pode ser resolvida, quando, na verdade, todas as ações que conhecemos podem ajudar na adaptação e na redução de seus efeitos, mas o que está em curso já não pode ser totalmente revertido.É essa gravidade, associada a múltiplos fatores e algumas incertezas (que precisam ser vistas como parte do fazer científico), que preci-sam ser devidamente comunicadas para que a sociedade compreen-da a dimensão do desafio ambiental no qual estamos imersos.soluções, também pode engajar pro lado errado, pode gerar inação, porque é uma coisa muito complexa, depende de múltiplas ações acontecendo ao mesmo tempo e enfrenta cenários de desigualdades muito grande. (ATIVISTA DO NORTE)Expedição "Sem Floresta Sem Água" no Brasil. Foto: © Gabriel Lindoso / Greenpeace VISUALIZAR AS CONSEQUÊNCIAS DA CRISE CLIMÁTICA ↓Gado em região de floresta desmatada no Mato Grosso. Foto: © Rodrigo Baleia / GreenpeaceVISUALIZAR AS CAUSAS E PRINCIPAIS RESPONSÁVEIS →O jornalismo precisa ter respeito e compreensão da representatividade das comunidades locais, além de incluir as vozes e valorizar as vivências dos povos. Além disso, necessita dialogar mais com as pessoas; o amanhã chegou, é hoje. Ouvir as pessoas que já estão reflorestando as mentes e seus territórios, re(e)xistindo em pé. (GF CNORTE, 02.07.2022)"PESQUISA 202275CONSIDERAÇÕES FINAIS A partir do exposto, percebe-se que as necessidades e lacunas mencionadas pelos ativistas das cinco regiões brasileiras podem ser traduzidas nesta pesquisa como orientações para elaboração de estratégias a serem colocadas em prática no jornalismo, so-bretudo naquele dito hegemônico. Os participantes da pesquisa concordam que é preciso trazer novos olhares sobre a pauta cli-mática, pois reconhecem o papel deste tipo de jornalismo, seja pela sua capacidade de abrangência, seja como forma de legi-timar ações locais ou de nicho (é quando a imprensa hegemônica cobre certos temas que percebem um maior interesse de “pessoas que estão fora de suas bolhas”, como alguns comentaram).Os ativistas desejam que esse jornalismo massivo se engaje para o enfrentamento das mudanças climáticas, chamando os públicos para a ação, até de forma explícita, e oferecendo infor-mações que considerem as especificidades locais e as cone-xões entre as regiões do país, bem como com a escala global. A ênfase na denúncia também é repetida: o jornalismo deveria assumir uma posição combativa ao expor os responsáveis pelos impactos ambientais/climáticos, em especial nas notícias sobre os eventos climáticos extremos. De acordo com os ativistas, é preciso cobrar os gestores pú-blicos e os setores produtivos (sendo que o mais criticado foi a agropecuária): “[...] penso que, falando de questões de crise climática, [é] difícil apontar culpados, ou não há interesse em apontar. [...] tu não vês na grande imprensa uma culpabiliza-ção das grandes empresas ou cobrança dos políticos. [...] falta comprar essa briga mesmo, cobrar grandes políticos e gran-des empresas” (ATIVISTA DO SUL).Como consumidores de jornalismo não hegemônico, principalmen-te especializado em temas ambientais, os ativistas mencionam [...] o jornalismo deveria assumir uma posição combativa ao expor os responsáveis pelos impactos ambientais/climáticos, em especial nas notícias sobre os eventos climáticos extremos.CONSIDERAÇÕES FINAIS 76JORNALISMO E ENGAJAMENTO CLIMÁTICOPESQUISA 202277que há esforços ocorrendo nesse sentido aqui no Brasil, mas eles deveriam ser replicados pelo jornalismo hegemônico (pensando em um engajamento mais efetivo da sociedade). Mesmo tendo reconhecido os vínculos dos grandes grupos de comunicação com os setores mais poluentes, através de patrocínios e veiculação de anúncios, por exemplo, seria uma mudança que poderia incidir sobre o conhecimento e posterior envolvimento com as ações que cercam a emergência climática.Outra questão recorrentemente referida por todos os ativistas é a necessidade de ouvir outras fontes, com ênfase nas locais, sejam elas ativistas ou afetados, como tentativa de aproximação do tema com a realidade local, visando gerar identificação e senti-mento de pertencimento pelas pessoas. As perspectivas a partir de fontes científicas, que representam grande parte das notícias que temos sobre o clima hoje, foram entendidas como impor-tantes para combater o negacionismo, porém não podem ser as únicas. Para eles, o jornalismo deve diversificar as perspecti-vas, visibilizando as iniciativas dos territórios e os movimentos que defendem a proteção da natureza, bem como àqueles que enfrentam os impactos ambientais e os efeitos das mudanças climáticas há mais tempo. Quando os vídeos foram apresenta-dos, ambos orientados pela perspectiva científica, a ausência de outras fontes, que aproximassem o problema do cotidiano dos públicos, foi bastante citada.A linguagem jornalística mereceu muita atenção dos ativistas quan-to à necessidade de ser menos formal, mais simples, de traduzir os termos técnicos, em especial os científicos, de buscar materializar/tornar mais palpável o significado dos números e demais dados de pesquisas por meio de recursos audiovisuais e das artes. Alguns chegaram a dizer que essa didatização das mudanças climáticas era uma das atribuições do jornalismo, afinal os ativistas reconhe-cem a dificuldade de acompanhar esse tema. Em alguns casos, a informação acessível e didática é posta como forma de enfren-tar a desinformação.A questão da imagem, provocada pela própria exibição de produtos audiovisuais, também foi trazida por ativistas. Na literatura, encon-tram-se evidências sobre as dificuldades do jornalismo representar as causas (retratar a emissão de gases de efeito estufa, por exem-plo) e acabar dando muita ênfase para os efeitos (fotografias de queimadas, enchentes, vendavais, etc.). Muitos reportaram que o desmatamento é trazido com frequência por ter imagens fortes, de derrubada de muitas árvores.A disparidade de acesso aos meios jornalísticos entre as regiões brasileiras também ficou evidente. Ativistas, muitos conhecedores das restrições das comunidades localizadas longe das capitais, exprimiram preocupação com o fato de o jornalismo nem sempre alcançar tais públicos, especialmente nas regiões Centro-Oeste, Norte e Nordeste.Quanto menos informações apuradas chegarem a essas comuni-dades, mais suscetíveis a informações enganosas essas pessoas ficarão. Mesmo não existindo ênfase sobre o negacionismo na im-prensa brasileira, chamou atenção que comentários sobre ceticis-mo, negacionismo e “fake news” apareceram durante os grupos de forma orgânica, o que sugere que a comunicação climática precisa ser analisada para além do campo jornalístico. Estudos orientados para esses canais difusores (aplicativos de mensagens e redes sociais), com ênfase na questão climática, poderiam averiguar se esta é uma impressão localizada, fruto da propagação midiática da expressão “fake news” ou se há uma veiculação consistente de mensagens falsas que está ocupando esse “vazio jornalístico”. Por ora, os participantes manifestaram o desejo de uma descentraliza-ção da cobertura jornalística, que é baseada no eixo Rio-São Paulo, entendendo que o enfoque regional seria um avanço importante para mobilizar as populações.A partir das subcategorias causas, efeitos e soluções, ganha destaque o olhar dos ativistas sobre as soluções (entendidas de forma crítica, com seus limites e sobreposições). A cobertura jornalística sobre clima, nacional e internacionalmente, tende a usar como gancho jornalístico muito mais os efeitos, que podem ser associados aos critérios de noticiabilidade vinculados à fac-tualidade/novidade, impacto e negatividade, sendo prevalente o enfoque sobre as consequências climáticas (LOOSE, 2020). A despeito da indicação de que há evidências de uma mudança na cobertura nacional orientada mais para o enfrentamento (RODAS; DI GIULIO, 2017) e que as soluções começam a ser debatidas após a cobertura dos efeitos, a alta dependência de agências de notícias e a falta de "[Há] essa dificuldade de acesso à informação, que faz com que muitas pessoas não consigam ter acesso a esse conteúdo jornalístico que é transmitido por diversos canais que a gente tem acesso [...] muitas delas estão se informando através do WhatsApp, de notícias através ali das fake news, que estão sendo difundidas, e a gente percebe que muitas delas acreditam, divulgam e pensam que é aquilo é realmente a verdade. (ATIVISTA DO NORDESTE)CONSIDERAÇÕES FINAIS 78JORNALISMO E ENGAJAMENTO CLIMÁTICOPESQUISA 202279profissionais especializados são impeditivos para uma real trans-formação de enfoque.Apesar disso, é preciso notar que, no contexto da pesquisa, um dos vídeos reproduzidos mencionava alternativas de ação, como redução de uso de transporte aéreo e consumo de carne, abandono de veículos à combustão, criação de redes de parques, promoção da agricultura urbana e hortas urbanas. Identificou-se, posterior-mente, que muitas dessas soluções foram compartilhadas pelos ativistas, podendo ser mais fixada pelos participantes por serem apresentadas no final do vídeo, junto com a perspectiva sobre justiça social. Ainda é importante registrar que, do ponto de vista do ativismo, faz sentido que haja priorização na ação (e não nos efeitos, como ainda ocorre na imprensa mainstream).Sob o olhar dos ativistas, as causas e os responsáveis pela crise climática quase nunca são apresentados, o que é visto como con-dição para a falta da compreensão adequada sobre a gravidade da crise ambiental e climática, a urgência na implementação de ações para mitigação e adaptação aos efeitos, em especial diante dos eventos extremos, como inundações ou secas. De fato, o jornalis-mo tem dado pouco espaço para discutir as origens da crise atual, seja porque é orientada para um presente imediato (e são poucos os formatos que permitem uma contextualização mais aprofunda-da), seja porque está diretamente ligada aos interesses políticos e econômicos que o mantém.Entre os atores que deveriam ser cobrados quanto à execução de políticas e fiscalizados quanto ao efetivo controle dos impactos, fo-ram citados os gestores públicos e os representantes das grandes empresas. Nesse âmbito, os ativistas requereram maior cobertura e exposição de outros paradigmas, para além das tecnologias, visto a atual preferência pela publicização de mudanças individuais e domésticas, cujas emissões de gases de efeito estufa são propor-cionalmente insignificantes para conter o aquecimento global.Os ativistas expressaram a vontade de serem mais visibilizadas abordagens a respeito das soluções, em especial sobre as ações e práticas locais das pessoas vulnerabilizadas, dos povos originá-rios e comunidades tradicionais, bem como das organizações da sociedade civil que há anos vêm pesquisando e experimentando tecnologias sociais além dos saberes acumulados por gerações sobre o funcionamento dos ecossistemas. Pontua-se a falta de um reconhecimento do que as populações tradicionais já vêm fazendo:Essa abordagem, segundo indicaram, pode ser feita através da apresen-tação de personagens contando a própria experiência e vivências. Acre-ditam que a exposição de histórias de pessoas comuns, das respectivas regiões do país, pode sensibilizar e motivar os públicos a se engajarem pelo clima ao proporcionar este espaço de identificação e pertencimento.A associação dos impactos socioambientais ao encarecimento do custo de vida é entendida como uma estratégia para auxiliar a compreensão, por um lado, daqueles que se beneficiam cultivando commodities em latifúndios, e por outro, daqueles que, na luta diária pela sobrevivência, igualmente não se inteiram da emergência climática nem da necessida-de de engajamento para frear o aumento da temperatura média global. O apelo econômico pode aproximar a população da discussão, mas não deve, segundo a maioria dos participantes, ser o único enquadramento para falar de mudanças climáticas."[...] deu muita visibilidade pra Greta Thunberg, enquanto isso nós temos também ativistas e militantes daqui da nossa região Amazônica que são ameaçados e tão... já têm uma longa caminhada, né? Em defesa do território. E... é puxar essa... é dar voz, né? Na verdade, dar voz não, né? Ecoar mais ainda o seu processo de defesa que já realiza há muito tempo aqui na região. (ATIVISTA DO NORTE)Onça Pintada em meio à área desmatada no Pantanal. Foto: © Carlos Eduardo Fragoso / Greenpeace VISUALIZAR AS CONSEQUÊNCIAS DA CRISE CLIMÁTICA ↓CONSIDERAÇÕES FINAIS 80JORNALISMO E ENGAJAMENTO CLIMÁTICOPESQUISA 202281Houve uma forte convergência dos ativistas sobre uma necessária “desnaturalização” de aspectos relacionados à discussão sobre jornalismo, engajamento e mudanças climáticas. Tal subcatego-ria reuniu falas que compartilham a ideia de algo que é tão repetido (e pouco ou nada questionado) que passa a ser “normalizado” pela sociedade. O jornalismo, assim como outros construtores sociais de representações, colabora com a cristalização de certas visões de mundo. Duveen (2003) frisa que não haveria comunicação sem a representação, tal é a importância da comunicação na constituição das representações sociais.Fante (2012, p. 162) analisou notícias de dois jornais principais gaúchos sobre o bioma Pampa "como um lugar para a exploração de recursos a serem fornecidos à sociedade, e não enquanto terri-tório de riquezas e bens naturais que é”, no âmbito da implantação da silvicultura industrial, através de monocultivos arbóreos de eucalipto e pinus, que ocuparam espaço significativo das tradicio-nais atividades econômico-produtivas, como a pecuária familiar, a grande responsável pela conservação desta vegetação campestre. Passados tantos anos, essa imprensa ainda se refere aos plantios arbóreos como florestas ou reflorestamento, mera representação do setor empresarial que desconhece o valor e o potencial das prá-ticas locais pampeanas, uma forma de apagamento do bioma que constitui 68% do estado.Nesta pesquisa, os ativistas denunciaram uma forma de apa-gamento através do foco exacerbado sobre a Amazônia em detrimento da cobertura a respeito da destruição da vegetação nativa em todos os biomas. A repetição de um único problema obscurece o papel da conservação dos modos de pensar e de viver nos outros territórios do país, agravando a limitada compreensão sobre as conexões entre as cinco regiões brasileiras, principalmen-te quando se fala das articulações entre clima e alteração nos regi-mes de chuvas. Tal aspecto também foi notado a partir da recente visibilidade que o clima tem recebido na imprensa: há um temor pelo esvaziamento da pauta, já que a cobertura cresceu quantitati-vamente, mas em termos qualitativos ainda deixa muito a desejar ao não contemplar os interesses dos distintos públicos provenien-tes das díspares regiões brasileiras.Como forma de despertar ecológico, os ativistas também falaram sobre a necessidade de mostrar que a economia não está acima da vida ecológica (sendo dependente dela). A constatação da falta de representatividade dos amazônidas nas notícias demonstra uma reação ao trabalho das organizações midiáticas. Mais do que a inclusão das suas pautas e dos habitantes como fontes, os ativistas do Norte buscam espaço para difundir visões de mundo e modos de viver ainda desconhecidos e invisibilizados. A ideia de proximida-de geográfica é extrapolada, pois se pretende mostrar uma forma de viver que difere da perspectiva hegemônica, como ditopor um representante de lá:A questão das (in)justiças sociais também permearam as falas dos ativistas. Revelou-se forte a ideia de que todos seremos afetados, mas uns mais do que os outros por situações de vulnerabilização social.Além disso, a desigualdade socioeconômica entre as regiões deve ser considerada também na produção das coberturas jornalísticas, de modo a buscar reverter a dificuldade de acesso ao jornalismo em relação a: 1) sinal da internet, em especial nos municípios distantes das capitais; 2) recursos para adquirir os dispositivos e custear a prestação do serviço; e 3) linguagem simples e próxima aos públicos. Não obstante o padrão noticioso em vigor nas regi-ões Sudeste e Sul tenha pretensões universalistas, os ativistas das demais regiões denunciaram esse silenciamento. Urge a prática da aproximação geográfica nas coberturas, levando em consideração as especificidades entre as regiões e as diferenças locais:""A impressão que dá, assistindo às reportagens, é que eles passaram voando e filmaram, de cima assim, né? Não desce, né? Não desce pra filmar de baixo, pra falar com as pessoas, pra saber o que está acontecendo aqui embaixo, porque tem muita árvore, tá acontecendo muito desmatamento, mas aqui embaixo tem pessoas também que estão sendo impactadas por isso. Então descer, né? Descer desse pedestal de que eu vou passar essa informação conforme eu entendo essa informação, [...] são dados muito complexos, são dados muito importantes que precisam chegar [...] em todo mundo, principalmente quem defende a floresta, quem tá nos territórios e nas comunidades nas beiras dos rios. (ATIVISTA DO NORTE)Criar conexão com a realidade da pessoa. Então um pantaneiro, um produtor rural vai estar vendo um outro produtor rural falando, o cara fala: "ah, esse ano não teve chuva aqui, minha plantação secou, os bichos tão com sede, e antes não era assim, eu moro aqui há 40 anos e não era assim", é... eu acho que é isso, colocar o lugar de falar de alguém que vai identificar.. (ATIVISTA DO CENTRO OESTE)CONSIDERAÇÕES FINAIS 82JORNALISMO E ENGAJAMENTO CLIMÁTICOPESQUISA 202283""Falta a gente ver as pessoas, a gente se identificar nessas notícias sobre mudanças climáticas. [...] e trazer identificações para gente né, eu acho que é importante trazer nível global, mas esse nível global, ele não comunica o local, e a gente precisa cada vez mais de um jornalismo local, que também é global [...] pra gente que é do norte da Amazônia se identificar com notícias sobre a gente também. (ATIVISTA DO NORTE)A gente já tá vendo pequenos locais em que estão sendo afetados pelo desmatamento da Amazônia, pelo descongelamento das geleiras, então é a gente fazer com que as pessoas percebam, que localidades com indígenas, povos quilombolas, os povos da floresta mesmo, já estão sendo afetados por essa mudança climática, eu acho que é essa mensagem que tem que passar. (ATIVISTA DO NORTE)Os ativistas reiteraram que, para o jornalismo conseguir engajar mais para o enfrentamento do clima, é preciso apresentar os atores e setores responsáveis, considerando o cumprimento das políticas públicas e/ou o desmonte delas, e as possíveis soluções e respos-tas, com foco na realidade do Sul Global:Finalmente pode-se avaliar que o público desta pesquisa (ativistas com até 35 anos que afirmam consumir notícias) reconhece a mul-tiplicidade de jornalismos existentes, algumas de suas limitações técnicas (como o tempo de uma notícia televisiva, por exemplo) e políticas (quando se referem aos interesses comerciais e vincu-lações com setores produtivos), mas endossa o papel-chave que o jornalismo hegemônico ainda exerce, sobretudo para alcançar vastos públicos. Porém, tal padronização, necessária para o alcance massivo, reduz significativamente a possibilidade de uma produção mais regionalizada, fortemente demarcada pelos diferentes GFs.De acordo o pressuposto do Jornalismo Ambiental, que trata do comprometimento do jornalismo com a mudança de pensamento (LOOSE; GIRARDI, 2017), os ativistas esperam que o jornalismo produzido sobre as mudanças climáticas avance na função de informar, sendo também didático e mobilizador, em consonância com o proposto por Moser (2010). Pode-se ainda associar tais percepções de como o jornalismo climático deveria ser às funções propostas por Bueno (2007) ao Jornalismo Ambiental: informativa, pedagógica e política. Por último, verificou-se que a proposta de Gunster (2017) de como o jornalismo pode contribuir ainda mais com o engajamento foi ratificada pelos ativistas consultados: 1) eles percebem a neces-sidade de pensar mensagens mais regionalizadas/localizadas; 2) sublinham a importância do enquadramento da justiça social; 3) avaliam que o jornalismo precisa ir além de informar, trazen-do valores culturais que possam impulsionar outras atitudes, crenças e comportamentos; e 4) buscando dar visibilidade às experiências e emoções de pessoas e comunidades que estão colaborando para enfrentar a emergência climática.VISUALIZAR AS SOLUÇÕES ↖Cacique Valto Datie com Placas Solares na Aldeia Dace Watpu Foto: © Otávio Almeida / GreenpeaceCONSIDERAÇÕES FINAIS 84JORNALISMO E ENGAJAMENTO CLIMÁTICOPESQUISA 202285RECOMENDAÇÕESDe modo a contribuir com avanços na produção jornalística climática que gere engajamento, listamos alguns aspectos no sentido de dar continuidade aos estudos nessa área no Brasil e também partilhar os resultados desse diagnóstico com outros jornalistas e comunicadores.Como já relatado, há poucos estudos voltados para a compreensão das informações jornalísticas sobre meio ambiente ou mudanças climáticas. Há alguns estudos de opinião, mas há inúmeras lacunas em termos de pesquisas qualitativas que respondam a questões complexas como a que esse estudo se propõe.Diante dos achados desta investigação, sugerem-se algumas pro-postas para outros estudos: ǿ Aprofundar estudos, por região, com amostras representati-vas, que tragam dados mais consistentes sobre os tipos de in-formações climáticas consumidas (considerando o reforço pela regionalização de conteúdos observados nos grupos focais, há demanda para esse tipo de conteúdo? Em qual formato?); ǿ Estudar a hibridação dos formatos e linguagens (até que ponto o jornalismo pode transmutar- se sem perder a credibilidade, decorrente do jornalismo centrado na informação?); ǿ Trabalhar com públicos diversos, de modo a verificar se as estratégias de comunicação apontadas pelos ativistas, de fato, alcançarão um público não engajado; ǿ Comparar as nuances sobre o funcionamento das assessorias em organizações sem fins lucrativos e naquelas com fins lucra-tivos, além de sistematizar as reais diferenças entre os produ-tos construídos a partir da lógica dos meios de comunicação. ǿ Material de apoio (manual ou guia) endereçado às redações dos meios hegemônicos, com atualizações desde a perspec-tiva dos respondentes desta pesquisa a partir da regionaliza-ção da cobertura climática; ǿ Incentivo às parcerias ou estruturação de redes para cir-culação de narrativas transmídias locais em outros meios, considerando a pluralidade de vozes. Dezenas de povos indígenas já têm jornalistas competentes para participar de um projeto de amplo alcance que dê visibilidade às suas visões de mundo e formas de viver. Essa proposta pode-ria se expandir para os territórios mais invisibilizados do Brasil, buscando o fomento do debate climático a partir da realidade local; ǿ Desenvolvimento de formações para jornalistas que atuam em veículos regionais/locais, de modo a discutir a inserção da perspectiva climática em diferentes pautas; ǿ Realizar eventos regionalizados sobre soluções e ações pró-clima, com caráter crítico, de modo a promover a pauta em contextos específicos, contribuir com a popularização de alternativas não hegemônicas; ǿ Desenvolver uma campanha institucional entre organizações e veículos jornalísticos de modo a tornar mais explícito o chamado para o engajamento climático.1. PESQUISAS FUTURAS2. AÇÕES PARA MELHORAR O JORNALISMO CLIMÁTICO86JORNALISMO E ENGAJAMENTO CLIMÁTICOPESQUISA 202287VISUALIZAR AS CONSEQUÊNCIAS DAS MUDANÇAS CLIMÁTICAS E AS PESSOAS IMPACTADAS ↑ Famílias Impactadas por Enchentes no Brasil em Minas Gerais. Foto: © Isis Medeiros / GreenpeacePESQUISA 20228988JORNALISMO E ENGAJAMENTO CLIMÁTICOREFERÊNCIAS BIBLIOGRÁFICASAKAS. 2022. Disponível em: <https://portal.comunique-se.com.br/jornalismo-esta-fazendo-para-ajudar-a-resolver-a-crise-do-clima> Acesso em: 17.08.2022.ANDRADE, Francisca Marli Rodrigues de; BARRETO, Tarssio Brito; HENRIQUES, Alen Batista. Rio de Janeiro and climate crisis: gover-nance, interactivity and discursive construction on Twitter. Ambien-te & Sociedade, v. 23, 2020.APPELGREN, Ester; JÖNSSON, Anna Maria. Engaging Citizens for Climate Change: Challenges for Journalism Digital Journalism, v. 9, n. 6, p. 755-772, 2021. doi: 10.1080/21670811.2020.1827965.BALBÉ, Alice; LOOSE, Eloisa Beling. Jornalismo, medo e alterações climáticas: articulações possíveis para pensar o enfrentamento dos riscos climáticos. Observatorio (OBS*), v. 14, n. 2, p. 38-55, 2020.BARDIN, Laurence. Análise de conteúdo. Lisboa: Edições 70, 1979.BOLSEN, Toby; SHAPIRO, Matthew A. The US News Media, Polariza-tion on Climate Change, and Pathways to Effective Communication. Environmental Communication, v. 12, n. 2, p. 149-163, 2017. doi: 10.1080/17524032.2017.1397039.BUENO, Wilson. Comunicação, jornalismo e meio ambiente: teoria e pesquisa. São Paulo: Majoara, 2007.CARVALHO, Anabela; VAN WASSEL, Margit; MAESEELE, Pieter Maeseele. Communication practices and political engagement with climate change: a research agenda, Environmental Communica-tion, 2016. doi: 10.1080/17524032.2016.1241815.CAVALCANTI, Giselli. O papel das ONGs ambientalistas na comuni-cação das mudanças climáticas: Um caminho para o engajamento? Dissertação de Mestrado. Programa de Pós- Graduação em Psicolo-gia da Universidade Federal do Rio Grande do Norte. Natal-RN, 2019.CHARRON, Jean; BONVILLE, Jean de. Natureza e transformação do jornalismo. Florianópolis: Insular; Brasília: FAC, 2016.DEJAVITE, Fábia A. INFOtenimento: informação + entretenimento no jornalismo. São Paulo: Paulinas/Sepac, 2006.DUVEEN, Gerard. Introdução – O poder das ideias. In: MOSCOVI-CI, Serge. Representações Sociais: investigações em psicologia social. Editado em inglês por Gerard Duveen; traduzido do inglês por Pedrinho Guareschi. Petrópolis, RJ: Vozes, 2003. p.07-28.FANTE, Eliege M. As representações sociais sobre o Bioma Pampa no jornalismo de referência sul-rio-grandense. Dissertação (Mes-trado em Comunicação e Informação). Porto Alegre, Universidade Federal do Rio Grande do Sul, 2012.FRANCO, Maria Laura P. B. Análise de Conteúdo. Líber Livro, 2008.GALLI, G. O jornalismo comunitário, a democracia e as identidades individuais e coletivas. Revista Alterjor, 23(1), p. 99-124, 2021.GIRARDI, Ilza; LOOSE, Eloisa Beling; SILVA, Jamille Almeida. O Jornalismo Ambiental na concepção de quem o faz: estudo com jornalistas da América Latina, Caribe, Portugal, Espanha e países africanos de Língua Portuguesa. Aturá Revista Pan-Amazônica de Comunicação, v. 2, n. 2, p. 48-66, 2018.GIRARDI, Ilza M. Tourinho; MASSIERER, Carine.; LOOSE, Eloisa Beling.; SCHWAAB, Reges. Caminhos e descaminhos do Jorna-lismo Ambiental. Comunicação e Sociedade, v. 34, n. 1, p.131-152, 2012.GLOBAL FUTURE. A Crisis in Common: How Eco-Anxiety is Shared Across our Country. 2021. Disponível em: <https://ourglobalfuture.com/wp-content/uploads/2021/10/A-Crisis-in-Common-1.pdf>. Acesso em: 19.08.2022.REFERÊNCIAS BIBLIOGRÁFICAS90JORNALISMO E ENGAJAMENTO CLIMÁTICOPESQUISA 202291GUNSTER, Shane. Engaging climate communication: Audiences, frames, values and norms. In: HACKETT, Robert; FORDE, Susan; GUNSTER, Shane; FOXWELL-NORTON, Kerrie. Journalism and Climate Crisis: Public Engagement, Media Alternatives. London: Routledge, 2017. p. 49-76.HULME, Mike. Why we disagree on climate change: Understanding controversy, inaction and opportunity. Cambridge: Cambridge Uni-versity Press, 2009.ITS-RIO; YALE UNIVERSITY. Mudanças climáticas na percepção dos brasileiros: relatório de análise. Rio de Janeiro, 2022. Disponí-vel em: <https://itsrio.org/wpcontent/uploads/2021/02/Apresenta%-C3%A7%C3%A3o-IBOPE_FINAL.pptx.pdf>. Acesso em: 12.05.2022.LAWS, Norman; CHOJNICKA, Joanna. “A Future to Believe in”: In-troducing Varieties of Advocacy Journalism. The Examples Sustai-nability and the Sanders Campaign. Journalism Studies, v.21, n. 9, p. 1261–1283, 2020. doi: 10.1080/1461670X.2020.1742773.LOOSE, Eloisa Beling. Jornalismo e mudanças climáticas desde o Sul: os vínculos do jornalismo não hegemônico com a colonialidade. Tese (Programa de Pós-Graduação em Comunicação). Universidade Federal do Rio Grande do Sul. Porto Alegre: UFRGS, 2021.LOOSE, Eloisa Beling. Jornalismo e riscos climáticos: percepções e entendimentos de jornalistas, fontes e leitores. Curitiba: Editora UFPR, 2020.LOOSE, Eloisa Beling; GIRARDI, Ilza Maria Tourinho . Interfaces entre o debate colonial e os estudos de jornalismo ambiental. Desenvolvimento e Meio Ambiente, v. 58, p. 319-333, 2021.LOOSE, Eloisa Beling; GIRARDI, Ilza Maria Tourinho. O Jornalismo Ambiental sob a ótica dos riscos climáticos. Interin, Curitiba, v. 22, n. 2, p. 154-172, 2017.MAPBIOMAS ALERTA. Monitor da Fiscalização do Desmatamento. Disponível em: <https://plataforma.alerta.mapbiomas.org/monitor--da-fiscalizacao>. Acesso em 04.05.2022MIGUEL, Katarini Giroldo. Narrativas socioambientais: afeto e experi-mentações jornalísticas na estética transmídia. Animus - Revista Interamericana de Comunicação Midiática, v. 18, n. 37, 2019. doi:10.5902/2175497734329.MOSER, Susanne C. Communicating climate change: history, chal-lenges, processes and future directions. WIREs Climate Change, 1, 31–53, 2010.NEELIMA, B. N.; REDDY, R. Uttama. Mass Media and Climate Change. International Journal of Research (IJR). v. 1, n. 2, p. 1-14, 2014.NEVEU, Érik. Sociologia do jornalismo. São Paulo: Edições Loyola, 2006.PERUZZO, Cicilia M Krohling. Revisitando os Conceitos de Comunicação Popular, Alternativa e Comunitária. In: XXIX Congresso Brasileiro de Ciências da Comunicação - Intercom. Brasília: UnB, 2006.RADAR CLIMÁTICO. Crisis alimentaria en el mundo: Conoce las causas y posibles soluciones. 2022. Disponível em: <https://conexioncop.com/radarclimatico>. Acesso em: 17.08.2022.RODAS, Caroline; DI GIULIO, Gabriela. Mídia brasileira e mudanças climáticas: uma análise sobre tendências da cobertura jornalística, abordagens e critérios de noticiabilidade. Desenvolvimento e Meio Ambiente, v. 40, p. 101-124, 2017. doi: 10.5380/dma.v40i0.49.S&P GLOBAL. Weather Warning: Assessing Countries’ Vulnerability To Economic Losses From Physical Climate Risks. 2022. Disponível em: < https://www.spglobal.com/_assets/documents/ratings/resear-ch/101529900.pdf>. Acesso em 04.05.2022SCHÄFER, Mike; PAINTER, James. Climate journalism in a changing media ecosystem: Assessing the production of climate change‐rela-ted news around the world. Wiley Interdisciplinary Reviews: Climate Change, v. 12, n. 1, p. 1-20, 2021.SCHÄFER, M; SCHLICHTING, I. Media representations of climate change: A meta-analysis of the research field. Environmental Commu-nication, 8(2), 142-160, 2014. doi: 10.1080/17524032.2014.9140 SEEG. Análise das emissões brasileiras de gases de efeito estufa e suas implicações para as metas climáticas do Brasil 1970-2020. 2021. Disponível em: <https://seeg-br.s3.amazonaws.com/Documen-tos%20Analiticos/SEEG_9/OC_03_relatorio_2021_FINAL.pdf>. Acesso em 04.05.2022.STODDART, Mark C. J.; SMITH, Jillian. The endangered Arctic, the Arctic as resource frontier: News media narratives of climate change and the North. Canadian Review of Sociology, vol. 53, n. 3, 2016. p. 316–336.APÊNDICE92JORNALISMO E ENGAJAMENTO CLIMÁTICOPESQUISA 202293RELAÇÃO DE PARTICIPANTES DOS GRUPOS FOCAISAté 19 anosDe 25 a 29 anosDe 25 a 29 anosDe 35 a 39 anosDe 30 a 34 anosDe 25 a 29 anosDe 25 a 29 anosDe 30 a 34 anosDe 25 a 29 anosDe 20 a 24 anosDe 25 a 29 anosDe 20 a 24 anosDe 25 a 29 anosDe 25 a 29 anosDe 25 a 29 anosDe 25 a 29 anosDe 25 a 29 anosDe 25 a 29 anosDe 30 a 34 anosDe 30 a 34 anosDe 25 a 29 anosDe 20 a 24 anosDe 30 a 34 anosDe 25 a 29 anosDe 20 a 24 anosDe 35 a 39 anosDe 30 a 34 anosParanáRio Grande do SulRio Grande do SulRio Grande do SulRio Grande do SulParanáParanáMinas GeraisRio de JaneiroMato Grosso do SulMinas GeraisSão PauloSão PauloMato GrossoParanáRio Grande do SulRio Grande do SulSanta CatarinaSanta CatarinaSão PauloRio de JaneiroRio de JaneiroRio de JaneiroMato Grosso do SulDistrito FederalMinas GeraisEspírito SantoBalsa NovaPorto AlegrePorto AlegreSão LeopoldoCapão da CanoaPonta GrossaCuritibaBelo HorizonteRio de JaneiroCampo GrandeSão Carlos*Porto Alegre*São PauloPrimavera do LestePontal do ParanáViamãoPorto AlegreFlorianópolisAtalantaSuzanoRio de JaneiroRio de JaneiroCachoeiras De MacacuCampo GrandeBrasíliaSão Carlos*VitóriaEnergia, Jovens, Democracia, OMG Arayara, Fridays for FutureMovimentos a favor da soberania alimentar, LGBT's, movimentos de esquerda.Eco Pelo Clima, Fridays for Future Brasil, 350.org, veganismo, promotora da agenda dos objetivos da ONUe do Acordo de Paris.Justiças socioambiental, econômica e de gênero.Movimento socioambientalistaEducação para a sustentabilidade/ feminismoMigração e Refúgio / Direitos HumanosAtivista pela Alimentação no Slow Food BrasilUnião da Juventude Comunista, Associado ao Sindicato Nacional ASSIBGE e Ativismo Ambiental.Feminismo; já atuou em Coletivo de Mulheres do Jornalismo, participou de mobilizações online, pesquisa sobre a temática no mestradoMovimento NegroEstudantilEducação e Meio AmbienteArticulado da AssociaçãoComunidade acadêmica, práticas extensionistas e apoio à movimentos dos povos à beira-marMovimento Negro e ComunistaAtua na Amada Massa, um coletivo de reparação social que tem uma padaria de fermentação natural vegana em Porto Alegre; e já me envolvi em movimentações de resistência a projetos que afetavam a causa ambientalista (exemplo da mina Guaíba e o PL 260, que liberava no RS agrotóxicos proibidos nos países de origem). Também apoio o movimento de comprar direto de quem faz e localmente.VeganismoAmbiental (na produção de mudas nativas) - atua através na preservação, restauração e conservação do meio ambiente através da ONG na qual trabalha.Crise Climática / Veganismo / Consumo Consciente / FeminismoMovimentos de Favelas e Causa AmbientalYouth Climate Leaders, Academia de Ação Climática e Climate ScienceCausas SocioambientaisGênero e Meio AmbienteUnião Vegana FeministaColetivos de Negros do Instituto de Química da USP São CarlosAgroecologia, Justiça Ambiental, Populações Tradicionais, Reforma Agrária010507091102121418262022162408100604031317192325272115FAIXA ETÁRIAFAIXA ETÁRIAESTADOESTADOCIDADECIDADECAUSAS EM QUE É ENGAJADOCAUSAS EM QUE É ENGAJADO* Cidade onde reside atualmenteAPÊNDICE94JORNALISMO E ENGAJAMENTO CLIMÁTICOPESQUISA 202295De 25 a 29 anosDe 25 a 29 anosDe 30 a 34 anosDe 30 a 34 anosDe 30 a 34 anosDe 25 a 29 anosDe 30 a 34 anosDe 20 a 24 anosDe 30 a 34 anosDe 25 a 29 anosDe 25 a 29 anosDe 30 a 34 anosAté 19 anosDe 20 a 24 anosDe 30 a 34 anosDe 20 a 24 anosDe 30 a 34 anosDe 25 a 29 anosDe 25 a 29 anosDe 30 a 34 anosDe 35 a 39 anosDe 30 a 34 anosDe 25 a 29 anosDe 25 a 29 anosDe 25 a 29 anosDe 30 a 34 anosDe 25 a 29 anosDe 25 a 29 anosDe 20 a 24 anosDe 20 a 24 anosDe 20 a 24 anosAté 19 anosAté 19 anosMato GrossoBahiaAcreAmapáAmazonasParáSão PauloGoiásPernambucoParáParáRondôniaParáParáRoraimaAcreBahiaAmazonasTocantinsDistrito FederalMato Grosso do SulBahiaMaranhãoCearáCearáMaranhãoDistrito FederalCearáParaíbaBahiaBahiaBahiaMato Grosso do SulBarra do GarçasFeira de SantanaRio BrancoMacapáTeféBelémSão PauloValparaisoRecifeSantarémBelémCacoalBelémSantarémBoa VistaRio BrancoNão InformadaManausGoianorte TocantinsBrasíliaCampo GrandeSalvadorPaço do LumiarFortalezaFortalezaSão LuísBrasíliaPentecosteBayeuxSalvadorFeira de SantanaSalvadorCampo GrandeMovimento Bem ViverAmazônia e Mudanças ClimáticasComitê Chico MendesPovo TradicionaisCáritas BrasileiraRede Jandyras / Youth Climate Leaders / Advocacy Climático / Justiça Racial, de Gênero, Climática, Ambiental e Social - Periferias e Comunidades tradicionais do território amazônico.Ambiental (defesa do Pantanal)Pessoas em situação de ruaLevante Popular da Juventude e Movimento Brasil PopularMovimento Pela Soberania Popular na Mineração - MAMPela cultura, comunicação e Juventudes em defesa da AmazôniaMeio ambiente, soberania alimentar,terra para todos CPTMov. Negro, Feminista, Ecossocialista, LGBTQIA+, estudantil e pela democratização da mídia.Movimento Pela Soberania Popular na Mineração - MAMCausas indígenas, juventude indígena eLevante popular da juventudeCáritas brasileira Serviço Pastoral dos MigrantesHortas Coletivas, Direito a Cidade,Puxirum do Bem ViverPastoral da JuventudeAtivismo Climático e Direitos Humanos.AmbientalMovimento NegroCuidado aos IdososPopulação em Situação de Rua e Luta por MoradiaNa sua participação dogrupo focal disse trabalhar com pesquisas voltadas para meioambientePedal das Minas São Luís Coletivo que pauta a Mobilidade Urbana Ativa por Bicicleta.LGBTQIA+Comunidade LGBTQIAP+ e Literatura RegionalLuta por Moradia e Justiça AmbientalFeminismo e RacismoEngajamundo, Clima de Eleição e Escola de AtivismoNa sua participação do grupo focal disse em uma ONG, desenvolvendo projetos com jovens agricultores e agricultoras rurais do Nordeste e outras partes do Brasil.Pantanal no Fridays For Future Brasil284652545648302945495759555058604751533133353741433932363842404434FAIXA ETÁRIAFAIXA ETÁRIAESTADOESTADOCIDADECIDADECAUSAS EM QUE É ENGAJADOCAUSAS EM QUE É ENGAJADOJORNALISMO E ENGAJAMENTO CLIMÁTICOPara perguntas sobre o relatório ou outras discussões, questões editoriais ou comerciais, entre em contato conosco:PARA QUESTÕES OU DISCUSSÕES SOBRE A PESQUISAPARA QUESTÕES EDITORIAIS OU COMERCIAISPARA ASSESSORIA DE IMPRENSAMarina ColeratoEDITORA marina@modefica.com.brEloisa Beling LooseCOORDENAÇÃO DE PEQUISA eloisa.beling@gmail.comcontato@modefica.com.brassessoria@modefica.com.brMODEFICA.COM.BRView publication stats | RESEARCH GATE |
1 ppe51 variants promote non-replicating Mycobacterium tuberculosis to grow at acidic pH by 2 selectively promoting glycerol uptake 3 Shelby J. Dechow1, Jacob J. Baker1, Megan R. Murto1, and Robert B. Abramovitch1* 4 5 1Department of Microbiology and Molecular Genetics, Michigan State University, East Lansing, 6 MI 48824. 7 8 9 10 11 12 13 14 15 *Corresponding Author: 16 Robert Abramovitch 17 E-mail: abramov5@msu.edu 18 Phone: (517) 884-5416 19 Fax: (517) 353-8957 20 21 22 23 .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint Abstract 24 In defined media supplemented with single carbon sources, Mycobacterium tuberculosis (Mtb) 25 exhibits carbon source specific growth restriction. When supplied glycerol as the sole carbon 26 source at pH 5.7, Mtb establishes a metabolically active state of nonreplicating persistence known 27 as acid growth arrest. We hypothesized that acidic growth arrest on glycerol is not a metabolic 28 restriction, but rather an adaptive response. To test this hypothesis, we conducted forward genetic 29 screens that identified several Mtb mutants that could grow under these restrictive conditions. All 30 of the mutants were mapped to the ppe51 gene and resulted in three amino acid substitution – 31 S211R, E215K, and A228D. Expression of the PPE51 variants in Mtb promoted growth at acidic 32 pH showing that the mutant alleles are sufficient to cause the dominant gain-of-function, enhanced 33 acid growth (eag) phenotype. Testing growth on other single carbon sources showed the PPE51 34 variants specifically enhanced growth on glycerol, suggesting ppe51 plays a role in glycerol 35 uptake. Using radiolabeled glycerol, enhanced glycerol uptake was observed in Mtb expressing 36 the PPE51 (S211R) variant, with glycerol overaccumulation in triacylglycerol. Notably, the eag 37 phenotype is deleterious for growth in macrophages, where the mutants have selectively faster 38 replication and reduced in virulence in activated macrophages as compared to resting 39 macrophages. Recombinant PPE51 protein exhibited differential thermostability in the WT or 40 S211R variants in the presence of glycerol, supporting the eag substitutions alter PPE51-glycerol 41 interactions. Together, these findings support that PPE51 variants selectively promote glycerol 42 uptake and that slowed growth at acidic pH is an important adaptive mechanism required for 43 macrophage pathogenesis. 44 45 .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint Introduction 46 During infection, Mycobacterium tuberculosis (Mtb) senses and adapts to a variety of 47 immune cues including hypoxia1,2, nutrient starvation3,4, pH changes5, and nitrosative and 48 oxidative stress6. Exposure to these stresses can promote Mtb to establish slowed growth or a non-49 replicating persistent (NRP) state. NRP bacteria are tolerant to immune and antibiotic-mediated 50 killing7-9, therefore understanding mechanisms underlying NRP may promote new methods to 51 shorten the course of TB therapy. 52 Following macrophage infection, Mtb senses the mildly acidic pH of the phagosome and 53 broadly remodels its gene expression10. Adaptations to acidic pH include the induction of the 54 PhoPR regulon, induction of ESX-1 secretion, and remodeling of central metabolism and cell 55 envelope lipids11. Defects in adaptations to acidic pH reduce Mtb virulence in macrophages and 56 animals12-15, therefore, pH dependent adaptations are required for Mtb virulence. 57 Previous studies conducted by our lab sought to understand the interplay of acidic pH and 58 Mtb central metabolism. We observed that Mtb exhibits selectivity of the carbon sources on which 59 it can growth at pH 5.7 relative to pH 7.0. For example, Mtb incubated at acidic pH with glycerol 60 as a sole carbon source is restricted for growth and establishes a viable, metabolically active state 61 of NRP called acid growth arrest16-18. Acid growth arrest is observed on a variety of other carbon 62 sources associated with glycolysis and TCA cycle. Interestingly, Mtb can resuscitate its growth at 63 acidic pH by addition of specific carbon sources, such as pyruvate, acetate, oxaloacetate [OA] and 64 cholesterol, which function at the intersection of glycolysis and the TCA cycle (a.k.a. the 65 anaplerotic node)18. This discovery suggests that the anaplerotic node is the location of a pH-66 dependent metabolic switch that may promote Mtb growth on permissive carbon sources during 67 pathogenesis at acidic pH, and that metabolic remodeling is required for pH adaptation11. 68 .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint It is puzzling that Mtb cannot grow at acidic pH on specific carbon sources, as Mtb is 69 provided with oxygen as a terminal electron acceptor and a carbon source that is well utilized at 70 pH 7.0. Thus, acid growth arrest is different from other NRP models, where the bacterium is 71 missing a key factor required for growth (e.g. oxygen or nutrients in the hypoxia or starvation 72 models of NRP). Therefore, we hypothesized that acid growth arrest is not an actual restriction on 73 growth, but an adaption by the bacterium to slow and arrest its growth. In a previous study, our 74 lab previously sought to identify genes regulating acid growth arrest by selecting for mutants 75 incapable of arresting their growth on minimal medium agar plates, buffered to pH 5.7 with 76 glycerol as the carbon source. From our screening efforts, novel missense mutations were 77 identified in ppe51 (H37Rv annotated Mtb gene, Rv3136) and were named enhanced acid growth 78 (eag) mutants 16. PPE51 is part of the PE and PPE mycobacterial protein family. Named for their 79 unique N-terminus motifs Pro-Glu (PE) and Pro-Pro-Glu (PPE), most of these proteins have 80 remained largely enigmatic in their functional roles19,20. However, a growing body of literature in 81 recent years has assigned diverse putative functional roles for PE and PPE proteins including 82 immune evasion20-24, calcium binding25, iron utilization26,27, Mg2+ and PO32− transport28, 83 fibronectin binding29, and lipase activity30,31. 84 Mounting evidence suggests that some PE and PPE proteins may play important roles in 85 Mtb nutrient acquisition. Examination of pe and ppe evolution reveals an expansion of this protein 86 family corresponding with Type VII or ESX secretion systems, where it is thought that ancestral 87 pe and ppe genes inserted into an esx gene cluster and expanded alongside this secretion system 88 during subsequent gene duplication events 22,32,33. Secretion via ESX provides a route for PE and 89 PPE proteins to access the cell surface and nutrients in the host cell milieu, which is supported by 90 high-throughput proteomic evidence showing direct surface localization of PE and PPE 91 .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint proteins26,28,34-36. Mtb contains five ESX secretion systems37, with ESX-5 contributing to the 92 majority of PE and PPE secretion32,34,38. Furthermore, ESX-5 and its cognate PE and PPE proteins 93 have been implicated in the uptake of fatty acids and possibly the utilization of other nutrient 94 substrates34. ESX-3-mediated PE and PPE proteins are thought to play a role in iron acquisition, 95 whereby they have been shown to be directly involved in mycobactin-mediated iron uptake and 96 heme uptake26,39,40. Taken together, these results provide direct examples of Mtb acquiring and 97 utilizing host resources through secretion of PE and PPE proteins. 98 Based on these findings showing a role for PPE proteins in transport and that PPE51 eag 99 variants could grow on glycerol, we hypothesized in our 2018 study16, “that these amino acid 100 substitutions may increase Mtb growth by modulating mycomembrane permeability, possibly 101 by modulating the channel size or specificity of PPE51, which may function as a porin to enhance 102 access to glycerol or other nutrients at acidic pH.” The goal of this study was to test this hypothesis 103 and further define the role of PPE51 in glycerol acquisition and pathogenesis. Notably, concurrent 104 with this study, recently published studies confirmed the hypothesis that PPE51 is an exported cell 105 surface-associated protein and linked to the nutrient acquisition of glycolytic carbon sources 28,41,42. 106 Here, we show that in a saturating forward genetic screen only three PPE51 variants, S211R, 107 E215K and A228D were isolated as eag mutants. The PPE51 variants specifically promoted 108 growth at pH 5.7 on glycerol and not other tested carbon sources, supporting the substitutions 109 selectively promote glycerol utilization. Radiolabeling studies show that the S211R variant has 110 enhanced uptake of glycerol and accumulation of triacylglycerol (TAG), showing that the variants 111 promote glycerol uptake. Differential thermal stability of WT versus S211R variant PPE51 112 proteins in the presences of glycerol, support the variant has direct and differential interactions 113 with glycerol, but similar interactions with the non-permissive substrate glucose. Structural 114 .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint modeling supports that PPE51 forms a structure homologous with bacterial nutrient transporters, 115 with the variants altering the predicted ligand specificity. These data are consistent with a model 116 where PPE51 promotes uptake of glycerol across the mycomembrane by acting like a porin and 117 that variants alter the conformation to enhance glycerol uptake. eag variants exhibit enhanced 118 replication and reduced virulence in activated macrophages, supporting a role for pH-dependent 119 slowed growth during macrophage pathogenesis. 120 121 Methods 122 Bacterial Strains and Growth Conditions 123 All experiments were performed with M. tuberculosis strains Erdman and CDC1551. Mtb was 124 grown at 37°C and 5% CO2 in vented T-25 culture flasks containing Middlebrook 7H9 media with 125 10% oleic acid-albumin-dextrose-catalase (OADC), 0.05% Tween-80, and 0.2% glycerol. For acid 126 stress and single carbon source experiments, MMAT defined minimal media was used as described 127 by Lee et al. 43: 1 g/L KH2PO4, 2.5 g/L Na2PO4, 0.5 g/L (NH4)2SO4, 0.17 g/L L-Asparagine 128 monohydrate, 10 mg/L MgSO4, 50 mg/L ferric ammonium citrate, 0.1 mg/L ZnSO4, 0.5 mg/L 129 CaCl2, and 0.05% Tyloxapol. MMAT media was buffered with 100 mM 3-(N-130 morpholino)propanesulfonic acid (MOPS) for experiments requiring pH 6.6-7.0 and 100 mM 2-131 (N-morpholino)ethanesulfonic acid (MES) for experiments requiring pH 5.5-6.5 44. For growth 132 curve experiments, Mtb was grown to mid-late log phase (OD600 0.6-1.0) and seeded in MMAT at 133 a starting OD600 of 0.05. Optical density measurements were conducted by removing 500 µL of 134 samples at each time point. Viability assays were performed in a similar manner with samples 135 being diluted 10-fold in PBS + 0.05% Tween-80 and plated for viable colony forming units (CFUs) 136 on 7H10 + 10% OADC agar plates. 137 .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint Genetic Screen and Sequencing 138 A wild type Erdman Mtb population of 4x109 bacteria was plated on MMAT agar plates (1 g/L 139 KH2PO4, 2.5 g/L Na2PO4, 0.5 g/L (NH4)2SO4, 0.17 g/L L-Asparagine monohydrate, 10 mg/L 140 MgSO4, 50 mg/L ferric ammonium citrate, 0.1 mg/L ZnSO4, 0.5 mg/L CaCl2, and 0.05% 141 Tyloxapol) supplemented with 10 mM glycerol as the sole carbon source and buffered to pH 5.7 142 with 100 mM MES 44. Plates were incubated at 37°C with spontaneous mutants appearing around 143 week eight and isolated for growth. Single-colony isolates were confirmed as enhanced acid 144 growth (eag) mutants under acidic conditions in liquid MMAT (pH 5.7) media amended with 10 145 mM glycerol. Whole genome sequencing (WGS) was performed on genomic DNA isolated from 146 mutants representing various levels of enhanced growth as well as a wild type Erdman control. 147 Samples were sequenced using the Illumina MiSeq in a 2x250-bp paired end format. Base calling 148 was done by Illumina Real Time Analysis v1.18.54, demultiplexed, and converted to FastQ using 149 Illumina Bcl2fastq v2.19.1. Low-quality bases were trimmed and adapter sequences were removed 150 using Trimmomatic (v0.36)45 and aligned sequence reads to the Erdman reference genome using 151 BWA-MEM46. SNPs and indels were identified using Genome Analysis ToolKit (GATK)47. 152 153 Generation and Analysis of ppe51 Knockout 154 The ppe51 gene was replaced with a hygromycin resistance (HygR) cassette in both Erdman and 155 CDC1551 Mtb strain backgrounds using a new chromosomal engineering system called ORBIT ( 156 for “oligonucleotide-mediated recombineering followed by Bxb1 integrase targeting”) that 157 combines site-specific recombination with homologous recombination48. An ORBIT 158 recombineering plasmid (pKM444) expressing RecT annealase and Bxb1 integrase from the 159 anhydrotetracycline (ATc)-inducible Ptet promoter and containing a kanamycin resistance (KanR) 160 .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint cassette was first transformed into Mtb, selected for KanR, induced with ATc, and generated into 161 electrocompetent cells. Electrocompetent cells were then transformed with a knockout integration 162 plasmid (pKM464) harboring HygR and targeting oligonucleotide. Hygromycin-resistant colonies 163 were isolated and cured of the kanamycin-containing recombineering plasmid. Genomic DNA was 164 extracted from transformants and the 5’ and 3’ junction sites of the knockout were confirmed by 165 PCR and sequencing using ORBIT target-specific and ppe51-specific primers (Figure S7B-C, 166 Table SX). Gene replacement was further verified via quantitative real time PCR (qPCR) (Figure 167 S7D). Δppe51 was complemented with WT and variant ppe51 from their native promoter and 168 confirmed by qPCR (Figure S7E). 169 170 pH-and-Glycerol dose response combination growth assays 171 Mtb cultures were incubated in a range of pH buffered MMAT media (pH 5.0-pH 7.0) at a starting 172 OD600 of 0.2 in 96-well plates. Cultures were treated with 2.5-fold dose-response (0.13-80 mM) 173 of glycerol and incubated over the course of 21 days, with growth assessed by optical density. 174 Bacterial viability was assessed by diluting wells 10-fold and plating for viable CFUs on 175 7H10 + 10% OADC agar plates. Optical density data was converted to percent of maximum well-176 growth and normalized based on no carbon control at pH 5.5 (0%) and maximum Mtb growth on 177 glycerol at pH 6.5 (100%). Each condition and time-point experiment was conducted in triplicate 178 and representative of multiple individual experiments. 179 180 Radiolabeled Glycerol Uptake Assay 181 Mtb Erdman cultures were pre-adapted for 3 days in MMAT media (pH 5.7 or pH 7.0) containing 182 10 mM glycerol. Following adaptation, Mtb was washed twice with PBS+0.05% Tween-80 and 183 .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint resuspended in the same buffered MMAT media amended with 10 mM glycerol and 6 μCi of [U-184 14C] Glycerol. Samples were removed over the course of 24 hours, fixed with 4% 185 paraformaldehyde, and assessed for total radioactivity using scintillation counting. Short-term 186 radiolabeling uptake experiments were also conducted in a similar manner as previously 187 described28,49 to compare previously published results. Mtb strains were pre-adapted for 3 days in 188 MMAT media buffered to pH 5.7 with 10 mM glycerol, harvested via centrifugation, and washed 189 twice with uptake buffer (50 mM Tris, pH 7.0, 15 mM KCl, 10 mM (NH4)2SO4 1mM MgSO4, 190 0.02% tyloxapol). Samples were resuspended in uptake buffer at a starting OD600 of 2.0 and 191 incubated for 30 minutes at 37°C in T-25 tissue culture flasks. Cultures were amended with 16 192 µM [U-14C] Glycerol, and 200 µL samples were taken at minutes 3, 6, 9, 18, and 36. Samples were 193 immediately filtered by centrifugation through a 0.22 µM Spin-X tube (Costar) and washed twice 194 with 4% paraformaldehyde. Total radioactivity of the filters was assessed by liquid scintillation 195 counting. All strains used for radiolabel uptake experiments were repeated in two biological 196 replicates. 197 198 Analysis of metabolism of radiolabeled lipids into Mtb lipids 199 Mtb Erdman cultures were pre-adapted as described above for the uptake experiments. Following 200 pre-adaptation, cultures were seeded at a starting OD600 of 0.2 in MMAT media (pH 5.7 or pH 7.0) 201 + 10 mM glycerol and set up in two biological replicates. Lipids were labeled with 6 μCi of [U-202 14C] Glycerol for 6 days, and samples were pelleted and washed with PBS before lipid extraction. 203 Total lipids were extracted and Folch washed as previously described18 and 14C-incorporation was 204 measured using scintillation counting. For thin layer chromatography (TLC), 5,000 counts per 205 minute (CPM) of each pH 5.7 sample and 10,000 CPM of each pH 7.0 sample was loaded on a 206 .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint 100-cm2 high-performance TLC silica gel 60 aluminum sheet (EMD Millipore) and analyzed with 207 a chloroform:methanol:water (90:10:1 v/v/v) solvent system50. Sulfolipids, TAG, and PDIM were 208 separated as previously described13,18,50 and quantified using a phosphor screen and Typhon imager 209 and ImageJ software51. 210 211 Replication during acid growth arrest 212 For measurement of replication during acid growth arrest, WT Mtb, Δppe51, and ppe51 native 213 variants in both CDC1551 and Erdman backgrounds carrying the pBP10 plasmid52 were inoculated 214 into MMAT media (pH 5.7 and pH 7.0) + 10 mM glycerol and in the absence of kanamycin. 215 Plasmid loss and percentage of bacteria still containing the pBP10 was determined by plating for 216 CFUs on 7H10 + 10% OADC agar plates ±25 µg/uL kanamycin. Rates of growth, death, and 217 cumulative bacterial burden were quantified using equations as previously described53. The Mtb 218 segregation constant (s = 0.18± 0.023) was used for all calculations. 219 220 Macrophage pathogenesis studies 221 Bone Marrow-derived macrophages (BMDMs) were extracted and infected with the panel of 222 complemented strains built into the CDC1551 Δppe51 knockout background at a multiplicity of 223 infection (MOI) of 1:1 using previously described methods54. BMDMs were activated by treating 224 with 100 units/mL IFN-γ overnight, followed by treatment with 10 ng/mL lipopolysaccharide 225 overnight. Infected BMDMs were lysed at days 0, 3, 6, and 9 and intracellular bacterial lysates 226 were serially diluted and enumerated on 7H10 + 10% OADC agar plates. Each strain for each 227 timepoint was performed in triplicate. BMDMs were also infected with CDC1551 strains 228 containing the pBP10 replication clock plasmid as described in the pBP10 in vitro experiments 229 .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint using the same macrophage infection methods described above with minor modifications. 230 BMDMs infected with pBP10-containing strains were lysed at days 0, 2, 4, 6, and 8 and 231 enumerated on 7H10 + 10% OADC agar plates ±25 µg/uL kanamycin selection. Calculations for 232 pBP10 plasmid loss and replication dynamics were performed as described in the in vitro pBP10 233 experiments. 234 235 Recombinant PPE51 Protein Expression and Purification. The ORF of PPE51 was amplified 236 using pET23::ppe51_FWD and pET23::ppe51_REV primers (Table S1) and cloned into the 237 pET23a+ vector containing a C-terminal polyhistidine (His)-tag. The cloned protein has a deletion 238 of the final four C-terminal amino acids. Transformants propagated in E. coli BL21(DE3) were 239 selected on LB agar plates containing Amp. The S211R mutation was introduced into the 240 pET::ppe51-WT construct using the site-directed mutagenesis primers PPE51-S211R_FWD and 241 PPE51-S211R_REV (Table S1) and the QuikChangeTM Site-Directed Mutagenesis Kit. Overnight 242 cultures were expanded into 250 mL of fresh LB media with Amp at an initial inoculum OD600 of 243 0.05 and grown to an OD600 of 0.6 at 37 °C with shaking at 200 rpm. Proteins were then induced 244 with 1 mM of isopropyl-ß-D-thiogalactopyranoside (IPTG) at 18°C for 20 hours. Culture was then 245 harvested via centrifugation at 4000 rpm for 25 minutes at 4°C. Pellets were then lysed for 30 246 minutes on ice with occasional vortexing in ice-cold lysis buffer (50 mM phosphate buffer [pH 247 7.6], 200 mM NaCl, 0.1% Triton-x-100, 0.1 mg/mL PMSF, 0.5 mg/mL lysozyme). Because 248 PPE51 possibly interacts with glycerol, glycerol was completely removed from all buffers used 249 during the purification process. Cells were further lysed by sonication and cell lysate was clarified 250 via centrifugation at 14,000 rpm for 30 minutes at 4°C. Supernatant was loaded onto a nickel ion-251 containing affinity resin column and bound overnight with shaking at 4°C. Protein was washed 252 .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint first with wash buffer containing no imidazole and a second time with wash buffer containing 50 253 mM imidazole. PPE51 protein was then eluted into (4) 1 mL fractions with elution buffer 254 containing 200 mM imidazole. Recombinant PPE51 was quantified using the Qubit assay. 255 256 PPE51 Protein Thermostability Assay 257 The thermostability assay was performed as previously described55 with 13.5 µL of 0.635 mg/mL 258 of batch-purified PPE51 samples aliquoted into PCR tube containing 1.5 µL of 100 mM glycerol, 259 yielding a final glycerol concentration of 10 mM. Samples were incubated for 20 minutes at room 260 temperature and transferred to PCR thermocyclers where they were incubated for an additional 5 261 minutes at the following temperatures: 35, 40, 45, 50, 55, 60, and 65°C. Samples were then 262 centrifuged at 4000 rpm for 10 minutes to pellet precipitated protein. After centrifugation, soluble 263 protein was removed from the tubes and detected in Western blots using mouse anti-His tag 264 monoclonal antibody followed by HRP-conjugated anti-mouse IgG secondary antibody. Enhanced 265 chemiluminescence (ECL) Western blotting substrate (Pierce) was used for Western blot 266 detection. The AI600 Chemiluminescent Imager was used to visualize and analyze Western blot 267 results. 268 269 Results 270 All isolated eag mutants have spontaneous mutations in ppe51 . During acid growth arrest in 271 minimal media, Mtb is provided all required nutrients including a metabolically utilized carbon 272 source and a terminal electron acceptor. This suggests that acid growth arrest is not due to a 273 physiological limitation presented by the growth environment but rather is a regulated process 274 whereby Mtb adapts to its acidic environment. A previously published forward genetic screen 275 .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint tested this hypothesis using a CDC1551 transposon mutant library containing >100,000 was plated 276 on MMAT defined minimal media agar with glycerol at pH 5.7 and resulted in 98 transposon (Tn) 277 mutants and two spontaneous WT mutants16. These mutants were isolated and confirmed as 278 enhanced acid growth (eag) mutants based on their ability to grow well compared to native WT 279 Mtb at pH 5.7 in liquid MMAT supplemented with glycerol 16. Interestingly, complementation 280 attempts with the Tn mutants did not restore growth arrest, and whole genome sequencing 281 identified spontaneous mutations in ppe51 in both Tn and WT mutant backgrounds 16. To repeat a 282 saturating screen, in the absence of transposon mutagenesis, a second forward genetic screen was 283 performed on MMAT agar buffered to pH 5.7 with glycerol, using a larger bacterial population 284 (4x109 bacteria) in the Erdman Mtb strain (Figure 1A). From the WT Erdman screen, 98 285 spontaneous eag mutants were isolated of which 52 were colony-purified and confirmed for 286 enhanced growth under acidic conditions in liquid MMAT containing glycerol (Figure 1B, Figure 287 S1A-D). The eag isolates exhibited an up to ~4-fold increase in growth compared to WT Erdman 288 which exhibited complete growth arrest (Figure 1B). Of these eag mutants, 24 were selected for 289 whole genome sequencing. Remarkably, all 24 isolates had single nucleotide polymorphisms 290 (SNPs) mapping to the ppe51 gene (Table 1). All mutations were non-synonymous (S211R, 291 A228D, and E215K) and were centrally located within a 50 bp region on the ppe51 gene (Figure 292 S2). The S211R and A228D variants were also identified in the prior Tn mutant CDC1551 screen, 293 with E215K being a novel mutation found in the new Erdman screen. 294 295 ppe51 mutations are sufficient to overcome growth arrest. Given that ppe51 variants exhibit 296 enhanced growth at acidic pH, we investigated the function of the variant alleles in the presence 297 of the WT ppe51 allele. Overexpression constructs of WT or mutant ppe51 were transformed into 298 .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint WT CDC1551 or WT Erdman Mtb strains carrying the native ppe51 allele. Overexpression strains 299 were grown in MMAT at pH 5.7 with glycerol as a carbon source. Overexpression of ppe51-S211R 300 and ppe51-A228D variants in WT Mtb resulted in significantly enhanced growth under acidic 301 conditions (Figure 2A). In contrast, overexpression of WT ppe51 and the empty vector exhibited 302 complete growth arrest at pH 5.7. The E215K allele was also not sufficient at overcoming growth 303 arrest which may explain why it has only been observed once across two independent forward 304 genetic screens. Additionally, all overexpression strains grew equally well at pH 7.0 (Figure S3), 305 showing that the observed growth phenotype is pH-specific. Interestingly, although the growth 306 phenotype with A228D overexpression resulted in enhanced growth at acidic pH, it grew at a 307 slower rate compared to S211R overexpression strains in both CDC1551 and Erdman 308 backgrounds. We examined individual variant alleles from the screen (Figure 1B) and observed 309 that S211R variants significantly grouped together at a higher rate of growth compared to A228D 310 and E215K (Figure 2B and 2C). Together, these results demonstrate that the eag mutations confer 311 a dominant, gain-of-function growth phenotype, and specific mutations are associated with 312 differential strength of the phenotype. 313 314 PPE51 variants selectively promote growth on glycerol. Based on the enhanced growth 315 phenotype that the variants exhibit at acidic pH, we hypothesized that this phenotype may be due 316 to ppe51 variants modulating mycomembrane permeability, resulting in enhanced nutrient uptake. 317 To test this hypothesis, we conducted an Ethidium Bromide (EtBr) assay looking at permeability 318 with WT Mtb overexpression constructs (empty vector, WT, and S211R) in both CDC1551 and 319 Erdman backgrounds. With the EtBr assay we did not observe differences in the rate of uptake 320 between the WT and eag overexpression strains in either CDC1551 or Erdman (Figure S4), 321 .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint suggesting that the growth phenotype is not due to a general increase in permeability. We then 322 hypothesized that the growth phenotype may be due to nutrient-specific uptake. We explored this 323 possibility by growing CDC1551 and Erdman overexpression strains (empty vector, ppe51 and 324 ppe51-S211R) in liquid minimal media (pH 5.7) in the presence of various growth-permissive (e.g. 325 pyruvate, acetate, cholesterol) and non-permissive (e.g. glucose, glycerol, propionate) carbon 326 sources18. After 20 days we found that enhanced acid growth was only observed with ppe51-S211R 327 in the presence of 10 mM glycerol, a normally non-permissive carbon source at pH 5.7, in both 328 CDC1551 and Erdman (Figure 2D, Figure S5, Figure S6A-B). As expected, all overexpression 329 strains exhibited enhanced growth on permissive carbon sources in both Mtb backgrounds. 330 Notably, the ppe51 variants specifically promote growth on glycerol, but not glucose, another 331 proposed nutrient associated with PPE51-dependent uptake28,41, demonstrating the variants are 332 selective for glycerol. 333 334 ppe51 is not required for survival during acid growth arrest. Transcriptional profiling studies 335 previously conducted show that ppe51 is significantly induced during acid growth arrest18.We 336 hypothesized that ppe51 may be required for Mtb to promote survival when exposed to acid growth 337 arresting conditions. To test this hypothesis, a ∆ppe51 knockout strain was generated in both 338 CDC1551 and Erdman Mtb using the mycobacteria-specific ORBIT system (Figure S7A)48. 339 Successful knockout of ppe51 was confirmed by sequencing the oriE and HygC junction sites of 340 ∆ppe51, PCR amplification of the entire knockout region, and RT-PCR (Figure S7B-D). 341 Complementation constructs containing the native ppe51 promoter were introduced into ∆ppe51 342 carrying WT ppe51, variant ppe51 (S211R, A228D, E215K) and a double variant 343 (S211R+A228D). An empty complementation vector was also introduced into ∆ppe51. The 344 .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint presence of these constructs was also confirmed via RT-PCR (Figure S7E). Growth curves of the 345 complementation constructs grown in growth arresting conditions showed that WT and empty 346 vector strains exhibit growth arrest at pH 5.7, whereas the variant complemented strains exhibited 347 enhanced growth in both CDC and Erdman (Figure 3A, Figure S8A). Additionally, the S211R, 348 A228D, and S211R+A228D variant complemented strains grow slightly better compared to 349 E215K in both CDC1551 and Erdman Mtb strain knockout backgrounds, which aligns with 350 previous overexpression growth curve and relative growth data (Figure 2A and B). At pH 7.0, all 351 WT and complemented strains showed similar levels of growth (Figure S8B and C). To examine 352 the role of ppe51 in survival, a viability assay was performed with the previously described panel 353 of strains. The empty vector and WT-complemented ∆ppe51 maintain viability over the course of 354 40 days at pH 5.7 (Figure 3B, Figure S9A). The complemented ppe51 variants also maintain 355 viability and replicate at pH 5.7. At pH 7.0, all WT and complemented strains maintain viability 356 and exhibit similar increases in CFUs over the course of 18 days (Figure S9B and C). 357 The stable viability of WT or ∆ppe51 mutant may be due to growth arrest or, alternatively, 358 balanced growth and death. To determine if the strains are truly growth arrested we examined 359 replication using a the pBP10 clock plasmid, transformed into WT Mtb, ∆ppe51, and eag variants 360 in both CDC1551 and Erdman backgrounds53. The strains were incubated in minimal media (pH 361 5.7 and 7.0) with glycerol for 40 adays. We observed that the WT and ∆ppe51 strains do not 362 replicate under acid growth arrest conditions in both strain backgrounds (Figure 3C, Figure S10B). 363 In contrast, we are able to observe high rates of replication in the eag variants at pH 5.7 (Figure 364 3C, Figure S10B). We then compared eag variants’ calculated cumulative bacterial burden (CBB) 365 to total CFUs counted on nonselective plates and observed that greater rates of replication in the 366 eag variants is associated with a high death rate, yielding a large difference between CBB and total 367 .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint CFUs (Figure 3D, Figure S10D-F). When in vitro pBP10 growth studies were conducted at pH 7.0 368 with all strains, we observed similar high rates of replication and plasmid loss across all strains 369 (Figure S10A and C). Interestingly, these results show that enhanced growth at acidic pH is driven 370 by higher replication, but this growth is offset somewhat by a higher death rate, supporting the 371 conclusion that faster replication at acidic pH is deleterious to Mtb survival. 372 373 Acidic pH limits glycerol uptake and PPE51 variants overcome this restriction. Pyruvate can 374 rescue growth on glycerol in a concentration-dependent manner at pH 5.7 18. However, it is 375 unknown whether glycerol concentration affects acid growth arrest. We hypothesized that acid 376 growth arrest may be driven by glycerol starvation and the PPE51 variants promote growth by 377 promoting enhanced uptake of glycerol. If this is the case, we would expect to see a dependence 378 of glycerol concentration and acidic pH on growth. To examine this, we examined checker-board 379 dose responses combining varying pH levels (pH 6.5-5.5) and glycerol concentrations (80 mM-380 0.13 mM) using the panel of isogenic strains. The standard concentration of glycerol used in our 381 acid growth arrest model is 10 mM. Growth in the wells was analyzed using optical density (OD600) 382 and data was normalized to wells containing the highest (100%) levels of growth and wells with 383 no carbon representing the lowest (0%) levels of growth. Growth assays were performed for 21 384 days, and the data shown is Day 14 which is representative for the duration of the experiment. 385 Interestingly, we found that growth arrest appears to be both pH and glycerol concentration-386 dependent, with growth partially rescued at high concentrations of glycerol (~80 mM) for WT, 387 ∆ppe51::pMV306 and ∆ppe51::pMV-ppe51 at pH 5.7 (Figure 4 and S11A). Additionally, we 388 observed higher levels of growth at lower glycerol concentrations (~0.82 mM) at pH 5.7 with the 389 complemented ppe51 variants compared to the empty vector and WT ppe51 complemented strains. 390 .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint Growth could also be rescued with high glycerol concentration (~32 mM) for variants at pH 5.5. 391 Interestingly, the presence of the double ppe51 variant (S211R+A228D) overcomes growth arrest 392 at pH 5.5 at even lower glycerol concentrations (~5.12 mM) compared to the single variants, 393 indicating that the presence of two eag point mutations confers a slight growth advantage during 394 acid growth arrest. Concentrations of glycerol below 0.33 mM do not rescue growth starting at pH 395 6.0 in any eag strains, which could be due to glycerol being fully consumed. Similarly, these 396 observations were also made in the native eag variants in both CDC1551 and Erdman, while WT 397 exhibited a reduced capacity for glycerol uptake (Figure S11B). Together, these findings suggest 398 that Mtb has reduced capacity to uptake glycerol in a pH-dependent manner, and that PPE51 399 variants function by promoting enhanced uptake of glycerol. 400 Based on these checkerboard results, ppe51 appears to restrict its growth on glycerol at 401 acidic pH. Additionally, WT Mtb has been shown to completely arrest its growth at pH 5.7 on 10 402 mM glycerol; however, it is able to maintain viability for up to 40 days, remains metabolically 403 active, and incorporate limited amounts of exogenous 14C-glycerol into lipids16. To further test the 404 hypothesis that Mtb restricts glycerol uptake at acidic pH and that eag variants promote enhanced 405 glycerol uptake, a radiolabeling experiment using 14C-glycerol was conducted with WT Erdman 406 and the ∆ppe51 complemented strains previously described. Strains were pre-adapted for three 407 days in MMAT (pH 5.7 or 7.0) with 10 mM glycerol and washed with PBS prior to radiolabeling 408 with 6 µCi of 14C-glycerol. Samples were collected over the course of 24 hours, washed, and 409 analyzed for radiolabel uptake by scintillation counting. All complemented strains containing a 410 ppe51 variant accumulated 14C-glycerol at a similarly increased rate of approximately 300% 411 compared to the WT Mtb strain (Figure 5A). These results are consistent with radiolabeling that 412 was conducted with the pVV16 overexpression empty vector, S211R overexpression strain, and 413 .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint native eag-S211R variant where we observed similar enhanced glycerol uptake at ~ 60% with 414 strains containing S211R compared to WT overexpression empty vector (Figure S12A). We also 415 looked at glycerol uptake with WT CDC1551, empty vector, and complemented S211R at pH 7.0. 416 We did not observe significant differences in glycerol uptake between strains, and the rate of 417 uptake was similar to the complemented ppe51 variant strains at pH 5.7 (Figure S12B). Together, 418 these results show that Mtb does restrict glycerol uptake at pH 5.7 regardless of whether ppe51 is 419 functionally intact. In contrast, strains containing ppe51 variants have significantly enhanced 420 glycerol uptake. 421 While the radiolabeling strongly indicated that glycerol was being taken up by the strains, 422 it did not answer whether glycerol was being metabolized by Mtb and incorporated into lipids or 423 binding to the mycomembrane without uptake across the plasma membrane. To address this 424 question, we performed lipid radiolabeling with 14C- glycerol. WT Erdman and ∆ppe51 425 complemented strains were pre-adapted for three days in the same culture conditions as the 426 previously described radiolabeled uptake experiment. The operon controlling sulfolipid synthesis 427 is induced in a phoPR-dependent and a pH-dependent manner18,50,56. Sulfolipid was observed to 428 specifically accumulate at pH 5.7 (Figure 5B) with no accumulation occurring at pH 7.0 (Figure 429 S12C). Triacylglycerol (TAG) has been shown to accumulate during periods of hypoxic and pH-430 stress18,57, and pathways involved in TAG synthesis play a role in reducing Mtb growth by 431 redirecting carbon flux away from the TCA cycle58. Interestingly, we found that TAG accumulated 432 specifically in the complemented S211R strain at pH 5.7 (Figure 5C). In contrast, we did see 433 similar TAG accumulation across all strains at pH 7.0 (Figure S12D). The observation of labeled 434 lipids in both growth arrested and growing Mtb at acidic pH, shows that glycerol is imported and 435 metabolized at acidic pH, with enhanced uptake in the S211R variant. 436 .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint PPE51 variants have selectively reduced growth in activated macrophages. ppe51 is induced in 437 a pH-dependent and phoP-dependent manner within 2 hours following phagocytosis by 438 macrophages10, suggesting that ppe51 is important for pathogenesis. We hypothesized that ppe51 439 or its eag variants may be required for pathogenesis, specifically in activated macrophages, where 440 the phagosome is acidified. To test this hypothesis, resting and activated primary murine bone 441 marrow-derived macrophages (BMDMs) were infected with WT CDC1551 and ∆ppe51 mutant 442 and complemented variant strains. In resting macrophages, we did not observe significant 443 differences in Mtb growth between the strains (Figure 6A), with all of the strains growing ~1.25-444 log over nine days. In contrast, in activated macrophages, while the WT, ∆ppe51 empty vector, 445 and ∆ppe51 WT complemented strain still exhibit ~1.25-log increase in growth, the ppe51 446 complemented variants show significantly lower growth (Figure 6B). These results show that 447 ppe51 variants are selectively attenuated for virulence in activated macrophage environment, 448 which is consistent with a pH-dependent phenotype that is observed in vitro. 449 Rohde et al. showed that rapid replication of intracellular Mtb is associated with greater 450 Mtb killing by the macrophage10. We observed in vitro that variants had enhanced death during 451 replication at acidic pH, and we hypothesized that the eag variants may be replicating faster than 452 the WT in macrophages but have lower CFUs due to enhanced death rates. To test this hypothesis, 453 we infected BMDMs with native CDC1551 WT, Δppe51, and A228D variant containing the 454 pBP10 plasmid as described previously. Infection was conducted over the course of 8 days with 455 cells lysed and plated for viable CFUs every 2 days. We observed an initial ~0.5 log decrease in 456 viable CFUs in both WT and Δppe51 around day 2 that is consistent with observations made by 457 Rohde et. al.59, and supports their findings that Mtb exhibits delayed adaptation to survive and 458 replicate within macrophages (Figure 6C). Both WT and Δppe51 then replicated over the course 459 .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint of 8 days inside activated BMDMs as evident by their ~1 log increase in CFUs starting at day 2. 460 In contrast, the A228D variant lacks this initial adaptation period and instead show a continual ~1 461 log decrease in CFUs over the course of 8 days. Calculating the CBB of the A228D variant shows 462 a large difference between the CBB and CFUs, demonstrating that the A228D variant is replicating 463 at a higher rate and dying at an even greater rate. These strains are able to replicate and survive 464 better in resting BMDMs compared to activated BMDMs (Figure S13). We conclude that slowed 465 growth in response to acidic pH inside activated macrophages is necessary for mycobacterial 466 survival and that the eag variants do not be sufficiently slow their growth inside macrophages, 467 resulting in enhanced killing. These results also support that the PPE51 variant is promoting uptake 468 of a carbon source during macrophage infection, suggesting that Mtb may metabolize glycerol 469 when growing in macrophages. 470 471 Differential thermal stability of PPE51 and the S211R variant proteins support direct 472 interactions between PPE51 and glycerol. We hypothesized that the eag variants promote PPE51 473 uptake of glycerol by altering PPE51 structure and its affinity for glycerol. Changes in the thermal 474 stability of the protein would provide evidence supporting this hypothesis. C-terminal his-tagged 475 recombinant PPE51 and PPE51 (S211R) variant proteins were expressed and purified from E. coli 476 (Figure S14A and B, Table S1). Glycerol was omitted from the reagents used in the purification 477 process and loading dye. In the absence of glycerol, we observed differential stability between the 478 WT and S211R PPE51 variants, with the WT and S211R proteins completely denaturing at 60 ºC 479 and 50 ºC, respectively, supportive of a significant structural change by the amino acid substitution 480 (Figure 7A and B). In glycerol, the WT protein exhibited enhanced stability, completely 481 denaturing at 65 ºC, a shift of 5 ºC, and the S211R protein completely denaturing at 60 ºC, a shift 482 .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint of 10 ºC. These findings support that glycerol/PPE51 interactions, with differential stability shifts 483 dependent on the S211R substitution. 484 We previously noted that PPE51 S211R did not promote the growth on glucose and therefore, 485 examined the thermal stability in glucose. We observed reduced stability of the WT protein in 486 glucose, completely denaturing at 55 ºC and did not observe any differences in stability with the 487 S211R protein in glucose, supporting the stability shifts are selectively dependent on glycerol 488 (Figure 7A and B). Together, these data show that glycerol selectively increases the thermal 489 stability of PPE51, with enhanced impact on the S211R variant, lending further support for a 490 mechanism whereby PPE51 directly binds glycerol for uptake and acquisition into the Mtb cell. 491 Based on the eag phenotype and differences in thermal stability, we hypothesized that these 492 substitutions may have a significant impact on protein structure and thus conducted in silico 493 modeling of PPE51 using the Iterative Threading ASSEmbly Refinement (I-TASSER) server for 494 protein structure and function prediction60. The best fit model of PPE51-WT had a moderately 495 high confidence score (C-score) of -0.86 on a scale of -5 (low confidence) to 2 (high confidence) 496 (Figure 7C). Interestingly, threading of the sequence against known transporter structures 497 produced a porin-like model with a possible channel. This model was matched to all structures in 498 the Protein Data Bank (PDB) library. The top 10 proteins from the PDB with the closest structural 499 similarity were all predicted to be nutrient transporter proteins (Figure S14C) An overlay of the 500 S211R variant (blue) with PPE51-WT model shows that the introduction of this substitution 501 confers a noticeable conformation shift in the predicted protein structure ( Figure 7C). For PPE51-502 WT, the top two predicted ligands were maltose and a monoacylglycerol derivative (78N) with 503 one predicted ligand binding site for maltose being within the 18 amino acid residue range of the 504 eag variants, located at residue 225 (Figure S14D). Additionally, the predicted top Gene Ontology 505 .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint terms for the molecular function, biological process, and cellular component are hexose:hydrogen 506 symporter activity (GO:0009679), transmembrane transport (GO:0055085), and integral to 507 membrane (GO:0016021), respectively (Figure S14E). While these in silico results are predictions, 508 they provide further support for a role with PPE51 acting as a nutrient transporter for Mtb. 509 Furthermore, all eag mutations mapped to a single alpha helix on the predicted I-TASSER model, 510 with S211R and E215K located at the top of the predicted channel and A228D located within the 511 center channel structure (Figure S14F) and the substitutions, altered the modeled substrate 512 interaction, further supporting our model for PPE51 variants acting to promote uptake of glycerol 513 by altering the protein structure and ligand interactions. 514 515 PDIM biosynthesis is disrupted in the ppe51 deletion strains. Surprisingly, we found that the 516 Δppe51 mutant generated in this study does not have the same growth or glycerol uptake compared 517 to previous studies that have generated similar knockouts of knockdowns of ppe5128,41,61. We also 518 found that the Δppe51 grew just as well as other strains at pH 7.0 on glycerol (Figures 3B and 519 S8A). This observation was previously made by Wang et. al., who showed that mutations in 520 phthiocerol dimycocerosates (PDIM) biosynthesis were responsible for permeabilizing the 521 mycomembrane and compensating for the loss of functional ppe5128. We sequenced the genomes 522 of Δppe51 mutants in both the CDC1551 and Erdman backgrounds and found that both Δppe51 523 mutants had evolved mutations in PDIM biosynthesis pathway genes (ppsC in Δppe51-CDC1551, 524 and ppsD in Δppe51-Erdman). We confirmed for loss of functional PDIM by radiolabeling it with 525 14C-glycerol and 14C-acetate for six days and extracting total lipids for TLC analysis. As expected, 526 we observed loss of functional PDIM accumulation in the Δppe51::pMV-EV compared to WT Mtb 527 radiolabeled at both pH 5.7 and pH 7.0 with 14C-acetate and 14C-glycerol, respectively (Figures 528 .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint 8B and S13). However, despite the occurrence of these PDIM mutations in the Δppe51 mutants, 529 no PDIM mutations were present in the sequenced eag variant mutants used in this study or WT 530 Mtb, highlighting that the eag variants remain gain-of-function, dominant mutants (Figure 2A), 531 that selectively promote uptake of glycerol (Figure 2D) and promote enhanced uptake of glycerol 532 and enhanced replication in vitro and macrophages (Figures 5A, 3C, 6C, S11B, S12A, S13). 533 Conclusions for the PPE51 knockout or eag variants in the ppe51 knockout background, must take 534 into account that these strains are also PDIM mutants. Overall, we did not observe any differences 535 in the eag mutants if they were in a WT or ppe51/PDIM mutant background, supporting that these 536 gain-of-function eag phenotypes are independent of PDIM levels. 537 538 Discussion 539 Mtb exhibits complex regulatory and physiological adaptations when grown in acidic 540 environments, including changes in growth rate. The underlying basis of slowed growth in mildly 541 acidic environments is still not fully resolved, but appears to be associated with metabolic and 542 redox stress, that may be linked to balancing cytoplasmic pH-homeostasis and respiration11. 543 Providing specific carbon sources, such as pyruvate or acetyl-CoA, relieve this metabolic stress 544 and enable Mtb to grow similarly well at acidic and neutral pH16,18. However, it has been puzzling 545 as to why Mtb cannot grow on glycerol at acidic pH, as it has a carbon source and oxygen, 546 everything it needs to grow. In this study, we found that Mtb limits uptake of glycerol at acidic pH 547 to restrict its growth and that mutations in ppe51 promote uptake of glycerol at acidic pH and 548 enable growth. That is, Mtb can grow well at acidic pH on glycerol, but has adapted instead to stop 549 growth. 550 .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint We further show that this pH-dependent metabolic adaptation is required for pathogenesis. 551 Selectively in activated macrophages, where the pH of the phagosome is more acidic, we observed 552 a virulence defect in strains expressing the eag variants. Notably, using a replication clock plasmid, 553 we found that eag variants have enhanced growth in macrophages, but even greater killing, the 554 balance of which results in reduced fitness. Thus, slowed growth in macrophages, in an activation 555 dependent manner is dependent on the restriction of metabolism at acidic pH, and that PPE51 556 variants overcome this restriction to the detriment of the pathogen. This finding supports that the 557 nutrient imported by the PPE51 variant is relevant to the macrophage environment. We showed 558 that the variants specifically promote uptake of glycerol, therefore, it is plausible that glycerol is a 559 key regulator of Mtb growth in the macrophage. It has been previously shown that Mtb can uptake 560 TAG in macrophages62, TAG is abundant in granulomas63, and Mtb exports the TAG lipase 561 LipY64, therefore, it is possible glycerol is released from TAG during infection, and restriction of 562 glycerol uptake plays an important role in slowing growth during infection. Studies examining the 563 interactions of PPE51 eag variants, LipY and glycerol metabolism genes during pathogenesis will 564 be undertaken to test this hypothetical model. 565 It is a striking finding that all of the eag mutants selected were in ppe51 and that they all 566 clustered with a highly conserved region of 18 amino acid residues (residues 211-228). Three 567 single amino acid substitutions (S211R, A228D, and E215K) greatly altered WT ppe51 function 568 and promoted growth under acid stress when given the non-permissible carbon source, glycerol. 569 S211R was able to confer the greatest enhanced growth, whereas A228D conferred moderate 570 enhanced growth and E215K exhibited the least amount of enhanced growth, comparatively 571 (Figure 2B). The growth phenotypes of the native mutant alleles were further recapitulated in 572 overexpression studies in a WT Mtb background as well as a Δppe51 background, where again we 573 .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint observed overall greater eag with the S211R variant compared to A228D and E215K (Figures 2A, 574 3A, S8A). Given that the phenotype was conserved in PDIM containing strains (the initially 575 isolated mutants and the overexpressors in the WT) and PDIM lacking strains (the ppe51 deletion 576 mutants), this demonstrates that the gain-of-function phenotype is independent of PDIM. This 577 region of PPE51 may play a key role in protein-substrate interactions, and indeed with recombinant 578 proteins, we observed differential stability in the variant protein and its interaction with glycerol. 579 Interestingly, the structural modeling showed substitutions in this region altered the predicted 580 ligand of the modeled transporter, supporting further study of this critical region for modulating 581 PPE51-ligand interactions. 582 Another key finding of this study is that glycerol uptake is restricted at acidic pH. Data 583 supporting this conclusion include the reduced uptake of radiolabeled glycerol at acidic pH as 584 compared to neutral pH (Figures 5A and S12AB), the dependence of glycerol concentration and 585 pH in regulating growth (Figures 4, S11A and S11B), and the ability of PPE51 variants to enhance 586 growth and glycerol uptake at acidic pH (Figures 5A and S12AB). How Mtb restricts glycerol 587 uptake is still not known, but it is puzzling that PPE51 is strongly induced at acidic pH and counter 588 to a model where PPE51 promotes in glycerol uptake, but Mtb restricts glycerol uptake at acidic 589 pH. This contradiction remains unresolved and points to a new known unknown of Mtb 590 metabolism restriction at acidic pH. Notably, growth on glycerol-containing mixtures can exceed 591 growth compared to growth on glycerol alone65, suggesting that Mtb may need to restrict glycerol 592 to regulate its growth while consuming other carbon sources it encounters during infection. 593 We identified that the eag variants selectively enabled growth on glycerol alone compared 594 to WT Mtb (Figure 2D). The identification of this carbon specificity with PPE51 eag variants 595 implies a putative role for PPE proteins in nutrient acquisition, a model that is strongly supported 596 .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint by data put forth by Ates et. al., Mitra et. al. and Wang et al. These studies showed that PE and 597 PPE proteins located at the cell envelope and cell surface play a vital role in nutrient uptake for 598 Mtb. Ates et. al. provides strong evidence that the type VII secretion system, ESX-5, is essential 599 for mycobacterial growth and nutrient uptake. In this study, essentiality of ESX-5 could be rescued 600 by altering cell wall lipid composition or introducing the M. smegmatis outer membrane porin, 601 mspA, which mediates cell wall permeability and influx of hydrophilic nutrients66-68. ESX-5 602 mutations in M. marinum result in significantly reduced growth on medium with Tween-40 or 603 Tween-80 as the sole carbon source, and the ESX-5 mutant strain exhibits significantly impaired 604 uptake of fluorescently-labeled fatty acids compared to WT and complemented strains34. These 605 data support ESX-5 facilitating the uptake of fatty acids to be used as a carbon source through the 606 secretion of PE and PPE proteins. In support of Ates’ ESX-5 substrate nutrient influx hypothesis, 607 Mitra and colleagues26 showed direct evidence tying PE and PPE proteins to iron acquisition. Mitra 608 identified Mtb transposon mutants that were resistant to a toxic heme analog26. The mutants were 609 in three previously uncharacterized genes of which two were PPE proteins, PPE36 and PPE6226. 610 Furthermore PPE62 was shown to be surface-accessible and predicted structure indicates that it 611 may form a β-barrel that resembles Haemophilus influenzae heme cell surface receptor26,69 and 612 that heme transport is facilitated into the cell by the periplasmic lipoprotein DppA70. Finally, the 613 Wang et al. study provided direct evidence that PPE51 is exported to the mycomembrane to 614 promote uptake of glycerol and glucose, possibly by acting like a porin. Notably, loss of function 615 ppe51 mutants have altered sensitivity to antibiotics, including pyrazinamide71 and meropenem72, 616 suggesting that PPE51 mediated impacts on carbon source uptake or mycomembrane permeability 617 play a role in drug susceptibility, supporting further studies of PPE51 as a target for potentiating 618 antibiotics. 619 .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint Here, we present a model that integrates the current understanding of PE and PPE nutrient 620 acquisition with our findings (Figure 9), wherein PPE51 embeds itself into the outermost layer of 621 the cell envelope and is surface accessible to glycerol28,42,73. Gene expression profiling data 622 supports induction of ppe51 by phoP and acidic pH15,18. Phylogenetic evidence shows that PPE51 623 is duplicated alongside ESX-532, which has been shown to mediate the secretion of most PE/PPE 624 proteins in M. marinum, including PPE5134,38. We propose that an unknown periplasmic nutrient 625 transporter helps mediate the import of glycerol across the plasma membrane and into the cell from 626 initial import by PPE51. pe/ppe families have high variation rates between Mycobacterium 627 tuberculosis complex (MTBC) genomes with ppe51 being the single exception in showing almost 628 no variation74. However, under the selective pressure of our screen (Figure 1A), we have shown 629 that we can select for mutations that enhance PPE51’s proposed uptake of glycerol (Figure 9). 630 Furthermore, our initial in silico modeling of PPE51 suggests that it can form a porin-like structure 631 consistent with a role in transport and ligand-binding sites for carbon nutrient sources (Figure 8C). 632 Based on these data, we further propose an eag mutant model, whereby the eag amino acid 633 substitutions introduce conformational changes that allow for a possible PPE51-porin structure to 634 widen or enhance the binding the glycerol, allowing enhanced transport through the 635 mycomembrane. 636 This study has focused on the role of the eag PPE51 variants, and the not the Dppe51 mutant, 637 due to confounding mutations in PDIM in the deletion strains. It is interesting that both deletion 638 mutants (in Erdman and CDC1551) evolved these mutations during the construction of the mutants 639 and suggests there may have been a selective advantage for the mutations. Indeed, Wang et al., 640 showed that ppe51 knockouts only had a glycerol uptake phenotype when the PDIM was restored 641 in the mutant. This finding is consistent with our observation that the Dppe51 mutants in this study 642 .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint did not have a growth defect in glycerol, presumably due to the lack of PDIM, whereas the ppe51 643 mutants in the Wang et al., study were defective for growth. Given the conservation of the eag 644 mutants in strains with or without PDIM, we conclude that PDIM level do not appreciably impact 645 the enhanced uptake of glycerol in eag variants. However, it is also possible that differences for 646 the PPE51 mutants between this study and the others may be driven by genomic differences. Both 647 Wang et. al and Korycka-Machała et. al. used the H37Rv Mtb strain for their knockout and 648 CRISPRi knockdown studies, respectively. However, sequence analysis of the region directly 649 upstream of ppe51 in both CDC1551 and Erdman compared to H37Rv shows an almost total 650 deletion of the ppe50 gene preceding ppe51. The ppe50 sequence is also not present anywhere else 651 in the CDC1551 or Erdman genome except for a matching 66 bp sequence that precedes ppe51 in 652 both genomes. The large sequence difference in the ppe51 promoter region between strains could 653 imply an additional reason why we see strong phenotypic growth differences between our 654 respective growth screens of ppe51 knockouts. 655 656 Acknowledgements 657 We thank members of the Abramovitch lab for critical reading of the manuscript and Prof. David 658 Sherman for sharing the pBP10 clock plasmid. This research was supported by a grant from the 659 NIH-NIAID (R01AI116605) and AgBioResearch. 660 661 Conflicts of Interest 662 R.B.A. is the founder and owner of Tarn Biosciences, Inc., a company that is working to develop 663 new TB drugs. 664 665 .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint Author Contributions 666 S.J.D., J.J.B., and R.B.A. conceived the project. S.J.D performed all of the experimental studies. 667 M.M. conducted thermal stability assay studies. S.J.D. and R.B.A. wrote the manuscript. All 668 authors reviewed the manuscript. 669 670 Figure and Table Legends 671 Figure 1. Genetic screen for mutants unable to arrest growth at acidic pH. 672 A) Schematic of the forward genetic screen used to acquire eag mutants. The appearance of distinct colony 673 growth after 8 weeks was indicative of mutants that were unable to arrest their growth at acidic pH. 674 B) Growth phenotypes of isolated mutants were determined by measuring Day 9 OD600 and compared to 675 OD600 from the initial inoculum. Each dot represents mutants that were isolated from MMAT agar plates 676 and confirmed for enhanced growth in liquid MMAT (pH 5.7). Pink-colored data points indicate mutants 677 chosen for whole-genome sequencing. The dotted line represents relative WT Erdman growth (< ratio of 678 1). 679 680 Figure 2. PPE51 variants drive the eag phenotype and exhibit phenotypic and carbon source-681 dependent growth differences. 682 A) Growth curve of WT Mtb (Erdman and CDC1551 strains) overexpressing ppe51 and eag variants. 683 Growth of WT overexpression strains (pVV16 [empty vector] and pVV-ppe51-WT) were compared to WT 684 carrying the overexpression constructs of the mutant alleles (S211R, A228D, and E215K) in minimal media 685 (pH 5.7 + glycerol). Overexpression of eag mutant alleles in WT Mtb results in significantly enhanced 686 growth under acidic conditions. 687 B) eag variants show distinct clustering of variant type based on relative growth. 688 C) Statistical analysis of growth differences between native eag strains containing S211R or A228D was 689 performed using an unpaired t-test (****P < 0.0001) with Welch’s correction. 690 .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint D) Analysis of the S211R variant growth on various carbon sources. CDC1551 overexpression strains were 691 grown in MMAT medium (pH 5.7) in the presence of various growth-permissive (i.e. pyruvate acetate, OA) 692 and non-permissive (i.e. glucose, propionate, lactate) carbon sources. ppe51-S211R (pink bars) growth is 693 carbon source specific and only exhibits enhanced growth on glycerol, a normally non-permissive carbon 694 source at pH 5.7. Growth on permissive carbon sources is not impacted by ppe51-S211R. The horizontal 695 dotted line indicates the starting density of 0.05 OD600. Similar results were observed in Erdman (Figure 696 S5). 697 698 Figure 3. Viability and replication dynamics of eag mutants. 699 A) Growth curve of Δppe51 complemented from its native promoter with the WT ppe51 allele or mutant 700 alleles and performed under acid growth arrest conditions: minimal media buffered to pH 5.7 with glycerol 701 as a carbon source. eag variants promote Mtb growth at pH 5.7, while Δppe51 complemented with WT 702 ppe51 maintains growth arrest. 703 B) Viability of Mtb strains as measured by colony-forming units (CFUs). 704 C) CDC1551 Mtb containing the native variant allele, A228D, continues to replicate during acid growth 705 arrest conditions, but WT Mtb and Δppe51 cease replication. To estimate replication dynamics of the 706 indicated strains, clock plasmid replication data was obtained from CFU counts (right axis, dotted lines). 707 CFUs of plasmid-free and plasmid-bearing strains were then used to calculate cumulative bacterial burden 708 (CBB) of total live, dead, or degraded Mtb (left axis, solid lines). 709 D) Replication dynamics of the native A228D (CDC1551) variant, comparing CBB (cumulative bacterial 710 burden), CFU (total CFUs from nonselective plating), and % pBP10 (percentage of bacteria carrying 711 plasmid). 712 713 Figure 4. Mtb restricts glycerol uptake at low pH. Growth of WT CDC1551, Δppe51 (empty vector), 714 and Δppe51 complemented strains in minimal media supplemented in a dose-dependent manner with 715 glycerol and buffered to one of five pH levels (pH 6.5, 6.2, 6.0, 5.7, or 5.5). All strains exhibit a reduced 716 .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint capacity for growth starting ~2 mM glycerol compared to higher glycerol concentrations. At decreasing 717 pH, WT, Δppe51(empty vector), and Δppe51::pMV-WT restrict their ability to uptake glycerol, whereas 718 any variant complement is able to maintain glycerol uptake. However, restricted growth can be rescued at 719 high concentrations of glycerol (~80 mM) at pH 5.7 for WT, Δppe51(empty vector), and Δppe51::pMV-720 ppe51, and pH 5.5 for variant complements. Growth analyses were performed at Day 14 following initial 721 inoculation with data being shown as percent of the maximum well-growth. All conditions were conducted 722 in triplicate and representative of multiple independent experiments. Similar data were observed in a native 723 eag mutant (Figure S11B). 724 725 Figure 5. eag variants exhibit enhanced 14C-glycerol uptake and incorporation into lipids. 726 A) Strains expressing eag variants uptake 14C-glycerol at an enhanced rate. Mtb was pre-adapted for 3 days 727 in MMAT (pH 5.7) with 10 mM glycerol and subsequently washed prior to the addition of radiolabeled 728 glycerol. 14C-glycerol uptake was measured using scintillation counting at various time-points over the 729 course of 24 hours. Significance was determined by two-way ANOVA (Tukey’s multiple comparisons test; 730 ****P < 0.0001) 731 B) Incorporation of 14C-glycerol into sulfolipids at acidic pH. Sulfolipid is indicated with an arrow and 732 accumulates at a similar rate in each strain. Strains were analyzed in duplicate with representative results 733 being shown. 734 C) Incorporation of 14C-glycerol into TAG at acidic pH. TAG are indicated with an arrow and are absent 735 from all strains except for Δppe51::pMV-S211R. Strains were analyzed in duplicate with representative 736 results being shown. 737 738 Figure 6. eag variants exhibit selectively enhanced replication and reduced survival in activated 739 macrophages. 740 A) BMDMs infected with the isogenic panel of CDC1551 Δppe51 complemented strains and WT 741 CDC1551. Growth is similar for all strains in resting macrophages, but in activated BMDMs, WT, 742 .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint Δppe51(empty vector), and Δppe51::pMV-ppe51 exhibit ~1.25 log increase in growth compared to variant 743 complements which show a lower log increase in growth (~0.25-1). Data shown was conducted in triplicate 744 and representative of three independent experiments. 745 B) Statistical analysis of growth differences between Δppe51 complemented strains at Day 9 in activated 746 BMDMs. Significance was determined by one-way ANOVA (Tukey’s multiple comparisons test; *P < 747 0.05, ***P < .001, ****P < 0.0001). 748 C) Activated BMDMs infected with native WT CDC1551, Δppe51, and A228D variant strains containing 749 the pBP10 replication clock plasmid. CFUs on selective plates were compared to CFUs on nonselective 750 plates and used to calculated frequency of plasmid-bearing bacteria (% pBP10), cumulative bacterial burden 751 (CBB) of total live and dead bacteria, and total enumerated colonies on nonselective plates (CFU). 752 753 Figure 7. Glycerol differentially interacts with recombinant WT PPE51 or S211R variant proteins. 754 A) Recombinant WT PPE51 and S211R proteins were assessed for thermostability under no glycerol, 10 755 mM glycerol, and 10 mM glucose conditions. The protein was preincubated at room temperature (RT) for 756 20 minutes and subjected to eight different temperature conditions as indicated for five minutes. Following 757 heating, samples were spun down to pellet the protein precipitate. Soluble protein was removed and 758 analyzed by Western Blotting. * represent the highest temperature where soluble protein was detected. 759 B) Signal intensity of individual bands was measured and normalized to the pET::ppe51-WT (no glycerol) 760 band at RT. Samples containing glycerol continue to show detectable signal intensity up to 55°C for 761 pET::ppe51-S211R and 60°C for pET::ppe51-WT. 762 C. in silico protein structure modeling and function prediction for PPE51. The peptide sequence of PPE51-763 WT (green) and PPE51-S211R (blue) were analyzed using the Iterative Threading ASSEmbly Refinement 764 (I-TASSER) approach60. Both WT and variant PPE51 were modeled without constraint and appear to form 765 a porin-like structure with an inner channel. PPE51-S211R is modeled against the WT to show the slight 766 conformational changes that occur with the introduction of this mutation. PPE51-WT and PPE51-S211R 767 .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint structures received C-scores of -0.86 and -1.24, respectively, which is a measure of structure confidence on 768 a range of -5 (low) to 2 (high)60. A model with C-score >-1.5 usually indicates a correct fold. 769 770 Figure 8. ppe51 knockout strains contain background mutations that disrupt PDIM biosynthesis. 771 A) Whole genome sequencing results of Δppe51 show a nucleotide insertion in ppsC and a point mutation 772 in ppsD in CDC1551 and Erdman backgrounds, respectively. 773 B) Incorporation of 14C-acetate into PDIM at acidic and neutral pH. PDIM is indicated with an arrow and 774 accumulates in the WT strain at both pH 5.7 and pH 7.0 but is absent in the ppe51 knockout mutant. Strains 775 were analyzed in duplicate with representative results being shown. 776 777 Figure 9. A proposed model for the role of ppe51 and eag variants in glycerol acquisition. Presented 778 is a hypothetical model, in which ppe51 expression is induced by PhoP under acidic conditions. PPE51 is 779 thought to be secreted through ESX-5 and embeds itself into the mycomembrane, making itself surface-780 accessible. At this interface, it could interact with glycerol and promote transport across the mycomembrane 781 (WT Pathway). PPE51 variants may function by having an altered channel opening or ligand binding 782 surface, allowing for enhanced glycerol transport across the mycomembrane and leading to the enhanced 783 growth phenotype observed during acid growth arrest (Mutant Pathway). 784 785 Table 1. Whole genome sequencing results of isolated colony variants. 786 787 References 788 1 Via, L. E. et al. Tuberculous granulomas are hypoxic in guinea pigs, rabbits, and nonhuman 789 primates. Infect Immun 76, 2333-2340, doi:10.1128/IAI.01515-07 (2008). 790 2 Converse, P. J. et al. Role of the dosR-dosS two-component regulatory system in 791 Mycobacterium tuberculosis virulence in three animal models. Infect Immun 77, 1230-792 1237, doi:10.1128/IAI.01117-08 (2009). 793 .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint 3 Munoz-Elias, E. J. & McKinney, J. D. Carbon metabolism of intracellular bacteria. Cell 794 Microbiol 8, 10-22, doi:10.1111/j.1462-5822.2005.00648.x (2006). 795 4 Loebel, R. O., Shorr, E. & Richardson, H. B. The Influence of Adverse Conditions upon 796 the Respiratory Metabolism and Growth of Human Tubercle Bacilli. Journal of 797 bacteriology 26, 167-200, doi:10.1128/JB.26.2.167-200.1933 (1933). 798 5 Vandal, O. H., Nathan, C. F. & Ehrt, S. Acid resistance in Mycobacterium tuberculosis. J 799 Bacteriol 191, 4714-4721, doi:10.1128/JB.00305-09 (2009). 800 6 Voskuil, M. I., Bartek, I. L., Visconti, K. & Schoolnik, G. K. The response of 801 mycobacterium tuberculosis to reactive oxygen and nitrogen species. Front Microbiol 2, 802 105, doi:10.3389/fmicb.2011.00105 (2011). 803 7 Wayne, L. G. & Sohaskey, C. D. Nonreplicating persistence of mycobacterium 804 tuberculosis. Annu Rev Microbiol 55, 139-163, doi:10.1146/annurev.micro.55.1.139 805 (2001). 806 8 Zhang, Y. Persistent and dormant tubercle bacilli and latent tuberculosis. Frontiers in 807 Bioscience 9, 1136, doi:10.2741/1291 (2004). 808 9 Deb, C. et al. A Novel In Vitro Multiple-Stress Dormancy Model for Mycobacterium 809 tuberculosis Generates a Lipid-Loaded, Drug-Tolerant, Dormant Pathogen. PLoS ONE 4, 810 e6077, doi:10.1371/journal.pone.0006077 (2009). 811 10 Rohde, K. H., Abramovitch, R. B. & Russell, D. G. Mycobacterium tuberculosis invasion 812 of macrophages: linking bacterial gene expression to environmental cues. Cell Host 813 Microbe 2, 352-364 (2007). 814 11 Baker, J. J., Dechow, S. J. & Abramovitch, R. B. Acid Fasting: Modulation of 815 Mycobacterium tuberculosis Metabolism at Acidic pH. Trends Microbiol, 816 doi:10.1016/j.tim.2019.06.005 (2019). 817 12 Vandal, O. H., Pierini, L. M., Schnappinger, D., Nathan, C. F. & Ehrt, S. A membrane 818 protein preserves intrabacterial pH in intraphagosomal Mycobacterium tuberculosis. Nat 819 Med 14, 849-854 (2008). 820 13 Abramovitch, R. B., Rohde, K. H., Hsu, F. F. & Russell, D. G. aprABC: a Mycobacterium 821 tuberculosis complex-specific locus that modulates pH-driven adaptation to the 822 macrophage phagosome. Mol Microbiol 80, 678-694, doi:10.1111/j.1365-823 2958.2011.07601.x (2011). 824 14 Perez, E. et al. An essential role for phoP in Mycobacterium tuberculosis virulence. Mol 825 Microbiol 41, 179-187 (2001). 826 15 Johnson, B. K. et al. The Carbonic Anhydrase Inhibitor Ethoxzolamide Inhibits the 827 Mycobacterium tuberculosis PhoPR Regulon and Esx-1 Secretion and Attenuates 828 .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint Virulence. Antimicrob Agents Chemother 59, 4436-4445, doi:10.1128/AAC.00719-15 829 (2015). 830 16 Baker, J. J. & Abramovitch, R. B. Genetic and metabolic regulation of Mycobacterium 831 tuberculosis acid growth arrest. Sci Rep 8, 4168, doi:10.1038/s41598-018-22343-4 (2018). 832 17 Baker, J. J., Dechow, S. J. & Abramovitch, R. B. Acid Fasting: Modulation of 833 Mycobacterium tuberculosis Metabolism at Acidic pH. Trends Microbiol 27, 942-953, 834 doi:10.1016/j.tim.2019.06.005 (2019). 835 18 Baker, J. J., Johnson, B. K. & Abramovitch, R. B. Slow growth of Mycobacterium 836 tuberculosis at acidic pH is regulated by phoPR and host-associated carbon sources. Mol 837 Microbiol 94, 56-69, doi:10.1111/mmi.12688 (2014). 838 19 Camus, J. C., Pryor, M. J., Medigue, C. & Cole, S. T. Re-annotation of the genome 839 sequence of Mycobacterium tuberculosis H37Rv. Microbiology 148, 2967-2973, 840 doi:10.1099/00221287-148-10-2967 (2002). 841 20 Cole, S. T. et al. Deciphering the biology of Mycobacterium tuberculosis from the 842 complete genome sequence. Nature 393, 537-544, doi:10.1038/31159 (1998). 843 21 Karboul, A. et al. Frequent Homologous Recombination Events in Mycobacterium 844 tuberculosis PE/PPE Multigene Families: Potential Role in Antigenic Variability. Journal 845 of Bacteriology 190, 7838-7846, doi:10.1128/jb.00827-08 (2008). 846 22 Akhter, Y., Ehebauer, M. T., Mukhopadhyay, S. & Hasnain, S. E. The PE/PPE multigene 847 family codes for virulence factors and is a possible source of mycobacterial antigenic 848 variation: perhaps more? Biochimie 94, 110-116, doi:10.1016/j.biochi.2011.09.026 (2012). 849 23 Mukhopadhyay, S., Nair, S. & Ghosh, S. Pathogenesis in tuberculosis: transcriptomic 850 approaches to unraveling virulence mechanisms and finding new drug targets. FEMS 851 Microbiol Rev 36, 463-485, doi:10.1111/j.1574-6976.2011.00302.x (2012). 852 24 Bhat, K. H., Ahmed, A., Kumar, S., Sharma, P. & Mukhopadhyay, S. Role of PPE18 853 Protein in Intracellular Survival and Pathogenicity of Mycobacterium tuberculosis in Mice. 854 PLoS ONE 7, e52601, doi:10.1371/journal.pone.0052601 (2012). 855 25 Strong, M. et al. Toward the structural genomics of complexes: crystal structure of a 856 PE/PPE protein complex from Mycobacterium tuberculosis. Proc Natl Acad Sci U S A 103, 857 8060-8065, doi:10.1073/pnas.0602606103 (2006). 858 26 Mitra, A., Speer, A., Lin, K., Ehrt, S. & Niederweis, M. PPE Surface Proteins Are Required 859 for Heme Utilization by Mycobacterium tuberculosis. mBio 8, doi:10.1128/mBio.01720-860 16 (2017). 861 27 Zhang, L. et al. Comprehensive analysis of iron utilization by Mycobacterium tuberculosis. 862 PLoS Pathog 16, e1008337, doi:10.1371/journal.ppat.1008337 (2020). 863 .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint 28 Wang, Q. et al. PE/PPE proteins mediate nutrient transport across the outer membrane of 864 <em>Mycobacterium tuberculosis</em>. Science 367, 1147, 865 doi:10.1126/science.aav5912 (2020). 866 29 Espitia, C. et al. The PE-PGRS glycine-rich proteins of Mycobacterium tuberculosis: a 867 new family of fibronectin-binding proteins? Microbiology (Reading) 145 ( Pt 12), 3487-868 3495, doi:10.1099/00221287-145-12-3487 (1999). 869 30 Sultana, R., Tanneeru, K. & Guruprasad, L. The PE-PPE domain in mycobacterium reveals 870 a serine alpha/beta hydrolase fold and function: an in-silico analysis. PLoS One 6, e16745, 871 doi:10.1371/journal.pone.0016745 (2011). 872 31 Garrett, C. K., Broadwell, L. J., Hayne, C. K. & Neher, S. B. Modulation of the Activity 873 of Mycobacterium tuberculosis LipY by Its PE Domain. PLoS One 10, e0135447, 874 doi:10.1371/journal.pone.0135447 (2015). 875 32 Gey van Pittius, N. C. et al. Evolution and expansion of the Mycobacterium tuberculosis 876 PE and PPE multigene families and their association with the duplication of the ESAT-6 877 (esx) gene cluster regions. BMC Evol Biol 6, 95, doi:10.1186/1471-2148-6-95 (2006). 878 33 Simeone, R., Bottai, D., Frigui, W., Majlessi, L. & Brosch, R. ESX/type VII secretion 879 systems of mycobacteria: Insights into evolution, pathogenicity and protection. 880 Tuberculosis (Edinb) 95 Suppl 1, S150-154, doi:10.1016/j.tube.2015.02.019 (2015). 881 34 Ates, L. S. et al. Essential Role of the ESX-5 Secretion System in Outer Membrane 882 Permeability of Pathogenic Mycobacteria. PLoS Genet 11, e1005190, 883 doi:10.1371/journal.pgen.1005190 (2015). 884 35 Sani, M. et al. Direct Visualization by Cryo-EM of the Mycobacterial Capsular Layer: A 885 Labile Structure Containing ESX-1-Secreted Proteins. PLoS Pathogens 6, e1000794, 886 doi:10.1371/journal.ppat.1000794 (2010). 887 36 Sampson, S. L. Mycobacterial PE/PPE proteins at the host-pathogen interface. Clin Dev 888 Immunol 2011, 497203-497203, doi:10.1155/2011/497203 (2011). 889 37 Gey Van Pittius, N. C. et al. The ESAT-6 gene cluster of Mycobacterium tuberculosis and 890 other high G+C Gram-positive bacteria. Genome Biol 2, Research0044, doi:10.1186/gb-891 2001-2-10-research0044 (2001). 892 38 Abdallah, A. M. et al. PPE and PE_PGRS proteins of Mycobacterium marinum are 893 transported via the type VII secretion system ESX-5. Mol Microbiol 73, 329-340, 894 doi:10.1111/j.1365-2958.2009.06783.x (2009). 895 39 Tufariello, J. M. et al. Separable roles forMycobacterium tuberculosisESX-3 effectors in 896 iron acquisition and virulence. Proceedings of the National Academy of Sciences 113, 897 E348-E357, doi:10.1073/pnas.1523321113 (2016). 898 .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint 40 Mitra, A., Ko, Y. H., Cingolani, G. & Niederweis, M. Heme and hemoglobin utilization by 899 Mycobacterium tuberculosis. Nat Commun 10, 4260, doi:10.1038/s41467-019-12109-5 900 (2019). 901 41 Korycka-Machała, M. et al. PPE51 Is Involved in the Uptake of Disaccharides by 902 Mycobacterium tuberculosis. Cells 9, 603, doi:10.3390/cells9030603 (2020). 903 42 Malen, H., De Souza, G. A., Pathak, S., Softeland, T. & Wiker, H. G. Comparison of 904 membrane proteins of Mycobacterium tuberculosis H37Rv and H37Ra strains. BMC 905 Microbiol 11, 18, doi:10.1186/1471-2180-11-18 (2011). 906 43 Lee, W., VanderVen, B. C., Fahey, R. J. & Russell, D. G. Intracellular Mycobacterium 907 tuberculosis exploits host-derived fatty acids to limit metabolic stress. J Biol Chem 288, 908 6788-6800, doi:10.1074/jbc.M112.445056 (2013). 909 44 Piddington, D. L. a. K. A. a. B. N. A. Growth of Mycobacterium tuberculosis in a defined 910 medium is very restricted by acid pH and Mg(2+) levels. Infection and immunity 68, 4518-911 -4522 (2000). 912 45 Lohse, M. et al. RobiNA: a user-friendly, integrated software solution for RNA-Seq-based 913 transcriptomics. Nucleic Acids Res 40, W622-627, doi:10.1093/nar/gks540 (2012). 914 46 Li, H. Aligning sequence reads, clone sequences and assembly contigs with BWA-MEM. 915 arXiv 1303.3997v1[q-bio.GN], doi:arxiv:1303.3997 (2013). 916 47 McKenna, A. et al. The Genome Analysis Toolkit: a MapReduce framework for analyzing 917 next-generation DNA sequencing data. Genome Res 20, 1297-1303, 918 doi:10.1101/gr.107524.110 (2010). 919 48 Murphy, K. C. et al. ORBIT: a New Paradigm for Genetic Engineering of Mycobacterial 920 Chromosomes. mBio 9, e01467-01418, doi:10.1128/mBio.01467-18 (2018). 921 49 Danilchanka, O. et al. An outer membrane channel protein of Mycobacterium tuberculosis 922 with exotoxin activity. Proceedings of the National Academy of Sciences of the United 923 States of America 111, 6750-6755, doi:10.1073/pnas.1400136111 (2014). 924 50 Asensio, J. G. et al. The Virulence-associated Two-component PhoP-PhoR System 925 Controls the Biosynthesis of Polyketide-derived Lipids inMycobacterium tuberculosis. 926 Journal of Biological Chemistry 281, 1313-1316, doi:10.1074/jbc.c500388200 (2006). 927 51 Schneider, C. A., Rasband, W. S. & Eliceiri, K. W. NIH Image to ImageJ: 25 years of 928 image analysis. Nature Methods 9, 671-675, doi:10.1038/nmeth.2089 (2012). 929 52 Bachrach, G. et al. A new single-copy mycobacterial plasmid, pMF1, from Mycobacterium 930 fortuitum which is compatible with the pAL5000 replicon. Microbiology (Reading) 146 ( 931 Pt 2), 297-303, doi:10.1099/00221287-146-2-297 (2000). 932 .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint 53 Gill, W. P. et al. A replication clock for Mycobacterium tuberculosis. Nature medicine 15, 933 211-214, doi:10.1038/nm.1915 (2009). 934 54 Johnson, B. K. & Abramovitch, R. B. Macrophage infection models for Mycobacterium 935 tuberculosis. Methods Mol Biol 1285, 329-341, doi:10.1007/978-1-4939-2450-9_20 936 (2015). 937 55 Park, Y. et al. Inhibition of CorA-Dependent Magnesium Homeostasis Is Cidal in 938 Mycobacterium tuberculosis. Antimicrobial Agents and Chemotherapy 63, 939 doi:10.1128/aac.01006-19 (2019). 940 56 Walters, S. B. et al. The Mycobacterium tuberculosis PhoPR two-component system 941 regulates genes essential for virulence and complex lipid biosynthesis. Mol Microbiol 60, 942 312-330, doi:10.1111/j.1365-2958.2006.05102.x (2006). 943 57 Sirakova, T. D. et al. Identification of a diacylglycerol acyltransferase gene involved in 944 accumulation of triacylglycerol in Mycobacterium tuberculosis under stress. Microbiology 945 (Reading, England) 152, 2717-2725, doi:10.1099/mic.0.28993-0 (2006). 946 58 Baek, S.-H., Li, A. H. & Sassetti, C. M. Metabolic regulation of mycobacterial growth and 947 antibiotic sensitivity. PLoS Biol 9, e1001065-e1001065, 948 doi:10.1371/journal.pbio.1001065 (2011). 949 59 Rohde, K. H., Veiga, D. F. T., Caldwell, S., Balázsi, G. & Russell, D. G. Linking the 950 Transcriptional Profiles and the Physiological States of Mycobacterium tuberculosis 951 during an Extended Intracellular Infection. PLoS Pathogens 8, e1002769, 952 doi:10.1371/journal.ppat.1002769 (2012). 953 60 Roy, A., Kucukural, A. & Zhang, Y. I-TASSER: a unified platform for automated protein 954 structure and function prediction. Nature Protocols 5, 725-738, doi:10.1038/nprot.2010.5 955 (2010). 956 61 Bellerose, M. M. et al. Common Variants in the Glycerol Kinase Gene Reduce 957 Tuberculosis Drug Efficacy. mBio 10, e00663-00619, doi:10.1128/mBio.00663-19 (2019). 958 62 Daniel, J., Maamar, H., Deb, C., Sirakova, T. D. & Kolattukudy, P. E. Mycobacterium 959 tuberculosis Uses Host Triacylglycerol to Accumulate Lipid Droplets and Acquires a 960 Dormancy-Like Phenotype in Lipid-Loaded Macrophages. PLoS Pathog 7, e1002093 961 (2011). 962 63 Guerrini, V. et al. Storage lipid studies in tuberculosis reveal that foam cell biogenesis is 963 disease-specific. PLoS Pathog 14, e1007223, doi:10.1371/journal.ppat.1007223 (2018). 964 64 Mishra, K. C. et al. Functional role of the PE domain and immunogenicity of the 965 Mycobacterium tuberculosis triacylglycerol hydrolase LipY. Infect Immun 76, 127-140, 966 doi:10.1128/IAI.00410-07 (2008). 967 .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint 65 de Carvalho, L. P. S. et al. Metabolomics of Mycobacterium tuberculosis Reveals 968 Compartmentalized Co-Catabolism of Carbon Substrates. Chemistry & Biology 17, 1122-969 1131, doi:https://doi.org/10.1016/j.chembiol.2010.08.009 (2010). 970 66 Stahl, C. et al. MspA provides the main hydrophilic pathway through the cell wall of 971 Mycobacterium smegmatis. Mol Microbiol 40, 451-464, doi:10.1046/j.1365-972 2958.2001.02394.x (2001). 973 67 Stephan, J. et al. The growth rate of Mycobacterium smegmatis depends on sufficient 974 porin-mediated influx of nutrients. Molecular Microbiology 58, 714-730, 975 doi:https://doi.org/10.1111/j.1365-2958.2005.04878.x (2005). 976 68 Danilchanka, O., Pavlenok, M. & Niederweis, M. Role of Porins for Uptake of Antibiotics 977 by <em>Mycobacterium smegmatis</em>. Antimicrobial Agents and 978 Chemotherapy 52, 3127, doi:10.1128/AAC.00239-08 (2008). 979 69 Zambolin, S. et al. Structural basis for haem piracy from host haemopexin by Haemophilus 980 influenzae. Nature Communications 7, 11590, doi:10.1038/ncomms11590 (2016). 981 70 Mitra, A., Ko, Y.-H., Cingolani, G. & Niederweis, M. Heme and hemoglobin utilization 982 by Mycobacterium tuberculosis. Nature Communications 10, doi:10.1038/s41467-019-983 12109-5 (2019). 984 71 Bellerose, M. M. et al. Common variants in the glycerol kinase gene reduce tuberculosis 985 drug efficacy. mBio in press (2019). 986 72 Xu, W. et al. Chemical Genetic Interaction Profiling Reveals Determinants of Intrinsic 987 Antibiotic Resistance in Mycobacterium tuberculosis. Antimicrob Agents Chemother 61, 988 doi:10.1128/AAC.01334-17 (2017). 989 73 Wolfe, L. M., Mahaffey, S. B., Kruh, N. A. & Dobos, K. M. Proteomic definition of the 990 cell wall of Mycobacterium tuberculosis. J Proteome Res 9, 5816-5826, 991 doi:10.1021/pr1005873 (2010). 992 74 McEvoy, C. R. E. et al. Comparative Analysis of Mycobacterium tuberculosis pe and ppe 993 Genes Reveals High Sequence Variation and an Apparent Absence of Selective 994 Constraints. PLoS ONE 7, e30593, doi:10.1371/journal.pone.0030593 (2012). 995 996 .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint Figure 1.CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint Figure 2.CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint Figure 3.CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint Figure 4.CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint Figure 5.CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint Figure 6.CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint Figure 7.CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint Figure 8.CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint Figure 9.CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint Table 1. Whole genome sequencing results of isolated eag variants.CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint Supplemental Figure 1. Enhanced acid growth confirmation of mutants isolated from WT Erdman genetic screen. Single colony isolates from Plate 1 (A), Plate 2 (B), Plate 3 (C), and Plate 4 (D) were grown in liquid MMAT (pH 5.7) with glycerol and compared to the WT for the enhanced acid growth phenotype. Each symbol represents the numbered colony isolated from the acid growth arrest plates. .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint Supplemental Figure 2. SNP sites in ppe51. SNP mapping identified three separate mutations within a 50 bp region of ppe51. Underlined and starred bases represent the SNP position (bp). S211R substitution had two SNPs as denoted by the guanine (G) underneath..CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint Supplemental Figure 3. Growth curve of pVV16 overexpression constructs (CDC1551 and Erdman) in minimal media at pH 7.0 with 10 mM glycerol..CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint Supplemental Figure 4. Accumulation of EtBr by Mtb and pVV16 overexpression constructs. pVV16 overexpression constructs (CDC1551 and Erdman) were incubated with Ethidium bromide (EtBr) for 90 minutes. EtBr fluorescence was measured every 3 minutes using the excitation wavelength (530 nm) and emission wavelength (590 nm). Samples were measured in triplicate. .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint Supplemental Figure 5. Analysis of the S211R variant growth on various carbon sources. Erdman overexpression strains were grown in MMAT medium (pH 5.7) in the presence of various growth-permissive (i.e. pyruvate acetate, OA) and non-permissive (i.e. glucose, propionate, lactate) carbon sources. ppe51-S211R (pink bars) growth is carbon source specific and only exhibits enhanced growth on glycerol, a normally non-permissive carbon source at pH 5.7. Growth on permissive carbon sources is not impacted by ppe51-S211R. The horizontal dotted line indicates the starting density of 0.05 OD600..CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint Supplemental Figure 6A. Growth curves of CDC1551 overexpression strains on individual carbon sources. Each strain was grown in duplicate for 20 days on their respective carbon source in minimal media buffered to pH 5.7..CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint Supplemental Figure 6B. Growth curves of Erdman overexpression strains on individual carbon sources. Each strain was grown in duplicate for 20 days on their respective carbon source in minimal media buffered to pH 5.7..CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint Supplemental Figure 7. Construction of ppe51 deletion mutant in Mtb CDC1551 and Erdman.A) Schematic of chromosomal ppe51 and the subsequent ORBIT48-promoted deletion of the ppe51 target gene, attP replacement, and plasmid integration containing hygromycin resistance for selection.B) PCR amplification of the 5’(oriE) and 3’ (HygC-out1/2) junctions of CDC1551 and Erdman Δppe51. Positive bands were confirmed by sanger sequencing.C) PCR analysis of the integration site of the payload plasmid (pKM464). pKM464 is 3082 bp which is consistent with the size of the bands observed in the ppe51 deletion mutants compared to WT ppe51 which is 1143 bp.D) qRT-PCR analysis confirming the ppe51 knockout in CDC1551 and Erdman. Total RNA was collected after samples were grown for six days in minimal media buffered to pH 5.7. Fold expression was normalized to the respective WT strains. Error bars represent the standard deviation of three technical replicates. Deletion mutants typically exhibited a non-specific primed Ct ~30 cycles compared to WT which had a Ct of ~20 cycles.E) qRT-PCR analysis confirming the presence of the ppe51 gene in complemented Δppe51. Fold expression was normalized to WT Mtb. Error bars represent the standard deviation of three technical replicates. Complemented strains had Ct ~14 cycles compared to Δppe51 which had non-specific primed Ct ~30 cycles..CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint Supplemental Figure 8. Growth of complemented Δppe51 strains in A) minimal media at pH 5.7 with 10 mM glycerol (Erdman), B) minimal media at pH 7.0 with 10 mM glycerol (CDC1551) and C) minimal media at pH 7.0 with 10 mM glycerol (Erdman). Error bars represent standard deviation of three technical replicates. .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint Supplemental Figure 9. Viability of complemented Δppe51 strains in A) minimal media at pH 5.7 with 10 mM glycerol (Erdman), B) minimal media at pH 7.0 with 10 mM glycerol (CDC1551) and C) minimal media at pH 7.0 with 10 mM glycerol (Erdman). Error bars represent standard deviation of three technical replicates. .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint Supplemental Figure 10. In vitro replication dynamics of CDC1551 eag variants (pH 7.0) and Erdman eag variants (pH 5.7 and pH 7.0). A) All CDC1551 Mtb strains, WT, Δppe51, and the native variant allele, A228D, continues to replicate at neutral pH in minimal media. To estimate replication dynamics of the indicated strains, plasmid frequency data was obtained from CFU counts (right axis, dotted lines). CFUs of plasmid-free and plasmid-bearing strains were then used to calculate cumulative bacterial burden (CBB) of total live, dead, or degraded Mtb (left axis, solid lines).B) Erdman strains containin the native variant alleles, S211R, A228D, and E215K continue to replicate at pH 5.7 in minimal media, albeit E215K exhibits a more reduced capacity for replication compared to S211R and A228D. WT and Δppe51 cease replication. Plasmid frequency (right axis, dotted lines) and cumulative bacterial burden (CBB) (left axis, solid lines) are shown.C) All Erdman Mtb strains, WT, Δppe51, and the native variant alleles (S211R, A228D, and E215K), continue to replicate at neutral pH in minimal media. Plasmid frequency (right axis, dotted lines) and cumulative bacterial burden (CBB) (left axis, solid lines) are shown.Replication dynamics of the native Erdman D) S211R variant, E) A228D variant, and F) E215K variant ( comparing CBB (cumulative bacterial burden), CFU (total CFUs from nonselective plating), and % pBP10 (percentage of bacteria carrying plasmid).. .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint Supplemental Figure 11A. Mtb shows growth restriction at low pH in Erdman. Growth of WT Erdman, Δppe51 (empty vector), and Δppe51 complemented strains in minimal media supplemented in a dose-dependent manner with glycerol and buffered to one of five pH levels (pH 6.5, 6.2, 6.0, 5.7, or 5.5). All strains exhibit a reduced capacity for growth starting ~2 mM glycerol compared to higher glycerol concentrations. At decreasing pH, WT, Δppe51(empty vector), and Δppe51::pMV-WT restrict their ability to uptake glycerol, whereas any variant complement is able to maintain glycerol uptake. However, restricted growth can be rescued at high concentrations of glycerol (~80 mM) at pH 5.7 for WT, Δppe51(empty vector), and Δppe51::pMV-ppe51, and pH 5.5 for variant complements. Growth analyses were performed at Day 14 following initial inoculation with data being shown as percent of the maximum well-growth. All conditions were conducted in triplicate and representative of multiple independent experiments..CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint Supplemental 11B.Mtb growth restriction and rescue at low pH is also observed in the native eag variants in CDC1551 and Erdman. WT CDC1551 and Erdman exhibit a slight rescuing of growth at high glycerol concentrations (~80 mM) in minimal media buffered to pH 5.7, consistent with what has been observed in the Δppe51 complemented strains. The native CDC1551 eag variant, A228D, and native Erdman eag variant exhibit a greater capacity for growth at pH 5.7 at lower concentrations of glycerol (~5mM) compared to their respective WT strains. Growth analyses were performed at Day 14 following initial inoculation with data being shown as percent of the maximum well-growth. All conditions were conducted in duplicate and representative of multiple independent experiments.. .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint Supplemental Figure 12. Glycerol uptake in native eag variants (pH 5.7), and radiolabeled uptake and incorporation into lipids at pH 7.0. A) The Erdman native eag S211R variant and the overexpressing pVV16-S211R variant uptake 14C-glycerol at a similar enhanced rate compared to WT Erdman which exhibits a more reduced capacity for radiolabeled uptake. Mtb was pre-adapted for 3 days in minimal media (pH 5.7) with 10 mM glycerol and subsequently washed prior to the additional of radiolabeled glycerol. 14C-glycerol uptake was measured at various timepoints using Scintillation counting over 24 hours. Significance was determined by two-way ANOVA (Tukey’s multiple comparisons test; ****P < 0.0001)B) Erdman WT, Δppe51 (empty vector) and, Δppe51::pMV-S211R were assessed for 14C-glycerol uptake in minimal media buffered to pH 7.0. Strains were pre-adapted for 3 days in minimal media buffered to pH 7.0 with 10 mM glycerol. All strains showed similar rates of glycerol uptake.C) Incorporation of 14C-glycerol into sulfolipids at neutral pH. Sulfolipid is absent at pH 7.0 which is consistent with previous observations made by Baker et. al. Strains were analyzed in duplicate with representative results being shown.D) Incorporation of 14C-glycerol into TAG at acidic pH. TAG is indicated with an arrow and are present in all strains. Strains were analyzed in duplicate with representative results being shown.. .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint Supplemental 13. Resting BMDMs infected with native WT CDC1551, Δppe51, and A228D variant strains containing the pBP10 replication clock plasmid. CFUs on selective plates were compared to CFUs on nonselective plates and used to calculated frequency of plasmid-bearing bacteria (% pBP10), cumulative bacterial burden (CBB) of total live and dead bacteria, and total enumerated colonies on nonselective plates (CFU).. .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint Supplemental 14. Protein expression of PPE51His and in silico modeling. Lysates of overexpressed His-tagged PPE51-WT and PPE51-S211R were run on a Talon resin column, and fractions were collected in (4) 1 mL aliquots. PPE51-WT and PPE51-S211R proteins were separated on 12% SDS-PAGE gels , which were either stained with Coomassie Blue dye (A) or used for western blots (B). Western blots were incubated with mouse anti-His tag monoclonal antibody followed by HRP-conjugated anti-mouse IgG secondary antibody. The molecular weights of the protein standards are shown on the left.C) Top ten PDB structures close to the target protein. TM-scores are a measurement of the structural similarity between the query structure and known structures in the PDB library in the range [0,1]. TM-scores >0.5 indicate a more correct topology. RMSD is the measurement of the average distance of residues between two structures.D) Ligand binding site prediction. C-score is the confidence score of the prediction in the range [0,1]. A higher score indicates a more reliable prediction. Ligand names are possible binding ligands foud in the BioLiP database.E) Gene ontology (GO) term prediction. Summary prediction of the most common GO terms occurring in three functional aspects (molecular function, biological process, and cellular component). GO-Score is a confidence score of the predicted GO term. A GO-Score >0.5 indicates a more reliable prediction.F) Modeling of the eag mutation sites in PPE51 shows that S211R (pink), E215K (orange), and A228D (blue) all occur on on the same alpha helix. . .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint Supplemental Figure 15. Incorporation of 14C-glycerol into PDIM at acidic and neutral pH. PDIM is indicated with a bracket and accumulates in the WT strain at both pH 5.7 and pH 7.0 but is absent in the ppe51 knockout mutant. A band consistent with TAG appears in Δppe51::pMV-S211R and is indicated by an arrow. Strains were analyzed in duplicate with representative results being shown.. Table S1. Mass Spectrometry results for bands associated with PPE51 induction. Sample NameClusterProtein NameProtein accession numbersProtein molecular weight (Da)Protein identification probabilityExclusive unique peptide countExclusive unique spectrum countTotal spectrum countInducedCluster of PPE family protein [Mycobacterium tuberculosis CDC1551] (AAK47561.1)PPE family protein [Mycobacterium tuberculosis CDC1551]AAK47561.1,KBN10067.1,WP_003416381.1,sp|P9WHY2.1|PPE51_MYCTO37,979.90100.00%51332PelletCluster of PPE family protein [Mycobacterium tuberculosis CDC1551] (AAK47561.1)PPE family protein [Mycobacterium tuberculosis CDC1551]AAK47561.1,KBN10067.1,WP_003416381.1,sp|P9WHY2.1|PPE51_MYCTO37,979.90100.00%71754.CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint Plasmid or Primer Name Characteristics or Sequence (5’à 3’) Reference Plasmids pVV16 KanR, HygR; E. coli-mycobacterial shuttle vector containing the hsp60 promoter BEI Resources, NIAID, NIH pKM444 KanR; Mycobacterial shuttle vector expressing the Che9c phage RecT annealase and the Bxb1 phage integrase from the Ptet promoter Murphy et. al. (2018) pKM464 HygR; Mycobacterial integration vector for deleting target gene, insertion of Bxb1 attB site from the PHyg promoter Murphy et. al. (2018) pMV306 KanR; Mycobacterial integration vector Stover et. al. (1991) pBP10 AmpR, KanR; Mycobacterial shuttle vector, used as a replication clock plasmid Bachrach et. al. (2000) pET-23a(+) AmpR; Bacterial expression vector carrying an N-terminal T7-Tag sequence and a C-terminal His-Tag sequence from the T7 promoter Rosenberg et. al. (1987); Studier et. al. (1990) pVV-ppe51-WT ppe51 PCR product ligated into BamHI and HindIII sites of pVV16 Baker et. al. (2018) pVV-ppe51-S211R ppe51 PCR product ligated into BamHI and HindIII sites of pVV16 Baker et. al. (2018) pVV-ppe51-A228D Ligated site-directed mutagenesis product of pVV16 and ppe51 This study pVV-ppe51-E215K Ligated site-directed mutagenesis product of pVV16 and ppe51 This study pMV::ppe51 ppe51+ its native promoter PCR product ligated into XbaI and EcoRI sites of pMV306 This study pMV::ppe51-S211R Ligated site-directed mutagenesis product of pMV306 and ppe51 This study pMV::ppe51-A228D Ligated site-directed mutagenesis product of pMV306 and ppe51 This study pMV::ppe51-E215K Ligated site-directed mutagenesis product of pMV306 and ppe51 This study pMV::ppe51-S211R+A228D Ligated site-directed mutagenesis product of pMV306 and ppe51-S211R This study pET::ppe51-WT ppe51 PCR product ligated into BamHI and HindIII sites of pET23a(+) This study pET::ppe51-S211R Ligated site-directed mutagenesis product of pET32a(+) and ppe51 This Study .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint ORBIT Oligonucleotide ppe51 (oligomer) ACGACACCGTATCCGCACAAATGTAAGGAGCTGAGACACAATGGATTTCGCACTGTTACCACCGGAAGTCGGTTTGTCTGGTCAACCACCGCGGTCTCAGTGGTGTACGGTACAAACCGTGATGGCCCACCCACCCGCGGCAGGGTAACCCGGCGCCTAACCGACAGGCGGCCCGTTGGGCGTAAACG This Study ORBIT primers oriE CCTGGTATCTTTATAGTCCTGTCG Murphy et. al (2018) HygC-out Murphy et. al (2018) PCR primers Seq-ppe51-For ATGGATTTCGCACTGTTACCACCGGA Baker et. al (2018) Seq-ppe51-Rev CTGTCGGTTAGTTACCCTGCCGC Baker et. al (2018) pMV306::ppe51-Fwd GGTACCAGATCTTTAAATGCCTGCCGCACAGAACCTC This study pMV306::ppe51-Rev GTCGACATCGATAAGCTTCGTTACCCTGCCGCGGGTG This study pET23a::ppe51-Fwd ATATATGGATCCATGGATTTCGCACTGTTACCACCGGAAG This study pET23a::ppe51-Rev ATATAAGCTTTGGGTGGGCCATCACCGTGA This study pMV306-conf-Fwd CGTATTACCGCCTTTGAGTGAG This study pMV306-conf-Rev GCAGTGAAGAGAATAGACCGG This study Site-directed mutagenesis primers ppe51-S211R-1-Fwd GCTGACGATTCCGAGATTCATCCCTGAGGAC This Study ppe51-S211R-1-Rev GTCCTCAGGGATGAATCTCGGAATCGTCAGC This study ppe51-S211R-2-Fwd GCTGACGATTCCGAGGTTCATCCCTGAGGAC This study ppe51-S211R-2-Rev GTCCTCAGGGATGAACCTCGGAATCGTCAGC This study ppe51-A228D-Fwd CGATTCCGAGCTTCATCCCTAAGGACTTCACCTTC This study ppe51-A228D-Rev GAAGGTGAAGTCCTTAGGGATGAAGCTCGGAATCG This study ppe51-E215K-Fwd CATATTCGCTGGATATGACACGGTAGGTGTGACG This study .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint Table S2. Plasmids and Primers used in this study. The Bxb1 phage attP sequence is in bold. Enzyme restriction cut sites are underlined. Site-directed mutagenesis sites are bold and underlined. ppe51-E215K-Rev CGTCACACCTACCGTGTCATATCCAGCGAATATG This study qRT-PCR primers RTpcr-ppe51-Fwd GAGCAAGCATACGCAATGAC RTpcr-ppe51-Rev AGTGTTCTGGCCGAAGAAG .CC-BY 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted May 19, 2021. ; https://doi.org/10.1101/2021.05.19.444820doi: bioRxiv preprint | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/344121502Interdependence of Approach and Avoidance Goals in Romantic Couples OverDays and MonthsArticle in The Journals of Gerontology Series B Psychological Sciences and Social Sciences · September 2020DOI: 10.1093/geronb/gbaa149CITATIONS2READS3076 authors, including:Some of the authors of this publication are also working on these related projects:Basel Study of Preterm Children View projectSelf-efficacy in mathematical development View projectJana NikitinUniversity of Vienna56 PUBLICATIONS 762 CITATIONS SEE PROFILEJenna WünscheDeutsches Zentrum für Altersfragen18 PUBLICATIONS 100 CITATIONS SEE PROFILEJanina Larissa BuehlerJohannes Gutenberg-Universität Mainz31 PUBLICATIONS 239 CITATIONS SEE PROFILERebekka WeidmannMichigan State University41 PUBLICATIONS 304 CITATIONS SEE PROFILEAll content following this page was uploaded by Jana Nikitin on 05 September 2020.The user has requested enhancement of the downloaded file.INTERDEPENDENCE OF APPROACH AND AVOIDANCE GOALS 1 Interdependence of Approach and Avoidance Goals in Romantic Couples Over Days and Months Jana Nikitin, PhD1, Jenna Wünsche, MSc1, Janina L. Bühler, PhD2, Rebekka Weidmann, PhD1, Robert P. Burriss, PhD1, Alexander Grob, PhD1 1 Department of Psychology, University of Basel 2 Department of Psychology, Heidelberg University Author Note Correspondence concerning this paper should be addressed to Jana Nikitin, Department of Psychology, University of Basel, Missionsstrasse 62a, 4055 Basel, Switzerland. Email: jana.nikitin@unibas.ch. Article Accepted for publication in The Journals of Gerontology: Series B This manuscript might not exactly replicate the final version published in the journal. It is not the copy of record. Please cite as: Nikitin, J., Wünsche, J., Bühler, J. L., Weidmann, R., Burriss, R. P., Grob, A., Interdependence of approach and avoidance goals in romantic couples over days and months, The Journals of Gerontology: Series B, gbaa149, https://doi.org/10.1093/geronb/gbaa149 INTERDEPENDENCE OF APPROACH AND AVOIDANCE GOALS 2 Abstract Objectives. Despite the centrality of people’s approach goals (i.e., approach toward positive outcomes) and avoidance goals (i.e., avoidance of negative outcomes) in romantic relationships, little is known about the interdependence of approach and avoidance relationship goals between partners. Assuming that short-term, state-level goals accumulate into general goal tendencies, the present research tested whether partners’ daily (i.e., state level) and aggregated daily (i.e., trait level) approach and avoidance goals are mutually predictive in the short term (after one day) and the long term (after 10-12 months). In addition, we explored whether goal interdependence unfolds differently across adulthood and in relationships of different duration. Method. Approach and avoidance goals were assessed daily on two 14-day measurement-burst occasions that were conducted 10-12 months apart. The sample consisted of N = 456 female-male couples (age: M = 33.6, SD = 13.8 years; relationship duration: M = 9.6, SD = 10.7 years). Results. We observed significant short- and long-term partner effects in the prediction of couple members’ approach and avoidance goals. These partner effects were restricted to trait level and they did not emerge at the state level. Almost all effects were independent of age and relationship duration. Discussion. The present research underscores the importance of disentangling state- and trait-level goal tendencies when investigating the interdependence of approach and avoidance goals within romantic relationships. Keywords: personality, goal interdependence, romantic relationships, motivation INTERDEPENDENCE OF APPROACH AND AVOIDANCE GOALS 3 Interdependence of Approach and Avoidance Goals in Romantic Couples Over Days and Months Approach and avoidance goals are crucial for understanding emotions, cognitions, and behaviors across the life span (Nikitin & Schoch, in press). Approach goals are defined as “striving for desired end-states”; avoidance goals are defined as “striving to move away from undesired end-states” (Scholer, Cornwell, & Higgins, 2019, p. 111). In the relationship domain, the desired end-states include relationship growth, intimacy, or closeness, whereas undesired end-states include hurt, rejection, or conflict (Elliot et al., 2006). Little is known about whether approach and avoidance goals interact in close relationships. Taking a transactional view on personality development (Caspi & Roberts, 2001; Neyer et al., 2014), the current study examines whether romantic partners’ approach and avoidance relationship goals are interdependent over the short and long term. Specifically, the present research seeks to answer (a) whether a person’s goals are predicted by the goals of the partner, (b) whether partners’ goals predict each other in the short and the long term, and (c) whether goal interdependence manifests differently across adulthood and relationship duration. Acknowledging that goals are “natural units of analysis for a personality psychology” (Little, 1989, p. 15), the present research provides insights into personality development in the context of romantic relationships. Approach and Avoidance Relationship Goals Approach and avoidance goals are thought to embody two motivational systems that operate in parallel (Gable et al., 2003). Whereas approach goals are associated with the processing of positive information, with positive emotions, and with approach behavior, avoidance goals are associated with the processing of negative information, with negative emotions, and with inhibited behavior (e.g., Nikitin & Freund, 2010, 2018; Schoch et al., 2015). Unsurprisingly, approach goals predict subjective well-being positively and avoidance goals negatively (see Nikitin & Schoch, in press). This is also the case in the context of INTERDEPENDENCE OF APPROACH AND AVOIDANCE GOALS 4 romantic relationships: Approach relationship goals are associated with increases in relationship satisfaction over time, whereas avoidance relationship goals are associated with decreases in relationship satisfaction over time (e.g., Impett et al., 2010; Kuster et al., 2017). Importantly, approach and avoidance goals do not simply mirror relationship satisfaction. Instead, they reflect individual differences in the motivation to obtain incentives and avoid threats in the relationship (Gable & Poore, 2008). Do a Person’s Goals Depend on the Goals of their Partner? It has been proposed that dynamic transactions between individuals and the social environment are responsible for personality development (Caspi & Roberts, 2001; McAdams & Olson, 2010; Neyer et al., 2014; Wagner et al., in press). An important social context for personality development is the romantic relationship (Finkel et al., 2014). Previous research has demonstrated that romantic partners co-develop (i.e., they develop in similar directions) in individual characteristics, such as self-esteem, subjective well-being, and attachment (Hudson et al., 2014; Orth et al., 2018). Two mechanisms of co-development in partners are being discussed in the literature: (1) shared environments and (2) reciprocal influence (Hudson et al., 2014; Orth et al., 2018). With respect to shared environments, partners tend to develop similarly because they share the same environmental factors (Orth et al., 2018). This mechanism does not require any interdependent co-regulation between the partners (i.e., mutual influence; e.g., Sbarra & Hazan, 2008). In the present study, we focus on reciprocal influence and examine whether romantic partners’ goals are mutually predictive across time. The transactive goal dynamic theory (TGD; Fitzsimons et al., 2015) suggests that the dyad represents “a single self-regulating system” (Fitzsimons et al., 2015, p. 648), in which goals, their pursuit, and their outcomes are interdependently related (Finkel et al., 2014; Fitzsimons et al., 2015). The core idea is that the output of one person’s goal pursuit inputs into the other person’s goal pursuit, so that partners’ goals affect each other. Although this INTERDEPENDENCE OF APPROACH AND AVOIDANCE GOALS 5 influence is possible in any relationship, it is particularly strong in romantic relationships because the partners have both the opportunity and the motivation to coordinate their goals (Finkel et al., 2014). Applying these ideas to approach and avoidance goals, it is likely that a person’s own approach and avoidance goals affect their partner’s approach and avoidance goals and vice versa. Research indicates that romantic partners tend to have similar levels of daily approach and avoidance goals (LaBuda et al., 2019). However, it remains unresolved whether partners’ approach and avoidance goals predict each other across time, and whether partners’ goals predict each other within a goal orientation (e.g., approach goals of one partner predict approach goals of the other partner) or between goal orientations (e.g., approach goals of one partner predict avoidance goals of the other partner). We assume that partners’ goals predict each other within a goal orientation rather than between goal orientations since approach goals are associated with processing approach-relevant (but not avoidance-relevant) information, whereas avoidance goals are associated with processing avoidance-relevant (but not approach-relevant) information (e.g., Schoch et al., 2015). We therefore consider approach goals of one partner to be a cue for approach (but not avoidance) goals of the other partner, whereas avoidance goals of one partner might be a cue for avoidance (but not approach) goals of the other partner (see also LaBuda et al., 2019). Does Goal Interdependence Occur in the Short and Long Term? Goals reflect both relatively stable personality traits and a person’s current state (Gable & Impett, 2012; McAdams & Olson, 2010). We investigate both levels of approach and avoidance goals and their interdependence between couple members. Specifically, we test whether persons’ state-level goals affect their partners’ state-level goals on the next day (short-term effects). Similarly, we test whether persons’ trait-level goals affect their partners’ trait-level goals 10-12 months later (long-term effects). In both cases, we control for partners’ INTERDEPENDENCE OF APPROACH AND AVOIDANCE GOALS 6 concurrent goal similarity because concurrent similarities in goals might be an artifact of selection processes rather than a coregulation mechanism (Hudson et al., 2014). Investigating both short-term and long-term interdependence of partners’ goals should provide information about the time scale at which interdependence unfolds within a relationship. It has been repeatedly argued that long-term trait development occurs due to short-term state processes (e.g. the TESSERA framework; Wrzus & Roberts, 2017). Accordingly, development in individual characteristics presumably results from repeated state experiences that develop into consistent patterns of thoughts, behaviors, and feelings. Building upon this framework, we assume that accumulated state experiences (i.e., aggregated daily levels) in approach and avoidance goals predict long-term differences in approach and avoidance goals. At the same time, we investigate short-term goal interdependence, by testing whether daily fluctuations in a person’s goals change their partner’s approach and avoidance goals on the next day. In other words, we test whether partners’ goals transmit in relatively short periods of time and, consequently, evoke change in one another. Does Goal Interdependence Manifest Differently with Age and Relationship Duration? Personality can change throughout the life span until old age (e.g., Roberts et al., 2006; Roberts & DelVecchio, 2000). Although there are no longitudinal studies on the development of approach and avoidance goals in adulthood, there is evidence from cross-sectional studies that the capacity for personality change might also hold true for approach and avoidance goals (e.g., Nikitin & Freund, 2019b). Similarly, the dynamic goal theory of marital satisfaction proposes that couple’s goals are characterized by dynamic changes across adulthood until old age (Li & Fung, 2011). It remains an open question whether the strength of interdependence in goals between partners differs across the life span. When investigating developmental processes in romantic relationships across the life span, two time variables are important: age and relationship duration. Although age and relationship duration are highly correlated, both might yield different effects on relationship INTERDEPENDENCE OF APPROACH AND AVOIDANCE GOALS 7 processes because they are based on distinct mechanisms. Specifically, developmental mechanisms that are associated with age, such as socioemotional selectivity (Carstensen et al., 1999), are not necessarily the same as relationship mechanisms that are associated with relationship duration, such as the honeymoon-is-over effect (e.g., Kurdek, 1998). This prominent mechanism suggests that high initial levels of relationship satisfaction decrease because couples are, for instance, confronted with a daily routine and a reality check of their positive illusions. In fact, research on personal growth in romantic relationships has revealed differential effects of age and relationship duration on the relationship processes that facilitate the movement toward an ideal self (Bühler et al., 2019). Thus, to disentangle the roles of age and relationship duration, we test age and relationship duration as separate moderators in our models. The Current Study To summarize, approach and avoidance goals are two fundamental goal orientations that operate in parallel and are influential in people’s daily lives and romantic relationships. Little is known about how approach and avoidance goals develop within individuals (i.e., actor effects) and are interdependent between partners (i.e., partner effects), and there has been no research on the moderation effects of age and relationship duration. The current study investigated these questions by looking at partners’ approach and avoidance goals in romantic relationships in the short term and the long term. We hypothesize the following short-term effects: H1a. Actor’s daily approach goals are positively predicted by their own and partner’s approach (but not avoidance) goals the day before. H1b. Actor’s daily avoidance goals are positively predicted by their own and partner’s avoidance (but not approach) goals the day before. We hypothesize the following long-term effects: INTERDEPENDENCE OF APPROACH AND AVOIDANCE GOALS 8 H2a. Actor’s approach goals at T2 are positively predicted by their own and partner’s approach (but not avoidance) goals at T1. H2b. Actor’s avoidance goals at T2 are positively predicted by their own and partner’s avoidance (but not approach) goals at T1. In addition, we explore whether the hypothesized processes differ as a function of age and relationship duration. The preregistered hypotheses and research questions including analysis codes are available on the Open Science Framework (https://osf.io/j5wvg/?view_only=bf4f45dee9d647db9b11425f304f9cc5). In addition to the preregistered hypotheses and research questions, we control for daily relationship satisfaction in all analyses to ensure that the effects of approach and avoidance goals are not explained by daily fluctuations in relational well-being. Moreover, we include partners’ trait approach and avoidance goals as predictors of actors’ daily approach and avoidance goals to explore partners’ trait goals as an additional source of differences in actors’ daily goals. These analyses were not preregistered but emerged during the review process as valuable additions. Methods Recruitment and Procedure The data came from a longitudinal online study of romantic couples (for more information, see Bühler et al., 2020). In the last two waves that were separated by an interval of 10-12 months, participants reported on their daily approach and avoidance relationship goals. We use data from these two waves and refer to them as T1 and T2. The ethics committee of the Department of Psychology at the University of Basel approved the study. Sample At T1, 1,481 participants at least partially responded to the surveys (including 672 couples in which both partners responded); at T2, 1,265 participants at least partially responded to the surveys (including 549 couples in which both partners responded). To identify the analysis samples for the short-term and long-term effects, we followed a stepwise INTERDEPENDENCE OF APPROACH AND AVOIDANCE GOALS 9 selection procedure (note that the presented selection criteria were applied on the couple level; if one dyad member did not fulfill a selection criterion, both members of the dyad were excluded from the analyses). In a first step, we excluded 20 same-gender couples because we relied on dyadic models for distinguishable dyads, and gender was used as the distinguishing variable. In a second step, we excluded 60 couples that separated between T1 and T2 or did not report on their relationship status at T2. The remaining sample consisted of 469 couples. In a third step, we applied the following two criteria: (1) To test the short-term effects, we selected couples in which both partners reported their approach and avoidance goals on at least two consecutive days in the T1 daily surveys. This criterion was met by 441 couples. (2) To test the long-term effects, we selected couples in which both partners reported at least once on their approach and avoidance goals during both T1 and T2. This criterion was met by 456 couples.1 In the following, we describe the T1 characteristics of couples that were included either in the short-term or the long-term analyses or both (N = 463 couples).2 At T1, the mean age was 33.6 (SD = 13.8, range 18–82, Mdn = 28.67) years and the mean relationship duration was 9.6 (SD = 10.7, range 1–53, Mdn = 5.76) years. The majority of this sample (54.5%) was single; 41.8% were married; 0.2% were cohabiting; 2.9% were separated or divorced; and 0.5% were widowed. Two thirds of the sample (66.9%) reported to be full- or part-time employed; 23.9% were in educational training; 2.2% were unemployed or for other reasons not active in the labor force (5.5%); and 20.5% reported some other employment status (e.g., working as a freelancer, voluntary work). The majority of the sample (61.6%) was resident in Germany; 24.6% were resident in Switzerland and 12.5% in Austria; 1.2 % reported living in another country. INTERDEPENDENCE OF APPROACH AND AVOIDANCE GOALS 10 Measures Approach and Avoidance Relationship Goals Approach and avoidance relationship goals were assessed on each of the 14 days at T1 and T2. The goals were measured with the Friendship Goals Scale (Elliot et al., 2006). The items were translated into German and reformulated so as to capture approach and avoidance goals in the romantic domain (Impett et al., 2008) and with respect to the daily relationship context (see Table S1). Previous research demonstrated configural, metric, and scalar measurement invariance of the scales between young, middle-aged, and older adults (Nikitin & Freund, 2019b) and high predictive validity for individual and relationship outcomes (for a summary see LaBuda et al., 2019). Approach and avoidance relationship goals were each assessed with four items that were rated on a seven-point scale (1 = very little; 7 = very much). For the present analyses, responses were summarized into daily mean scores for the short-term effects. For the long-term effects and to account for general partner tendencies in the short-term effects, we created wave-specific mean scores (i.e., across the 14-day survey intervals of both T1 and T2). The aggregation into wave-specific mean scores is based on the idea that personality states are momentary manifestations or expressions of personality traits. Consequently, aggregated states across measurements should approximate a trait (e.g., Rauthmann et al., 2019). Daily Cronbach’s alphas for the approach-goals scale ranged between a = .81 and a = .90; for the avoidance-goals scale between a = .71 and a = .86 (except for day 1 of T1 and T2 where α = .67 and .69, respectively). Controls We controlled for relationship satisfaction which was also assessed on a daily level. Participants responded to the question “All in all, how satisfied were you with the relationship with your partner today?” on a five-point scale (1 = very dissatisfied; 5 = very satisfied). INTERDEPENDENCE OF APPROACH AND AVOIDANCE GOALS 11 Statistical Approach In the following, we describe an expanded version of the preregistered analytical approach. To test the short-term effects, we used a multi-level approach for distinguishable dyads (Bolger & Laurenceau, 2013), in which person observations (Level 1) were nested within dyads (Level 2).3 We specified separate models to predict daily approach and avoidance goals. To quantify the amount of variation within female and male partners’ daily approach and avoidance goals as well as the amount of between-couple variability in both partners’ approach and avoidance goals, we first estimated intercept-only models. In a next step, and to test the hypothesized short-term effects, we used the analytical approach presented in Equation 1. An equivalent model was specified for daily avoidance goals. To obtain separate effect estimates for both couple members, we followed a dual-intercept approach and included two dummy-coded variables that distinguished between observations of female and male partners. (1) !""#$%&ℎjk = )*+%,*j × .F-0j + +%,*j × .M-0j + )*+%,*j × .F-1j(!&1$#!""#$%&ℎj[k-1]) ++%,*j × .M-1j(!&1$#!""#$%&ℎj[k-1]) +)*+%,*j × .F-2j(3%#14*#!""#$%&ℎj[k-1]) ++%,*j × .M-2j(3%#14*#!""#$%&ℎj[k-1]) +)*+%,*j × .F-3j(!&1$#!5$67%4&*j[k-1]) ++%,*j × .M-3j(!&1$#!5$67%4&*j[k-1]) +)*+%,*j × .F-4j(3%#14*#!5$67%4&*j[k-1]) ++%,*j × .M-4j(3%#14*#!5$67%4&*j[k-1]) +)*+%,*j × .F-5j(!&1$#8%169)%&16$4j[k-1]) ++%,*j × .M-5j(!&1$#8%169)%&16$4j[k-1]) +)*+%,*j × .F-6j(3%#14*#8%169)%&16$4j[k-1]) ++%,*j × .M-6j(3%#14*#8%169)%&16$4j[k-1]) + *F-jk + *M-jk INTERDEPENDENCE OF APPROACH AND AVOIDANCE GOALS 12 where .F-0j = γF00 + )*+%,*j × =F01(>*%43%#14*#!""#$%&ℎj) +)*+%,*j × =F02(>*%43%#14*#!5$67%4&*j) + ?F0j .F-1j = γF10 + ?F1j .F-2j = γF20 + ?F2j .F-3j = γF30 + ?F3j .F-4j = γF40 + ?F4j .F-5j = γF50 .F-6j = γF60 .M-0j = γM00 + +%,* × =M01(>*%43%#14*#!""#$%&ℎj) ++%,* × =M02(>*%43%#14*#!5$67%4&*j) + ?M0j .M-1j = γM10 + ?M1j .M-2j = γM20 + ?M2j .M-3j = γM30 + ?M3j .M-4j = γM40 + ?M4j .M-5j = γM50 .M-6j = γM6 At Level 1, approach goals of female and male partners in couple j on a given day k are predicted by a partner-specific intercept level (βF-0j, βM-0j), by a time-lagged effect of their approach (βF-1j, βM-1j) and avoidance (βF-3j, βM-3j) goals, as well as by a time-lagged effect of the approach (βF-2j, βM-2j) and avoidance (βF-4j, βM-4j) goals from the corresponding partner. We also included a time-lagged effect of both couple members’ own (βF-5j, βM-5j) and their partner’s relationship satisfaction (βF-6j, βM-6j) as covariates. The time lag was specified as k-1. All predictor variables were person-mean centered. Finally, the terms eF-jk and eM-jk captured time- and partner-specific deviations from predicted levels of approach goals on a given day and they were allowed to covary between partners. INTERDEPENDENCE OF APPROACH AND AVOIDANCE GOALS 13 At Level 2, the estimated intercepts as well as the actor and partner effects of approach and avoidance goals were allowed to randomly vary between couples (uF0j through uM4j) from their sample-level fixed effects (γF00 through γM40). We included average partner levels of approach (γF01, γM01) and avoidance goals (γF02, γM02) as grand-mean centered predictors of the intercept to account for between-couple differences in the general tendency of partners to report higher or lower approach and avoidance goals than other partners. This procedure allowed us to disentangle the within-person and between-couple sources of partner effects on daily goal orientations. Gender differences in the estimated parameters were tested by including gender as an interaction term (Kashy & Donnellan, 2008). If partners did not significantly differ in their estimated effects, we report only one effect for both partners (obtained from the interaction model); if there was a significant interaction effect, we report separate estimates for female and male partners (obtained from the dual-intercept model). To explore our research questions, we tested the moderation effects of age and relationship duration by including them as time-invariant interacting variables for the short-term actor and partner effects of approach and avoidance goals. The moderating effects of age and relationship duration were tested in separate models. For the long-term effects, we followed a generalized least squares approach for distinguishable dyads. For these analyses, we exclusively relied on wave-specific mean scores obtained from the daily reports of T1 and T2. Equation 2 illustrates the model for the long-term prediction of both partners’ approach goals. An equivalent model was specified for the prediction of avoidance goals. Again, we followed a dual-intercept approach to obtain separate effects for female and male partners. According to this model, female and male partners’ average T2 levels of approach goals in couple j are predicted by an intercept estimate for female and male partners (βF0, βM0); as well as by their own approach (βF1, βM1) and avoidance (βF3, βM3) goals and by their partner’s approach (βF2, βM2) and avoidance (βF4, βM4) goals at T1. We accounted for INTERDEPENDENCE OF APPROACH AND AVOIDANCE GOALS 14 relational well-being by including couple members’ own (βF5, βM5) and their partner’s relationship satisfaction (βF5, βM5) at T1 as control variables. All predictor variables were grand-mean centered at the sample-level average of T1. Finally, the error terms eF-j and eM-j capture female and male partners’ deviations from predicted levels of approach goals at T2 and were allowed to covary between members of a dyad. As in the first set of analyses, we tested for gender differences by including gender as an interaction term and explored our research questions by including age and relationship duration as moderator effects. Separate estimates for female and male partners are only presented if the interaction effect of gender was significant. (2) !""#$&ℎ@AB = )*+%,*j × .F0 + +%,*j × .M0 + )*+%,*j × .F1C!&1$#!""#$%&ℎ@DjE ++%,*j × .M1C!&1$#!""#$%&ℎ@DjE +)*+%,*j × .F2C3%#14*#!""#$%&ℎ@DjE ++%,*j × .M2C3%#14*#!""#$%&ℎ@DjE +)*+%,*j × .F3C!&1$#!5$67%4&*@DjE ++%,*j × .M3C!&1$#!5$67%4&*@DjE +)*+%,*j × .F4C3%#14*#!5$67%4&*@DjE ++%,*j × .M4C3%#14*#!5$67%4&*@DjE +)*+%,*j × .F5C!&1$#8%169)%&16$4@DjE ++%,*j × .M5C!&1$#8%169)%&16$4@DjE +)*+%,*j × .F6C3%#14*#8%169)%&16$4@DjE ++%,*j × .M6C3%#14*#8%169)%&16$4@DjE + *F-j + *M-j The analyses were conducted in R version 3.6.1 (R Core Team, 2019) using the nlme (Pinheiro et al., 2019) and the dyadR (Garcia & Kenny, 2019) packages. INTERDEPENDENCE OF APPROACH AND AVOIDANCE GOALS 15 Results We present the results of the above described expanded analytical approach. The results of the preregistered analyses are displayed in Tables S6-S11. Regarding the primary hypotheses (H1 and H2), both analytical strategies point in the same direction. The expanded analyses, however, provide valuable additional insights, which is why they are presented in the main body of the paper. Table 1 reports the means, standard deviations, and zero-order correlations of the main study variables for female and male partners. Short-Term Effects Preliminary Analyses According to the intercept-only models, female and male partners’ daily reports of approach (eF = 1.08; eM = 1.00) and avoidance goals (eF = 0.79; eM = 0.85) substantially varied across the 14-day survey interval. Moreover, the intra-class correlations for female and male partners’ approach (ICCF = .55; ICCM = .56) and avoidance goals (ICCF = .64; ICCM = .62) indicate that a large percentage of the total variation in the daily reports is due to between-couple differences. Descriptive correlations for female and male partners’ daily approach goals, avoidance goals, and relationship satisfaction at T1 are presented in Table 2. The results of the multi-level models for testing the hypothesized short-term effects are presented in Table 3. Approach Relationship Goals (H1a) In the model to predict daily approach goals, we observed significant time-lagged actor effects of approach goals in both couple members (as hypothesized) and found a significant time-lagged actor effect of avoidance goals among men (not hypothesized): Couple members reported higher daily levels of approach goals if they had reported higher than (personal) average levels of approach goals on the previous day and, only among male partners, if they also reported higher than (personal) average levels of avoidance goals on the INTERDEPENDENCE OF APPROACH AND AVOIDANCE GOALS 16 previous day. Although we did not find the hypothesized time-lagged partner effects of approach goals, we found a significant positive partner effect for mean levels of approach goals: Couple members reported higher daily levels of approach goals if they had a partner who generally reported higher than (sample) average levels of approach goals. All effects appeared irrespective of the inclusion of relationship satisfaction as a covariate. Avoidance Relationship Goals (H1b) In the model to predict daily avoidance goals, we observed (as hypothesized) significant time-lagged actor effects of avoidance but not approach goals. Contrary to our hypothesis, we did not identify any significant time-lagged partner effects of avoidance goals. However, we found significant positive partner effects for mean levels of approach and avoidance goals. That is, couple members reported higher levels of daily avoidance goals if they also reported higher than (personal) average levels of avoidance goals on the previous day and if they had a partner who reported higher than (sample) average levels of approach and avoidance goals. These effects appeared irrespective of relationship satisfaction. Long-Term Effects Preliminary Analyses Descriptive changes in approach goals, avoidance goals, and relationship satisfaction between T1 and T2 as well as correlated changes in these variables within and between partners can be found in Table 4. The results of the generalized least squares models for testing the hypothesized long-term effects are presented in Table 5. Approach relationship goals (H2a) As hypothesized, we found significant actor and partner effects of approach goals but no significant actor and partner effects of avoidance goals. That is, couple members reported higher levels of approach goals at T2 if they and their partner reported higher than (sample) average levels of approach goals at T1. These effects appeared irrespective of relationship satisfaction. INTERDEPENDENCE OF APPROACH AND AVOIDANCE GOALS 17 Avoidance Relationship Goals (H2b) We observed a significant actor effect of avoidance but not approach goals (as hypothesized), but we did not find a significant partner effect of avoidance goals. That is, couple members reported higher levels of avoidance goals at T2 if they reported higher than (sample) average levels of avoidance goals at T1 but not when their partners reported higher than (sample) average levels of avoidance goals. We did not identify any significant effects of approach goals. These effects appeared irrespective of relationship satisfaction. Moderation of Age and Relationship Duration We tested whether the actor and partner effects of approach and avoidance goals on the short- and long-term scale were moderated by age and relationship duration. The results from these analyses are displayed in Tables S2-S5. Age Of the 20 age moderations that were tested (Tables S2 and S3), only two were significant. Both moderation effects were positive and they were observed for mean partner avoidance goals. In the prediction of daily avoidance goals, age amplified the positive partner effect of mean avoidance in both couple members (γ02 x Age = 0.01, t[6073] = 2.25, p = .025), while in the prediction of approach goals at T2, age buffered the negative partner effect of mean avoidance among women (βF4 x Age = 0.01, t[906] = 2.04, p = .041). Relationship Duration Of the 20 relationship-duration moderations that were tested (Tables S4 and S5), only one was significant. This moderation effect was positive and it emerged in the prediction of daily avoidance goals, suggesting that relationship duration amplified the positive partner effect of mean avoidance goals (γ02 x Relationship Duration = 0.01, t[6032] = 2.34, p = .019). Discussion This study tested short- and long-term interdependence of approach and avoidance goals in romantic couples across adulthood. The study supported the majority of our INTERDEPENDENCE OF APPROACH AND AVOIDANCE GOALS 18 preregistered hypotheses (11 of 16 predicted associations) and provides five main findings. (1) We found consistent actor effects at the state and trait level. (2) The majority of these effects emerged within a goal orientation (i.e., within approach or avoidance goals). (3) Partner effects also occurred but only at the trait level. (4) Almost all of the tested associations were independent of age and relationship duration. (5) The findings were largely independent of relationship satisfaction, ruling out that approach and avoidance goals merely mirror relational well-being. We discuss these findings and their implications against the background of our initial questions. Do a Person’s Goals Depend on the Goals of their Partner? We draw two conclusions in answer to this question: First, it seems that a person’s approach and avoidance goals more strongly reflect an actor effect than a partner effect; second, partner effects might be more influential at the trait rather than the state level. We found consistent short- and long-term actor effects of approach and avoidance goals. In other words, goals not only predicted goals on the subsequent day but aggregated daily levels predicted goals 10-12 months later. These findings support the idea that development in individual characteristics results from repeated state experiences that develop into consistent patterns of thoughts, behaviors, and feelings (e.g., Wrzus & Roberts, 2017). Approach and avoidance goals are associated with specific information processing (see Nikitin & Schoch, in press) that influence how people integrate experiences in their relationship into global perceptions of the relationship (Gable & Poore, 2008). As expected, we found that these processes occurred mainly within a goal orientation (in contrast to between goal orientations), supporting and extending previous evidence that approach and avoidance goals are part of two distinct motivational systems that operate in parallel (Gable et al., 2003). There was one unexpected short-term association between approach and avoidance goals: Approach goals were positively associated with previous-day avoidance goals (only for INTERDEPENDENCE OF APPROACH AND AVOIDANCE GOALS 19 men). It is possible that stronger avoidance goals signal stronger threats to the relationship, which might lead to attempts to reconnect with the partner (social reconnection hypothesis; Maner et al., 2007). Supporting this explanation, Nikitin and Freund (2019a) found that in potentially difficult social situations, both approach and avoidance motivation were positively associated with the orientation toward the partner. The question remains why in the present study this was the case only for men. Rather than speculate about possible gender differences, we encourage future research to test the robustness of this finding. Does Goal Interdependence Occur in the Short and Long Term? The present findings suggest that goal interdependence occurs in the short and long term but appears to be limited to the partner’s trait level. Specifically, actors’ daily approach goals were predicted by their partners’ trait-level but not state-level approach goals and actors’ daily avoidance goals were predicted by their partners’ trait-level but not state-level avoidance goals (there was also a positive but weak effect of partners’ trait-level approach goals on actor’s daily avoidance goals). The finding that daily fluctuations in approach and avoidance goals were not transmitted to the partner is surprising given that approach and avoidance goals of one partner have been found to have consequences for the emotions and relationship satisfaction of the other partner (Impett et al., 2010). We see at least two explanations why higher-than-usual levels of one person’s goals did not predict their partner’s goals on the next day: (1) When reporting their own goals, partners might rely on knowledge about their partners’ general motivational tendencies rather than on perceptions of the daily fluctuations of their partners’ goals (for a similar explanation of partners’ emotional empathy, see Rauers et al., 2013); (2) it is possible that motivational attempts of one partner need to accumulate over longer periods of time before their effects on the other partner’s daily and long-term motivations manifest. Alternatively, it is possible that partners’ goals affect each other across minutes or hours rather than days. Although it has been shown that coregulation processes can be observed on different time scales (e.g., minutes, days, and weeks; for a INTERDEPENDENCE OF APPROACH AND AVOIDANCE GOALS 20 discussion see Hudson et al., 2014), studies assessing goals several times a day are needed to test the possibility that goal interdependence between partners is an immediate process. In the long term, and in line with our hypotheses, approach goals were positively associated between partners. The association between partners’ approach goals as found in the present study might be a long-term mechanism of positive partnership development, namely the accumulation of positive outcomes of a person’s approach goals that leads to approach goals in their partner 10-12 months later. Supporting this interpretation, one study has shown that partners of individuals with relatively high approach goals became more committed to and more satisfied with their relationship over time (Impett et al., 2010). It is an open question why partners’ avoidance goals were not associated in the long term. One possible interpretation is that avoidance goals more strongly reflect individual dispositions than interpersonal processes. This is in line with results from attachment research that suggest attachment anxiety (which is positively associated with avoidance goals; see Nikitin & Freund, 2010) might be the result of negative self-relevant (but not others-relevant) working models (e.g., Hudson et al., 2014). Does Goal Interdependence Manifest Differently with Age and Relationship Duration? The current study found no consistent evidence regarding interaction effects of age or relationship duration on the interdependence of approach and avoidance goals between partners (we refrain from discussing the few significant effects as they might reflect a Type I error). Although approach and avoidance goals have outcomes that vary by age (Nikitin & Freund, 2018, 2019b), the current findings suggest that the interdependence of approach and avoidance goals within relationships might be independent of age and relationship duration (see also Kuster et al., 2017). It is therefore an interesting question for future research to explore why the interdependence of approach and avoidance goals within a couple occurs across age and relationship duration. INTERDEPENDENCE OF APPROACH AND AVOIDANCE GOALS 21 Strengths and Limitations Although the present study has clear strengths such as the longitudinal design combined with the daily assessment, the inclusion of both couple members in the analyses, and a heterogeneous sample with respect to sociodemographic background, it has also several limitations. We used an aggregated state measure of goals as an approximation of a trait. Although this approach is widespread (for a summary, see Rauthmann et al., 2019), there is also doubt that aggregates capture the same construct as a trait measure. In fact, Rauthmann and colleagues (2019) demonstrated that the two measures can differ in their nomological validity. Thus, the present results should be interpreted with caution. Similarly, the aggregated measure is only an approximation of the accumulation hypothesis (i.e., that repeated state experiences develop into consistent patterns of thoughts, behaviors, and feelings; Wrzus & Roberts, 2017) and it did not include any goal-related processes or actual behavior that might contribute to goal development. In addition, studying accumulation processes might need longer time periods than 14 days. Our research is a starting point for the study of goal interdependence, and we hope that it will stimulate future research focusing on specific underlying processes. Although the sample was age-heterogeneous, most of the participants were relatively young. Half of the participants were younger than 29 years and in their current relationship for less than six years. The lack of age and relationship-duration effects might at least partly be explained by the uneven age distribution. Relatedly, given that this study was internet based, participants were required to have a certain level of computer expertise, which might have led to a sample selectivity. This might reduce the generalizability of the present findings. Finally, we did not investigate any implications of the interdependence of approach and avoidance goals between partners. Although we could rule out that approach and avoidance relationship goals merely mirror relationship satisfaction, it is an open question INTERDEPENDENCE OF APPROACH AND AVOIDANCE GOALS 22 whether (a lack of) interdependence in goals affects relationship outcomes. For example, Impett et al. (2010) found that both partners need to be equally committed to pursuing approach relationship goals to succeed and thrive over time. Conclusion The present research demonstrates that actor effects are, in general, more pronounced than partner effects in the prediction of goals: We found all hypothesized (and some not hypothesized) actor effects. Partner effects were less frequent and appeared more on the trait- than state-level. Thus, although goals are driven by people’s close relationships (Fitzsimons et al., 2015), it seems that state-level goal orientations of one person do not necessarily evoke the same goal orientation in the partner. Instead, the interdependence of goal orientations might reflect accumulation processes in a relationship across time. More research is needed to explore the differential mechanisms and implications of the short-term and long-term goal interdependencies of romantic partners. INTERDEPENDENCE OF APPROACH AND AVOIDANCE GOALS 23 References Ackermann, R., & Kenny, D. A. (2016). APIMPower: An interactive tool for Actor-Partner Interdependence Model power analysis [Computer software]. https://robert-a-ackerman.shinyapps.io/apimpower/. Bolger, N., & Laurenceau, J.-P. (2013). Intensive longitudinal methods: An introduction to diary and experience sampling research. Guilford Press. Bühler, J. L., Weidmann, R., Kumashiro, M., & Grob, A. (2019). Does Michelangelo care about age? An adult life-span perspective on the Michelangelo phenomenon. Journal of Social and Personal Relationships, 36, 1392–1412. https://doi.org/10.1177/0265407518766698 Bühler, J. L., Weidmann, R., Wünsche, J., Burriss, R. P., & Grob, A. (2020). Daily responsiveness, expectations, and self-disclosure: How the average levels and within-person variability of three relationship components mediate personality-relationship transactions in romantic couples. European Journal of Personality. https://doi.org/10.1002/per.2255 Carstensen, L. L., Isaacowitz, D. M., & Charles, S. T. (1999). Taking time seriously: A theory of socioemotional selectivity. American Psychologist, 54, 165–181. https://doi.org/10.1037/0003-066X.54.3.165 Caspi, A., & Roberts, B. W. (2001). Personality development across the life course:The argument for change and continuity. Psychological Inquiry, 12(2), 49–66. https://doi.org/10.1207/S15327965PLI1202_01 Elliot, A. J., Gable, S. L., & Mapes, R. R. (2006). Approach and avoidance motivation in the social domain. Personality and Social Psychology Bulletin, 32, 378–391. https://doi.org/10.1177/0146167205282153 Finkel, E. J., Hui, C. M., Carswell, K. L., & Larson, G. M. (2014). The suffocation of marriage: Climbing mount Maslow without enough oxygen. Psychological Inquiry, 25, INTERDEPENDENCE OF APPROACH AND AVOIDANCE GOALS 24 1–41. https://doi.org/10.1080/1047840X.2014.863723 Fitzsimons, G. M., Finkel, E. J., & VanDellen, M. R. (2015). Transactive goal dynamics. Psychological Review, 122, 648–673. https://doi.org/10.1037/a0039654 Gable, S. L., & Impett, E. A. (2012). Approach and avoidance motives and close relationships. Social and Personality Psychology Compass, 6, 95–108. https://doi.org/10.1111/j.1751-9004.2011.00405.x Gable, S. L., & Poore, J. (2008). Which thoughts count? Algorithms for evaluating satisfaction in relationships. Psychological Science, 19, 1030–1036. https://doi.org/10.1111/j.1467-9280.2008.02195.x Gable, S. L., Reis, H. T., & Elliot, A. J. (2003). Evidence for bivariate systems: An empirical test of appetition and aversion across domains. Journal of Research in Personality, 37, 349–372. https://doi.org/10.1016/S0092-6566(02)00580-9 Garcia, R. L., & Kenny, D. A. (2019). dyadr: Dyadic Data Analysis (version 0.0.0.9000). Hudson, N. W., Fraley, R. C., Brumbaugh, C. C., & Vicary, A. M. (2014). Coregulation in romantic partners’ attachment styles. Personality and Social Psychology Bulletin, 40, 845–857. https://doi.org/10.1177/0146167214528989 Impett, E. A., Gordon, A. M. M., Kogan, A., Oveis, C., Gable, S. L., & Keltner, D. (2010). Moving toward more perfect unions: Daily and long-term consquences of approach and avoidance goals in romantic relationships. Journal of Personality and Social Psychology, 99, 948–963. https://doi.org/10.1037/a0020271 Impett, E. A., Strachman, A., Finkel, E. J., & Gable, S. L. (2008). Maintaining sexual desire in intimate relationships: The importance of approach goals. Journal of Personality and Social Psychology, 94, 808–823. https://doi.org/10.1037/0022-3514.94.5.808 Kashy, D. A., & Donnellan, M. B. (2008). Comparing MLM and SEM approaches to analyzing developmental dyadic data: Growth curve models of hostility in families. In N. A. Card, J. P. Selig, & T. D. Little (Eds.), Modeling dyadic and interdependent data in INTERDEPENDENCE OF APPROACH AND AVOIDANCE GOALS 25 the developmental and behavioral sciences (pp. 165–190). Routledge. https://doi.org/10.4324/9780203892763-12 Kurdek, L. A. (1998). The nature and predictors of the trajectory of change in marital quality over the first 4 years of marriage for first-married husbands and wives. Journal of Family Psychology, 12, 494–510. https://doi.org/10.1037/0893-3200.12.4.494 Kuster, M., Backes, S., Brandstätter, V., Nussbeck, F. W., Bradbury, T. N., Sutter-Stickel, D., & Bodenmann, G. (2017). Approach-avoidance goals and relationship problems, communication of stress, and dyadic coping in couples. Motivation and Emotion, 41, 576–590. https://doi.org/10.1007/s11031-017-9629-3 LaBuda, J. E., Gere, J., & Impett, E. A. (2019). Perceptions of a romantic partner’s approach and avoidance motives: Accuracy, bias, and emotional cues. Journal of Personality and Social Psychology, Advance online publication. https://doi.org/10.1037/pspp0000256 Lane, S. P., & Hennes, E. P. (2018). Power struggles: Estimating sample size for multilevel relationships research. Journal of Social and Personal Relationships, 35, 7–31. https://doi.org/10.1177/0265407517710342 Laurenceau, J. P., Kleinman, B. M., Kaczynski, K. J., & Carver, C. S. (2010). Assessment of relationship-specific incentive and threat sensitivities: Predicting satisfaction and affect in adult intimate relationships. Psychological Assessment, 22(2), 407–419. https://doi.org/10.1037/a0019231 Li, T., & Fung, H. H. (2011). The dynamic goal theory of marital satisfaction. Review of General Psychology, 15, 246–254. https://doi.org/10.1037/a0024694 Little, B. (1989). Personal projects analysis: Trivial pursuits, magnificent obsessions, and the search for coherence. In D. M. Buss & N. Cantor (Eds.), Personality psychology: Recent trends and emerging directions (pp. 15–31). Springer. Maner, J. K., DeWall, C. N., Baumeister, R. F., & Schaller, M. (2007). Does social exclusion motivate interpersonal reconnection? Resolving the “porcupine problem.” Journal of INTERDEPENDENCE OF APPROACH AND AVOIDANCE GOALS 26 Personality and Social Psychology, 92, 42–55. https://doi.org/10.1037/0022-3514.92.1.42 McAdams, D. P., & Olson, B. D. (2010). Personality development: Continuity and change over the life course. Annual Review of Psychology, 61(1), 517–542. https://doi.org/10.1146/annurev.psych.093008.100507 Neyer, F. J., Mund, M., Zimmermann, J., & Wrzus, C. (2014). Personality-relationship transactions revisited. Journal of Personality, 82(6), 539–550. https://doi.org/10.1111/jopy.12063 Nikitin, J., & Freund, A. M. (2010). When wanting and fearing go together: The effect of co-occurring social approach and avoidance motivation on behavior, affect, and cognition. European Journal of Social Psychology, 40, 783–804. https://doi.org/10.1002/ejsp.650 Nikitin, J., & Freund, A. M. (2018). Feeling loved and integrated or lonely and rejected in everyday life: The role of age and social motivation. Developmental Psychology, 54, 1186–1198. https://doi.org/10.1037/dev0000502 Nikitin, J., & Freund, A. M. (2019a). Who cares? Age-differential effects of social approach and avoidance motivation on responsiveness to others. Personality and Social Psychology Bulletin, 45, 182–195. https://doi.org/10.1177/0146167218781335 Nikitin, J., & Freund, A. M. (2019b). Individual differences in habitual social goals and daily well-being: The role of age and relationship closeness. European Journal of Personality, 33, 337–358. https://doi.org/10.1002/per.2190 Nikitin, J., Gong, X., Schoch, S., & Freund, A. M. (2019). Social motives, attributions and expectations as predictors of the decision to participate in a speed-dating event. Motivation and Emotion, 43, 610–624. https://doi.org/10.1007/s11031-019-09762-0 Nikitin, J., & Schoch, S. (2020). Social approach and avoidance motivations. In R. J. Coplan, J. C. Bowker, & L. J. Nelson (Eds.), The handbook of solitude: Psychological perspectives on social isolation, social withdrawal, and being alone (2nd ed., pp. 202–INTERDEPENDENCE OF APPROACH AND AVOIDANCE GOALS 27 223). Wiley-Blackwell. https://doi.org/10.1002/9781118427378 Orth, U., Erol, R. Y., Ledermann, T., & Grob, A. (2018). Codevelopment of well-being and self-esteem in romantic partners: Disentangling the effects of mutual influence and shared environment. Developmental Psychology, 54, 151–166. https://doi.org/10.1037/dev0000400 Pinheiro, J., Bates, D., DebRoy, S., Sarkar, D., & Team, R. C. (2019). nlme: Linear and Nonlinear Mixed Effects Models (R package version 3.1-141). https://cran.r-project.org/package=nlme R Core Team (Institution/Organization). (2019). R: A Language and environment for statistical computing. https://www.r-project.org Rauers, A., Blanke, E., & Riediger, M. (2013). Everyday empathic accuracy in younger and older couples: Do you need to see your partner to know his or her feelings? Psychological Science, 24, 2210–2217. https://doi.org/10.1177/0956797613490747 Rauthmann, J. F., Horstmann, K. T., & Sherman, R. A. (2019). Do self-reported traits and aggregated states capture the same thing? A nomological perspective on trait-state homomorphy. Social Psychological and Personality Science, 10, 596–611. https://doi.org/10.1177/1948550618774772 Roberts, B. W., & DelVecchio, W. F. (2000). The rank-order consistency of personality traits from childhood to old age: a quantitative review of longitudinal studies. Psychological Bulletin, 126, 3–25. https://doi.org/10.1037//0033-2909.126.1.3 Roberts, B. W., Walton, K. E., & Viechtbauer, W. (2006). Patterns of mean-level change in personality traits across the life course: A meta-analysis of longitudinal studies. Psychological Bulletin, 132(1), 1–25. https://doi.org/10.1037/0033-2909.132.1.1 Sbarra, D. A., & Hazan, C. (2008). Coregulation, dysregulation, self-regulation: An integrative analysis and empirical agenda for understanding adult attachment, separation, loss, and recovery. Personality and Social Psychology Review : An Official Journal of INTERDEPENDENCE OF APPROACH AND AVOIDANCE GOALS 28 the Society for Personality and Social Psychology, Inc, 12(2), 141–167. https://doi.org/10.1177/1088868308315702 Schoch, S., Nikitin, J., & Freund, A. (2015). Why do(n’t) you like me? The role of social approach and avoidance motives in attributions following social acceptance and rejection. Motivation and Emotion, 39, 1–13. https://doi.org/10.1007/s11031-015-9482-1 Scholer, A. A., Cornwell, J. F. M., & Higgins, E. T. (2019). Should we approach approach and avoid avoidance? An inquiry from different levels. Psychological Inquiry, 30, 111–124. https://doi.org/10.1080/1047840X.2019.1643667 Thorson, K. R., West, T. V., & Mendes, W. B. (2018). Measuring physiological influence in dyads: A guide to designing, implementing, and analyzing dyadic physiological studies. Psychological Methods, 23, 595–616. https://doi.org/10.1037/met0000166 Wagner, J., Orth, U., Bleidorn, W., Hopwood, C. J., & Kandler, C. (n.d.). Towards an integrative model of sources of personality stability and change. Current Directions in Psychological Science. Wrzus, C., & Roberts, B. W. (2017). Processes of personality development in adulthood: The TESSERA framework. Personality and Social Psychology Review, 21, 253–277. https://doi.org/10.1177/1088868316652279 INTERDEPENDENCE OF APPROACH AND AVOIDANCE GOALS 29 Footnotes 1 We used the online tool APIMpower (Ackermann & Kenny, 2016) to conduct a post-hoc power analysis for the hypothesized effects. Assuming strong actor effects (standardized estimate = .50) and small partner effects (standardized estimate = .15; e.g., Orth et al., 2018), the power to detect actor effects was >.99 in both analysis samples. The power to detect partner effects was .94 for short-term effects and .95 for long-term-effects. The APIMPower tool was designed to detect the power of actor and partner effects within a dyadic structural equation framework. To our knowledge, there is currently no open-source software to run power analyses for dyadic multi-level models. For complex simulation-based approaches in Mplus and SAS see Lane and Hennes (2018) and Thorson, West, and Mendes (2018). 2 Regarding the primary study variables (i.e., age, relationship duration, approach goals, and avoidance goals), couples that were included in both analysis samples (i.e., short-term and long-term sample) did not significantly differ from couples that were only part of one of these samples. Specifically, the differences were as follows: age, t(65) = 0.11, p = .914; relationship duration t(62) = 0.04, p = .965; approach goals, t(68) = 1.23, p = .223; and avoidance goals, t(67) = -1.21, p = .232. We found, however, that couples who were only part of one of the two samples were significantly less satisfied with their relationship, t(63) = 3.19, p = .002. 3 Based on recommendations by Bolger and Laurenceau (2013), we did not follow a three-level approach (i.e., observations nested within persons nested within couples) for two main reasons. First, there would be no random variation at the person level because the role within the dyad (i.e., male vs. female) was already included as a variable in the model itself. Second, couple members are commonly measured at the same (or at least a similar) time point. Hence, the levels of time and person were cross-qualified within dyads. INTERDEPENDENCE OF APPROACH AND AVOIDANCE GOALS 30 Funding The data analyzed here originate from the Processes in Romantic Relationships and Their Impact on Relationship and Personal Outcomes (CouPers) study, a two-year longitudinal online study of romantic couples. The study was conducted at the University of Basel, Switzerland, between 2016 and 2018 and was funded by the Swiss National Science Foundation (SNSF 100019_162697). Acknowledgments The authors thank Sabrina Brunner, Fabienne Fend, and Rahel Hütten for research assistance. A minimal dataset including the data underlying the findings described here is available upon request from FORSbase, the archive of the Swiss Centre of Expertise in the Social Sciences. INTERDEPENDENCE OF APPROACH AND AVOIDANCE GOALS 31 Table 1 Means, Standard Deviations, and Correlations of the Study Variables for Female and Male Partners Female partner Male partner Correlation Variables M SD M SD Cohen’s d 1 2 3 4 5 6 7 8 1. Approach goals T1 3.76 1.32 3.79 1.29 -.02 .43 .67 .30 .74 .48 .27 .04 .00 2. Avoidance goals T1 2.78 1.41 3.12 1.38 -.24 .66 .41 -.01 .49 .72 -.03 -.02 .00 3. Rel. satisfaction T1 4.26 0.58 4.22 0.59 .07 .24 -.03 .57 .24 -.05 .70 .04 .03 4. Approach goals T2 3.76 1.28 3.86 1.33 -.08 .71 .48 .19 .43 .62 .35 .00 -.04 5. Avoidance goals T2 2.82 1.47 3.08 1.44 -.18 .54 .71 -.02 .67 .42 -.04 .02 .01 6. Rel. satisfaction T2 4.20 0.61 4.18 0.62 .03 .17 -.05 .65 .22 -.05 .58 .00 .00 7. Age T1 32.79 13.63 34.38 13.98 -.12 -.01 .01 .04 -.06 .01 -.03 .96 .79 8. Rel. duration T1 9.60 10.73 9.60 10.73 0 -.04 .00 -.01 -.09 -.04 -.01 .80 1 Note. N = 463 couples. Rel. = Relationship. The presented values for approach goals, avoidance goals, and relationship satisfaction refer to both partners’ wave-specific average scores at T1 and T2 (i.e., across the 14-day survey interval). Relationship duration represents the mean report across both couple members at T1. Hence, the values are identical between female and male partners. Correlations above the diagonal INTERDEPENDENCE OF APPROACH AND AVOIDANCE GOALS 32 belong to male partners; correlations below the diagonal belong to female partners. Correlations on the diagonal (gray) are between-partner correlations. Cohen’s d estimates indicate standardized differences between female and male partners and they were calculated for dependent samples; positive estimates indicate higher mean levels among female partners; negative estimates indicate higher mean levels among male partners. Correlations and Cohen’s ds in bold are significant (p < .05). INTERDEPENDENCE OF APPROACH AND AVOIDANCE GOALS 33 Table 2 Descriptive Correlations for Female and Male Partners’ Daily Approach Goals, Avoidance Goals, and Relationship Satisfaction Across the 14-Day Survey Interval of T1 Raw scores Person-mean centered scores Variables 1 2 3 1 2 3 1. Daily approach goals .37 .61 .28 .25 .47 .23 2. Daily avoidance goals .58 .32 .04 .41 .08 .07 3. Daily rel. satisfaction .27 .03 .36 .31 .12 .21 Note. N = 441 couples. Rel. = Relationship. Correlations above the diagonal belong to male partners; correlations below the diagonal belong to female partners. Correlations on the diagonal (gray) are between-partner correlations. Correlations in bold are significant (p < .05). INTERDEPENDENCE OF APPROACH AND AVOIDANCE GOALS 34 Table 3 Estimated Parameters for the Models to Predict Daily Approach and Avoidance Goals Daily approach goals Daily avoidance goals Fixed effects Estimate 95% CI p Estimate 95% CI p Interceptfemale 3.75 [3.67, 3.84] <.001 2.68 [2.55, 2.81] <.001 Interceptmale 3.75 [3.67, 3.84] <.001 3.13 [3.00, 3.25] <.001 Lagged actor approach 0.12 [0.08, 0.15] <.001 0.00 [-0.03, 0.03] .981 Lagged partner approach 0.01 [-0.02, 0.04] .386 0.01 [-0.01, 0.04] .212 Lagged actor avoidancefemale 0.02 [-0.04, 0.07] .617 0.12 [0.08, 0.15] <.001 Lagged actor avoidancemale 0.10 [0.05, 0.16] <.001 0.12 [0.08, 0.15] <.001 Lagged partner avoidance 0.02 [-0.02, 0.06] .236 0.02 [-0.01, 0.04] .298 Lagged actor satisfaction -0.02 [-0.06, 0.02] .248 -0.02 [-0.05, 0.01] .212 Lagged partner satisfaction 0.02 [-0.02, 0.05] .416 -0.03 [-0.06, -0.01] .023 Mean partner approach 0.38 [0.30, 0.47] <.001 0.09 [0.001, 0.18] .049 Mean partner avoidance 0.06 [-0.02, 0.14] .160 0.36 [0.27, 0.44] <.001 Gender x Intercept -0.04 [-0.12, 0.05] .393 -0.23 [-0.32, -0.14] <.001 INTERDEPENDENCE OF APPROACH AND AVOIDANCE GOALS 35 Table 3 (continued) Daily approach goals Daily avoidance goals Fixed effects Estimate 95% CI p Estimate 95% CI p Gender x Lagged actor approach 0.02 [-0.01, 0.06] .199 0.01 [-0.01, 0.04] .385 Gender x Lagged partner approach 0.02 [-0.01, 0.05] .244 0.00 [-0.02, 0.03] .894 Gender x Lagged actor avoidance -0.04 [-0.08, -0.004] .033 -0.03 [-0.06, 0.01] .150 Gender x Lagged partner avoidance -0.02 [-0.06, 0.02] .257 -0.02 [-0.05, 0.01] .116 Gender x Lagged actor satisfaction -0.04 [-0.08, 0.001] .056 -0.01 [-0.04, 0.02] .582 Gender x Lagged partner satisfaction 0.02 [-0.02, 0.06] .432 0.03 [-0.002, 0.06] .064 Gender x Mean partner approach -0.05 [-0.13, 0.04] .300 -0.07 [-0.16, 0.02] .112 Gender x Mean partner avoidance 0.06 [-0.02, 0.14] .130 0.06 [-0.03, 0.14] .174 Random effects Female Male Female Male Intercept 1.21 1.18 1.33 1.28 Lagged actor approach 0.21 0.22 0.13 0.14 Lagged partner approach 0.00 0.05 0.00 0.04 Lagged actor avoidance 0.20 0.19 0.25 0.22 INTERDEPENDENCE OF APPROACH AND AVOIDANCE GOALS 36 Table 3 (continued) Random effects Female Male Female Male Lagged partner avoidance 0.17 0.22 0.15 0.07 Residual between-partner correlation .28 .11 Note. N = 441 couples. CI = confidence interval. We report unstandardized estimates. Estimates with p-values in bold are significant (p < .05). Actor and partner approach goals, avoidance goals, and relationship satisfaction were included as lagged person-mean centered predictor variables. Mean levels of actor and partner approach and avoidance goals were included as grand-mean centered predictor variables. Gender was effect-coded (1 = female, -1 = male). We report the same/only one fixed effect estimate for both couple members if no significant interaction effect of gender emerged. INTERDEPENDENCE OF APPROACH AND AVOIDANCE GOALS 37 Table 4 Descriptive Changes and Change Correlations for Female and Male Partners’ Approach Goals, Avoidance Goals, and Relationship Satisfaction Between T1 and T2 Change(T1 à T2) Female partner Change(T1 à T2) Male partner Change(T1 à T2) Correlation Variables M SD Cohen’s d M SD Cohen’s d 1 2 3 1. Approach goals 0.00 0.99 0.00 0.06 0.95 0.05 .29 .61 .24 2. Avoidance goals 0.04 1.10 0.03 -0.05 1.05 -0.04 .55 .22 .05 3. Rel. satisfaction -0.06 0.50 -0.10 -0.04 0.50 -0.07 .17 -.01 .40 Note. N = 456 couples. Rel. = Relationship. The presented changes in approach goals, avoidance goals, and relationship satisfaction refer to differences between a person’s wave-specific average score at T1 and T2 (i.e., across the 14-day survey interval). T1 and T2 were separated by 10–12 months. Change correlations above the diagonal belong to male partners; change correlations below the diagonal belong to female partners. Change correlations on the diagonal (gray) are between-partner change correlations. Cohen’s ds were calculated for dependent samples. Change correlations and Cohen’s ds in bold are significant (p < .05). INTERDEPENDENCE OF APPROACH AND AVOIDANCE GOALS 38 Table 5 Estimated Parameters for the Models to Predict Approach and Avoidance Goals Across Months Approach goals T2 Avoidance goals T2 Effects Estimate 95% CI p Estimate 95% CI p Intercept 3.82 [3.75, 3.89] <.001 2.94 [2.87, 3.02] <.001 Actor approach T1 0.70 [0.63, 0.76] <.001 0.05 [-0.02, 0.13] .182 Partner approach T1 0.07 [0.01, 0.14] .027 0.05 [-0.02, 0.13] .156 Actor avoidance T1 0.01 [-0.05, 0.07] .728 0.69 [0.62, 0.75] <.001 Partner avoidance T1 -0.03 [-0.09 0.03] .257 0.04 [-0.03, 0.10] .289 Actor satisfaction T1 0.03 [-0.08, 0.15] .572 -0.09 [-0.22, 0.04] .183 Partner satisfaction T1 -0.02 [-0.14, 0.09] .704 -0.02 [-0.15, 0.11] .777 Gender x Intercept T1 -0.03 [-0.08, 0.02] .207 -0.02 [-0.08, 0.04] .568 Gender x Actor approach T1 -0.04 [-0.11, 0.03] .230 0.08 [-0.002, 0.16] .057 Gender x Partner approach T1 0.02 [-0.05, 0.09] .626 -0.01 [-0.09, 0.07] .788 Gender x Actor avoidance T1 0.01 [-0.05, 0.08] .705 -0.06 [-0.13, 0.01] .106 Gender x Partner avoidance T1 -0.01 [-0.08, 0.05] .739 0.02 [-0.05, 0.09] .607 INTERDEPENDENCE OF APPROACH AND AVOIDANCE GOALS 39 Table 5 (continued) Approach goals T2 Avoidance goals T2 Effects Estimate 95% CI p Estimate 95% CI p Gender x Actor satisfaction T1 -0.04 [-0.18, 0.10] .561 0.02 [-0.14, 0.17] .824 Gender x Partner satisfaction T1 0.03 [-0.10, 0.17] .638 -0.01 [-0.16, 0.14] .911 Residual between-partner correlation .32 .26 Note. N = 456 couples. CI = confidence interval. T1 and T2 were separated by 10–12 months. We report unstandardized estimates. Estimates with p-values in bold are significant (p < .05). Actor and partner approach goals, avoidance goals, and relationship satisfaction were included as grand-mean centered predictor variables. Gender was effect-coded (1 = female, -1 = male). We report the same/only one fixed effect estimate for both couple members if no significant interaction effect of gender emerged. View publication stats | RESEARCH GATE |
DOI: http://dx.doi.org/10.7769/gesec.v14i1.1513 ISSN: 2178-9010 Revista GeSec São Paulo, SP, Brasil v. 14, n. 1, p. 277-305,2023 Proposições para o Portal da Transparência do Poder Executivo do Estado de Santa Catarina com foco na área da saúde Proposals for the Transparency Portal of the Executive Power of the State of Santa Catarina focusing on the health area Gilceu Ferreira1 Fabiano Maury Raupp2 Resumo O objetivo deste artigo consiste em propor diretrizes para o aperfeiçoamento da transparência de dados e informações públicas estaduais, com foco na área de Saúde, por meio do Portal da Transparência do Poder Executivo de Santa Catarina. Trata-se de uma pesquisa descritiva e qualitativa. Um modelo de análise foi desenvolvido, baseado em atributos da qualidade da informação, para coletar, diagnosticar e propor conteúdos. Foram considerados os dados dos setores de comunicação e ouvidoria da Secretaria Estadual de Saúde, dos 57 pedidos de informação da Lei de Acesso à Informação (LAI), do ano de 2019 da área da Saúde, e dos 26 portais da transparência das Unidades Federativas (UF). Utilizando a análise documental e de conteúdo, aliado a técnica de triangulação dos dados coletados, foram geradas propostas de melhorias para o aspecto informacional da área da Saúde. A primeira proposta foi o desenvolvimento de uma área temática por meio de páginas contendo informações da Saúde. A segunda recomenda a convergência dos conteúdos existentes a partir de mecanismos de seleção, como filtros de consultas do Portal, para compor a área temática. A terceira propõe 38 conteúdos para serem incorporados na área temática proposta. A proposta de organizar e 1 Mestre em Administração pela Universidade do Estado de Santa Catarina, Av. Madre Benvenuta, 2037, Itacurubi, CEP: 88035-001, Florianópolis - SC. E-mail: gilceuf@gmail.com Orcid: https://orcid.org/0000-0001-9761-3193 2 Doutor em Administração pela Universidade Federal da Bahia, Av. Madre Benvenuta, 2037, Itacurubi, CEP: 88035-001, Florianópolis - SC. E-mail:fabianoraupp@hotmail.com Orcid: https://orcid.org/0000-0001-9533-2574 Proposições para o Portal da Transparência do Poder Executivo do Estado de Santa Catarina com foco na área da saúde Revista Gestão e Secretariado (GeSec), São Paulo, SP, v. 14, n. 1, 2023, p. 277-305. 278 incorporar conteúdos poderá aperfeiçoar a transparência da área da Saúde, indicando prioridades relacionadas à prestação de serviços. Ao serem implementadas poderão trazer facilidades no acesso, convergência e disponibilidade de conteúdos na Internet para atender as demandas informacionais dos cidadãos e facilitar o exercício do controle social. Palavras-chave: Transparência Pública. Saúde. Qualidade da Informação. Website. Abstract This purpose of thir article is propose guidelines for improving the transparency of state public data and information focusing on the Health area, through the Transparency Portal of the Executive Power of the State of Santa Catarina. This is a descriptive and qualitative research. An analysis model was developed, based on information quality attributes to collect, diagnose and propose content. Data from the communication and ombudsman sectors of the State Department of Health, the 57 requests for information from the Access to Information Law (LAI), for the year 2019 in the Health Area, and the 26 transparency portals of the Federative Units (UF) were considered. Using document and content analysis, combined with the triangulation technique of the data collected, proposals for improvements were generated for the informational aspect of the Health area. The first consists of the development of a thematic area through pages containing health information. The second recommends the convergence of existing content from selection mechanisms, as the application of filters in the Portal's queries, to compose the thematic area. The third proposes 38 contents to be incorporated in the proposed thematic area. The proposal to organize and incorporate contents may improve transparency of the Health area, indicating priorities related to the provision of services. When implemented, they can facilitate access, convergence and availability of contents on the Internet to attend the informational demands of citizens and facilitate the exercise of social control. Keywords: Public Transparency. Health. Information Quality. Website. Introdução A informação e o conhecimento são aspectos fundamentais para o crescimento e manutenção das organizações, seja no âmbito interno como na relação com a população, tendo em vista que se vive em um ambiente dinâmico e em constante mudança. A importância destes aspectos configurou o surgimento da chamada Sociedade da Informação, Sociedade do Proposições para o Portal da Transparência do Poder Executivo do Estado de Santa Catarina com foco na área da saúde Revista Gestão e Secretariado (GeSec), São Paulo, SP, v. 14, n. 1, 2023, p. 277-305. 279 Conhecimento, Nova Economia ou outros nomes que representam um novo estágio atual de desenvolvimento da sociedade (Castells, 1999; Akutsu & Pinho, 2002; Pinho, 2008). Um dos pressupostos da Sociedade da Informação está relacionado as Tecnologias de Informação e Comunicação (TICs), como a Internet e os portais eletrônicos, que possibilitaram disponibilizar informações, dados e serviços para a sociedade, usufruindo de facilidades da tecnologia, como o formato digital, a velocidade de divulgação, as novas formas de interação, o custo reduzido na divulgação e a ampliação de acesso com uso de conteúdos eletrônicos. Dentre os benefícios, as TICs permitiram aos cidadãos melhorias no acesso às informações a respeito da gestão governamental (Akutsu & Pinho, 2002). As TICs estão inseridas no setor público e relacionados com o acesso à informação pública, que consiste em um direito dos cidadãos de obterem informações produzidas e guardadas pelo Poder Público, a menos que exista uma justificativa superior de interesse público que negue o acesso (Bento, 2015). O acesso às informações públicas foi orientado por declarações e pactos internacionais que colaboraram para o desenvolvimento de legislações nacionais que amparassem o direito à informação pública para os cidadãos (Angélico, 2015). Nas últimas décadas, importantes marcos legais favoreceram para que o país avançasse na garantia desse direito. Entre eles, a Constituição Federal (CF) de 1988 definiu dispositivos para assegurar o acesso à informação pública. Com a Lei Complementar n.º 131, de 27 de maio de 2009, foram estabelecidas formas para assegurar a transparência. A promulgação da Lei Federal n.º 12.527, de 18 de novembro de 2011, o Brasil passou a contar com uma Lei de Acesso à Informação (LAI), regulamentando o acesso às informações públicas previsto na CF (Angélico, 2015; Bento, 2015; Heinen, 2015). A LAI estabeleceu, em conformidade com os princípios básicos da administração pública, cinco diretrizes para assegurar o direito de acesso às informações públicas. A primeira refere-se a observância da publicidade como preceito geral e do sigilo como exceção. A segunda faz menção à divulgação de informações de interesse público, independentemente das solicitações da população. A utilização de meios de comunicação que sejam viabilizados pelas TICs consiste na terceira. A quarta diretriz estabelece o fomento ao desenvolvimento da cultura de transparência na administração pública e a última, orienta que ocorra o desenvolvimento do controle social da administração pública. As diretrizes da LAI orientam a divulgação de informações por meio da utilização das TICs como uma forma para desenvolver a transparência pública, abrindo possibilidades para melhorar o controle social (BRASIL, 2011). Proposições para o Portal da Transparência do Poder Executivo do Estado de Santa Catarina com foco na área da saúde Revista Gestão e Secretariado (GeSec), São Paulo, SP, v. 14, n. 1, 2023, p. 277-305. 280 Dentro destas perspectivas, a pesquisa apresenta o seguinte problema: quais diretrizes para o aperfeiçoamento da transparência de dados e informações públicas estaduais, com foco na área de Saúde, por meio do Portal da Transparência do Poder Executivo de Santa Catarina, podem ser propostas? O objetivo do artigo consiste em propor diretrizes para o aperfeiçoamento da transparência de dados e informações públicas estaduais, com foco na área de Saúde, por meio do Portal da Transparência do Poder Executivo de Santa Catarina. A área de Saúde atende a direitos sociais e universais da população por meio da prestação de serviços de forma gratuita pelos estados e municípios (BRASIL, 1988). A cidadania está ligada aos direitos sociais e universais que passaram a ser garantidos a partir da CF de 1988, estabelecendo-se assim deveres do Estado para com seus membros (Hullen, 2018). A área de Saúde apresenta complexidade no financiamento dos serviços para poder atender com plenitude e qualidade o atendimento dos direitos dos cidadãos, o que necessita de financiamentos suficientes e uma gestão adequada dos recursos obtidos (Saldiva & Veras, 2018; Pinto & Adrião, 2006), sendo que o percentual mínimo de aplicação para os estados é de 12% da receita resultante de impostos e deve ser gasto em ações e serviços públicos de saúde (BRASIL, 1988; BRASIL, 2012). Nos estudos anteriores pesquisados, a geração de proposições de conteúdos para os portais da transparência aparece como parte integrante de algumas publicações, mas foi pouco identificado estudos dirigidos que utilizaram atributos da qualidade da informação como requisito para esse fim. Constatou-se uma tendência de publicações voltadas para o aspecto avaliativo dos portais, valendo-se dos atributos da qualidade da informação (Lima et al., 2009; Alegretti & Platt Neto, 2010; Freire, 2014; Almeida, 2015; Jahns & Raupp, 2016; Máchová & Lnenicka, 2016; Souza et al., 2019; Curto-Rodríguez, 2020; Silva et al., 2020). A pesquisa traz uma contribuição teórica, pois estabelece uma forma sistematizada de identificar e gerar proposições com a utilização de atributos da qualidade da informação. As pesquisas voltadas para melhorias informacionais nos portais da transparência necessitam ser sistematizadas, para que a sociedade possa ter maiores informações acerca do que ser faz dentro do aparelho do Estado e contribuir para com o controle social. Acredita-se que uma população que tenha maior e melhor acesso dos dados e informações do Estado consiga compreender melhor o seu funcionamento e ter melhores condições para exercer o controle social. Proposições para o Portal da Transparência do Poder Executivo do Estado de Santa Catarina com foco na área da saúde Revista Gestão e Secretariado (GeSec), São Paulo, SP, v. 14, n. 1, 2023, p. 277-305. 281 Referencial Teórico-Empírico 2.1 Definindo transparência pública O conceito de transparência é cada vez mais encontrado na literatura e no uso cotidiano da sociedade, sendo natural que seu significado seja fluido, principalmente pelo fato do seu conceito ser usado em diferentes aspectos relacionados ao fluxo das informações. Na definição de Porumbescu (2015), o conceito de transparência se refere a amplitude com que as partes interessadas externas têm acesso regular as informações sobre o funcionamento das organizações públicas. Um objetivo principal da transparência é melhorar a compreensão do público sobre o que seu governo está fazendo e o porquê, por meio da divulgação das informações governamentais (Porumbescu, 2015). Angélico (2015) define transparência a partir dos mecanismos que possibilitam o acesso à informação pública seja pelas iniciativas do governo que são realizadas de uma maneira proativa ou, quando o mesmo é provocado para responder solicitações de informação encaminhadas pelos cidadãos, que são materializadas em pedidos de informação que seguem procedimentos formais até o cidadão receber a sua reivindicação, exclusiva e pessoal. A transparência não se confunde com a publicidade e com a publicação, sendo mais abrangente que ambas. A publicação é uma forma de exercício da publicidade, essa última deixa mostrar algo. Já a transparência vai além de apenas mostrar e está associado aquilo que deve se visível ou acessível de forma clara. Ela objetiva que o cidadão possa ter um melhor entendimento para acompanhar projetos, metas, indicadores, avaliar a eficiência dos serviços prestados e compreender outros conteúdos do Estado (Heinen, 2015). Platt Neto et al. (2009) explanam também que a transparência é um conceito mais amplo do que publicidade, pois uma informação pode ser pública, mas não ser relevante, confiável, tempestiva e compreensível. Além da garantia do atendimento das normas legais, as iniciativas de transparência na administração pública constituem uma política de gestão responsável que favorece o exercício da cidadania pela população (Platt Neto et al., 2009). Os autores citam atributos da qualidade da informação (confiabilidade, tempestividade e compreensibilidade) para o atendimento da transparência pública e consequente atendimento das necessidades dos cidadãos. Para que a transparência seja efetiva, a qualidade das informações disponibilizadas deve ser assegurada, visto que a divulgação de informações de má́ qualidade impossibilita a realização do efetivo acompanhamento as ações de governo (Souza et al., 2019). Proposições para o Portal da Transparência do Poder Executivo do Estado de Santa Catarina com foco na área da saúde Revista Gestão e Secretariado (GeSec), São Paulo, SP, v. 14, n. 1, 2023, p. 277-305. 282 Para os conteúdos de finanças públicas, a literatura normalmente remete a transparência fiscal, conceito que pode ser visto a partir da formulação realizada pelos economistas Kopits e Craig (1998) da International Monetary Fund (IMF), que enfatizam a importância da abertura de informações financeiras (políticas fiscais, contas públicas) e do próprio funcionamento do governo (estrutura e funções) para que os cidadãos (eleitorado) e os mercados financeiros possam compreender e avaliar a situação econômica e financeira do Estado, não apenas dos seus impactos no presente, mas também no futuro. Enfatiza algumas qualidades da transparência, como ser compreensível, comparável, tempestiva, confiável e abrangente. No Brasil, a transparência fiscal se tornou um tema de grande importância a partir do debate e promulgação da Lei Complementar n.º 101, de quatro de maio de 2000, denominada de Lei de Responsabilidade Fiscal (LRF), na qual o termo “transparência da gestão fiscal” aparece de forma explícita em uma seção desta lei. A transparência fiscal eleva seu grau de importância ao ponto de ser considerada como um princípio de gestão na LRF para assegurar a disponibilização de informações de cunho financeiro, aqui entendida de maneira ampla os aspectos inerentes as contas públicas, orçamento e outros relatórios previstos na LRF. 2.2 Trajetória legal A promulgação da “Constituição Cidadã”, assim conhecida a CF de 1988, consolidou um novo sistema de controle do Estado fundamentado nos princípios da legalidade, moralidade, finalidade pública, motivação, impessoalidade, publicidade e eficiência, nos quais a publicidade abriu caminho para que a sociedade tivesse conhecimento a respeito dos atos da administração pública (Zuccolotto & Teixeira, 2019). Foram criadas condições para o processo de abertura do aspecto informacional, principalmente quando se observam os ditames legais estabelecidos no art. 5 (incisos XIV e XXXIII), art. 37 (parágrafos 1º e 3º, inciso II) e no art. 26 (parágrafo 2º), que de forma geral buscam estabelecer o direito ao acesso à informação púbica para a população e a obrigação estatal de fornecê-las. Parte-se dos preceitos da Lei Maior do país para contextualizar os relacionados com a transparência pública, conforme Quadro 1. Proposições para o Portal da Transparência do Poder Executivo do Estado de Santa Catarina com foco na área da saúde Revista Gestão e Secretariado (GeSec), São Paulo, SP, v. 14, n. 1, 2023, p. 277-305. 283 Legislação Dispositivos Importantes Lei Complementar n° 101/2000 (Lei de Responsabilidade Fiscal - LRF) Artigo 1º: A responsabilidade da gestão fiscal tem como pressuposto a ação planejada e transparente; Artigo 48: Ampla divulgação em meio eletrônico de instrumentos da transparência da gestão fiscal, como os planos, orçamentos, prestações de contas, Relatório de Gestão Fiscal. Lei Complementar n° 131/2009 (Lei da Transparência): Acrescenta dispositivos na LRF Artigo 48: incentivo à participação popular e realização de audiências públicas durante a elaboração do planejamento orçamentário (planos, lei de diretrizes orçamentárias e orçamentos); disponibilização, em tempo real, de informações da execução orçamentária e financeira em meios eletrônicos de acesso público; Artigo 48-A: disponibilização para a população informações a respeito da despesa e receita pública. Decreto Federal 10.540/2020 (atualizou e revogou o Decreto Federal 7.185/2010), regulamentando o artigo 48, § 1º, inciso III e § 6º da LRF Artigo 2º: conceitos, como do Siafic, da liberação em tempo real, execução orçamentária; Artigo 3º e 6º: requisitos de qualidade para o Siafic; Artigo 7º e 8º: determinação de disponibilizar informações em meio eletrônico para a sociedade; detalhamento de informações de despesa e receita pública para serem disponibilizadas. Lei Federal n.º 12.527/2011 (Lei de Acesso à Informação – LAI) Artigo 3º: estabelece diretrizes, como a divulgação de informações de interesse público sem necessitar solicitações, o fomento ao desenvolvimento da cultura da transparência na administração pública para o controle social; Artigo 8º: relação de alguns dados e informações para serem divulgados e requisitos dos portais eletrônicos; Artigo 10º: determina a previsão de abertura de pedidos de informação para solicitação pelo requerente, independente de apresentar motivações para a solicitação. Quadro 1 – Principais Legislações Nacionais da Transparência Pública Fonte: Elaborado com base em Brasil (2000), Brasil (2009), Brasil (2010), Brasil (2011), Brasil (2020). A obrigação legal da transparência teve como ponto inicial a publicação de informações fiscais dos entes públicos previstos na LRF, que evidenciou a importância da transparência e relacionou os conteúdos a serem divulgados, como os relatórios fiscais, os planos e orçamentos. Ocorreram outros aprimoramentos normativos em termos de requisitos de disponibilização dos dados e informações e incentivos a participação popular, trazidos pela Lei da Transparência, os Decretos Federais n.º 7.185/2010 e n.º 10.540/2020. A LAI vem posteriormente ampliar a divulgação de novos conteúdos, além de estabelecer mecanismos para a solicitação de pedidos de informação e definir diretrizes visando fomentar a transparência pública e do controle social. 2.3 Estudos anteriores A quantidade de publicações totalizou 297 nas bases escolhidas (EBSCO com 109, SCOPUS com 62, BDTD com 116, Busca Livre com 10). Para facilitar as atividades, utilizou-se o gerenciador de referências Mendeley nas publicações das bases EBSCO e SCOPUS. Já para a busca de teses e dissertações na base BDTD contou com auxílio de planilha eletrônica, gerada a partir do portal no formato “CSV”. Na segunda seleção foram verificados os títulos, resumos e a leitura dos artigos relacionados. O critério estabelecido foi selecionar àquelas Proposições para o Portal da Transparência do Poder Executivo do Estado de Santa Catarina com foco na área da saúde Revista Gestão e Secretariado (GeSec), São Paulo, SP, v. 14, n. 1, 2023, p. 277-305. 284 voltadas ao setor público e a partir de avaliação preliminar, verificou-se os aspectos temáticos que se aproximassem da pesquisa. A seleção resultou em 52 publicações (EBSCO com 18, SCOPUS com 14, BDTD com 14, Busca Livre com 6), selecionando aquelas com estudos semelhantes aos tratados nesta pesquisa, dentro de uma perspectiva de proposição de conteúdo, incluindo os relacionados à área da Saúde, e aquelas associadas às dimensões/atributos da qualidade da informação. Como resultado, identificaram-se 13 publicações mais aderentes, apresentadas no Quadro 2, explanadas na sequência, que foram consolidadas e sintetizadas em duas abordagens. Periódico Autores Ano Título Revista Gestão em Análise Silva, A. A. P. S.; Monteiro, D. A. A.; Reis, A O. 2020 Qualidade da Informação dos dados governamentais abertos: análise do portal de dados abertos brasileiro. El profesional de la información, Curto-Rodríguez, R. 2020 Transparencia operativa de las comunidades autónomas españolas mediante sus portales de datos abiertos. Encontros Bibli: revista eletrônica de biblioteconomia e ciência da informação Souza, R. J.; Barbosa, R. B.; Cabral, A. C. A; Santos, S. M. 2019 Avaliação da Qualidade das Informações disponibilizadas no Portal da Transparência do Estado do Ceará. Repositório da Universidade de Brasília (UNB) Barbosa, M. C. B. 2018 A Lei de Acesso à informação no Brasil: análise da convergência entre os princípios da Lei e o conteúdo das solicitações dos usuários. Sociedade, Contabilidade e Gestão Calvosa, M. V. D.; Silva, T. A.; Krakauer, P. V. C. 2017 Portais Eletrônicos utilizados nos Municípios fluminenses: Análise das Ações Inovadoras. Journal of Theoretical and Applied Electronic Commerce Research Máchová, R.; Lnenicka, M. 2017 Evaluating the Quality of Open Data Portals on the National Level. Revista Gestão.Org Carossi, D.F.; Teixeira Filho, J.G.A. 2016 Uma análise dos Pedidos de Acesso à Informação encaminhados a uma instituição de ensino superior. Revista Universo Contábil Jahns, F. T.; Raupp, F.M. 2016 Transparência do Poder Executivo dos Estados Brasileiros. Repositório da Universidade de Brasília (UNB) Gama, J. R. 2015 Instrumentos de Transparência e acesso às informações públicas: um estudo das demandas por informações contábeis nas universidades federais. Repositório da Universidade Federal da Paraíba (UFPB) Almeida, S. E. C. 2015 Lei de Acesso à Informação e transparência: proposta de inclusão de informações no portal transparência UFPB. Repositório da Universidade de Brasília (UNB) Freire, F. R. F. 2014 Desafios para a Transparência Pública: Um estudo com os usuários do Portal da Transparência do Governo Federal. Revista Catarinense da Ciência Contábil Allegretti, D.D.S.; Platt Neto, O. A. 2010 Funcionalidades, limitações e potencialidades do Portal da Transparência do Rio Grande do Sul. Cadernos de Saúde Pública Lima, C. R. A.; Schramm, J. M. A.; Coeli, C. M.; Silva, M. E. M. 2009 Revisão das dimensões de qualidade dos dados e métodos aplicados na avaliação dos sistemas de informação em saúde. Quadro 2 – Relação de Publicações Fonte: Dados da pesquisa (2020). Proposições para o Portal da Transparência do Poder Executivo do Estado de Santa Catarina com foco na área da saúde Revista Gestão e Secretariado (GeSec), São Paulo, SP, v. 14, n. 1, 2023, p. 277-305. 285 A partir dos levantamentos das publicações apresentadas pode-se consolidar em duas abordagens as linhas de pesquisa sobre a temática do estudo, sumarizadas no Quadro 3. Abordagem Descrição Autores Avaliativa de portais eletrônicos, podendo utilizar atributos da qualidade da informação e apresentar proposições de melhorias. Avalia as condições e níveis de transparência dos portais eletrônicos, com alguns estudos utilizando atributos da qualidade da informação. Apresentam proposições gerais, como para usabilidade e organização dos portais. - Allegretti, D. D. S.; Platt Neto, O. - Curto-Rodríguez, R. - Freire, F. R. F. - Jahns, F.T.; Raupp, F.M. - Lima, C. R. A.; Schramm, J. M. A.; Coeli, C. M. Silva, M. E. M. - Máchová, R.; Lnenicka, M - Silva, A. A. P. S.; Monteiro, D. A. A.; Reis, A O. - Souza, R. J.; Barbosa, R. B.; Cabral, A.C.A; Santos, S.M. Investigativa de demandas informacionais da sociedade, sem utilizar atributos da qualidade da informação, podendo apresentar proposições de melhorias Investiga demandas de informação da sociedade, a partir de pedidos de informação da LAI e implementações em portais, sem o uso dos atributos da qualidade da informação. Apresentam proposições gerais, incluindo para os conteúdos. - Almeida, S.E.C. - Barbosa, M. C. B - Calvosa, M.V. D.; Silva, T.A.; Krakauer, P. V. C. - Carossi, D. F.; Teixeira Filho, J. G. A. - Gama, J. R. Quadro 3 – Abordagem dos Estudos Anteriores Fonte: Dados da pesquisa (2020). Considerando as duas abordagens identificadas nos estudos anteriores, o presente estudo está mais próximo da segunda abordagem, pois busca investigar demandas da sociedade a partir de pedidos de informação e da pesquisa em portais para propor alternativas para aperfeiçoamento do Portal. No entanto, traz uma forma diferenciada de estruturar as proposições, além de utilizar fontes de dados primária (questionário aplicado nas secretarias) para identificar outras demandas da população. Assim, infere-se que estudos sistematizados para a proposição de conteúdos nos portais da transparência, utilizando os atributos da qualidade da informação, são pouco utilizadas nesta concepção propositiva. A geração de propostas acaba ocorrendo como parte integrante dos estudos ou de uma forma geral, que até podem trazer em sua essência a qualidade da informação, mas pouco observam o relacionamento direto e sistematizado entre as dimensões/atributos da qualidade da informação e o conteúdo a ser proposto. Procedimentos Metodológicos 3.1 Tipos de pesquisa e contexto O estudo empreendido é descritivo, com abordagem qualitativa, realizado por meio de Proposições para o Portal da Transparência do Poder Executivo do Estado de Santa Catarina com foco na área da saúde Revista Gestão e Secretariado (GeSec), São Paulo, SP, v. 14, n. 1, 2023, p. 277-305. 286 estudo de caso, bibliográfico e documental. O contexto envolveu o Estado de Santa Catarina (SC), mais especificamente o Poder Executivo. As organizações mencionadas na pesquisa são a Secretaria de Estado a Fazenda (SEF), gestora do Portal entre os anos de 2012 e 2019, a Controladoria Geral do Estado (CGE), por meio da Ouvidoria Geral do Estado (OGE), que assumiu a gestão do Portal em meados de 2019 e realiza a gestão dos pedidos de informação da LAI, e a Secretaria de Estado da Saúde (SES) com o objetivo de contribuírem com proposições. Com relação aos dados primários, a população segue uma amostra intencional direcionada para a SES, mais especificamente para os setores que recebem demandas informacionais da sociedade, neste caso os setores de comunicação e Ouvidoria Geral, a fim identificar proposições de conteúdos, respondidas por meio de questionário eletrônico. Os dados secundários referem-se aos pedidos de informação da LAI sobre a área da Saúde feitos ao Estado de SC no ano de 2019 e os dados dos 26 portais da transparência das Unidades Federativas (UF) disponíveis na Internet. 3.2 Modelo de análise O modelo de análise utilizou como referência os atributos da qualidade da informação apresentados por Arouck (2011), Brasil (2011), Trindade, Oliveira e Becker (2011) e Platt Neto et al. (2009), além do Vocabulário Controlado de Governo Eletrônico (VCGE) (Brasil 2016) e da estrutura das informações (Cucciniello et al., 2014) como referências para aplicar classificações em conteúdos. O Quadro 4 apresenta o modelo de análise com a explicação das dimensões e categorias, dentro de uma visão de buscar a identificação, a compreensão e a proposição de conteúdos. Dimensões Categorias Base Teórica Identidade Objetiva definir uma nomenclatura e classificação do conteúdo. Conteúdo Identificação do conteúdo pela definição de um nome. Brasil (2016): VCGE (Saúde: Assistência Hospitalar e Ambulatorial, Medicamentos e Aparelhos, Combate a Epidemias, Defesa e Vigilância Sanitária, Outros em Saúde). Cucciniello et al. (2014): Estrutura das Informações (Institucional, Gestão Financeira, Prestação de Serviços). Classificação Identificação do tipo de melhoria (novo ou existente) e da categorização do conteúdo (estrutura das informações, assunto, detalhamento do assunto, sistema educacional, segmento da saúde). Utilidade Aborda a utilidade da divulgação do conteúdo e as fontes da divulgação. Relevância Aspectos que demonstram a utilidade e importância do conteúdo. Arouck (2011): Confiabilidade, Credibilidade, Importância, Relevância, Utilidade, Valor Informativo. Brasil (2011): Autencidade, Integridade Proposições para o Portal da Transparência do Poder Executivo do Estado de Santa Catarina com foco na área da saúde Revista Gestão e Secretariado (GeSec), São Paulo, SP, v. 14, n. 1, 2023, p. 277-305. 287 Confiabilidade Formas de demonstrar a credibilidade do conteúdo divulgado (sistema, órgão, autor, entre outros). Trindade, Oliveira e Becker (2011): Acurácia, Fonte, Valor, Confiabilidade, Relevância. Platt Neto et al. (2009): Confiabilidade, Utilidade, Relevância. Compreensibilidade Voltada à apresentação do conteúdo quanto a linguagem, formas, formatos e local de divulgação. Linguagem Trata das nomenclaturas e informações para o entendimento da composição do conteúdo. Arouck (2011): Clareza, Compreensibilidade, Formato, Simplicidade, Interpretabilidade. Brasil (2011): Disponibilidade Trindade, Oliveira e Becker (2011): Apresentação, Clareza, Formato, Inteligibilidade, Mídia, Interface, Interpretabilidade. Platt Neto et al. (2009): Apresentação, Compreensibilidade, Divulgação, Linguagem. Apresentação Relacionada às formas (consulta, texto, gráfico, arquivo), formatos de obtenção (pdf, csv, etc) e local da divulgação do conteúdo. Temporalidade Relacionada aos aspectos temporais do conteúdo para atualização e manutenção da disponibilização do mesmo. Tempestividade Indica a periodicidade requerida de atualização do conteúdo (diário, mensal, anual, etc.). Arouck (2011): Atualidade, Tempestividade. Trindade, Oliveira e Becker (2011): Atualidade, Frequência, Idade, Período, Tempestividade, Temporalidade. Platt Neto et al. (2009): Tempestividade. Série Histórica Estabelece o período de tempo que inicia a divulgação do conteúdo (tempo passado para início da série histórica). Quadro 4 – Modelo de Análise Fonte: Dados da pesquisa (2020). Trata-se de um modelo desenvolvido com o objetivo de coletar, diagnosticar e propor conteúdos. Para isto, trabalha com uma dimensão para classificar o conteúdo e criar uma identidade para o mesmo, facilitando atividades de coleta e análise de conteúdo, que requerem consolidação e organização de dados e informações. As demais dimensões foram elaboradas a partir da teoria da qualidade da informação, especificamente relacionada aos seus atributos, no sentido de definir alguns atributos que permitam conhecer, descrever e gerar proposições de conteúdos para serem aplicados em portais. O modelo é composto por quatro dimensões e oito categorias. A dimensão identidade contém um nome (conteúdo), um atributo do tipo de melhoria (conteúdo novo ou existente), e categorias comuns às áreas para classificação: estrutura das informações (agrupamento geral), assunto (nível médio de agrupamento) e detalhamento do assunto (nível analítico), e categorias específicas da área da Saúde. As dimensões da utilidade, compreensibilidade e temporalidade são dimensões que objetivam a compreensão de alguns aspectos e características dos dados e informações, sendo cada uma segmentada em duas categorias de análise. A utilidade possui as categorias relevância e confiabilidade. A compreensibilidade apresentada a linguagem e a apresentação como suas categorias. E a temporalidade, apresenta as categorias tempestividade e série histórica. Proposições para o Portal da Transparência do Poder Executivo do Estado de Santa Catarina com foco na área da saúde Revista Gestão e Secretariado (GeSec), São Paulo, SP, v. 14, n. 1, 2023, p. 277-305. 288 3.3 Etapas da pesquisa, instrumentos de coleta e técnicas de análise As etapas 1 e 2 da pesquisa voltaram-se para a caracterização do Portal e o diagnóstico da situação dos conteúdos da área da Saúde. A etapa 3 envolveu a identificação e levantamento de conteúdos gerados por demandas da sociedade a partir de algumas fontes de dados. A etapa 4 gerou propostas de aperfeiçoamento contendo alternativas para melhorias informacionais das áreas a partir do conjunto de dados e informações coletados. A partir da definição das etapas da pesquisa, vinculadas a cada objetivo específico, procedeu-se a definição dos instrumentos de coleta e análise dos dados. O Quadro 5 apresenta as etapas, e as respectivas técnicas de coleta e de análise de dados previstos. Etapas da Pesquisa Técnicas de Coleta Técnicas de Análise Etapa 1: Caracterização do Portal da Transparência a) Pesquisa Documental: realizada por meio de documentos publicados, legislações, informações e dados do Portal na Internet. a) Análise Documental: teve como resultado de um texto descritivo a partir da análise de documentos. Etapa 2: Descrição do Conteúdo da Área a) Pesquisa Documental: realizada por meio de documentos publicados, informações e dados do Portal na Internet da área da Saúde. a) Análise Documental: teve como resultado de um texto descritivo a partir da análise de documentos e dados da Saúde. Etapa 3: Identificação de demandas da sociedade: - Coleta de Pedidos de Informação (LAI) (demandas geradas por pedidos da LAI) - Coleta de Questionários na Secretaria. (demandas informadas pela secretaria) - Coleta em Portais da Transparência das Unidades Federativas (demandas desenvolvidas em portais estaduais – benchmarking) a) Pesquisa Documental: utilizada a base de dados estadual de pedidos de informação da LAI (2019) da área da Saúde (planilha eletrônica). b) Questionário: encaminhado questionário (arquivo word) para a Secretaria de Estado da Saúde, aos setores de assessoria de comunicação e ouvidoria. c) Pesquisa Documental e Observação Sistemática: acessado os portais da transparência (checklist e protocolo de observação). a) Análise documental: para cada área foi gerada três bases de dados (planilhas eletrônicas) com a relação e dados dos conteúdos: pedidos de informação da LAI, pesquisa nos portais e questionário aplicado. b) Análise de Conteúdo: para cada área, as três bases de dados tiveram os conteúdos descritos, classificados e depois associados para formar uma base de dados consolidada (planilhas eletrônicas - base de dados final). Etapa 4: Propostas de Aperfeiçoamento Triangulação: combinação do conjunto de informações das etapas anteriores Triangulação: teve como resultado um texto descritivo a partir da análise das etapas anteriores Quadro 5 – Etapas da Pesquisa, Técnicas de Coleta e Instrumentos de Análise de Dados Fonte: Dados da pesquisa (2020). De modo particular, estruturou-se a análise de conteúdo a partir das etapas teóricas citadas por Moraes (1999), como segue: a) Preparação da base de dados: preparação das três bases de dados (planilhas eletrônicas) iniciais da área, conhecendo, revisando e definindo campos, dados previstos para Proposições para o Portal da Transparência do Poder Executivo do Estado de Santa Catarina com foco na área da saúde Revista Gestão e Secretariado (GeSec), São Paulo, SP, v. 14, n. 1, 2023, p. 277-305. 289 composição, padrões de colunas das bases, além de outras questões inerentes a cada conjunto de dados: 1) base dos pedidos de informação da LAI: necessidade de identificar os pedido de informação de interesse geral, excluindo os pedidos de demandas particulares e pontuais que não se aplicam no contexto da pesquisa; 2) base dos portais das UF: necessidade de preencher e revisar os dados dos registros dos protocolos de observação de cada busca realizada; 3) base da coletada de dados das organizações: necessidade de preencher e revisar os dados recebidos dos questionários dos setores da Saúde; b) Categorização do Conteúdo: atividade para classificar os conteúdos em categorias, que seguem a dimensão “identidade” do modelo de análise. A primeira e mais sintética partiu da estrutura das informações (informações institucionais, gestão financeira, prestação de serviços), seguida de duas categorias vinculadas a cada estrutura (assunto e detalhamento do assunto), representando níveis detalhados, que foram sendo criadas na análise, agrupando os conteúdos conforme sua semelhança. Outras classificações foram específicas da área e classificadas de forma independente; c) Consolidação das Bases de Dados: atividade para unificar as três bases de dados de cada área, criando uma base consolidada (planilha eletrônica). O objetivo foi gerar um agrupamento geral e único a partir das três fontes de dados (pedidos da LAI, questionário das secretarias e pesquisa nos portais das UF); d) Descrição e Interpretação do Conteúdo: atividade final para interpretar e definir os conteúdos a partir da associação daqueles comuns e semelhantes (gerar uma única proposição) a partir das bases de dados consolidadas. A descrição ocorreu a partir das categorias do modelo de análise. Caracterização, Diagnóstico e Análise da Realidade Estudada 4.1 O portal e os conteúdos da saúde A partir do Portal da Transparência do Poder Executivo de Santa Catarina (www.transparencia.sc.gov.br) pode-se acessar e navegar diversas funcionalidades, como a pergunta cidadã, o mapa da transparência, a galeria de quadros e os menus que representam as quatro grandes áreas (Receita, Despesa, Responsabilidade Fiscal e Gestão Estadual) que permitem acessar outras funcionalidades. Os dados e informações do Portal são provenientes de alguns sistemas de gestão do Estado, como o Sistema Integrado de Planejamento e Gestão Fiscal (SIGEF) e de páginas textuais e relatórios gerados pelas áreas governamentais (SANTA Proposições para o Portal da Transparência do Poder Executivo do Estado de Santa Catarina com foco na área da saúde Revista Gestão e Secretariado (GeSec), São Paulo, SP, v. 14, n. 1, 2023, p. 277-305. 290 CATARINA, 2018). Os conteúdos da área Saúde estão presentes no Portal por meio de funcionalidades identificadas de maneira direta na página principal (quadros de despesa, COVID-19) ou quando dispostas nos menus, como no Gestão Estadual. No entanto, a grande maioria dos conteúdos necessita da seleção e aplicação de filtros para direcionar para as respectivas áreas, seja pelo nome da área, pelos órgãos e unidades gestoras da estrutura, pela função da despesa e outros parâmetros tendo em vista que os dados dos sistemas e relatórios têm suas particularidades, padrões próprios e atendem de forma ampla as áreas governamentais. A seleção para realizar o agrupamento ou detalhamento das consultas pesquisadas é realizada geralmente pela estrutura (órgãos ou unidades gestoras) da Secretaria de Estado da Saúde (SES) e os fundos associados da Saúde (Fundo Estadual da Saúde; Fundo Estadual de Apoio aos Hospitais Filantrópicos, Hemosc, Cepon e Hospitais Municipais; Fundo Catarinense para o Desenvolvimento da Saúde). O Quadro 6 apresenta a relação sumarizada dos conteúdos identificados da Saúde, agrupados de acordo com as informações do modelo de análise. Estrutura Conteúdos Informações Institucionais Informações dos Servidores; Situação de Cargos de Provimento Efetivo; Legislação dos Servidores; Informações Organizacionais (estrutura organizacional, contatos e competências); Informações dos Bens Móveis; Informações dos Bens Imóveis Gestão Financeira Licitações; Contratos; Diárias pagas aos servidores; Receita Orçamentária arrecadada pelo órgão; Despesa Orçamentária do Exercício e Restos a Pagar; Documentos de empenho, liquidação e pagamento; Relação de pagamentos por fornecedores; Relação de pagamentos do Cartão de Pagamento (CPESC); Pagamentos Extraorçamentários; Acompanhamento da aplicação do percentual de aplicação de 12% das receitas de impostos.; Informações dos Orçamentos Públicos em Saúde (SIOPS); Extratos bancários de contas correntes; Informações do Planejamento (PPA, LDO e LOA) Prestação de Serviços Obras Públicas; Convênios com Municípios e Entidades; Acompanhamento do Plano Plurianual (PPA); Informações da Prestação de Contas Anual do Governo (Balanço Geral); Lista de Espera do Sistema Único de Saúde (SUS); Informações de Relatórios de Gestão da Saúde; Informações da Covid-19 Quadro 6 – Conteúdos da Saúde Sumarizados Fonte: Elaborado com base em Santa Catarina (2020). 4.2 Pedidos de informação da LAI Os pedidos de informação da LAI do ano de 2019 da Saúde representam uma fonte de dados para identificar demandas da sociedade oriundas da transparência passiva. De acordo com a LAI, qualquer interessado pode realizar o pedido de informação aos órgãos públicos que devem viabilizar alternativa para encaminhamento dos pedidos por meio dos seus sítios na Internet (BRASIL, 2011; SANTA CATARINA, 2020). Proposições para o Portal da Transparência do Poder Executivo do Estado de Santa Catarina com foco na área da saúde Revista Gestão e Secretariado (GeSec), São Paulo, SP, v. 14, n. 1, 2023, p. 277-305. 291 As bases de dados (planilhas eletrônicas) recebidas constaram de 52 pedidos da Saúde. O perfil dos solicitantes dos 52 pedidos foi caracterizado por 46,2% (24 pedidos) de pessoa física, 34,6% (18 pedidos) de pessoa jurídica e 19,2% não informaram (10 pedidos). Os solicitantes do sexo feminino abriram 21 pedidos (40,4%), 13 pedidos foram do sexo masculino (25%) e 18 pedidos não foram informados (34,6%). A localização geográfica dos solicitantes, por estado, apresentada na Figura 13, teve como predominância o estado de SC com 36,5% e do PR com 30,8%, totalizando 67,3% dos pedidos. O restante dos pedidos (32,7%) ficou distribuído em 5 UF (DF, SP, Rio de Janeiro – RJ, MG, Rio Grande do Sul – RS). Em relação às cidades, Florianópolis com 30,8% (16 pedidos) e Curitiba com 28,8% (15 pedidos) tiveram o maior número de solicitantes (59,6%), ficando o restante dos pedidos distribuídos em 11 cidades (40,4%), variando entre 5 e 1 pedidos para cada cidade. Após o relato do perfil dos solicitantes, para conhecer e organizar os pedidos, aplicou-se o modelo de análise. Foi realizado a identificação do conteúdo (definição de um nome), classificação do tipo de melhoria e a categorização, que seguiu as categorias definidas da estrutura das informações e do Vocabulário Controlado de Governo Eletrônico (VCGE). Além disso, foram definidas outras categorias durante o processo de análise de conteúdo (categorias assunto e detalhamento do assunto). A realização dos agrupamentos favoreceu a identificação de dados e informações semelhantes e com características próximas. Os dados da categorização de todos os 52 pedidos de informação mostraram que 67,3% (35 pedidos) estão associados à Prestação de Serviços, 25% (13 pedidos) à Gestão Financeira e 7,7% (4 pedidos) para as Informações Institucionais. Uma categorização adicional foi aplicada com base no VCGE da área da Saúde, aqui chamada Segmento da Saúde, gerando o seguinte perfil de dados: Medicamentos e Aparelhos com 38,5% (20 pedidos), Assistência Hospitalar e Ambulatorial com 32,7% (17 pedidos), Defesa e Vigilância Sanitária com 13,5% (7 pedidos) e Outros em Saúde representaram 15,4% (8 pedidos). A partir do agrupamento dos pedidos da Saúde, que resultou em 19 conteúdos, a composição da categoria Prestação de Serviços apresentou um percentual de 68,4% (13 conteúdos), mantendo-se com maior representatividade. As categorias informações institucionais e gestão financeira tiveram a mesma representatividade (15,8% - 3 conteúdos cada). Com relação a categorização do Segmento da Saúde, a Assistência Hospitalar e Ambulatorial apresentou o percentual de 42,1% (8 conteúdos), Medicamentos e Aparelhos (21,1% - 4 conteúdos), Defesa e Vigilância Sanitária (10,5% - 2 conteúdos) e Outros em Saúde (26,3% - 5 conteúdos). Na relação dos 19 conteúdos, 26,3% (5 conteúdos: informações de licitações nos Proposições para o Portal da Transparência do Poder Executivo do Estado de Santa Catarina com foco na área da saúde Revista Gestão e Secretariado (GeSec), São Paulo, SP, v. 14, n. 1, 2023, p. 277-305. 292 hospitais, quantidade de servidores e dados funcionais, dados gerais e estatísticos de leitos, legislação relacionado aos servidores da saúde, informações da lista de espera) estão voltados para melhoria de conteúdos existentes e 73,7% (14 conteúdos) para a proposições de novos. Uma outra questão se refere à descrição das constatações identificadas durante a análise de conteúdo dos pedidos das áreas a partir dos atributos da qualidade da informação. A descrição dos principais relatos e inferências encontrados a partir dos atributos da qualidade da informação foram os seguintes para a área da Saúde: 1. Relevância: observado o interesse pela estrutura da saúde em seus aspectos gerais e também relacionada a um determinado hospital ou local com alguma especialidade ou referência de tratamento. Outro interesse foi para a obtenção de dados estatísticos de algumas doenças, medicamentos e de pacientes atendidos em determinado procedimento ou que sofreram alguma doença. Constaram pedidos para atender pesquisas acadêmicas da temática da saúde. Constatou-se também o interesse por questões financeiras voltadas a aquisição de medicamentos, como as notas fiscais das operações, os estoques, datas e valores; 2. Confiabilidade: verificado que a maioria das reivindicações remetem diretamente a SES, com poucos casos para diretorias. A informação do nível mais alto da hierarquia (secretaria) já demonstra ser suficiente para demostrar a confiabilidade; 3. Linguagem: uma forma de identificar este atributo foi a partir das palavras mais repetidas nos textos dos pedidos. Foi elaborada uma Word Cloud contendo uma lista de palavras com a ferramenta Atlas.TI® como forma de identificar palavras recorrentes e utilizadas pela população, constatando-se em ordem crescente: Medicamentos (59 palavras); Informações (45); Saúde (41); Santa Catarina (36); Estado (35); Acesso (28); Hospitais (27); Secretaria (27). O destaque foram as palavras “medicamentos” e “hospitais” que são específicas da área da Saúde dentro do grupo das oito mais repetidas. Além delas, tiveram outras com menor volume dentro de uma linguagem da área: pacientes, doença, procedimentos, leitos e tratamento. Identificou-se, assim, e fica sugerido o uso de nomenclaturas e descrições que utilizam as palavras mencionadas da área da Saúde dentro da linguagem próxima dos cidadãos; 4. Apresentação: verificado as demandas envolvendo quantidades, valores, listas, fabricantes, datas, como nas solicitações de informações de medicamentos e de notas fiscais, sugerem formas tabuladas para geração a obtenção dos dados. Os solicitantes pediram para encaminhar em planilhas com formatos “xls”, “csv” e “xml” (notas fiscais eletrônicas) havendo citações para ser em formato aberto, como previsto na LAI; 5. Tempestividade: constatado que os pedidos de informações foram demandados em Proposições para o Portal da Transparência do Poder Executivo do Estado de Santa Catarina com foco na área da saúde Revista Gestão e Secretariado (GeSec), São Paulo, SP, v. 14, n. 1, 2023, p. 277-305. 293 sua maioria por mês e ano, sendo considerado o ano como a situação vigente da solicitação. Observou-se necessidade de dados que remetem uma atualização diária (estoque de medicamentos, oferta de leitos, por exemplo) que requer um menor tempo de atualização. Os conteúdos que envolveram dados quantitativos de um determinado período normalmente foram solicitados por mês e ano, sendo a última com maior volume; 6. Série Histórica: verificado que a maioria dos pedidos contemplaram séries históricas entre os anos de 2015 e 2019. A maior série foi um pedido de dados e informações entre 1998 e 2018 (20 anos) e alguns pedidos demandaram que os conteúdos fossem gerados a partir do final da década de 2000 (em torno de 10 anos). Infere-se assim que a disponibilização de séries históricas a partir de 5 anos tende a contemplar grande parte das solicitações, onde se teve demandas a partir de 2017 (3 anos) com certa recorrência nesta faixa. A predominância dos pedidos de séries históricas envolveu solicitação de quantidades, valores e listagem de dados. 4.3 Informações da Secretaria da Saúde Foi estruturado um questionário contendo uma parte inicial com a descrição da pesquisa (seção: Apresentação da Pesquisa), seguido da relação de conteúdos divulgados no Portal com a funcionalidade e o endereço de acesso dos mesmos (seção: O que o Portal divulga hoje?), e na parte final um espaço para o registro das contribuições dos setores por meio do preenchimento de quadros de proposição de conteúdo, contendo 10 questões para guiar a formulação de cada conteúdo a partir do modelo de análise (seção: Contribua com a Saúde). O Quadro 7 apresenta duas categorias de assunto e quatro de detalhamento de assunto, criadas a partir da análise das informações dos questionários. Estrutura das Informações Assunto Detalhamento do Assunto Informações Institucionais Recursos Humanos Concursos Prestação de Serviços Informações dos Serviços Medicamentos Procedimentos Vigilância Sanitária Quadro 7 – Categorias Desenvolvidas da Saúde na Análise dos Questionários Fonte: Dados da pesquisa (2021). Foram recebidas quatro proposições de conteúdos que após classificadas mostraram que 75% (3 conteúdos) se referem à categoria de prestação de serviços e 25% (1 conteúdo) para as Informações Institucionais. A composição na categoria Segmento da Saúde ficou distribuída igualmente em quatro categorias (25% cada): Assistência Hospitalar e Ambulatorial, Defesa e Vigilância Sanitária, Medicamentos e Aparelhos, Outros em Saúde. Proposições para o Portal da Transparência do Poder Executivo do Estado de Santa Catarina com foco na área da saúde Revista Gestão e Secretariado (GeSec), São Paulo, SP, v. 14, n. 1, 2023, p. 277-305. 294 Nas melhorias propostas todos se enquadraram em novos conteúdos para disponibilização. O Quadro 8 relaciona os conteúdos da Saúde com a categorização aplicada e a origem dos mesmos. Conteúdo Categorização Proposição Recebida Concursos e Seleções Informações Institucionais / Recursos Humanos / Concursos / Outros em Saúde R1 SC Transplantes Prestação de Serviços / Informações dos Serviços / Procedimentos / Assistência Hospitalar e Ambulatorial R2 Assistência Farmacêutica Prestação de Serviços / Informações dos Serviços / Medicamentos / Medicamentos e Aparelhos R3 Vigilância em Saúde Prestação de Serviços / Informações dos Serviços / Vigilância Sanitária / Defesa e Vigilância Sanitária R4 Quadro 8 – Conteúdos da Saúde da Pesquisa com a SES Fonte: Dados da pesquisa (2021). Utilizando a categoria Assunto foram explanados os conteúdos. Em Recursos Humanos estão agrupadas as informações dos concursos e seleções para ingresso de servidores e das Organizações Sociais da Saúde (OSS). A categoria prestação de serviços dispõem de conteúdos envolvendo os transplantes do Estado, como informações dos procedimentos e estatísticas de doadores e transplantados. Contém assuntos da assistência farmacêutica dos medicamentos, que apresentam informações de medicamentos, protocolos, listas e também conteúdos da vigilância em saúde estadual, como os alertas sanitários, os cuidados básicos em virtude das enchentes, os produtos irregulares notificados (alimentos, cosméticos, medicamentos, entre outros). Achados semelhantes não foram identificados em estudos anteriores. 4.4 Portais da transparência estaduais Os dados coletados nos portais da transparência estaduais configuram uma fonte de dados secundária para identificar conteúdos já desenvolvidos nos demais estados e o Distrito Federal, que buscaram atender necessidades informacionais da população. Os portais foram acessados pela Internet entre os meses de fevereiro e abril de 2021. Foram identificados 80 achados de conteúdos da área da Saúde, que se constituíram em uma base de dados que foi estruturada e analisada. O resultado das classificações está apresentado no Quadro 9, partindo-se das três categorias macros (Informações Institucionais, Gestão Financeira, Prestação de Serviços), gerando 8 categorias de assunto e 17 de detalhamento do assunto. Proposições para o Portal da Transparência do Poder Executivo do Estado de Santa Catarina com foco na área da saúde Revista Gestão e Secretariado (GeSec), São Paulo, SP, v. 14, n. 1, 2023, p. 277-305. 295 Estrutura da Informação Assunto Detalhamento do Assunto Informações Institucionais Estrutura Organizacional Saúde Estadual Gestão Financeira Despesas COVID-19 Organizações Sociais da Saúde Saúde Estadual Informações Financeiras Organizações Sociais da Saúde Informações Fiscais COVID-19 Impactos das Medidas de Proteção COVID-19 Receitas COVID-19 Prestação de Serviços Ações e Resultados COVID-19 Doenças Saúde Estadual Prestação de Serviços Informações dos Serviços Informações dos Servidores Dados dos Hospitais Medicamentos Organizações Sociais da Saúde Procedimentos Saúde Estadual Quadro 9 – Categorias Desenvolvidas da Saúde na Análise dos Portais Fonte: Dados da pesquisa (2021). A composição dos 80 achados na Saúde (base de dados inicial) em termos percentuais a partir das três categorias macros retratou que a Prestação de Serviços representou 66,2% (53 achados) do total, seguida da Gestão Financeira com 40% (24 achados) e das Informações Institucionais com 3,8% (3 achados). Na categorização do Segmento da Saúde a distribuição ocorreu em cinco categorias, sendo a de volume maior a de Combate a Epidemias (58,8% - 47 achados) onde se concentram os assuntos da COVID-19. As outras foram: Assistência Hospitalar e Ambulatorial com 25% (20 achados), Medicamentos e Aparelhos com 5% (4 achados), Defesa e Vigilância Sanitária com 1,2% (1 achado) e os demais achados foram categorizados em Outros em Saúde (10% - 8 achados). Tais achados evidenciam uma relação com a categoria teórica Utilidade (Arouck, 2011; Brasil, 2011; Platt Neto et al., 2009; Trindade, Oliveira e Becker, 2011), cujo significado remete à utilidade do conteúdo e as fontes de divulgação. Propostas de Aperfeiçoamento 5.1 Desenvolvimento da área temática da Saúde A recomendação aqui exposta visa propor uma forma de organizar a gama dos conteúdos das áreas objetivando convergir em um ponto comum. Recomenda-se o desenvolvimento de páginas temáticas para as áreas para que os cidadãos possam acessar todo os conteúdos divulgados das áreas em um único local. A utilização de áreas temáticas foi observada em alguns portais estaduais, embora com poucos conteúdos, mas demonstra uma Proposições para o Portal da Transparência do Poder Executivo do Estado de Santa Catarina com foco na área da saúde Revista Gestão e Secretariado (GeSec), São Paulo, SP, v. 14, n. 1, 2023, p. 277-305. 296 tendência em convergir assuntos por temas nos quais a Saúde faz parte. Os achados foram identificados nos portais da transparência do Espírito Santo, Paraná, Maranhã e São Paulo (ESPÍRITO SANTO, 2021; PARANÁ, 2021; MARANHÃO, 2021; SÃO PAULO, 2021), e demonstram um relação com as categorias teóricas Compreensibilidade (Arouck, 2011; Brasil, 2011; Platt Neto et al., 2009; Trindade, Oliveira e Becker, 2011) e Utilidade do conteúdo (Arouck, 2011; Brasil, 2011; Platt Neto et al., 2009; Trindade, Oliveira e Becker, 2011). Outra constatação desta tendência são as implementações de grupos, áreas ou temas, assim chamados estes conjuntos de assuntos nos portais de dados abertos, em que a Saúde faz parte. No portal de dados abertos do Distrito Federal, por exemplo, é explanado que a utilização de temas (Saúde, Educação, Transporte) é realizada para criar e gerenciar os dados por assuntos, ajudando quem consulta a encontrar dados de assuntos comuns (DISTRITO FEDERAL, 2021). A recomendação das páginas temáticas representa uma melhoria para o Portal agrupar a quantidade e diversidade dos conteúdos das áreas. Atualmente, as áreas estão dispostas no menu gestão estadual do Portal, onde cada uma é um submenu com três funcionalidades de conteúdos: Lista de Espera no SUS, Relatório de Gestão, Orçamentos Públicos em Saúde (SANTA CATARINA, 2020). Este formato se torna insuficiente para organizar a divulgação da série de conteúdos sugeridos. Recomenda-se reestruturar o acesso criando um submenu “Áreas Temáticas”, dentro do menu gestão estadual, agrupando funcionalidades (links) com o nome das áreas que direcionariam para as páginas temáticas de cada uma. Além disso, recomenda-se criar outras formas para facilitar o acesso para estas páginas, como pela utilização de quadros na página principal, que forneceriam uma visualização destacada e acesso rápido. Nesta primeira parte da proposta de aperfeiçoamento, a sugestão visa convergir os conteúdos de cada área para locais próprios e individualizados, que seriam as áreas temáticas das áreas, instrumentalizada por uma página da Saúde. Cada local agruparia diversas funcionalidades para direcionar os conteúdos da área, podendo conter informações e números importantes. A página temática faria a integração com funcionalidades do Portal (menu, página principal, páginas internas de conteúdos da área, ferramenta de busca, mapa do Portal) e com páginas externas de conteúdos da área, tornando-se um ponto concentrado para disponibilizar os dados e informações de áreas temáticas governamentais. Proposições para o Portal da Transparência do Poder Executivo do Estado de Santa Catarina com foco na área da saúde Revista Gestão e Secretariado (GeSec), São Paulo, SP, v. 14, n. 1, 2023, p. 277-305. 297 5.2 Convergência dos conteúdos existentes para a área temática Saúde O Quadro 10 relaciona os conteúdos existentes no Portal, organizados pela área e subdivisão temática, e as ações recomendadas de seleção que foram propostas a partir da experiência do pesquisador no processo de navegação e busca de dados e informações no Portal. Subdivisão Temática Conteúdos Recomendação Seleção Informações Institucionais Informações dos Servidores; Informações dos Bens Móveis; Informações dos Bens Imóveis. Desenvolver automatização da consulta existente com filtros da área. Informações Institucionais Legislação dos Servidores. Relacionar documentos e disponibilizar em local próprio da área. Informações Institucionais Informações Institucionais (estrutura organizacional, contatos e competências). Selecionar o conteúdo da área e desenvolver relatório próprio. Direcionar para página externa da área (estrutura organizacional). Informações Institucionais Situação de Cargos de Provimento Efetivo. Selecionar o conteúdo da área e desenvolver relatório próprio. Gestão Financeira Licitações; Contratos; Diárias pagas aos servidores; Receita Orçamentária arrecada pelo órgão; Despesa Orçamentária do Exercício e Restos a Pagar; Documentos de empenho, liquidação e pagamento; Relação de pagamentos por fornecedores; Relação de pagamentos do Cartão de Pagamento (CPESC); Pagamentos Extraorçamentários; Extratos bancários de contas correntes. Desenvolver automatização da consulta existente com filtros da área. Gestão Financeira Informações do Planejamento (PPA, LDO e LOA). Selecionar o conteúdo da área e desenvolver relatório próprio. Gestão Financeira Acompanhamento da aplicação do percentual de aplicação de 12% das receitas de impostos.; Informações dos Orçamentos Públicos em Saúde (SIOPS). Direcionar para página externa, reavaliando segmentações dos conteúdos. Prestação de Serviços Obras Públicas; Convênios com Municípios e Entidades; Estoque e Almoxarifado; Acompanhamento do Plano Plurianual (PPA). Desenvolver automatização da consulta existente com filtros da área. Prestação de Serviços Informações da Prestação de Contas Anual do Governo (Balanço Geral). Selecionar o conteúdo da área e desenvolver relatório próprio. Prestação de Serviços Lista de Espera do Sistema Único de Saúde (SUS); Informações de Relatórios de Gestão da Saúde; Informações da Covid-19. Direcionar para página externa, reavaliando segmentações dos conteúdos. Quadro 10 – Conteúdos Existentes para a Área Temática Saúde Fonte: Elaborado com base em Santa Catarina (2020). O Quadro 10 apresenta uma proposta da composição da área temática a partir dos conteúdos existentes, sem necessitar modificação da organização atual do Portal, mas aplicando processos de seleção dos dados e informações. As recomendações de seleção apresentadas sugestionam que ocorram desenvolvimentos automatizados de Tecnologia da Informação e Comunicação (TIC) nas funcionalidades para deixar os dados e informações Proposições para o Portal da Transparência do Poder Executivo do Estado de Santa Catarina com foco na área da saúde Revista Gestão e Secretariado (GeSec), São Paulo, SP, v. 14, n. 1, 2023, p. 277-305. 298 selecionados para cada área, ou seja, as consultas seriam replicadas já filtradas para a área temática. Outra recomendação de seleção seria para identificar e divulgar o conteúdo da área que hoje é disponibilizado junto com outros em arquivos (parte de um consolidado), para serem apresentados em um relatório próprio. A funcionalidade que abrigaria poderia explanar o documento de origem, se houvesse necessidade. Além disso, os documentos eletrônicos das áreas, presentes em funcionalidades gerais, poderiam ser selecionados e divulgados em uma funcionalidade própria, como no caso das legislações dos servidores. Por fim, os conteúdos que estão sendo direcionados para páginas externas fariam parte do conjunto, podendo ser avaliado a segmentação dos mesmos. 5.3 Inclusão de novos conteúdos e melhorias nos existentes A última parte da proposta de recomendação apresenta um conjunto de novos conteúdos e de melhorias nos existentes da Saúde que poderão ser avaliados para divulgação no Portal. Recomenda-se que os conteúdos sejam integrados nas áreas temáticas e dispostos nas subdivisões temáticas para que o acesso dos cidadãos seja facilitado. Os dados e informações propostos são resultantes da consolidação de três fontes de coleta de dados: pedidos de informações da LAI do ano de 2019, informações da Secretaria de Estado da Saúde, pesquisa nos portais das UF, sendo as duas últimas realizadas em 2021. A temática da prestação de serviços predominou nas proposições com 57,9% (22 conteúdos) do total, ficando as demais distribuídas na gestão financeira (31,6% - 12 conteúdos) e nas informações institucionais (10,5% - 4 conteúdos). A classificação resultou na seguinte composição: Assistência Hospitalar e Ambulatorial (15 conteúdos – 39,5%), Combate a Epidemias (9 – 23,7%), Outros em Saúde (7 – 18,4%); Medicamentos e Aparelhos (4 – 10,5%); Defesa e Vigilância Sanitária (3 – 7,9%). As proposições de melhorias de novos conteúdos representaram 71,1% (27 conteúdos) e de conteúdos existentes (28,9% - 11 conteúdos). Estes achados evidenciam uma relação com a categoria teórica Utilidade (Arouck, 2011; Brasil, 2011; Platt Neto et al., 2009; Trindade, Oliveira e Becker, 2011). Os 38 conteúdos definidos (11 melhorias e 27 novos) foram assim descritos: 22 conteúdos de Prestação de Serviços (57,9%): incluem dados gerais e estatísticos de hospitais, procedimentos, leitos, das unidades estaduais da saúde, informações de estatísticas de doenças e eventos vitais, a respeito do estoque de medicamentos e insumos, sobre a vigilância sanitária, em relação a conteúdos da COVID-19 (distribuição de insumos, vacinação, testes aplicados, Proposições para o Portal da Transparência do Poder Executivo do Estado de Santa Catarina com foco na área da saúde Revista Gestão e Secretariado (GeSec), São Paulo, SP, v. 14, n. 1, 2023, p. 277-305. 299 doações) e outros conteúdos da Saúde; 12 contéudos de Gestão Financeira (31,6%): englobam os gastos com medicamentos, a disponibilização de notas fiscais eletrônicas da Saúde, licitações nos hospitais e para a COVID-19, as informações financeiras das unidades de saúde mantidas por Organizações Sociais da Saúde (OSS); e 4 conteúdos de Informações Institucionais (10,5%): versam sobre a estruturação da saúde catarinense, legislação dos servidores da Saúde, quantidade de servidores e dados funcionais da Saúde e informações de concursos e seleções No atributo relevância retratam-se aspectos da importância e utilidade da divulgação do conteúdo, cada qual com suas particularidades. Na Confiabilidade consta como recomendação que seja apresentado o órgão central responsável pela informação (secretaria) havendo conteúdos incluindo a diretoria, setor ou sistema de geração dos dados. Na Linguagem encontram-se títulos para as chamadas dos conteúdos a partir de palavras da área e que procuraram trazer uma proximidade e entendimento de um vocabulário cidadão. Recomenda-se também, no momento de divulgação, incluírem frases curtas ou pequenos textos nas funcionalidades dos conteúdos para explaná-los, seguindo uma linguagem cidadã e com palavras cotidianas da área. No atributo apresentação constam alternativas de dois tipos de locais para ficarem disponíveis os conteúdos: uma página do Portal (desenvolvida com seu layout) ou uma página externa que estaria em outro ambiente e o Portal faria o direcionamento. Na Apresentação foram incluídas proposições dos tipos de páginas, sendo a tipo texto aquela com imagens, mapas, gráficos, textos, dashboards e outras funcionalidades e a do tipo consulta aquela que apresenta variedade de filtros, visualização de dados tabulados, gráficos e outros recursos associados. Incluído também sugestão para download das informações, focando nos formatos pdf e csv que atendem as necessidades de dados tabulados e de textos/relatórios descritivos. A plataforma de dados abertos do Estado seria a solução para disponibilizar outros formatos, abrindo possibilidade de criar acesso para a mesma a partir das páginas de cada conteúdo. A Tempestividade sugerida previu a atualização anual, mensal ou diária, variando de acordo com a necessidade de cada conteúdo, observada nas constatações na pesquisa. A Série Histórica teve também variações conforme o conteúdo, ficando concentrada dentro dos últimos 5 anos, o que proporciona a divulgação de séries recentes e incluem as demandas informacionais geradas dos pedidos de informação (ano de 2019 da pesquisa). O Quadro 11 apresenta uma síntese das propostas para serem avaliadas e desenvolvidas no Portal. Proposições para o Portal da Transparência do Poder Executivo do Estado de Santa Catarina com foco na área da saúde Revista Gestão e Secretariado (GeSec), São Paulo, SP, v. 14, n. 1, 2023, p. 277-305. 300 Proposta Diretriz para Aperfeiçoamento 1 - Desenvolvimento da Área Temática da Saúde - Criação de locais exclusivos (páginas temáticas) para a área da Saúde; - Agrupamento de todos os conteúdos das áreas, por subdivisões temáticas, nas páginas temáticas, com nomes que direcionam para páginas do Portal ou externas, onde abrigam os conteúdos; - Divulgação de algumas informações principais das áreas nas páginas temáticas. 2 - Convergência dos Conteúdos Existentes para a Área Temática - Implementação de processos de seleção dos conteúdos da Saúde para facilitar ao acesso e divulgação; - Desenvolvimento dos conteúdos selecionados nas páginas temáticas. 3 – Inclusão de Novos Conteúdos e Melhorias nos Existentes - Proposição de desenvolver melhorias em 11 conteúdos existentes do Portal; - Proposição de inclusão de 27 novos conteúdos no Portal; - Desenvolvimento dos conteúdos propostos nas páginas temáticas. Quadro 11 – Síntese das Propostas de Aperfeiçoamento Fonte: Dados da pesquisa (2021). A primeira e a segunda proposta trazem alternativas para que os cidadãos consigam visualizar, buscar e acessar facilmente os conteúdos. A primeira propõe o desenvolvimento da área temática da Saúde a partir da criação de páginas individualizadas para que se possa convergir, organizar e comportar a relação dos conteúdos a serem divulgados. Segue uma tendência observada em alguns portais da transparência e de dados abertos que organizam os conteúdos por temas, neste caso a Saúde. Este desenvolvimento facilitará a inclusão de dados e informações das áreas, como os recomendados neste estudo. A segunda proposta recomenda alternativas para que ocorram processos de seleção dos dados e informações existentes da área de Saúde no Portal, como a implementação de filtros automatizados nas consultas do Portal. O desenvolvimento destas medidas possibilitará que a população encontre com facilidade os conteúdos presentes da Saúde. A convergência de dados e informações atualmente disponíveis pode ser o conjunto inicial de conteúdos para as áreas temáticas uma vez que a tendência é implementações menos complexas. A terceira proposta explorou demandas da sociedade demonstrando a existência de uma variedade de conteúdos que podem ser implementados ou melhorados na gestão financeira, nas informações institucionais e de modo especial, em virtude do volume gerado e da aproximação com o dia a dia da população, nos conteúdos classificados na prestação de serviços da área. Citam-se as informações dos hospitais, de medicamentos e outros dados estatísticos da prestação de serviços que de forma geral agregarão um conjunto de conteúdos diferenciados para o Portal, expandindo a sua utilidade para a sociedade. Conclusões A pesquisa foi norteada pelo objetivo de propor diretrizes para o aperfeiçoamento da transparência de dados e informações públicas estaduais, relacionadas à área da Saúde, por Proposições para o Portal da Transparência do Poder Executivo do Estado de Santa Catarina com foco na área da saúde Revista Gestão e Secretariado (GeSec), São Paulo, SP, v. 14, n. 1, 2023, p. 277-305. 301 meio do Portal da Transparência do Poder Executivo de Santa Catarina. A primeira fonte de dados teve como origem os pedidos da informação da LAI do ano de 2019. A predominância dos pedidos foi para conteúdos que solicitaram os gastos e os estoques de medicamentos, podendo este último dispor de uma consulta apresentando a disponibilidade e o local que determinados medicamentos estariam sendo ofertados à população, facilitando o processo de divulgação antecipada e o conhecimento da situação de cada medicamento (disponibilidade de estoque e a previsão de abastecimento, no caso de ausência). A segunda fonte de dados identificou demandas de conteúdos na Secretaria de Estado da Saúde, por meio do encaminhamento de questionário para os setores de Ouvidoria e Comunicação. Faz-se menção a série histórica de dados estatísticos de doações e transplantes realizados pelo Estado nos últimos anos, conteúdo que pode ser divulgado para mostrar à sociedade a prestação deste serviço de destaque estadual. A terceira fonte de dados foi resultante de observações sistemáticas realizadas nos portais da transparência das Unidades Federativas (UF), tendo um volume grande de assuntos relacionados à COVID-19. O modelo de análise foi utilizado para classificar e apresentar as características de cada conteúdo a partir dos atributos da qualidade da informação. Na análise documental e de conteúdo das três fontes de dados ficou caracterizado que o volume maior de demandas informacionais está relacionado com a Prestação de Serviços. Demostra uma carência de conteúdos neste aspecto bastante importante no qual o Estado transmite aos cidadãos como estão sendo realizados e entregues os serviços públicos e os resultados da prestação de serviços. A pesquisa apresentou um conjunto de propostas, que foram divididas em três partes. A primeira propõe o desenvolvimento de áreas temáticas, criando páginas individualizadas para comportar, organizar e convergir o conjunto dos conteúdos de cada área. A criação das subdivisões temáticas (Gestão Financeira, Informações Institucionais, Prestação de Serviços), dentro de cada página temática, consiste em uma alternativa para agrupar a diversidade e quantidade de dados e informações das áreas. A segunda proposta trata da convergência dos conteúdos existentes para incorporarem nas áreas temáticas a partir de um conjunto de recomendações de processos de seleção dos dados e informações existentes, como a aplicação de filtros automatizados nas consultas do Portal. A terceira proposta recomenda um conjunto de novos conteúdos e melhorias em existentes para serem incorporados no Portal (nas páginas temáticas) que foram elaboradas a partir da coleta dos dados dos pedidos de informação da LAI, das informações das secretarias estaduais e da pesquisa nos portais das UF. As propostas reuniram alternativas para o aperfeiçoamento dos conteúdos da Saúde. Proposições para o Portal da Transparência do Poder Executivo do Estado de Santa Catarina com foco na área da saúde Revista Gestão e Secretariado (GeSec), São Paulo, SP, v. 14, n. 1, 2023, p. 277-305. 302 Do ponto de vista prático, orientam o desenvolvimento de uma página temática de Saúde que conteriam informações, a relação e o acesso dos conteúdos. Na proposta estão elencados conteúdos que poderão ser avaliados para incorporação no Portal, seja pela seleção de conteúdos presentes hoje ou pelo desenvolvimento de novos ou melhorias nos existentes. Outra reflexão que a pesquisa traz é para que os entes públicos observem, de forma sistematizada e periódica, as demandas dos pedidos de informação da LAI, voltadas ao interesse público, com vistas à uma provável implementação nos mecanismos de transparência ativa, o que pode minimizar as demandas recorrentes e semelhantes. As recomendações para estudos futuros podem ser dirigidas para realizar a pesquisa em outras áreas do governo estadual, como Segurança Pública, Defesa Civil, Agricultura e Assistência Social, aplicando a sistematização utilizada no presente estudo. Outro foco é a realização de pesquisas no âmbito municipal que podem ser aplicadas nas áreas governamentais. Estudos para explorar e compreender demandas informacionais de diversos setores da sociedade, como universidades, veículos de imprensa, observatórios sociais e outras Organizações Não Governamentais (ONG), que podem não estar chegando aos canais de comunicação direto do Estado e nos pedidos de informação da LAI, são recomendados para serem realizados em pesquisas futuras. Referências Akutsu, L., & Pinho; J. A. G. (2002). Sociedade da informação, accountability e democracia delegativa: investigação em portais de governo no Brasil. Revista de Administração Pública, 36(5), 723-746. Allegretti, D. D. S., & Platt Neto, O. A. (2010). Funcionalidades, limitações e potencialidades do Portal da Transparência do Estado do Rio Grande do Sul. Revista Catarinense da Ciência Contábil, 9(26), 79-95. Almeida, S. E. C. (2015). Lei de Acesso à Informação e transparência: proposta de inclusão de informações no portal transparência UFPB (Dissertação de Mestrado). Universidade de Federal da Paraíba, João Pessoa, Brasil. Angélico, F. (2015). Lei de Acesso à Informação: reforço ao controle democrático. São Paulo: Estúdio Editoras.com. Arouck, O. (2011). Atributos da Qualidade da Informação (Dissertação de Mestrado). Universidade de Brasília, Brasília, Brasil. Barbosa, M. C. B. (2018). A Lei de Acesso à informação no Brasil: análise da convergência entre os princípios da Lei e o conteúdo das solicitações dos usuários (Dissertação de Mestrado). Universidade de Brasília, Brasília, Brasil. Bento, L. V. (2015). Acesso a informações públicas: princípios internacionais e o direito Proposições para o Portal da Transparência do Poder Executivo do Estado de Santa Catarina com foco na área da saúde Revista Gestão e Secretariado (GeSec), São Paulo, SP, v. 14, n. 1, 2023, p. 277-305. 303 brasileiro. Curitiba: Juruá. Constituição da República Federativa do Brasil de 1988. 1988. Brasília, DF: Presidência da República, 1988. Recuperado de http://www.planalto.gov.br/ccivil_03/constituicao/constituicaocompilado.htm. Calvosa, M. V. D., Silva, T.A., & Krakauer, P. V. C. (2017). Portais Eletrônicos utilizados nos Municípios fluminenses: Análise das Ações Inovadoras. Sociedade, Contabilidade e Gestão, 2(2), 63-79. Carossi, D. F., & Teixeira Filho, J. G. A. (2016). Uma análise dos Pedidos de Acesso à Informação encaminhados a uma instituição de ensino superior. Revista Getão.Org, 14, 255-264. Castells, M. (1999). A sociedade em rede: A Era da Informação: economia, sociedade e cultura. São Paulo: Paz e Terra. Cucciniello, M., Bellè, N., & Valloti, G. (2014). Assessing public preferences and the level of transparency in government using an exploratory approach. Social Science Computer Review, 1-16. Curto-Rodríguez, R. (2020). Transparencia operativa de las comunidades autónomas españolas mediante sus portales de datos abiertos. El profesional de la información, 29(1), 1-13. Decreto Federal n° 7185, de 27 de maio de 2010. 2010. Recuperado de http://www.planalto.gov.br/ccivil_03/_ato2007-2010/2010/decreto/d7185.htm. Decreto Federal n° 10.540, de 5 de novembro de 2020. 2020. Recuperado de http://www.planalto.gov.br/ccivil_03/_ato2019-2022/2020/decreto/D10540.htm. Distrito Federal. Portal de Dados Abertos do Distrito Federal. 2021. Recuperado de http://www.dados.df.gov.br. Freire, F. R. F. (2014). Desafios para a Transparência Pública: um estudo com os usuários do Portal da Transparência do Governo Federal (Dissertação de Mestrado). Universidade de Brasília, Brasília, Brasil. Gama, J. R. (2015). Instrumentos de Transparência e acesso às informações públicas: um estudo das demandas por informações contáveis nas universidades federais (Tese de Doutorado). Universidade de Brasília, Brasília, Brasil. Heinen, J. (2015). Lei de Acesso à Informação: Lei n° 12.527/2011. Belo Horizonte: Fórum. Hullen, A. C. N. (2018). Cidadania e direitos sociais no Brasil: um longo percurso para o acesso aos direitos fundamentais. Revista de la Secretaría del Tribunal Permanente de Revisión, 6(11), 213-227. Jahns, F. T., & Raupp, F. M. (2016). Transparência do Poder Executivo dos Estados Brasileiros. Revista Universo Contábil, 12(3), 65-72. Kopits, G., & Craig, J. (1998). Transparency in government operations. Occasional Paper. International Monetary Fund. Washington DC, 158. Proposições para o Portal da Transparência do Poder Executivo do Estado de Santa Catarina com foco na área da saúde Revista Gestão e Secretariado (GeSec), São Paulo, SP, v. 14, n. 1, 2023, p. 277-305. 304 Lei Complementar n° 101, de 04 de maio de 2000. 2000. Recuperado de http://www.planalto.gov.br/ccivil_03/leis/lcp/lcp101.htm. Lei Complementar n° 131, de 28 de maio de 2009. 2009. Recuperado de http://www.planalto.gov.br/ccivil_03/leis/lcp/lcp131.htm. Lei n° 12.527, de 18 de novembro de 2011. 2011. Recuperado de http://www.planalto.gov.br/ccivil_03/_ato2011-2014/2011/lei/l12527.htm. Lei Complementar n° 141, de 13 de janeiro de 2012. 2012. Recuperado de http://www.planalto.gov.br/ccivil_03/leis/lcp/lcp141.htm. Lima, C. R. A., Schramm, J. M. A., Coeli, C. M., & Silva, M. E. M. (2009). Revisão das dimensões de qualidade dos dados e métodos aplicados na avaliação dos sistemas de informação em saúde. Cadernos de Saúde Pública, 10. Máchová, R., & Lnenicka, M. (2016). Evaluating the Quality of Open Data Portals on the National Level. Journal of Theoretical and Applied Electronic Commerce Research, 12(1), 21-41. Maranhã. Portal da Transparência do Estado do Maranhão. 2021. Recuperado de http://www.transparencia.ma.gov.br. Ministério do Planejamento, Orçamento e Gestão. (2016). VCGE Vocabulário de Governo Eletrônico 2.1.0 detalhado. Brasília: Ministério do Planejamento, Orçamento e Gestão. Recuperado de https://www.gov.br/governodigital/pt-br/governanca-de-dados/vocabulario-controlado-do-governo-eletronico. Moraes, R. (1999). Análise de conteúdo. Revista Educação (PUC-RS), 22(37), 7-32. Paraná. Portal da Transparência do Estado do Paraná. 2021. Recuperado de http://www.transparencia.pr.gov.br. Pinho, J. A. G. de. (2008). Investigando portais de governo eletrônico de estados no Brasil: muita tecnologia, pouca democracia. Revista de Administração Pública, 42(3), 471-493. Pinho, J. A. G. de, & Sacramento, A. R. S. (2009). Accountability: já podemos traduzi-la para o português? Revista de Administração Pública, 43(6), 1343-1368. Pinto, J. M. de R., & Adrião, T. (2006). Noções gerais sobre o financiamento da educação no Brasil. EccoS Revista Científica, 8(I), 23-46. Platt Neto, O. A., Da Cruz, F., Ensslin, S. R., & Ensslin, L. (2009). Publicidade e Transparência das Contas Públicas: Obrigatoriedade e Abrangência desses Princípios na Administração Pública Brasileira. Contabilidade Vista & Revista, 18(1), 75-94. Porumbescu, G. A. (2015). Using transparency to enhance responsiveness and trust in local government: Can it work? State and Local Government Review, 47(3), 205-213. Saldiva, P. H. N., & Veras, M. (2018). Gastos Públicos com saúde: breve histórico, situação atual e perspectivas futuras. Estudos Avançados, 32(92), 47-61. Proposições para o Portal da Transparência do Poder Executivo do Estado de Santa Catarina com foco na área da saúde Revista Gestão e Secretariado (GeSec), São Paulo, SP, v. 14, n. 1, 2023, p. 277-305. 305 Santa Catarina. Portal da Transparência do Poder Executivo de Santa Catarina. 2020. Recuperado de http://www.transparencia.sc.gov.br. Santa Catarina. Manual de Navegação do Portal da Transparência do Poder Executivo de Santa Catarina. 2018. Recuperado de http://www2.sef.sc.gov.br/transparenciasc/manual/Manual.pdf. São Paulo. Portal da Transparência do Estado de São Paulo. 2021. Recuperado de http://www.transparencia.sp.gov.br. Silva, A. A. P. S., Monteiro, D. A. A., & Reis, A O. (2020). Qualidade da Informação dos dados governamentais abertos: análise do portal de dados abertos brasileiro. Revista Gestão em Análise, 9(1), 31-47. Souza, R. J., Barbosa, R. B., Cabral, A. C. A., Santos, S. M. (2019). Avaliação da qualidade das informações disponibilizadas no portal da transparência do estado do Ceará. Encontros Bibli: revista eletrônica de biblioteconomia e ciência da informação, 24(54), 91-104 Trindade, A. L. B., Oliveira, M., & Becker, G. V. (2011). Análise dos atributos para avaliação da Qualidade da Informação nos ambientes de intranet para apoio à gestão do conhecimento. Revista Eletrônica de Administração, 70(3), 776–801. Zuccolotto, R., & Teixeira, M. A. C. (2019). Transparência: aspectos conceituais e avanços no contexto brasileiro. Brasília: Enap. Zuccolotto, R., Teixeira, M. A. C., & Riccio, E. L. (2015). Transparência: reposicionando o debate. Revista Contemporânea de Contabilidade, 12(25), 137-158. Submetido em: 12.12.2022 Aceito em: 11.01.2023 | RESEARCH GATE |
http://www.rasi.vr.uff.br RASI, Volta Redonda/RJ, v. 8, n. 1, pp. 07-22, Jan./Abr. 2022 Análise das práticas de logística reversa (LR) na região Sul Fluminense, após a implementação da Política Nacional de Resíduos Sólidos (PNRS) Janinne da Silva Bitencourt (UFF) - janinnebitencourt@id.uff.br Cecilia Toledo Hernández (UFF) - ctoledo@id.uff.br Recebido em 23/01/2021 Aceito em 17/09/2021 Resumo: A Política Nacional de Resíduos Sólidos (PNRS) é um marco brasileiro em relação a gestão dos resíduos sólidos. Neste contexto, a legislação, não somente a brasileira, se torna cada vez mais rigorosa em relação ao descarte de produtos no final de seus ciclos de vida. Este artigo utilizou o critério de Leite (2012) que classificava os programas em econômicos, de serviço ao cliente, legais, de cidadania corporativa e de imagem corporativa. Neste sentido, paira a dúvida se realmente o programa legal está sendo considerado importante nas empresas. Com base neste contexto, o objetivo da pesquisa é avaliar a importância de motivadores da logística reversa (LR) em empresas da Região Sul Fluminense, no ramo automobilístico, após a implementação da PNRS. O estudo tem caráter aplicado e exploratório e a coleta de dados foi feita por meio de questionários. A fim de analisar as informações obtidas, utilizou-se o método Analytic Hierarchy Process (AHP). Nesta pesquisa, o programa econômico foi considerado o mais importante seguido do programa legal. Outro estudo feito na Região Sul Fluminense também teve como destaque o programa econômico. Entretanto, em grau de priorização, o programa legal não foi considerado um motivador de LR, sendo que na atual pesquisa este se encontra em segundo lugar em termos de importância. Como recomendação de pesquisa futura, a análise pode ser feita em outras regiões e ramos. Palavras-Chave: Economia Circular; Logística Reversa; Multiple Criteria Decision Making (MCDM); Política Nacional de Resíduos Sólidos (PNRS). Analysis of reverse logistics practices (LR) in the South Fluminense region, after the implementation of the National Solid Waste Policy (PNRS) Abstract: The National Solid Waste Policy (PNRS) is a Brazilian landmark in relation to solid waste management. In this context, legislation, not only the Brazilian one, becomes increasingly strict regarding the disposal of products at the end of their life cycles. This article uses Leite's (2012) criteria, which classifies programs in economics, customer service, legal, corporate citizenship, and corporate image. In this sense, the question remains whether the legal program is really being considered important in companies. Based on this context, the objective of the research is to evaluate the importance of reverse logistics (RL) motivators in companies in the South Fluminense region, in the automobile sector, after the implementation of the PNRS. The study has an applied and exploratory character and data collection was done through questionnaires. To analyze the information, the Analytic Hierarchy Process (AHP) method was used. In this research, the economic program was considered the most important followed by the legal program. Another study carried out in the South Fluminense region also highlighted the economic program. However, in terms of prioritization, the legal program was not considered a RL motivator, and in the current research it is ranked second in terms of importance. As a recommendation for future research, an analysis can be done in other regions and branches. Keywords: Circular Economy; Reverse logistic; Multiple Criteria Decision Making (MCDM); National Solid Waste Policy (PNRS). R. Desembargador Ellis Hermydio Figueira, 783, Bloco A, sl. 218, Aterrado. 27213-415 - Volta Redonda, RJ – Brasil www.uff.br Copyright © 2020 RASI. Todos os direitos, até mesmo de tradução, são reservados. É permitido citar parte de artigos sem autorização prévia, desde que seja identificada a fonte. Bitencourt & Hernández |8 10.20401/rasi.8.3.623 Avaliado pelo sistema Double Blind Review Editor: Marcelo G. Amaral RASI, Volta Redonda/RJ, v. 8, n. 1, pp. 07-22, Jan./Abr. 2022 http://www.rasi.vr.uff.br Análise das práticas de logística reversa (LR) na região Sul Fluminense, após a implementação da Política Nacional de Resíduos Sólidos (PNRS). 1. Introdução Temáticas relacionadas a sustentabilidade adquirem notório destaque com a atual propensão de crescimento populacional, com maiores montantes de desperdícios, derivados de processos industriais, poluição da água, do ar ou da terra e com uma crescente consciência ambiental na sociedade, devido a legislações mais rigorosas e produtos com menores ciclos de vida (Bhandari et al., 2019). Neste contexto, vários assuntos como a logística reversa (LR) e a economia circular (EC) vem sendo discutidos. Alguns estudos buscam tornar mais claro o conceito de economia circular ao fazer uma revisão sistemática da literatura (Farooque et al., 2019; Geisendorf & Pietrulla, 2018; Geissdoerfer et al., 2017). Nestes estudos, é mostrada uma gradativa preocupação ambiental ao redor do mundo, e muitos destes destacam a China, como um país que mostra uma progressiva preocupação com a temática em diversos ramos da economia (Geng et al., 2012; Ness, 2008; Su et al., 2013). No cenário brasileiro, a Política Nacional de Resíduos Sólidos (PNRS), sancionada em 2010, deu grande notoriedade a EC, pois esta marca a gestão dos resíduos sólidos no Brasil (Oliveira et al., 2019). Percebe-se em todo este novo cenário, que o conceito de LR se relaciona com a EC, sendo um de seus princípios e sua relevância aumenta à medida que ela conceitua e clarifica a gestão de retorno de produtos. Desta maneira, uma parte crítica do gerenciamento das organizações é o fluxo reverso, tratado como um fator estratégico. Na perspectiva brasileira, ainda há a necessidade de maior disseminação do conceito da EC, pois falta divulgação local e nacional do modelo, o que ainda coloca o país em posição inferior de pesquisas em relação a muitos países da Europa e da China (Oliveira et al., 2018). Diante do ainda estágio inicial de pesquisa nestes assuntos no cenário nacional, o artigo buscou responder a seguinte questão: Quais são os programas de LR priorizados após a PNRS no ramo automobilístico? Desta maneira, o objetivo geral desta pesquisa é avaliar programas e atividades neste ramo, relacionados à LR na região Sul Fluminense, através do método de auxílio multicritério a decisão, Analytic Hierarchy Process (AHP). Entretanto, a pesquisa se delimita espacialmente, pois analisa somente empresas de uma região específica. A LR é importante para o fortalecimento das empresas, porém ainda existem muitas barreiras na implementação da LR, especialmente no ramo automobilístico (Lima, 2016; Vieira et al., 2020). As principais barreiras no contexto brasileiro, são internas, como econômicas, de infraestrutura e gestão das empresas e externas as organizações, como falta de controle e fiscalização em relação a legislação. Desta maneira, conhecer as práticas de LR no ramo automobilístico é importante para uma melhor gestão do fluxo reverso, tanto para os gestores privados e públicos, no sentido de analisar se os programas legais vêm ganhando notoriedade e quais necessitam de maior atenção. Assim, estruturar um canal de fluxo reverso é relevante por questões ambientais, bem estar social e no aspecto econômico, uma vez que pode gerar ganhos de competividade nos produtos nacionais (Lima, 2016). Este trabalho se estrutura em seis seções, sendo a seção 2, uma fundamentação teórica sobre a economia circular e especificamente sobre a LR, a seção 3, um tratamento das questões metodológicas da pesquisa, a seção 4, uma exposição dos resultados obtidos com a utilização do método AHP, a seção 5, a discussão dos resultados e a seção seis, uma Bitencourt & Hernández |9 10.20401/rasi.8.3.623 Avaliado pelo sistema Double Blind Review Editor: Marcelo G. Amaral RASI, Volta Redonda/RJ, v. 8, n. 1, pp. 07-22, Jan./Abr. 2022 http://www.rasi.vr.uff.br apresentação das principais conclusões e sugestões de trabalhos futuros, seguidos das referências bibliográficas. 2. Referencial Teórico A fundamentação teórica tem como objetivo permitir a análise do tema objeto de estudo. Para isto, foram analisados diferentes aspectos sobre o tema de pesquisa como a sustentabilidade, a economia circular e lei de resíduos sólidos no Brasil, a LR e sua relação com a simbiose industrial, além de programas de LR e o uso da tomada de decisão com múltiplos critérios. 2.1. Sustentabilidade Dentro da temática da sustentabilidade, caracterizada como o processo pelo qual a atividade humana minimamente influencia o ecossistema terrestre, uns dos principais conceitos sustentáveis é o tripé da sustentabilidade que estabelece três dimensões para as organizações, a econômica, a ambiental e a social (Batista et al., 2018). A dimensão econômica se concentra em construir capital financeiro. Gerar capital social preocupar-se com a responsabilidade interna ou externa das empresas, caracteriza a dimensão social da sustentabilidade. A dimensão ambiental teria como objetivo, minimizar o uso de recursos naturais (Elkington, 1998). Dentro desta temática, alguns conceitos emergiram e são semelhantes a EC, apesar de possuírem algumas diferenças entre si, como o conceito “cradle to cradle”, as cadeias de suprimentos fechadas, a ecologia industrial, a LR, que será estudada neste artigo, e a biomimética (Geisendorf & Pietrulla, 2018). Dentro do contexto sustentável, a principal definição de economia circular é dada pela fundação Ellen MacArthur Foundation (2015), que a define como um sistema restaurativo e regenerativo intencionalmente (Farooque et al., 2019; Geisendorf & Pietrulla, 2018). Outra famosa definição se dá ao contrapor-se a economia linear, modelo de “pegar, fazer e descartar”, pois, este modelo é ineficiente em relação ao tripé da sustentabilidade (Ghisellini et al., 2016; Ness, 2008). Percebe-se com o tripé da sustentabilidade, que implementar o desenvolvimento sustentável nas empresas não é de simples operacionalização. Assim, é necessário para isto, mudanças na tomada de decisão e no modo de produção e consumo (Lang et al., 2012). Para enfrentar todas estas barreiras da sustentabilidade, Escrig-Olmedo et al. (2019), expuseram a necessidade de uma ampla geração de conhecimento em relação aos principais princípios da sustentabilidade e a interação entre diferentes gerações (a passada, a presente e a futura). Para vencer os desafios e gerar maior conhecimento no campo da sustentabilidade, o desenvolvimento sustentável deve ser operacionalizado, surgindo para isto alguns conceitos mais estruturados como a EC e a LR (Geissdoerfer et al., 2017). 2.2. Economia Circular e Lei de Resíduos Sólidos no Brasil O conceito de EC é relativamente novo e teve sua definição proposta por Boulding (1966) com o livro The Economic of the Coming Spaceship Earth (Geisendorf & Pietrulla, 2018). Adicionalmente, a EC tem diferentes níveis de implementação ao redor do mundo (Ghisellini et al., 2016).A China, em relação a implementação da EC, se destaca mundialmente com vários estudos nesta área de estudo (Wang & Wang, 2020).Parte desta grande influência chinesas e deve especialmente a promoção da Lei de Economia Circular, Bitencourt & Hernández |10 10.20401/rasi.8.3.623 Avaliado pelo sistema Double Blind Review Editor: Marcelo G. Amaral RASI, Volta Redonda/RJ, v. 8, n. 1, pp. 07-22, Jan./Abr. 2022 http://www.rasi.vr.uff.br homologada em 2008 que impulsiona o fechamento de processos produtivos diversos (Machado et al., 2012). Nos Estados Unidos, a Agência de Proteção Ambiental dos Estados Unidos (EPA) é parte essencial da implementação da EC, uma vez que regulariza a gestão de resíduos domésticos, perigosos ou não, com base na lei de Conservação e Recuperação de Recursos (RCRA) (EPA, 2002). Na União Europeia, há uma diretiva que regula a gestão de resíduos, a diretiva 2008/98/EU, responsável por reutilizar, reciclar e recuperar energia. Em relação ao cenário brasileiro, o Brasil ainda precisa ter uma maior disseminação do conceito da EC, uma vez que falta divulgação local e nacional do modelo, o que coloca o país ainda em posição inferior de pesquisas em relação a muitos países da Europa e da China (Oliveira et al., 2018). A EC possui uma grande importância no mundo e no Brasil para se implantar processos sustentáveis e de acordo com a sua mais famosa definição, dada por Ellen MacArthur Foundation (2015), a EC é um sistema restaurador e regenerativo, que percebe os produtos ao final de sua vida útil, não como resíduos, assim como no sistema linear, mas como recursos. Assim, para efetiva implementação do sistema circular, destaca-se a gestão de resíduos como um dos principais processos do conceito. A PNRS foi o marco na gestão dos resíduos sólidos no Brasil, pois preencheu uma lacuna na legislação brasileira ambiental baseado na gestão internacional e suas legislações. Esta política brasileira destaca a responsabilidade compartilhada pelo ciclo de vida do produto, acordos setoriais e LR (GRS-UFPE, 2014). A responsabilidade compartilhada envolve todos os envolvidos no ciclo de vida dos produtos, os fabricantes, importadores, distribuidores e comerciantes que devem se preocupar com a destinação final e o impacto do resíduo, sendo a LR fundamental neste processo (CNI, 2019). Entretanto, a PNRS enfrenta muitos desafios em sua implementação ao longo do país. Falta de fundos governamentais, falta de capacidade técnica e administrativa, falta de conscientização do público em relação à geração de resíduos e insuficiência da estratégia de descentralização do governo federal na implementação da agenda (Cetrulo et al., 2018). 2.3. Logística Reversa (LR) De acordo com o Council of Supply Chain Management Professionals, a LR é uma área especializada da logística, que foca no movimento de produtos após a venda, incluindo os produtos que são devolvidos. No que tange ao aumento da conscientização, em relação a temáticas sustentáveis e as necessidades dos clientes cada vez mais inovadores, a LR se tornou tão importante quanto a logística em seu fluxo direto (Vlachos, 2016). Assim, alguns dos motivadores principais para se implementar a LR nos diversos processos produtivos são fatores financeiros, necessidades dos clientes e legislações mais rigorosas (Vlachos, 2016). Alguns autores buscaram relacionar a LR como um pilar da EC, de modo a destacar a importância que a LR tem na implementação do modelo circular. Estes dois conceitos são tidos como semelhantes, porém o conceito de EC é mais amplo, uma vez que não cobre só o lado reverso nas cadeias de suprimentos (Beiler et al., 2020). Ao relacionar os conceitos de EC e LR, os ramos da economia mais estudados nesta relação de conceitos são o da construção civil, de embalagens, de reciclagem, incluindo os catadores, do poliestireno, comida e bebida, ramo automobilístico e de equipamentos elétricos eletrônicos. Ainda são poucas as publicações em EC separadamente da LR ao perceber mais trabalhos que se concentram em estudar somente a LR, e não como ela se inclui na EC (Cosenza et al., 2020). Bitencourt & Hernández |11 10.20401/rasi.8.3.623 Avaliado pelo sistema Double Blind Review Editor: Marcelo G. Amaral RASI, Volta Redonda/RJ, v. 8, n. 1, pp. 07-22, Jan./Abr. 2022 http://www.rasi.vr.uff.br O conceito mais difundido e que se insere na ecologia industrial (EI), é a simbiose industrial (SI), que é essencial para fechar os ciclos produtivos, ou seja, é de grande importância na implementação da LR. A EI se baseia na comparação da natureza e do ambiente industrial (Trevisan et al., 2016). Já a SI é definida como a troca que ocorre entre materiais, energia, água e coprodutos (Ferreira et al., 2021). Assim, como a LR, a SI destaca a importância da colaboração entre os diferentes stakeholders, sejam eles, funcionários e gestores privados, sociedade, governo (Mota & Paula, 2020; Pinheiro et al., 2020). Assim, é essencial que o setor público e privado crie um ambiente que seja propício para a prática da SI e LR. Além destes conceitos que se relacionam de alguma forma, motivadores da LR, barreiras e aplicação de métodos de tomada de decisão em alguns ramos específicos da economia são alguns dos principais temas estudados por Abbas e Farooquie (2020) e Agrawal e Singh (2019). 2.4. Programas de LR: Motivadores e Barreiras Razões econômicas, ambientais, legislação e necessidade de se destacar no mercado competitivo são alguns dos motivadores para se implementar a LR. Entretanto, muitas são as barreiras desta implementação (Brito & Dekker, 2004). Demajorovic et al. (2012), destacam a importância de motivadores econômicos, de imagem e de serviço ao cliente. Couto e Lange (2017) definem alguns aspectos importantes, como legislação e normatização, aspectos tributários e financeiros, controle governamental e participação da população, para se implementar a LR no contexto brasileiro. Realizou-se uma pesquisa bibliográfica a partir das bases de dados bibliográficas científico-técnicas, Web of Science e Scopus, após a homologação da PNRS, em 2010. As palavras-chaves contidas no título, palavras-chaves ou resumo utilizadas para recuperar os artigos são: Reverse Logistics, Drivers ou Reverse Logistics Programs ou Reverse Logistics Barriers. Isso resultou em mais de 60 artigos. Foi realizado um processo de triagem, em que os títulos e resumos foram analisados para assegurar que seu tema principal estava adequado ao escopo da pesquisa. O conjunto final de artigos usados neste estudo para identificar as práticas de LR foram de 35 artigos. Percebe-se que os diferentes motivadores de LR, faz com que exista diferença nos objetivos dos programas estabelecidos por alguns autores. Adicionalmente, com esta pesquisa bibliográfica, percebeu-se que todos os autores classificavam seus programas seguindo a lógica dos cinco programas, estudados por Leite (2012), Hernández (2010), Hernández et al. (2012), Mangla et al. (2012), Eltayeb et al. (2010), Govindan e Bouzon (2018) e Kumar e Putnam (2008), definem programas econômicos, de serviço ao cliente, legais, de cidadania corporativa e de imagem corporativa, assim como Leite (2012). 3. Método A pesquisa pode ser classificada como aplicada quanto a sua natureza, ou seja, tem por objetivo gerar conhecimentos para uma aplicação prática ao solucionar problemas específicos (Gerhardt & Silveira, 2009). No caso deste artigo, a pesquisa terá aplicação prática ao analisar motivadores de LR na região Sul Fluminense. Na classificação quanto aos objetivos, seu caráter é exploratório uma vez que tem por finalidade fazer com que o problema se torne mais explícito (Gil, 2008). A pesquisa pode ser definida como qualitativa, pois os dados são coletados segundo uma escala qualitativa e toda Bitencourt & Hernández |12 10.20401/rasi.8.3.623 Avaliado pelo sistema Double Blind Review Editor: Marcelo G. Amaral RASI, Volta Redonda/RJ, v. 8, n. 1, pp. 07-22, Jan./Abr. 2022 http://www.rasi.vr.uff.br a análise numérica posterior se dá a partir desta. O estudo seguiu os quatro passos apresentados na Figura 1. Figura 1. Etapas da Pesquisa Fonte: elaboração própria (2021). Além da classificação da pesquisa, uma etapa importante da metodologia é a descrição dos procedimentos metodológicos e o seu desenvolvimento. Segundo Creswell (2007), uma etapa importante no procedimento de pesquisa é o instrumento de coleta e análise de dados. O instrumento de coleta de dados foi o questionário, baseado em Hernández (2010), que contava com questões abertas em uma primeira fase e com questões fechadas, para julgamentos da análise multicritério, em uma segunda fase. O objetivo inicial, para o levantamento qualitativo, foi a obtenção de informação abrangente sobre as características dos programas e atividades de LR. A segunda fase é aquela na qual os profissionais de LR definiram quais programas e atividades estão presentes em suas empresas. No caso deste artigo, a amostra foi selecionada com um professor do curso de especialização de um centro universitário de Volta Redonda, de maneira presencial. Esta amostra contou com 3 profissionais especializados na área de logística, que representaram três montadoras automobilísticas importantes da Região Sul Fluminense. Deste modo, a amostragem foi feita por conveniência e consome pouco tempo quando comparada com outras técnicas Segundo Jonathas e Paiva (2020), o automobilístico nesta região conta com cinco montadoras, MAN Latin America, PSA Peugeot Citroën, Renault-Nissan, Hyundai e Jaguar Land Rover. Apesar dos especialistas da amostra serem acessíveis e dispostos a cooperar, uma das desvantagens desta coleta foi a existência de vieses nas respostas, de maneira a utilizar práticas de LR destacados na literatura para melhor aproveitamento dos questionários. Adicionalmente, a amostra foi formada somente por especialistas da região Sul Fluminense, uma vez que o panorama depois da homologação da PNRS, exposto neste artigo, será comparado com panorama anterior explicitado pelo estudo de Hernández (2010) que também utilizou a região Sul Fluminense como referência. Para definir os programas e atividades de LR que comporiam o questionário, foi utilizado o critério de Leite (2012). Segundo classificação do autor, o programa econômico visa otimizar a obtenção de lucros, o de serviço ao cliente visa destacar a empresa no mercado e diferenciá-la quanto aos serviços prestados aos clientes, o legal visa cumprir com as legislações existentes, o de cidadania corporativa responsabiliza as organizações como socialmente responsáveis e o de imagem tem como objetivo reforçar a imagem das empresas no mercado. Nestes programas, foram definidas atividades de LR que são realizadas de acordo com modelo proposto por Hernández (2010) na Tabela 1. Bitencourt & Hernández |13 10.20401/rasi.8.3.623 Avaliado pelo sistema Double Blind Review Editor: Marcelo G. Amaral RASI, Volta Redonda/RJ, v. 8, n. 1, pp. 07-22, Jan./Abr. 2022 http://www.rasi.vr.uff.br Tabela 1. Programas e Atividades de LR Programas de LR Atividades ou alternativas de LR Programas Econômicos (PE) -Materiais retornados ao processo produtivo. (A1) -Reuso de embalagens e venda como matéria-prima para outros processos. (A2) -Revenda de produtos em mercados secundários. (A3) -Reciclagem. (A4) -Existência de registro dos custos gerados pelas devoluções. (A5) -Desembolso por ações sociais e meio ambientais. (A6) -Despesas por treinamento de funcionários. (A7) -Custos para operar o canal reverso (coleta, seleção, transporte, armazenagem). (A8) -Custos para desenvolver novas tecnologias. (A9) Programas de Imagem (PI) -Propaganda como empresa responsável quanto aos seus produtos e processos. (A10) -Desenvolvimento de novas tecnologias para aproveitar os materiais reciclados. (A11) -Destino adequado aos resíduos. (A12) Programas de Cidadania Corporativa (PC -Projetos sociais. (A13) -Projetos educacionais. (A14) - Criação de emprego para operar o canal reverso. (A15) Programas de Serviço ao Cliente (PS) -Parcerias com stakeholders. (A16) -Políticas de retorno liberais. (A17) -Fidelização de clientes. (A18) -Retornos bem definidos. (A19) ProgramasLegais (PL) -Responsabilidade das empresas pelo destino correto de seus produtos no fim da vida útil. (A20) -Estabelecimento de níveis mínimos de recuperação a serem cumpridos pelas empresas. (A21) Fonte: elaboração própria (2021). Nota-se pela Tabela 1, que diferentes alternativas foram estabelecidas para os cinco programas de LR estabelecidos por Leite (2012). Após a coleta de dados, foi medida a importância das práticas de LR, com o uso do AHP, devido a sua maior simplificação e conveniência. O AHP, um método de auxílio multicritério à decisão, foi desenvolvido por Saaty (1980) e tem como objetivo resolver problemas complexos, mas é considerada uma técnica de simples manejo. O AHP é um dos métodos de auxílio multicritério à decisão utilizado para resolver problemas complexos, mas de simples manejo. Para aplicar o AHP foram seguidas as seguintes etapas baseadas em Saaty (2010): 1. Representação da hierarquia: O problema teve como objetivo analisar a priorização dos programas de LR na indústria automotiva. Para isto foi decomposto em dois níveis: critérios (programas de LR) e alternativas (atividades de LR desenvolvidas pelas empresas). 2. Comparação de pares: Comparação entre si e em pares de todos os critérios e das alternativas com respeito a cada critério. Para a construção destas matrizes de comparações pareadas, utilizou a escala de Saaty que utiliza uma escala de importância entre 1 e 9. Nesta escala, 1 significa que os elementos comparados contribuem de igual forma para o objetivo e na medida que os valores aumentam a importância de um com respeito ao outro também aumenta até o 9, onde um elemento tem importância absoluta sobre outro. Bitencourt & Hernández |14 10.20401/rasi.8.3.623 Avaliado pelo sistema Double Blind Review Editor: Marcelo G. Amaral RASI, Volta Redonda/RJ, v. 8, n. 1, pp. 07-22, Jan./Abr. 2022 http://www.rasi.vr.uff.br 3. Método dos autovalores: Nesta etapa, foram estimados os pesos dos elementos de decisão (peso dos critérios/programas e das alternativas/atividades), utilizando-se a síntese hierárquica. 4. Avaliação da consistência das comparações por pares: Foi calculado o índice de consistência das matrizes que não deve exceder 0,10. 5. Agregação das prioridades: As prioridades relativas foram agregadas para se avaliar os resultados e desta forma obter um ranking dos critérios/programas e das alternativas/atividades através das prioridades globais. Este último resultado com a prioridade global de critérios/programas de LR e alternativas/atividades de LR é analisado e comparado com estudo de Hernández (2010). 4. Resultados Foi realizada a comparação par a par dos critérios, a fim de medir a importância de cada um dos programas de LR. Essa comparação, em formato de matrizes será apresentada e discutida nesta seção. De acordo com Lima (2016), estudos como este é fundamental para o debate sobre a implementação da LR no ramo automobilístico brasileiro, de modo a contribuir para o avanço do tema e para possíveis soluções que se adequem ao contexto brasileiro. 4.1. Ramo Automobilístico Para a aplicação do método AHP, foi estruturada a matriz hierárquica da Figura 2. No primeiro nível, encontra-se o objetivo, seguido dos critérios do modelo e das alternativas. Os critérios são os programas de LR identificados pelos respondentes e as alternativas são as atividades de LR realizadas nas empresas analisadas. Figura 2. Árvore hierárquica com os critérios e alternativas no ramo automobilístico Fonte: elaboração própria (2021). Como pode ser observado na Figura 2, no ramo automobilístico foram destacados 3 programas e 5 atividades Bitencourt & Hernández |15 10.20401/rasi.8.3.623 Avaliado pelo sistema Double Blind Review Editor: Marcelo G. Amaral RASI, Volta Redonda/RJ, v. 8, n. 1, pp. 07-22, Jan./Abr. 2022 http://www.rasi.vr.uff.br Como etapa seguinte foram necessárias desenvolver 4 matrizes de comparação (uma matriz para a comparação dos critérios). A Tabela 2 mostra a comparação entre os critérios/programas de LR estabelecidos para o ramo automobilístico Tabela 2. Matriz comparativa entre os critérios (Programas de LR) Matriz Comparativa PE PL PI Vetor de prioridades Programa Econômico-PE 1 4 9 0,71707 Programa Legal-PL 1 4 0,21717 Programa de Imagem-PI 1 0,06577 Fonte: elaboração própria (2021). Observa-se pela Tabela 2, que o vetor de prioridades obtido dita que o programa econômico tem uma prioridade maior em relação aos outros, seguido do programa legal. Para verificar se o modelo é consistente, Saaty e Sodenkamp (2010), destacam a Taxa de Consistência (CR), que é determinado pela razão entre o Índice de Consistência e o Índice de Consistência Aleatório. A matriz será considerada consistente se a proporção resultante for menor que 10%. Para a comparação dos critérios/programas de LR da Tabela 2, obteve-se um índice de consistência de 0,03548, ou seja aproximadamente 3,55%, valor menor que o valor de 10% ou 0,1, o que mostra a consistência da matriz. De igual forma foram elaboradas outras três matrizes que permitiram a comparação entre as alternativas/atividades de LR com respeito a cada critério/programa de LR. O resultado de cada vetor de prioridade é mostrado na Tabela 3. Tabela 3. Prioridade de Cada Alternativa com Respeito aos Programas Matriz Comparativa PE PL PI Reuso, revenda e reciclagem- A2, A3, A4 0,29582 0,15070 0,08913 Custos por atividade de LR- A8, A9 0,52842 0,15070 0,03601 Propaganda e novas tecnologias- A10, A11 0,03254 0,02548 0,53418 Destino adequado aos resíduos- A12 0,05914 0,02548 0,30376 Responsabilidade e cumprimento da lei- A20, A21 0,08409 0,28560 0,03693 Fonte: elaboração própria (2021). Percebe-se pela Tabela 3, que: - As atividades A8, A9, custos para operar e desenvolver novas tecnologias no canal reverso, são as mais destacadas ou prioritárias quando analisados os programas com objetivos econômicos. - Quando avaliadas as atividades com respeito ao programa legal, as mais importantes são as A20 e A21, relacionadas à responsabilidade por cumprimento da lei. - Com respeito ao programa de imagem, as atividades de maior peso são as relacionadas à propaganda e desenvolvimento de novas tecnologias, atividades A10 e A11. O valor do CR para as matrizes ficou entre 0,02982 e 0,09739, menores que o valor 0,1, mostrando valores de CR consistentes. Por último foi calculado o desempenho global de cada programa de LR no ramo automobilístico, além do desempenho de cada alternativa como mostrado na Tabela 4. Bitencourt & Hernández |16 10.20401/rasi.8.3.623 Avaliado pelo sistema Double Blind Review Editor: Marcelo G. Amaral RASI, Volta Redonda/RJ, v. 8, n. 1, pp. 07-22, Jan./Abr. 2022 http://www.rasi.vr.uff.br Tabela 4. Prioridades dos programas de LR Critérios/Alternativas Prioridades Programas Econômicos - PE 0,71706 Programas Legais - PL 0,21717 Programas de Imagem - PI 0,06577 Reuso, revenda e reciclagem - A2, A3, A4 0,28222 Custos por atividade de LR - A8, A9 0,49603 Propaganda e novas tecnologias - A10, A11 0,06553 Destino adequado aos resíduos - A12 0,07523 Responsabilidade e cumprimento da lei - A20, A21 0,08099 Fonte: elaboração própria (2021). Percebe-se pela Tabela 4, que algumas observações podem ser feitas em relação ao ramo automobilístico. Foram destacados somente três programas como existentes na empresa. O programa econômico foi considerado o mais importante, seguido do programa legal. Adicionalmente, a atividade A8 e A9, custos para operar o canal reverso e custos por desenvolver novas tecnologias relacionadas com o programa econômico, são os que têm maior importância. Estes resultados mostram que o programa econômico ainda é o que possui maior importância, porém, o programa legal vem ganhando destaque, o que pode servir para uma comparação do panorama da LR antes e depois da PNRS. Desta maneira, é importante conhecer as práticas de LR que estão sendo utilizadas neste ramo, uma vez que a implementação da LR gera vários benefícios sociais. Os resultados podem servir para estruturar políticas públicas que fomentem o maior diálogo entre todos os agentes envolvidos, montadoras, recicladores, consumidores finais e governo. Adicionalmente, conhecer as práticas de LR no ramo automobilístico também podem atrair maiores investimentos, o que beneficia as empresas e o governo, de modo a aumentar o apoio governamental, ainda considerado pouco (Lima, 2016). 5. Discussão dos Resultados Pode-se perceber que o programa legal quando comparado com outros estudos realizados em empresas brasileiras mostrou uma mudança significativa na importância. Quando comparado ao estudo de Hernández (2010), o programa econômico teve maior destaque. Entretanto, em grau de priorização, o programa legal não aparecia como um programa estabelecido para o ramo automobilístico, sendo o programa de cidadania, considerado o segundo mais importante. Entretanto, neste artigo, este se encontra em segundo lugar em termos de importância. Assim, nota-se que o programa legal teve sua relevância aumentada, o que pode indicar a importância da instalação da lei da PNRS. Adicionalmente, tem-se o estudo de Leite et al. (2006), que inclui o ramo automobilístico nos ramos estudados. Neste estudo, o programa econômico também foi destacado, com 50% de priorização, ao considerar quatro critérios, o econômico, a competitividade, a legislação e o meio ambiente, que incluiu aspectos de imagem e cidadania. A Tabela 5mostra os critérios/programas priorizados no estudo de Hernández (2010) e neste artigo, para o ramo automobilístico em específico. Bitencourt & Hernández |17 10.20401/rasi.8.3.623 Avaliado pelo sistema Double Blind Review Editor: Marcelo G. Amaral RASI, Volta Redonda/RJ, v. 8, n. 1, pp. 07-22, Jan./Abr. 2022 http://www.rasi.vr.uff.br Tabela 5. Prioridades dos programas de LR relacionadas a indústria automobilística no estudo de Hernández (2010) e neste artigo Critérios/Programas Vetor de prioridades (Hernandez 2010) Vetor de prioridades (Neste estudo) Programas econômicos - PE 0,73064 0,71707 Programas de imagem - PI 0,08096 0,06577 Programas de cidadania - PC 0,18839 Não relatado Progama Legal - PL Não relatado 0,21717 Fonte: elaboração própria (2021). Ao analisar a Tabela 5, nota-se que se mantêm três programas priorizados para este ramo. Entretanto, no estudo de 2010, os priorizados foram os critérios/programas econômicos, de imagem e de cidadania e no atual, o econômico, o legal e o de imagem. Nota-se que em 2010, o programa legal nem foi relatado como existente. Em relação as atividades priorizadas no estudo de Hernández (2010), foi priorizado para o ramo automobilístico, a atividade de desenvolvimento de embalagens retornáveis e no atual estudo, as atividades relacionadas com coleta, seleção, transporte, armazenagem e o desenvolvimento de novas tecnologias e seus respectivos custos. Em todo este cenário, a PNRS, homologada em 2010, preencheu uma lacuna na legislação brasileira ambiental baseado na gestão internacional e suas legislações. A principal mudança em relação a gestão dos resíduos se deu com o destaque da responsabilidade compartilhada pelo ciclo de vida do produto, acordos setoriais, LR, sustentabilidade operacional e financeira e envolvimento de padrões sustentáveis de produção e consumo (GRS-UFPE, 2014). Segundo esta lei, a LR seria parte importante para fazer o correto gerenciamento de resíduos e otimização do ciclo de vida dos produtos e estaria definida como um conjunto de procedimentos e atos que possam restituir os resíduos sólidos para as empresas (Luna & Viana, 2019). Desta maneira, a LR, principal instrumento da PNRS, é importante na gestão de resíduos e em vários outros ramos, além do automobilístico. Barbosa et al. (2019), destacam que gestores brasileiros na indústria de laticínio possuem conhecimento superficial quando se trata de desenvolvimento sustentável e citam como barreira principal, o alto custo de investimento. Esta barreira se enquadra nas principais dificuldades de implementação da PNRS, como falta de fundos governamentais, falta de capacidade técnica e administrativa, falta de conscientização do público em relação à geração de resíduos, falta de articulação da PNRS com outras políticas, falta de instrumentos de controle e cobrança e necessidade de maior articulação entre os entes federativos (Cetrulo et al., 2018; GRS-UFPE, 2014). 6. Considerações Finais Ao analisar os resultados obtidos na seção 4, nota-se que o programa legal se torna mais relevante com a legislação brasileira cada vez mais rigorosa devido ao marco na gestão de resíduos, a PNRS. Entretanto, o Brasil ainda tem uma legislação defasada em relação a outros países. O objetivo do artigo foi atingido ao se avaliar a importância dos motivadores de LR, no ramo automobilístico e na Região Sul Fluminense, após a implementação da PNRS. Percebe-se um maior destaque ao programa legal ao comparar esta pesquisa com a realizada Bitencourt & Hernández |18 10.20401/rasi.8.3.623 Avaliado pelo sistema Double Blind Review Editor: Marcelo G. Amaral RASI, Volta Redonda/RJ, v. 8, n. 1, pp. 07-22, Jan./Abr. 2022 http://www.rasi.vr.uff.br por Hernández (2010). Este fato mostra que a questão da sustentabilidade ganha maior relevância, ao destacar como diferentes ramos da economia regional prioriza programas de LR e como estas fazem a gestão de seus resíduos sólidos. Neste cenário, práticas de LR mudaram depois da implementação da PNRS, uma vez que antes de 2010, faltava uma política que regulamentava e padronizava as tratativas no fim de vida útil dos produtos. Kumar e Putnam (2008), destacaram algumas importantes atividades na implementação da LR na indústria automobilística tais como, reciclagem, reuso, revenda, custos para desenvolver novas tecnologias, destino adequado aos resíduos e envolvimento dos diversos stakeholders. Desta maneira, estas atividades se correspondem em grande parte com as encontradas no artigo que foram reuso, revenda e reciclagem, custos por atividades de LR, assim como para desenvolver novas tecnologias, destino adequado aos resíduos e cumprimento da legislação. De fato, os dois estudos se diferenciam nas atividades de envolvimento de diversos stakeholders, não encontrada no artigo e cumprimento da legislação, não encontrada no estudo de Kumar e Putnam (2008). Como limitação, o estudo analisou uma região específica, a região Sul Fluminense e somente um ramo da economia, o automobilístico. A análise foi realizada com base nos questionários preenchidos por pessoas de determinadas empresas, o que confere subjetividade ao estudo, tendo o AHP função essencial em relação a este fator. Como recomendação de pesquisa futura, a análise pode ser feita em outras regiões e ramos, ou até mesmo na mesma região e ramos, mas em outras empresas para até mesmo ser feito um comparativo entre os estudos. Referências Bibliográficas Abbas, H., & Farooquie, J. A. (2020). Reverse Logistics Practices in Indian Pharmaceutical Supply Chains: A Study of Manufacturers. International Journal of Logistics Systems and Management, 35(1), 72–89. Agrawal, S., & Singh, R. K. (2019). Analyzing Disposition Decisions for Sustainable Reverse Logistics: Triple Bottom Line Approach. Resources, Conservation and Recycling, 150(March), 104448. Batista, L., Gong, Y., Pereira, S., Jia, F., & Bittar, A. (2018). Circular Supply Chains in Emerging Economies – A Comparative Study of Packaging Recovery Ecosystems in China and Brazil. International Journal of Production Research, 7543. Barbosa, B. B. C. C., Pereira, M. M. O., Antunes, L. G. R., Antonialli, L. M., & Ferreira, M. C. (2019). A Inovação Ambiental em Laticínios Segundo seus Gestores: Um Estudo na Região do Alto Paranaíba-MG. Revista de Administração, Sociedade e Inovação, 5(3), 38–54. Bhandari, D., Singh, R. K., & Garg, S. K. (2019). Prioritisation and Evaluation of Barriers Intensity for Implementation of Cleaner Technologies: Framework for Sustainable Production. Resources, Conservation and Recycling, 146(February), 156–167. Beiler, B. C., Ignácio, P. S. A., Pacagnella Júnior, A. C., Anholon, R., & Rampasso, I. S. (2020). Reverse Logistics System Analysis of a Brazilian beverage Company: An Exploratory Study. Journal of Cleaner Production, 274, 122624. Bitencourt & Hernández |19 10.20401/rasi.8.3.623 Avaliado pelo sistema Double Blind Review Editor: Marcelo G. Amaral RASI, Volta Redonda/RJ, v. 8, n. 1, pp. 07-22, Jan./Abr. 2022 http://www.rasi.vr.uff.br Brito, M. P., & Dekker, R. (2004). A Framework for Reverse Logistics. ERIM Report Series Research in Management, 3–27. Cetrulo, T. B., Marques, R. C., Cetrulo, N. M., Pinto, F. S., Moreira, R. M., Mendizábal-Cortés, A. D., & Malheiros, T. F. (2018). Effectiveness of Solid Waste Policies in Developing Countries: A Case Study in Brazil. Journal of Cleaner Production, 205, 179–187. CNI. (2019). Economia Circular: Caminho Estratégico para a Indústria Brasileira. 68. Cosenza, J. P., Andrade, E. M. D., & Assunção, G. M. De. (2020). A Circular Economy as an Alternative for Brazil’s Sustainable Growth: Analysis of the National Solid Waste Policy. Revista de Gestao Ambiental e Sustentabilidade, 9(1), 1–30. Couto, M. C. L., & Lange, L. C. (2017). Análise dos Sistemas de Logística Reversa no Brasil. Engenharia Sanitaria e Ambiental, 22(5), 889–898. Creswell, J. W. (2007). Projeto de Pesquisa: Métodos Qualitativo, Quantitativo e Misto. 2ª edição,Porto Alegre: Artmed. Demajorovic, J., Huertas, M. K. Z., Boueres, J. A., Silva, A. G., & Sotano, A. S. (2012). Logística Reversa: Como as Empresas Comunicam o Descarte de Baterias e Celulares? Revista de Administração de Empresas, 52(2), 165–178. Elkington, J. (1998). Accounting for the Triple Bottom Line. Measuring Business Excellence, 2(3), 18–22. Eltayeb, T. K., Zailani, S., & Filho, W. L. (2010). Green Business Among Certified Companies in Malaysia Towards Environmental Sustainability: Benchmarking on the Drivers, Initiatives and Outcomes. International Journal of Environmental Technology and Management, 12(1), 95–125. Escrig-Olmedo, E., Fernández-Izquierdo, M. A., Ferrero-Ferrero, I., Rivera-Lirio, J. M., & Muñoz-Torres, M. J. (2019). Rating the Raters: Evaluating how ESG Rating Agencies Integrate Sustainability Principles. Sustainability (Switzerland), 11(3). Farooque, M., Zhang, A., Thürer, M., Qu, T., & Huisingh, D. (2019). Circular Supply Chain Management: A Definition and Structured Literature Review. Journal of Cleaner Production, 228(April), 882–900. Ferreira, N. de M., Santos, L. C., & Silva, M. L. P. da. (2021). Stakeholders e o Gerenciamento de Coprodutos: Logística Reversa e Simbiose Industrial na Redução de Custos. Brazilian Journal of Development, 7(3), 33182–33201. Geissdoerfer, M., Savaget, P., Bocken, N. M. P., & Hultink, E. J. (2017). The Circular Economy – A New Sustainability Paradigm? Journal of Cleaner Production, 143, 757–768. Geisendorf, S., & Pietrulla, F. (2018). The Circular Economy and Circular Economic Concepts — A Literature Analysis and Redefinition. Thunderbird International Business Review, 60(5). Bitencourt & Hernández |20 10.20401/rasi.8.3.623 Avaliado pelo sistema Double Blind Review Editor: Marcelo G. Amaral RASI, Volta Redonda/RJ, v. 8, n. 1, pp. 07-22, Jan./Abr. 2022 http://www.rasi.vr.uff.br Geng, Y., Fu, J., Sarkis, J., & Xue, B. (2012). Towards a National Circular Economy Indicator System in China: An Evaluation and Critical Analysis. Journal of Cleaner Production, 23(1). Gerhardt, T. E., Silveira, D. T. (2009). Métodos de Pesquisa. Universidade Aberta do Brasil - UAB/UFRGS. Porto Alegre: Editora da UFRGS. Ghisellini, P., Cialani, C., & Ulgiati, S. (2016). A Review on Circular Economy: The Expected Transition to a Balanced Interplay of Environmental and Economic Systems. Journal of Cleaner Production, 114, 11–32. Gil, A. C. (2007). Como Elaborar Projetos de Pesquisa. 6ª edição, São Paulo: Atlas. Govindan, K., & Bouzon, M. (2018). From a Literature Review to a Multi-Perspective Framework for Reverse Logistics Barriers and Drivers. Journal of Cleaner Production, 187, 318–337. GRS-UFPE. (2014). Análise das Diversas Tecnologias de Tratamento e Disposição Final de Resíduos Sólidos Urbanos no Brasil, Europa, Estados Unidos e Japão. 188. http://www.bndes.gov.br/SiteBNDES/export/sites/default/bndes_pt/Galerias/Arquivos/produtos/download/aep_fep/chamada_publica_residuos_solidos_Relat_Final.pdf Hernández, C. T. (2010). Modelo de Gerenciamento da Logística Reversa Integrado às Questões Estratégicas das Organizações. Hernández, C. T., Marins, F. A. S., & Castro, R. C. (2012). Modelo de Gerenciamento da Logística Reversa. Gestao e Produçãoo, 19(3), 445–456. Hernández, C. T., Marins, F. A. S., Rocha, P. M., & Durán, J. A. R. (2010). Using AHP and ANP to Evaluate the Relation Between Reverse Logistics and Corporate Performance in Brazilian Industry. Brazilian Journal of Operations & Production Management Volume, 7(2), 47–62. Jonathas, R., & Paiva, A. D. (2020). O Cluster Automotivo Sul Fluminense. 10–23. Kumar, S., & Putnam, V. (2008). Cradle to Cradle: Reverse Logistics Strategies and Ppportunities Across Three Industry Sectors. International Journal of Production Economics, 115(2), 305–315. Lang, D. J., Wiek, A., Bergmann, M., Stauffacher, M., Martens, P., Moll, P., Swilling, M., & Thomas, C. J. (2012). Transdisciplinary Research in Sustainability Science: Practice, Principles, and Cchallenges. Sustainability Science, 7(SUPPL. 1), 25–43. Leite, P. R. (2012). Direcionadores Estratégicos Em Programas de Logística Reversa no Brasil. Revista Alcance, 19(2), 182–201. https://doi.org/10.14210/alcance.v19n2.p182-201. Leite, P. R., De Brito, E. P. Z., Macau, F., & Povoa, Â. (2006). O Papel dos Ganhos Econômicos e de Imagem Corporativa na Estruturação dos Canais Reversos. Revista Eletrônica de Gestão Organizacional, 5–21. Lima, N. N. H. (2016). Logística Reversa no Setor Automobilístico Brasileiro: Uma Bitencourt & Hernández |21 10.20401/rasi.8.3.623 Avaliado pelo sistema Double Blind Review Editor: Marcelo G. Amaral RASI, Volta Redonda/RJ, v. 8, n. 1, pp. 07-22, Jan./Abr. 2022 http://www.rasi.vr.uff.br Aplicação para o Estado de São Paulo. Luna, R. A., & Viana, F. L. E. (2019). O Papel da Política Nacional dos Resíduos Sólidos na Logística Reversa em Empresas Farmacêuticas. Revista de Gestão Social e Ambiental, 13(1), 40–56. Machado, H. H. D. S., Sgorlon, J. G., Altoé, S. P. S., Meneguetti, K. S., Oliveira, J. C. D., Martins, C. H., & Tavares, C. R. G. (2012). A Gestão dos Resíduos Sólidos Industriais Aplicadas a Países Desenvolvidos e em Desenvolvimento. Congreso Latino-Americano de Ecologia Urbana. Mangla, S., Madaan, J., & Chan, F. T. S. (2012). Analysis of Performance Focused Variables for Multi-Objective Flexible Decision Modeling Approach of Product Recovery Systems. Global Journal of Flexible Systems Management, 13(2), 77–86. Mota, M. C. S. A. R. C., & Paula, E. V. (2020). Identifying Barriers and Proposing a Roadmap to Develop Industrial Symbiosis Projects. Revista Administração da UFSM, 11(2), 471–488. Ness, D. (2008). Sustainable Urban Infrastructure in China: Towards a Factor 10 Improvement in Resource Productivity through Integrated Infrastructure Systems. 15, 288–301. Oliveira, C. T., Luna, M. M. M., & Campos, L. M. S. (2019). Understanding the Brazilian Expanded Polystyrene Supply Chain and its Reverse Logistics Towards Circular Economy. Journal of Cleaner Production, 235, 562–573. Oliveira, F. R., França, S. L. B., & Rangel, L. A. D. (2018). Challenges and Opportunities in a Circular Economy for a Local Productive Arrangement of Furniture in Brazil. Resources, Conservation and Recycling, 135(January), 202–209. Pinheiro, A. B., Chagas, A. C. C., Santos, S. M. S., Silva, S. M., & Lopes, F. M. D. A. (2020). Obstáculos para a Adoção da Simbiose Industrial no Distrito Industrial de Maracanaú: Insights para o Desenvolvimento Sustentável. Revista de Casos e Consultoria, 1–24. Saaty, T. L., & Sodenkamp, M. (2010). The Analytic Hierarchy and Analytic Network Measurement Processes: The Measurement of Intangibles. 1(1), 91–166. https://doi.org/10.1007/978-3-540-92828-7_4 Su, B., Heshmati, A., Geng, Y., & Yu, X. (2013). A Review of the Circular Economy in China: Moving from Rhetoric to Implementation. Journal of Cleaner Production, 42, 215–227. Trevisan, M., Nascimento, L. F., Madruga, L. R. da R. G., Neutzling, D. M., Figueiró, P. S., & Bossle, M. B. (2016). Ecologia Industrial, Simbiose Industrial e Ecoparque Industrial: Conhecer para Aplicar. Sistemas & Gestão, 11(2), 204–215. Vlachos, I. P. (2016). Reverse Logistics Capabilities and Firm Performance: The Mediating Role of Business Strategy. International Journal of Logistics Research and Applications, 19(5), 424–442. Bitencourt & Hernández |22 10.20401/rasi.8.3.623 Avaliado pelo sistema Double Blind Review Editor: Marcelo G. Amaral RASI, Volta Redonda/RJ, v. 8, n. 1, pp. 07-22, Jan./Abr. 2022 http://www.rasi.vr.uff.br Vieira, B. O., Guarnieri, P., Silva, L. C., & Alfinito, S. (2020). Prioritizing Barriers to be Solved to the Implementation of Reverse Logistics of e-Waste in Brazil under a Multicriteria Decision aid Approach. Sustainability (Switzerland), 12(10). Wang, J., & Wang, Z. (2020). Strengths, Eeaknesses, Opportunities and Threats (Swot) Analysis of China’s Prevention and Control Strategy for the Covid-19 Epidemic. International Journal of Environmental Research and Public Health, 17(7). | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/336531043The Anticipation of Afterlife as Based on Current Physics of InformationArticle · October 2019CITATIONS7READS18,1362 authors, including:Some of the authors of this publication are also working on these related projects:Biliary excretion of drugs View projectElectronic music as a way to explore xenolinguistics View projectDirk K F MeijerUniversity of Groningen632 PUBLICATIONS 18,945 CITATIONS SEE PROFILEAll content following this page was uploaded by Dirk K F Meijer on 14 October 2019.The user has requested enhancement of the downloaded file.1 The Anticipation of Afterlife as Based on Current Physics of Information Dirk K. F. Meijer* * Em. Professor, University of Groningen, PhD, The Netherlands. Mail: meij6076@planet.nl Abstract Individual consciousness is considered as an expression of an underlying, non-local, quantum field, that exhibits holographic properties. This paper pays special attention to the interfacing of this field of universal consciousness and our personal brain in relation to potential afterlife. We postulate a toroidal event horizon workspace of the brain that allows symmetric 4-D to 3-D quantum information flux and holographic 2-D personal memory integration. The geometry of a 3-D brain being embedded in a 4-D realm may explain the phenomena of functional brain binding, qualia, intuition, serendipity, synchronicity, extra-sensory perception, and well-established Psi-phenomena. Brain function is conceptualized as guided by zero-point energy field (ZPE) derived pilot waves that support consciousness also in the absence of neuronal activity, such as in near death experiences (NDE). Its toroidal organization exhibits quaternionic dynamics and thereby allows an opening to 4-D geometry and thus to universal consciousness and ZPE. This personal holographic workspace, that is associated but not reducible to the brain, collects active information in a 2-D ”brain event horizon”, an internal and fully integral model of the self. At death of the material body, this personal knowledge mental domain uncouples from the body, yet it is conserved because quantum information can never be destroyed. In NDE this uncoupling is only temporal but reveals universal consciousness in a fully transparent manner, since in this condition non-neuronal information processing is maintained by fractal semi-harmonic frequencies from ZPE that reflect an entangled personal register of each conscious being. The proposed concept, therefore contradicts the tentative, promissory materialist solution to the mind-body problem. Instead, it substantiates the notion that brain can act as a kind of ‘receiver’ by filtering of (sub)conscious states, through holographic resonance of universal consciousness with specific coherent oscillation domains in the body. Yet, it is recognized that our self-consciousness can also act as a damping filter for information from this universal knowledge field. The latter aspect of a ” dual filter theory” is apparently removed at states of modified brain function such as NDE, deep meditation and use of psycho-mimetic drugs, that expose us to an unknown cosmic perspective. The presence of a mental field-receptive resonant workspace, might be termed our “supervening double”, (or “soul”, not implying religious doctrine), and provides an interpretation framework for widely reported, but poorly understood transpersonal conscious states. It even implies that death should not be seen as a transition to another state of existence, but rather that all of us are already belong to such an eternal domain in our present life. Therefore, the present model implies the potential for the survival of individual life, qualifying us as designated survivors and eternal beings. 2 Key Words: Afterlife, Holo-fractal brain model, Universal consciousness, ZPE field, Field- receptive workspace of human brain, Scale invariant consciousness, Mental aspect of brain function, Brain attractors, Mind –body problem, Transpersonal experiences, Near-death experience, Out of body experience, Qualia of conscious perception, Dual filter theory Introduction Following the publication of the recent interview of the present author by Epoch Times, Israel, (Meijer, 2019), some readers asked for a more detailed outline on the item of immortal consciousness and potential afterlife. This paper addresses this important subject that, as always, raises quite some international interest and discussion, as treated in the following (see also Further Reading list at the end of this paper). Man is a symbol and metaphor creating being, who apprehends the world through the power of imagination. The desire to give an enabling place to an “inevitable“ individual death has led mankind to many powerful immortality and afterlife symbols. However, cell biology shows that humans are programmed to die and daily experience shows that we cannot escape fatal wear of our body. An incredible amount of thought has been written about afterlife and potential (im)mortality. There are many books and also academic dissertations about this subject with some really excellent! (van Bergen, 2002, Vidal, 2012, see also the “Further Reading List“, following the references). Today, molecular and genetic aspects of current longevity research shed a spectacular light on the first attempts to master immortality. In addition, we entered the era of digital immortality (Meijer, 2013). Current transhumanism advocates that human/machine hybrids (cyborgs) will be designed and the universe might be finally recreated by super-intelligent civilizations, providing a cyclic process of rebirth of our universe. (Bostrom, 2003; Lloyd, 2006; Tipler, 1994). Yet, is this all wishful thinking or do we have a real perspective for some sort of afterlife in a cosmic setting? In other words, do we have to wait for completion of such technological developments (up to millions of years) or can we rely on more immediate answers to the question of” afterlife”. In this paper we postulate that the ten thousand of faithfully recorded Near Death Experiences (NDE) and reincarnation cases clearly have explanatory power and reveal the true nature of our individual consciousness as connected to universal consciousness and potential eternal being. It is shown, on the basis of quantum physics, that dying may be only virtual and should be interpreted in the framework of conservation of information, (Meijer, 2013). While studying the nature of consciousness (Meijer and Geesink, 2016, 2017), we encountered reports about a non-material existence, a continuation of consciousness even while the brain was without cortical function and seemed to have stopped functioning, as discussed in many books on NDE, (for example by Pim van Lommel, 2010: ”Consciousness beyond Life”). We argued that if herein an alternate reality was being described, it would be theoretically possible that the NDE reports might contain explanations for 3 (self)-consciousness, or useful hints about the nature of a reality, ordinarily not perceived in personal experience or by objective science. At the instant of a severe trauma-inducing NDE’s, such as a heart attack or violent vehicle collision, the suffering individuals report being puzzled about what has happened, as they are suddenly viewing their environment from a different perspective. Typically, they are now located above their body, so that they no longer see through their eyes, but see the environment from a different location. Figure 1. An integrated scheme depicting the Construction of Reality, with its material (right part of the figure) and mental (left), aspects. This concept assumes a central quantum information field, that provides the very basis for creation of our universe and dynamically evolves further through cyclic feed-back processes from the present reality, in which natural (among others human) and artificial intelligence play crucial roles in observation and participation (Meijer, 2012). They also then start to notice that the fundamental nature of their sight has changed, as they seem to see through objects, can focus onto the vibrational atomic level of reality and obtain wide vistas across the whole cosmos, and even may experience panoramic views over 360 degrees. Individuals who have been born blind also report experiencing complex visual representations, colors and perceptions of distinct objects, that they never could have observed before, during their OBE (Ring, and Cooper, 1999). An excellent case has been made for the reality of discarnate consciousness by Greyson, 2019, and Pandarakalam, 2019. 4 Here, we hypothesize that conscious perception during normal function of body and brain occurs not solely in the brain itself, but is realized by, permanently, receiving conscious states through resonant interfacing with an individual part of a dynamic universal consciousness (Fig. 1) in which the whole boy is embedded. The differential qualities of perception thus are derived externally from this universal field of consciousness and not from intrinsic stored information in the brain. It is further proposed that the discrete EMF wave frequencies that we receive, inform our 3D reality through resonance with an underlying 4D universal field of consciousness (Fig. 1).We have earlier shown that information in the universe is fundamental and with energy and matter forms the building blocks for the architecture of reality (Meijer, 2012). One of the most interesting features of the OBE is the feeling that usual linear time no longer runs, and during the OBE it is reported that one can have simultaneously views of the past, current environment and, remarkably, also future scenes. Hiller, 2011, related this absence of running time to a sort of frozen time, realizing that formulation of Special Relativity predicts that photons do not experience time, and that NDE experience reveals a world of pure light. Figure 2: The tripartite concept of consciousness: our 3-D world (domain 1, green) is permanently imbedded in a universal consciousnees field revealed by NDE experiences (domain 2, yellow). All is part of the domain 3 supposed to be consisting of the grat mind of an external observer (modified from Hiller, 2011) 5 A conjecture that seems best suited for the existence of a material brain within a universal consciousness domain is that there initially existed a domain of pure consciousness in which nothing else was contained, no light, no space, not even time. Such an initial domain is generally discussed in science in relation to a primordial external observer. It also has been associated, in common with most religions, with the presence of God. Thus, assuming such an external observer or Great Mind, a tripartite domain theory can be conceived (Tiller, 2011), as depicted in Fig. 2. Such a concept is consistent with many NDE reports: the feeling of an omnipresent and powerful being, (see Domain 3}, in addition to a world of universal consciousness (the 2nd Domain) and the physical world we occupy and normally perceive (Domain 1). Of note this concept implies that domain 2 should be seen as primary and that our world is permanently imbedded in universal or cosmic consciousness. A triadic dimension distinction model for consciousness was also proposed by Neppe and Close, 2017. Reflection about the origin of the fabric of reality, such as the current Big Bang theories that postulate some kind of perturbation of quantum fluctuations, logically invite the hypothesis of an “initiator of events” or in other words an “external conscious observer”. However, current physics seems quite reluctant to allow the involvement of a “privileged observer”, or even, alternatively, the work of an alien intelligent species. Yet, such theories are certainly not less likely than the current mainstream multiverse speculations, that have an infinite number of solutions and an obvious lack of testability. Modern Physics and its Implications for Afterlife As we all realize, the individual life is finite, despite the current advanced measures to prolong life: death remains ultimately unavoidable. However, there are theorists, such as some quantum physicists (Amoroso,1999,2003; Goswami,1990; Berkovich-Ohana, 2014; Hameroff and Chopra, 2013; Wolf, 1985, 1996,1999, 2008; Kastrup, 2016; Bleicher, 2012b; King, 2003; Burke and Persinger, 2013; Lanza,2012; Carter, 2012; Schwarz et al.,, 2005, Schwarz and Begley, 2018; Schwartz, 2019; Pregnolato and Pereira, 2016 ), who at this point come to a very different conclusion: namely, that true permanent death is in fact, impossible! Our physical body, indeed, cannot escape its demise, but according to them, the information that determines our overall personality and reflects our life’s experience will be retained according to the law of conservation of information. Although, at first glance, this seems to be an unrealistic idea, it is perhaps somewhat more compelling when one realizes that our perception of the full range of reality is very limited and thus provides a very inadequate representation of reality! In the underlying knowledge field which is described by quantum physics, the preservation of information is postulated. Therefore, in the framework of the present essay, it certainly makes sense to discuss such alternative scientific perspectives (see also Amoroso, 1999, 2006; Forberg, 2006, 2007, Greenfield, 2003, Lanza, 2012, Kastrup, 2016, 2017, 2018, 2019; Pregnolato and Pereira, 2016). In the 20th century, the standard model of physics, about the structure of matter and its building blocks, the elementary particles, was developed. This model, at the micro level, includes quantum mechanical theory. However, quantum mechanics is not compatible with the description of the macro world of the Universe, as described in Einstein's cosmological Theory of Relativity. There are now desperate attempts 6 to develop a "theory of everything" that would include modeling of the cosmos and quantum micro-world in a single set of equations. One example of this is the, so called, “Super String theory” or, even better, its successor the M-theory that even assumes at least 10 dimensions. The basic idea of the String theory is, that the real building blocks of elementary particles, such as electrons. protons, photons and quarks, are much smaller elements that can best be described as a sort of tiny strings or loops that vibrate at different frequencies in a discrete manner. This theory implies that at the smallest physical level can be conceived of as an assembly of vibrating entities! In addition, quantum physics tells us that material particles at the same time can behave as oscillators (waves), by which they can produce a sort of poised state of waves/particles. Interestingly, particles that belong to each other in term of their properties (e.g. their polarization and rotation or spin) can be correlated with regard to each other over huge distances: if one changes the spin of one particle, the paired particle spin is altered too to maintain their system integrity. This phenomenon termed “quantum entanglement” was characterized by Einstein as “spooky action at a distance.” Thus, the structure of reality may be more flexible and intricately interconnected over vast distances than classically thought. Nature can therefore be seen as an enduring interaction of waves/ particles carrying their inherent physical status information. System status information exchanged is not only used for a deeper form of intra-system coordination, but probably also for the creation of interpretive meaning by conscious observers. A number of prominent scientists who laid the foundations for the hypothesis of the quantum brain: the Nobel Prize winners Wigner, 1983 (quantum physicist) and Eccles, 1992 (neurologist), the quantum physicists Bohm, 1980, Bohm and Hiley, 1993, Bohm and Peat , 2008; Stapp, 2017; Goswami, 1990 and Wolf, 1985, 1999, 2008 and the neurologist Pribram, 2014, as well as the psychiatrist Carl Jung and the mathematician Penrose, 2014, have theorized that our brain may functions as an interface between the individual and a “collective consciousness” that is stored in a supposed universal quantum field. Ervin Laszlo, 2006, stated that the universe, in this way, exhibits a kind of universal memory and that all experiences are stored in, what he calls, the Akashic record or field, a term that is also encountered in the Eastern religions. The basis for the existence of this field was derived by Laszlo from the, so called, quantum vacuum domain, also called the zero-point energy field (see later). Recent experiments showed teleportation of particles, based on sending complete information on a particular particle over a wide distance by which a solid particle is created at a distant site: this implies that matter may arise from information. The renowned scientist Anton Zeilinger, 2003 therefore concluded that information is more fundamental than matter and energy. It became also clear that the abovementioned universal information field can also be seen as the source and origin of reality, underlying the design of the Universe. Quantum entanglement implied to physicists, such as David Bohm, a reason to postulate a "quantum wholeness", an idea that expresses the concept that everything in the universe is connected ("entangled), see Meijer, 2012, Meijer and Korf, 2013). Yet, these aspects of reality are hidden from ordinary perception. Instead, our individual consciousness is so dominated by the normal but overwhelming sensory input that we are rarely aware of this kind of connectedness. According to Bohm, such quantum 7 phenomena are hidden due to fact that they are part of an underlying unseen quantum information field that is always present and manifested everywhere (and thus has non-local character), what we term the universal field of consciousness. Figure 3: Elements of quantum physics: uncertainty of position of particles, wave / particle duality as demonstrated in the double-slit experiment (upper part), as well as entanglement (non-locality) of particles at great distances, the phenomenon of coherence/decoherence and superposition of waves (lower part). Such a universal quantum field may do so through the exchange of information energy that is continually reallocated in a universal dynamic process. Building and decomposition of matter are, in fact, caused by absorption and release of (virtual) photons, and matter can thus, more poetically, be envisioned as "light captured by magneto- gravity." The virtual particles of the field can in this manner also provide an information exchange between the physical reality and the non-zero local field to which they belong by the phenomenon of resonance. Even the force of gravity, a yet poorly understood phenomenon, may be related to this field, because of its retarding effect on the movement of material: the, so-called, inertia (Setterfield, 2017). Penrose, 2014) suggests that quantum gravity at the Planck-scale level is instrumental in the capture of quantum information in our brain (Meijer and Korf, 2013). Infinite Consciousness: An Information Field? The universal quantum energy field can also be envisioned as a giant hologram ( for theory see, Susskind, 2015 and the Dutch scientist 't Hooft, 2012), that permeates everything in the universe. Numerous 8 authors support the idea that the universe has a holographic structure (Talbot, 2006; Anjamrooz et al, 2011; Aurich et al, 2008; Bjerve, 2016; Batiz and Milonvanovic,2017; Luminet, 2016; Haramein et al., 2016; Brown, 2019; Leffert, 2019). It is considered by the earlier mentioned physicist and systems theorist Laszlo as a universal information field (Laszlo, 2007, 2012). Consequently, each of us is supposed to be in contact with the field through wave interactions. This implies that we are in fact permanently connected to and embedded in a general energetic information field, which penetrates all animate and non-animate material (Meijer, 2012, Meijer et al., 2019). We will now explore the possibility that also personal information may be stored in this domain and that this can be instrumental in the realization of commonly unexpected modes of immortality. In other words: immortality and potential afterlife could be conceived as a modality of conservation of information. Zero-point Energy Field and Stochastic Dynamics One of the main challenges in consciousness research is widely known as the hard problem of consciousness. In order to tackle this problem, an approach from theoretical physics, called stochastic electrodynamics (SED), is utilized which goes one step beyond quantum theory and sheds new light on the reality behind matter (Keppler, 2012, 2016) Figure 4: Left: The universal, all pervading "zero energy field" pictured as a vacuum with fluctuations of quantum waves or particle/antiparticles in a fractal setting ( B and inset) that are supposed to display a vortex character (A). Through superposition and photon polarization information storage is possible. The zero-point quantum fluctuations 9 are reflected by creation and annihilation of particle/anti-particle pairs, in which the anti-particles travel in a reversed time mode. Charged particles moving in the ZPE- field obtain a toroidal form of energy flux that may form the basis for the creation of pilot waves that guide reality in our 3-D world (D). The collective wave patterns that obtain permanent feed- back from our reality, constitute the overall wave function of the Universe ( depicted in C). According to this approach, matter is a resonant oscillator that is orchestrated by an all-pervasive stochastic radiation field, called zero-point field (ZPF, see Fig. 4). In addition to these works there is also a very interesting SED-based approach to inertia and gravity, elaborated by Rueda and Haisch (Rueda and Haisch 1998; 2005). SED is based on the conception that the vacuum is filled with a real, all-pervasive stochastic radiation field, called zero-point field (ZPF), which can be pictured as an infinite sea of light and an ocean of pure energy. The properties of matter are not intrinsic but acquired by dynamic interaction with the ZPF, which in turn picks up information about the material system as soon as an ordered state, i.e., a stable attractor, is reached. These principles apply also to macroscopic biological systems. From this perspective, long-range correlations in the brain, such as neural gamma synchrony, can be interpreted in terms of order phenomena induced and stabilized by the ZPF, suggesting that every attractor in the brain goes along with an information state in the ZPF (Fig. 5). How can the, so called, hard problem in consciousness theories be approached? On closer inspection, the hard problem turns out to be tightly linked to the western way of thinking that adheres to the idea of a matter-dominated universe. This mindset proves extremely obstructive not only to the progress in consciousness research, but also to the deeper understanding of the physical world. In the following section the pioneering work of Keppler, 2012 and 2016 is summarized. How ZPF affects matter As explicated in the former section, SED regards matter as immersed in an all-pervasive stochastic background field with which it interacts permanently and unavoidably, thus acquiring a stochastic motion. This motion can be studied for various systems. As an example, we may take a closer look at the hydrogen atom, which is composed of a proton and an electron. In classical physics this system is unstable and collapses within a tiny fraction of a second due to the fact that the orbiting electron emits radiation and loses energy. However, within SED the situation changes significantly since the electron is no longer surrounded by a void. Rather, the electron is now able to perform a dynamic interaction with the background field, which results in an exchange of energy between the material system and the ZPF). Indeed, it can be shown analytically and numerically (Cole and Zou 2003; 2004a; 2004b; De la Peña and Cetto 2006; Cavalleri et al. 2010,) that there are certain dynamic situations in which the average power absorbed by the atomic electron compensates its average radiated power. These situations are characterized by quantization conditions and correspond exactly to the stationary states predicted by quantum theory, i.e., the stability of matter goes necessarily hand in hand with the quantum behavior of matter and both are a consequence of the interaction with the ZPF. A closer look behind the scenes of matter from an SED point of view reveals that not only the stability of matter but also its spatial structure and three-dimensional conformation are governed by the ZPF. Hence, SED is able to provide a clearer and more intuitive understanding of structure formation, in such a way that a quantum mechanical orbital, which reflects the probability density of finding an electron in a specific region around the nucleus, is associated with a stable attractor of the stochastic interaction process between the electron and the ZPF (Rodriguez, 2012). In other words, every stationary state of matter is characterized by an” individual dance pattern” that comes into being under direction of the ZPF. External stimuli, such as the presence of a magnetic field, can cause transitions 10 between different attractors, i.e., an external stimulus or a perturbation can prompt the system to follow a new dance pattern, ( Fig. 5). How matter affects ZPF So far, we have dealt with the impacts of the ZPF on matter. In the second step, we now have a look at the impacts of matter on the ZPF. This is very important because it must be considered that matter and ZPF exert a mutual influence, i.e., not only the ZPF affects the dynamics of matter, but the latter also affects the dynamics of the ZPF. Figure 5: Filtering of resonance frequencies from the ZPE field (modified from Keppler, 2015), by a phase-locking mechanism yields the qualia for our brain function. In reverse, the quantum brain leaves fingerprints in the ZPE field and the resulting dynamic and permanently updated ZPE constitutes the very substrate for consciousness and life processes (inset below right). From the study of simple nonlinear systems (De la Peña and Cetto 2001; 2006) one can learn that the ZPF is modified as soon as the system reaches a stable attractor, (Fig. 5). This aspect reminds us of modern versions of the guiding pilot-wave theory of David Bohm ,1980; Bohm and Peat, 2008), in which a number of authors propose a back reaction from our world to the supposed implicate order. Such a bidirectional flux of active information would imply that free-will based decisions of humans are transmitted to and integrated in the ZPE realm, meaning that the field is never static or deterministic but that intelligent species participate in the dynamic character of universal consciousness (Meijer and Geesink, 2017, Meijer, 2018). The free field with the initially random phase adapts itself to the new situation in such a way that the relevant frequency components involved in the maintenance of the equilibrium become highly correlated (De la Peña and Cetto, 2001). In other words, the 11 formation of a stable attractor results in a de-randomization of the local ZPF. This amounts to imprinting an information state on the ZPF. Different attractors are associated with different ZPF configurations and, hence, different information states. Consciousness in the Universe is Scale Invariant and Present in All There Is It was previously postulated by us that consciousness in the entire universe arises through, scale invariant, nested toroidal coupling of various energy fields, an aspect that may include quantum error correction. In the brain of the human species, this takes the form of the proposed holographic workspace, that collects active information in a mental workspace coined the ”brain- event horizon”, representing an internal and fully integral model of the self, (Meijer and Geesink, 2017). This brain-supervening workspace is equipped to convert integrated coherent wave energies into attractor type/standing waves that guide the related cortical template to a higher coordination of reflection and action as well as network synchronicity, as known to be required for conscious states. In relation to its scale-invariant global character, support was found for a universal information matrix, that was extensively described earlier by David Bohm, 1980; Bohm and Peat, 2008, as a supposed implicate order. Figure 6: Modeling of brain/mind relation in a 4+1-dimensional space-time framework (4+1 implies 4 spatial dimensions and one single dimension of time), on the basis of energy trajectories in a nested toroidal geometry. The 12 opposing forces of Dark energy (diverging force) and Gravity (converging force) as well as discrete wave frequencies of electromagnetic fields, are instrumental in the generation and compression of individual life information. The human brain may receive quantum wave information directly derived from the Planck space-time level (left above) through quantum gravity mediated wave reduction, as well as through resonance with the ZPE field (right above). Our brain can perceive only 3+1 dimensions with a one-directional arrow of time. The material brain and its 4+1-D supervening field-receptive mental workspace should be seen as an integral whole, until bodily death of the organism. The 4th spatial dimension allows individual self-consciousness since an extra degree of freedom is required for self-observation and reflection, while in the mental context the time dimension is symmetrical, allowing to integrate past and future- anticipating events. The 4th spatial dimensions is also assumed to accommodate the bidirectional flow of information between the domains of self-consciousness and universal consciousness. Bottom-up information flow from the Planck scale, combined with top-down information conjugation from the ZPE field, constitute the event horizon of the brain, also integrating gravitational and dark energy related force fields, and supervenes the physical brain. Event horizons of brain and whole body are depicted in red ellipse and circle respectively. It plays a role in a spectrum of space-time theories in current physics. The presence of a field-receptive resonant workspace, associated with, but not reducible to, our brain, may provide an interpretation framework for widely reported, but poorly understood transpersonal conscious states and also the supposd algorithmic origin of life. It also points out the deep connection of mankind with the cosmos and our major responsibility for the future of our planet. Consciousness can therefore be defined, as a state of a semi-stable system that has developed in a cooperative and cyclic operating mode so that it has become “causally self-observant”. Thereby, it can not only predict aspects of the local environment, but also can integrate memorized information and future-directed projections into a personal worldview that serves individual survival, development and social communication (Forshaw, 2016a,b). Yet, in the present paper a much wider context for consciousness is offered, in which our individual mind is seen as a part of a larger universal consciousness, being instrumental in the entire fabric of reality. This concept is based on our earlier consideration of an extended mind (Meijer, 2015). The Connecting Principle of Quantum Information in the Material Universe We usually talk about two seemingly separate worlds: that of material particles and that of a hidden wave world with its force fields, such as gravity and dark energy. The special feature of the work of Verlinde, 2011, 2016, is that the author brings the two aspects together in the form of quantum information as the most fundamental building block of the universe, following the concepts of John Wheeler, 1994 and more recently of Zeilinger, 2003 a,b as also pointed out earlier by Meijer, 2012, 2014. Matter and thus particles can therefore be seen as condensations of force fields that interact and both can be described with quantum information, that is actually a form of energy (see later). The special property of the torus operator is to bring the various types of field information together in integrating the various types of energy flow. Lefferts, 2019, in his recent impressive essay on Cosmometry, conceived the torus dynamics as reflecting the processes of enfolding and unfolding of information in a Bohmian context. 13 In quantum theory, energy is quantized and thus consists of discrete vibrational units (vibrating strings or loops). The space is also quantized according to the theory, thus divided into small space parts. This matrix of such space units is usually called space foam, bearing units that function as operators. Known examples of such elements are twistors (Penrose, 2014) related to nested torus geometry. Such units are supposed to operate on every fractal scale, from very small (Planck scale) to very large (black holes), and can be conceived as the collection points of the various force fields: gravity-, dark energy-, zero-point energy-, electromagnetic-, and Higgs fields etc. (see later also the wormhole matrix hypothesis of Haramein et al., 2018, Brown, 2019). In this manner, such operators integrate quantum information and store it on the edge of each fractal unit, that in the case of the black hole was called the "event horizon". The leading principle is that every object is fully described with information gathered on a virtual screen around the object (its event horizon). Therefore, quantum information, like energy, is never lost. Verlinde, 2011, among many physicists, also used the holographic principle, invented by the Nobel laureate 't Hooft (see for holography aspects see also Sieb, 2016, 2018; Batiz, 2107; Alfonso-Faus, 2011 Talbot, 2006; Anjamrooz et al, 2011; Aurich et al, 2008; Bjerve, 2016; Batiz and Milonvanovic,2017; Luminet, 2016; Harameein et al., 2016; Brown, 2019; Leffert, 2019). The entire universe and also galaxies, suns, planets and even living systems are to be regarded as toroidal organized information fields, each projecting digital information on their respective event horizons. In this respect, it has been experimentally demonstrated recently that: 1) information is in fact a form of energy: when information is removed from a quantum system, energy is released in the form of heat (entropy), (Bérut et al, 2012; Toyabe et 2010; Peterson, 2016). 2) this also applies to the quantum world. Binary units (bits, say a kind of yes/no questions) are then Qbits, but now information can mix (superpose) and can show entanglement with other states of quantum information (Lloyd, 2007, Nielsen and Huang, 2000). 3) the suggestion of Verlinde, 2011, 2016 and earlier Zeilinger, 2000, 2003, is that the information is intrinsic to matter (and even the source of it!). Thus, information always arises from interactions of wave/particles and according to classical information theory, information/entropy represents the potential to ask yes/no questions in such an event with regard to a particular system (see Lloyd, 2007, Meijer, 2013). According to these concepts, information is in fact the sum of expected information obtained from such yes/no questions. An example is DNA in our cells which in itself contains a lot of potential information (digitally expressed in Bits), yet is only clearly expressed in the cell in relation with RNA and proteins. The intrinsic (hidden), information of an object is therefore the result of the entanglement of the stored (individual) information from the various constituting particles, providing a sort of global information by converting all of this information into a coherent information matrix, that is dynamic in time (Keppler, 2013, 2016). Some link this matrix with the so called zero-point energy field, (Laszlo, 2007, 2012; Setterfield, 2002; Nation et al, 2012). 14 That we cannot directly perceive this information aspect, is traditionally ascribed to a hidden 4th spatial dimension (not the dimension time!), which cannot be observed in our 3D world, but can be mathematically derived. Such supposedly 4th dimension could also explain the creation of dark matter in our 3-D world through selective wave exclusion in the ZPE field, according to the so-called Casimir effect (Wongyun, 2013; Green and Levin, 2007). Recently it has also become clear that even space-time itself may be derived from the above-mentioned quantum fluctuation field and, in particular, through the entanglement of quantum information that is locked in. Instrumental in this respect are a sort of “short cuts” in space, that connect one part of the space with another, via a so-called wormhole structure (Fig.7). This concept is called EP = EPR conjecture. (Maldacena and Susskind, 2013; Susskind, 2016; van Raamsdonk, 2010). This wormhole concept (geometrically quite similar to the central channel of the torus) was already known from the physics of black holes, but now appears to be present at every fractal scale in the universe up to the Planck scale where it constitutes the aforementioned quantum foam (Haramein, 2016; Ford and Roman, 2000; Lloyd, 2007; Loll, 2011, Wikipedia/quantum foam). Figure 7: The flow of information in the universe from micro- to macro-levels (bottom to top) conceived as a nested toroidal operation that is fractal and scale-invariant and is initiated in a knowledge realm underlying the known wormhole matrix (quantum foam) at the Planck scale. Supposed quantized string activities produce elementary particles, atoms, molecules and life systems. The latter contain dedicated holographic memory spaces at the cellular and organ level. The human brain integrates internally and externally guided, conscious states. Further fractal and self-similar properties in a quantum fluid universe provide the architecture of cosmic macro-structures. from Meijer and Geesink, 2019). 15 Torus as a Geometric Model for Information Flow in the Universe A well- known example to illustrate the use of a metaphoric and archetypic figure in art and science is the Ouroboros: a serpent that seems to swallow its own tail, bringing together the extremities of its body in a circular mode, (Meijer, 2017, see Fig. 8, below). From ancient times it was seen as a symbol of eternal return and immortality. More recently, scientists such as Primack, 2006 projected on the snake body the relative size of the defined structures in the universe, going from the extremes of the smallest (so called, Planck) scale to the largest such as black holes and the entire universe. In this, humans seem to take an intermediate (almost anthropomorphic) position. (see Fig. 8, left). Yet, what do we actually really learn from this representation of cosmological scaling and the position that we, as human species, take in this respect? Evidently, it could be understood as a necessary complementary process, enabling people to internalize and integrate knowledge (a sort of enfolding) in a more holistic way, information that can be later unfolded spontaneously and then centered in the here and now. The process of "looking at" the cosmic Ouroboros, pictured below, can be seen as an explanatory trap in the case that it is only considered in a one-dimensional plane, even as depicted here as two- dimensional figure. However, no attention was drawn to any potentially circular dynamics around the serpent's body, at each point along its length. The particular explanation, clearly, avoids mention of the axis through the center of the circle along which the observer is located, as we know from the torus geometry treated above. Even more curiously the Ouroboros as pictured here, in a superficial way, takes the form of a zero, a container for nothing. The mystery of nothing, of course, is a current preoccupation of physics, whilst having long been a preoccupation in relation to spiritual insight. "Nothing" is also that with which many are now faced in contemplating their socio-economic future. So, what's missing? In the circular figure, below, we recognize a flat torus (a circling circle), with its known central channel-like axis or central hole. It is in this sense that the empty central portion of the Ouroboros constitutes a form of neglected invitation to an "inplanatory process" (see later), effectively performed going "through" the plane of the Ouroboros as presented (Meijer, 2017). How can we conceive the Ouroboros differently: not as an intoxicating bite or a suffocating swallowing of itself, showing a circular process of self -death, but rather from a completely different perspective. This is realized by asking the question: what can the tail offer to the mouth so that the figure turns into a symbol of resurrection and ultimate physical rebirth as alchemists and philosophers like Jung told us long ago? Here we should go into deeper layers of the toroidal inner structure (in-planation) that provides a superposition of wave information, recurrent coupling, and thereby information mirroring in a process of “physical self-reflection”. In the framework of the cosmic Ouroboros in Fig. 8, we now ask which element in nature can connect the various material shapes of the universe (often with striking spiral features) at the various scales? (see Fig. 11 in more detail) 16 Figure 8: The Cosmic Ouroboros A: The hollow ring structure can be interpreted as empty, but alternatively can be seen as a flat torus. B: The alternative toroidal representation of the Ouroboros. The central hole is depicted here as a black hole spherical structure, on its surface bearing the basic units of information (in 0 and 1) that are projected on the black hole event horizon. Each Q-bit is composed of four Planck surface units. A possible answer to this puzzle thus lies in the recognition of the central role of physical information, through which any of these structures can be described. The tail portion entails the smallest quantized (Planck) scale, supposedly composed of the smallest units of information: the Q-bit. The mouth portion of the serpent depicts the largest structure of black holes at centers of all galaxies, that not only gravitationally swallow information but, according to recent cosmological research, also radiates this information at their surface, called the event horizon. Here the smallest scale of the Universe meets the largest one known! In this model, the universe is conceived as a “living” structure that provides event horizon-mediated back-radiation of conserved information (in Bits or Q-bits) from huge numbers of black holes that are distributed throughout the entire universe. The radiated, entangled, information is supposed to provide a holographic projection of all there is, including our own planet and its life forms. In the foreseen, ultimate, fate of the universe, a final single black hole, that contains the total of all generated information, will pass the accumulated information to a, so called, connected white hole (Fig. 10). The supposed final black hole can in this manner function as a dedicated instrument for rebirth of our universe, in a cyclic operating mode (see Meijer, 2015) 17 Now, realizing this, we may no longer see the cosmic Ouroborus as a “suicidal machine”, but rather as an archetype that indeed symbols resurrection and eternity. Various dynamics of in-planation are fruitfully suggested by the dynamics of a torus as indicated above: its paradoxical nature as a mirror calls for self-reflexive dynamics. It can even be argued that, in addition, it exhibits contraction/relaxation loops, discussed earlier, in which the torus turns inside out in a vibrating mode, inviting strange loop trajectories. The "strangeness" of such a loop, see Hofstadter, 2007, further suggests that the toroidal Ouroboros is better understood as embodying a cognitive twist, relating "inside" to "outside" as with a Möbius strip. A potentially interesting representation of the Ouroboros is a 7-coloured hypersphere which can be rotated through its toroidal center, offering a comprehensible illustration of such 4-D twistedness in 3 dimensions. Equally charming is the sense in which the thereby produced extra 6 dimensions of spacetime are sometimes conjectured to take the form of a 6-dimensional Calabi-Yau manifold that is build up from basic toroidal structures (see Fig. 10). This muti-dimensional space representation in physics, led to the idea of mirror symmetry, notably associated with so-called string or membrane-theory (M-theory). There is also the suggestive indication that the abovementioned 6 dimensions might be significantly associated with the 6-line coding of the I-Ching hexagrams. Electromagnetic Aspects of Dynamic Models of Consciousness Many scientists have earlier suggested that basic information reaches our brain from outside (Persinger et al.,2008, 2015; Grof, 1987; Jahn and Dunne, 2004; Tonneau, 2004; Wolf, 1985, 1999, 2008, Taneichi, 2015), since the nervous system may also function as a receiver of subliminal signals. One could regard this as a physically defined “extrasensory perception”. In other words, we have to take into account a “sixth” sense in the form of a vibrational, resonance sensitive macromolecular apparatus in each of our cells (Hameroff and Tuzcinsky, 2015). These receivers act as vibrational, resonance sensitive elements in cells and act as receptors and as emitters of quantum information, which functions as resonant oscillators with specific resonance frequencies, which are coupled with a natural quantum field (Rouleau, 2014; Cifra et al., 2010; Meijer and Geesink, 2019; Bokkon e al., 2009, 2013; Dotta, 2013; Pereira and Furlan,2007, Pereira, 2015; Persinger, 2008, Persinger and Lavallee, 2010). The particular cellular sensors are composed of flexible three- dimensional structures of proteins, oligo-nucleotides and elements of the cell skeleton, that mutually communicate through discrete wave resonances and are sensitive to fluxes of photons, phonons, excitons and related quasi particles such as polarons (solitons) and polaritons. This bio-sensing apparatus of the cell, situated in an apparently electromagnetic cell, was tentatively called ” electrome” (de Loof, 2016), and is under the continuous influence of natural occurring internal as well as external electromagnetic fields (Meijer and Geesink, 2017, 2018, 2019). In these recent studies we reported on a fractal series of discrete EMF frequencies that influence a wide range of animate and animate systems, are identical to frequencies that have been reported in cosmological studies and therefore likely represent a spectrum of EM frequencies originating in zero-point energy field. 18 We hold that this overall harmonic matrix of EM resonance produces quantum coherent vibration domains in aqueous brain compartments and intrinsic ions (Geesink and Meijer, 2019), proteins, DNA, in addition to membrane associated structures such as ion-channel and microtubular proteins. The particular macro-coherent wave connections are instrumental in creation of the sub-conscious brain function that maintains its information processing ability in the absence of neuronal transmission, for example in near death conditions. Yet this field-type of consciousness is absolutely required for bidirectional information transfer with the associated memory workspace and thus with the 3D to 4D interfacing with the universal consciousness realm (see Fig. 12), as it is reported in such ND experiences. The latter all-pervading zero-point energy field potentially may resonant interaction with either life systems (Setterfield, 2002; Laszlo, 2007; Keppler, 2012; Caligiuri, 2015), or with physically defined mental dimensions (Grof, 1987; Jahn and Dunne, 2004; Beichler, 2012b). Also, bio-photonic type of communication (Bokkon and DÁngiulli, 2009; Dotta, 2013), gravitationally sensing of information present at the Planck scale (Penrose, 2014) and even information projected from event horizons of black holes have been implied in such holo-fractal phenomena (Maldacena and Susskind, 2013, Pourhassan et al., 2013). Figure 9. Current models of consciousness on the basis of long-range electromagnetic fields that may explain the simultaneous binding of distant brain nuclei involved in integral perception processes 19 In this respect it is worthwhile to mention that, based on quite solid evidence, the brain has been described as an electromagnetic workspace (McFadden, 2007; Pocket, 2012; John, 2001, Fig. 9). The universal force of electromagnetism controls all biological response” as Hawking, 2010 noted in “A Brief History of Time”. Indeed, living systems are under the continuous influence of electromagnetic fields and it is proposed in the present paper that the native, non-trivial, photon/electron vibrations exhibited by such scalar fields are shared with resonating proteins and nucleotides that control cell function throughout the hierarchy of living system. Of note, this inferred collective field concepts may constitute an interpretation framework for poorly understood phenomena such as mental states such as intuition, telepathy, far distance observation as well as near death experiences (see Radin, 1997) as well as near death experiences (Beichler, 2012c; Bókkon et al, 2013) and Psi phenomena (Radin, 1997; Beichler, 2012b; Rousseau, 2011), to mention only some of the many studies available on this topic (see later). Evidence for a Supervening Resonant Mental Workspace in our Brain Function Subjective conscious experience exhibits a unitary and integrated nature that seems fundamentally at odds with the fragmented functional architecture of the brain that have been identified in neurophysiological studies, an issue which has come to be known as the binding problem. It is generally agreed that the supposed boundaries between these classical ” brain compartments” are arbitrary and that intentional and especially emotional factors influence the relative involvement of deeper layers of mind (also see for this aspect Schwartz et al, 2005, Tammietto and de Gelder, 2010; Jahn and Dunne 2004, Rousseau ,2011, Meijer and Geesink, 2017; Beichler, 2012a). This has become evident in a large variety of conscious states as influenced, for example: by emotional feelings, meditation, hypnosis, vivid dream states, rhythmic sound exposure, use of psycho-active agents and also in life- threatening events that induce near death experiences. Furthermore, we postulate earlier that a dedicated part of the total brain activity is employed for the dynamic and ongoing construction of an integral personal universe/worldview (Meijer and Korf, 2014), in which consciousness represents a meta-phenomenon (Linton, 2015). Such an integrated representation of the outer world should not only include our individual ”autobiography” but also the intrinsic interactions with the external world, including the physical laws that determine it. It requires that we recognize our memories as true events as having relevance and significance for the present. The latter recognition aspect can only be realized if we see the present not only as a product of the past, but also as the anticipated (simulated) outcomes of multiple future projections. In other words there is no worldview of the present without probabilistic projection of our potential future and the latter implicitly signifies our long-term freedom of choice. Such an internal impression of the self, however, should be permanently validated via an integral and versatile external “state of art” of our Self that also should include potential subliminal and unconscious interactions, including time-retrograde (backward) projections of future events. The latter could be called ”remembering of the future”, a process that have been physically 20 defined and experimentally demonstrated earlier by Aharonov, 2010, in so called soft-stimulation quantum experimentation. We submitted therefore, (Meijer and Geesink, 2017), that an external memory workspace is operating in the human brain in a 4-D setting (Fig.6). The latter can take into account the hidden interaction with all, natural, forces/fields and also can integrate symmetric time and thus an aspect of backward causation. The proposed field-sensitive information workspace could function as a non-material and wave field-like simulation domain for a spectrum of mental representations. These may undergo a superposition with the internal worldview, in order to monitor the quality of our individual being (Fig. 6). This “software-like” mental program should exhibit an extremely fast response time, make immediate selection within a spectrum of multiple simulations possible. It also should offer the integral organism an optimal qualitative and quantitative impression of the current state of the whole body as embedded in its environment and its development. We envision such a monitoring system as supervening the basic neuronal communication networks, in order to generate a global type of a mental field (Meijer and Geesink, 2017, Meijer et al., 2019). By this guiding space-time domain, synchronic oscillations in the cortico-thalamic region of the brain would be induced through resonant coherent and condensed electromagnetic vibrations (standing waves) and/or attractor type of quantum information. The latter could, for example, be derived from non-linear and bidirectional interaction with the earlier mentioned stochastic zero-point energy field (Keppler, 2012; Caligiuri, 2015). 21 Figure 10: Toroidal wormhole structure of the universe (A) with fractal features (C) that encompasses the aspects of Dark Energy (expanding force) and Gravity (compressing force) as well as Information-energy unfolding and enfolding (B). The Black Hole /White Hole transition may imply twisted ( strange loop) features as also present in multi-dimensional space models (D) Cosmological Aspects of the Photonic Background Field The findings on semi-harmonic EMF frequencies in life systems, may also be related to our previous suggestion (Meijer and Geesink, 2016, 2018) that the apparent musical harmony- like aspect of electromagnetic background field may reveal hidden variables of the so called implicate order as supposed to underlie our daily reality by David Bohm, 1980, 1987. The latter was later ascribed to zero-point energy field by Lazslo, 2007, 2012). If so, this implies that bio-molecular brain processes are coupled to a fine-scale structure of the universe, as suggested earlier by Penrose, 2014, and further defined in more detail by Haramein et al., 2016, as a unified, micro-wormhole entangled matrix on the Planck scale. Energy flow in the Universe may exhibit a toroidal patten throughout the various scales of its fractal structure (Fig. 11). It is of great interest also that recent studies indicate that some black holes may predate the Big Bang on the basis of Bohmian quantum trajectories, do not exhibit a singularity and can function as a portal to another universe, by which implicitly the so called loss of information paradox is overcome (Gambini and Pullin, 2013). One of the models that was constructed presents the three-dimensional universe floating as a membrane (or brane) in a “bulk universe” that has four dimensions. The 4-D black hole would have an “event horizon” just like the known 3-D ones. The event horizon should be defined as the boundary between the inside and the outside of a black hole. In a 3-D universe, the event horizon appears as a two-dimensional surface. So, in a 4- D universe, the event horizon would be a 3-D object called a hypersphere (Pourhasan et al., 2013). The 4-D character of reality has been proposed earlier in many physics studies (Sieb, 2016, 2018; Sirag, 1993, Smythies, 2003; Carter, 2014; Hardy, 2016;Tozzi andPeters, 2016b). The present paper, may directly contribute to an answer on the famous question of Chalmers, 1995, 2019: how can something immaterial like subjective experience and self-consciousness arise from a material brain? Yet, in order to rephrase this question we may wonder that if consciousness is indeed the most fundamental aspect of reality (Zeilinger, 2003; Davies, 2007, Meijer, 2012), how does consciousness result in the manifestation of matter? The, so called, panpsychism (every material object contains specific information), at first sight, may be a logical solution to Chalmers question, as put forward in the information integration concept of Tononi et al, 2015 2016. Yet the latter may rather be seen as a reductionist approach in the line of current materialistic physics, since one fails to envision the issue from the point of view of consciousness as primary (Goswami, 1990; Kastrup, 2016, 2017, 2018; Keppler, 2012, 2013, 2016; Struppa et al, 2002; Pereira et al, 2015, Pregnolato and Pereira, 2016). 22 Figure 11. Toroidal geometry shows identified structures in the whole cosmos, from macro-(left above) to micro (right below) scales of the fabric of reality. The inset, left below, depicts the supposed nested toroidal geometry of the human body, heart and brain. Our brain model may also provide the potential for extra-sensory inter-individual communication and correlated brain signals (Hasson, 2012, Wackerman et al, 2003, Radin, 2004, Richards et al, 2005, Standish et al, 2004, Pizzi et al, 2004), since the postulated holographic workspace, projects its active information from a dedicated virtual screen in our brain, as described for event horizons in current cosmology. These projected multi-level fractal wave structures thereby contribute to the integral wave function of our universe (Fig. 4). The latter, in the form of the ultimate compressed information, in turn provides the necessary recipe for further evolution and a future rebirth of our universe, also operating in a toroidal rebound context, as treated earlier (see Fig 10, Meijer, 2012, 2015). Transcendental Experiences and the Resonant Mental Workspace A closely related aspect of a event horizon memory workspace is labeled with the term psi phenomena such as clairvoyance, distant viewing, telepathy, psychokinesis and near death experiences (Tiller, 1990, Grof, 1987; Jahn and Dunne, 2004; Hameroff, 2012, Hameroff and Chopra, 2013; Rousseau, 2011, Lanza, 2012, Phipps, 2012, Carter, 2012, Venselaar, 2011). The latter aspect is documented in thousands of international reports and nowadays open to scientific inquiry (Lake, 2015; Greyson, 2010, 2019; Schwartz et al, 2005; Bókkon et al, 2013, Hardy, 2016; Kastrup, 23 2016; Schwarz, 2019). The specific components of this experience, often with a long- lasting psychic impact on the recovered patient, include so called out of the body experiences, tunnel visions, and a remarkable clear and holistic state of awareness, verbally reported by the particular patients in retrospect. This is claimed to occur in the absence of the cortical activity that are normally shown in EEG’s during vivid dream states. The typical descriptions of a felt dissociation from the body, in addition to an experienced total life-panorama (Lake, 2015, Greyson, 2010, 2019; Schwartz et al, 2005; Pereira and Reddy, 2016), as reported by a part of the NDE cases, may point at the existence of the radiant resonance mind field (Meijer, 2013), as considered in the present study. Such a ”personal double” (Vitiello, 2001) that can largely influence cognitive brain function, is somehow persisting while other brain functions seem defective. This overall picture may indeed point at a supervening field character that is not directly dependent on normal brain activity and is open to the input of huge amounts of personal information apparently originating from an external data source. In our model (Meijer and Geesink, 2017,) such a potential universal knowledge field would involve a fourth spatial dimension (see Fig. 12). The particular holographic type of consciousness, situated in the particular event horizons as a sort of bordering memory domains, can effectively function as a nested information workspace, that in humans is instrumental in constructing a mental model of reality for internal use in each individual, thus functioning as a global reference system. It is instrumental in the overall monitoring of each individual (of any intelligent species that inhabits our universe). In addition, it should therefore be involved in the fine tuning with and updating of a supposed universal consciousness. The feature of quantum back-propagation can effectively implement a form of adaptive error correction (Goncalves, 2017), as has also been shown in the generation of cyclic/toroidal chaos by Hopfield networks (Akhmet and Fen, 2014). The present model thus postulates a mental workspace in close connection to our brain. This raises a number of central questions: 1) is there a physical basis for such a workspace or should it be seen as an intrinsic mind/matter dualistic approach 2) If such a disembodied aspect of consciousness exists, how does it communicate with the physical brain? 3) If our brain is supervened by an updated inner projection of the outside world, and thereby tends to reduce “surprise”, how does this affect free will (choice). With regard to the potential dualistic aspect, we stipulate that we consider our model as non-dualistic and, in this sense, agree with the Operational Architecture model of Fingelkurz and Fingelkurtz, 2010, 2014 on separate phenomenal and neuro-physical aspects of consciousness, stating that both aspects have an ontological relation but are not reducible to each other. We respect the related concept of Libet, 1994, 1996, of a mind field, especially in relation to the aspect of retro-causality in conscious experience, but tend to disagree on his suggestion that “if such a field is defined by physics, it should be entirely different from presently known physical mechanisms” (see discussion by Pockett, 2012, 2013). The implicit suggestion of a non-material and extra-corporal mental workspace, that supervenes our neural system and provides the dominant part of self-consciousness and may operate in addition to our daily experienced conscious state is supported by earlier and also more recent observations in fNMR studies that long term memory is not correlated with scaled sizes of the brain. 24 Savants with normal brain size can demonstrate a huge, disproportional, memory space (entire novels and even contents of telephone books are memorized in detail). Hydrocephalic patients that have only 5% of normal brain volume (micro-cephaly) can show quite normal intelligence and social behavior (Forsdyke, 2014). Other striking examples are patients with a largely destroyed forebrain that maintain a quite normal life (Sasal et al, 2016). Even the known split-brain patients that seem to develop two different types of consciousness in, the isolated, right and left halves of the brain, in fact show this aspect (McGillchrist, 2016). In addition, split-brain patients with disconnected hemispheres even perform better at some cognitive tests (see Sasai et al, 2016). In more general terms, the aspect of non-material mental aspects of consciousness has been strongly pursued from neurological (Nagel, 2012), bio-physical (Keppler, 2016), philosophical (Kastrup, 2016), quantum-physical (Henry, 2005) and evolutionary viewpoints (Grandpierre, 2014). A recent paper of Tononi (2016), reviewing the integration of information consciousness concept, mentioned an interesting view of Sullivan (1996), that another type of consciousness becomes manifest in meditative states that in fact can be considered as rather information content-less, and could reveal a normally hidden part of consciousness that is normally masked or filtered away by the busy default activity of our brain (see also Jahn and Dunne, 2004; Rousseau, 2011; Martin, 2013 ; Schwartz et al, 2005). Hardy, 2016, takes a space-time approach by positioning individual consciousness and the Self in a hyper-dimension in which death is the just the severing of the link between this domain and the brain/body, leading to an independent holographic semantic field on a personal basis. The latter resembles the proposal of Irwin, 2014, seeing consciousness as a quantized space-time language that can be described by quasicrystal mathematics of the E8 geometry. It was also suggested that that self-consciousness could continue outside the body, but remains at the level of Planck-scale geometry, as related to generation of biophotons in which visual imageries are coupled to long term visual memory an NDE seems driven by visual processes (Bókkon et al, 2013). It is of interest that Tozzi and Peters, 2016b, gave their work a much broader context by applying their hyperspace toroidal model and the Borsuk-Ulam theorem to a cosmological interpretation of evolution of our universe, in which the start of the universe is envisioned as a loss of dimensions from a multi-dimensional symmetric manifold, picturing a pre-big bang scenario. Merali, 2008, earlier commented on the doughnut-shaped universe concept (Aurich et al, 2008, see Fig. 11). The striking similarity with the 3-D brain as a “personal universe” (Meijer and Korf, 2013), supervened from a 4D- toroidal hyperspace domain, again points to a scale invariant symmetry breaking as a dominant feature of reality. In this respect it was put forward earlier that a hyperspherical universe, in which the present universe is nested in surrounding a toroidal hypersphere, can directly be derived from Einstein’s relativity theory on the basis of a re-interpretation of the Klein-Gordon equation, as performed by the famous Italian mathematician Fantappié (see Galloni, 2012, Chiatti, 2007). As mentioned above, we tentatively add to this configuration a toroidal fourth dimension of self-consciousness in continuous contact with an extended consciousness or awareness continuum that is 25 defined by us and many others as universal consciousness. The latter aspect rejects the usual framing of a mental workspace as a dualistic concept, since we envision our proposal of the extended brain as being derived from universal consciousness, as the very source of all that exists (see Goswami, 1990, Meijer and Raggett, 2014). A similar discussion was raised in relation to the very elegant consciousness model of Fingelkurz et al, 2010, 2014, called nested operational architectonics of the brain. In this model it was postulated that an electromagnetic brain field (see also Kida et al, 2016, for a fractal representation) connects a mind-subjective space-time to a distant physical space-time. Figure. 12: The mental workspace (bottom middle) as part of universal consciousness in a 4-dimensional domain. A temporal break connection is at stake in NDE experience and permanent disconnection occurs at bodily death This all-pervading system represents a non-local and retro-causal space memory, operating via recursive information feed-back/feed forward processing of polarizable electromagnetic quanta of the vacuum (Brown, 2018). We may speculate that this represents a pre- big bang information domain, in the framework of biological cosmogenesis. Such a pilot wave system may also have exerted an ordering influence during (pre)-biological evolution, at all scales of the universe. In our model, the fractal geometry of the torus takes a central position in brain physiology, as was also put forward by Tozzi et al, 2015, 2016, 2017; Knierim and Zhang,2012. The toroidal generation (scattering) of quasi wave/particles such as polaritons and polarons (solitons) is a crucial aspect of information integration of the various force fields that are involved. Solitons have been proposed widely 26 as axonal information carriers in the brain (Wikipedia), in which myelinated nerve fibres function as light guides for solitons and bio-photons in this manner bridging electromagnetic and spin aspect of neural information processing (Kumar, 2016). The Event Horizon Consciousness Model in Relation to Afterlife and Immortality With regard to the second point on the communication item, we propose a multi-factorial informational connection on the basis of quantum holographic principles (see Fig. 6 and 12). Yet, all of the various communication mechanisms mentioned here are clearly interrelated. They include long range and bidirectional correlations of the supervening mental workspace with the physical brain through phase-locked quantum resonance, that is instrumental in the holonomic sharing of quantum information through 4-D to 3-D projection and phase-conjugation (Mitchhell and Staretz, 2011). A similar multifactorial process has been proposed in relation to information conservation by event horizons of black holes (Van Raamsdonk, 2010; Pourhasan et al, 2013; Verlinde 2011, 2016). If consciousness and in particular self-consciousness is related to permanent contact with the 4-D hypersphere information workspace (see also Prakash et al, 2008, Sieb, 2016, 2018, Taneichi, 2015), by which in fact probability is converted to meaning, this can be envisioned as a “storyline”, implying that consciousness can observe itself from a geometric distance. This aspect is crucial in the understanding of the earlier suggestion made by Penrose, that consciousness contains an intrinsic non-computable component. Zizzi and Pregnolato, 2013, stated in this respect that to approach the non-computable aspect, a sort of mathematical meta-language will be required. In other words: if self-consciousness observes the brain memory and awareness states, by this very act it changes both of them. The only possibility to address this paradox seems the potential to link or integrate past and future events in the framework of quantum approaches. Relevant examples are the transactional interpretation of quantum physics of Cramer, 2004 and the, so called, soft quantum wave stimulation concept of Aharonov et al., 2010, 2013 in which future states interact with the past states to create the present, (treated in Meijer, 2012). In the bidirectional flow of energy according to re-interpretation of the Gordon-Klein mass/momentum/energy equation of Fantappié and the retro-causality considerations of Auletta and Ellis, (reviewed by Meijer, 2012) such a mechanism is plausible. It is of interest that recent calculations of Song, 2007, indicate that consciousness should, at least partly, reside outside the brain and that it is applied to the brain than rather than generated by it, as also discussed earlier by Tonneau, 2004; Pregnolata and Pereira; Pereira and Reddy, 2016;Beichler, 2012 c; Pereira, 2015; Lanza, 2012; Lazlo, 2007, 2012; Berkovitch- Ohana, 2014). This is so, since the conscious activity of an observer observing the change of an observable is, per definition mathematically not computable (Song, 2007). 27 Transcendental aspects of NDE: Preview of Afterlife? The phenomenon of near-death experience also raises deep philosophical questions. This experience shows that consciousness, under those circumstances, apparently can observe a physical world without involvement of the normal senses. There are even documented cases in which during the near-death experience, observations were made by the patient at great distances from the body, that later were verified as being absolutely correct. Figure 13: Transpersonal experiences: the researchers (bottom right), the techniques for them to arouse or experience (top right) and the postulate of the internal filters in our brains (top left). Personal information from a supposed quantum field, is experienced as clairvoyance or described as an inner voice (called “Daemon” in classical Greece philosophy). The question arises whether our normal sensory perceptions act as a filter for this kind of conscious extra-sensory perceptions. Transpersonal experiences has been studied by Greek philosophers referring to the “Daemon” as an inner advising voice, and more recently by a spectrum of renown scientists (Fig.13). From the testimony of many persons reporting NDE, it is inferred that their perception during near-death experience is faster, broader, more realistic and pure of nature than during an ordinary dream. Also during the use of certain drugs, meditation and rapid change of gravity (falling down and traffic accidents), "out of the body" and NDE-like experiences are generated, but they do not have the quality and character 28 content of the expanded NDE experiences as they were in many cases reported (see Fig.14). Recently an interesting study was presented (Venselaar, 2011), in which the NDE is described fully physically as a 5- stage process that includes a) separation of photonic consciousness from the body, b)a journey to a tunnel of light after strong contraction involving mini- black hole and wormhole modalities and observation of cosmic structures, c)being at the presence of light at a border, having a life review and experiencing specific light shifts, d) a subsequent return to the body and e) final unification with the physical body. “Life after life” is pictured as a being of light, due to the fact that damaged or dying atoms of our physical body release countless photons (light energy). Since our daily consciousness is connected to the brain, these exotic photons, are supposed to be transformed into another broader transcendent consciousness, in which such photons are indestructible. The particular photonic body is capable to 'carry' information about who we were and are. Figure 14: The general characteristics of a Near Death Experience The panorama of life: better than memory .... Impressive is that during a near death experience, the patient is confronted with a full and bright panorama of his/her entire life, in which he/she is not only relives the life from his/her own ego perspective, but also from the perspective of each person with whom one has been in contact during one’s lifetime. This experience from the perspective of the other, according to many of the NDE patients is complete, including the thoughts and feelings of every person who crossed one’s path in life. Of note, the majority of NDE witnesses, subsequent to their profound experience, live a more conscious life, being 29 more open to others and more interested in the big questions of life,( see Greyson, 2019; van Lommel; Carter,2012 Bokkon et al., 20132001, 2007; Pregnolato and Pereira, 2018, (Fig. 14). Is NDE only due to previously stored images in the brain? The extraordinary near-death experience of Vicky Noratuk showed that even blind people can see again during a near death experience! From her birth she had a shrunken eyeball and optic nerve and an undeveloped visual cortex. So even if there would be a yet unknown activity of the brains during the near-death experience, she did not possess a healthy sense organ that could transfer her visual impressions to her consciousness through the brain, nor a memory for detailed images. After she had her experience, she described this as "it felt like the place where all knowledge is". A faithfully executed study with 31 similar cases of blind vision (Ring and Cooper, 1997), revealed that blind persons, including those blind from birth, do report classic NDEs of the kind common to sighted persons. Further citing the article: “Thus, what we have here is an adumbration of a process that begins with Mind, fully independent of brain becoming self-referential, that is, becoming identified with consciousness itself, and then converting this noumenal consciousness into a dualistic modality that generates the familiar phenomenal world. What we have called transcendental awareness is at least the beginning of the reversal of that process by which, even though the traces of an everyday dualism remain, the individual is enabled, however temporarily, to experience the world from a perspective independent of brain functioning and the operation of the senses. Each of these theories formally entails such a state of awareness, and specifically in blind persons, during NDEs or OBEs”. The modern NDE research thus leads, just as investigation of reincarnation (see below), to the hypothesis that consciousness continues to exist in some form after death. Death, according to this hypothesis, is only an illusion. It is, in this consideration, only a transitional state to another form of conscious life, but without the experience of a physical body. According to some quantum physicists, such as Alan Wolf, (1996,1999, and 2008), dying is more precisely, not a transition to another phase, but rather every individual is at every moment of one’s life already represented in the non-material dimension of the aforementioned quantum information field (Fig. 12). In other words, conservation of individual information is a continuous process and is permanent. Many authors in consciousness studies conclude, after really exhaustive and scrupulous research, that in a manner not understood, some people (or detailed data on them) persist, at least temporarily, after their death. This, in some way, leads to the transfer of personal information to young children that become beware of this information and interpret it as their former life (Braude, 2003; Phipps, 2012; Carter, 2012, and Wikipedia on Reincarnation). It is clear that we arrive here at the borders of knowledge and that we experience something that is far beyond our present human understanding. Nevertheless, the phenomenon of children report is, in itself, established in thorough research, and despite the absence of a satisfactory explanation, is recognized as such in scientific circles, albeit without the notion that this represents a process of rebirth or reincarnation. An impressive report of a “skeptic” scientist is depicted in the link about reincarnation, that provides an interesting view on this possible modality of 30 immortality (Phipps, 2012). The notion that can be inferred from the sections above, namely that consciousness of all living entities belongs to a collective "web of information", should implicitly be supplemented with the reverse side of this concept: that individual consciousness , through this interfacing, is just as much an expression of the universal (nonlocal) consciousness. In this respect we are, according to this quantum concept, truly part of the unity of nature, as Spinoza already postulated. These phenomena could also be related an underlying non-local information field because they concern influences at a distance or messages from the future experienced by people. The central question is how for the necessary information for these phenomena can reach the brain and how these extra-sensory communication / perceptions occur from a mechanistic point of view. The latter is of great importance for future investigation, because it may provide answers to study the, well known, evading aspects at sequential PSI observations, as well how the signal strength of ESP experiences could be improved. Both aspects could be related to the inbuild filters in our brain that should protect us against to much disturbing information (see Jahn and Dunne, 2007, Radin, 1996, 2006, Kastrup, 2016, 2017 ,2018). The entity that we define as consciousness, the soul, or the self, (that which makes me who I am), does not stop existing just because someone has entered the period beyond death” (Parnia, 2013). Therefore, since science cannot explain the source of the most basic characteristic of “waking” consciousness, the sense of self, this opens the door for a new hypothesis. Does “waking consciousness” independently take shape as the result of brain-based abilities that arrive around ages two to three? Or does it result from adaptation of an already present “consciousness” to life on earth as cognitive brain faculties become active and our experiences expand? Of note, the idea that self-awareness may originate outside the brain is not new. The idea of the “self” was our most prized, yet unexplainable, possession reaches back to Greek philosophy. Parnia, 2011 seems to convey immortality on the new entity. As such, it reasonably might be expected to possess a unique identity of its own. Would this then continue to serve waking consciousness as our perceived self-identity? Remember that we normally are not aware of memories stored in our right cerebral hemisphere hence we forget the source of our self-identity. Personal information from a supposed quantum field, is experienced as clairvoyance or id also described as an inner voice (called “Daemon” in classical Greece philosophy. As early as 1994, the Nobel Laureate and Australian neurophysiologist Sir John Eccles, 1994 claimed, “The odds are 10 to the power 10.000 against the uniqueness of the individual self being derived from the genetic uniqueness that built the associated brain. The uniqueness of the individual self must therefore arise from some “external source”. But is this also true for the personal wave information, that is the non-material aspect of these atoms/elementary particles. In other words, is this information destroyed? Some quantum physics claim that information cannot be lost (the law of conservation of information) and that due to entanglement of the wave type of these building blocks, the personal structure and functional organization of individual life will survive bodily death (see the references of Wolf, 1989, 1999, 1990 Hameroff and Chopra, 2013). As the consequence of the persistent storage of this information in the universal quantum field, together with the total imprint of our personal experiences (our complete biography) in this domain, the personal 31 profile would be retained and would also be available for resurrection (Tipler, 1996, Hameroff and Chopra, 2013) and, in principle, may be maintained for a later reincarnation. Could consciousness exist outside the body after death? The latter authors believe it can: “According to the Orch- OR consciousness model, under normal conditions in an intact, healthy brain, consciousness occurs as frames or snapshots extending through multiple spatiotemporal levels from networks to neurons to microtubules to quantum forces (see also Meijer and Korf, 2013), down to and including Planck scale geometry. When the blood stops flowing and metabolic energy can no longer drive microtubule quantum coherence, quantum information relating to the subject’s conscious experience and memory isn’t necessarily lost or destroyed, but may dissipate to the universe at large, remaining entangled as a unified soul-like entity grounded in Planck scale geometry. If the body is resuscitated, the quantum information can return, and the subject may report an NDE or OOB experience. If the body is not resuscitated and the patient dies, the entangled quantum information constituting the subject’s consciousness and memory may persist in spacetime geometry, perhaps entering an embryo in the context of reincarnation”. Could the universe – empty spacetime geometry – conceivably host consciousness on the loose? There is ample energy in the form of zero- point fluctuations, so the question is whether information can be registered in the nothingness of spacetime, and transcend from Planck scale to biological scale. Many will see this daring hypothesis as an empty speculation or, at best, an understandable illusion, but for others it will provide wide perspectives for potential answers to an ancient quest of mankind, regarding individual survival and the reality of afterlife (Carter, 2012). Extra-sensory Perception: Information from Another Domain? Examples are, the fairly normal, (trans)-personal feelings of intuition, serendipity, synchronicity as well as feelings of "high", and dreams (Fig. 13 and 15). But, also, more "transcendent," or paranormal aspects such as channeling, out of body experiences and near-death experiences (NDEs) might be discussed in this framework (Radin,2006; Jahn and Dunne, 2006; Braude, 2003; Kastrup,2016; Carter, 2012). The general question here is whether the so called "normal" and "beyond normal" experiences are so sharply separated and whether they do not constitute a continuity of reality. The quantum consciousness models treated above may also be instrumental in the explanation of a number of so-called transcendental experiences in the category of PSI phenomena (see Fig. 15,) such as clairvoyance, telepathy, remote viewing and psychokinetic. These phenomena could also be related an underlying non-local information field because they concern influences at a distance or messages from the future, experienced by people. The central question is how for the necessary information for these phenomena can reach the brain and how these extra-sensory communication/ perceptions occur from a mechanistic point of view. It is important to note that much of PSI / ESP research in this field is carried out at a scrupulously accurate manner (Schwarz, 2019), to exclude any form of statistical bias and misinterpretation. Yet, this research 32 is often focused on the direct detection of, so called, paranormal phenomena and less on the underlying physical mechanisms of information transfer. The latter is of great importance for future investigation, because it may provide answers to study the well-known evading aspects at sequential PSI observations, as well how the signal strength of ESP experiences could be improved. Both aspects could be related to the inbuilt filters in our brain that should protect us against too much disturbing information (see Jahn and Dunne, 2007, Radin, 1996, 2006,). Figure 15: Various forms of extra- sensory perception and Psi phenomena How Information Is Received in Brain: a “Dual Filter” Hypothesis The modern scientific understanding of mental phenomena asserts that the brain is the sole causal agency of mind. This view is substantiated by the correlations ordinarily observed between subjective mental states and objective brain states. Yet, there is an increasing amount of evidence suggesting that, under certain extraordinary circumstances, the correlation between peak subjective experiences and brain-states breaks. This strongly indicates either that the brain is not the sole causal agency of mind, or (and more likely) that it is not a causal agency at all. An alternative hypothesis for the relationship between the mind and the brain was put forward by Kastrup, 2016, 2017, 2018, that is entirely consistent with current neuroscience data and increasingly supported by the latest scientific evidence. The latter author deserves credit for highlighting the crucial 33 aspect of idealism again in science philosophy (see for a splendid dialogue with some renown materialists Kastrup, 2019). The particular hypothesis explains not only why brain states are, ordinarily, tightly correlated to mind states, but also how, under extraordinary conditions, subjective experience can occur independently of the brain. The theory offers a rational, evidence-based, yet fundamentally different perspective on the nature of consciousness, life, identity, and death than that offered by materialism. The following section provides a compilation of Kastrup’s ideas. There is an undeniable correlation between brain states and subjective experience. Alterations of consciousness accompanying physical trauma to the brain, as well as the use of anesthetics and psychiatric drugs, are also examples of the tight link between mind and brain that many of us are personally familiar with. Laboratory studies have provided evidence that this correlation is even more specific than one could infer from direct experience: particular conscious experiences have been linked to specific neuronal activation patterns in the brain (Metzinger, 2000). Experiments with Transcranial Magnetic Stimulation (TMS) have also demonstrated that deactivation of specific brain regions correlates tightly to specific changes in subjective experience (Pascual-Leone et al., 2002).Therefore, any theoretical hypothesis purporting to explain the ontological status of mind must be able to explain why and how subjective experience seems, ordinarily, so tightly correlated to brain processes. Much of the neuronal processing in our heads, entailing the exact same kind of neurons that otherwise lead to awareness, is completely unconscious (Eagleman, 2011). Such seemingly insurmountable difficulty in logically deducing the qualities of experience from the properties of matter is called ‘the explanatory gap’ (Levine 1999), or ‘the hard problem of consciousness’ (Chalmers 2003, 2019). Now, if consciousness is primary and irreducible, then the brain cannot be the causal agency of mind; mind must exist a priori, ontologically preceding the brain. How can we then explain the empirical observation that, ordinarily, mind states correlate tightly to brain states . The hypothesis here is that the function of the brain is to localize consciousness, pinning it to the space-time locus of the physical body. In doing so, the brain modulates conscious perception in accordance with the position and perspective of the body in space-time. Indeed, there would be clear survival advantages for the brain to evolve to do just that: by localizing and modulating subjective experience according to the space-time locus of the body, the brain coaxes mind to identify itself with the body and, therefore, contribute actively to the body’s survival. According to the so called ‘filter’ hypothesis of mind-brain interaction, as earlier proposed by Jahn and Dunne, 2004, (see Fig. 16), no subjective experience is ever created by the brain, but merely selected by it according to the position and perspective of the body in space-time. This selection process is akin to a ‘filtering out’ of conscious experience: analogous to how an analog radio receiver selects, from among the variety of stations present concurrently in the broadcast signal, that which one wants to listen to; all other stations being filtered out and never reaching the awareness of the listener. As such, all subjective experiences exist a priori, irreducibly; the brain merely selects those that are useful for the survival of the physical body. The brain activation patterns that ordinarily correlate to conscious experience reflect the 34 filtering process at work: They are analogous to the circuit oscillations in the radio’s tuner, which correlate tightly to the sounds the radio produces. The presence of such circuit oscillations obviously does not mean that the radio is generating the broadcast signal itself, but merely selecting a subset of information from a preexisting signal. Analogously, brain activation patterns do not mean that the brain is generating the correlated conscious experience, but merely selecting it from a broader, irreducible superset (see again Fig.5). Figure 16: The dual filter hypothesis: The human brain filters conscious information states such as qualia from the collective consciousness field (left), but at the same time it filters subluminal information by attitudinal behavior, modified from Jahn and Dunne, 2004, 2007 Therefore, the ordinarily observed correlation between brain and mind states is a direct and necessary consequence of this selective filtering out of subjective experience: when the filtering mechanism (i.e. the brain) is interfered with – physically, as in a blow to the head, or chemically, as during anesthesia– the filtering process that modulates our conscious experience is perturbed, so that corresponding perturbations of experience follow ( Fig. 5). In conclusion: the hypothesis offered here remains consistent with all observed correlations between subjective experience and measurable brain states. For there to be a survival advantage in capturing an otherwise unbound consciousness within the space-time confines determined by the brain, consciousness must have materially-irreducible causal efficacy on brain function. In other words, there must be downward causation from consciousness towards brain 35 structure and/or activity, otherwise consciousness would be merely a useless ‘spectator’ providing no survival advantage. As it turns out, there is indeed significant empirical evidence that downward causation does occur. Experiments have been performed in which subjects could physically alter their own neuronal wiring – thereby reversing previously diagnosed brain pathologies – simply by directing their conscious intent (Schwartz and Begley, 2004). This surprising effect is known as ‘self-directed neuroplasticity,’ and it suggests that conscious intent is not only ontologically independent from, but can also causally affect, brain activity and structure, thereby potentially tilting the survival fitness of an organism. The first aspect of a ‘dual filter’ hypothesis, as proposed in the present paper, implies that consciousness, in its unfiltered state, is unbound. As such, consciousness is fundamentally unitary and non-individualized. The emergence of multiple, separate, and different conscious perspectives, or egos, is a consequence of the filtering process: different egos, occupying different points in the fabric of space-time, retain awareness of different subsets of a universal superset of all potential subjective experiences; the rest being filtered out. It is the differences across subsets that give each ego its idiosyncratic characteristics, personal history, and sense of personal identity. The part of the universal superset of subjective experiences that is filtered out becomes, then, the unconscious mind of the respective ego. Since each ego retains only an infinitesimally small part of the universal superset. Importantly, the ‘filter’ hypothesis predicts that one can conceivably have experiences that do not correlate to one’s brain states. Since here the brain is seen merely as a mechanism for filtering out experiences, it is conceivable that, when this mechanism is interfered with so as to be (partially and/or temporarily) deactivated, one’s subjective experience could delocalize, expand beyond the body in time and space, and perhaps even beyond time and space as such. In other words, the ‘filter’ hypothesis predicts that transpersonal, nonlocal experiences can conceivably happen when particular brain processes are (partially and/or temporarily) deactivated. The key element of this second prediction of the ‘dual filter’ hypothesis is that non-local, transpersonal experiences are predicted to correlate precisely to certain reductions of brain activity. Moving now to the second key prediction of the ‘dual filter’ hypothesis, there is indeed a broad pattern of empirical evidence associating nonlocal, transpersonal experiences with procedures that reduce brain metabolism (Fig. 17): a) Fainting caused by asphyxiation or other restrictions of blood flow to the brain is known to sometimes induce intense transpersonal experiences and states of non-locality. The highly dangerous ‘chocking game,' played mainly by teenagers worldwide, is an attempt to induce such experiences through partial strangulation, often at the risk of death (Neal, 2008). Erotic asphyxiation is a similar game played in combination with sexual intercourse. The effect has been described as 'a lucid, semi-hallucinogenic state [which,] combined with orgasm, [is said to be] no less powerful than cocaine' (Shuman, 2007) b) Pilots undergoing G-force induced loss of consciousness (G-LOC) – where blood is forced out of the brain, significantly reducing its metabolism – report experiences similar to the notoriously nonlocal and transpersonal NDE (Whinnery and Whinnery 1990); 36 Figure 17: Various conditions that reduce overall brain activity can promote perception of universal consciousness or mystic experiences c) The technique of “Holotropic Breathwork”, as well as more traditional Yogic breathing practices, use a form hyperventilation to achieve a similar effect: They increase blood alkalinity levels, thereby constricting blood vessels in the brain and causing hypoxia and dissociation (Rhinewine and Williams, 2007). This, in turn, reportedly leads to significant transpersonal, non-local experiences (Taylor, 1994). d) Psychedelic substances have been known to induce highly complex, intense, non-local, transpersonal experiences (Strassman et al. 2008). It had always been assumed that they do so by exciting the parts of the brain correlated to such experiences. Yet, a very recent study has shown that psychedelics actually do the opposite. The study reported that ‘profound changes in consciousness were observed after [the administration of the psychedelic], but surprisingly, only . in cerebral blood flow were seen. (Carhart-Harris et al. 2012a) Indeed, the researchers ‘observed no increases in cerebral blood flow in any region (Carhart-Harris et al. 2012b). Even more striking, they reported that ‘the magnitude of this decrease [in brain activity] predicted the intensity of the subjective effects.’ (Carhart-Harris et al. 2012:1) In other words, the intensity of the experience was inversely proportional to the activation of the brain, precisely as predicted by the ‘filter’ hypothesis; 37 e) The use of Transcranial Magnetic Stimulation (TMS) can inhibit cortical function in highly localized areas of the brain by extinction of the associated electromagnetic fields. When the neuronal activity in the angular gyrus of a patient with epilepsy was inhibited in this way, Out of Body Experiences (OBEs) were reportedly induced (Blanke, 2002). f) If the trend above is consistent, we should be able to extrapolate it further: Brain damage, through deactivating certain parts of the brain, should also induce non-local, transpersonal experiences under the right circumstances. And indeed, this has been reported. Two prominent examples are the case of neuroanatomist Dr. Jill Bolte Taylor, 2009, who underwent a profound transpersonal experience as a consequence of a stroke, and a systematic study recently carried out in Italy (Urgesi et al. 2010). In the Italian study, patients were evaluated before and after brain surgery for the removal of tumors. Statistically significant increases in feelings of self-transcendence were reported after the surgery. g) Treated before, Near-Death Experiences are the ultimate example of non-local, transpersonal experiences associated with not only reduced, but absent brain activity. Evidence for the validity of NDEs continues to be collected under scientific protocols, and has been mounting (Greyson, 2019, Greyson and Kelly, 2009). The pattern here is not only clear, but striking. The most complex, coherent, intense, non-local, and transpersonal experiences people report are associated precisely with reductions, or even elimination, of brain metabolism. This is consistent with the ‘filter’ hypothesis discussed here, and contradicts the materialist assumption. Conclusions of this section The broad pattern that associates peak transpersonal, non-local experiences with reductions in brain metabolism seems to contradict the tentative, promissory materialist solution to the mind-body problem. Instead, it substantiates the notion that the brain is a kind of ‘filter’ of consciousness, which selects from a universal superset of irreducible subjective experiences those which correlate with the space-time locus of the body. It is reasonable to think that there were survival advantages for the brain to evolve this capacity, which the empirical evidence for downward causation further substantiates. This ‘filter’ hypothesis explains how traditional techniques for the attainment of transpersonal insight work: by reducing the activity of certain brain regions, they (partially and/or temporarily) take the filtering mechanism offline, allowing consciousness to de-clench and expand beyond the space-time locus of the body. From this perspective, physical death is not the end of consciousness, but its liberation. Final Conclusions and Summary In treating hard problems in consciousness conceptions (Chalmers, 2010, 2019), we should formulate the proper scientific question: why do we propose a mental consciousness on basis of a 4th dimension? In the present work, we advocate an extra spatial dimension together with symmetrical time concept, thus a bidirectional arrow of time. We also emphasize the importance of proper definitions: one should clearly differentiate between consciousness in general (awareness due to integration of information from the environment that all living beings have in order to survive and that is for 90% automatic or subconscious) 38 and self-consciousness representing the ability that we can be conscious of being conscious and can perform introspection. The latter requires an extra degree of freedom, like a certain distance to ourselves and thus an extra 4th spatial dimension. The memory workspace that is conceptualized in the present study is called mental, since it is not observable and according to known principles of physics can be localized in a 4th dimensional (see Fig. 11). Why is such a construction necessary anyhow? This is in order to communicate with the supposed mental, and 4th dimensional universal consciousness. In fact, the event horizon workspace takes care of interfacing a purely mental domain with a 3-D material brain. The 4-D downscaling to 3-D world occurs by known holographic principles: our 3-D brain only collects the shadows of a 4-D mental world (Deli et al, 2019). In order to treat NDE experiences anyway, we depend on material verbal report of NDE experiencers. At the same time some of us can feel the 4-D world in meditative introspective conditions and other special conditions as treated above, that is, if the normal filter for such experiences of our material brain is somehow removed (see previous section on alternative filter hypothesis). The mechanism of holographic communication occurs by wave inference of photons and solitons as known from ZPE field. This latter is an immaterial and not observable wave phenomenon that exhibits mental properties. It can be conceived as a reality-steering implicate order, that entertains pilot waves and their implicit back reaction. The latter implies a non-static and permanently updated universal consciousness to which all individuals contribute trough bidirectional information flux. Self-consciousness of humans, thus, depends on contact with a 4-D personal workspace that we conceive as mental and, as such, is an integral part of 4-D universal consciousness. This pictures how our 3-D material brain can interact and communicate with the 4th dimension. Therefore, the definition of Fingelkurz, 2010, 2014.: it can be seen as a mental workspace associated with the brain but not reducible to the brain (not produced by the brain!). The latter implies a potential independence of this personal memory workspace from the material brain Fig. 17), that becomes manifest if connection with the material brain is temporarily disconnected (NDE) or permanently broken down (Death) (see Fig. 11). In reverse, we can faithfully hypothesize that NDE experience is a direct and strong demonstration of the existence of universal consciousness, since it displays individual life panorama’s and deep visions not explainable by material modalities. Here is what the NDE reports tell us about space and time: • Time does not run in the usual way; instead there is an experience of an eternal now. • Time does have representation of past, present, and future states, but not as a unidirectional; flow, such as we experience here. Instead, different times correspond to different locations in a multidimensional space which is reported to be higher than the usual four of space-time. • An observer located within the universal consciousness, is able to simultaneously view segments of the past, present, and future at will (the future is at times reported to be fuzzy). • Space has a form of reality in the universal consciousness realm, because entities are experienced in perception to be separated from each other, and at varying degrees of separation. 39 • One of the remarkable features of the function of consciousness operating within the universal domain is that our integral mental workspace or” the soul” is able to move about and across any “distance,” however great it appears to be, with apparent instantaneity. Figure 18: The mental workspace associated with the brain is an integral part of a cyclic or spiral arrangement of vibratory elements from the micro- (right part) to macro level (left part), indicating that cosmic consciousness can be conceived as producing matter and vice versa through a common field : the spin network at the Planck scale (from Meijer, 2014, 2015) The main thesis of this article on the nature of space, time, and consciousness is that in science and philosophy the dominant paradigm of materialism should accommodate ongoing research that implies materialism is incomplete for explaining all of reality. This, in spite of the recent claim that “consciousness is a state of matter” (Tegmark, 2015). For this peculiar concept the author did make a special invention: that of a supposed “perceptronium”, being a hypothetical substance that feels subjectively self-aware and stores and processes information……., and, ultimately, may turn out as Tegmark’s own “discovery of mind”. Thus, the treated reports from NDE/OBEs and paranormal phenomena imply that there is a reality hidden from ordinary materialistic sense perception that informs our brain and mind to form true knowledge. The latter process requires a physically defined interfacing of a 4-D realm with our 3-D world. This unique 40 feature can be realized by a toroidal event horizon workspace that allows symmetric 4-D to 3-D information flux and holographic 2-D personal memory integration. A further pre-requisite for information transfer between the sub-conscious brain and the associated workspace is non-neuronal information processing under NDE conditions at a flat EEG. This combination of physical aspects of NDE (mental/physics bridging) presented here, is broadly supported in current literature with respect to harmonic brain oscillations (Atasoy, 2018, Deli et al, 2019), the concept of a mental workspace associated with the brain (Fingelkurz, 2016, Hardy, 2018) and the aspect of a fourth spatial dimension in brain function, (Tozzi and Peters, 2016) as well as that of toroidal geometry of brain function (Tozzi and Peters, 2017, Deli et al. 2019 and Atasoy et al, 2016). In his thesis on the beginning and end of our Universe, Vidal (2012) stated: “But how can we imagine to seriously care for such an issue as cosmological immortality? We can summarize five steps towards it. The first is to realize that your individual death is normal and inevitable in the long term. The second is to develop psychologically, and fulfill all your needs to grow the hierarchy of needs up to the need of self-transcendence. You then surpass yourself to become compassionate and identify with the process of cosmic evolution. Even if you accept individual death, you still refuse death as a whole, namely the idea that nothing would continue to evolve after the predictable death of your body, society, species, Sun, galaxy and universe. You then set the immortality of the evolutionary process as a goal”. Finally, Schwarz, 2019, in a lucid report recently postulated: “In mind/body research particularly, the issue of consciousness has attained a new prominence, as evidenced by the growing number of placebo studies, research on meditation mindfulness and its effects on the brain, and its practical use in things like posttraumatic stress disorder (PTSD); studies on the role of psychophysical self-regulation in healing; prospective research on near-death experiences (NDE); studies suggesting the reality of reincarnation; and research into the relationship of genius and spirituality. The great irony is: from where does an Einstein, a Newton, a Planck, a Ramanujan, a Jung, a Salk come? The answer, as each of them has said quite clearly, is that their great insight came in a special state of consciousness, when all things seemed interconnected and interdependent and out of space and time. Those individuals who produce extraordinary research do so not by force of intellect or will alone, although these are important, but because they have had nonlocal intuitional insights at the same time that there was a crisis. What makes these key figures revolutionaries, then, is not just the quality of their work. They are also revolutionaries because of the source, mechanism unknown, from which their information derives. At the deepest level the process by which the information is obtained is as revolutionary as the information itself. Second, the already recorded 13 million near-death experiencers will increase in number thanks to increasingly sophisticated acute care medicine and cardiopulmonary resuscitation………We are moving into a new world!” Dirk K F Meijer, Groningen, 11-10-2019 41 References Aharonov Y, Popescu S, Rohrlich D, Skrzypczyk P. 2013. Quantum Cheshire Cats. New J. of Physics; 15, 1-8. Aharonov Y, Popescu S, Tollaksen J. 2010. A time-symmetric formulation of quantum mechanics, Physics Today, November 2010; 27-32. Akhmet MA, Fen M O, 2014. Generation of cyclic/toroidal chaos by Hopfield Neural Networks Neurocomputing,; 145, 230-239. Alfonso-Faus A, 2011. Cosmology, holography, the brain and the quantumvacuum, https://arxiv.org/ftp/arxiv/papers/1102/1102.0443.pdf Amoroso R L 1999. An Introduction to the Noetic Field Theory: the Quantization of Mind. Noetic Journal; vol.2, 18-37. Amoroso, R L, 2003 The Physical Basis of Consciousness: implications for Noetic Medicine and Transpersonal Psychology.in: Brain and Consciousness. Rio de Janeiro, Brazil Anjamrooz et al. The Cellular Universe: A New Cosmological Model Based on The Holographic Principle. Int. J. phys. Sciences, 2011; 6, 2175-2183. Atasoy S, Deco G, Kringelbach ML, Pearson J, 2018. Harmonic Brain Modes: A Unifying Framework for Linking Space and Time in Brain Dynamics. Neuroscientist. Jun;24(3):277-293. doi: 10.1177/1073858417728032. Epub Sep 1. 2017 Auletta G, Ellis GFR and Jaeger L, 2008. Top-down causation by information control: from a philosophical problem to a scientific research programme. J. R. Soc. Interface; vol. 5, no. 27, 1159-1172. Aurich, R, Janzer HS, Lustig S. & Steiner F, 2008. Do we live in a small universe? Class. Quant,; Grav://arxiv.org/abs/0708.1420. Batiz Z, Milovanovic D, 2017. Quantum Holography and Agency: Toward a Formalism of Schema QD. NeuroQuantology; 15, 45-59 Beichler JE. The Evolutionary Imperative of Consciousness Journal of Biological Physics, 2012a; Volume 38, pp. 113-120. Beichler, J E. “The neurophysics of mind and consciousness: A theoretical model of emergence, evolution, psi and survival.” A presentation made at the SMN/SSE Conference at An Grianan AEC, Ireland, 2012b, “http://www.neurocosmology.net”. Beichler, J E. The neurophysical basis of mind and consciousness: An electromagnetic model of the neuron. A presentation at the annual Spring Meeting of the OSAPS, Ohio State 68 University, April 2012c. “http://www.neurocosmology.net”. Berkovich-Ohana A, Glicksohn J. The Consciousness State Space (CSS)-a Unifying Model for Consciousness and Self. Front. Psychol; 2014. Bérut A B, Arakelyan A, Petrosyan A, Ciliberto S, Dillenschneider R, Lutz E. Experimental verification of Landauer’s principle linking information and thermodynamics, N AT U R E Letters 2012; 483, p 187-190. Bieberich E. Introduction in the Fractality Principle of Consciousness and the Sentyon Postulate. Cognit Comput., 2012; 4, 13-28. Bjerve, A. The Fractal-holographic Universe, 2016. http://holofractal.net/the-holofractographic Blanke, O. et al. 2002. “Stimulating illusory own-body perceptions: The part of the brain that can induce out-of-body experiences has been located.” Nature 419: 269–270. Bohm D, FD Peat, 2008.Science, Order and Creativity, Routledge, 2nd ed. 2000 (transferred to digital printing, Routledge): ISBN 0-415-17182-2. Bohm D, 1980. Wholeness and the implicate order. London: Routledge & Kegan Paul, 42 Bohm, D, Hiley BJ, 1993. The undivided universe. An ontological interpretation of quantum theory,. London: Routledge. Bókkon I, D'Angiulli A. Emergence and transmission of visual awareness through optical coding in the brain: a redox molecular hypothesis on visual mental imagery. Biosci. Hypotheses 2, 2009; 226–232.10.1016/j.bihy.2009.01.009. Bókkon I, Mallick BN, Tuszynski J A, 2013. Near death experience: a multidisciplinary hypothesis. Frontiers in Human Neuroscience; 7, 1-11. Bostrom N, 2003. Are You Living In a Computer Simulation? Philosophical Quarterly, Vol. 53, No. 211, 243-255. Braude SE, 2003. Immortal Remains. The Evidence for Life after Death., Rowman & Littlefield Publ., Oxford, UK Brown W, 2019. Unified Physics and the Entanglement Nexus of Awareness. NeuroQuantology, 17, 40-52 Burke RC, Persinger MA, 2013. Convergent Quantitative Solutions Indicating the Human Hippocampus as Singularity and Access to Cosmological Consciousness.;11,1-7 Caligiuri LM, 2015. Zero-Point Field, QED Coherence, Living Systems and Biophoton Emission. Open Journal of Biophysics; 5: 21-34. Capra F, 1996. THE WEB OF LIFE. A New Scientific Understanding of Living Systems, Anchor Books, New York Carhart-Harris, R. L. et al.. (2012). Neural correlates of the psychedelic state as determined by fMRI studies with psilocybin. Proceeding of the National Academy of Sciences of the United States of America, 109 (6), 2138–2143. Carhart-Harris, R. L. et al.. (2016). Neural correlates of the LSD experience revealed by multimodal neuroimaging. Proceeding of the National Academy of Sciences of the United States of America (PNAS Early Edition), doi: 10.1073/pnas.1518377113. Carter C, 2012. Science and the Afterlife Experience. Evidence for the Immortality of Consciousness. Inner Traditions, Rochester Carter P J, 2014. Quantum Spacetime and Consciousness. NeuroQuantology; 12: 46-75. Cavalleri, G., Barbero, F., Bertazzi, G., Cesaroni, E., Tonni, E., Bosi, L., … Gillies, G. T. (2010). A quantitative assessment of stochastic electrodynamics with spin (SEDS): physical principles and novel applications. Frontiers of Physics in China, 5(1). Chalmers D J, 1995. "Facing Up to the Problem of Consciousness". Journal of Consciousness Studies ; 2 (3):200-219. http://consc.net/papers/facing.pdf. Chalmers D J, 2018. The Meta-problem of Consciousness. J. of Consciousness Studies, 25, 6-61 Chiatti L, 2007. Fantappié-Arcidiacono Theory of Relativity Versus Recent Cosmological Evidences : A Preliminary Comparison, , https://arxiv.org/ftp/physics/papers/0702/0702178.pdf. Chown M, 2007. The Never-Ending Days of Being Dead, Faber & Faber, London Cifra M,. Fields JZ, Farhadi A. 2010. Electromagnetic cellular interactions. Progress in Biophysics and Molecular Biology; 1-24, doi:10.1016/j.pbiomolbio.2010.07.003. Cole, D C and Zou, Y, 2004. Analysis of orbital decay time for the classical hydrogen atom interacting with circular polarized electromagnetic radiation. Phys. Rev. E 69, 016601, Cole D C, Zou, Y., "Quantum mechanical ground state of hydrogen obtained from classical electrodynamics", Phys. Lett. A 317, 14-20 (2003). Cramer J (1988). An Overview of the Transactional Interpretation. International Journal of Theoretical Physics 27, 227. Damasio A, Carvalho, G B, 2013. "The nature of feelings: Evolutionary and neurobiological origins". Nature Reviews. Neuroscience; 14 (2): 143–52. Davies P C W, 2014. Does quantum mechanics play a non-trivial role in life?" BioSystems; 78: pp. 69–79 43 Davies P C W, 2007. The implications of a Cosmological Information Bound for Complexity, Quantum Information and Nature, http://power.itp.ac.cn/~mli/pdavies.pdf of Physical Law, http://power.itp.ac.cn/~mli/pdavies.pdf. De la Peña L, Cetto AM, 1994. Quantum Phenomena and the Zero point Radiation Field. Foundations of Physics; 24.6: 917-48 De la Peña, L., & Cetto, A. M. (2006). The Foundations of Linear Stochastic Electrodynamics. Foundations of Physics, 36, 350-368. http://dx.doi.org/10.1007/s10701-005-9020-1 De Loof A, 2016. The cell's self-generated “electrome”: the biophysical essence of the immaterial dimension of Life?; https://www.researchgate.net/publication/304712940 Deli F, Tozzi A, Peters JF, 2017. Relationship between short and fast brain timescales. Cogn. Neurodyn., 11, 539-552 Dotta, BT. 2013. Ultraweak photon emission in cells: coupling to molecular pathways, applied magnetic fields, and potential non-locality, thesis Eagleman, D. (2011). Incognito: The Secret Lives of the Brain. New York: Canongat Eccles JC, 1992. Evolution of consciousness. Proc Natl Acad Sci USA.;89:7320-4 Fingelkurts AA, Fingelkurts AA, Neves CFH, 2010. Emergent Monism, Biological Realism, Operation and Brain-Mind Problem. Physics of Life Reviews; 7, 264-268. Fingelkurts AA, Fingelkurts AA, 2014. Present moment, past and Future: Mental Kaleidoscope. Front. Psychol. May,. Firstenberg O, Peyronel T, Liang QY, Gorshkov AV, Lukin V, Vuletic MD, 2013. Attractive photons in a quantum nonlinear medium. Nature (London) 502, 71–75 https://dspace.mit.edu/openaccess-disseminate/1721.1/91605 Forberg J L, 2006.Quantum Consciousness and your Immortality, Forberg J L, 2007. Quantum physics and your immortality. Xlibris Corp Ford L H, Roman TA, 2003.. Negative energy, wormholes and warp drive, Sc. Am.; pp. 84-91. Forsdyke D R, 2014. Long-term memory: scaling of information to brain size. Front. Human Neuroscience; 8, 1-4. Forshaw S D, 2016. Memory of a universal mind. Galactic structures and their existence as cognitive operators in an emerging universally conscious machine. ResearchGate; 309209078. Forshaw S D, 2016. The Third State. Towards a quatum theory of consciousness. NeuroQuantology; 14, : 49-61. Galloni M, 2012. The heresy of Fantappié and Teillard and the converging revolution. Syntropy,; 1: 79-84. Gambini R and Pullin J, 2013. Loop Quantization of the Schwarzschild Black Hole, Phys. Rev. Lett.; 110: 211301 Geesink JH, Meijer D K F, 2016. Quantum Wave Information of Life Revealed: An Algorithm for Electromagnetic Frequencies that Create Stability of Biological Order, with Implications for Brain Function and Consciousness. NeuroQuantology, ; 14: 106-125, file:///C:/Users/Dick/Documents/911-2447-1-PB.pdf Gonçalves, Ó. F., Rêgo, G., Oliveira-Silva, P., Leite, J., Carvalho, S., de Souza-Queiroz, J., . . . Boggio, P. S. (2017). Is the relationship between mind wandering and attention culture-specific? Psychology & Neuroscience, 10(2), 132-143.http://dx.doi.org/10.1037/pne0000083 Goswami, A 1990. Consciousness in Quantum Physics and The Mind-Body Problem. The Journal of Mind and Behavior; 11, 1, p. 75-96. Grandpierre A, 2014. Biologically Organized QuantumVacuum and the Cosmic Origin of Life. Analecta Husserliana,; 116, chapter 1. Green B R, Levin J, 2007. Dark energy and stabilization of extra dimensions,; arXiv:0707.1062v2 [hep-th]. Greyson B, 2010. Seeing Dead People Not Known to Have Died: “Peak in Darien” Experiencesa Anthropology and Humanism; Vol. 35, Issue 2: pp 159–171, Greyson, 2019. Nature of Mind and Consciousness. Is consciousness produced by the brain? https://www.youtube.com/watch?v=2aWM95RuMqU 44 Grof S. 1987. Beyond the Brain; Birth, Death and Transcendence in Psychotherapy; New York: State University of New York Press. Hameroff S, Penrose R. 2014. Consciousness in the universe: A review of the 'Orch OR' theory. Phys Life Rev,; Mar 11(1):39-78. Hameroff S, Tuszynsky J, 2010. Quantum states in protein assemblies: The essence of Life?; Hameroff S, 2012. How quantum brain biology can rescue conscious free will. Front. Integrative Neuroscience; 6: 1-17 Hameroff S and Chopra D, 2013, The Quantum Soul: A Scientific Hypothesis. In:Exploring Frontiers of Brain-Mind Relationship. Mindfulness in Behavioral Health, Springer Haramein N, Brown, WD and Val Baker A, 2016. The United Spacememory Network: From Cosmogenesis to Consciousness. NeuroQuantology; 14: 1-15. Hardy C H 2016. ISS Theory: Cosmic Consciousness, Self, and Life Beyond Death in a Hyperdimensional Physics. J. Consciousness Exploration and Research, 7,; issue 11: 1012-1035. Hasson U, 2012. Ghazanfar AA, Galantucci B, Garrod S, Keysers C,. Brain to brain coupling: A mechanism for creating and sharing a social world Trends Cogn. Sci.,;16: 114-12. Henry R C, 2005. The Mental Universe, Nature; 436(7): 29. Hofstadter, D. (2007). I Am a Strange Loop. Irwin K, 2014. A New Approach to the Hard problem of Consciousness: A quasi-crystalline language of primitive units of consciousness in quantized spacetime. J. Consciousness Exploration & Research; 5: 483-497.. Jahn R G and Dunne B J ,2007. A modular model of mind/matter manifestations. Explore, 3:311-24, reprinted from J. Scientific. Exploration, 2001 Jahn RG, Dunne B J. 2004. Sensors, Filters, and the Source of Reality. Journal of Scientific Exploration; 18(4): 547–570. John E R, 2001. A field theory of consciousness. Consciousness and Cognition; 10:184-213 Kaku M, 2005. Parallel Worlds. The science of alternative universes and our future in the cosmos, Kastrup B, 2016. The Idealist View of Consciousness After Death, Journal of Consciousness Exploration & Research| December 2016 | Vol. 7 | Issue 11 | pp. 900-909 Kastrup B, 2017. Self-transcendence correlates with brain function impairment. Journal of Cognition and Neuroethics 4 (3):33-42 https://philpapers.org/rec/KASSCW Kastrup B, 2018. The Universe in Consciousness. J. of Consciousness Studies, 25, 125-155 Kastrup B, 2019. Conflating Abstraction with Emperical Observation: the False Mind -Matter Dichotomy. Constructive foundations, 13, nr3, 341-360 Keppler J A, 2013.A new perspective on the functioning of the brain and the mechanisms behind conscious processes. Frontiers in Psychology, Theoretical and Philosophical Psychology; pp1-6. Keppler J A, 2012. Conceptual framework for consciousness based on a deep understanding of matter. Philos. Study; 2: 689–703. Keppler J A. 2016. On the universal mechanism underlying conscious systems and the foundations for a theory of mind. Open J. of Philosophy; 6: 346-367. Kida T, Tanaka E, Kakigi R, 2016. Multi-dimensional dynamics of human electromagnetic brain activity. Front. Human Neuroscience; 9: 1-20. King C C 2003. Chaos, Quantum-transactions and Consciousness, NeuroQuantology,; 1(1): 129-162. Knierim JJ, Zhang K, 2012. Attractor Dynamics of Spatially Correlated Neural Activity in the Limbic System Ann. Rev. Neurol. 10.1146/2012; 062111-150351. Lake J, 2015. The near death experience: implications for a more complete theory of consciousness. Quantum Biosystems; 6: 131-138. 45 Lanza R L,2012. Does Death Excist ? http://www.robertlanza.com/does-death-exist-new-theory-says-no-2/ Laszlo E, 2012. Cosmic Symphony: A Deeper Look at Quantum Consciousness,; http://www.huffingtonpost.com/ervin-laszlo/cosmic-symphony-a-deeper_b_532315.html. László E, 2007. The Akashic Field. New York: Dutton,. Lehar S, 2003. Harmonic Resonance Theory: An Alternative to the "Neuron Doctrine" Paradigm of Neurocomputation to Address Gestalt properties of perception. Perception;;April 32: 423-448 http://cns-alumni.bu.edu/~slehar/webstuff/hr1/hr1.html. Levine, J. 1983. “Materialism and qualia: the explanatory gap”. Pacific Philosophical Quarterly, 64: 354-361. Levine, J "Conceivability, Identity, and the Explanatory Gap" in Stuart R. Hameroff, Alfred W. Kaszniak and David Chalmers (eds.), Towards a Science of Consciousness III: The Third Tucson Discussions and Debates, The MIT Press, 1999,. pp 3-12. Libet B, 1994. A testable field theory of mind-brain interaction. J. Consciousness Studies; 1: 119–126.. Libet B, 1996. Conscious Mind as a Field J. theor. Biol.; 178: 223–224. Linton O, 2015. Consciousness as a meta-phenomenon,; http://quantum-mind.co.uk/consciousness-as-a-meta-phenomenon/. Lloyd N J, 2007. Black hole computers,. https://www.scientificamerican.com/.../black-hole-computers-2007. Lloyd S, 2006. Programming the Universe: A Quantum Computer Scientist Takes On the Cosmos, Knopf Doubleday Publishing Group (Random House). Loll R. 2011. Quantum Gravity at the Planck scale: getting a hold on spacetime foam,; Luminet J P, 2016. The holographic Universe,; https://arxiv.org/abs/1602.07258. Maldacena J, Susskind K, 2013. Cool horizons for entangled black holes, High Energy Physics - Theory (hep-th),; DOI: http://arxiv.org/pdf/1306.0533v2.pdf. Martin F, Carminati F, Galli Carminati, 2013. Quantum Information Theory Applied to Unconscious and Consciousness. NeuroQuantology; 11: 16-33. McFadden J, 2007. Conscious electromagnetic (CEMI) field theory. NeuroQuantology; 5(3):262–270. Meijer D K F, 2012. The Information Universe. On the Missing Link in Concepts on the Architecture of Reality. Syntropy Journal,;1:1-64. http://www.sintropia.it/english/2012-eng-1-1.pdf. Meijer D K F, 2013. Information: What Do You Mean? Syntropy Journal,;3:1-49, http://www.lifeenergyscience.it/english/2013-eng-3-01.pdf Meijer D K F, 2013. Immortality: Myth or Becoming Reality? On the Conservation of Information. Syntropy Journal,; 3: 166-203.http://www.lifeenergyscience.it/english/2013-eng-3-04.pdf. Meijer D K F, Korf J (2013) Quantum modeling of the mental state: the concept of a cyclic mental workspace. Syntropy Journal 2013 nr 2: pp1-41: http://www.lifeenergyscience.it/english/2013-eng-1-1.pdf Meijer D K F, Raggett S, 2014. Quantum Physics in Consciousness Studies. The Quantum Mind Extended, Quantum Mind;http://quantum-mind.co.uk/wp-content/uploads/2014/11/Quantum-Ph-rev-def-2.pdf Meijer D K F. 2014. The Extended Brain. Cyclic information flow in a quantum physical realm. NeuroQuantology; 12-2: 180-200. Meijer D K F. 2015. The Universe as a Cyclic Organized Information System. An Essay on the Worldview of John Wheeler. NeuroQuantology; 1: 1-40. Meijer D K F, Geesink J H. 2016. Phonon Guided Biology. Architecture of Life and Conscious Perception are mediated by Toroidal Coupling of Phonon, Photon and Electron Information Fluxes at Discrete Eigenfrequencies. NeuroQuantology; 16, issue 4:718-755 Meijer D K F, Geesink J H, 2017. Protein folding: the Role of Long- and Short Range Electromagnetic Pilot Mechanisms.; available at: ResearchGate, march, 46 https://www.researchgate.net/publication/315706536_The_Folding_of_a_Life_Proteins_On_the_role_of_long-and_short_range_electromagnetic_pilot_mechanisms Meijer D K F, 2017. Science and Art Modeled as a Flux of Information Using Toroidal Geometry, Open Journal of Philosophy, 8, 365-400. doi: 10.4236/ojpp.2018.84026. Meijer D K F, 2019. Universal Consciousness. Collective Evidence on the Basis of Current Physics and Philosophy of Mind. Part 1. ResearchGate, https://www.academia.edu/37711629/Universal_Consciousness_Collective_Evidence_on_the_Basis_of_Current_Physics_and_Philosophy_of_Mind._Part_1 Meijer D.K.F, Geesink J. H, 2019. Life and Consciousness are Guided by a Semi-Harmonic EM Background Field. NeuroQuantology ,Vol. 17, Issue 4, 37-44 https://www.neuroquantology.com/index.php/journal/article/view/2074 Meijer D K F, Sonderkamp T, Jerman I, Melkikh, Sbitnev,V I. 2019. Superfluid quantum space Consciousness Arises from Hydrodynamic Superfluid Quantum Space as a Conformal Mental Attribute of Space-time. In preparation Meijer D K F, 2019. Do thoughts really come only from the brain: the deeper connective principles of consciousness. ResearchGate. https://www.researchgate.net/publication/334490501 Merali Z, 2008. Doughnut-shaped Universe bites back. Astronomers say Universe is small and finite. Nature; doi:10.1038/news.2008.854. Metzinger, T. (2004[2003]). Being No One: The Self-Model Theory of Subjectivity, 2nd Edn. Cambridge, MA: MIT Press. Mitchell ED, Staretz R, 2011 The Quantum Hologram and the Nature of Consciousness. Journal of Cosmology.;14:1-35. Nagel T, 2012. Minds and Cosmos. Why the materialist neo-darwinian conception of nature is almost certainly false,; Oxford Univ. Press. New York. Nation P, Jahansson J, Blencowe M, and Nori F, 2012. Stimulating uncertainty: amplifying the quantum vacuum with superconducting circuits. Review of Modern Physics; 81(1):1. Neal, R. M. 2008. The Path to Addiction: And Other Troubles We Are Born to Know. Bloomington, IN: AuthorHouse. Neppe V M, Close E R, 2017. Fifty discoveries that changed the world: why the triadic dimension distinction vortical paradigm makes a difference. IQ Nexus Journal, vol. 9 no.2 Nielsen MA, Chuang I, 2000. Quantum calculation and quantum information. Cambridge University Press,; http://www.michaelnielsen.org/qcqi/QINFO-book-nielsen-and-chuang-toc-and-chapter1-nov00.pdf. Pandarakalam J, 2019. Scientific evidence for discarnate existence and a search for the sacred, Neuroquuantology, 17, 45-59 Parnia S, Spearpoint K, de Vos G, Fenwick P, et al. 2014. AWARE - AWArenes during REsuscitation – a prospective study. Resuscitation; 85:1799–805 Parnia, S, 2017. Understanding the cognitive experience of death and the near-death experience. QJM: An International Journal of Medicine, 2017, 67–69 Pascual-Leone A, Walsh V, Rothwell J. 2000. Transcranial magnetic stimulation in cognitive neuroscience—virtual lesion, chronometry, and functional connectivity. Curr. Opin. Neurobiol. 10:232–37 Penrose R ,2010. Cycles of Time. An Extraordinary New View of the Universe. London: Bodley Head. Penrose R. On the Gravitization of Quantum Mechanics 1: Quantum State Reduction". Foundations of Physics 2014; 44: 557–575. Pereira A, Furlan FA. 2007. Biomolecular Information, Brain Activity and Cognitive Function. In: ARBS Annual Rev of Biomedical Sciences; 9: 12-29. 47 Pereira C, Reddy JSK, 2016. Near-Death Cases desegrating Non-locality/ Disembodyment vai Quantum Mediated Consciousness. J. Consciousness Exploration & Research; 7: 951-968. Pereira C, 2015. Electromagnetic Radiation, a Living Cell and the Soul, A Collated Hypothesis. NeuroQuantology; 13: 426-438. Persinger MA, Lavallee CF, 2010. Theoretical and experimental evidence of macroscopic entanglement between human brain activity and photon emissions: implications for quantum consciousness and future applications. J. Cons. Explor. Res.; 1: 785–807. Persinger M A. 2008. On the possible representation of the electromagnetic equivalents of all human memory within the earth’s magnetic field: implications for theoretical biology. Theor. Biol. Insights; 1: 3–11. Peterson JPS, Sarthour RS, Souza AM, Oliveira I S, Goold J, Modi K, Soares-Pinto DO, Céleri LC. 2016. Experimental demonstration of information to energy conversion in a quantum system at the Landauer limit. / Proceedings of the Royal Society A-Mathematical Physical and Engineering Sciences; Vol. 472, No. 2188, 20150813, 01.04. Phipps C, 2012. Death, Rebirth and everything in between. A scientific exploration. http://integral-options.blogspot.nl/2009/06/carter-phipps-death-rebirth-and.html Pizzi R, Fantasia A, Gelain F, Rossetti D, Vescovi A, 2004. Non-local correlations between separated neuronal networks. Proc SPIE. 5436:107-17. Pockett S, 2013. Field theories of consciousness. Scholarpedia; 8: 4951. Pockett S, 2012. The electromagnetic field theory of consciousness: a testable hypothesis about the characteristics of conscious as opposed to non-conscious fields. Journal of Consciousness Studies; 19 (11-12): 191-223. Pourhasan R, Afshordi N, Mann RB, 2013. Out of the White Hole. The Holographic Origin for the Big Bang. arXiv:1309.1487v2 [hep-th] 22 Sep. Prakash R. 2008. Global Workspace Model of Consciousness and its Electromagnetic correlates. Ann. Indian Acad. Neurol.; 11: 146-153. Pregnolato M, Pereira A, 2016. On the Possible Existence of Quantum Consciousness After Brain Death. J. Consciousness Exploration & Research; 7: 969-991. Pribram KH, 2014. Consciousness Reassessed. Mind and Matter; 2: 7-35. philpapers.org/rec/PRICR. Primack, J., & Abrams, N. E. (2006). The View from the Centre of the Universe. Discovering our Extraordinary Place in the Cosmos. London: Fourth Estate. http://www.chronon.org/reviews/view_from_centre.html Radin D I and Nelson R ,2006. Entangled Minds. Extrasensory experiences in the quantum reality. New York: Simon & Schuster. Radin D I, 2004. Event-related electroencephalographic correlations between isolated human subjects. J Altern Complement Med.;10(2):315-2 Rueda A. & Haisch, B. 1998. Contribution to inertial mass by reaction of the vacuum to accelerated motion Foundations of Physics, Vol. 28, No. 7, pp. 1057-1108 Rueda A & Bernard Haisch B, 2005. Gravity and the Quantum Vacuum Inertia Hypothesis Annalen der Physik, Vol. 14, No. 8, 479-498 Rhinewine, J P and Williams, O. J. 2007. “Holotropic Breathwork: The Potential Role of a Prolonged, Voluntary Hyperventilation Procedure as an Adjunct to Psychotherapy.” The Journal of Alternative and Complementary Medicine 13 (7): 771-776. Rodriguez, E., George, N., Lachaux, J. P., Martinerie, J., Renault, B., & Varela, F. J. (1999). Perception’s Shadow: Long Distance Synchronization of Human Brain Activity. Nature, 397, 430- 433. http://dx.doi.org/10.1038/17120 48 Richards TL, Kozak L, Johnson LC, Standish LJ, 2005. Replicable functional magnetic resonance imaging evidence of correlated brain signals between physically and sensory isolated subjects. J Altern Complement Med;11(6):955-63. Ring K, Cooper C. 1997. Near-Death and Out-of-Body Experiences in the Blind: A Study of Apparent Eyeless Vision,;http://www.newdualism.org/nde-papers/Ring/Ring-Journal%20of%20Near-Death%20Studies_1997-16-101-147.pdf. Rouleau N, Dotta B T, 2014. Electromagnetic fields as structure-function zeitgebers in biological systems: environmental orchestrations of morphogenesis and consciousness. Frontiers in Integrative Neuroscience; 8: 1-9. Rousseau D, 2011. Understanding spiritual awareness in terms of anomalous information access. The Open Information Science Journal; 3: 40-53. Schwartz JM, Stapp HP, Beauregard M, 2005. Quantum physics in Neuroscience and Psychology: A Neurophysical Model of Mind-Brain Interaction. Philos Trans R Soc Lond B Biol Sci.; 360(1458):1309-27. Schwarz and Begley, 2003. The Mind and the Brain: Neuroplasticity and the Power of Mental Force, Kindle Publ. UsA Schwartz, S A, 2013. Crossing the threshold: Non-local consciousness and the burden of proof. Explore, 9, 77-81 Setterfield B, 2002. Exploring the vacuum. J. Theoretics.; http://www.journaloftheoretics.com/links/papers/setterfield.pdf. Shoup R . 2011. Understanding Retrocausality. Can a Message Be Send to the past?; https://www.researchgate.net/publication/234885500_Understanding_Retrocausality-Can_a_Message_Be_Sent_to_the_Past. Shuman, G. D. (2007). Last Breath: A Sherry Moore Novel. New York, NY: Simon & Schuster. Sieb R., 2016. Human Conscious Experience is Four-dimensional and has a Neural Correlate, Modeled by Einstein’s Special Theory of Relativity. NeuroQuantology; 14, 630-644. Simon and Schuster, New York Sieb R, 2018. Space-time intervals underlie human conscious experience, gravity and theory of everything. NeuroQuntology, 16,49-64 Sirag S P. 1993. Consciousness, a hyperspace view; http://www.williamjames.com/Theory/Consciousness.pdf. Smythies J, 2003. Space, time & consciousness – Journal of Consciousness Studies; 10, (3): 47-5. Standish LJ, Kozak L, Johnson LC, Richards T, 2004. Electroencephalographic evidence of correlated event-related signals between the brains of spatially and sensory isolated human subjects. J Altern Complement Med.;10(2):307-14. Stapp, H. (2017) Retrocausation in quantum mechanics and the effects of minds on the creation of physical reality, AIP Conference Proceedings 1841, 040001 (2017); https://doi.org/10.1063/1.4982777. Struppa D C, Kafatos M, et al 2002. Category Theory as the Language of Consciousness Noetic Journal Vol. 3 No. 3, July Sullivan PR, 1995. Contentless consciousness and information processing theories of mind. Philosophy, Psychiatry & Psychology; 2: 51-59. Susskind L, 2016. Copenhagen vs Everett, Teleportation, and ER=EPR,; https://arxiv.org/abs/1604.02589. Talbot M, 2006. The Universe as a Hologram: Does Objective Reality Exist or is the Universe a Phantasm?; http://homepages.ihug.co.nz/~sai/hologram.html Tammiettto M, de Gelder B, 2010. Neural basis of the non-conscious perception of emotional signals. Nature Reviews Neurosciences; 1-13. Taneichi T, 2015. Physical Consciousness Outside the Brain: Parasite Fermion Model for Substance of Consciousness. NeuroQuantology; 13: 134-140. 49 Taylor, K. 1994. The Breathwork Experience: Exploration and Healing in Nonordinary States of Consciousness. Santa Cruz, CA: Hanford Mead. Tegmark M, 2008. The mathematical universe. Found Phys; 38:101-150, arXiv:0704.0646 [gr-qc]. Tegmark, 2015. Consciousness as a State of Matter. https://arxiv.org/abs/1401.1219. Tiller W A 1999. Towards a predictive model of subtle domain connections in the physical domain of reality. J. Scientific Exploration,; 13: 41-67. Tippler F J , 1994. The Physics of Immortality. MacMillan Tonneau F, 2004. Consciousness outside the head. Behavior and Philosophy; 32: 97-123. Tononi G, Boly, M, Massimini M, Koch C, 2016. Integrated information theory: from consciousness to its physical substrate. Nature Reviews Neuroscience; pp 1-12. Tononi G, Koch C, 2015. Consciousness: here, there and everywhere? Philosophical Transactions R. Soc.; 370. Toyabe S, Sagawa T Ueda M Meneyuki, Sano M, 2010. Supplementary information for information heat engine: converting information to energy by feedback control,; arXiv:1009.5287v2 [cond-mat.stat-mech]. Tozzi A, Peters J F, Jaušovec N, 2017. Repetitive Modular Oscillation underlies human brain electric activity; http://biorxiv.org/content/biorxiv/early/2016/08/31/072538.full.pdf. Tozzi A, Peters J F 2015. Brain Activity on a Hypersphere. Quantitative Biology Neurons and Cognition,; arXiv:1512.00036 [q-bio.NC]. Tozzi A, Peters J F. Towards a fourth spatial dimension of brain activity. Cognitive Neurodynamics 2016b; 10 (3): 189-99 PMID: 27275375. Urgesi, C. et al. 2010. “The Spiritual Brain: Selective Cortical Lesions Modulate Human Self Transcendence.” Neuron 65: 309–319. Van Bergen C, 2002. Verlangen naar Onsterfelijkheid (in Dutch), Damon, Budel, Belgium Van Lommel P, 2007. Consciousness Beyond Life: The Science of the Near-Death Experience, HarperCollins. ISBN 978-0061777264. Van Lommel P, van Wees R, Meyers V, Elfferich I, 2001. "Near-Death Experience in Survivors of Cardiac Arrest: A prospective Study in the Netherlands", The Lancet, 358(9298):2039–45, doi:10.1016/S0140-6736(01)07100-8. Van Raamsdonk M V, 2010. Building up spacetime with quantum entanglement. Gen Relativ Gravit; 42: 2323, https://arxiv.org/pdf/1005.3035.pdf. Venselaar M, 2011. The near death decoded in five phases theory. http://www.nderf.org/NDERF/Books/maureen_v_research.htm Verlinde E P. 2016. Emergent gravity and Dark Universe,; https://arxiv.org/abs/1611.02269. Verlinde E P. 2011. On the Origin of Gravity and the Laws Newton,; JHEP. arXiv:1001.0785Bibcode:2011JHEP...04..029V. doi:10.1007/JHEP04(2011)029. Vidal C, 2012. THE BEGINNING AND THE END. Meaning of Life in a Cosmological Perspective. Chapter 10.4: Voyage to five immortalities. Vidal C, 2010. Computational and biological analogies for understanding fine-tuned parameters in physics. Foundation of Science; 15(4): 375-393. Vitiello G. 2001. My Double Unveiled: The Dissipative Quantum Model of Brain,; John Benjamins Publishing. Wackermann J, Seiter C, Keibel H, Walach H, 2003. Correlations between brain electrical activities of two spatially separated human subjects. NeurosciLett.;336(1):60-4. Wheeler J A and Feynman, 1994. At Home in the Universe, New York Institute of Physics Wilczek, F. (2008). The Lightness of Being. Mass, Ether, and the Unification of Forces. New York: Basic Books. Wigner, E. P., 1983. Remarks on the mind-body problem. In: Quantum Theory And Measurement (John Wheeler and Zurek Wojciech, eds.) Princeton University Press, Princeton, NJ, p. 99 50 Wolf F A, 2008. Is the Mind of God Found in Quantum Field Theory? http://www.fredalanwolf.com/myarticles/Quantum%20Field%20Theory.pdf Wolf F A, 1999. THE QUANTUM PHYSICAL COMMUNICATION BETWEEN THE SELF AND THE SOUL Noetic Journal Volume 2, No. 2 Wolf F A, 1985. The Quantum Physics of Consciousness: Towards a New Psychology. Integr. Psychiatry; 3:235: 4-25, Wongjun. 2013. Casimir Dark energy, stabilization of extra dimensions and Gass-Bonet term,. Zeilinger A, 2003. Quantum Teleportation. Scientific Am.;Febr. 8-16. Update : http://www.univie.ac.at/qfp/publications3/pdffiles/2003-24.pdf. Zeilinger A, 2003. Why the Quantum? It from Bit? A Participatory Universe? Three Far-Reaching, Visionary Challenges from John Archibald Wheeler and How They Inspired a Quantum Experimentalist,; Charles L. Harper, Templeton Foundation Press. Zizzi P, Pregnolato M, 2013. Non-Turing Computable Origin of Cellular Life. Journal of Consciousness Exploration & Research | December; Vol. 4, Issue 10: 1062-1065. Further Reading: Alexander E, 2012. Proof of Heaven. Lennart Sane Agency Cave S, (2012). Immortality. The quest to live forever and how it drives civilization. Crown, New York Chown M, (2007). The Never-Ending Days of Being Dead, Faber & Faber, London Currivan J, (2017). The Cosmic Hologram. In-formation at the center of creation. Inner Traditions, Vermont, USA Elsaesser V E, (1997). On the other side of Life. Exploring the phenomenon of Near-Death Experience. Perseus Publ. Cambridge, Mass. USA Fontana D, (2005). Is there an Afterlife? A comprehensive overview on the evidence. Books, London Greenfield S, (2003). Tomorrow’s People. How 21-st century technology is changing the way we think and feel, Penquin, London Laszlo E, Peake A, (2014). Immortal Mind. Science and continuity of consciousness beyond the brain. Inner Traditions, Vermont, USA Leffert, M (2019) Cosmometry. Exploring the holo-fractal nature of the cosmos. Cosmometria Publishing, Santa Cruz, USA Lloyd S, (2006). Programming the Universe: A Quantum Computer Scientist Takes on the Cosmos, Knopf Doubleday Publishing Group (Random House). Peake A, (2008). Is there life after death ? The extraordinary science of what happens when we die. Arcturus Publ., , Penguin/Allen Lane, London 51 Pickover C A, (2007). Immortality. A Beginners Guide. Thunder’s Mouth Press, New York Radin D. (1997). The Conscious Universe. The Scientific Truth of Psychic Phenomena. New York, HarperEdge. Talbot M, (2006). The Universe as a Hologram: Does Objective Reality Exist or is the Universe a Phantasm?; http://homepages.ihug.co.nz/~sai/hologram.html Tippler F J ( 1994). The Physics of Immortality. MacMillan, USA Van Lommel P, (2007). Consciousness Beyond Life: The Science of the Near-Death Experience, HarperCollins. ISBN 978-0061777264. Venselaar M, 2011. The near death decoded in five phases theory. http://www.nderf.org/NDERF/Books/maureen_v_research.htm Weiner J, (2010). Long for this World. The Strange Science of Immortality, Ecco, New York Wolf F A, (1996). The Spiritual Universe. How Quantum Physics proves the Existence of the Soul Special journal issues on Afterlife can be found in: Journal of Consciousness Exploration & Research| December 2016| Volume 7 | Issue 11 | pp. 862-1244 Scientific GOD Journal | May 2017 | Volume 8 | Issue 5 | pp. 293-421 View publication stats | RESEARCH GATE |
Nilai Budaya dalam Cerita Rakyat Toraja Ratnawati) Balai Bahasa Ujung Pandang Abstrak Toraja adalah salah satu suku yang mendiami wilayah pegunungan di Sulawesi Selatan. Penduduknya yang berjumlah sekitar 450.000 jiwa masih tinggal di Kabupaten Toraja Induk dan Kabupaten Toraja Utara. Umumnya, penduduk ini menganut agama Kristen, sebagian lagi memeluk agama Islam, serta sebagiannya lagi masih ada yang menganut kepercayaan animisme yang dikenal dengan Aluk To Dolo. Kepercayaan Aluk To Dolo inilah yang mendasari pelaksanaan berbagai upacara yang memerlukan persembahan hewan kurban dalam jumlah nominal tinggi dalam kehidupan masyarakat Toraja. Tulisan ini memaparkan nilai budaya dalam Cerita Rakyat Toraja. Nilai budaya yang menonjol dalam Cerita Rakyat Toraja sebagian besar dipengaruhi oleh kepercayaan Auk To Dolo yang mencakupi hubungan manusia dengan Sang Pencipta, hubungan manusia dengan alam, hubungan manusia dengan sesama manusia, dan hubungan manusia dengan dirinya sendiri. Tulisan ini bertujuan menambah wawasan tentang kebudayaan Toraja yang selanjutnya dapat meningkatkan pemahaman terhadap salah satu kebudayaan yang ada dan berkembang di Sulawesi Selatan. Kata Kunci: Nilai budaya, Cerita Rakyat Toraja 1. Pendahuluan Sastra dan kebudayaan, baik sebagai satu kesatuan, maupun secara terpisah, yaitu ‘sastra’ dan ‘kebudayaan’ selalu dikaitkan dengan nilai-nilai positif. Artinya, sastra dan kebudayaan, yang dengan sendirinya dihasilkan melalui aktivitas manusia itu sendiri, berfungsi untuk meningkatkan kehidupan. Karya sastra sebagai katharsis (Aristoteles), aesthetic function (Mukarovsky), lango (Zoetmulder), aktivitas manusia sebagai pencerahan (abad Pertengahan), dan berbagai ) Magister Pendidikan, Pembantu Pimpinan pada Balai Bahasa Ujung Pandang Nilai Budaya Dalam Cerita Rakyat Toraja (Ratnawati) 49 definisi sastra sebagai hiburan, pada dasarnya menunjuk pada fungsi kemanusiaan tersebut (Ratna, 2005:9). Usaha pembinaan dan pengembangan kebudayaan nasional tidak dapat dilepaskan dengan upaya penggalian sumber-sumber kebudayaan daerah yang banyak tersebar di seluruh pelosok nusantara. Dalam konteks ini, kebudayaan daerah merupakan sumber potensial bagi terwujudnya kebudayaan nasional, sekaligus memberi corak dan karakteristik kepribadian bangsa. Cerita rakyat merupakan bagian integral dari sastra pada umumnya terutama sastra daerah yang ada dan berkembang di nusantara ini. Ia tumbuh dan berkembang di tengah-tengah masyarakat penikmatnya. Cerita rakyat umumnya dituturkan dan hanya tersimpan dalam ingatan orang-orang tua atau pencerita. Namun, sekarang ini tidak sedikit pula cerita rakyat itu ada yang sudah ditulis, bahkan diterbitkan. Cerita Rakyat Toraja, seperti halnya sastra daerah lain, juga berkembang secara lisan dari satu generasi ke generasi berikutrnya. Sastra daerah ini tetap terpelihara oleh masyarakatnya karena mereka yakin bahwa di dalamnya sarat dengan nilai-nilai luhur yang sangat bermanfaat bagi kehidupan umat manusia. Pengungkapan nilai-nilai budaya dalam cerita rakyat perlu dilakukan agar nilai-nilai luhur yang ada dalam suatu masyarakat dapat diketahui oleh masyarakat lain sehingga dapat meningkatkan saling pengertian antarmasyarakat. Selain itu, nilai-nilai budaya bagi generasi muda dapat menjadi filter terhadap unsur-unsur luar yang belum tentu menguntungkan. Mabasan – Vol. 3 No. 2 Juli—Desember 2009: 48--65 50 2. Pembahasan 2.1 Eksistensi Cerita Rakyat Cerita Rakyat Toraja bagi masyarakat Toraja tidak hanya untuk didengar dan diceritakan secara turun-temurun, tetapi juga cerita itu dianggap pernah terjadi pada masa lampau sehingga dapat memengaruhi tingkah laku mereka. Jadi, mereka taat kepada larangan/perintah yang berkaitan dengan cerita itu, misalnya mengapa dilarang bertengkar kalau sedang menumbuk padi, mengapa kayu cendana pantang dijadikan sebagai kayu bakar, mengapa dilarang menunjuk pelangi dengan jari telunjuk, mengapa kerbau putih pantang dimakan, dan sebagainya. Selain itu, cerita juga berkaitan erat dengan lingkungan alam sekitarnya. Sejumlah cerita menampilkan nama-nama tempat dan gunung, atau mengapa sesuatu diberi nama demikian. Benda-benda berupa batu dan benda pusaka seperti emas, kain selendang, cangkul, dan lain-lain dianggap sebagai bukti kebenaran cerita. Pada beberapa cerita, penutur meyakini bahwa apa yang diungkapkan dalam cerita itu benar-benar terjadi. Tokoh-tokoh cerita yang dipercayai kebenarannya dianggap sebagai leluhur yang menurunkan anak cucu dari generasi zaman lampau ke generasi berikutnya yang dengan penuh keajaiban menghadapi tantangan hidup. Menurut William R. Bascom (dalam Abdullah, 1999:3--4), cerita rakyat dapat dibagi ke dalam beberapa golongan besar, yaitu: a. Mite adalah cerita yang oleh masyarakat pemiliknya dipercaya sebagai sesuatu yang sungguh-sungguh terjadi pada zaman dahulu. Mite diajarkan untuk dipercaya karena dianggap memiliki kekuatan untuk menjawab ketidaktahuan, keragu-raguan, atau ketidakpercayaan dan sering dikaitkan dengan teknologi dan ritual. Mite merupakan perwujudan dogma dan biasanya dianggap suci. Nilai Budaya Dalam Cerita Rakyat Toraja (Ratnawati) 51 b. Legenda adalah cerita yang dianggap benar-benar terjadi, baik oleh pencerita maupun oleh pendengarnya. Waktu kejadiannya di zaman yang lebih muda. Legenda dapat bersifat sekuler atau suci dan tokoh utamanya adalah manusia. c. Dongeng dianggap sebagai cerita rekaan, tidak dianggap sebagai dogma atau sejarah, dan tidak dipermasalahkan kebenaran tentang kejadian peristiwanya, sering hanya untuk hiburan, tetapi memiliki fungsi penting seperti dikesankan oleh dongeng-dongeng yang mengandung nasihat. Dongeng tidak terikat tempat dan waktu, berkisah tentang kehebatan peri dan dewa, tentang petualangan manusia dan binatang. Berdasarkan pendapat tersebut, Cerita Rakyat Toraja digolongkan ke dalam tiga bentuk, yaitu mite, legenda, dan dongeng. Cerita Rakyat Toraja biasanya dituturkan secara lisan pada berbagai kesempatan, yaitu (1) pada waktu pelaksanaan pesta adat seperti pesta kematian, pesta pengucapan syukur, dan pesta perkawinan; (2) pada waktu ada pertemuan, misalnya pertemuan di balai desa, pertemuan rumpun keluarga, dan pada waktu mempersiapkan penyambutan tamu (pembesar) secara adat; (3) pada saat mengadakan kerja bakti massal atau pada saat dalam perjalanan sebagai perintang-rintang waktu; dan (4) pada waktu seseorang menanyakan asal-usul suatu benda, nama tempat, sejarah perjuangan daerah, hubungan Tana Toraja dengan kerajaan yang ada di Sulawesi Selatan pada zaman dahulu, dan sejarah perjanjian Tana Toraja dengan daerah lainnya. Mabasan – Vol. 3 No. 2 Juli—Desember 2009: 48--65 52 2.2 Nilai Budaya Nilai budaya adalah konsep abstrak mengenai masalah dasar yang sangat penting dan bernilai dalam kehidupan manusia (KBBI, 2008:963). Sejalan dengan definisi tersebut, Koentjaraningrat (1984:25) mengatakan bahwa nilai budaya itu adalah tingkat pertama kebudayaan ideal atau adat. Nilai budaya adalah lapisan paling abstrak dan luas ruang lingkupnya. Tingkat ini adalah ide-ide yang mengonsepsikan hal-hal yang paling bernilai dalam kehidupan masyarakat. Selain itu, sistem nilai terdiri atas konsepsi yang hidup dalam alam pikiran sebagian besar warga masyarakat mengenai hal-hal yang harus mereka anggap bernilai dalam kehidupan. Oleh karena itu, suatu sistem nilai budaya biasanya berfungsi sebagai pedoman tertinggi bagi perilaku manusia. Nilai budaya dalam cerita rakyat dapat diketahui berdasarkan tema dan amanat yang terkandung dalam cerita tersebut. Selain itu, nilai juga dapat diketahui dari para tokoh yang berperan di dalam cerita itu. 2.3 Nilai Budaya dalam Cerita Rakyat Toraja Nilai budaya dalam Cerita Rakyat Toraja dapat dibedakan ke dalam empat golongan besar, yaitu yang berkaitan dengan hubungan manusia dengan Sang Pencipta, hubungan manusia dengan alam, hubungan manusia dengan sesama manusia, dan hubungan manusia dengan dirinya sendiri. Berikut ini dipaparkan keempat hal tersebut. 2.3.1 Nilai Budaya yang Berkaitan dengan Hubungan Manusia dengan Sang Pencipta Dalam hubungan dengan Sang Pencipta, masyarakat Toraja bersumber pada kepercayaan Aluk Todolo. Berdasarkan ajaran Aluk Todolo, kesatuan alam ini dengan segala isinya mempunyai kewajiban Nilai Budaya Dalam Cerita Rakyat Toraja (Ratnawati) 53 mengabdi dan memuja Tuhan yang menciptakannya yaitu kepada tiga hal, yaitu: a. Puang Matua adalah sebagai Sang Pencipta semesta alam yang tertinggi dan yang paling utama. b. Deata-Deata adalah sebagai pemelihara seluruh ciptaan Puang Matua. Deata terbagi atas tiga golongan, yaitu Deata Tangngana Langi (Sang pemelihara di langit, yaitu dewa yang menguasai seluruh isi langit dan cakrawala. Yang kedua, Deata Kapadanganua (Sang pemelihara permukaan bumi) yaitu dewa atau deata yang menguasai seluruh apa yang terdapat di atas muka bumi. Yang ketiga, Deata Tangngana Padang (Sang pemelihara isi bagian tengah tanah) yaitu dewa yang menguasai isi tanah, laut dan sungai. c. Tomembali Puang atau biasa disebut tololo sebagai sang pengawas dan pemberi berkat kepada manusia turunannya. Hal tersebut ditampilkan dalam dua cerita seperti berikut ini. Iamoto ke memalaqi tu Aluk Todolo dao Toraya todiomai sae lako totemo belanna mangkamo napondok nenek todolon tu allo pura di bokoq. (Serreq Datu, 1986:184) Itulah sebabnya orang yang masih memeluk Aluk Todolo selalu melaksanakan kegiatan pemujaan kepada dewa yang memberi rezeki kepada umat manusia. Dao tanete todayang ke makaren-karen tontongbang tu Gonggang tikuqbiq-kuqbiq puduqna mukkun mangngando sia mssambayang langngan …. (Gonggang ri Sadoqkoq, 1986:119) Mabasan – Vol. 3 No. 2 Juli—Desember 2009: 48--65 54 Di bukit Todayang setiap sore gonggang dengan mulut komat-kamit mengucapkan mantra, memanjatkan doa ke hadapan Sang Pencipta …. Berdasarkan kepercayaan tersebut, ketiga hal itu harus disembah/dipuja dengan tingkat dan cara yang berbeda, yaitu: a. Pemala langngan Puang matua artinya pemujaan dan persembahan kepada sang pencipta sebagai upacara pemujaan dan persembahan yang paling tinggi yang harus dilakukan dengan kurban persembahan berupa kerbau, babi, dan ayam. b. Pemala lako deata artinya pemujaan dan persembahan kepada sang pemelihara sebagai upacara yang menengah harus dilakukan dengan kurban persembahan babi dan ayam c. Pemala lako Tomembali Puang/Todolo artinya pemujaan dan persembahan kepada Sang Pengawas Manusia turunannya sebagai upacara yang rendah yang harus dilakukan dengan kurban persembahan babi atau ayam. Secara eksplisit cara pemujaan ditampilkan sebagai berikut. Indeto nani umpogauq kapemalaran mangngando langnga Totumampana nasorongngi tu susinna: Bolu sitammu uraqna Kalosi ponno isinna Kapoq maqdua lallang Manuk sukku maelona (Gonggang ri sadoqkoq, 1986:119-120) Nilai Budaya Dalam Cerita Rakyat Toraja (Ratnawati) 55 Pada saat-saat tertentu di sana ia melaksanakan upacara pemujaan memajatkan doa ke hadapan Puang Pakombong dengan menyajikan antara lain sebagai berikut. Daun sirih yang bertemu uratnya Buah pinang yang bernas Kapur sirih yang putih bersih Ayam yang tak bercacat cela Kepercayaan Aluk Todolo juga mendasari pelaksanaan berbagai upacara yang dibedakan atas dua golongan besar, yaitu Rambu Tukaq berupa upacara keselamatan dan pengucapan syukur dan Rambu Soloq berupa upacara kematian/pemakaman. Berikut ini beberapa upacara Rambu Tukaq dalam Cerita Rakyat Toraja. a. Upacara kelahiran Anna maqaluk todolopa tu tomatuanna, inang lao lako toq bubun maqpiong. (Sadukung, 1986:197) Demikian pula dalam memperingati hari kelahiran anak, bagi penganut aluk todolo masih berlaku kebiasaan pergi ke pinggir sumur untuk mengadakan sesajen berupa lemang. b. Upacara syukuran karena ternaknya berkembang biak Inan dukato den pissan napogauqi Gonggang kapemalaran disanga “Massambeq Tedong” iamo tu maqkurre sumangaq belanna tontongbang membaqkaq tu tedongna. Lan kapemalaran Mabasan – Vol. 3 No. 2 Juli—Desember 2009: 48--65 56 iate billaq tu dipake unniraq dukuq manuk. (Gonggang ri Sadoqkoq, 1986:121) Di tempat ini suatu waktu Gonggang mengadakan upacara massambeq tedong (upacara syukuran karena kerbau sudah berkembang biak dengan menggunakan billaq (sembilu) untuk mengiris daging ayam yang dijadikan sesajen. c. Upacara pengakuan dosa Iatu penassananna Pasauq napopaqbutti iamo maqbuaq untunna tedong namangngaku langnga totumpana. (Saleq sola Pasauq, 1986:148) Sebagai bukti penyesalan itu, Pasauq mengadakan upacara maqbua dan di dalam pesta itu ia dengan jujur mengakui dosanya serta memohon ampun kepada Tuhan Mahakuasa. d. Berbagai macam pesta adat Belangna kasugiranna tau lan tondok iato, pembudami umpogauq kapemalaran susinna: maqtadoran, massuraq tallang, maqparekke para, murauk, laqpaq kasalle, metangdoq, sia pemalaran pengaqna. (Padang di Rura, 1986: 142) Karena kekayaan mereka itu berlebihan dan serba berkecukupan, akhirnya mereka sering mengadakan pesta adat secara besar-besaran seperti maqtadoran, massuraq tallang, maqparekke para, merauk, laqpaq kasalle, metangdoq, dan lain-lain. Nilai Budaya Dalam Cerita Rakyat Toraja (Ratnawati) 57 Jika dibandingkan dengan upacara rambu tukaq, sepertinya masyarakat Toraja lebih mengagungkan upacara rambu soloq. Berikut ini upacara Rambu Soloq dalam Cerita Rakyat Toraja. a. Upacara kematian Upacara ini diungkapkan dalam beberapa cerita, yaitu dalam cerita Bunga Alluq dan Dolitau, Sangbidang, Babuqsolong, Massudilalong dan Lebonna, Batu Tomate, serta Bokkoqbokkoq. Salah satu kutipan upacara kematian sebagai berikut. Gannaq pentallun ditambai tu Sokko namebali, diraqtag kumua ia tu Babuqsolong patalomo lako siuluqna. Belanna ia te siuluqna, tedong lana tunuanmo indoqna tu natanggaq anna natalomo Babuqsolong, maqkatampakanna Babuqsolong mannamo umpantunuanni tu indoqna tonna dialuk. Dadi ia tu siuluqna taeqmo tedong natunuan indoqna belanna puramo natalo Babuqsolong Setelah tiga kali berturut-turut nama Sokko mebali dipanggil dan selalu menyahut, diputuskanlah bahwa Babuqsolong menang dan kerbau sebanyak enam puluh empat ekor untuk pesta almarhum ibunya di lapangan upacara adalah miliknya semua. Jadi, kerbau saudara-saudaranya untuk persediaan pesta ibunya sudah habis semuanya. Akhirnya, hanyalah Babuqsolong sendiri yang memenuhi kewajiban sebagai anak yang mengabdi kepada kematian ibunya karena dia sendirilah yang dapat memotong kerbau dalam pesta itu. Mabasan – Vol. 3 No. 2 Juli—Desember 2009: 48--65 58 Dalam kepercayaan Aluk Todolo dikenal juga Pemali Sukaran Aluk (Larangan-Larangan dan pantangan dalam aturan-aturan agama) a. Larangan untuk bertengkar pada saat menumbuk padi Iamoto namepasan tu tomatua nakua taeq naden sigaga kelangkiqlu toq issong belanna taeq namembaqkaq tu boqboq. Napasan duka tomatua kumua matakuq tu dalleq tama banua kesigagabangngi tu tau lan banua. (Anak Yatim Piatu, 1986:149) Itulah sebabnya orang-orang tua selalu berpesan kepada anak-anaknya atau cucunya bahwa kalau sedang menumbuk padi jangan sekali-kali bertengkar karena beras itu tidak mau bertambah atau pun berkembang dan rezeki tidak mau datang. b. Pantangan mengucapkan kata-kata tertentu Berikut ini kutipan-kutipan dalam dua cerita, yaitu Polo Padang, Datu Lumuran yang mengisahkan tentang pantangan mengucapkan kata-kata tertentu. Mebalimi Polo Padang nakua, “Kamu poleq sola tallu tu sae umpepurai dalleku ke bongi. Dadi inang malolo tu Puang Matua, totemo lakupobaineko.” Nakuami tinde adak dara, masussakan sibali tolino iatu tolino biasa ia manglambe sia mekambullung. (Polo Padang, 1986:125—126) Polo padang mengatakan bahwa Tuhan Maha Adil, rupanya kamulah yang selalu menghabiskan tanamanku setiap malam. Sekarang aku ingin menikahimu. Putri menjawab, “Agaknya Nilai Budaya Dalam Cerita Rakyat Toraja (Ratnawati) 59 sukar bagi kami kawin dengan manusia karena manusia kadang-kadang mengucapkan kata-kata tabu yang pantang bagi kami. Tirambanni tu ambeqna diong sulluk natangmengkilala nasimpolo maqkada nakua, “wa, pida Pasuloan ungkatteneinaq.” …Ia te katangmekilalanna Batara Kassa male mengkaranduk lako Datu Lumuran .... (Datu Lumuran, 1986:106—107). Dengan tidak sadar, Batara Kassa tiba-tiba berteriak, “Wah, Pida, Pasuloan mengencingi saya.” …. Batara Kassa yang sudah terlanjur melanggar pantangan Datu Lumuran menyampaikan permohonan ampun, tetapi tidak diterima …. 2.3.2 Nilai budaya yang berkaitan dengan hubungan manusia dengan alam Berkaitan dengan alam, masyarakat Toraja mengenal tiga alam, yaitu alam atas, alam tengah, alam bawah. Hal tersebut terdapat dalam pembagian tugas deata (Dewa), yaitu Deata-Deata sebagai pemelihara seluruh ciptaan Puang matua. Deata terbagi atas tiga golongan, yaitu Deata tangngana Langi (Sang pemelihara di langit, yaitu dewa yang menguasai seluruh isi langit dan cakrawala. Yang kedua, Deata Kapadanganua (sang pemelihara permukaan bumi) yaitu dewa ata deata yang menguasai seluruh apa yang terdapat di atas muka bumi. Yang ketiga, Deata Tangngana Padang (Sang pemelihara isi bagian tengah tanah) yaitu dewa yang menguasai isi tanah, laut dan sungai. a. Dunia atas, yang dikuasai oleh Deata tangngana Langi dalam cerita dapat dikutip sebagai berikut. Mabasan – Vol. 3 No. 2 Juli—Desember 2009: 48--65 60 Umbai paqporainnamo Puang ungkombong langiq na lino, anna sirampean tolino tu anak tampakku, laqbiraka sipadadiammo batiq. Tanglatapomadiong pena, apa belanna torro tolino ia dadi lasule sola tallui rokko lino anna digenteq tomanurun di langiq, tosongloq di batara. (PPG 1986:128—129) Tuhan telah menjodohkan anakku yang bungsu dengan Polo Padang dari bumi sehingga tidak boleh diceraikan oleh siapa pun juga. Karena Polo padang dari manusia dari bumi maka mereka bertiga akan kembali ke bumi dan mereka inilah yang disebut tomanurung, artinya orang diturunkan ke bumi. b. Dunia bawah, yang dikuasai Deata Tangngana Padang dalam cerita dapat dikutip sebagai berikut. Tikaririk bongi buqtumi tu tau diong mai Tokengko umpeparei tu pare. Natiroi Marampio Padang taqpa narok doke bulawan nasampe tu bulawan lamkalena. Attu ia dukato larimi iatu tau rokko pagtana. (Marampio padang sola Datu Nakkaq, 1986:170) Pada malam harinya Marampio Padang ini pergi menunggui padinya. Tiada berapa lama menunggu, munculah orang melalui lubang dari bawah bumi (manusia dunia bawah) mencuri padinya itu. Apa iate Datu Lumuran mebali nakua, “Apa tu mupokadanna apa taeqbang nalamaqdin dadi, belanna tantu muissan sia lamutandai kumua; iate tau akute ludiongnaqpi toq mata wai naya tu iko daoko menggantananna.” Mebali Batara Kassaq, Nilai Budaya Dalam Cerita Rakyat Toraja (Ratnawati) 61 “mentuqna tu alasanmu taeqnasang kuporai sangngadinna inang lakupobaineko.” (DLM, 1986: 106) Datu lumuran menjawab, “apa yang kamu harapkan mustahil terjadi. Dunia kita berbeda, saya hidup di dalam air, sedangkan kamu hidup di darat.” Batara Kassa menjawab, “Saya tidak peduli semua itu. Yang penting engkau harus menjadi istriku.” 2.3.3 Nilai budaya yang berkaitan dengan hubungan manusia dengan manusia lain Nilai budaya yang berkaitan dengan hubungan antara satu pihak dengan pihak lain dalam cerita, yaitu setia pada janji, tolong-menolong, menghormati tamu. a. Setia pada janji Berikut ini tanda kesetiaan pada janji yang diungkapkan oleh seorang pria dalam bentuk ucapan bahkan dibuktikan dengan cara bunuh diri. Kumua, pura bassemoq Lebonna, mangka sibole-bole Pada tuo, pada mate Pada sangrontaq inaya, pada ditambuttanai Naia tu Lebonna, masai allomo manteqna Rontaq rondon tobatangna Lamentiromoko tau, mengkita sanda mairiq Lamanteq todamoq aku, lasangrontaq inaamoq (Massudilalong sola Lebonna, 1985:140) Bahwa aku telah berjanji dengan Lebonna Sehidup semati dan sepenanggungan Satu lubang kami berdua Mabasan – Vol. 3 No. 2 Juli—Desember 2009: 48--65 62 Lebonna telah pergi Dia sudah mendahului Sekarang kamu menyaksikan, yang hadir jangan kaget Sekarang aku berangkat, menyusul dia yang lebih dahulu b. Tolong-menolong Tolong-menolong merupakan sikap terpuji dan patut dicontoh dalam menerapkan suatu konsep kehidupan yang damai sejahtera. Makhluk sebagai ciptaan hadir dengan segala kelebihan dan kekurangan masing-masing. Oleh karena itu, sepatutnyalah dalam kehidupan ini kita menggunakan kelebihan yang dimiliki untuk membantu pihak lain. Pada gilirannya nanti, kelemahan kita akan ditutupi oleh kelebihan pihak lain. Nilai tolong-menolong dalam cerita dapat diketahui lewat kutipan berikut. Naupuiqmi donga nakua, “Lasiangkarangkiq kedenkiq nerampoi sussa, susinnato kenalambinduqkanaq paqdiq baqtu kedenkiq nerampoi sussa (tattiuq sola donga, 1986:109) Selanjutnyam rusa itu mengungkapkan bahwa apabila ada di antara mereka yang kena musibah atau kesusahan, maka kedua pihak harus saling menolong dan merasakannya bersama-sama. c. Menghormati tamu Menghormati tamu dalam cerita dapat diketahui berdasarkan kutipan berikut ini. ”Apara kamu mipogauq nasusira te tu kasugiranmi.” Nakuami tinde pia biung, “Memalaq-malaq bangkan belanna den tau sae inde banuangki, iamo kipemalaqi.” (Serreg Datu, 2986:284) Nilai Budaya Dalam Cerita Rakyat Toraja (Ratnawati) 63 “Apa gerangan yang kamu perbuat sehingga kamu menjadi kaya?” Jawab anak yatim itu, ”Kami haya menghormati dan melayani dengan sungguh-sungguh kalau ada orang yang datang di rumah kami.” 2.3.4 Nilai budaya yang berkaitan dengan hubungan manusia dengan dirinya sendiri Nilai budaya yang berkaitan dengan hubungan manusia dengan dirinya yaitu kerja keras dan sabar. a. kerja keras Mepasanmi tu ambeqna nakua, “Pemeloibangmi tu tengkamu lako lalan den oupaq musalamaq rampo lako tu lamuninna undakaq kande. Manarangkiq ussaro mase lako tau. Mebalikiq keditambaikiq, sia dituruq sola didama kenasuakiq tau. Namentuqnato rakkaq sangpulota umpatuokiq diolu padangna tau sia ditiro melo kementengkakiq den oupaq naden salamaq. (Sadoqdongna, 1986:162) Berkatalah ayahnya, “Baik-baiklah dalam perjalanan, semoga engkau selamat tiba di tempat tujuan. Pandai-pandailah engkau menarik hati orang. Menyahutlah bila engkau dipanggil dan rajinlah mengerjakan apa yang diperintahkan. Buruk baik nasib yang engkau alami terletak pada kedua belah tanganmu. Ikutlah perbuatan yang baik sehingga kamu mendapat keberuntungan dan keselamatan.” Mabasan – Vol. 3 No. 2 Juli—Desember 2009: 48--65 64 b. sabar Moi susito, ia tu Matadua saqbaraqbang dikkag. Ia kedisengkei kappabang dikkaq. (Tallu Tosiunuq, 1986:179) Namun demikian, si Mata Dua tetap sabar bahkan kalau dimarahi oleh orang tuanya, ia tetap tenang dan diam saja. 3. Penutup Cerita Rakyat Toraja bagi masyarakat Toraja tidak hanya untuk didengar dan diceritakan secara turun-temurun, tetapi juga cerita itu dianggap pernah terjadi pada masa lampau sehingga dapat memengaruhi tingkah laku mereka. Sesuai dengan analisis yang dilakukan terhadap Cerita Rakyat Toraja disimpulkan nilai budaya dalam Cerita Rakyat Toraja yang berkaitan dengan hubungan manusia dengan sang Pencipta dan hubungan manusia dengan alam sebagian besar didasarkan pada kepercayaan Aluk Todolo. Berikut ini cakupan nilai budaya yang berkaitan dengan empat hubungan, yaitu: a. Hubungan manusia dengan Sang Pencipta berupa kewajiban memuja dan menyembah, tata cara pemujaan, upacara keselamatan dan kematian, serta larangan/pantangan yang harus dijauhi; b. Hubungan manusia dengan alam berupa alam atas, alam tengah, dan alam bawah; c. Hubungan manusia dengan sesama manusia berupa kesetiaan terhadap janji, tolong menolong, dan menghormati tamu; d. Hubungan manusia dengan dirinya sendiri berupa kerja keras dan sabar. Nilai Budaya Dalam Cerita Rakyat Toraja (Ratnawati) 65 Daftar Pustaka Koentjaraningrat. (1984). Kebudayaan, Mentalitas, dan Pembangunan. Jakarta: PT Gramedia L.T. Tangdilintin. (1974). Toraja dan Kebudayaannya. Ujung Pandang: Kantor Cabang II Lembaga Sejarah dan Antropologi -----------------------. (1978). Sejarah dan Pola-Pola Hidup Toraja. Tana Toraja: Yayasan Lepongan Bulan Ratna, Nyoman Kutha. (2005). Sastra dan Cultural Studies. Yogyakarta: Pustaka Pelajar Sikki, Muhammad et al. (1986). Struktur Cerita Rakyat Toraja. Jakarta: Pusat Pembinaan dan Pengembangan Bahasa. Tim Penyusun Kamus Pusat Bahasa. (2008). Kamus Besar Bahasa Indonesia. Jakarta: PT Gramedia Pustaka Utama | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/332106338Protrusion of the Androconial Organs of Lycorea haliaatergatis Doubleday [1847] (Nymphalidae) from a CloudForest at Veracruz State, Mé....Article in Southwestern Entomologist · March 2019DOI: 10.3958/059.044.0136CITATION1READS1385 authors, including:Some of the authors of this publication are also working on these related projects:Catálogo de tipos y ejemplares de la familia Castniidae conservados en el Museo Argentino de Ciencias Naturales ViewprojectSystematics, Biology and Biogeography of Neotropical Lepidoptera (Papilionoidea & Hesperioidea) View projectFernando Hernández-BazUniversidad Veracruzana55 PUBLICATIONS 204 CITATIONS SEE PROFILEJulio César Castañeda OrtegaUniversidad Veracruzana24 PUBLICATIONS 4 CITATIONS SEE PROFILELiliana Lara-CapistránUniversidad Veracruzana52 PUBLICATIONS 176 CITATIONS SEE PROFILEJorge M. GonzalezAustin Achieve Public Schools181 PUBLICATIONS 1,352 CITATIONS SEE PROFILEAll content following this page was uploaded by Jorge M. Gonzalez on 31 March 2019.The user has requested enhancement of the downloaded file.327 VOL. 44, NO. 1 SOUTHWESTERN ENTOMOLOGIST MAR. 2019 SCIENTIFIC NOTE Protrusion of the Androconial Organs of Lycorea halia atergatis Doubleday [1847] (Nymphalidae) from a Cloud Forest at Veracruz State, México Fernando Hernández-Baz1*, Julio César Castañeda-Ortega1, Liliana Lara-Capistrán2, María Del Carmen Núñez-Camargo2, and Jorge M. González3 Abstract. The protrusion of androconial organs of Lycorea halia atergatis Doubleday [1847] is reported herein. Androconial setae are inserted in the intersegmental membrane between the 8th tergite and the male genitalia. The butterfly species is typically found in cloud forests along the central mountains of Veracruz, Mexico. Resumen. Se reporta la protrusión de los órganos androconiales en forma de penachos extendidos que se insertan medialmente en la membrana intersegmental entre el octavo tergito y los genitales masculinos en la especie Lycorea halia atergatis Doubleday [1847], típica de bosques nublados de las montañas centrales de Veracruz, México. Introduction Although insects are diverse, they are similar in overall design (Evans 1984). All insects depend on sensorial systems that allow them to identify food sources and appropriate mates, and this seems particularly important in taxa whose main activities are done at night (Renou 2014). To succeed at this, many insects rely on chemoreception that provides information that mediates ecological interactions of several kinds (González et al. 2018). Thus, many insects develop a complex intra- and inter-specific communication system involving olfactory cues (Hernández-Roldán et al. 2014, González et al. 2018). Courtship and mating are complex and variable in many insects, and particularly in butterflies (Rutowski 1991, Estrada et al. 2011). But, although many females play an important role in courtship and mating behavior, Lepidoptera males possess complex structures connected to glands or glandular cells, as well as non-glandular hairs, and are relevant during courtship (Varley 1962, Birch and Hefetz 1987, Boppré and Vane-Wright 1989, Scoble 1992). The structures have always intrigued researchers who have tried to elucidate their function (Boppre and Vane-Wright 1989, Zilli 1997, Hernández-Roldán et al. 2014). They also have a diverse array of forms, from scattered hairs and scales to complex organs associated with ________________________ 1Facultad de Biología-Xalapa, Universidad Veracruzana CP 91000 Circuito Gonzalo Aguirre s/n. Zona Universitaria, Xalapa, Veracruz, México, E-mail: ferhbmx@yahoo.com.mx; fhernandez@uv.mx 2Doctoral Student in Environmental Management, Universidad Popular Autónoma de Veracruz and Facultad de Ciencias Agrícolas, Universidad Veracruzana, CP 91000 Circuito Gonzalo Aguirre s/n. Zona Universitaria Xalapa, Veracruz, México. 3Department of Plant Sciences, California State University, Fresno, CA 93740, (Research Associate, McGuire Center for Lepidoptera and Biodiversity), USA, E-mail: gonzalez.jorge.m@gmail.com *Corresponding author e-mail: ferhbmx@yahoo.com.mx; fhernandez@uv.mx 328 wings and/or different parts of the body, and especially the abdomen (Boppre and Vane-Wright 1989, Speidel and Naumann 1995, Speidel et al. 1996). They also might involve modifications of the exoskeleton (Zilli and Di Giulio 1996). Morphologically, they also are very diverse and vary from simple hair tufts to complex eversible organs with elaborate methods of secretion, gathering, and releasing of pheromones (Boppre and Vane-Wright 1989, Hall and Harvey 2002). Males of many lepidopterans possess cutaneous epidermal structures that frequently are connected to glandular cells. The structures are considered modified non-glandular scales and known as androconial organs, sometimes called coremata, and are considered secondary sexual structures in moths (Davenport and Conner 2003). The main function of the organs that could be anywhere on the body of the male is to produce and disseminate pheromones that influence sexual behavior of females. Once released by females, sex pheromones are perceived by sensilla in the antennae males use to locate females. Once the female is detected, the male gets closer to the female by walking and fluttering his wings and eventually starts courtship. In front of the female, the male extrudes his androconial organs and places them repeatedly toward the female, appeasing her to allow copulation (Zagatti 1989). There are 2,940 species of Lepidoptera recorded in the State of Veracruz, Mexico, but there might be as many as 6,500 species (Hernández-Baz and Iglesias-Andrew 2001). Of those, 729 Papilionoidea have been reported, with 328 in the family Nymphalidae. In this family, the genus Lycorea Doubleday [1847] is represented in the state by two species: Lycorea ilione albescens (Distant 1876) and Lycorea halia atergatis Doubleday, [1847] (Hernández-Baz et al. 2010, Luis-Martínez et al. 2011). On 20 February 2017, we visited the Teocelo Gorge (Barranca de Teocelo, Puente Teocelo) (19°24´10.15¨N / 96°59´12.07¨) 1,160 m above sea level at the Municipality of Teocelo, Veracruz, Mexico. Few relic areas of mountainous cloud forest remain there and are mixed with perennial lower tropical forest and coffee (Coffea arabica L., Rubiaceae) and banana (Musa sp.; Musaceae) plantations. Two male Lycorea halia atergatis Doubleday [1847] (Nymphalidae: Danainae) were collected and brought to our laboratory for identification (Hernández-Baz et al. 2010, Luis-Martínez et al. 2011). Both specimens were flying over an Aloysia virgata (Ruiz y Pav.) (Verbenaceae) bush, an exotic plant used sometimes as an ornamental that has flowers with abundant nectar with attractive smell that attracts many visitors. When put into an envelope, one L. halia atergatis protruded a pair of organs in the shape of hair pencils that immediately opened as a plume of modified scales or androconial setae. The plumes were inserted into the intersegmental membrane between the 8th tergite and male genitalia. The plumes, each containing about 144 long scales were black and had a diameter of about 17 mm when completely protruded (Figs. 1a-c). Starting with this note, we intend to provide a detailed analysis of types and shapes of androconial organs of Mexican lepidopterans in mountainous cloud forests. Acknowledgment This work is part of the project “The Lepidoptera of Veracruz, México: an analysis of its richness, ecology, and conservation” sponsored by PRODEP-SEP-MEXICO and Universidad Veracruzana and developed by the Cuerpo Académico Entomología y Parasitología, code: UV-CA-421. Insects were collected under 329 scientific license # FAUT-0194 provided by Secretaría de Medio Ambiente Recursos Naturales, México. Our gratitude to Bonnie Pendleton and the anonymous reviewers for their comments and corrections that helped improve the final manuscript. Fig. 1. Androconial organs of Lycorea halia atergatis (Nymphalidae: Danainae): silvery ejectable membrane under male valvae supporting the androconial organs (1a), protrusion of black androconial setae resembling a paintbrush (1b), and androconial setae extended and forming a plume (1c). References Cited Birch, M. C., and A. Hefetz. 1987. Extrusible organs in male moths and their role in courtship behavior. Bull. Entomol. Soc. Am. 33: 222-229. Boppré, M., and R. I. Vane-Wright. 1989. Androconial systems in Danainae (Lepidoptera): functional morphology of Amauris, Danaus, Tirumala, and Euploea. Zool, J. Linnean Soc. 97: 101-133. Davenport, J. W., and W. E. Conner. 2003. Dietary alkaloids and the development of androconial organs in Estigmene acrea. J. Insect Sci. 3(3). 330 Estrada, C., S. Schulz, S. Yildizhan, and L. E. Gilbert. 2011. Sexual selection drives the evolution of anaphrodisiac pheromones in butterflies. Evolution 65: 2843-2854. Evans, H. E. 1984. Insect Biology: A Textbook of Entomology. Addison-Wesley Pub. Co., USA. González, J. M., D. Camino, S. Simon, and A. Cusumano. 2018. Semiochemical exploitation of host-associated cues by seven Melittobia parasitoid species: behavioral and phylogenetic implications. Front. Ecol. Evol. 5: 172. Hall, J., and D. Harvey. 2002. A survey of androconial organs in the Riodinidae (Lepidoptera). Zool. J. Linnean Soc. 136: 171-197. Hernández-Baz, F., and L. Iglesias-Andrew. 2001. La diversidad de orden Lepidoptera en el estado de Veracruz, México: una síntesis preliminar. Cuadernos de Biodiversidad, Universidad de Alicante, España 7: 7-10. Hernández-Baz, F., J. Llorente B., A. Luis M., and I. Vargas F. 2010. Las mariposas de Veracruz. Secretaria de Educación de Veracruz, Consejo Veracruzano de Ciencia y Tecnología, Gobierno del estado de Veracruz, Universidad Veracruzana. Hernandéz-Roldán, J., R. Bofil, L. Dapporto, M. Munguira, and R. Vila. 2014. Morphological and Chemical analysis of male scant organs in the butterfly genus Pyrgus (Lepidoptera: Hesperidae). Organisms Diversity and Evolution. DOI 10.1007/s13127 014 0170x Luis-Martínez, M. A., J. E. Llorente-Bousquets, I. Vargas-Fernández, and F. Hernández-Baz. 2011. Mariposas diurnas Papilionoidea y Hesperioidea (Insecta: Lepidoptera), pp. 339-354 + apéndices. En A. Cruz Aragón [Coord.], La Biodiversidad de Veracruz, Vol. II. Comisión Nacional para el Conocimiento y Uso de la Biodiversidad, Gobierno del Estado de Veracruz. Renou, M. 2014. Pheromones and general odor perception in insects. In C. Mucignat-Caretta [ed.], Neurobiology of Chemical Communication. CRC Press/Taylor and Francis, Boca Raton, FL. https://www.ncbi.nlm.nih.gov/ books/NBK200986/?report=reader#_NBK200986_pubdet_ Rutowski, R. L. 1991. The evolution of male mate-locating behavior in butterflies. Am. Nat. 138: 1121-1139. Scoble, M. J. 1992. The Lepidoptera: Form, Function, and Diversity. Natural History Museum and Oxford University Press, Oxford. Speidel, W., and M. C. Naumann. 1995. Further morphological characters for a phylogenetic classification of the Noctuidae (Lepidoptera). Beitr. Entomol. 45: 119-135. Speidel, W., H. Fänger, and M. C. Naumann. 1996. The phylogenetic of the Noctuidae (Lepidoptera). Syst. Entomol. 21: 219-251. Varley, J. C. 1962. A plea for a new look at Lepidoptera with special reference to the scent distributing organs of male moths. Trans Soc. Br. Entomol. 15: 29-40. Zagatti, P. 1989. Feromonas sexuales de Lepidoptera: biología, química y aplicaciones. Revista Palmas 10: 47-58. Zilli, A. 1997. The unusual male brush apparatus of Hypopteridia (Lepidoptera: Noctuidae). Eur. J. Entomol. 94: 503-510. Zilli, A., and Di. A. Diulio. 1996. Diversitá degli organi androconiali nei Mythimnini italiani; morfología comparata ed implicazioni filogenetiche (Lepidoptera, Noctuidae). Fragm. Entomol. 28: 97-147. View publication stats | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/330664359Synthesis of heterocycles via aerobic oxidationArticle in ARKIVOC: archive for organic chemistry · March 2019DOI: 10.24820/ark.5550190.p010.843CITATIONS0READS2251 author:Some of the authors of this publication are also working on these related projects:C-H functionalization View projectCopper-catalyzed cross couplings View projectCarlo SambiagioJanssen Pharmaceutica19 PUBLICATIONS 2,248 CITATIONS SEE PROFILEAll content following this page was uploaded by Carlo Sambiagio on 07 March 2019.The user has requested enhancement of the downloaded file.DOI: https://doi.org/10.24820/ark.5550190.p010.843 Page 1 ©ARKAT USA, Inc The Free Internet Journal for Organic Chemistry Review Archive for Organic Chemistry Arkivoc 2018, part vi, 0-0 to be inserted by editorial office Synthesis of heterocycles via aerobic oxidation Carlo Sambiagio Department of Chemical Engineering and Chemistry, Micro Flow Chemistry & Synthetic Methodology, Eindhoven University of Technology, Den Dolech 2, 5612 AZ Eindhoven, The Netherlands Email: C.Sambiagio@tue.nl Received 12-11-2018 Accepted 01-26-2019 Published on line 03-03-2019 Abstract Research in the field of aerobic oxidation has recently gained much attention from both academia and industry. The main reason for this has to be found in the inherent greenness of such methods. Although much investigation has been undertaken to develop new methods and reactions and to elucidate the mechanistic aspects of this chemistry, its applications in the synthesis of heterocycles have been relatively limited. In this review, the recent developments in this field are summarized. Keywords: Aerobic oxidation, heterocyclic synthesis, reaction mechanism, oxidative couplings, oxygen Arkivoc 2018, vi, 0-0 Sambiagio, C. Page 2 ©ARKAT USA, Inc Table of Contents 1. Introduction 2. Csp3-H bond Oxygenation 2.1 Oxygenation / condensation with amines 2.2 Oxygenation / C-O bond formation 3. Oxidative Isocyanide Insertion 3.1 Formation of C-C + C-heteroatom bonds 3.2 Formation of two C-heteroatom bonds 3.3 Isocyanide insertion / acyl migration 4. Oxidative CO Insertion 5. Oxidative Friedlander annulation – Acridine Synthesis 6. Oxidative N-Heteroatom Bond Formation 6.1 Oxidative N-N bond formation 6.1.1 Synthesis of pyrazoles and indazoles 6.1.2 Synthesis of triazoles 6.2 Oxidative N-O, N-S, and N-Se bond formation 7. Oxidative C-Heteroatom Bond Formation 7.1 Oxidative C-N bond formation 7.2 Oxidative C-O and C-S bond formation 8. Intermolecular Oxidative Cyclization with Unsaturated Moieties 8.1 Synthesis of imidazole derivatives via cyclization of amidines with electron-poor alkenes 8.2 Synthesis of benzofurans via cyclization of phenols with alkynes 8.3 Miscellaneous cyclizations with alkynes 9. Conclusions 1. Introduction Aerobic oxidation chemistry (i.e. the use of O2 as terminal oxidant in a chemical reaction) is a research field for which a steadily growing interest has been observed during the last 2 decades (Figure 1, left).1-2 The increasing interest in green and sustainable chemistry from both academia and industry surely has an important role in this. As oxidations (in a general sense) are widespread reactions, the possibility to use an abundant, cheap, benign, and atom economic reagent for these transformations is a considerable motivation for all researchers in this field. Due to its clear advantages, O2 has long been employed in several industrial processes, mostly for the production of bulk chemicals, such as cumene hydroperoxide (for the production of phenol and acetone) and cyclohexanone (production of polyamides).3 For these compounds, produced in very large scale and with a low market price, the use of oxygen is the only economically viable option. However, the employment of O2 in chemical processes also has several drawbacks, which somewhat limit its applications. The inefficient mass-transfer of the gas into solution calls for high pressures to be used to speed up the reaction. On the other hand, a high O2 pressure in the presence of a flammable solvent at high temperatures represents an important safety hazard. Many aerobic industrial processes for the synthesis of bulk chemicals, therefore, utilize synthetic air (5-10% O2 in N2) instead of pure O2. The low reaction rate achieved in this way (due to the low Arkivoc 2018, vi, 0-0 Sambiagio, C. Page 3 ©ARKAT USA, Inc concentration of oxygen) is generally compensated by the implementation of recycling streams at the end of the reaction, overall achieving a good compromise between reaction rate/conversion, safety, and cost. While this is generally a good solution for the production of bulk chemicals, which can count on dedicated plant reactors, 2-4 this is not the case for the fine chemical and pharmaceutical industry, where the smaller scale does not justify a dedicated plant, and the higher price of the final products also allows the use of more expensive, but easier to handle and process-safer oxidants.4 Nonetheless, as most other oxidant are also characterized by a low atom economy, high content in metals, and a certain toxicity, now research is being carried out on the development of alternative methods to utilize O2 in a more controlled and safer manner, so to allow its use even in the fine chemicals and pharma industry. For example, continuous-flow chemistry in microreactors is now being often employed to improve mass transfer between gas and liquid, and reduce ignition hazard, due to the low concentration of oxygen in every segment of the flow reactor, and the absence of a headspace filled with pressurized gas.4-6 These improvements will on the long run make aerobic oxidation methods more attractive and ultimately broaden their applications. Aerobic oxidations can be broadly divided into oxygenase- and oxidase-type reactions, which take their name from the corresponding enzymes that catalyze these processes in biological systems. Simply put, in oxygenase chemistry the O2 acts as a reactant, and oxygen atoms are incorporated into the molecule (either an intermediate or the final product). For example, on C-H bonds, oxygenase results in the formation of alcohols, ketones/aldehydes, or carboxylic acids. In oxidase chemistry, the oxygen only acts as an electron acceptor, and no oxygen atom is incorporated into the product. For example, oxidase processes occur when a metal catalyst, at the end of a catalytic cycle in an inactive state (low oxidation state), is re-oxidized to its active form (higher oxidation state) by donating electrons to O2. This is typically the case in, for example, cross-dehydrogenative couplings (CDCs) between two R-H moieties to form R-R bonds. Between these two reactivity modes, aerobic oxidations can in theory be applied to the formation of virtually any type of C-C, C-heteroatom, or heteroatom-heteroatom bonds. This versatility is a great advantage in synthesis, and aerobic oxidation methods have been applied to a large number of reactions in the last decades, in particular alcohol dehydrogenation, oxygenation of methyl/methylene groups to carbonyl derivatives, and inter- and intramolecular CDCs. Among its applications, the synthesis of heterocycles is particularly attractive, because of the variety of bonds that can be formed via aerobic oxidation strategies within the same heterocyclic core. Due to the importance of heterocycles in organic, inorganic, material, and medicinal chemistry, the development of novel retrosynthetic pathways for their synthesis, or the improvement of existing pathways into greener alternatives, is of high interest. Although this specific sub-area of aerobic oxidation has not been extensively explored to date, an increasing number of publications in the field has been observed from the first decade of the 2000’s, with peaks in the period 2012-2014 (Figure 1, right). This trend reflects an increasing interest from the chemistry community, and further exploration of the field is expected in the near future. This review summarizes the recent advancements (since approximately 2010) on the use of O2 as oxidant in heterocyclic synthesis. In this review the literature will not be divided according to the heterocyclic cores formed, as many strategies can be applied to the synthesis of a variety of heterocycles. Instead, the main different strategies will be discussed separately, showing all the possibilities offered in synthesis by each specific approach. Arkivoc 2018, vi, 0-0 Sambiagio, C. Page 4 ©ARKAT USA, Inc Figure 1. Left: publications containing the words “aerobic oxidation” in the period 2000-2018, SciFinder search 7th Jan. 2019. Right: publications containing the concepts “heterocycle synthesis” and “aerobic oxidation” associated in the period 2000-2018, SciFinder search 7th Jan. 2019. 2. Csp3-H bond Oxygenation While many of the aerobic processes for the synthesis of heterocycles involve oxidase-type mechanisms, one common exception is the oxygenation of Csp3-H bonds as a crucial step in the formation of the heterocycle. Two strategies can be in particular distinguished: the oxidation of C-H bonds to carbonyls, followed by in-situ condensation with an amine (new C-N bond in the heterocycle); and the oxygenation of C-H bonds, followed by radical coupling of the newly installed oxygen with activated carbon moieties, leading to new C-O bonds. 2.1 Oxygenation / condensation with amines The condensation of a carbonyl and an amine to give an imine is a common reaction encountered in the synthesis of aza-heterocycles. An interesting approach is the in-situ formation of the carbonyl reactant from the corresponding precursors. The formation of ketones or aldehydes, or analogous intermediates, can be directly achieved via Cu-/Fe-catalyzed aerobic oxidation of alcohols (oxidase) or direct oxygenation of activated (e.g. benzylic) methylenes (oxygenase), according to Scheme 1. Another possible method is the iodine-catalyzed Kornblum oxidation, which provides access to carbonyls from the corresponding methylenes in an indirect way. In this case the oxygen atom in the carbonyl comes from DMSO. Being an indirect oxidation, heterocyclic synthesis involving this process will not be treated here.7-8 The carbonyl obtained under aerobic conditions can then be transformed in-situ into an imine. In the synthesis of heterocycles, this step generally does not constitute the ring-closing step, so that this process is intermolecular, and is typically followed by an intramolecular nucleophilic attack to the imine carbon or a cyclization at the methylene alpha to the nitrogen, furnishing the desired aza-heterocycle (Scheme 1). Arkivoc 2018, vi, 0-0 Sambiagio, C. Page 5 ©ARKAT USA, Inc Scheme 1. Possible pathways for the synthesis of aza-heterocyclic cores via Csp3-H oxygenation followed by condensation with amines. The synthesis of quinazolinones and other diheterocycles (2-5) has been reported by Zhou/Yin making use of 2-methylaza-heterocycles9 or phenylacetic acids10 as aldehyde precursors, according to Scheme 2A-B. While the first method is more atom economic, it has the limitation of only working on 2-methylaza-arenes as aldehyde precursors, due to its mechanism (see Scheme 2, bottom).9, 11 The second method required prefunctionalized, though easily accessible, precursors, but a larger variety of aryl and heteroaryl substituents could be employed.10 A variety of other oxidants tested with this methods resulted in no reaction. The use of 2-methylaza-arenes was also reported by Yan for the synthesis of functionalized quinolines 8 and pyrroloquinoxalines 9 ( Scheme 2C).12 Very notably, it was shown that apart from 2-methylpyridines and 2-methylquinolines, also 4-methylpyridine was effective in the reaction, although 4-methylquinoline was not. In this protocol, after the formation of the imine moiety, an electrocyclization reaction takes place, leading to ring closure. Aromaticity is then re-established by oxidation. While this method is limited to activated picolines as reactants, the introduction of an alkyl chain in the pyrroloquinoxaline scaffold is possible using aliphatic alcohols as reactants. These are oxidized to the corresponding aliphatic aldehyde, which undergo cyclisation giving the desired scaffold 10. The aerobic oxidation of N-aryl pyrroles 7 with alcohols has been reported using superstoichiometric Fe,13 but noteworthy is the metal-free version reported by Jiang in 2017 ( Scheme 2D).14 The mechanism of the Cu-catalyzed oxidation of 2-methylpyridines has been studied in details by Maes and co-workers,11 and is at the base of the lack of reactivity of 3-methylpyridines observed by Zhou/Yin.9 The reaction is initiated by an imine-enamine tautomerism, which cannot occur in the 3-methyl isomer. Reaction of the enamine tautomer with Cu/O2 results in the formation of the aldehyde intermediate, or a peroxide intermediate which directly reacts with the amine.9 The mechanism of the decarboxylative protocol is suggested to occur via aerobic oxidation of the methylene alpha to the carboxyl moiety, followed by decarboxylation, leading to the corresponding benzaldehydes ( Scheme 2, bottom).10 Arkivoc 2018, vi, 0-0 Sambiagio, C. Page 6 ©ARKAT USA, Inc Scheme 2. Synthesis of aza-heterocycles via Csp3-H oxygenation / condensation starting from picolines or phenylacetic acids. When benzylic amines are used, a different mechanism becomes available. In such cases, the benzylic carbon of the amine can be involved in an intramolecular oxidative cyclization, furnishing the heterocycle. The synthesis of imidazole-based compounds 11-13 was reported via this mechanism by Zhang and Ji/Wang ( Scheme 3).15-16 Arkivoc 2018, vi, 0-0 Sambiagio, C. Page 7 ©ARKAT USA, Inc Scheme 3. Synthesis of aza-heterocycles via Csp3-H oxygenation / condensation involving benzylamines as reactants. 2.2 Oxygenation / C-O bond formation Although the formation of imines in-situ from carbonyls is a particularly recurrent strategy for the synthesis of heterocycles under aerobic conditions, oxygenase methods can also be applied to the synthesis of O-containing heterocycles, where the aerobic oxygen remains in the final product. Chiba and co-workers were able to synthesize a range of dihydrooxazoles 16 from N-alkylamidines 14 via a Cu-catalyzed process ( Scheme 4A).17 The reaction occurred via oxidation of the amidine moiety, generating a nitrogen radical, followed by a 1,5-hydrogen radical shift, leading to a tertiary carbon radical. Cu-catalyzed oxygenation in this position and subsequent cyclization resulted in the dihydrooxazoles 16. 1,3-Benzooxazines 17 could be prepared using the same strategy from N-(2-isopropylphenyl)amidines 15 ( Scheme 4A).17 In this case a 1,6-hydrogen radical shift was postulated for the formation of the intermediate carbon radical. Related strategies were developed by Jiao et al. for the synthesis of di- or trisubstituted oxazoles 18-19.18-19 Both protocols involve the formation of imine intermediates, followed by oxygenation at the alpha position and final cyclization-dehydrogenation to furnish the oxazole core ( Scheme 4B-C). Arkivoc 2018, vi, 0-0 Sambiagio, C. Page 8 ©ARKAT USA, Inc Scheme 4. Synthesis of aza-heterocycles via Csp3-H oxygenation / C-O bond formation. 3. Oxidative Isocyanide Insertion Isocyanides have recently replaced the isoelectronic carbon monoxide as a one-carbon synthon in a variety of protocols, including for the synthesis of heterocycles,20-21 due to their lower toxicity and their easier handling. Similarly to the oxidative insertion of CO (see Section 4), an important strategy for the synthesis of a variety of heterocycles is the oxidative insertion of isocyanides into a metallacycle (usually Pd), followed by reductive elimination. Only the terminal carbon of the isocyanide is thus incorporated into the ring, and an amino-substituted heterocycle is obtained. In this case the oxidant is necessary to restore the active Pd(II) species from Pd(0) after the reductive elimination. Non-oxidative versions of this reaction were reported before 2010, making use of organic halides as substrates and not requiring any oxidant ( Scheme 5, left). A number of different heterocyclic structures can be easily obtained using this method, by formation of either a C-C and a C-heteroatom bond, or two C-heteroatom bonds ( Scheme 5, right). These protocols are discussed in Sections 3.1 and 3.2. Arkivoc 2018, vi, 0-0 Sambiagio, C. Page 9 ©ARKAT USA, Inc A different reactivity pathway has also been recently discovered. Here an initial isocyanide insertion is followed by an acyl migration, resulting in both the C and N atoms of the isocyanide being incorporated into the ring. This reactivity has very different implications and mechanism, and is treated in Section 3.3. Scheme 5. Left: oxidative vs non-oxidative isocyanide insertion. Right: examples of synthetic applications. 3.1 Formation of C-C + C-heteroatom bonds The first example of aerobic oxidative insertion of isocyanides into a heterocyclic ring was reported in 2011 by Zhu and co-workers for the synthesis of 4-aminoquinazolines 21.22 The authors reported the reaction between N-arylamidines 20 and isocyanides in the presence of Pd(OAc)2 under an O2 atmosphere, leading to the desired products in generally good yields (Scheme 6A). It is worth noting that the use of oxygen gave almost double yields of the desired product than other common oxidants, such as oxone and benzoquinone (BQ). The reaction worked well with both aliphatic and aromatic isocyanides, although stability issues somewhat hampered the use of the latter. As higher yields were obtained when electron-donating groups were present on the phenyl ring, the authors proposed an electrophilic aromatic substitution as the Csp2-H activation step (Scheme 6, right).22 An analogous mechanism was later suggested for the synthesis of benzoxazinoindazoles 23 and indazoloquinaoxalines 24 from indazole derivatives 22.23 While the use of oxygen was efficient for the former, the latter were obtained in higher yields employing CuCl2 as oxidant (Scheme 6B). Scheme 6. Oxidative isocyanide insertion leading to C-C + C-heteroatom bond formation. Arkivoc 2018, vi, 0-0 Sambiagio, C. Page 10 ©ARKAT USA, Inc 3.2 Formation of two C-heteroatom bonds The use of O2 as oxidant for isocyanide insertions between two heteroatoms was first reported by Maes/Orru/Ruijter in 2012 for the synthesis of 2-aminobenzimidazoles 25, and other heterocyclic cores were synthesized under similar reaction conditions (26-31, Scheme 7A).24 A notable aspect is the variety of Scheme 7. Oxidative isocyanide insertion leading to C-heteroatom + C-heteroatom bond formation. Arkivoc 2018, vi, 0-0 Sambiagio, C. Page 11 ©ARKAT USA, Inc (hetero)aromatic precursors tolerated in the synthesis of 2-aminobenzimidazole derivatives. Again high yields were obtained from electron-rich starting materials (substituent on Ar) using only 1% Pd, while strong electron-withdrawing groups required the use of higher catalyst loadings. Concerning the scope in isocyanides, a few aliphatic substituents were tolerated such as tBu, cyclohexyl, pentyl and iPr. The use of iPrNC as reactant resulted in certain amounts of side-products 32-33, derived from the insertion of two and three equivalents of isocyanides.24 A similar insertion of two molecules of isocyanide was also reported by Jiang and co-workers. In this case, a change in catalyst and reaction temperature proved crucial to selectively obtain the mono- or di-insertion product (34 and 26, Scheme 7B).25 A large variety of isocyanides was suitable for both transformations, including functionalized aliphatic, benzylic and aromatic isocyanides. Moreover, several other N,O heterocycles, including dibenzoxazepines 39, were prepared using similar conditions. Several other reports on the synthesis of related heterocycles were recently published,26-27 and an original decarboxylative variation was reported by Sharada in 2015. Here the authors employed isatoic anhydrides 40 as starting materials, which undergo decarboxylation-amination and isocyanide insertion in a one-pot fashion, to provide 2-amino-4-quinazolinones 28 (Scheme 7C).28 The use of molecular sieves was essential for obtaining good yields. The same transformation was achieved using Ag2CO3 as oxidant by Cai.29 A comparison of the results obtained for this reaction with different oxidants is reported in Scheme 7C. Although higher yields were obtained with Ag or Cu oxidants rather than O2, the use of two equivalents of the metal makes these options considerably more wasteful and more expensive. The proposed mechanism for the isocyanide insertion between two heteroatoms (Scheme 7, bottom)24 is only slightly different from the one shown in Scheme 6, however other mechanisms have been suggested to account for the formation of deamination side-products, sometimes observed in small amounts.25 3.3 Isocyanide insertion / acyl migration In 2014, while designing a catalytic system able to tolerate strongly coordinating heterocyclic substituents in group-directed C-H functionalization,30 Yu et al. discovered a new reactivity pattern in isocyanide insertion. Subjecting N-methoxy amides 41 to Pd-catalyzed aerobic isocyanide insertion, instead of the expected amino-substituted heterocycles, compounds 42 were isolated in high yields ( Scheme 8).31 These are formed by an initial isocyanide insertion, followed by a 1,3 acyl migration process, resulting in the incorporation of both the C and N atoms into the ring. Counter-intuitively, the reaction performed much better under air than under O2. This was later proved to be due to the faster oxidation of the isocyanide to isocyanate under O2, and the concomitant inhibition of the insertion reaction.32 Detailed mechanistic investigation of this process has been undertaken by Stahl32 and Wang.33 Arkivoc 2018, vi, 0-0 Sambiagio, C. Page 12 ©ARKAT USA, Inc Scheme 8. Synthesis of isoindolinone derivatives 42 via oxidative isocyanide insertion / acyl migration. 4. Oxidative CO Insertion As mentioned in the previous section, oxidative carbonylations catalyzed by transition metals have been explored even before the use of related isocyanides. As these transformations have already been reviewed,34-35 and generally follow the same mechanism as described for isocyanides, these will not be treated in depth here. Nonetheless, a few recent applications have appeared in the literature that are worth mentioning. As the mixture CO/O2 can generate considerable safety hazard, particular attention must be given to the reaction conditions (in particular pressure and ratio of the two gases), to avoid explosive mixtures. While CO insertions are typically catalyzed by Pd, Kaneda and co-workers disclosed the use of hydrotalcite-supported Au nanoparticles (Au@HT) as a heterogeneous catalytic system for the carbonylation of aminophenols, leading to benzoxazolinones and benzimidazolones 43-44 ( Scheme 9A).36 The solid catalyst could be recycled and re-used for a few runs without a reduction in yield, and no leaching was detected from the heterogeneous system. Interesting cascade reactions were reported under Pd catalysis by Gabriele/Mancuso/Della Ca’ for the synthesis of 8-membered lactams 46 and fused furo[3,4-b]indol-1-ones 48 ( Scheme 9B-C).37-38 The same authors also reported the Pd-catalyzed carbonylation of aminoalcohols/phenols in an ionic liquid ( Scheme 9D).39 Notably, after separation of the products, the ionic liquid medium (with the catalyst) could be recycled several times. A continuous-flow version of this reaction was recently reported by Kappe ( Scheme 9E).40 While in batch the use of a large excess of CO or the use of air instead of pure O2 is required to provide a certain operational safety, the use of flow allowed the use of equal pressures (and lower excesses than in batch) of CO and pure O2 in a safe manner. A range of 5- and 6-membered heterocycles were synthesized with this flow process. Arkivoc 2018, vi, 0-0 Sambiagio, C. Page 13 ©ARKAT USA, Inc Scheme 9. Synthesis of heterocycles via oxidative carbonyl insertion. 5. Oxidative Friedlander Annulation – Acridine Synthesis The Friedlander annulation of 2-acyl or 2-formylanilines with ketones for the synthesis of quinolines has been known for over a century.41-42 An interesting application of this chemistry in combination with aerobic oxidation has been recently explored for the formation of acridine cores. The role of oxygen here is limited to the dehydrogenation of the aliphatic ring introduced with the Friedlander reaction, and therefore not strictly involved in the synthesis of the heterocycle itself. Despite not being a general, common strategy for the synthesis of heterocycles, the importance of the acridine core and the lack of other general aerobic oxidation strategies for the synthesis of this specific moiety justify the addition of a small section in this review. Wang first reported in 2016 a metal-free method employing both TBHP and O2 as oxidants, under acidic conditions (54, Scheme 10).43 The same protocol could be applied for the synthesis of other heterocycles, employing 3-amino-2-acylbenzofuranes (55). Basing their studies on this report, Wu and Deng developed analogous processes under Pd catalysis, which avoided the use of TBHP ( Scheme 10).44-45 The postulated mechanism for this process involves an acid-catalyzed dehydrative coupling of 2-acylanilines and cyclohexanone (Friedlander annulation), followed by dehydrogenation, leading Arkivoc 2018, vi, 0-0 Sambiagio, C. Page 14 ©ARKAT USA, Inc to the desired acridine core. This mechanism is supported by the formation of quinolines 56 when using non-aromatizable ketones,44 and by the isolation of the oxygenated intermediate 57.43 Scheme 10. Aerobic oxidative synthesis of acridine cores via Friedlander annulation. 6. Oxidative N-Heteroatom Bond Formation A variety of 1,2-diheterocycles and related triheterocycles can be easily obtained via aerobic oxidative N-heteroatom bond formation, several examples of which have recently appeared in the literature. Oxidative N-N, N-O, N-S and N-Se bond formation have all been investigated for the synthesis of a range of 4- to 6-membered heterocyclic moieties. 6.1 Oxidative N-N bond formation Oxidative N-N bond formation is not a major research field in organic synthesis, but has been reported in the past to occur using strong oxidants or stoichiometric/superstoichiometric amounts of Cu.46-47 Since the 2000s, several reports have appeared on the use of catalytic Cu/O2 combination as a greener alternative. These methods can be exploited for the synthesis of pyrazoles, indazoles, or triazoles, which are traditionally synthesized using reagents already containing a N-N bond (i.e. hydrazines). Several variations for the synthesis of such heterocycles have been reported with the Cu/O2 methodology, often in a cascade or one-pot fashion. In such cases, an initial intermolecular reaction (typically non-oxidative) provides a suitable coordinating N,N substrate, which, upon coordination with Cu, can be oxidatively coupled by a reductive elimination-like process. The role of oxygen is presumably the reoxidation of the resulting reduced Cu species, allowing catalyst turnover (oxidase-type). While the second step (oxidative N-N bond formation) seems to be fairly easy under these conditions, the overall outcome of the reaction can be very dependent on the starting materials used and the coupling method. Arkivoc 2018, vi, 0-0 Sambiagio, C. Page 15 ©ARKAT USA, Inc 6.1.1 Synthesis of pyrazoles and indazoles. Pyrazoles are typically synthesized by reaction of hydrazines, hydrazones or analogues with alkynes, diols/diketones, or -unsaturated carbonyls.48 A variety of different pathways and synthons can be exploited for a N-N bond-forming process. In 2010 Glorius disclosed the serendipitous formation of pyrazoles upon treatment of enaminones 58 with Cu(OAc)2 in acetonitrile.46 Although this reaction required 1.5 eq. of Cu and was actually sensitive to O2 atmosphere, the protocol was later improved to employ only catalytic amounts of Cu (10%) and O2 as oxidant.49 This reaction worked efficiently with a range of substituted enaminones and nitriles, either aliphatic or aromatic, proving a versatile method for the synthesis of 1,3,4,5-tetrasubstituted pyrazoles 59 ( Scheme 11A). Nonetheless, the reaction outcome in aerobic conditions varied considerably depending on the substrate and the nitrile used, and the yields obtained were comparable or lower than with stoichiometric Cu. Another important pyrazole synthesis was published by Jiang in 2014, occurring via a Cu-catalyzed multicomponent reaction between oximes, anilines and formaldehyde, furnishing 1,3-disubstituted (mostly diaryl) pyrazoles 61 ( Scheme 11B).50 The reaction under N2 gave good yields of the corresponding pyrazoline 62, suggesting O2 to be only involved in the dehydrogenation of the ring, rather than the N-N bond formation itself. This process is here likely promoted by the oxime moiety, which acts as an internal oxidant. A similar approach was independently described by Cui and co-workers. In this case, the cyclization of tosylimines and oxime esters also resulted in pyrazolines, but also required superstoichiometric Cu salts.51 Although the reaction was performed under air, aerobic oxygen was probably not essential for the process, due to the high amounts of Cu used. For the synthesis of indazoles 63, the Glorius group reported a tandem process involving Rh-catalyzed group-directed30 C-H amidation (intermediate 64) and subsequent Cu-catalyzed aerobic N-N bond formation ( Scheme 11C).52 An analogous cascade process, only employing Cu as catalyst for both the directed C-H sulfonamidation and the N-N coupling was reported in 2014 by Zhu, but under inert conditions (the azide was suggested to act as both coupling partner and oxidant).53 Another indazole synthesis was reported by Chen and Tang in 2016 starting from ortho-aminobenzonitriles ( Scheme 11D).54 Ketimines 65 were prepared from the nitriles via reaction with a Grignard or organolithium reagent, and subsequent Cu-catalyzed oxidation resulted in the desired indazoles 66. As a range of organometallic species and N-substituents could be tolerated under these conditions, this method appears complementary to Glorius’s procedure, and several compounds were prepared in a multigram scale.54 When a N-aryl substituted substrate was employed, the aerobic oxidation also resulted in a cascade ortho hydroxylation of the N-aryl ring (67).55 Also in this instance the reaction proved robust and good-yielding even at gram scale. It is noteworthy that the use of indazole as a pre-formed directing group was inefficient for the hydroxylation, as well as the amino group alone.55 Arkivoc 2018, vi, 0-0 Sambiagio, C. Page 16 ©ARKAT USA, Inc Scheme 11. Synthesis of pyrazole and indazole derivatives via oxidative N-N bond formation. 6.1.2. Synthesis of triazoles. 1,2,4-Triazoles are easily formed by addition of amidines to nitriles (or analogous reactants), followed by intramolecular oxidative N-N coupling. In 2009 this strategy was employed by Nagasawa for the synthesis of triazolopyridines 68 and triazoles 69 in generally good yields ( Scheme 12A).56 This cascade protocol worked efficiently with a variety of aromatic nitriles, but yields were limited when using cyanamides, likely due to difficult addition to the nitrile in such compounds.57 Given the interest in the 2-amino-[1,2,4]triazolo[1,5-a]pyridine scaffold in medicinal chemistry, researchers at Roche modified Nagasawa’s protocol employing guanidinic precursors 71 ( Scheme 12B).57 This method allowed the efficient synthesis of 2-aminotriazolopyridines 72 and analogues 73-76 from diazaheterocyclic guanidines. Another very similar procedure to Nagasawa’s was reported by Zhang, this time applied to the synthesis of trisubstituted triazoles 77. Importantly, the authors also reported the synthesis of disubstituted triazoles 69 using aliphatic nitriles (1 example only), typically not reactive under these conditions ( Scheme 12C).58 The synthesis of 1,2,4-trazoles 69 by condensation or dimerization of amidines or imidates (both derived from nitriles) was reported by Fu59 and Beifuss60 ( Scheme 12D-E). These processes result in the same intermediate obtained from the addition of amidines to nitriles (70), but the long reaction time, the tendency of aromatic amidines to dimerize, and/or the complex reaction mixture employed make these methods not highly efficient, despite synthetically useful yields of 60-Arkivoc 2018, vi, 0-0 Sambiagio, C. Page 17 ©ARKAT USA, Inc 70% could be achieved, depending on the substituents. Limited yields were also obtained using an anaerobic protocol starting from nitriles/hydroxylamine combinations.61 Scheme 12. Synthesis of 1,2,4-triazole derivatives via oxidative N-N bond formation. The synthesis of 1,2,3-triazoles can be accomplished via Cu-catalyzed intramolecular coupling between a hydrazine and another N moiety in the substrate, as described by Nagasawa for the synthesis of 1,2,3-triazolopyridines 80 ( Scheme 13A).62 The reaction could be performed either in one step starting from 2-acylpyridine hydrazones 78, or in a one-pot manner starting from the corresponding 2-acylpyridines 79. High yields were Arkivoc 2018, vi, 0-0 Sambiagio, C. Page 18 ©ARKAT USA, Inc obtained with both protocols. Mechanistic studies suggested that the re-oxidation of the Cu(I) species formed during the reaction was the rate-determining step, and that the acetate anion acted as both a ligand and a base to deprotonate the hydrazone.62 An interesting protocol for the synthesis of 2,4,5-triaryl 1,2,3-triazoles 82 was disclosed by Punniyamurthy, via dimerization of bisarylhydrazones 81 ( Scheme 13B). In this process the Cu catalyst is also involved in the initial C-C bond formation.63 Although one example of asymmetric 1,2,3-triazole was shown, the above method is clearly more suitable for the synthesis of symmetric products. A complementary method, starting from hydrazonoketones 84 and an ammonium source, which allows the synthesis of variedly functionalized triazoles, was reported by Lin in 2015 ( Scheme 13C).64 With this procedure the C4 and C5 positions of the triazole can be functionalized with a variety of carbonyls (esters, ketones, amides), aryl, or alkyl groups (85), although the yields are strongly dependent on the substitution pattern. Scheme 13. Synthesis of 1,2,3-triazole derivatives via oxidative N-N bond formation. 6.2 Oxidative N-O, N-S, and N-Se bond formation Driven by the research on oxidative N-N bond-forming processes, the formation of other heteroatom-heteroatom bonds started to be investigated under similar catalytic conditions. Analogously to 1,2,4-triazoles, 1,2,4-oxadiazoles 86 could be prepared by addition of amides to nitriles, followed by oxidative N-O coupling, as described by Hong and co-workers ( Arkivoc 2018, vi, 0-0 Sambiagio, C. Page 19 ©ARKAT USA, Inc Scheme 14A).65 The use of molecular sieves and ZnI2, together with an excess of nitrile, were crucial for an efficient coupling/cyclization. Despite with variable yields (20-94%), the reaction shows a broad scope in nitriles, in particular heteroaromatic and aliphatic. A couple of medicinally relevant compounds (a precursor of Ataluren, a drug against cystic fibrosis, and an inhibitor of penicillin-binding protein PBP2a, 87 and 88) could also be synthesized in decent yields, highlighting the synthetic utility of the method. Scheme 14. Synthesis of various heterocyclic cores via oxidative N-O, N-S and N-Se bond formation. Arkivoc 2018, vi, 0-0 Sambiagio, C. Page 20 ©ARKAT USA, Inc The addition of amidines to isothiocyanates, followed by intramolecular N-S coupling, results in 5-amino-1,2,4-thiadiazoles 89, another important heterocyclic core ( Scheme 14B).66 Although the cascade process gave lower yields than the intramolecular coupling starting from the intermediate imidoylthioureas, the reaction could still be completed in short time (2-6 h) and without else than Cu(OTf)2 and Cs2CO3, required for the N-S coupling. An extension of this protocol was recently reported by Jiang and Wu, who employed 2-aminopyridines as substrates to synthesize thiadiazolopyridines 90 ( Scheme 14C).67 The reaction was fairly sensitive to the substitution pattern of both coupling partners, but high yields could be obtained for several examples. Benzoisothiazolones 9268 and benzothiadiazinones 9569 can be prepared in high yields from 2-mercaptobenzamides 91 or 2-mercaptophenylureas 94 under aerobic CuI catalysis, as demonstrated respectively by Kanai/Kuninobu and Dogan ( Scheme 14D-E). N-aryl and N-alkyl substituents are tolerated in both protocols, and one example of benzothiazetidine 93 was also reported in 84% yield.68 A couple of interesting metal-free protocols for similar couplings have also been reported. The synthesis of 1,2,3-thiadiazole derivatives 97 was reported by Li and Jiang using iodine as a SET catalyst ( Scheme 14F),70 while an oxidative N-Se bond formation strategy for the synthesis of 1,2,4-selenadiazoles 98 was recently disclosed by Ji/Wang and co-workers ( Scheme 14G).71 The reaction of isocyanides with selenium powder results in the corresponding isoselenocyanates 101, which undergo coupling with the amidines and oxidative cyclization, giving the selenadiazoles. Considering the lack of metal catalyst in this process, the role of oxygen was suggested to be related to the oxidation of the anionic organic intermediate 102 formed upon addition of the imidine to the isoselenocyanate, leading to a Se-centered radical (103). This radical might then undergo cyclization via different pathways ( Scheme 14).71 7. Oxidative C-Heteroatom Bond Formation 7.1 Oxidative C-N bond formation The formation of Csp2-N bonds through oxidative C-H bond amination is a very common strategy for the synthesis of aza-heterocycles, and many examples have now been investigated under aerobic conditions. This transformation typically requires the use of transition metal catalysts, and many reports using Pd and other noble metals, as well as a variety of oxidants, were reported.72 More recently, the use of Cu and other base metals have instead become predominant, and some procedures under metal-free conditions have also been successfully applied. Due to the many examples reported in the recent literature, only a selection will be reported in this section. In 2008 Buchwald reported the synthesis of benzimidazoles 105 from amidines 104 using catalytic Cu(OAc)2 and acetic acid as additive under an O2 atmosphere ( Scheme 15A).73 Similar procedures were then employed for the synthesis of pyrido[1,2-a]benzimidazoles 107 by Zhu/Zhang74 and Maes,75 and for the synthesis of 1,4-dihydropyrazine moieties 109 by Fu ( Arkivoc 2018, vi, 0-0 Sambiagio, C. Page 21 ©ARKAT USA, Inc Scheme 15B-C).76 The mechanism of this reaction is not certain, and different views are reported in the literature, but the role of the acid seems to be related to the formation of a more electrophilic Cu intermediate, as more electron-withdrawing carboxylates were shown to be more effective in some cases.75 Several other syntheses have been reported using similar protocols in the presence of a base. Examples are imidazobenzimidazoles 111 and benzoimidazoquinazolinones 113, prepared in high yields despite the long reaction time (2 days or more), as reported by Fu ( Scheme 15D-E).77-78 The Maes group reported the synthesis of purine derivatives 117 and 118 via aerobic formation of two different C-N bonds. While the formation of the C4-N9 bond in Scheme 15F-top required the use of a Fe catalyst in large amounts and 24 h reaction time in most cases,79 the C5-N7 bond in Scheme 15F-bottom could be easily formed with Cu, either using O2 (18 h reaction) or tBu2O2 (2 h reaction) as oxidant.80 Despite the longer reaction time required when using O2, the yields obtained in this case are often higher than those achieved with tBu2O2. The synthesis of pyrazoles, indazoles, and similar compounds, which can be also prepared by oxidative N-N bond formation (Section 6.1), have been reported using this strategy, either under metal catalysis81-83 or in metal-free conditions84-85 ( Scheme 15G-H). Hydrazine derivatives need to be employed as starting materials in these cases, which have to be prepared beforehand. Arkivoc 2018, vi, 0-0 Sambiagio, C. Page 22 ©ARKAT USA, Inc Scheme 15. Synthesis of various aza-heterocycles via oxidative C-N bond formation. Although no thorough mechanistic investigations have been yet reported for these transformations, several authors reported preliminary studies involving kinetic isotope effect (KIE), radical-trapping experiments, or electron paramagnetic spectroscopy (EPR) studies. On the basis of these, two main mechanisms have been proposed. One involves the formation of free C- or N-centered radicals, and has been suggested for both metal-free and metal-catalyzed reactions.79, 83-84 In the case of metal-catalyzed reactions no obvious coordination between metal and substrate has been suggested, and the metal seems to act mostly as an oxidant for a double single electron-transfer (SET) process, as described in Scheme 16-top. The second mechanism is often proposed only based on radical-trapping experiments. The absence of an inhibition is interpreted in terms of coordination of the two coupling moieties to the metal, (e.g. electrophilic aromatic substitution, or a different mechanism), and the absence of free radicals (Scheme 16-bottom). Variations of these mechanisms can be postulated for different substrates. Arkivoc 2018, vi, 0-0 Sambiagio, C. Page 23 ©ARKAT USA, Inc Scheme 16. Examples of radical and non-radical mechanisms proposed for the Cu-catalyzed oxidative C-N bond formation in heterocyclic synthesis. 7.2 Oxidative C-O and C-S bond formation Oxidative cross-dehydrogenative couplings for the formation of C-O and C-S bonds in heterocycles have also been explored, albeit less frequently than C-N bonds. In particular, benzoxazoles have been prepared using a variety of protocols, which can be divided into two main strategies. In 2008 Nagasawa disclosed a Cu-catalyzed aerobic C-O coupling leading from N-phenylamides to benzoxazoles 122, according to Scheme 17A.86-87 Despite being fairly substrate-sensitive, the reaction proceeded well giving decent yields of a variety of 2-aryl and 2-alkyl benzoxazoles. The reaction was suggested to occur with a mechanism similar to the formation of benzimidazoles 105 reported by Buchwald in the same year ( Scheme 15A), i.e. via an electrophilic aromatic substitution as in Scheme 16-bottom. A second strategy involves, instead of the activation of an Ar-H bond, the nucleophilic attack of an OH group to an imine. The required intermediate 123 can be formed in-situ from different substrates,88-90 as shown in Scheme 17B. Importantly, the use of o-aminothiols as substrates also proved amenable to cyclization, furnishing benzothiazoles 124 using Methods 1 and 2 (Scheme 17B). Note that Method 1 involves a strategy described in Section 2.1, but was reported here for comparison. Isoxazoles 126 and 1,3,4-oxadiazoles 128 can also be synthesized via dehydrogenative C-O couplings, as shown by Sun91 and Patel,92 respectively (Scheme 17C-D). A metal-free synthesis of 2-aminobenzothiazoles 130 starting from isothiocyanates 129 and a variety of amines (primary, secondary, aromatic) was developed by Fan and Zhang in 2017, employing iodine as catalyst (Scheme 17E).93 This protocol is particularly interesting, as it is complementary to the Pd-catalyzed insertion of Arkivoc 2018, vi, 0-0 Sambiagio, C. Page 24 ©ARKAT USA, Inc isocyanides (cf. Section 3), which only allows the introduction of mono-substituted amino groups in the heterocycle (and mostly N-alkyl substituents). The reaction was proposed to be initiated by the addition of the amine to the isothiocyanate, resulting in the intermediate thiourea 131, which then cyclizes upon formation of a S-I bond (Scheme 17E, right). This reaction worked similarly using other oxidants or stoichiometric iodine. Scheme 17. Synthesis of various heterocycles via oxidative C-O and C-S bond formation. 8. Intermolecular Oxidative Cyclization with Unsaturated Moieties A variety of substrates can be oxidatively coupled with unsaturated moieties leading to formal [3+2] cycloaddition products. As different mechanistic pathways can lead to the same overall transformation, these reactions are somewhat difficult to classify, and some of the processes presented in other sections might also fall in this category. Here a few specific strategies will be treated, divided according to substrate and mechanism. Arkivoc 2018, vi, 0-0 Sambiagio, C. Page 25 ©ARKAT USA, Inc 8.1 Synthesis of imidazole derivatives via cyclization of amidines with electron-poor alkenes The aerobic coupling/cyclization of 2-aminopyridines with α,β-unsaturated carbonyls to give 3-aroylimidazo[1,2-a]pyridines 132 was first reported by Hajra in 2014 ( Scheme 18A).94 The reaction was catalyzed by Cu salts under an O2 atmosphere, and gave good yields (74-86%) of the desired compounds with complete regioselectivity. The reaction was shown to proceed via a non-catalyzed Michael addition of the 2-aminopyridine to the α,β-unsaturated carbonyl to give the intermediate shown in Scheme 18-bottom, followed by an oxidative cyclization and aromatization. The main limitation of this protocol was the necessity to use chalcones (or dibenzylideneacetone) as coupling partners. Aliphatic unsaturated ketones did not give any product, and benzylideneacetone gave only 26% yield of the corresponding heterocycle.94 Other similar reports appeared in the following years, including a multicomponent process where the formation of the unsaturated carbonyl occurred in-situ via an aldol reaction, reported by Kumar,95 and the extension of the procedure to cynnamaldehydes by Bharate and co-workers (133, Scheme 18B).96 Hachem and Gree subsequently further extended the scope to γ-difluoroenones 134, albeit with limited yields ( Scheme 18C).97 In 2015 an alternative catalytic method for this reaction was reported by Li et al., who introduced the use of Fe/I2 as a double catalyst for the cyclization of amidines with chalcones, furnishing imidazoles 138 ( Scheme 18D).98 The use of I2 as co-catalyst facilitates the formation of the second C-N bond after the initial Michael addition. This made higher yields possible to achieve, and enlarged the substrate scope of the reaction. Fe/I2 and Cu/I2 catalytic systems were therefore subsequently employed for similar reactions.99-100 In particular, an interesting advancement was brought forward by Nguyen and co-workers, who performed the coupling of 2-aminopyridines and 2-aminopirimidines with cyclohexenones, previously not reactive in this type of transformation (139, Scheme 18E).100 Notably, the use of I2 or Fe alone in this case only resulted in the formation of the Michael adduct, without cyclization taking place. The cyclization of aminopyridines with nitroalkenes under photoredox catalysis by Eosin Y, reported in 2016 by Singh and co-workers, is a variation of the reaction with chalcones described above. Under these conditions, however, a reversed selectivity was observed, and 3-arylimidazopyridines 140 were obtained instead of the corresponding 2-aryl derivatives ( Scheme 18F).101 This reaction was also suggested to occur via a Michael addition (pyridine as nucleophile), although the origin of the different regioselectivity is not clear. Arkivoc 2018, vi, 0-0 Sambiagio, C. Page 26 ©ARKAT USA, Inc Scheme 18. Cyclization of amidines/aminopyridines with electron-poor alkenes for the synthesis of imidazole and pyridoimidazole derivatives. 8.2 Synthesis of benzofurans via cyclization of phenols with alkynes The synthesis of benzofurans via oxidative cyclization of phenols and alkynes has been known for quite some time. This reaction can be performed in different ways, for example in the presence of a transition metal catalyst with stoichiometric Cu as oxidant,102-103 or using N-phenoxyamides as internal oxidant,104 but some interesting examples have been disclosed that make use of O2 in the presence of catalytic Cu or Au. A catalytic amount of Cu(II) salt and a stoichiometric amount of ZnCl2 as Lewis acid under an O2 atmosphere was demonstrated by Jiang to be a useful combination for this reaction in 2013.105 The reaction Arkivoc 2018, vi, 0-0 Sambiagio, C. Page 27 ©ARKAT USA, Inc performed efficiently with symmetric diarylalkynes and with arylalkylalkynes, in the latter case in a regioselective way (141 and 142, Scheme 19A). The same transformation was later reported by Guo using only a cationic Au(I) catalyst, affording similar results.106 Important follow-up research on this transformation revealed that ZnCl2 alone can promote the oxidative cyclization acting as both Lewis acid and oxidant in the absence of other reagents.107 Satoh and Miura recently reported the synthesis of benzofurans from salicylaldehydes via an Ir-catalyzed decarbonylative coupling with alkynes ( Scheme 19B).108 This reaction occurs with a completely different mechanism from the above coupling, furnishing overall much lower yields and requiring more complex reaction mixtures. Nonetheless, a dialkylalkyne could be coupled with a 29% yield, and the reaction with unsymmetric arylalkylalkyne gave the opposite regioisomer to the one obtained by Jiang with Cu/ZnCl2 (cf. 142 and 143). Scheme 19. Formation of benzofurans via aerobic cyclization of alkynes with phenols or salicylaldehydes. 8.3 Miscellaneous cyclizations with alkynes Halogenated imidazopyridines 144 and related compounds 145-148 could be prepared in good yields (60-80%) by cyclization of 2-aminopyridines with haloalkynes ( Scheme 20A).109 As haloalkynes are prone to react via nucleophilic substitution/addition, or to dimerization, this is a remarkable transformation. Either iodo, bromo, or chloroalkynes proved suitable for this transformation, and a variety of aminopyridines and aminodiazines could be efficiently coupled. A non-radical Arkivoc 2018, vi, 0-0 Sambiagio, C. Page 28 ©ARKAT USA, Inc mechanism was suggested for this reaction, involving migratory insertion of the alkyne into a N-Cu bond, followed by reductive elimination.109 Scheme 20. Formation of various 5-membered heterocycles via aerobic cyclization of alkynes with different substrates. Hu and co-workers reported an interesting synthesis of pyrrolo[2,1,5-cd]indolizines 150 via cyclization between indolizines 149 and diarylalkynes under Pd catalysis ( Scheme 20B).110 The substrate scope for this reaction is limited to electron-poor indolizines (EWG in position 2 or 3) and to diarylalkynes (unsymmetrical alkynes give regioisomeric mixtures). Arkivoc 2018, vi, 0-0 Sambiagio, C. Page 29 ©ARKAT USA, Inc The synthesis of trisubstituted oxazoles 151 via a formal [2+2+1] cycloaddition between alkynes, nitriles and water under Cu catalysis was disclosed by Jiang in 2012 ( Scheme 20C).111 Both diaryl and dialkyl alkynes were well tolerated, and regioselective products were obtained with unsymmetrical alkynes; moreover, a variety of nitriles (allylic, aromatic and aliphatic) could be incorporated into the final products. Although the mechanism was not studied in detail, some experiments suggested an initial nucleophilic attack of the nitrile to the Cu-activated alkyne, followed by nucleophilic attack of water to the cationic intermediate 152, and finally intramolecular C-O bond formation.111 As for other reactions of this kind (Section 7), the role of oxygen is the re-oxidation of the reduced Cu species at the end of the catalytic cycle. 9. Conclusions This review summarizes some of the recent developments in the synthesis of heterocycles making use of catalytic aerobic oxidation. Such oxidative processes can be applied to the formation of C-C, C-heteroatom, and heteroatom-heteroatom bonds, sometimes contributing to the formation of more than one bond at once. Because of this, a very large array of moieties can be accessed with these methods, and N-, O-, S-, and Se-containing heterocycles can be easily prepared, often with good yields and selectivity. The generally broad scope, combined with the low price and toxicity, and the high atom-economic nature of oxygen make it a very versatile and appealing reagent for the synthesis of heterocycles, when compared to other commonly used oxidants (e.g. hypervalent iodine species, peroxides, stoichiometric Cu or Ag salts). The current state of the research in aerobic oxidation led to the identification of specific catalytic systems, such as Cu/O2, Fe/O2 or I2/O2, which can be applied, with relatively little variation, to a number of processes, as can be seen in this review. These systems are now also being studied from a mechanistic perspective, and can be considered a starting point for future aerobic oxidation research projects. As mentioned in the introduction, practical limitations to the use of gaseous O2 exist, which somewhat prevent its use in the fine chemicals industry. Nonetheless, research in this field is very active, and adequate technology for the broader and safer use of aerobic oxidation processes might become available in the near future. Acknowledgements The author acknowledges the European Union for a Marie Curie European post-doctoral fellowship (FlowAct, Grant No. 794072). References 1. Allen, S. E.; Walvoord, R. R.; Padilla-Salinas, R.; Kozlowski, M. C. Chem. Rev. 2013, 113, 6234-6458. https://doi.org/10.1021/cr300527g 2. Sterckx, H.; Morel, B.; Maes, B. U. Angew. Chem. Int. Ed., doi:10.1002/anie.201804946. https://doi.org/10.1002/anie.201804946 Arkivoc 2018, vi, 0-0 Sambiagio, C. Page 30 ©ARKAT USA, Inc 3. Stahl, S. S.; Alster, P. L. Liquid Phase Aerobic Oxidation Catalysis: Industrial Applications and Academic Perspectives. Wiley-VCH: 2016, 10.1002/9783527690121. https://doi.org/10.1002/9783527690121 4. Hone, C. A.; Roberge, D. M.; Kappe, C. O. ChemSusChem 2017, 10, 32-41. https://doi.org/10.1002/cssc.201601321 5. Gemoets, H. P. L.; Su, Y.; Shang, M.; Hessel, V.; Luque, R.; Noël, T. Chem. Soc. Rev. 2016, 45, 83-117. https://doi.org/10.1039/C5CS00447K 6. Pieber, B.; Kappe, C. O. In Organometallic Flow Chemistry, Noël, T., Ed. Springer International Publishing: Cham, 2016, 10.1007/3418_2015_133 pp 97-136. 7. Yaragorla, S.; Vijaya Babu, P. Tetrahedron Lett. 2017, 58, 1879-1882. https://doi.org/10.1016/j.tetlet.2017.04.007 8. Yang, L.; Shi, X.; Hu, B. Q.; Wang, L. X. Asian J. Org. Chem. 2016, 5, 494-498. https://doi.org/10.1002/ajoc.201600041 9. Li, Q.; Huang, Y.; Chen, T.; Zhou, Y.; Xu, Q.; Yin, S.-F.; Han, L.-B. Org. Lett. 2014, 16, 3672-3675. https://doi.org/10.1021/ol501454j 10. Chen, X.; Chen, T.; Ji, F.; Zhou, Y.; Yin, S.-F. Catal. Sci. Technol. 2015, 5, 2197-2202. https://doi.org/10.1039/C4CY01618A 11. Sterckx, H.; De Houwer, J.; Mensch, C.; Caretti, I.; Tehrani, K. A.; Herrebout, W. A.; Van Doorslaer, S.; Maes, B. U. W. Chem. Sci. 2016, 7, 346-357. https://doi.org/10.1039/C5SC03530A 12. Pang, X.; Wu, M.; Ni, J.; Zhang, F.; Lan, J.; Chen, B.; Yan, R. J. Org. Chem. 2017, 82, 10110-10120. https://doi.org/10.1021/acs.joc.7b01575 13. Pereira, M. d. F.; Thiéry, V. Org. Lett. 2012, 14, 4754-4757. https://doi.org/10.1021/ol302006b 14. Li, J.; Zhang, J.; Yang, H.; Gao, Z.; Jiang, G. J. Org. Chem. 2017, 82, 765-769. https://doi.org/10.1021/acs.joc.6b02501 15. Tan, Z.; Zhao, H.; Zhou, C.; Jiang, H.; Zhang, M. J. Org. Chem. 2016, 81, 9939-9946. https://doi.org/10.1021/acs.joc.6b02117 16. Cai, Z.-J.; Wang, S.-Y.; Ji, S.-J., Org. Lett. 2012, 14, 6068-6071. https://doi.org/10.1021/ol302955u 17. Wang, Y.-F.; Chen, H.; Zhu, X.; Chiba, S. J. Am. Chem. Soc. 2012, 134, 11980-11983. https://doi.org/10.1021/ja305833a 18. Pan, J.; Li, X.; Qiu, X.; Luo, X.; Jiao, N. Org. Lett. 2018, 20, 2762-2765. https://doi.org/10.1021/acs.orglett.8b00992 19. Xu, Z.; Zhang, C.; Jiao, N. Angew. Chem. Int. Ed. 2012, 51, 11367-11370. https://doi.org/10.1002/anie.201206382 20. Vlaar, T.; Ruijter, E.; Maes, B. U. W.; Orru, R. V. A. Angew. Chem. Int. Ed. 2013, 52, 7084-7097. https://doi.org/10.1002/anie.201300942 21. Lang, S., Chem. Soc. Rev. 2013, 42, 4867-4880. https://doi.org/10.1039/c3cs60022j 22. Wang, Y.; Wang, H.; Peng, J.; Zhu, Q. Org. Lett. 2011, 13, 4604-4607. https://doi.org/10.1021/ol201807n 23. Vidyacharan, S.; Murugan, A.; Sharada, D. S. J. Org. Chem. 2016, 81, 2837-2848. https://doi.org/10.1021/acs.joc.6b00048 Arkivoc 2018, vi, 0-0 Sambiagio, C. Page 31 ©ARKAT USA, Inc 24. Vlaar, T.; Cioc, R. C.; Mampuys, P.; Maes, B. U. W.; Orru, R. V. A.; Ruijter, E. Angew. Chem. Int. Ed. 2012, 51, 13058-13061. https://doi.org/10.1002/anie.201207410 25. Liu, B.; Yin, M.; Gao, H.; Wu, W.; Jiang, H. J. Org. Chem. 2013, 78, 3009-3020. https://doi.org/10.1021/jo400002f 26. Vlaar, T.; Bensch, L.; Kraakman, J.; Velde, C. M. L. V.; Maes, B. U. W.; Orru, R. V. A.; Ruijter, E. Adv. Synth. Catal. 2014, 356, 1205-1209. https://doi.org/10.1002/adsc.201301129 27. Vlaar, T.; Orru, R. V. A.; Maes, B. U. W.; Ruijter, E. J. Org. Chem. 2013, 78, 10469-10475. https://doi.org/10.1021/jo401924h 28. Vidyacharan, S.; Chaitra, N. C.; Sagar, A.; Sharada, D. S. Synth. Comm. 2015, 45, 898-907. https://doi.org/10.1080/00397911.2014.992076 29. Ji, F.; Lv, M.-F.; Yi, W.-B.; Cai, C. Org. Biomol. Chem. 2014, 12, 5766-5772. https://doi.org/10.1039/C4OB00484A 30. Sambiagio, C.; Schönbauer, D.; Blieck, R.; Dao-Huy, T.; Pototschnig, G.; Schaaf, P.; Wiesinger, T.; Zia, M. F.; Wencel-Delord, J.; Besset, T.; Maes, B. U. W.; Schnürch, M. Chem. Soc. Rev. 2018, 47, 6603-6743. https://doi.org/10.1039/C8CS00201K 31. Liu, Y.-J.; Xu, H.; Kong, W.-J.; Shang, M.; Dai, H.-X.; Yu, J.-Q. Nature 2014, 515, 389. https://doi.org/10.1038/nature13885 32. Tereniak, S. J.; Stahl, S. S. J. Am. Chem. Soc. 2017, 139, 14533-14541. https://doi.org/10.1021/jacs.7b07359 33. Dang, Y.; Deng, X.; Guo, J.; Song, C.; Hu, W.; Wang, Z.-X. J. Am. Chem. Soc. 2016, 138, 2712-2723. https://doi.org/10.1021/jacs.5b12112 34. Gadge, S. T.; Bhanage, B. M. RSC Advances 2014, 4, 10367-10389. https://doi.org/10.1039/c3ra46273k 35. Gabriele, B.; Mancuso, R.; Salerno, G. Eur. J. Org. Chem. 2012, 2012, 6825-6839. https://doi.org/10.1002/ejoc.201200794 36. Noujima, A.; Mitsudome, T.; Mizugaki, T.; Jitsukawa, K.; Kaneda, K. Green Chem. 2013, 15, 608-611. https://doi.org/10.1039/c2gc36851j 37. Mancuso, R.; Raut, D. S.; Marino, N.; De Luca, G.; Giordano, C.; Catalano, S.; Barone, I.; Andò, S.; Gabriele, B. Chem. Eur. J. 2016, 22, 3053-3064. https://doi.org/10.1002/chem.201504443 38. Acerbi, A.; Carfagna, C.; Costa, M.; Mancuso, R.; Gabriele, B.; Della Ca', N. Chem. Eur. J. 2018, 24, 4835-4840. https://doi.org/10.1002/chem.201706067 39. Mancuso, R.; Raut, D. S.; Della Ca', N.; Fini, F.; Carfagna, C.; Gabriele, B. ChemSusChem 2015, 8, 2204-2211. https://doi.org/10.1002/cssc.201500343 40. Chen, Y.; Hone, C. A.; Gutmann, B.; Kappe, C. O. Org. Process Res. Dev. 2017, 21, 1080-1087. https://doi.org/10.1021/acs.oprd.7b00217 41. Cheng, C.-C.; Yan, S.-J., In Organic Reactions, Dauben, W. G. Ed. John Wiley & Sons: 1982; Vol. 28, pp 37-201. https://doi.org/10.1002/0471264180.or028.02 42. Shiri, M.; Zolfigol, M. A.; Kruger, H. G.; Tanbakouchian, Z. In Advances in Heterocyclic Chemistry, Katritzky, A. R., Ed. Academic Press: 2011; Vol. 102, pp 139-227. 43. Senadi, G. C.; Dhandabani, G. K.; Hu, W.-P.; Wang, J.-J. Green Chem. 2016, 18, 6241-6245. https://doi.org/10.1039/C6GC02396G Arkivoc 2018, vi, 0-0 Sambiagio, C. Page 32 ©ARKAT USA, Inc 44. Wan-Lu, M.; Meirong, W.; Hui-Jing, L.; Deng-Ming, H.; Yi-Yun, Z.; Chao-Yi, L.; Ying, L.; Yan-Chao, W. Adv. Synth. Catal. 2017, 359, 4250-4257. https://doi.org/10.1002/adsc.201700715 45. Xiangui, C.; Yanjun, X.; Cheng, L.; Fuhong, X.; Guo-Jun, D. Eur. J. Org. Chem. 2017, 2017, 577-581. 46. Neumann, J. J.; Suri, M.; Glorius, F. Angew. Chem. Int. Ed. 2010, 49, 7790-7794, and refs. therein. https://doi.org/10.1002/anie.201002389 47. Mohan, D. C.; Ravi, C.; Rao, S. N.; Adimurthy, S. Org. Biomol. Chem. 2015, 13, 3556-3560. https://doi.org/10.1039/C5OB00018A 48. Fustero, S.; Simón-Fuentes, A.; Sanz-Cervera, J. F. Org. Prep. Proced. Int. 2009, 41, 253-290. https://doi.org/10.1080/00304940903077832 49. Suri, M.; Jousseaume, T.; Neumann, J. J.; Glorius, F. Green Chem. 2012, 14, 2193-2196. https://doi.org/10.1039/c2gc35476d 50. Tang, X.; Huang, L.; Yang, J.; Xu, Y.; Wu, W.; Jiang, H. Chem. Comm. 2014, 50, 14793-14796. https://doi.org/10.1039/C4CC06747A 51. Wu, Q.; Zhang, Y.; Cui, S. Org. Lett. 2014, 16, 1350-1353. https://doi.org/10.1021/ol500094w 52. Yu, D.-G.; Suri, M.; Glorius, F. J. Am. Chem. Soc. 2013, 135, 8802-8805. https://doi.org/10.1021/ja4033555 53. Peng, J.; Xie, Z.; Chen, M.; Wang, J.; Zhu, Q. Org. Lett. 2014, 16, 4702-4705. https://doi.org/10.1021/ol502010g 54. Chen, C.-y.; Tang, G.; He, F.; Wang, Z.; Jing, H.; Faessler, R. Org. Lett. 2016, 18, 1690-1693. https://doi.org/10.1021/acs.orglett.6b00611 55. Cheng-yi, C.; Fengxian, H.; Guangrong, T.; Han, D.; Zhaobin, W.; Dawei, L.; Lujiang, D.; Roger, F. Eur. J. Org. Chem. 2017, 2017, 6604-6608. 56. Ueda, S.; Nagasawa, H. J. Am. Chem. Soc. 2009, 131, 15080-15081. https://doi.org/10.1021/ja905056z 57. Bartels, B.; Bolas, C. G.; Cueni, P.; Fantasia, S.; Gaeng, N.; Trita, A. S. J. Org. Chem. 2015, 80, 1249-1257. https://doi.org/10.1021/jo502536t 58. Wang, F.; You, Q.; Wu, C.; Min, D.; Shi, T.; Kong, Y.; Zhang, W. RSC Advances 2015, 5, 78422-78426. https://doi.org/10.1039/C5RA15919A 59. Xu, H.; Jiang, Y.; Fu, H. Synlett 2013, 24, 125-129. 60. Sudheendran, K.; Schmidt, D.; Frey, W.; Conrad, J.; Beifuss, U. Tetrahedron 2014, 70, 1635-1645. https://doi.org/10.1016/j.tet.2014.01.019 61. Xu, H.; Ma, S.; Xu, Y.; Bian, L.; Ding, T.; Fang, X.; Zhang, W.; Ren, Y. J. Org. Chem. 2015, 80, 1789-1794. https://doi.org/10.1021/jo502709t 62. Tasuku, H.; Satoshi, U.; Takahiro, O.; Norihiko, T.; Kensuke, O.; Hideko, N. Chem. Eur. J. 2014, 20, 4156-4162. https://doi.org/10.1002/chem.201302997 63. Guru, M. M.; Punniyamurthy, T. J. Org. Chem. 2012, 77, 5063-5073. https://doi.org/10.1021/jo300592t 64. Wu, L.; Guo, S.; Wang, X.; Guo, Z.; Yao, G.; Lin, Q.; Wu, M. Tetrahedron Lett. 2015, 56, 2145-2148. https://doi.org/10.1016/j.tetlet.2015.03.019 65. Rao, K. M.; Gyum, K. W.; Kyungjae, M.; You, H. S. Eur. J. Org. Chem. 2016, 2016, 438-442. 66. Kim, H.-Y.; Kwak, S. H.; Lee, G.-H.; Gong, Y.-D. Tetrahedron 2014, 70, 8737-8743. https://doi.org/10.1016/j.tet.2014.09.023 Arkivoc 2018, vi, 0-0 Sambiagio, C. Page 33 ©ARKAT USA, Inc 67. Yu, W.; Huang, Y.; Li, J.; Tang, X.; Wu, W.; Jiang, H. J. Org. Chem. 2018, 83, 9334-9343. https://doi.org/10.1021/acs.joc.8b01292 68. Wang, Z.; Kuninobu, Y.; Kanai, M. J. Org. Chem. 2013, 78, 7337-7342. https://doi.org/10.1021/jo401056g 69. Doğan, Ş. D. Tetrahedron 2017, 73, 2217-2224. https://doi.org/10.1016/j.tet.2017.02.063 70. Fan, W.; Li, Q.; Li, Y.; Sun, H.; Jiang, B.; Li, G. Org. Lett. 2016, 18, 1258-1261. https://doi.org/10.1021/acs.orglett.6b00079 71. Fang, Y.; Zhu, Z.-L.; Xu, P.; Wang, S.-Y.; Ji, S.-J. Green Chem. 2017, 19, 1613-1618. https://doi.org/10.1039/C6GC03521C 72. Henry, M. C.; Mostafa, M. A. B.; Sutherland, A. Synthesis 2017, 49, 4586-4598. https://doi.org/10.1055/s-0036-1588536 73. Brasche, G.; Buchwald, S. L. Angew. Chem. Int. Ed. 2008, 47, 1932-1934. https://doi.org/10.1002/anie.200705420 74. Wang, H.; Wang, Y.; Peng, C.; Zhang, J.; Zhu, Q. J. Am. Chem. Soc. 2010, 132, 13217-13219. https://doi.org/10.1021/ja1067993 75. Masters, K.-S.; Rauws, T. R. M.; Yadav, A. K.; Herrebout, W. A.; Van der Veken, B.; Maes, B. U. W. Chem. Eur. J. 2011, 17, 6315-6320. https://doi.org/10.1002/chem.201100574 76. Lu, J.; Jin, Y.; Liu, H.; Jiang, Y.; Fu, H. Org. Lett. 2011, 13, 3694-3697. https://doi.org/10.1021/ol2013395 77. Wang, X.; Jin, Y.; Zhao, Y.; Zhu, L.; Fu, H. Org. Lett. 2012, 14, 452-455. https://doi.org/10.1021/ol202884z 78. Xu , H.; Fu , H., Chem. Eur. J. 2012, 18, 1180-1186. https://doi.org/10.1002/chem.201102794 79. Maes, J.; Rauws, T. R. M.; Maes, B. U. W., Chem. Eur. J. 2013, 19, 9137-9141. https://doi.org/10.1002/chem.201301248 80. Morel, B.; Franck, P.; Bidange, J.; Sergeyev, S.; Smith, D. A.; Moseley, J. D.; Maes, B. U. W. ChemSusChem 2017, 10, 624-628. https://doi.org/10.1002/cssc.201601483 81. Hu, J.; Chen, S.; Sun, Y.; Yang, J.; Rao, Y., Org. Lett. 2012, 14, 5030-5033. https://doi.org/10.1021/ol3022353 82. Li, X.; He, L.; Chen, H.; Wu, W.; Jiang, H., J. Org. Chem. 2013, 78, 3636-3646. https://doi.org/10.1021/jo400162d 83. Lan, C.; Tian, Z.; Liang, X.; Gao, M.; Liu, W.; An, Y.; Fu, W.; Jiao, G.; Xiao, J.; Xu, B. Adv. Synth. Catal. 2017, 359, 3735-3740. https://doi.org/10.1002/adsc.201700669 84. Hu, J.; Xu, H.; Nie, P.; Xie, X.; Nie, Z.; Rao, Y., Chem. Eur. J. 2014, 20, 3932-3938. https://doi.org/10.1002/chem.201304923 85. Yu, J.; Lim, J. W.; Kim, S. Y.; Kim, J.; Kim, J. N., Tetrahedron Lett. 2015, 56, 1432-1436. https://doi.org/10.1016/j.tetlet.2015.01.183 86. Ueda, S.; Nagasawa, H., Angew. Chem. Int. Ed. 2008, 47, 6411-6413. https://doi.org/10.1002/anie.200801240 87. Ueda, S.; Nagasawa, H., J. Org. Chem. 2009, 74, 4272-4277. Arkivoc 2018, vi, 0-0 Sambiagio, C. Page 34 ©ARKAT USA, Inc https://doi.org/10.1021/jo900513z 88. Cho, Y. H.; Lee, C.-Y.; Ha, D.-C.; Cheon, C.-H. Adv. Synth. Catal. 2012, 354, 2992-2996. https://doi.org/10.1002/adsc.201200684 89. Yang, D.; Yan, K.; Wei, W.; Tian, L.; Shuai, Y.; Li, R.; You, J.; Wang, H. Asian J. Org. Chem. 2014, 3, 969-973. https://doi.org/10.1002/ajoc.201402085 90. Chen, X.; Ji, F.; Zhao, Y.; Liu, Y.; Zhou, Y.; Chen, T.; Yin, S.-F. Adv. Synth. Catal. 2015, 357, 2924-2930. https://doi.org/10.1002/adsc.201500515 91. Sun, Y.; Abdukader, A.; Zhang, H.; Yang, W.; Liu, C., RSC Advances 2017, 7, 55786-55789. https://doi.org/10.1039/C7RA11436B 92. Guin, S.; Ghosh, T.; Rout, S. K.; Banerjee, A.; Patel, B. K. Org. Lett. 2011, 13, 5976-5979. https://doi.org/10.1021/ol202409r 93. Xu, Y.; Li, B.; Zhang, X.; Fan, X., J. Org. Chem. 2017, 82, 9637-9646. https://doi.org/10.1021/acs.joc.7b01683 94. Kamarul, M.; Avik, K. B.; Subhajit, M.; Adinath, M.; Alakananda, H., Adv. Synth. Catal. 2014, 356, 1105-1112. https://doi.org/10.1002/adsc.201300900 95. Kaswan, P.; Pericherla, K.; Saini, H. K.; Kumar, A., RSC Advances 2015, 5, 3670-3677. https://doi.org/10.1039/C4RA13056A 96. Bharate, J. B.; Abbat, S.; Bharatam, P. V.; Vishwakarma, R. A.; Bharate, S. B., Org. Biomol. Chem. 2015, 13, 7790-7794. https://doi.org/10.1039/C5OB00776C 97. Hariss, L.; Hadir, K. B.; El-Masri, M.; Roisnel, T.; Grée, R.; Hachem, A., Beilstein J. Org. Chem. 2017, 13, 2115-2121. https://doi.org/10.3762/bjoc.13.208 98. Zhu, Y.; Li, C.; Zhang, J.; She, M.; Sun, W.; Wan, K.; Wang, Y.; Yin, B.; Liu, P.; Li, J. Org. Lett. 2015, 17, 3872-3875. https://doi.org/10.1021/acs.orglett.5b01854 99. Li, J. J.; Song, C.; Cui, D.-M.; Zhang, C., Org. Biomol. Chem. 2017, 15, 5564-5570. https://doi.org/10.1039/C7OB01018D 100. Nguyen, T. B.; Corbin, M.; Retailleau, P.; Ermolenko, L.; Al-Mourabit, A., Org. Lett. 2015, 17, 4956-4959. https://doi.org/10.1021/acs.orglett.5b02340 101. Yadav, S.; Srivastava, M.; Rai, P.; Tripathi, B. P.; Mishra, A.; Singh, J.; Singh, J., New J. Chem. 2016, 40, 9694-9701. https://doi.org/10.1039/C6NJ02365G 102. Kuram, M. R.; Bhanuchandra, M.; Sahoo, A. K. Angew. Chem. Int. Ed. 2013, 52, 4607-4612. https://doi.org/10.1002/anie.201210217 103. Zhu, R.; Wei, J.; Shi, Z., Chem. Sci. 2013, 4, 3706-3711. https://doi.org/10.1039/c3sc51489g 104. Liu, G.; Shen, Y.; Zhou, Z.; Lu, X., Angew. Chem. Int. Ed. 2013, 52, 6033-6037. https://doi.org/10.1002/anie.201300881 105. Zeng, W.; Wu, W.; Jiang, H.; Huang, L.; Sun, Y.; Chen, Z.; Li, X., Chem. Comm. 2013, 49, 6611-6613. https://doi.org/10.1039/c3cc42326c 106. Liao, J.; Guo, P.; Chen, Q., Catal. Commun. 2016, 77, 22-25. https://doi.org/10.1016/j.catcom.2016.01.009 107. Sreenivasulu, C.; Gopi Krishna Reddy, A.; Satyanarayana, G., Org. Chem. Front. 2017, 4, 972-977. https://doi.org/10.1039/C6QO00531D 108. Yamane, S.; Hinoue, T.; Usuki, Y.; Itazaki, M.; Nakazawa, H.; Hayashi, Y.; Kawauchi, S.; Miura, M.; Satoh, T. Chem. Eur. J. 2018, 24, 7852-7855. Arkivoc 2018, vi, 0-0 Sambiagio, C. Page 35 ©ARKAT USA, Inc https://doi.org/10.1002/chem.201801245 109. Gao, Y.; Yin, M.; Wu, W.; Huang, H.; Jiang, H., Adv. Synth. Catal. 2013, 355, 2263-2273. https://doi.org/10.1002/adsc.201300157 110. Hu, H.; Li, G.; Hu, W.; Liu, Y.; Wang, X.; Kan, Y.; Ji, M., Org. Lett. 2015, 17, 1114-1117. https://doi.org/10.1021/ol503681n 111. Li, X.; Huang, L.; Chen, H.; Wu, W.; Huang, H.; Jiang, H., Chem. Sci. 2012, 3, 3463-3467. https://doi.org/10.1039/c2sc21041j Author’s Biography Carlo Sambiagio studied chemistry at the University of Cagliari (IT), acquiring his Master's degree in 2011. He then moved to the University of Leeds (UK), completing his PhD on Cu-catalyzed cross-couplings in 2015, under the supervision of Prof. Patrick McGowan. After that he spent two years as a postdoc in Prof. Bert Maes’s group in Antwerp (BE), mostly exploring aerobic oxidation chemistry. Carlo has recently moved to the Eindhoven University of Technology (NL) as a EU Marie Curie fellow to work on flow chemistry with Prof. Timothy Noël. His interests generally include metal-catalyzed transformations and the development of new methodologies for organic synthesis. View publication stats | RESEARCH GATE |
Research ArticleSynthesis, Characterization, X-RayCrystallography, and Antileishmanial Activities ofN-Linked and O-Linked GlycopyranosidesHaroon ur Rashid,1 Sher Wali Khan,2 Momin Khan,3,4 Akhtar Nadhman,5Noor Rehman,2 Muhammad Tariq,2 and Sammer Yousuf61Department of Chemistry, Sarhad University of Science and Information Technology, Peshawar, Khyber Pakhtunkhwa, Pakistan2Department of Chemistry, Shaheed BB University, Sheringal, Upper Dir, Khyber Pakhtunkhwa, Pakistan3Department of Biotechnology, Quaid-I-Azam University, Islamabad 45320, Pakistan4Department of Microbiology, Institute of Basic Medical Sciences, Khyber Medical University, Peshawar, Pakistan5Institute of Integrative Biosciences (IIB), CECOS University of Information Technology and Emerging Sciences, Peshawar,Khyber Pakhtunkhwa, Pakistan6H. E. J. Research Institute of Chemistry, International Center for Chemical and Biological Sciences, University of Karachi,Karachi 75270, PakistanCorrespondence should be addressed to Sher Wali Khan; sherwalikhn@yahoo.co.ukReceived 1 August 2017; Revised 14 November 2017; Accepted 7 December 2017; Published 30 January 2018Academic Editor: Vinod Kumar TiwariCopyright © 2018 Haroon ur Rashid et al. This is an open access article distributed under the Creative Commons AttributionLicense, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properlycited.Novel 𝑁-linked 5a–e and 𝑂-linked glycopyranosides 7a–e were synthesized in high yield from commercially available L-tartaricacid containing two asymmetric centers and 𝐶2 axis of symmetry. The compound L-tartaric acid was completely protectedand then partially hydrolyzed to get the monoester, which upon treatment with different amino and hydroxyl derivatives ofglycopyranoses gave the desired amides and esters. The synthesized derivatives were purified by chromatography and characterizedby spectroanalytical techniques. The structure of compound 7c in the series was supported by X-ray analysis. Leishmanicidalactivities of compounds 5a–e and 7a–e were investigated which showed moderate to good activities.1. IntroductionLeishmania has long been known to human beings especiallyin African and Indian populations since the mid eighteenthcentury [1–3]. In 1903, Leishman and Donovan discoveredthe protozoan in spleen tissues of patients in India, whichis now known as Leishmania donovani [4, 5]. Nicolle andComte reported infection by Leishmania in dogs in 1908 [6].Leishmaniasis is a vector-borne disease carried by sand flyand caused by protozoa of the genus Leishmania. Recentreports confirm that people in about 88 countries are affectedby Leishmania and 350 million are at risk, and more or less2 million new cases are reported each year [7]. In spite ofthe socioeconomical significance of this tropical infectiousdisease, the discovery of new potential drugs against it isunderway [8, 9].Glycolipid and glycoprotein oligosaccharides present oncell surfaces are known to play an important role in variousbiological processes, that is, cellular recognition, adhesion,tumor metastasis, and cell growth regulation [10]. Onemajor class of glycoprotein oligosaccharides consists of 𝑁-oligosaccharides linked to asparagine by an amide bond [11,12]. Amides are usually prepared by thermolysis of carboxylicacids with amines, by coupling of carboxylic acids andamines in the presence of a coupling reagent [13] or by priorconversion of the carboxylic acid into a more electrophilicderivative [14]. The concept to add a chiral aglycone moietysuch as tartaric acid to prepare glycosides is an active areaHindawiJournal of ChemistryVolume 2018, Article ID 9648710, 9 pageshttps://doi.org/10.1155/2018/96487102Journal of Chemistryof research. L-tartaric acid is cheap and readily availablestarting material; it has two chiral centers with two carboxylicacid groups which can be easily derivatized. Derivativesof tartaric acid containing ester and amide linkages havepromisingly useful biological applications, for example, anti-inflammatory, antifungal, thrombin inhibitors, antimicrobial,and 𝛽-secretase inhibitory activities [15–20].Hence, glyco derivatives of tartaric acid have potentialas curative agents for various infectious diseases. Herein, wereport an efficient and scalable synthesis of 𝑁-linked (5a–e)and 𝑂-linked glycopyranosides (7a–e) through anomeric 𝑁-and 𝑂-acylation, respectively, using commercially availableL-tartaric acid as a starting material.2. ExperimentalAll chemicals were of highest purity available and used as sup-plied. D-glucose, D-mannose, D-galactose, D-glucosamine,and D-galactosamine hydrochloride were purchased fromFluka and Sigma Aldrich. Dry solvents like methanol,dichloromethane, chloroform, and n-hexane were obtainedby distillation using standard procedures or by passagethrough a column of anhydrous alumina using equipmentfrom anhydrous engineering (University of Bristol, UK)based on Grubbs’ design. Reactions under anhydrous con-ditions were carried out under nitrogen gas using three-waystopcock and rubber septa.Liquid reagents, solutions, or solvents were added viasyringe or cannula through rubber septa. Solid reagents wereadded via Schlenk type adapters. All reactions were moni-tored by TLC on Kieselgel 60 F254 (Merck); ethyl acetate/n-hexane and methanol/chloroform were used as eluent. Chro-matograms were detected under UV light (𝜆max 254 and365 nm) and by charring with 10% sulfuric acid in ethanol,ninhydrin, and vanillin, respectively. Column chromatogra-phy was performed using silica gel [Merck, 230−400 mesh(40−63 𝜇m)]. Extracts were concentrated under reducedpressure using both rotary evaporator (bath temperature upto 40∘C) at a pressure of either 15 mmHg (diaphragm pump)or 0.1 mmHg (oil pump), as appropriate and a high vacuumline at room temperature. Melting points were determined indegree Celsius (∘C) using Gallenkamp digital melting pointapparatus and are uncorrected. IR spectra were recorded ona Bruker IFS 66 (FT-IR), Nicolet 205 FT-IR; Nicolet 360smart orbit (ATR); Thermo Scientific Nicolet 6700 FT-IR andShimadzu Fourier Transform Infrared SpectrophotometerModel 270. Solid samples were taken in KBr pellets andoils were used in NaCl cells for recording their spectra.1HNMR spectra were recorded on an NMR Bruker apparatusat 300 MHz and Varian 400 MHz INOVA instrument. 13CNMR spectra were recorded on NMR Bruker apparatus at75 MHz and Varian 100 MHz INOVA instrument. Chemicalshifts are quoted in parts per million from SiMe4 or residualsolvent proton signals and coupling constants (𝐽) given inHertz. Multiplicities are abbreviated as b (broad), s (singlet),d (doublet), t (triplet), q (quartet), m (multiplet), or combina-tions thereof. Positive ion Matrix Assisted Laser DesorptionIonization Time-of-Flight (MALDI-TOF) mass spectra wererecorded using an HP-MALDI instrument using gentisic acidas matrix material. Optical rotations were measured on anATAGO, AP-100 Automatic polarimeter. Single crystal X-ray diffraction data were collected on a Bruker Smart APEXII, CCD detector diffractometer [21]. Data reductions wereperformed using the SAINT program. The structures weresolved by direct methods [22] and refined by full-matrix leastsquares on F2 by using the SHELXTL-PC package [23]. Thefigures were plotted with the aid of ORTEP program [24].2.1. General Procedure for the Synthesis of N-Linked Gly-copyranosides (5a–e). In a round bottom flask (100 mL),monoester 1 (0.61 g, 3 mmol), EDC (0.74 g, 3 mmol × 1.3eq.), and catalytic amount of DMAP were placed undernitrogen. CH2Cl2 (30 mL) was added as a solvent. After30 min glycopyranosyl amine 4a–e (3 mmol) was added andthe mixture was stirred for 18 h. The byproduct urea wasremoved by extraction of the reaction mixture with ethylacetate or chloroform and water (30 mL × 2). The crudeproduct was purified by column chromatography using ethylacetate : hexane (3 : 7) as eluent.2.2. (2R,3R,5R,6R)-2-(Acetoxymethyl)-6-((4R,5R)-5-(methox-ycarbonyl)-2,2-dimethyl-1,3-dioxolane-4-carboxamido)tetra-hydro-2H-pyran-3,4,5-triyl Triacetate 5a. Yield: 72%. Color-less oil [𝛼]25D = 12.5. (IR 𝜐 cm−1): 3347 (NH), 2955 (CH),1742 (COOCH3), 1701 (CONH), 1299 (C-O-C). 1HNMR(400 MHz, CDCl3): 𝛿 (ppm): 7.18 (d, J = 9.3 Hz, NH), 5.23,5.18, 5.05, 4.96 (pseudo t, J ∼ 9.2 Hz in each, CH), 4.65 (d,J = 5.4 Hz, CH), 4.59 (d, J = 5.4 Hz, CH), 4.30 (dd, J = 12.5,3.6 Hz, CH), 4.21 (ddd, J = 10.4, 3.6, 2.1 Hz, CH), 3.95 (dd, J =12.5, 2.1 Hz, CH), 3.75 (s, OCH3), 1.94, 1.92, 1.91, 1.89 (s, CH3),1.56 (s, CH3), 1.50 (s, CH3). 13C NMR (100 MHz, CDCl3): 𝛿(ppm): 171.2 (CONH), 170.7 (COOCH3), 170.0, 169.9, 169.5,169.2 (CO), 113.3 (qt C), 81.8 (CH), 76.3 (CH), 75.2 (CH), 73.3,72.5, 70.0, 67.8 (CH), 61.2 (CH), 52.4 (OCH3), 25.9, (CH3),25.2 (CH3), 20.2, 20.1 (3), (CH3). HRMS-ESI for C22H31NO14Na: [M + Na]+ calcd: 556.1745, found: 556.1621. Anal. Calc.For C22H31NO14: C, 49.53; H, 5.86; N, 2.63; O, 41.99 FoundC, 49.43; H, 5.56; N, 2.78; O, 42.10.2.3. (2R,3R,5R,6R)-2-(Acetoxymethyl)-6-((4R,5R)-5-(methox-ycarbonyl)-2,2-dimethyl-1,3-dioxolane-4-carboxamido)tetra-hydro-2H-pyran-3,4,5-triyl Triacetate 5b. Yield: 69%. Color-less oil [𝛼]25D = 13.3. (IR 𝜐 cm−1): 3337 (NH), 2925 (CH),1732 (COOCH3), 1703 (CONH), 1292 (C-O-C). 1HNMR(400 MHz, CDCl3): 𝛿 (ppm): 7.47 (d, J = 9.3 Hz, NH), 5.34,5.31, 5.24 (pseudo t, J = 10.0, Hz in each, CH), 4.95 (t, J =3.9 Hz, CH), 4.93 (d, J = 5.5 Hz, CH), 4.80 (d, J = 5.5 Hz,CH), 4.28 (dd, J = 12.4, 3.5 Hz, CH), 4.10 (ddd, J = 10.3, 3.6,2.1 Hz, CH), 3.95 (dd, J = 12.4, 3.5 Hz, CH), 3.77 (s, OCH3),2.34, 2.23, 2.13, 1.99 (s, CH3), 1.52 (s, CH3), 1.46 (s, CH3).13C NMR (100 MHz, CDCl3): 𝛿 (ppm): 172.2 (CONH), 170.9(COOCH3), 170.1, 169.6, 169.3, 169.1 (CO), 114.3 (qt C), 81.4(CH), 76.5 (CH), 75.1 (CH), 73.1, 72.3, 70.1, 67.9 (CH), 61.8(CH), 52.9 (OCH3), 25.5 (CH3), 25.1 (CH3), 20.1, 20.0 (3)(CH3). HRMS-ESI for C22H31NO14 Na: [M + Na]+ calcd:556.1745, found: 556.1627. Anal. Calc. For C22H31NO14: C,Journal of Chemistry349.53; H, 5.86; N, 2.63; O, 41.99 Found C, 49.43; H, 5.56; N,2.78; O, 42.10.2.4. (2R,3R,5R,6R)-2-(Acetoxymethyl)-6-((4R,5R)-5-(methox-ycarbonyl)-2,2-dimethyl-1,3-dioxolane-4-carboxamido)tetra-hydro-2H-pyran-3,4,5-triyl Triacetate 5c. Yield: 65%. Color-less oil [𝛼]25D = 27.1. (IR 𝜐 cm−1): 3339 (NH), 2928 (CH),1736 (COOCH3), 1700 (CONH), 1290 (C-O-C). 1HNMR(400 MHz, CDCl3): 𝛿 (ppm): 7.37 (d, J = 9.3 Hz, NH), 5.42,5.40, 5.36 (pseudo t, J ∼ 6.5 Hz in each, CH), 4.98 (t, J =11.9 Hz, CH), 4.97 (d, J = 5.6 Hz, CH), 4.90 (d, J = 5.6 Hz,CH), 4.48 (dd, J = 12.7, 3.4 Hz, CH), 4.12 (ddd, J = 10.4, 3.6,2.1 Hz, CH), 3.92 (dd, J = 12.4, 3.5 Hz, CH), 3.84 (s, OCH3),2.11, 2.10, 2.09, 1.99 (s, CH3), 1.54 (s, CH3), 1.48 (s, CH3).13C NMR (100 MHz, CDCl3): 𝛿 (ppm): 171.8 (CONH), 170.2(COOCH3), 170.3, 169.5, 169.4, 169.2 (CO), 114.2 (qt C), 81.5(CH), 76.1 (CH), 75.2 (CH), 73.3, 72.2, 70.1, 67.5 (CH), 61.9(CH), 52.5 (OCH3), 26.5 (CH3), 25.7 (CH3), 20.4, 19.5 (3)(CH3). HRMS-ESI for C22H31NO14 Na: [M + Na]+ calcd:556.1745, found: 556.1643. Anal. Calc. For C22H31NO14: C,49.53; H, 5.86; N, 2.63; O, 41.99 Found C, 49.43; H, 5.56; N,2.78; O, 42.10.2.5. (2R,3S,5R,6R)-5-Acetamido-2-(acetoxymethyl)-6-((4R,5R)-5-(methoxycarbonyl)-2,2-dimethyl-1,3-dioxolane-4-car-boxamido)tetrahydro-2H-pyran-3,4-diyl Diacetate 5d. Yield:67%. Gel like. [𝛼]25D = 43.3. (IR 𝜐 cm−1): 3343 (NH), 2948(CH), 1746 (COOCH3), 1710 (CONH), 1293 (C-O-C).1HNMR (400 MHz, CDCl3): 𝛿 (ppm): 7.69 (d, J = 9.5 Hz,NH), 7.57 (d, J = 9.3 Hz, NH), 5.43 (t, J = 10.0 Hz, 1H), 5.41-5.32(m, CH), 5.37, 4.99 (pseudo t, J ∼ 10.0 Hz in each, CH), 4.96(d, J = 5.5 Hz, CH), 4.94 (d, J = 5.5 Hz, CH), 4.44 (dd, J = 12.7,3.4 Hz, CH), 4.22 (ddd, J = 10.4, 3.6, 2.1 Hz, CH), 3.91 (dd, J =12.4, 3.5 Hz, CH), 3.86 (s, OCH3), 2.21, 2.20, 2.19, 2.13 (s, CH3),1.57 (s, CH3), 1.49 (s, CH3). 13C NMR (100 MHz, CDCl3): 𝛿(ppm): 172.8 (CONH), 171.2 (COOCH3), 170.2, 169.8, 169.5,169.1 (CO), 114.1 (qt C), 81.8 (CH), 76.1 (CH), 75.3 (CH), 73.2,72.1, 70.0, 67.8 (CH), 61.4 (CH), 52.2 (OCH3), 26.2 (CH3), 25.8(CH3), 20.3, 19.3 (3) (CH3). HRMS-ESI for C22H32N2O13 Na:[M + Na]+ calcd: 555.1904, found: 555.1619. Anal. Calc. ForC22H32N2O13: C, 49.62; H, 6.06; N, 5.26; O, 39.06 Found C,49.32; H, 6.16; N, 5.21; O, 39.16.2.6. (2R,3S,5R,6R)-5-Acetamido-2-(acetoxymethyl)-6-((4R,5R)-5-(methoxycarbonyl)-2,2-dimethyl-1,3-dioxolane-4-car-boxamido)tetrahydro-2H-pyran-3,4-diyl Diacetate 5e. Yield:68%. Gel like. [𝛼]25D = 47.1. (IR 𝜐 cm−1): 3341 (NH), 2942 (CH),1745 (COOCH3), 1711 (CONH), 1290 (C-O-C). 1HNMR(400 MHz, CDCl3): 𝛿 (ppm): 7.67 (d, J = 9.6 Hz, NH), 7.55(d, J = 9.3 Hz, NH), 5.41 (t, J = 10.8 Hz, CH), 5.40 (m, CH),5.38, 4.97 (pseudo t, J ∼ 6.9 Hz in each, CH), 4.93 (d, J =5.6 Hz, CH), 4.91 (d, J = 5.5 Hz, CH), 4.41 (dd, J = 12.7, 3.4 Hz,CH), 4.21 (ddd, J = 10.4, 3.6, 2.1 Hz, CH), 3.90 (dd, J = 12.4,3.5 Hz, CH), 3.83 (s, OCH3), 2.23, 2.22, 2.20, 2.16 (s, CH3),1.56 (s, CH3), 1.46 (s, CH3). 13C NMR (100 MHz, CDCl3): 𝛿(ppm): 172.2 (CONH), 171.3 (COOCH3), 170.5, 169.5, 169.2,169.1 (CO), 113.4 (qt C), 81.6 (CH), 76.4 (CH), 75.2 (CH), 73.4,72.3, 70.1, 67.3, (CH), 62.4 (CH), 53.2 (OCH3), 26.2 (CH3),25.7 (CH3), 20.7, 19.5 (3) (CH3). HRMS-ESI for C22H31NO14Na: [M + Na]+ calcd: 555.1904, found: 555.1719. Anal. Calc.For C22H32N2O13: C, 49.62; H, 6.06; N, 5.26; O, 39.06 FoundC, 49.32; H, 6.16; N, 5.21; O, 39.16.2.7. General Procedure for the Synthesis of 𝑂-Linked Glycopy-ranosides (7a–e). Hydrazinium acetate (1.10 g, 12 mmol) wasadded to a stirred solution of anomeric mixture of acetylatedmonosaccharides 1a–e (3.12 g, 8 mmol) in DMF (40 mL). Thereaction mixture was heated at 55∘C for 30 min under anitrogen atmosphere. After completion, the reaction mixturewas diluted with H2O (100 mL) and extracted with EtOAc(40 mL × 3). The combined organic layers were washed withwater and brine (40 × 2), dried over anhydrous MgSO4.The solvent was removed under reduced pressure and thecrude product was used without further purification. To theresulting hemiacetal (1.04 g, 3 mmol), monoester 1 (0.61 g,3 mmol), EDC (0.74 g, 3 mmol × 1.3 eq.), and catalytic amountof DMAP were added under nitrogen with CH2Cl2 (30 mL).The reaction mixture was stirred for 18 h; the byproduct ureawas removed by extraction of the reaction mixture with ethylacetate or chloroform and water (30 mL × 3). The crudeproduct was purified by column chromatography using ethylacetate: n-hexane (3 : 7) as eluent.2.8. (4R,5R)-4-Methyl-5-((2S,3R,5R,6R)-3,4,5-triacetoxy-6-(acetoxymethyl)tetrahydro-2H-pyran-2-yl)-2,2-dimethyl-1,3-dioxolane-4,5-dicarboxylate 7a. Yield: 72%. Colorless oil[𝛼]25D = 26.5. (IR 𝜐 cm−1): 2991 (CH), 1742 (CO), 1366 (C-O-C). 1HNMR (400 MHz, CDCl3): 𝛿 (ppm): 6.45 (d, J = 4.0 Hz,1H, H-1𝛼), 5.56 (d, J = 8.0 Hz, 1H, H-1𝛽), 5.48, 5.45, 4.93(pseudo t, J ∼ 9.8 Hz in each, CH), 4.91 (d, J = 6.3 Hz, CH),4.88 (d, J = 6.3 Hz, CH), 4.32 (dd, J = 12.5, 3.6 Hz, CH), 4.12(ddd, J = 10.4, 3.6, 2.1 Hz, CH), 3.93 (dd, J = 12.5, 2.1 Hz, CH),3.86 (s, OCH3), 2.19, 2.17, 2.12, 2.04 (s, CH3), 1.55 (s, CH3),1.46 (s, CH3). 13C NMR (100 MHz, CDCl3): 𝛿 (ppm): 172.6(COOCH3), 171.2, 169.5, 169.3, 169.0 (CO), 113.4 (qt C), 92.3(CH), 92.1 (CH), 76.3 (CH), 75.4 (CH), 74.2, 73.3, 72.1, 70.3(CH), 56.6 (CH), 53.4 (OCH3), 26.6 (CH3), 25.9 (CH3), 20.3(2), 19.8, 19.3 (CH3). HRMS-ESI for C22H30O15 Na: [M + Na]+calcd: 557.1585, found: 557.1452. Anal. Calc. For C22H30O15: C,49.44; H, 5.66; O, 44.90 Found C, 49.39; H, 5.70; O, 44.81.2.9. (4R,5R)-4-Methyl-5-((2S,3R,5R,6R)-3,4,5-triacetoxy-6-(acetoxymethyl)tetrahydro-2H-pyran-2-yl)-2,2-dimethyl-1,3-dioxolane-4,5-dicarboxylate 7b. Yield: 74%. Colorless gel[𝛼]25D = 21.6. (IR 𝜐 cm−1): 2993 (CH), 1744 (CO), 1361 (C-O-C). 1HNMR (400 MHz, CDCl3): 𝛿 (ppm): 6.47 (d, J = 3.9 Hz,1H, H-1𝛼), 5.53 (d, J = 8.3 Hz, 1H, H-1𝛽), 5.47, 4.93 (pseudot, J ∼ 6.9 Hz in each, CH), 5.50 (t, J = 8.9 Hz, CH), 4.94 (d,J = 6.3 Hz, CH), 4.88 (d, J = 6.3 Hz, CH), 4.27 (dd, J = 12.5,3.6 Hz, CH), 4.14 (ddd, J = 10.4, 3.6, 2.1 Hz, CH), 3.96 (dd, J =12.5, 2.1 Hz, CH), 3.86 (s, OCH3), 2.19, 2.17, 2.12, 2.04 (s, CH3),1.55 (s, CH3), 1.45 (s, CH3). 13C NMR (100 MHz, CDCl3): 𝛿(ppm): 172.3 (COOCH3), 171.7, 169.4, 169.3, 169.1 (CO), 113.3(qt C), 92.7 (CH), 92.4 (CH), 76.2 (CH), 75.6 (CH), 74.4, 73.5,4Journal of Chemistry72.2, 70.2, (CH), 56.5 (CH), 53.2 (OCH3), 25.6 (CH3), 24.9(CH3), 20.6 (2), 19.7, 19.5 (CH3). HRMS-ESI for C22H30O15Na: [M + Na]+ calcd: 557.1585, found: 557.1466. Anal. Calc.For C22H30O15: C, 49.44; H, 5.66; O, 44.90 Found C, 49.39;H, 5.70; O, 44.81.2.10. (4R,5R)-4-Methyl-5-((2R,3R,5R,6R)-3,4,5-triacetoxy-6-(acetoxymethyl)tetrahydro-2H-pyran-2-yl)-2,2-dimethyl-1,3-dioxolane-4,5-dicarboxylate 7c. Yield: 70%. White foam m.p= 151–153∘C. [𝛼]25D = 17.3. (IR 𝜐 cm−1): 2995 (CH), 1745 (CO),1336 (C-O-C). 1HNMR (400 MHz, CDCl3): 𝛿 (ppm): 6.42(d, J = 3.7 Hz, 1H, H-1𝛼), 5.51 (d, J = 4.0 Hz, 1H, H-1𝛽), 5.46,5.43 (pseudo t, J ∼ 5.6 Hz in each, CH), 4.96 (t, J = 9.8 Hz,CH), 4.93 (d, J = 5.3 Hz, CH), 4.89 (d, J = 5.3 Hz, CH), 4.26(dd, J = 12.5, 3.6 Hz, CH), 4.13 (ddd, J = 10.4, 3.6, 2.1 Hz,CH), 3.93 (dd, J = 12.5, 2.1 Hz, CH), 3.76 (s, OCH3), 2.15,2.14, 2.12, 2.11 (s, CH3), 1.56 (s, CH3), 1.46 (s, CH3). 13C NMR(100 MHz, CDCl3): 𝛿 (ppm): 171.3 (COOCH3, 170.7, 169.6,169.3, 169.1 (CO), 113.4 (qt C), 92.4 (CH), 92.1 (CH), 76.1(CH), 75.6 (CH), 74.2, 73.2, 72.1, 70.5 (CH), 56.7 (CH), 53.3(OCH3), 26.6 (CH3), 25.8 (CH3), 20.2 (2), 19.8, 19.4 (CH3).HRMS-ESI for C22H30O15 Na: [M + Na]+ calcd: 557.1585,found: 557.1475. Anal. Calc. For C22H30O15: C, 49.44; H, 5.66;O, 44.90 Found C, 49.39; H, 5.70; O, 44.88.2.11. (4R,5R)-4-((2S,3R,5S,6R)-3-Acetamido-4,5-diacetoxy-6-(acetoxymethyl)tetrahydro-2H-pyran-2-yl)-5-methyl-2,2-dim-ethyl-1,3-dioxolane-4,5-dicarboxylate 7d. Yield: 67%. Color-less oil [𝛼]25D = 45.0. (IR 𝜐 cm−1): 3342 (NH), 2981 (CH), 1739(CO), 1360 (C-O-C). 1HNMR (400 MHz, CDCl3): 𝛿 (ppm):7.87 (d, J = 9.5 Hz, NH), 6.49 (d, J = 3.7 Hz, 1H, H-1𝛼), 5.57(d, J = 8.3 Hz, 1H, H-1𝛽), 5.46 (m, CH) 5.43, 4.96 (pseudo t,J ∼ 7.4 Hz in each, CH), 4.94 (d, J = 6.3 Hz, CH), 4.83 (d, J =6.3 Hz, CH), 4.37 (dd, J = 12.5, 3.6 Hz, CH), 4.23 (ddd, J = 10.4,3.6, 2.1 Hz, CH), 3.93 (dd, J = 12.5, 2.1 Hz, CH), 3.87 (s, OCH3),2.13, 2.11, 2.10, 2.04 (s, CH3), 1.57 (s, CH3), 1.49 (s, CH3).13C NMR (100 MHz, CDCl3): 𝛿 (ppm): 171.6 (COOCH3),170.3, 169.3, 169.2, 169.0 (CO), 113.6 (qt C), 92.6 (CH), 92.2(CH), 76.2 (CH), 75.1 (CH), 74.5, 73.3, 72.6, 70.1 (CH), 57.9(CH), 53.5 (OCH3), 26.4 (CH3), 25.6 (CH3), 20.1 (2), 19.5,19.2, (CH3). HRMS-ESI for C22H31NO14 Na: [M + Na]+ calcd:556.1745, found: 556.1614. Anal. Calc. For C22H31NO14: C,49.53; H, 5.86; N, 2.63; O, 41.99 Found C, 49.55; H, 5.76; N,2.60; O, 41.90.2.12. (4R,5R)-4-((2S,3R,5S,6R)-3-Acetamido-4,5-diacetoxy-6-(acetoxymethyl)tetrahydro-2H-pyran-2-yl)-5-methyl-2,2-dim-ethyl-1,3-dioxolane-4,5-dicarboxylate 7e. Yield: 63%. Color-less oil [𝛼]25D = 33.3. (IR 𝜐 cm−1): 3348 (NH), 2990 (CH), 1741(CO), 1356 (C-O-C). 1HNMR (400 MHz, CDCl3): 𝛿 (ppm):7.86 (d, 𝐽 = 9.7 Hz, NH), 6.52 (d, 𝐽 = 3.9 Hz, 1H, H-1𝛼), 5.67(d, 𝐽 = 8.7 Hz, 1H, H-1𝛽), 5.46 (m, CH), 5.43, 4.96 (pseudot, 𝐽 ∼ 6.0 Hz in each, CH), 4.93 (d, J = 5.4 Hz, CH), 4.82(d, 𝐽 = 5.3 Hz, CH), 4.38 (dd, 𝐽 = 12.5, 3.6 Hz, CH), 4.33(ddd, 𝐽 = 10.4, 3.6, 2.1 Hz, CH), 3.93 (dd, 𝐽 = 12.5, 2.1 Hz,CH), 3.85 (s, OCH3), 2.14, 2.12, 2.10, 2.06 (s, CH3), 1.56 (s,CH3), 1.46 (s, CH3). 13C NMR (100 MHz, CDCl3): 𝛿 (ppm):171.4 (COOCH3), 170.3, 169.5, 169.4, 169.2 (CO), 114.4 (qt C),93.4 (CH), 92.7 (CH), 76.3 (CH), 75.2 (CH), 74.5, 73.4, 72.6,70.1 (CH), 58.7 (CH), 53.5 (OCH3), 26.6 (CH3), 25.6 (CH3),20.1 (2), 19.4, 19.1 (CH3). HRMS-ESI for C22H31NO14 Na:[M + Na]+ calcd: 556.1745, found: 556.1616. Anal. Calc. ForC22H31NO14: C, 49.53; H, 5.86; N, 2.63; O, 41.99 Found C,49.55; H, 5.81; N, 2.60; O, 41.91.3. Results and DiscussionAs outlined in synthetic Schemes 1 and 2, our proposedstrategy was to transform five different glycopyranoses toglycopyranosyl amines 4a–e which would serve as nucle-ophiles. In sugars, the anomeric position is comparativelymore reactive due to the presence of the endocyclic oxygenatom. Owing to its higher reactivity, microwave assistedKochetkov amination was tried. The Kochetkov aminationis used to introduce amino functionality at the anomericposition in unprotected sugars [25]. However, when glucosewas reacted under microwave with Kochetkov reagents,that is, ammonium carbonate in DMSO, a mixture ofamino and untreated glucose was obtained after prolongedlipophilization. In the case of glucosamine and galactosaminehydrochlorides, which already contain amino functionality,to ensure chemoselectivity in coupling with tartaric acid,a relative long route was adopted involving protecting theentire hydroxyl and amino functionalities on these sugars.Thus per acetylation [26], followed by anomeric bromination[27], azide formation [28] and reduction of azides affordedthe target acetylated glycopyranosyl amines 4a–e.Following the same coupling procedure with increasedtime (Table 2), 𝑂-linked glycopyranosides 7a–e were alsoprepared in 63–74% yields, starting from acetylated monosac-charides 1a–e, Scheme 3.It was observed that during per acetylation, a mixtureof both 𝛼 and 𝛽 anomers is formed in a 3 : 1 ratio. Inthe bromination step, the thermodynamically more stable𝛼 anomer was obtained exclusively. In the resultant bromoanalogue when left for longer time (two days), slight decom-position was observed. Therefore the product was used inthe next step immediately. The reaction of bromide 3a withsodium azide in DMF proceeded in low yield; however undersonication, the azide analogue was produced in excellentyield. The substitution involves inversion of configuration atthe anomeric site and thus the 𝛼-glycopyranosyl bromidegave the corresponding 𝛽-glycopyranosyl azide. The azideintermediates were reduced immediately under catalytichydrogenation in the next step. The hydrogenation reactionfurnished the desired amines in quantitative yield, whichwere stable when completely dried under vacuum. Once aseries of glycopyranosyl amines 4a–e were generated, thenext step was to synthesize the monoester 1 starting fromcommercially available L-tartaric acid. Both the hydroxyl andcarboxyl groups were protected followed by partial hydrolysiswith one equivalent of lithium hydroxide; however the yieldwas variable. By the use of a stronger base sodium hydroxide,the yield was improved. It is important to mention that theformation of monoester 1 from the diester does not dependJournal of Chemistry5OHOOHOHOAcOOAcOAcOAcOOAcBrOAcOOAcOAcOOAcOOOOOHOOAcOOAcHNOOOOO(d) D-Glucosamine·HCl(e) D-Galactosamine·HCl1a–1e2a–2e3a–3e4a–4e5a–5e(a)(b)(c)(d)(e)(a) D-Glucose(b) D-Galactose(c) D-Mannose+２2２2２2２2２2２2２1２1２1２1２1２1２3２3２3２4２3２3２3２4２4２4２4．3．（2２1, ２3 = H; ２2, ２4 = OH２1, ２4 = OH; ２2, ２3 = H２1, ２4 = H; ２2, ２3 = OH２1, ２3 = H; ２2 = OH; ２4 = N（2·HCl２1 = OH; ２2, ２3 = H; ２4 = N（2·HCl２4Scheme 1: Reagents and conditions: (a) Ac2O, pyridine, cat. DMAP, N2, r.t., overnight; (b) HBr/AcOH (30%), CH2Cl2, N2, r.t., 2 h; (c) NaN3,DMF, sonication, r.t., 15 min; (d) Pd/C (10%), H2, MeOH, 30 min; (e) EDC, CH2Cl2, cat. DMAP, N2, r.t., 18 h.OOOOOOOOOHHOOHOHOOOOOHO(1)(a)(b)Scheme 2: Reagents and conditions: (a) MeOH, H+, reflux, 2 h, 2-methoxypropene, CSA, DMF, N2, r.t., overnight; (b) NaOH (1 eq), MeOH,r.t., 2 h.on which ester group of the two is cleaved as both groupsare homotopic and therefore monohydrolysis of any of thesegroups will lead to identical molecules. The same strategy wasthen used for all of the compounds in the series.With glycopyranosyl amines and monoester in hand, itwas decided to use coupling reagents to form peptides 5a–e.Different coupling conditions were tried to generate the targetamides, Table 1. DCC in the presence of a DMAP catalystfurnished the target amide in low yield. EDC when used in 1.3equivalents for 12 h afforded the amides in good yield and thereaction was reproducible. Once coupling conditions wereoptimized, the same coupling conditions were used for therest of the series to furnish the respective amides, Table 1.The coupling reaction was high yielding and the com-pounds were purified by column chromatography. IR spec-tral data of the compounds 5a–e exhibited the appearanceof characteristic (NH) stretching bands in the range of3347–3331 cm−1 and disappearance of NH2 stretching. TwoC=O stretching bands ranging from 1742 to 1736 and 1701to 1692 cm−1 were assigned to ester and amide, respec-tively. 1HNMR spectral data shows characteristic doubletsfor the NH protons in the range of 7.69–7.18 ppm witha coupling constant of 9.6–9.3 Hz. The anomeric protonsappeared as pseudotriplets in the range of 5.42–5.36 ppmwith coupling constant of 12.2–6.5 Hz which confirms theformation of the 𝛽-anomer. In 13C NMR spectra, peaksranging from 172.8–171.2 ppm were assigned to C=O (amide)and 170.9–170.2 ppm to C=O (ester), respectively. Elementalanalysis of the glycopyranosyl amides 5a–e confirmed thestructures of the compounds. The exact mass of the com-pounds was confirmed by HRMS-ESI.The anomeric acetyl group was hydrolyzed selectivelyusing hydrazinium acetate, in DMF [29–31]; the resultinghemiacetal was washed with water and brine and used with-out further purification. To the crude hemiacetal, monoesterwas added using EDC as a coupling agent. When compound6a was coupled with monoester 1 using 1.3 eq. EDC, DMAPwith increased reaction time afforded compound 7a in 72%yield. Once the reaction conditions were optimized, the sameconditions were used for the rest of the series to furnish therespective esters, Table 2.The synthesized compounds were purified by columnchromatography and characterized by spectroscopic tech-niques. IR spectral data of the compounds 7a–e showedcharacteristic (C=O) stretching bands ranging from 1745 to1739 cm−1 in addition to C-O-C stretching in 1366–1356.The NH stretching bands 7d-e appeared in 3348–3342 cm−1.1HNMR spectral data showed doublets 7d-e, for the NHprotons in the range of 7.87–7.86 ppm with coupling con-stant of 9.7–9.5 Hz. Both 𝛼- and 𝛽-anomers were foundto be 1 : 3 except compound 7c where 𝛼 and 𝛽 ratio was1 : 1 as confirmed by 1HNMR spectroscopy. The anomericprotons appeared as doublets in the range of 5.67–5.43 ppmwith coupling constant of 8.3–8.0 Hz for the 𝛽-anomers6Journal of ChemistryOAcOOAcOOOOOHOOAcOOAcOOOOOOOAcOAcOOAcOH(a)(b)6a–6e7a–7e1a–1e２2２2２2２1２1２1２3２3２4２4２4２3+Scheme 3: Reagents and conditions: (a) N2H4/AcOH (1.5 eq.), DMF, 55∘C, N2, 30 min; (b) EDC, CH2Cl2, cat. DMAP, N2, rt, 18 h.Table 1: Different coupling condition for amide synthesis.EntryStarting materialReagents/conditionsProduct (isolated yield%)14aHA1.2 eq. DCC, cat. DMAP, CH2Cl2, r.t., 8 h4624aHA1 eq. EDC, cat. DMAP, CH2Cl2 r.t., 8 h3634aHA1.1 eq. EDC, cat. DMAP, CH2Cl2 r.t., 12 h7044aHA1.3 eq. EDC, cat. DMAP, CH2Cl2 r.t., 12 h7655aHA1.3 eq. EDC, cat. DMAP, CH2Cl2 r.t., 12 h6166aHA1.3 eq. EDC, cat. DMAP, CH2Cl2 r.t., 18 h79and in 6.52–6.42 ppm with coupling constant of 4.0–3.7 Hzfor the 𝛼-anomers, respectively. The coupling constant ofanomeric hydrogen for mannopyranosyl 7c was 4.0 Hz inboth the anomers. Finally the synthesis of glycopyrano-sides was supported by one of the crystal structure for𝛼-anomer, O-(2,3,4,6-tetra-O-acetyl-𝛼-D-mannopyranosyl)-(4R,5R)-2,2-dimethyl-1,3-dioxolane-4,5-dicarboxylate 7c byX-ray crystallography (Figure 1).Crystal system = monoclinic. Torsion angles for com-pound 7c are C5-O1-C1-O13 = −62.9(2)∘, C5-O1-C1-C2= 54.9(2)∘, C15-O13-C1-O1 = −76.4(2) (2)∘, C15-O13-C1-C2 = 160.28(19)∘, C13-O8-C2-C3 = −103.8(2)∘, C13-O8-C2-C1 = 136.00(19)∘, O1-C1-C2-O8 = 73.3(2)∘, O13-C1-C2-O8 = −165.60(16)∘, O1-C1-C2-C3 = −46.9(2)∘, O13-C1-C2-C3 = 74.2(2)∘, C11-O6-C3-C4 = −149.8(2)∘, C11-O6-C3-C2 = 88.7(3)∘, O8-C2-C3-O6 = 49.8(2)∘, C1-C2-C3-O6 = 167.17(17)∘, O8-C2-C3-C4 = −69.0(2)∘, C1-C2-C3-C4 = 48.4(3)∘, C9-O4-C4-C3 = −119.6(3)∘, C9-O4-C4-C5 = 123.8(3)∘, O6-C3-C4-O4 = 67.2(2)∘, C2-C3-C4-O4= −172.34(18)∘, O6-C3-C4-C5 = −176.24(17)∘, C2-C3-C4-C5 = −55.8(2)∘, C1-O1-C5-C6 = 176.43(19)∘, C1-O1-C5-C4 = −61.4(2)∘, O4-C4-C5-O1 = 176.08(18)∘, C3-C4-C5-O1 = 60.3(2)∘, O4-C4-C5-C6 = −65.2(3)∘, C3-C4-C5-C6= 179.0(2)∘, C7-O2-C6-C5 = −153.1(2)∘, O1-C5-C6-O2 =−68.2(3)∘, C4-C5-C6-O2 = 172.1(2)∘, C6-O2-C7-O3 = 1.9(5)∘,C6-O2-C7-C8 = −175.2(3)∘, C4-O4-C9-O5 = −7.1(6)∘, C4-O4-C9-C10 = 174.8(3)∘, C3-O6-C11-O7 = 3.7(4)∘, C3-O6-C11-C12 = −175.5(2)∘, C2-O8-C13-O9 = 6.4(3)∘, C2-O8-C13-C14= −171.9(2)∘, C1-O13-C15-O12 = −5.2(4)∘, C1-O13-C15-C16= 173.8(2)∘, C19-O15-C16-C15 = 131.8(4)∘, C19-O15-C16-C17= 11.0(5)∘, O12-C15-C16-O15 = −7.6(4)∘, O13-C15-C16-O15 =173.5(2)∘, O12-C15-C16-C17 = 108.9(4)∘, O13-C15-C16-C17 =−70.1(3)∘, C19-O14-C17-C18 = 130.6(5)∘, C19-O14-C17-C16 =11.5(6)∘, O15-C16-C17-O14 = −13.4(5)∘, C15-C16-C17-O14 =−133.2(4)∘, O15-C16-C17-C18 = −131.8(3)∘, C15-C16-C17-C18= 108.4(3)∘, C22-O11-C18-O10 = −2.3(9)∘, C22-O11-C18-C17= 178.0(4)∘, O14-C17-C18-O10 = −31.2(7)∘, C16-C17-C18-O10= 82.9(6)∘, O14-C17-C18-O11 = 148.5(4)∘, C16-C17-C18-O11 =C14C13C2C5C4C6C7C8C16C17C18C19C20C21C15C1C3C9C10C11C22C12O9O8O6O5O7O4O1O13O2O3O11O10O14O15O12Figure 1: Crystal structure of 𝛼-anomer, O-(2,3,4,6-tetra-O-acetyl-𝛼-D-mannopyranosyl)-(4R,5R)-2,2-dimethyl-1,3-dioxolane-4,5-dicarboxylate 7c.−97.4(4)∘, C17-O14-C19-O15 = −5.3(7)∘, C17-O14-C19-C21 =109.3(8)∘, C17-O14-C19-C20 = −123.6(6)∘, C16-O15-C19-O14= −4.3(6)∘, C16-O15-C19-C21 = −121.1(8)∘, and C16-O15-C19-C20 = 113.7(6)∘.Journal of Chemistry7Table 2: Coupling conditions and isolated yield for 5a–e and 7a–e.EntryStarting materialReagents/conditionsProduct (isolated yield%)𝛼 : 𝛽ratio[𝛼]25D14aHA1.3 eq. EDC, cat. DMAP, r.t., 12 h720 : 112.524bHA1.3 eq. EDC, cat. DMAP, r.t., 12 h690 : 113.334cHA1.3 eq. EDC, cat. DMAP, r.t., 12 h650 : 127.144dHA1.3 eq. EDC, cat. DMAP, r.t., 12 h670 : 143.354eHA1.3 eq. EDC, cat. DMAP, r.t., 12 h680 : 147.166aHA1.3 eq. EDC, cat. DMAP, r.t., 18 h721 : 326.576bHA1.3 eq. EDC, cat. DMAP, r.t., 18 h741 : 321.686cHA1.3 eq. EDC, cat. DMAP, r.t., 18 h701 : 117.396dHA1.3 eq. EDC, cat. DMAP, r.t., 18 h671 : 345.0106eHA1.3 eq. EDC, cat. DMAP, r.t., 18 h631 : 333.3𝑐 = 24 mg/2 mL in dichloromethane.Table 3: % inhibition of compounds 5a–e and 7a–e against L. tropica leishmania.Compound𝐿. 𝑡𝑟𝑜𝑝𝑖𝑐𝑎aCompound𝐿. 𝑡𝑟𝑜𝑝𝑖𝑐𝑎a5a0.65 ± 0.017a1.35 ± 0.045b0.68 ± 0.097b0.73 ± 0.125c0.81 ± 0.167c0.63 ± 0.185d0.68 ± 0.277d0.77 ± 0.085e0.89 ± 0.117e0.66 ± 0.18Standard drug IC50(𝜇g/mL ± SD)Amphotericin B0.56 ± 0.20aPercentage inhibition activity: 100 = (nonsignificant; 0.95−0.80 = low; 0.79−0.70 = moderate; 0.69−0.60 = good; below 0.59−0.56 = significant activity).4. Material and Methods4.1. Antileishmanial Activity. Antileishmanial activities ofcompounds 5a–e and 7a–e were assayed by Zhai et al.’smethod (1999) [32] using a preestablished culture of Leishma-nia. Triple-N media slants were overlaid with 199 media forleishmanial growth. NNN medium was prepared by mixing4 g of agar in 100 mL of distilled water. The mixture wasthen dissolved and sterilized by autoclaving at 121∘C and thenallowed to cool to 55∘C. 15–20 mL of defibrinated sheep bloodwas aseptically added to the mixture with gentle rolling withglass beads; 1 ampoule of gentamicin was mixed in blood andmixed with the agar mixture. Slopes of culture medium wereprepared by dispensing 2-3 mL of the blood agar mixture intosterile tubes that were then set in a slant position until theagar completely solidified. To prepare 199 medium differentconstituents were mixed in 1000 mL of distilled water. The pHwas adjusted to 7.4 and medium was filtered, sterilized, andkept at 37∘C for 24 h to check sterility. Preestablished cultureof Leishmania tropica KWH23 was inoculated in 199 mediumin Triple-N medium slants and incubated at 24∘C for 6-7 daysand the results are summarized in Table 3.5. ResultsAll the synthesized N-linked 5a–e and O-linked glycopyra-nosides 7a–e were tested for their antileishmanial activityusing Leishmania tropica KWH23 promastigotes for in vitroscreening. The results are shown in Table 3. Compounds 5a,5b, and 5d showed good activity while the compounds 5c and5e exhibited low activity.AmongthesynthesizedO-linkedglycopyranosides7a–7e, compounds 7c and 7e showed good activity; 7band 7d showed moderate activity. Compound 7a showednonsignificant activity.6. ConclusionMost of the synthesized compounds 5a–e and 7a–e showedmoderate to good activities against Leishmania tropicaKWH23 promastigotes. Compounds 5a, 5b, and 5d were themost active. Among the synthesized compounds, 7a–7e, 7c,and 7e were more active than 7a, 7b, and 7d. Drug resistancehas been reported in various species of Leishmania againstvarious antileishmanial drugs like antimonials, amphotericinB, pentamidine, miltefosine, and so forth. In this scenariothese synthetic compounds may be further explored forresearch in medicinal chemistry and drug designing. Thesecompounds may prove to be good candidates against leish-maniasis. A number of other new compounds can also besynthesized while following the same N- and O-acylationsynthetic routes. Such compounds can act as valuable gly-codrugs as many of the glycosides and their derivatives havebeen reported to be biologically active compounds.8Journal of ChemistryConflicts of InterestThe authors confirm that this article’s content has no conflictsof interest.AcknowledgmentsThis work was financially supported by Higher EducationCommission (HEC), Pakistan, under International ResearchSupport Initiative Program (IRSIP) and National ResearchProgram (NRPU) for Universities.Supplementary MaterialsSupplementary data include NMR (1H and 13C), IR, and MSspectra of the synthesized compounds (5a– 5d and 7a–7e).1HNMR of compound 5a. 13CNMR of compound 5a. ESI-MS of compound5a. FT-IR of compound5a. 1HNMRof compound5b.13CNMR of compound5b. ESI-MSof compound 5b. 1HNMR of compound 5c. 13CNMR ofcompound 5c. 1HNMR of compound 5d. 13CNMR of com-pound 5d. ESI-MS of compound 5d. 1HNMR of compound7a. 13CNMR of compound 7a. ESI-MS of compound 7a.1HNMR of compound 7b. 13CNMR of compound 7b. ESI-MS of compound 7b. ESI-MS of compound 7b backup.1HNMR of compound 7c, new. 1HNMR of compound 7c.13CNMR of compound7c.13CNMR of compound7b,new. ESI-MS of compound 7c. Mass analysis of compound7c. 1HNMR of compound 7d. 13CNMR of compound 7d.ESI-MS of compound 7d. Mass analysis of compound 7d.1HNMR of compound7e. 13CNMR of compound7e.(Supplementary Materials)References[1] N. Singh, B. B. Mishra, S. Bajpai, R. K. Singh, and V. K. Tiwari,“Natural product based leads to fight against leishmaniasis,”Bioorganic & Medicinal Chemistry, vol. 22, no. 1, pp. 18–45, 2014.[2] P. Dwivedi, K. B. Mishra, B. B. Mishra, N. Singh, R. K. Singh,and V. K. Tiwari, “Click inspired synthesis of antileishmanialtriazolyl O-benzylquercetin glycoconjugates,” GlycoconjugateJournal, vol. 32, no. 3-4, pp. 127–140, 2015.[3] B. B. Mishra, J. K. Gour, N. Kishore, R. K. Singh, V. Tripathi, andV. K. Tiwari, “An antileishmanial prenyloxy-naphthoquinonefrom roots of plumbago zeylanica,” Natural Product Research(Formerly Natural Product Letters), vol. 27, no. 4-5, pp. 480–485,2013.[4] W. B. Leishman, “On the possibility of the occurrence of try-panosomiasis in India,” British Medical Journal, vol. 1, no. 2213,pp. 1252–1254, 1903.[5] C. Donovan, “On the possibility of the occurrence oftrypanosomia-sis in India,” British Medical Journal, vol. 2,pp. 79–82, 1903.[6] C. Nicolle and C. Comte, “Origine canine du Kala-azar,” BulletinSo-ci´et´e Pathologie Exotique, vol. 1, pp. 299–301, 1908.[7] WHO (World Health Organisation), “Communicable diseasesur-veillance and response,” http://www.who.int/emc/diseases/leish/leisdis1.html.[8] S. L. Croft, “The current status of antiparasite chemotherapy,”Parasitology, vol. 114, no. 1, pp. S3–S15, 1997.[9] H. D. Engers, R. Bergquist, and F. Moddaber, “Progress onvaccines against parasite,” Developments in Biological Standard-ization, vol. 87, pp. 73–84, 1996.[10] H. Hojo and N. Nakahara, “Recent Progress in the Solid-phaseSynthesis of Glycopeptide,” Current Protein & Peptide Science,vol. 1, no. 1, pp. 23–48, 2000.[11] H. Lis and N. Sharon, “Protein glycosylation: structural andfunctional aspects,” European Journal of Biochemistry, vol. 218,no. 1, pp. 1–27, 1993.[12] A. Varki, “Biological roles of oligosaccharides: all of the theoriesare correct,” Glycobiology, vol. 3, no. 2, pp. 97–130, 1993.[13] S.-Y. Han and Y.-A. Kim, “Recent development of peptidecoupling reagents in organic synthesis,” Tetrahedron, vol. 60, no.11, pp. 2447–2467, 2004.[14] R. B. Greenwald, H. Zhao, J. Xia, and A. Martinez,“Poly(ethylene glycol) transport forms of vancomycin: along-lived continuous release delivery system,” Journal ofMedicinal Chemistry, vol. 46, no. 23, pp. 5021–5030, 2003.[15] C. Dai, D. Li, J. Popovici-Muller et al., “2-(2-Aminothiazol-4-yl)pyrrolidine-based tartrate diamides as potent, selective andorally bioavailable TACE inhibitors,” Bioorganic and MedicinalChemistry Letters, vol. 21, no. 10, pp. 3172–3176, 2011.[16] J.-B. Behr, T. Gourlain, A. Helimi, and G. Guillerm, “Design,synthesis and biological evaluation of hetaryl-nucleosidederivatives as inhibitors of chitin synthase,” Bioorganic &Medicinal Chemistry Letters, vol. 13, no. 10, pp. 1713–1716, 2003.[17] A. Dahlgren, J. Branalt, I. Kvarnstr¨om, I. Nilsson, D. Musil,and B. Samuelsson, “Synthesis of potential thrombin inhibitors.Incorporation of tartaric acid templates as P2 proline mimetics,”Bioorganic and Medicinal Chemistry, vol. 10, no. 5, pp. 1567–1580, 2002.[18] H. B. K¨uc¸¨uk, A. Yusufoˇglu, E. Mataraci, and S. D¨osler, “Syn-thesis and biological activity of new 1,3-dioxolanes as potentialantibacterial and antifungal compounds,” Molecules, vol. 16, no.8, pp. 6806–6815, 2011.[19] S. Li, G. Liao, J. Xiao et al., “Derivatives of substituted tartaricacid and usage for preparing beta-secretase inhibitors,” US8268884 B2, 2012.[20] B. B. Mishra and V. K. Tiwari, “Natural products: an evolvingrole in future drug discovery,” European Journal of MedicinalChemistry, vol. 46, no. 10, pp. 4769–4807, 2011.[21] Siemens, SMART and SAINT, Siemens Analytical X-RayInstruments Inc., Madison, Wisconsin, USA, 1996.[22] A. Altomare, G. Cascarano, C. Giacovazzo, and A. Guagliardi,“Completion and refinement of crystal structures with SIR92,”Journal of Applied Crystallography, vol. 26, no. 3, pp. 343–350,1993.[23] G. M. Sheldrick, SHELXTL-PC (Version 5.1), Siemens analyticalinstruments, Inc., Madison, Wisconsin, Wis, 1997.[24] C. K. Jhnson, ORTEPII Report ORNL-5138, Oak Ridge NationalLaboratory, Tennessee, Tenn, USA, 1976.[25] L. M. Likhosherstov, O. S. Novikova, V. A. Derevitskaja, andN. K. Kochetkov, “A new simple synthesis of amino sugar 𝛽-d-glycosylamines,” Carbohydrate Research, vol. 46, no. 1, pp. C1–C5, 1986.[26] P. J. Garegg, “Saccharides of biological importance: challengesand opportunities for organic synthesis,” Accounts of ChemicalResearch, vol. 25, no. 12, pp. 575–580, 1992.[27] H. J. Berthold, S. Franke, J. Thiem, and T. Schotten, “Ex postglycoconjugation of phthalocyanines,” The Journal of OrganicChemistry, vol. 75, no. 11, pp. 3859–3862, 2010.Journal of Chemistry9[28] V. V. Sureshbabu, R. Venkataramanarao, and H. P. Hemantha, “Afacile synthesis of C-terminal neoglycopeptides: incorporationof urea moiety between sugars and peptides employing curtiusrearrangement,” International Journal of Peptide Research andTherapeutics, vol. 14, no. 1, pp. 34–40, 2008.[29] T.RenandD.Liu,“Synthesisoftargetablecationicamphiphiles,” Tetrahedron Letters, vol. 40, no. 43, pp. 7621–7625,1999.[30] R. R. Schmidt, “New methods for the synthesis of glycosides andoligosaccharides—are there alternatives to the koenigs-knorrmethod?” Angewandte Chemie International Edition, vol. 25, no.3, pp. 212–235, 1986.[31] G. Excoffier, D. Gagnaire, and J.-P. Utille, “Coupure s´electivepar l’hydrazine des groupements ac´etyles anom`eres de r´esidusglycosyles ac´etyl´es,” Carbohydrate Research, vol. 39, no. 2, pp.368–373, 1975.[32] L. Zhai, M. Chen, J. Blom, T. G. Theander, S. B. Christensen, andA. Kharazmi, “The antileishmanial activity of novel oxygenatedchalcones and their mechanism of action,” Journal of Antimicro-bial Chemotherapy, vol. 43, no. 6, pp. 793–803, 1999. | RESEARCH GATE |
CATASTROPHIC MATERIALS FAILUREDepartment of Materials and Metallurgical EngineeringBangladesh University of Engineering and Technology, DhakaMME ���: Crystal Defects, Deforma�on and FractureGroup name: Crack�������, ��������������, �������Coursework submi�ed to Dr. Mamun Al RashedLevel � Term �A Broken Dream:The Space Shuttle Challenger DisasterOn the morning of January ��, ����, The Challenger Space Shu�le was destroyed a�er �� seconds of launch from the Florida Space Sta�on.• Death of � crew members.• Destruc�on of ��.� Billion• American Space program was set back a decade.Synopsis:Damages:Failure Sequence:Preven�ve Measures:Use Materials that maintain their elas�city even at low temperature.Duc�le to Bri�le Transi�on Temperature (DBTT) should have been known for the material which is used.Strangest Disaster in History: Boston Molasses Tank FailureOn January ��, ����, a large molasses tank in Boston exploded with a great force. The tank contained more than � million liters molasses. Almost �� - �� feet high wave of molasses rushed through the streets at an es�mated �� km/hr.Synopsis:Failure Analysis: Damages:Preven�ve Measures:�� people are dead��� people are severely injuredProperty damages are es�mated to ���� million Manganese in steel composi�on should follow the standard to be duc�le.Bri�leness should be avoided in low carbon steel to stop rapid propaga�on of crack. Stess capacity of steel should be tested with the maximum load.The steel was too weak (about ��% too thin) to withstand the stress of that much molasses.Low manganese in steel was responsible for the bri�leness due to duc�le-to-bri�le transi�on temperature. The rubber made O-ring was designed to elongate and seal the tube to prevent gas flow. (The lowest tested temperature of O-ring was �� °F) But due to low temperature that day(��°F) the o ring turned out to be bri�le and when the pressure released, it is not wide enough to seal the tube and prevent the gas flow.The gas flows out of the chamber and explode.Neumann bands are observed near primary fracture which indicates rapidly propaga�ng crack in the tank.Impact EnergyBrittleDuctile15°C 4°C TemperatureFig: Ductile to brittle transition of Boston Molasses tank Failure date temp.Transi�on temp.Mining Went Wrong: The Alexander L. Kielland Oil Rig Capsized On the evening of �� March,����, The “Alexander L. Kielland”, a ��,���-ton semi-submersible rig opera�ng in the North Sea capsized.• Death of ��� crew members.• Destruc�on of a ��,���-ton rig• �� people barely survived.Synopsis:Damages:Failure Sequence:Preven�ve Measures:While connec�ng the hydrophone to D� brace via fillet welding, a fa�gue crack ini�ated in the welding region.Eventually the final fracture of D� brace was a bri�le fracture.At last, the remaining braces connec�ng the column D to the rest of the structure fractured due to duc�le overload failure. • Employing experienced workers for welding• Increased inspec�on• Heat treatment a�er welding, to reduce stress• Modifica�on of design of the rig. | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/349900043A Critical Review of Recent Research of Free Vibration and Stability ofFunctionally Graded Materials of Sandwich PlateArticle in IOP Conference Series Materials Science and Engineering · March 2021DOI: 10.1088/1757-899X/1094/1/012081CITATIONS26READS3943 authors:Some of the authors of this publication are also working on these related projects:New prosthetic socket for all amputees View projectBuckling of Sandwich Combined Plate View projectEmad K. NjimThe Papua New Guinea University of Technology23 PUBLICATIONS 241 CITATIONS SEE PROFILEMuhannad Al-WailyUniversity Of Kufa172 PUBLICATIONS 2,933 CITATIONS SEE PROFILESadeq Hussein BakhyUniversity of Technology, Iraq59 PUBLICATIONS 425 CITATIONS SEE PROFILEAll content following this page was uploaded by Muhannad Al-Waily on 08 March 2021.The user has requested enhancement of the downloaded file.IOP Conference Series: Materials Science and EngineeringPAPER • OPEN ACCESSA Critical Review of Recent Research of Free Vibration and Stability ofFunctionally Graded Materials of Sandwich PlateTo cite this article: Emad Kadum Njim et al 2021 IOP Conf. Ser.: Mater. Sci. Eng. 1094 012081 View the article online for updates and enhancements.This content was downloaded from IP address 37.236.166.13 on 08/03/2021 at 21:09Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distributionof this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.Published under licence by IOP Publishing LtdINTCSET 2020IOP Conf. Series: Materials Science and Engineering1094 (2021) 012081IOP Publishingdoi:10.1088/1757-899X/1094/1/0120811A Critical Review of Recent Research of Free Vibration and Stability of Functionally Graded Materials of Sandwich Plate Emad Kadum Njim1, Muhannad Al-Waily2, Sadeq H Bakhy1 1 University of Technology, Mechanical Engineering Department, Baghdad, Iraq 2 Department of Mechanical Engineering, Faculty of Engineering, University of Kufa, Iraq E-mail: 20093@uotechnology.edu.iq Abstract. In the past few decades, due to the unique material properties of functionally graded materials (FGM's), they have been used in various engineering industries. This article aims to introduce an overview of the existing literature on the area of application, stability, and free vibration analysis of FGM structures conducted by some recent research studies and to provide a comprehensive overview of the development, application, different numerical representation of materials, demonstrating procedures and arrangement technique and solution method of FGM rectangular plate. It focuses on the influence of many parameters on natural frequencies and buckling loads, such as aspect ratio, power-law index, porosity distribution throughout the thickness of the plate, and face sheet thickness. This research also involves various analyses and numerical techniques for vibration and buckling analysis of the FGM sandwich plate. Furthermore, some important notes and suggestions are put forward for future work trails in this field. It is found that there is an exceptionally restricted path to investigate the same above analysis for the FGM sandwich plate with the porous metal dependent on various parameters such as gradient index, aspect ratio, face sheet thickness, porous factor, FGM layers thickness, and the number of layers. Keywords. Applications of FGM, Free vibration, Buckling, Power-law index, FGM Plates, analytical techniques, FEA. 1. Introduction In the progress of science and innovation, materials have undertaken significant work. The engineering of modern composite materials has had a significant impact on design and construction technology. The special engineering application of raw materials in various inorganic and organic compounds has played an essential role in developing polymers, alloys, structural parts, etc. [1]. Functionally graded materials (FGM) are a class of advanced materials whose material properties (such as mechanical and thermal properties) change continuously from one surface to another, thereby eliminating stress concentration in laminated composites. Changes in phase distribution can be transformed into their volume or weight fraction, configuration, and geometry. The change in volume fraction may occur directly over the entire thickness of the part or in any other direction (such as the coordinates of the plate or beam). The ceramic area has good heat resistance due to its low expansion capability. The ductile metal parts can avoid failure due to stress due to rapid temperature changes in a short time [2,3]. INTCSET 2020IOP Conf. Series: Materials Science and Engineering1094 (2021) 012081IOP Publishingdoi:10.1088/1757-899X/1094/1/01208121.1. History of FGMs plate In 1972, Shen and Bever [4] first proposed the concept of gradual material composition for composite and polymer materials. In order to maintain mechanical properties and reduce the effects of thermal stress, most of these materials are used as coating materials. Functionally graded materials (FGMs) were developed in Japan in the 1980s for thermal insulation purposes [5]. Since then, FGMs have received extensive attention in various engineering applications and manufacturing industries as high-grade structural materials for heat insulation such as aerospace, nuclear reactors, automobiles, airplanes, spacecraft biomedicine, and steel industries. FGM was initially designed as a thermal insulation material for aerospace structures and fusion reactors [6]. At present, the focus of the material development activities of composite materials and FGM includes improving material performance, supporting optimized structural design, reducing manufacturing costs continuously, and the ability to operate reliably [7]. According to the composition stages' distribution, FGM can be divided into continuous or discontinuous gradual composition changes. Similarly, it can be divided into thin and overall FGM according to manufacturing technology [8, 9]. 1.2. Areas of application of FGMs The essential characteristics of FGM make it the first choice for almost engineering applications. Also, they are artificially produced. The original form of FGM exists in nature; bones, teeth, human skin, and bamboo trees can be considered organic forms of FGM. Due to its excellent thermal and mechanical properties, functionally graded materials (FGMs) are widely used in various fields and are likely to be used for other purposes. The most important applications include energy, aerospace, automobile, biomedical, defense, electrical/electronics, marine, Opto-Electronics, sport, thermoelectrics, and bioengineering [10]. Figure 1 shows the different types of FGMs and their application areas [11]. Figure 1. Areas of applications for the three types of FGM [12]. 2. Mathematical Idealization of FGMs Although FGM is very non-uniform, it is beneficial to idealize it as a continuum so that its mechanical properties change smoothly in spatial coordinates. In order to analyze FGMS effectively, a homogenization scheme must be adopted to simplify its complex heterogeneous microstructure. Through this idealization, mathematical model representation to identify some problems can be acquired, which will also help include and improve numerical techniques of FGM structures. It is vital that the conveyance of materials in the FG structure can be intended for different spatial particulars. A typical FGM represents a modern composite material with a recommended conveyance of the volume portion of the constituent stages. It is commonly expected that the material properties follow the progressive change in thickness in a nonstop way. Two sorts of varieties are usually used in the literature, and they include the vast majority of the current analysis models in the fabrication work. INTCSET 2020IOP Conf. Series: Materials Science and Engineering1094 (2021) 012081IOP Publishingdoi:10.1088/1757-899X/1094/1/01208132.1. The power law Material properties and volume content vary along the thickness direction. FGM is usually designed to assume that one of the outermost layers is treated as metal and the other layer is ceramic. The combination of FGM defines the characteristics of plates and beams. Assuming that the FG plate follows a power law change, it can be defined as follows [13], ( ) ( ) (1) The constituent volume fraction of the FGM plate is supposed to change continually along the thickness direction and obey power-law distribution in the following, ( ) ( ) ( ) (2) In equation (2), & are the corresponding material characteristic values of the ceramic and metal components of the FG plate, respectively. The total volume fraction of constituents are expressed as: ( ) ( ) 1, where and are the volume fractions of metals and ceramics respectively, and k is the power-law exponent, which is a non-negative variable parameter, where ϵ [0, ∞). The value of equal to zero represents a fully ceramic plate, whereas infinite indicates a fully metallic plate. The variations may be seen in Figure 2, which reveals that material properties with exponential gradation usually lie between those obtained with power-law exponents, k = 0.2,0.5,1,2, and 5. For our current formulation, the material properties (such as Young's modulus and mass density) vary along with the thickness, but it is assumed that the Poisson's ratio is constant. a. Young’s modulus Variation b. Mass densities variation Figure 2. Material properties representation of the FG plate using the power-law index [14]. 2.2. The exponential law It has been found many research articles that used the exponential function to express the material properties variation with the thickness of the FG plate as follows [15], ( ) ( )( ) ( ( ) )), ( ( )) (3) Where E (z) is the modulus of elasticity, indicates the value of the coefficient of thermal expansion and the property of thermal conductivity of the FGM plate or beam with a thickness of h. INTCSET 2020IOP Conf. Series: Materials Science and Engineering1094 (2021) 012081IOP Publishingdoi:10.1088/1757-899X/1094/1/0120814The representation of Young’s modulus in the direction of the EFGM plate's thickness is plotted in Figure 3. Figure 3. Young modulus variation along with the thickness of the FG plate using exponential function [16]. 2.3. Sigmoid law Sometimes, it is unwise to use low power functions to express FG beams and plates' characteristics and thickness changes. When ensuring the continuity of material properties over the entire thickness range, stress concentration appears in the beam's interface layer, but the change is not smooth [17]. To solve this type of problem, two power indices are used. In a study done by Chung and Chi [18], two power functions were used to express the change in volume fraction to prevent rapid stress changes through the interface. The following equations represent the variation of volume fraction using two power-law indices [19], ( ) ( ) fo ( ) ( ) ( ) fo ( ) (4) By using the rule of mixture, Young's modulus of the Sigmoid FGM can be calculated by, ( ) ( ) [ ( )] fo ( ) ( ) ( ) [ ( )] fo ( ) (5) Figure 4 shows the variation of FGM volume for different values of p by employing the sigmoid function. Figure 4. Variation of Young’s modulus and the thickness of the FG plate using the sigmoid function [20]. INTCSET 2020IOP Conf. Series: Materials Science and Engineering1094 (2021) 012081IOP Publishingdoi:10.1088/1757-899X/1094/1/01208152.4. Mori-Tanaka scheme This method is suitable for composite materials with a well-defined continuous matrix and gradually discontinuous particle-phase microstructure regions. Similarly, this method also considers the influence of the elastic field between adjacent inclusions and their interaction with the components. The effective bulk modulus Kz and shear modulus Gz are calculated according to the following formula [21]. ( )( ) (6) ( )( ) (7) ( ) ( ) (8) The modulus of elasticity z and Poisson’s atio are evaluated using the Bulk modulus and the modulus of shear as indicated below, [22], ( ) (9) ( ) (10) The heat conductivity and thermal expansion parameter αz are as follows, ( )( ) (11) (12) 3. Literature review on the free vibration of functionally graded rectangular plates Rectangular plates are widely used in many engineering applications due to their ability to dealing with the various loads' conditions (mechanical and thermal). The instability of the plate is caused by the in-plane compressive stress caused by these loads. For FGM plates, the stability problem is susceptible to boundary type conditions and material thickness variations [23]. When studying the vibration of FGM plates and beams, many researchers used frequent parameters that affect the fundamental natural frequency, Batra and Vel [24] provided an accurate solution for free vibration analysis of 3D of FGM plates. They assumed that the plate is made of an isotropic material with material properties varying in the thickness direction only. Farajollah Zare Jouneghani et al. [25] used the first-order shear deformation theory to study an FG porous shell's free vibration problem. Vyacheslav N. et al. [26] provided 3D modeling of free vibration and static response of functionally graded materials (FGM) sandwich plates. J. Woo [27] studied the dynamic response of an FGM thin plate using nonlinear analysis. The thin rectangular plate has an impact force locally distributed and has no or no elastic foundation. Prapot and Nuttawit [28] described the flexural vibration analysis of a functionally graded sandwich plate resting on an elastic foundation under arbitrary boundary conditions: Chebyshev collocation technology, the governing equation of free vibration problem is derived, and some of the crucial influences on the relationship between shear deformation and rotational inertia are given. Nuttawit Wattanasakulpong et al. [29], based on the improved coupling stress theory, the vibration state of dimension-dependent functionally graded sandwich microbeams with different boundary conditions was studied. Tran Van Liena et al. [30] used the dynamic stiffness method to study the free and forced vibration analysis of the multi-crack FGM multi-span continuous beam and explained the crack's influence on the vibration of the functionally graded beam. A.W. Leissa [31] used combinations of boundary conditions to investigate FGM rectangular plates' free vibration problems. INTCSET 2020IOP Conf. Series: Materials Science and Engineering1094 (2021) 012081IOP Publishingdoi:10.1088/1757-899X/1094/1/0120816A. F. Mota et al. [32] studied the mechanical behavior of porous functionally graded nanocomposite materials and analyzed nanocomposite functionally graded materials with different porosity distributions. Hassen Ait Atmane et al. [33] conducted a free vibration analysis on a porous functionally graded material beam resting on an elastic foundation. Mechab et al. [34] proved the effects of pores on the dynamic analysis of functionally graded materials nanoplates depend on Winkler–Pasternak foundation. Adda Hadj Mostefa and Merdaci Slimane [35] studied the influence of material property distribution and porosity on the natural frequency of FGM sandwich panels and explained different boundary conditions. Mesut Simsek [36], the free vibration frequency of rectangular plates has been evaluated in research and modern fields under different possible combinations of classical boundary conditions. Baferani et al. [37] described a mathematical model of free vibration of FG thick rectangular plates placed on an elastic foundation. Senthil S.Vela and R.C. Batrab [24] used Mori-Tanaka and self-consistent methods to study the accurate three-dimensional solutions of free vibration and forced vibration of a simply supported functionally graded rectangular plate. Also, Malekzadeh [38] combined 3D shear theory to study the dynamic response of thick functionally graded plates resting on elastic foundation. Liu et al. [39, 40] explained the influence of in-plane material inhomogeneity on the fundamental frequency of the FGM plate. In the mathematical model representation, the kinematic relations relying on the classical plate theory (CPT) were considered, and the analysis of rectangular plates using the (Rayleigh-Ritz) method was mentioned in [41-43]. According to different combinations of classical boundary conditions, rectangular plates' free vibration frequency is described in [44-46]. Reddy [47] proposed a general formula for FG plates using the 3D shear deformation theory. F. Liu and Liew [40] the free vibration of a medium-thickness rectangular plate is studied by the differential orthogonal parameter technique. Matsunaga [48] estimated the natural frequency and buckling stress of FG plates based on the two-dimensional high-order shear deformation theory. Hosseini Hashemi et al. [49] introduced the precise structural arrangement of 3D elasticity theory to study the free vibration of FG simply supported rectangular plates. 4. Functionally graded sandwich plates The wide application of sandwich structures in automotive, marine development, transportation, and aviation companies has attracted many considerations, and specific researchers have performed continuous static and dynamic inspections on them. Due to the outstanding performance of high strength-to-weight ratio, the use of sandwich structures in the field of micro auxiliary frames is constantly developing [50]. Therefore, in a wide range of FGM material types and uses, it is crucial to explore the static and dynamic behavior of auxiliary personnel with FGM, such as beams and plates [51, 52]. Consider the plate comprised of homogeneous hardcore and FGM face sheets [53-58]. The material non-uniformity of the FGMs is assumed as follows [59], { ( ) ( ) (13) Where and are the thickness of the plate core and each FG sheet, respectively. The material heterogeneity of the whole FG structure can be evaluated by using the Voigt rule. Therefore, the material inhomogeneity of the sandwich plate P varies with the thickness coordinates as, ( ) { ( ) ( ) ( ) ( ) ( ) (14) INTCSET 2020IOP Conf. Series: Materials Science and Engineering1094 (2021) 012081IOP Publishingdoi:10.1088/1757-899X/1094/1/01208175. Research on vibration of FGM structures with porosities Because of technical issues during the manufacture, porosities and microvoids can be made inside the FGM plate, which may essentially decrease the quality of materials. The assembling strategies for FGMs are a creating region; the sintering strategy is the most utilized because of its cost-saving advantage proportion. In any case, the sintering cycle empowers the development of microvoids or porosities [9]. Despite the critical improvement observed recently here, porosity stays an incessant imperfection in FGMs. As related above, porosities are outcomes of the assembling cycle and decrease the material's quality. Henceforth, the effect of this type of imperfection is necessary to consider the porosity effect on unique attributes of FGM structures conveying porosities [60]. Yan Qing Wang and Jean W. Zu contemplated vibration practices of practically evaluated rectangular plates with porosities and moving in a warm domain [61]. Saidi Hayat and Sahla Meriem [62] utilized vibration analysis of functionally graded plates with porosity made out of a mixture of Aluminum (Al) and Alumina (Al2O3) installed in an elastic medium. Xiang-Yu Zhang et al. [63] altered the topological design, porousness, and mechanical conduct of additively fabricated functionally graded porous metallic biomaterials. Free vibration of Timoshenko beams with porosities was carried out by Wattanasakulpong and Chaikittiratana, who thought about unevenly conveyed porosities in the model [29]. Y.H Dang et al. [64] additionally gave the conversation on free vibration attributes inferable from porosities occurring inside FGM tests of the graphene fortified permeable nanocomposite barrel-shaped shell with a spinning motion. Concerning porosity distributions, Nguyen et al. [65] studied the mechanical conduct of porous FGPs. For this reason, they considered two diverse porosity appropriations, shifting both through the thickness direction (to be specific, the even and uneven distributions). Zhang and Wang [66] created eight different porous material structures with deferent pore appropriations, including gradient distributions, and exposed them to some mechanical tests to assess significant materials properties like Young's modulus. Functionally graded porous materials join the qualities of both FGMs and porous materials. Beyond the great rigidity–weight ratio, the exceptional mechanical properties they present to clarify why these particular materials are broadly utilized in a broad scope of various fields [67]. Notwithstanding incredible advancements in assembling measures, the arrangement of micro-voids or porosities is as yet a reality [68], and in some particular applications, this can be even alluring and intended for. Notwithstanding the particular case, the material's strength will become lower as a result of these pores, which should be remembered for mechanical conduct contemplates [69]. Three types of porosity distributions through the thickness proposed by Kim et al. [70] and applied in numerous studies, along with the ones evolved by using Coskun et al. [71] and by Zhao et al. [72]. The closing was inspired inside the uniform distribution referred to Merdaci [73], whose studies recognize the case of a typical functionally graded ceramic/metal square plates considering deferent porosity distributions through the thickness. Moreover, Chakraverty and Pradhan [14, 74] have studied the free vibration of thin FG rectangular plates in the presence of complicated environments. Nuttawit and Variddhi [75] investigated the porosity parameter that influences the frequency parameter of FGM restrained ends beams using a combination of linear and nonlinear analysis. 6. Analytical solution of free vibration of fg rectangular plate The static and dynamic response of FG plates had been studied by many researchers based on specific plate theories defined below, Generally, in vibration problems, the displacement fields of the deformed beam (or plate) can be decided via shear deformation beam (or plate) theories. Instead of classical beam (or plate) theory, exceptional varieties of deformation theories may also be observed in the open literature. It might also occur because of the reality that the classical plate (or beam) idea neglects transverse shear deformation outcomes. To have a look at the vibration characteristics, the displacement fields of deformed FG beams and plates are considered with recognition to transverse displacement. Higher-order deformation beam theories of various forms are assumed by taking transverse shear deformations in the case of FG beams, whereas the classical plate theory (CPT) is taken into consideration within the case of FG plates with different geometries. A new quasi-three dimensional high shear deformation theory for the vibration of the functionally graded plate was considered by [76]. INTCSET 2020IOP Conf. Series: Materials Science and Engineering1094 (2021) 012081IOP Publishingdoi:10.1088/1757-899X/1094/1/0120818Baferani et al. [77] have analyzed the dynamic response of functionally graded thin plates. The governing equations of motion are obtained based on the classical plate theory, and the effects of aspect ratio, thickness, length ratio, power-law index, and boundary conditions on the vibration characteristics of functionally graded rectangular plates are discussed in detail. The nonlinear bending analysis is presented by Shen [78] for a simply supported functionally graded plate subjected to a transverse uniform or sinusoidal loads in thermal environments. Yang and Shen [79] have offered free and forced vibration analyses for initially stressed functionally graded plates in thermal surroundings. 6.1. Classical plate theory (CPT) The two-dimensional plate theories can be categorized into two types: (1) classical plate theory, in which the transverse shear deformation consequences are neglected, and (2) shear deformation plate theories. In the two-dimensional theory, the free vibration, thermal, and stability problems of the FGM structures, the displacement is represented in terms of thickness, while the lateral displacement is independent of the lateral (or thickness) coordinates. The results of the mathematical model in the coupled governing equation are independent of lateral displacement. Therefore, the analytical solution of this type of equation may be simpler than the solution procedure of three-dimensional elasticity theory [80] The easiest method concept is the Classical Plate Theory (CPT), which is an extension of the Kirchhoff (classical) plate theory to laminated composite plates. The conventional Kirchhoff model is no longer regarded as the effect of shear deformation; consequently, it is solely relevant for thin plate analysis. The classical plate principle was once at the beginning developed for homogeneous isotropic plates and was later extended to laminated composite and FGM plates. Most researchers hire CPT in the analysis alongside with the solution of FG plates of various shapes, viz. rectangular, elliptic, and triangular alongside with specific complicating results (elastic foundation, thermal environment, and piezoelectricity). Based on the physical neutral surface property, Zhang and Zhou [81] used classical plate theory to analyze FG thin plates neglecting the tension-bending coupling effect in the mathematical formulation for deflection problems. Liu et al. [39] studied the analysis of CPT-based FGM plates, especially when the material properties changed through the thickness, the effect of this configuration on the panel's natural frequency. Gene ally, the classical o Ki chhoff’s plate theo y is based on the following assumptions [82, 83], The thickness of the plate is small in contrast to different dimensions. The normal stresses in the route transverse to the plate are taken to be negligibly small. The effect of rotatory inertia is negligible. The normal to the undeformed middle surface remains straight, and the normal to the deformed middle surface remains unstretched in length. Using the CPT, the displacement fields of FG plates across the plate thickness at a distance z away from the middle surface are [84, 85], ( ) ( ) ( ) ( ) (15) Where, are the displacement of a point on the reference plane in the x, y, and z directions, respectively, and w represents the lateral deflection of the points on the mid-plane (x-y plane). The Kirchoff model is not considered the effect of shear deformation due to bending and plane elongation. 6.2. First-order shear deformation theory When a shear correction component is needed to compensate for the difference between the proper stress state and the assumed normal stress state, the first-order shear deformation theory (FSDT) is used. It was developed by Mindlin [86], and Reissner [87] accounts for the effect of shear deformation, but it violates the traction-free boundary conditions at the ends of the desired surface [88, INTCSET 2020IOP Conf. Series: Materials Science and Engineering1094 (2021) 012081IOP Publishingdoi:10.1088/1757-899X/1094/1/012081989]. Aghdam et al. [90] combined the iterative technique with the Kantorovich method and based on the theory of first-order shear deformation to investigate a static analysis of the bending of medium-thickness FG plate. Therefore, shear correction parameters are needed to compensate for the difference between the actual stress and the assumed normal stress state. Hosseini-Hashemi et al. [49] introduced the FOST, to study the free vibration of FGM plates based on the kinematics and constitutive equations for the proposed model to calculate the natural frequency, assuming that the properties of FG plate vary through the thickness and have a gradient index along with the part thickness orientation. M. Karami Khorramabadi et al. [91] used both FOST and TSDT to investigate the free vibration problem of simply supported FG plates and discussed the differences between the two theories on the dynamic response of FG plates. Nguyen et al. [92] developed a new model of free vibration for FGM plates employing the assumptions of FOST in the closed-form solution. According to the fi st-o de shea defo mation plate theo y [49]; the displacement field can be exp essed as, ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) (16) Where, , , and denote the displacements at the mid-plane of the plate along the x, y, and z directions, and and represent the normal transverse rotations about the y and x-axes, respectively. 6.3. A simple higher-order shear deformation theory To avoid using the shear correction thing and obtain a higher prediction of the transverse shear deformation and regular strains in FG plates, Higher-order shear deformation plate theories (HSDTs) have been proposed. In general, HSDTs can be developed based totally on higher-order variants of the in-plane displacements [93-95] or each in-plane and transverse displacements [96, 97] (i.e., quasi-3D theories). Using the simple and complete theory of trigonometric high-order, the bending and vibration of FGM parts are discussed to study the influence of normal transverse strain on deflection and stress [98]. Fakhari et al. [99] used thermal, electrical, and mechanical loads to propose a new nonlinear mathematical model based on high-order shear deformation theory to evaluate FG plates' natural frequency and stability with piezoelectric layers bonded through the part surface. However, HSDTs are enormously computational due to many unknowns (e.g., theories employing Neves et al. [100] with nine unknowns). Among different HSDT, Whitney and Sun [101] introduced the second-order shear deformation formulation and the TSDT of Lo et al. [102] with (11 unknowns), Kant [103] with six unknowns, Bhimaraddi and Stevens [104] with five unknowns, and Hanna and Leissa [57] with four unknowns, furthermore the TSDT proposed by Chen and Reddy [105] with five unknowns is the most generally utilized model in the investigation of FG vessels due to its acceptable results. Reddy [106] with (eleven unknowns) while Jha et al., [107] with (12 unknowns), Talha and Singh [108], and Natarajan and Manickam [109] with (13 unknowns). Many modifications are performed on this concept, such as a quasi-3D (hyperbolic, sinusoidal, simple, and higher-order shear deformation theory). Khalili and Mohammadi [110] used an improved high-order sandwich plate theory to analyze sandwich plates' free vibration with FGM face sheets in various thermal environments. According to fundamentals of the shear deformation theory, both the axial and lateral displacement (ux & uz) at any point of the plate or beam can be represented as [111], ( ) ( ) ( ) ( ) ( ) ( ) ( ) (17) Where, and represent the axial and lateral displacement of any point on the neutral axis, and v is a special function, which depends on the actual value of the shear strain on the symmetry axis, f (z) represents the geometric function to be calculated with assistance both of strain and stress resulting from transverse action across the plate thickness, and the derivative of the lateral displacement concerning x is the component . By selecting the correct form of the structural characteristic function f(z), the various mathematical formulation can be obtained, classical plate theory (CBT), INTCSET 2020IOP Conf. Series: Materials Science and Engineering1094 (2021) 012081IOP Publishingdoi:10.1088/1757-899X/1094/1/01208110first-order shear deformation theory FSDT, higher-order shear deformation theory HSDT, exponential shear deformation plate theory (ESDT, sinusoidal shear deformation plate theory SSDT, as mentioned in Simsek, ( ) for these, are given below [36], ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) (18) 6.4. Zeroth-order shear deformation theory To predict actual results for the free vibration and buckling load problems associated with FGM structures, a Zeroth-order shear deformation principle was utilized. This concept was fi st explored through Shimpi [112] for isotropic plates and later stretched out by Ray [113] for laminated composite plates. They affect the use of shear forces used to be taken into accounts in the Zeroth-order as an alternative of rotational displacements of transverse shear deformation as in present shear deformation theories. The zeroth-order shear deformation theory contains the equal five unknowns in the first-order shear deformation theory; however, it satisfies the traction-free boundary conditions on the plate's top and bottom surfaces besides requiring any shear correction factor. Using the zero-order shear deformation theory, equations of motions and the analytical solutions are derived. The frequency parameter of the free vibration analysis of functionally graded plates treated with nanoparticles mounting on elastic foundations was carried out [114,115]. Based on the CPT, the linear constitutive relations of an FG plate such as the bending and twisting moments respectively on a plate element in the pure bending case can be written as [116], ∫ ( )∫ ( ) ( ) ∫ ( )∫ ( ) ( ) ∫ ∫ ( ) (19) Then, (20) ∫ ( ) ( ) (21) ( ) ( ) { ( )} ( ) (22) where, is the stiffness coefficient of the functionally graded plate. And, ∫ ( ) Io is the moment of inertia of the FGM plate, which can be expressed in terms of the volume fraction index as, ∫ (( ) ( ) ) ∫{( )( ) } ∫ ( ) ( ) (23) INTCSET 2020IOP Conf. Series: Materials Science and Engineering1094 (2021) 012081IOP Publishingdoi:10.1088/1757-899X/1094/1/01208111Where, , are young modulus for ceramic and metal respectively; , are the mass density for ceramic and metal respectively; is Poisson's ratio for the FGM plate, and is the thickness of the FGM plate. Substituting Eqs. (7), (12) and (13) into Eq. (10), the solutions can be obtained from (( ) ( ) { ( )} ( ))( ) (( ) ) (24) 7. Stability analyses of fg plates The stability research of functionally graded material plates involves calculating the buckling loads under different boundary and load conditions. Generally, there are two types of solutions. The first is the three-dimensional (3D) elastic theory, used to determine stability status, and the eigenvalue problem needs to be calculated. However, there is no report on using three-dimensional elastic theory (a high-precision analysis technique) for buckling analysis of FGM plates so far. The second is the two-dimensional (2D) plate theory [117]. Rectangular plates are used as much as possible in industrial applications and tend to withstand various loading conditions, such as mechanical or thermal loads. These loads will generate in-plane compressive stresses, which will lead to instability of the plate. For isotropic and FG plates, the buckling problem is critical for boundary type conditions and material thickness variations [118]. This article covers various studies on the buckling of FGM structures, as follows: Ashraf Zenkour et al. [119] used third-order deformation theory and nanobeam resting on Pasternak's foundation to study the influence of the shear deformation and the slenderness ratio on the buckling behavior of the functionally graded part. Huang and Li [120] presented various beam theories to investigate functionally graded columns' stability according to more than a few types of load conditions. Wang et al. [121] explained that the classical plate theory (CPT) in which the transverse shear deformation consequence neglected; therefore, the classical plate theory usually underestimates deflection and overestimates the natural frequencies and buckling loads for thick plates. Samsam Shariat et al. [122] investigated the Buckling evaluation of functionally graded plates subjected to uniaxial loading. Birman [123] studied the buckling problem of functionally graded composite rectangular plates subjected to uniaxial compression. Lee et al. [124] studied Post-buckling analysis of functionally graded plates' concern to compressive and thermal loads. Wu L. [125] developed a new model of a simply supported rectangular functionally graded plate and studied the effect of aspect ratio and temperature gradient on buckling analysis. Czechowski L. et al. The buckling and post-buckling studies of step-variable FGM boxes were carried out [126]. Javaheri [127] studied the buckling of functionally graded plates under plane gradient compressive loads. Bekir Akgöz and his co-authors used the strain gradient theory and introduced the exact solution of FG microbeam stability analysis [128]. Fekrar et al. [129] proposed a new refined theory using the Navier method and including four unknown functions with in-plane loading to study the mechanical buckling of FG simply supported hybrid plates. The correspondence relationship between the deflection, buckling load, and frequency of functionally graded thin materials and corresponding homogeneous plates has been studied by Li Shirong et al. [130]. Shi-Rong Li et al. have learned about correspondence relations between deflection, buckling load, and frequencies of thin functionally graded materials and those of corresponding homogeneous plates [130]. Soldatos [131] presented 2D deformation theory based on Hamilton's principle and Lagrange multipliers to examine the stability of the homogeneous monoclinic plate. Oyekoya et al. [132] introduced a new analytical solution for the buckling of FG structures and used the finite element method to validate the analysis results. Also, a great deal of study has been performed by Bodaghi M et al. [133] to solve an analytical model with the assistance of higher-order shear deformation plate theory to find a solution for buckling analysis of thick FG rectangular plates. M.M. Najafizadeh and M.R. Eslamia evaluated the response of FG plates with various geometrical considerations subjected to compressive load [134]. B. Sidda Reddy et al. used the theory of high-order shear deformation to analyze the buckling of functionally graduated material plates [135]. Farzad Ebrahimi and Fateme Mahmood [136] presented a modified couple stress concept for buckling evaluation of higher-order inhomogeneous microbeams with porosities. The difficulty of the new INTCSET 2020IOP Conf. Series: Materials Science and Engineering1094 (2021) 012081IOP Publishingdoi:10.1088/1757-899X/1094/1/01208112principle of shear deformation for various engineering parts, such as laminated composite plates, is studied with the aid of [137]. Zenkour [138] presented a complete evaluation of the buckling and free vibration of a simply supported FG sandwich panel consisting of a homogeneous core and FG upper and lower parts. Elias Y. Ali and Yared S. Bayleyegn [139] presented a detailed study that includes an analytical and numerical calculation for buckling analysis of FG rectangular plates subjected to compression load in one direction. Akhavan et al. [140] evaluated the buckling analysis of a rectangular Mindlin plate under uniform load resting on the Pasternak elastic foundation and explained the influence of parameters such as foundation stiffness coefficient, boundary conditions, and thickness ratio on the buckling behavior. Meiche et al. [141] modified the new mathematical model of the dynamic response for the FG sandwich plate based on the hyperbolic shear deformation theory. Shen et al. [142] studied the post-buckling of sandwich panels with FGM panels and temperature-dependent characteristics. Kiani et al. [51] studied the stability and dynamic characteristics of functionally graded sandwich panels placed on Pasternak elastic foundations using different boundary conditions. Nguyen and Tung [143] introduced CPT and derived a mathematical model from studying FG plates' buckling analysis using different loads and aspect ratios. Park and Kim [144] developed a numerical model based on FSDT to study FGM plates' post-buckling response under thermal loads. Lee and Kim [145] studied the post-buckling of FGM panels in response to thermal environments. As an application, Zhang and Zhou [146] studied sandwich plates' equilibrium behavior made of functionally graded materials. They proposed the free vibration, buckling, and deflection analysis of FG structures based on physically neutral surfaces. Sobhy [147] analyzed the critical buckling load and free vibration of an exponential sandwich plate based on an elastic foundation under uniform shear under different boundary conditions. Hessameddin Yaghoobi and Pooria Yaghoobi [148] studied the buckling behavior of asymmetric FGM structure placed on an elastic foundation using the structural equation of FSDT and considered various boundary mechanics, thermodynamics, and thermodynamic analysis conditions. Based on the element-free Ritz procedure and FSDT, the buckling behavior of stable plates and plates with notches at the core was studied by Zhao et al. [149]; it was found that the influence of the volume fraction index affects the buckling temperature positively. Jalali et al. [150] used a new spectral method to study the effect of the volume fraction index on the stability degree and the results of thermal stresses owing to the buckling of sandwich round plates with variable thickness. It was concluded that the buckling load parameter would increase with a make bigger in volume fraction index and a decrease in the FG core to homogenous face sheet thickness ratio. Yu et al. [151] used an imperfect FGM plate, combined with the first-order shear deformation theory, and studied the buckling response using an extended isogeometric analysis method. Latifi [85] used Fourier series expansion to analyze the stability analysis of rectangular FG plates under various boundary conditions. S. Sirinivas and A. K. Rao [152] conducted an investigation on bending, vibration, and buckling of simply supported thick orthotropic rectangular plates and laminates. Also, Na and Kim [153,154] investigated functionally graded composite rectangular plates subjected to uniaxial compression. The bending and buckled parts of FGM plates can be studied for various aspect ratios. Although the plate's buckled arrangement can be demonstrated in each load case, the shear and biaxial buckling styles of the linear gradient functionally graded plate at the aspect ratio = 1% have been obtained and given in [155]. In [156, 157], an alternative method of deriving the buckling analysis equilibrium equation based on virtual displacement is mentioned. Using the energy method, M. Mohammadi et al. [158] proposed a general analysis program for the stability analysis of a thick FG rectangular plate with two simply supported opposite edges in uniaxial compression .M. Darvizeh et al. [159] studied the buckling behavior of simply-supported composite plates subjected to compressive loading condition by using the total potential energy technique in conjunction with the Rayleigh-Ritz method. Javaheri and Eslami [160,161] introduced classical and higher-order plate theories to investigate FG rectangular plates' thermal and mechanical buckling. Wu et al. [125] using fast-converging finite double Chebyshev polynomials, the post-buckling response of FG plates is obtained by thermomechanical load analysis. The mathematical formula is based on FOST and von-Karman nonlinear kinematics. The stability equation of the thin plate can be obtained by the INTCSET 2020IOP Conf. Series: Materials Science and Engineering1094 (2021) 012081IOP Publishingdoi:10.1088/1757-899X/1094/1/01208113variational approach. To identify the stability of the FG structural part, the buckling force can be found by solving the equilibrium equation in the form of a diaphragm, resulting in the following bending stress [162,163], (25) (26) ( ) (27) 8. Numerical investigation To verify the accuracy of the analytical solutions, numerical methods are usually used. There are many mathematical methodologies utilized for problem-solving [164-177]; however, the most exact is Finite Element Analysis (FEA) and Meshless methods [178-191]. FEA is a versatile and necessary strategy for assessing most of the design problems permits the investigation of the different structures without improving and including complex relations [192,205]. Furthermore, hybrid numerical mathematical techniques include the Rayleigh-Ritz finite element method (FEM) [206-219], and the differential orthogonal method is used to solve free vibration and buckling problems exposed to dynamic and static problems [220-232]. The dynamic behavior of the FG plate has been studied by many researchers in the world utilizing analytical analysis and various numerical methods. Below, a summary of numerical techniques is used in most research studies. Numerical results of Mantari et al. [233] were bound to the same layer of the functionally graded sandwich plate, while Zenkour and Sobhy [234,235] dealt with the same and different parts of the FGM sandwich plate. Many studies have been proposed to perform buckling analysis of FGM structures with general boundary conditions. The stability analysis in the thermal environment of Kiani and Eslami [236] depends on the details of the Galerkin procedure combining with the free KP-Ritz arrangement developed by Liew et al. [237] in this category. Chi and Chung used FEA to describe FGM plates' dynamic response under various loads [238]. Pradyumna and Bandyopadhyay [239] developed a new mathematical model based on the principles of the higher-order formulation; to solve the buckling problem in FGM structures. Additionally, buckling analysis of FG simply-supported plates was carried out by Nguyen-Xuan et al. [240] using a higher-order finite element formulation. Khalili et al. [241] introduced Rayleigh-Ritz and differential orthogonal methods to study the dynamic response characteristics of functionally gradual structures under fluctuating loads. Jafari and Eftehari [242] proposed a new version of coupled finite element and differential orthogonal to study the dynamic characteristics of beams under dynamic loads. Zhao et al. [243] proposed an element-free KP-Ritz method to analyze the free vibration of metal-ceramic FG plates whose material properties vary continuously throughout the thickness of the plate layers. Zhu and Liew [244] used the local Kriging meshless method to analyze metal and ceramic FG plates' free vibration. The comprehensive study includes the static and buckling analysis of the laterally loaded FG plate using the FEA formulation described in [245] Dozio [246] presented 2-D Ritz models using FEA to study the dynamic behavior of FG sandwich plates with homogenous face sheet. A finite strip element for the analysis of variable thickness rectangular thick plates was investigated by [247]. K. M. Liew [248] presented an adequate solution for free vibration problems of multilayers plate using a mesh-free Galerkin method. Civalek [249] proposed static and dynamic numerical solutions for the rectangular thin plate problem. To analyze static and dynamic in two and three-dimensional elasticity problems of FG structure, a standard FEM technique method has been consolidated by Song et al. [250]. An investigation of FG beams and plates dependent on first-order shear deformation theory using the finite element method model is developed by Chakraborty and Pradhan. [251]. Kant and Khare [252] used high-order shear deformation theory to analyze free vibration and tested the good performance of thin and thick plates and shells. A mixed finite element formulation with a bilinear shape function is used to solve free vibration problems in FGM structures such as the analysis of thin plates resting on an elastic foundation that has been done by many INTCSET 2020IOP Conf. Series: Materials Science and Engineering1094 (2021) 012081IOP Publishingdoi:10.1088/1757-899X/1094/1/01208114 esea ches [253]. Given the fi st and thi d-order shear deformation plate theories, an analysis of free vibrations of FG plates has been introduced by Ferreira et al. [254]. Shufrin and Eisenberger used a new numerical technique to study the stability and vibration of shear-deformable FG plates based on first-order and high-order analysis [255]. Finite element models based on the third-order shear deformation theory were presented to analyze FG plates' static and dynamic analysis by Reddy [256]. Navier arrangements acquired for a simply supported square plate under sinusoidally distributed load were presented, including the effect of shear deformation [257]. Kim and Hoa [258] experimentally determined the dynamic mechanical behavior of the composite plate under biaxial load and performed a numerical analysis. Sundararajan et al. [259] modified a combination of FSDT and correlation analysis with a high grade of node physical element representation conducting on a flexible quadrilateral plate made of FG to perform static and dynamic analysis. Malekzadeh and Shojaee [260] used eight-node physical elements and mathematical formula models to check the accuracy of FSDT on the dynamic response of printed FG parts subjected to heat movement. For static and vibration analysis of FGM structures recently, meshless technology has been widely used in different engineering analyses due to its flexibility [261]. Parandvar and Farid [262] proposed a new finite element modal technique that includes the development of system uniformity methods to deal with the free vibration of functionally graded plates. A modified Kirchhoff plate theory for free vibration analysis of FGM plates using the mesh-free method has been proposed by Vuong Nguyen Van Do [263]. Qian et al. [264,265] analyzed free and forced vibrations of both homogeneous and FG thick plates with the higher-order shear and normal deformable plate theory by using the meshless local Petrov–Galerkin method. Zhu and Liew [266] have developed a meshless method for FGM structure based on FSDT analysis using Kriging interpolation and von-Karman nonlinearity. Wang and Luo [267] used a meshless collocation and an element-free Galerkin method for the 3D free vibration of FGM sandwich plates. Yas and Aragh [268] used the generalized differential orthogonal method to analyze the four parameters and proposed a flexible solution for FGM cylindrical part. For the same arrangement, Pandey and Pradyumna [269] proposed commonly used numerical techniques to study the free vibration of FG sandwich plates installed in thermal surroundings. Hosseini Hashemi et al. [270] used both three-dimensional elasticity solutions and a finite element model to investigate two types of load (in-plane and out-of-plane) free vibrations for thick FGM simply supported rectangular plates, [271-273]. 9. Discussion Functionally graded material is a high-quality material that will revolutionize the manufacturing world in the 21st century. There are many roadblocks to understanding this target. Cost is a transcendent issue, with a tremendous section of the cost expended on the powder preparing and manufacturing strategy. In this work, a basic outline for study static and dynamic response of rectangular and sandwich plate with FGM core and metal face sheets for the selected models utilizing both the proposed analytical solution and numerical is completed to get the natural frequency and critical buckling load problem of sandwich structures. The FGM layers are graded throughout the thickness metal through changing more than a few parameters included but not restrained to gradient index, face sheet thickness, FGM thickness, aspect ratio, number of FGM platelayers, etc. The materials are assumed to be distributed to upper and lower plate parts and, the FGM part comprises various materials, for example, ceramic and aluminum. 10. Concluding remarks It can be inferred from various applications that beams and plates' vibration research is a significant field. The dynamic response includes evaluating frequency parameters, mode shape, stability, and buckling load in various complex environments. Likewise, various analysis and calculation techniques can be used to evaluate vibration characteristics. Existing literature has reviewed various studies on the free vibration and buckling analysis of FG plates. People try their best to remember all the essential contributions in the current field of interest, focusing on the most relevant works available to research engineers studying FG plate structures. The general comments of the present writing overview are as follows: INTCSET 2020IOP Conf. Series: Materials Science and Engineering1094 (2021) 012081IOP Publishingdoi:10.1088/1757-899X/1094/1/012081151. Many researchers have made great efforts to study the free vibration and the buckling in beams, rectangular plates, and sandwich plates, but it should be noted that compared with the analysis mentioned above methods that there are limited trails to investigate same above analysis for FGM sandwich plate with the porous metal taking in account some important parameters such as power-law index, thickness ratio, face sheet thickness, porous ratio, FGM core thickness and the number of layers. 2. It is found that a three-dimensional analytical solution for FG plates can be utilized to check the accuracy of various 2D plate theories and finite element formulations, but in most cases, there is major trouble to distinguish the numerical the mathematical representation for the selected models and get results. By checking the 3-D elasticity solution results, the dynamic response due to the 2D shear deformation theory is verified and accurate. No correlation has been found between the accuracy of the results and verification with experimental work. 3. For CPT, the influence of transverse shear is eliminated. Therefore, it is only used for thin FG plates. In most 2D theories, the influence of transverse shear and normal transverse deformation is considered, so it is suitable for predicting dynamic response for both thin and thick FG plates. 4. For a fixed power-law exponent, the natural frequency of the FG plate increases with the increase of the aspect ratio, and for a fixed aspect ratio, the natural frequency of the FG plate decreases with the increase of the power-law exponent. Having assessed a large segment of the FGM research accessible, it is evident that virtually all the works conducted have been purely analytical or with numerical simulation, and there is an apparent lack in the experimental work. 5. Due to the wide application of FGM structures in engineering industries, subsequently, further work should be done to improve the process control for in general FGM manufacture improvement to suit execution which is connected with free vibration and the critical buckling load and investigate this subject in many types of research to distinguish and sufficient safe and bring down the cost of FGM. 6. For future work, it is found that a 3D thermoelastic solution has not been implemented. Few people consider the effect of temperature changes on the performance of FGM boards. To expand the thermal analysis of various structures, including nonlinear effects, detailed research must be conducted. 7. For free vibration and buckling load calculations, aspect ratio and various material distributions play a vital role in distinguishing the frequency and buckling load parameter values of FG rectangular plates. Also, maybe noticed in most studies of free vibration that as the aspect ratio increases, the FG rectangular plate's frequency parameters also increase. This fact is because as the aspect ratio increases, the plate stiffness becomes greater. 11. References [1] G E Knoppers, J W Gunnink, J Van Den Hout and W Van Vliet 2005 The Reality of Functionally Graded Material Products (In Intelligent Production Machines and Systems- First I PROMS Virtual Conference: Proceedings and CD-ROM set), (Elsevier) pp 467 [2] Minoo Naebe and Kamyar Shirvanimoghaddam 2016 Functionally Graded Materials: a Review of Fabrication and Properties (Applied materials today) vol 5 pp 223–245 [3] Jha D K, Kant T and Singh R K 2013 A Critical Review of Recent Research on Functionally Graded Plates (Composite Structures) vol 96 pp 833–849 [4] M Shen and B Bever 1972 Gradients in Polymeric Materials (Journal of Materials Science) vol 7 no 7 pp 741–746 [5] M Koizumi and M Niino 1995 Overview of FGM research in Japan (Mrs Bulletin) vol 20 no 1 [6] S Suresh and A Mortensen 1998 Fundamentals of functionally graded materials (Materials Today) vol 1 no 4 p 18 [7] B Kieback, A Neubrand and H Riedel 2003 Processing Techniques for Functionally Graded Materials (Materials Science and Engineering) pp 81–105 [8] J J Lannutti 1994 Functionally Graded Materials: Properties,mPotential and Design INTCSET 2020IOP Conf. Series: Materials Science and Engineering1094 (2021) 012081IOP Publishingdoi:10.1088/1757-899X/1094/1/01208116 Guidelines (Composites Engineering) vol 4 pp 81–94 [9] Shahistha A, Varghese B and Baby A 2014 A Review on Functionally Graded Materials (The International Journal of Engineering and Science) vol 3 no 6 pp 90–101 [10] W Pompea, H Worch, M Epple, W Friess, M Gelinsky, P Greil, U Hempele, D Scharnweber, K Schulte 2003 Functionally Graded Materials for Biomedical Applications (Materials Science and Engineering) vol A362 pp 40–60 [11] E Mülle , C D aša , J Schilz and W A Kaysser 2003 Functionally Graded Materials for Sensor and Energy Applications (Materials Science and Engineering: A) vol 362 pp 17–30 [12] Cherradi N, Kawasaki A and Gasik M 1994 Worldwide Trends In Functional Gradient Materials Research And Development (Composites Engineering) vol 4 no 8 pp 883–894 [13] H J Xiang and J Yang 2008 Free and Forced Vibration of a Laminated FGM Timoshenko Beam of Variable Thickness Under Heat Conduction (Composites Part B: Engineering) vol 39 no 2 [14] S Chakraverty and K K Pradhan 2014 Free Vibration of Exponential Functionally Graded Rectangular Plates in Thermal Environment with General Boundary Conditions (Aerospace Science and Technology) vol 36 pp 132–156 [15] Metin Aydogdu and Vedat Taskin 2007 Free Vibration Analysis of Functionally Graded Beams with Simply Supported Edges (Materials and Design) vol 28 no 5 pp 1651–1656 [16] Reddy K S K and Kant T 2014 Three Dimensional Elasticity Solution for Free Vibrations of Exponentially Graded Plates (Journal of Engineering Mechanics) vol 140 no 7 [17] Delale F and Erdogan F 1983 The Crack Problem for a Nonhomogeneous Plane (ASME Journal of Applied Mechanics) vol 50 pp 609–614 [18] Chung Yen-Ling and Chi, S H 2001 The Residual Stress of Functionally Graded Materials (Journal of the Chinese Institute of Civil and Hydraulic Engineering) vol 13 pp 1–9 [19] Thang P T, Nguyen T T and Lee J 2016 Closed-Form Expression for Nonlinear Analysis of Imperfect Sigmoid-FGM Plates with Variable Thickness Resting on Elastic Medium (Composite Structures) vol 143 pp 143–150 [20] Bhandari M and Purohit K 2015 Response of Functionally Graded Material Plate under Thermomechanical Load Subjected to Various Boundary Conditions (International Journal of Metals) vol 2015 [21] T Mori and K Tanaka 1973 Average Stress in Matrix and Average Elastic Energy of Materials with Misfitting Inclusions (Acta Metallurgica) vol 21 pp 571–574 [22] Y Benveniste 1987 A New Approach to the Application of Mori–Tanaka’s Theory of Composite Materials (Mechanics of Materials) vol 6 pp 147–157 [23] Thai H T and Choi D H 2012 A Refined Shear Deformation Theory for Free Vibration of Functionally Graded Plates on Elastic Foundation (Composites Part B: Engineering) vol 43 no 5 pp 2335–2347 [24] Vel S S and Batra R C 2004 Three-Dimensional Exact Solution for the Vibration of Functionally Graded Rectangular Plates (Journal of Sound and Vibration) vol 272 no 3–5 pp 703–730 [25] Farajollah Zare Jouneghani, Rossana Dimitri, Michele Bacciocchi and Francesco Tornabene 2017 Free Vibration Analysis of Functionally Graded Porous Doubly-Curved Shells Based on the First-Order Shear Deformation Theory (Applied Sciences) vol 7 no 12 [26] Vyacheslav N and Tomasz Sadowski 2020 Free Vibrations and Static Analysis of Functionally Graded Sandwich Plates with Three- Dimensional Finite Elements (Meccanica) vol 55 [27] J Woo, S A Meguid and L S Ong 2006 Nonlinear Free Vibration Behavior of Functionally Graded Plates (Journal of Sound and Vibration) vol 289 no 3 pp 595–611 [28] Prapot Tossapanon and Nuttawit Wattanasakulpong 2017 Flexural Vibration Analysis of Functionally Graded Sandwich Plates Resting on Elastic Foundation with Arbitrary Boundary Conditions: Chebyshev Collocation Technique (Journal of Sandwich Structures and Materials) vol 22 no 2 pp 156–189 [29] Nuttawit Wattanasakulpong and Arisara Chaikittiratana 2015 Flexural Vibration of Imperfect Functionally Graded Beams Based on Timoshenko Beam Theory: Chebyshev Collocation Method (Meccanica) vol 50 pp 1331–1342 INTCSET 2020IOP Conf. Series: Materials Science and Engineering1094 (2021) 012081IOP Publishingdoi:10.1088/1757-899X/1094/1/01208117[30] Tran Van Liena, Ngo T ong Đuc and Nguyen Tien Khiem 2019 Free and Forced Vibration Analysis of Multiple Cracked FGM Multi Span Continuous Beams using Dynamic Stiffness Method (Latin American Journal of Solids and Structures) vol 16 no 2 [31] A W Leissa 1973 The Free Vibration of Rectangular Plates (Journal of Sound and Vibration) vol 31 no 3 pp 257–293 [32] A F Mota and M A R Loja 2019 Mechanical Behavior of Porous Functionally Graded Nanocomposite Materials (Journal of Carbon Research) vol 5 no 2 [33] Hassen Ait Atmane A Tounsi and F Bernard 2017 Effect of Thickness Stretching and Porosity on Mechanical Response of A Functionally Graded Beams Resting on Elastic Foundations (International Journal of Mechanics and Materials in Design) vol 13 pp 71–84 [34] Belaїd Mechab, Ismail Mechab, Sami Benaissa, Mohammed Ame i and Boualem Serier 2016 Probabilistic Analysis of Effect of the Porosities in Functionally Graded Material Nanoplate Resting on Winkler–Pasternak Elastic Foundations (Applied Mathematical Modelling) vol 40 no 2 [35] Adda Hadj Mostefa and Merdaci Slimane 2020 Influence of Porosity on the Analysis of Sandwich Plates FGM Using of High Order Shear–Deformation Theory (Frattura ed Integrità Strutturale) vol 51 pp 199–214 [36] M Simsek 210 Fundamental Frequency Analysis of Functionally Graded Beams by using Different Higher-Order Beam Theories (Nuclear Engineering and Design) vol 240 no 4 pp 697–705 [37] Baferani A H, Saidi A R and Ehteshami H 2011 Accurate Solution for Free Vibration Analysis of Functionally Graded Thick Rectangular Plates Resting on Elastic Foundation (Composite Structures) vol 93 pp 1842–1853 [38] Malekzadeh P 2009 Three-Dimensional Free Vibration Analysis of Thick Functionally Graded Plates on Elastic Foundations (Composite Structures) vol 89 pp 367–373 [39] D YLiu, C YWang and W Q Chen 2010 Free Vibration of FGM Plates with in-Plane Material Inhomogeneity (Composite Structures) vol 92 no 5 pp 1047–1051 [40] F L Liu and K M Liew 1999 Vibration Analysis of Discontinuous Mindlin Plates by Differential Quadrature Element Method (Journal of Vibration and Acoustics, Transactions of the ASME) vol 121 no 2 pp 204–208 [41] A Alibeigloo and K M Liew 2014 Free Vibration Analysis of Sandwich Cylindrical Panel with Functionally Graded Core using Three-Dimensional Theory of Elasticity (Structures) vol 113 [42] Pravin Kulkarni, Ashwinkumar Dhoble and Pramod Padole 2018 Review of Research and Recent Trends in Analysis of Composite Plates (Sådhanå) vol 43 [43] Dawe D J, LAM S S E and Azizian Z G 1992 Nonlinear Finite Strip Analysis of Rectangular Laminates under End Shortening, using Classical Plate Theory (International Journal for Numerical Methods in Engineering) vol 35 pp 1087–1110 [44] S Yin, T Yu and P Liu 2013 Free Vibration Analyses of FGM Thin Plates by Isogeometric Analysis Based on Classical Plate Theory and Physical Neutral Surface (Advances in Mechanical Engineering) vol 5 [45] Y-W Kim 2005 Temperature Dependent Vibration Analysis of Functionally Graded Rectangular Plates (Journal of Sound and Vibration) vol 284 no 3–5 pp 531–549 [46] A Allahverdizadeh, M H Naei and M N Bahrami 2008 Nonlinear Free and Forced Vibration Analysis of Thin Circular Functionally Graded Plates (Journal of Sound and Vibration) vol 310 no 4 pp 966–984 [47] J N Reddy 2000 Analysis of Functionally Graded Plates (International Journal for Numerical Methods in Engineering) vol 47 pp 663–684 [48] Hiroyuki Matsunaga 2008 Free Vibration and Stability of Functionally Graded Plates According to A 2–D Higher–Order Deformation Theory (Composite Structures) vol 82 no 4 pp 499–512 [49] S Hosseini-Hashemi, H Rokni Damavandi Taher, H Akhavan and M Omidi 2010 Free Vibration of Functionally Graded Rectangular Plates Using First–Order Shear Deformation INTCSET 2020IOP Conf. Series: Materials Science and Engineering1094 (2021) 012081IOP Publishingdoi:10.1088/1757-899X/1094/1/01208118 Plate Theory (Journal of Applied Mathematical Modelling) vol 34 no 5 pp 1276–1291 [50] Hadji L, Atmane H A, Tounsi A, Mechab I and Bedia E A 2011 Free Vibration of Functionally Graded Sandwich Plates using Four–Variable Refined Plate Theory (Applied Mathematics and Mechanics) vol 32 pp 925–942 [51] Y Kiani, E Bagherizadeh and M R Eslami 2011 Thermal and Mechanical Buckling of Sandwich Plates with FGM Face Sheets Resting on the Pasternak Elastic Foundation (Proceedings of the Institution of Mechanical Engineers, Part C: Journal of Mechanical Engineering Science) vol 226 no 1 pp 32–41 [52] T Anderson 2003 A 3–D Elasticity Solution for a Sandwich Composite with Functionally Graded Core Subjected to Transverse Loading by a Rigid Sphere (Composite Structures) vol 60 no 3 [53] Jie Cui, Taoran Zhou, Renchuan Ye, Oleg Gaidai, Zichao Li and Shenghui Tao 2019 Three– Dimensional Vibration Analysis of a Functionally Graded Sandwich Rectangular Plate Resting on an Elastic Foundation Using a Semi–Analytical Method (Materials) vol 12 [54] Q Li, V Lu and K Kou 2008 Three–Dimensional Vibration Analysis of Functionally Graded Material Sandwich Plates (Journal of Sound and Vibration) vol 311 no 1–2 pp 498–515 [55] Fiorenzo A Fazzolari 2015 Natural Frequencies and Critical Temperatures of Functionally Graded Sandwich Plates Subjected to Uniform and Non–Uniform Temperature Distributions (Composite Structures) vol 121 pp 197–210 [56] A Bhimaraddi and L K Stevens 1984 A Higher Order Theory for Free Vibration of Orthotropic, Homogenous, and Laminated Rectangular Plates (ASME Journal of Applied Mechanics) vol 51 no 1 pp 195–198 [57] N F Hanna and A W Leissa 1994 A Higher Order Shear Deformation Theory for the Vibration of Thick Plates (Journal of Sound and Vibration) vol 170 pp 545–555 [58] Kirigulige M S, Kitey R and Tippur H V 2005 Dynamic Fracture Behavior of Model Sandwich Structures with Functionally Graded Core: a Feasibility Study (Composites Science and Technology) vol 65 no 7–8 pp 1052–1068 [59] Abdelaziz H H, Atmane H A, Mechab I, Boumia L, Tounsi A and Abbas A B E 2011 Static Analysis of Functionally Graded Sandwich Plates using an Efficient and Simple Refined Theory (Chinese Journal of Aeronautics) vol 24 no 4 pp 434–448 [60] Liao X L, Xu W F, Wang Y L, Jia, B and Zhou G Y 2009 Effect of Porous Structure on Mechanical Properties of C/PLA/Nano-HA Composites Scaffold (Transactions of nonferrous Metals Society of China) vol 19 pp s748–s751 [61] Yan Qing Wang and Jean W Zu 2017 Vibration Behaviors of Functionally Graded Rectangular Plates with Porosities and Moving in Thermal Environment (Aerospace Science and Technology) vol 69 pp 550–562 [62] Saidi Hayat and Sahla Meriem 2019 Vibration Analysis of Functionally Graded Plates with Porosity Composed of a Mixture of Aluminum (Al) and Alumina (Al2O3) Embedded in an Elastic Medium (Frattura ed Integrità Strutturale) vol 50 pp 286–299 [63] Xiang-Yu Zhang, Gang Fang, Sander Leeflang, Amir A Zadpoor and Jie Zhou 2019 Topological Design, Permeability and Mechanical Behavior of Additively Manufactured Functionally Graded Porous Metallic Biomaterials (Acta Biomaterialia) vol 84 pp 437–452 [64] Y H Dang, Y H Li, D Chen and J Yang 2018 Vibration Characteristics of Functionally Graded Graphene Reinforced Porous Nanocomposite Cylindrical Shells with Spinning Motion (Composites Part B: Engineering) vol 145 pp 1–13 [65] Nguyen, N V, Nguyen H X, Lee S and Nguyen-Xuan H 2018 Geometrically Nonlinear Polygonal Finite Element Analysis of Functionally Graded Porous Plates (Advances in Engineering Software) vol 126 pp 110–126 [66] Zhang Y and Wang J 2017 Fabrication of Functionally Graded Porous Polymer Structures using Thermal Bonding Lamination Techniques (Procedia Manufacturing) vol 10 pp 866–875 [67] Kiani, Y and Eslami M R 2012 Thermal Buckling and Post–Buckling Response of Imperfect Temperature–Dependent Sandwich FGM Plates Resting on Elastic Foundation (Archive of INTCSET 2020IOP Conf. Series: Materials Science and Engineering1094 (2021) 012081IOP Publishingdoi:10.1088/1757-899X/1094/1/01208119 Applied Mechanics) vol 82 pp 891–905 [68] Rezaei A S and Said A R 2015 Exact Solution for Free Vibration of Thick Rectangular Plates Made of Porous Materials (Composite Structures) vol 134 pp 1051–1060 [69] Merdaci S, Belmahi S, Belghoul H and Hadj Mostefa A 2019 Free Vibration Analysis of Functionally Graded Plates FG with Porosities (International Journal of Engineering Research & Technology) vol 8 no 3 pp143–147 [70] Kim J, Zur K K and Reddy J 2019 Bending, Free Vibration and Buckling of Modified Couples Stress-Based Functionally Graded Porous Micro–Plates (Composite Structures) vol 209 [71] Coskun S, Kim J and Toutanji H 2019 Bending, Free Vibration and Buckling Analysis of Functionally Graded Porous Micro–Plates Using a General Third–Order Plate Theory (Journal of Composites Science) vol 3 no 1 [72] Zhao J, Wang Q, Deng X, Choe K, Zhong R and Shuai C 2019 Free Vibrations of Functionally Graded Porous Rectangular Plate with Uniform Elastic Boundary Conditions (Composites Part B: Engineering) vol 168 pp 106–120 [73] S Merdaci 2019 Free Vibration Analysis of Composite Material Plates “Case of a Typical Functionally Graded FG Plates Ceramic/Metal” with Porosities (Nano Hybrids and Composites) vol 25 pp69–83 [74] S Chakraverty and K K Pradhan 2014 Free Vibration of Functionally Graded Thin Rectangular Plates Resting on Winkler Elastic Foundation with General Boundary Conditions using Rayleigh–Ritz Method (International Journal of Applied Mechanics) vol 6 no 4 [75] Nuttawit Wattanasakulponga and Variddhi Ungbhakorn 2014 Linear and Nonlinear Vibration Analysis of Elastically Restrained Ends FGM Beams with Porosities (Aerospace Science and Technology) vol 32 no 1 pp 111–120 [76] S Jafari Mehrabadi and B Sobhani Aragh 2014 Stress Analysis of Functionally Graded Open Cylindrical Shell Reinforced by Agglomerated Carbon Nanotubes (Thin–Walled Structures) vol 80 pp 130–141 [77] Baferani A H, Saidi A R and Jomehzadeh E 2015 An Exact Solution for Free Vibration of Thin Functionally Graded Rectangular Plates (Proceedings of the Institution of Mechanical Engineers, Part C: Journal of Mechanical Engineering Science) vol 225 no 3 pp 526–536 [78] Shen H S 2007 Nonlinear Thermal Bending Response of FGM Plates Due to Heat Conduction (Composites Part B: Engineering) vol 38 no 2 pp 201–215 [79] Yang J and Shen H S 2002 Dynamic Response of Initially Stressed Functionally Graded Rectangular Thin Plates (Composite Structures) vol 54 no 4 pp 497–508 [80] S A Ambartsumyan, J E Ashton and T Cheron 1970 Theory of Anisotropic Plates (Technomic Publishing Company) [81] Li S R, Zhang J H and Zhao Y G 2007 Nonlinear Thermomechanical Post–Buckling of Circular FGM Plate with Geometric Imperfection (Thin–Walled Structures) vol 45 no 5 pp 528–536 [82] Zhang D and Zhou Y 2008 A Theoretical Analysis of FGM Thin Plates Based on Physical Neutral Surface (Computational Material Science) Vo 44 pp 716–720 [83] Snehashish Chakraverty and Karan Kumar Pradhan 2016 Vibration of Functionally Graded Beams and Plates (Science Direct) [84] Chi S and Chung Y 2006 Mechanical Behavior of Functionally Graded Material Plates under Transverse Load Part I: Analysis (International Journal of Solids and Structures) vol 43 no 13 [85] Latifi M, Farhatnia F and Kadkhodaei M 2013 Bucking Analysis of Rectangular Functionally Graded Plates under Various Edge Conditions using Fourier Series Expansion (Europe an Journal of Mechanics–A/Solids) vol 41 pp 16–27 [86] R D Mindlin 1951 Influence of Rotatory Inertia and Shear on Flexural Motions of Isotropic, Elastic Plates (ASME Journal of Applied Mechanics) vol 18 pp 31–38 [87] E Reissner 1945 The Effect of Transverse Shear Deformation on the Bending of Elastic Plates (ASME Journal of Applied Mechanics) vol 12 pp 69–77 [88] Carrera E and Brischetto S Robaldo A 2008 Variable Kinematic Model for the Analysis of Functionally Graded Material Plates (AIAA Journal) vol 46 no 1 pp 194–203 INTCSET 2020IOP Conf. Series: Materials Science and Engineering1094 (2021) 012081IOP Publishingdoi:10.1088/1757-899X/1094/1/01208120[89] Duc D N and Cong P H 2015 Nonlinear Vibration of Thick FGM Plates on Elastic Foundation Subjected to Thermal and Mechanical Loads Using the First–Order Shear Deformation Plate Theory (Cogent Engineering) vol 2 no 1 pp 1–17 [90] R Benferhat, T Hassaine Daouadji and M Said Mansour 2016 Free Vibration Analysis of FG Plates Resting on an Elastic Foundation and Based on The Neutral Surface Concept Using Higher–Order Shear Deformation Theory (Comptes Rendus Mecanique) vol 344 no 9 pp 631–641 [91] M Karami Khorramabadi, M M Najafizadeh, J Alibabaei Shahraki and P Khazaeinejad 2008 Effect of Shear Theories on Free Vibration of Functionally Graded Plates (World Academy of Science, Engineering and Technology) vol 24 [92] Trung–Kien Nguyen, Karam Sab and Guy Bonnet 2008 First–Order Shear Deformation Plate Models for Functionally Graded Materials (Composite Structures) vol 83 no 1 pp 25–36 [93] M Aydogdu 2006 Comparison of Various Shear Deformation Theories for Bending, Buckling, and Vibration of Rectangular Symmetric Cross–ply Plate with Simply Supported Edges (Journal of Composite Materials) vol 40 [94] D G Zhang 2013 Modeling and Analysis of FGM Rectangular Plates Based on Physical Neutral Surface and High Order Shear Deformation Theory (International Journal of Mechanical Sciences) vol 68 pp 92–104 [95] Hosseini-Hashemi S, Fadaee M and Taher H R D 2011 Exact Solutions for Free Flexural Vibration of Levy–Type Rectangular Thick Plates Via Third Order Shear Deformation Plate Theory (Applied Mathematical Modelling) vol 35 no 2 pp 708–727 [96] Nuttawit Wattanasakulpong ,Gangadhara Prusty and Donald W Kelly 2011 Thermal Buckling and Elastic Vibration of Third–Order Shear Deformable Functionally Graded Beams (International Journal of Mechanical Sciences) vol 53 no 9 pp 734–743 [97] Shimpi R P and Patel H G 2006 Free Vibrations of Plate using Two Variable Refined Plate Theory (Journal of Sound and Vibration) vol 296 pp 979–999 [98] A K noor, W S Burton 1989 Assessment of Shear Deformation Theories for Multilayered Composite Plates (Applied Mechanics Reviews) vol 42 pp 1–13 [99] Fakhari V, Ohad A and Yousefian P 2011 Nonlinear Free and Forced Vibration Behavior of Functionally Graded Plate with Piezoelectric Layers in Thermal Environment (Composite Structures) vol 93 no 9 pp 2310–2321 [100] Neves A M A, Ferreira A J M, Carrera E, Roque C M C, Cinefra M and Jorge R M N 2012 A Quasi–3D Sinusoidal Shear Deformation Theory for the Static and Free Vibration Analysis of Functionally Graded Plates (Composites Part B: Engineering) vol 43 no 2 pp 711–725 [101] J M Whitney and C T Sun 1973 A Higher Order Theory for Extensional Motion of Laminated Composites (Journal of Sound and Vibration) vol 30 no 1 pp 85–97 [102] K H Lo, R M Christensen and E M Wu 1977 A High–Order Theory of Plate Deformation Part 1: Homogeneous Plates (Journal of Applied Mechanics) vol 44 no 4 pp 663–668 [103] T Kant 1982 Numerical Analysis of Thick Plates (Computer Methods in Applied Mechanics and Engineering) vol 31 no 1 pp 1–18 [104] A Bhimaraddi and Stevens L K 1984 A Higher Order Theory for Free Vibration of Orthotropic, Homogeneous, And Laminated Rectangular Plates (Journal of Applied Mechanics) vol 51 no 1 [105] J N Reddy and C D Chin 1998 Thermoelastical Analysis of Functionally Graded Cylinders and Plates (Journal of Thermal Stresses) vol 21 no 6 pp 593–626 [106] Reddy J N 2011 A General Nonlinear Third–Order Theory of Functionally Graded Plates (International Journal of Aerospace and Light weight Structures) vol 1 no 1 pp 1–21 [107] Jha D K, Kant T and Singh R K 2013 Free Vibration Response of Functionally Graded Thick Plates with Shear and Normal Deformations Effects (Composite Structures) vol 96 pp 799– 823 [108] M Talha and B N Singh 2010 Static Response and Free Vibration Analysis of FGM Plates Using Higher Order Shear Deformation Theory (Applied Mathematical Modelling) vol 34 no 12 INTCSET 2020IOP Conf. Series: Materials Science and Engineering1094 (2021) 012081IOP Publishingdoi:10.1088/1757-899X/1094/1/01208121[109] Natarajan S and Manickam G 2012 Bending and Vibration of Functionally Graded Material Sandwich Plates using an Accurate Theory (Finite Elements in Analysis and Design) vol 57 [110] Mohammadi Y and Khalili S M R 2012 Free Vibration Analysis of Sandwich Plates with Temperature–Dependent Properties of the Core Materials and Functionally Graded Face Sheets (Mechanics and Properties of Composed Materials and Structures, Advanced Structured Materials) vol 31 pp 183–197 [111] S M R Khalili and Y Mohammadi 2012 Free Vibration Analysis of Sandwich Plates with Functionally Graded Face Sheets and Temperature–Dependent Material Properties: A New Approach (European Journal of Mechanics A/Solids) vol 35 pp 61–74 [112] Shimpi R P 1999 Zeroth–Order Shear Deformation Theory for Plates (AIAA Journal) vol 37 no 4 [113] Ray M C 2003 Zeroth–Order Shear Deformation Theory for Laminated Composite Plates (Journal of Applied Mechanics) vol 70(3) pp 374–380 [114] Bounouara F, K H Benrahou, I Belkorissat and A Tounsi 2016 A Nonlocal Zeroth–Order Shear Deformation Theory for Free Vibration of Functionally Graded Nanoscale Plates Resting on Elastic Foundation (Steel and Composite Structures) vol 20 no 2 pp 227–249 [115] Huu-Tai Thai and Dong-Ho Choi 2014 Zeroth-Order Shear Deformation Theory for Functionally Graded Plates Resting on Elastic Foundation (International Journal of Mechanical Sciences) vol 78 pp 35–43 [116] Muhsin J Jweeg 2016 A Suggested Analytical Solution for Vibration of Honeycombs Sandwich Combined Plate Structure (International Journal of Mechanical & Mechatronics Engineering) vol 16 no 4 [117] Bateni M, Kiani Y and M R Eslami 2013 A Comprehensive Study on Stability of FGM Plates (International Journal of Mechanical Sciences) vol 75 pp 134–144 [118] H Mozafari and A Ayob 2012 Effect of Thickness Variation on the Mechanical Buckling Load in Plates Made of Functionally Graded Materials (Procedia Technology) vol 1 pp 496–504 [119] Ashraf Zenkour, Farzad Ebrahimi and Mohammad Reza Barati 2019 Buckling Analysis of a Size–Dependent Functionally Graded Nanobeam Resting on Pasternak's Foundations (International Journal of Nano Dimension) vol 10 no 2 pp 141–153 [120] Y Huang and X F Li 2010 Buckling of Functionally Graded Circular Columns Including Shear Deformation (Materials & Design) vol 31 no 7 pp 3159–3166 [121] Wang C M and Reddy J N, Lee K H 2000 Shear Deformable Beams and Plates: Relationships with Classical Solutions (Elsevier) [122] B A Samsam Shariat and MR Eslam 2007 Buckling of Thick Functionally Graded Plates under Mechanical and Thermal Loads (Composite Structures) vol 78 no 3 pp 433–439 [123] V Birman 1995 Buckling of Functionally Graded Hybrid Composite Plates (Engineering mechanics) pp 1199–1202 [124] YY Lee, X Zhao and J N Reddy 2010 Postbuckling Analysis of Functionally Graded Plates Subject to Compressive and Thermal Loads (Computer Methods in Applied Mechanics and Engineering) vol 199 no 25–28 pp 1645–1653 [125] L Wu 2004 Thermal Buckling of A Simply Supported Moderately Thick Rectangular FGM Plate (Composite Structures) vol 64 no 2 pp 211–218 [126] Czechowski L and Kołakowski Z 2019 The Study of Buckling and Post–Buckling of a Step– Variable FGM Box (Materials) vol 12 no 6 [127] R Javaheri and M R Eslami 2002 Buckling of Functionally Graded Plates under In–plane Compressive Loading (Journal of applied mathematics and mechanics) vol 82 no 4 [128] Bekir Akgöz and Ömer Civalek 2013 Buckling Analysis of Functionally Graded Microbeams Based on the Strain Gradient Theory (Acta Mechanica) vol 224 pp 2185–2201 [129] Fekrar A, El Meiche N, Bessaim A, Tounsi A and Adda Bedia E A 2012 Buckling Analysis of Functionally Graded Hybrid Composite Plates Using a New Four Variable Refined Plate Theory (Steel and Composite Structures) vol 13 no 1 pp 91–107 [130] Shi-Rong Li, Bouazza Fahsi Abdelouahed Tounsi and Samy R Mahmoud 2015 Correspondence Relations Between Deflection, Buckling Load, and Frequencies of Thin Functionally Graded INTCSET 2020IOP Conf. Series: Materials Science and Engineering1094 (2021) 012081IOP Publishingdoi:10.1088/1757-899X/1094/1/01208122 Material Plates and Those of Corresponding Homogeneous Plates (Journal of Applied Mechanics) vol 82 no 11 [131] Soldatos K P 1992 A Transverse Shear Deformation Theory for Homogeneous Monoclinic Plates (Acta Mechanica) vol 94 pp195–220 [132] D U M Oyekoya and A M El-Zafranyet 2009 Buckling and Vibration Analysis of Functionally Graded Composite Structures using the Finite Element Method (Composite Structures) vol 89 no 1 pp 134–142 [133] Bodaghi M and Saidi A R 2010 Levy–Type Solution for Buckling Analysis of Thick Functionally Graded Rectangular Plates Based on the Higher–Order Shear Deformation Plate Theory (Applied Mathematical Modelling) vol 34 no 11 pp 3659–3673 [134] M M Najafizadeh and M R Eslamia 2002 Buckling Analysis of Circular Plates of Functionally Graded Materials under Uniform Radial Compression (International Journal of Mechanical Sciences) vol 44 pp 2479–2493 [135] B Sidda Reddy, J Suresh Kumar, C Eswara Reddy and K Vijaya Kumar Reddy 2013 Buckling Analysis of Functionally Graded Material Plates Using Higher Order Shear Deformation Theory (Journal of Composites) vol 2013 [136] Farzad Ebrahimi and Fateme Mahmood 2018 A Modified Couple Stress Theory for Buckling Analysis of Higher Order Inhomogeneous Microbeams with Porosities (Proceedings of the Institution of Mechanical Engineers, Part C: Journal of Mechanical Engineering Science) vol 233 no 8 [137] Duc N D and Tung H V 2011 Mechanical and Thermal Postbuckling of Higher Order Shear Deformable Functionally Graded Plates on Elastic Foundations (Composite Structures) vol 93 no 11 pp 2874–2881 [138] Zenkour and A M 2005 A Comprehensive Analysis of Functionally Graded Sandwich Plates: Part 2–Buckling and Free Vibration (International Journal of Solids and Structures) vol 42 no 18–19 [139] Elias Y Ali and Yared S Bayleyegn 2019 Analytical and Numerical Buckling Analysis of Rectangular Functionally–Graded Plates under Uniaxial Compression (Proceedings of the Annual Stability Conference Structural Stability Research Council St Louis, Missouri) [140] H Akhavan, Sh Hosseini Hashemi, H Rokni Damavandi Taher, A Alibeigloo and Sh Vahabi 2009 Exact Solutions for Rectangular Mindlin Plates under in–Plane Loads Resting on Pasternak Elastic Foundation Part I: Buckling Analysis (Computational Materials Science) vol 44 no 1 [141] noureddine El Meiche, Abdelouahed Tounsi, noureddine Ziane, Ismail Mechab and El Abbes Adda Bedia 2011 A New Hyperbolic Shear Deformation Theory for Buckling and Vibration of Functionally Graded Sandwich Plate (International Journal of Mechanical Sciences) vol 53 no 10 pp 237–247 [142] Hui-Shen Shen and Shi-Rong Li 2008 Post–Buckling of Sandwich Plates with FGM Face Sheets and Temperature–Dependent Properties (Composites Part B: Engineering) vol 39 no 2 [143] Ho ang Van Tung and Nguyen Dinh Duc 2009 Nonlinear Analysis of Stability for Functionally Graded Plates under Mechanical and Thermal Loads (Composite Structures) vol 92 no 5 [144] J S Park and J H Kim 2006 Thermal Postbuckling and Vibration Analyses of Functionally Graded Plates (Journal of Sound and Vibration) vol 289 no 1–2 pp 77–93 [145] Lee Y H, Bae S I and Kim J H 2016 Thermal Buckling Behavior of Functionally Graded Plates Based on Neutral Surface (Composite Structures) vol 137 pp 208–214 [146] Da-Guang Zhang and Hao-Miao Zhou 2015 Nonlinear Bending Analysis of FGM Circular Plates Based on Physical Neutral Surface and Higher–Order Shear Deformation Theory (Aerospace Science and Technology) vol 41 pp 90–98 [147] M Sobhy 2013 Buckling and Free Vibration of Exponentially Graded Sandwich Plates Resting on Elastic Foundations under Various Boundary Conditions (Composite Structures) vol 99 [148] Hessameddin Yaghoobi and Pooria Yaghoobi 2013 Buckling Analysis of Sandwich Plates with FGM Face Sheets Resting on Elastic Foundation with Various Boundary Conditions: An INTCSET 2020IOP Conf. Series: Materials Science and Engineering1094 (2021) 012081IOP Publishingdoi:10.1088/1757-899X/1094/1/01208123 Analytical Approach (Meccanica) vol 48 pp 2019–2035 [149] X Zhao, Y Y Lee and K M Liew 2009 Mechanical and Thermal Buckling Analysis of Functionally Graded Plates (Composite Structures) vol 90 no 2 pp 161–171 [150] S K Jalali, M H Naei and A Poorsolhjouy 2010 Thermal Stability Analysis of Circular Functionally Graded Sandwich Plates of Variable Thickness using Pseudo–Spectral Method (Materials and Design) vol 31 no 10 pp 4755–4763 [151] Tiantang Yu, Tinh Quoc Bui, Shuohui Yin, Duc Hong Doan, C T Wu, Thom Van Do and Satoyuki Tanaka 2016 On the Thermal Buckling Analysis of Functionally Graded Plates with Internal Defects using Extended Isogeometric Analysis (Composite Structures) vol 136 [152] S Sirinivas and A K Rao 1970 Bending, Vibration, and Buckling of Simply Supported Thick Orthotropic Rectangular Plates and Laminates (International Journal of Solids and Structures) vol 6 no 11 pp 1463–1481 [153] Na K S and Kim J H 2004 Three–Dimensional Thermal Buckling Analysis of Functionally Graded Materials (Composites Part B: Engineering) vol 35 no 5 pp 429–437 [154] Na K S and Kim J H 2006 Thermal Postbuckling Investigations of Functionally Graded Plates Using 3–D Finite Element Method (Finite Elements in Analysis and Design) vol 42 no 8–9 [155] Ehab N Abbas, Muhsin J Jweeg and Muhannad Al-Waily 2018 Analytical and Numerical Investigations for Dynamic Response of Composite Plates Under Various Dynamic Loading with the Influence of Carbon Multi-Wall Tube Nano Materials (International Journal of Mechanical & Mechatronics Engineering) vol 18 no 06 pp 1–10 [156] Jumaa S Chiad, Muhannad Al-Waily and Mohsin Abdullah Al-Shammari 2018 Buckling Investigation of Isotropic Composite Plate Reinforced by Different Types of Powders (International Journal of Mechanical Engineering and Technology) vol 09 no 09 [157] Muhannad Al-Waily 2015 Analytical and Numerical Thermal Buckling Analysis Investigation of Unidirectional and Woven Reinforcement Composite Plate Structural (International Journal of Energy and Environment) vol 6 no 2 [158] M Mohammadi, A R Saidi and E Jomehzadeh 2010 A Novel Analytical Approach for the Buckling Analysis of Moderately Thick Functionally Graded Rectangular Plates with Two Simply Supported Opposite Edges (Proceedings of the Institution of Mechanical Engineers C) vol 224 no 9 [159] M Darvizeh, A Darvizeh, R Ansari and C B Sharma 2004 Buckling Analysis of Generally Laminated Composite Plates (generalized differential quadrature rules versus Rayleigh–Ritz method) (Composite Structures) vol 63 pp 69–74 [160] R Javaheri and M R Eslami 2002 Thermal Buckling of Functionally Graded Plates Based on Higher Order Theory (Journal of Thermal Stresses) vol 25 no 7 pp 603–625 [161] R Javaheri and M R Eslami 2002 Thermal Buckling of Functionally Graded Plates (AIAA Journal) vol 40 no 1 pp 162–169 [162] Muhannad Al-Waily, Mohsin Abdullah Al-Shammari and Muhsin J Jweeg 2020 An Analytical Investigation of Thermal Buckling Behavior of Composite Plates Reinforced by Carbon Nano Particles (Engineering Journal) vol 24 no 3 [163] Mohsin Abdullah Al-Shammari and Muhannad Al-Waily 2018 Analytical Investigation of Buckling Behavior of Honeycombs Sandwich Combined Plate Structure (International Journal of Mechanical and Production Engineering Research and Development) vol 08 no 04 [164] Muhsin J Jweeg 1983 Application of Finite Element Analysis to Rotating Fan Impellers (Doctoral Thesis, Aston University) [165] Muhsin J Jweeg, Ali S Hammood and Muhannad Al-Waily 2012 A Suggested Analytical Solution of Isotropic Composite Plate with Crack Effect (International Journal of Mechanical & Mechatronics Engineering) vol 12 no 05 [166] S H Bakhy, S S Hassan, S M Nacy, K Dermitzakis and A H Arieta 2013 Contact Mechanics for Soft Robotic Fingers: Modeling And Experimentation (Robotica) [167] Mohsin Abdullah Al-Shammari, Emad Q Hussein and Ameer Alaa Oleiwi 2017 Material Characterization and Stress Analysis of a Through Knee Prosthesis Sockets (International Journal of Mechanical & Mechatronics Engineering) vol 17 no 06 INTCSET 2020IOP Conf. Series: Materials Science and Engineering1094 (2021) 012081IOP Publishingdoi:10.1088/1757-899X/1094/1/01208124[168] Ameer A Kadhim, Muhannad Al-Waily, Zaman Abud Almalik Abud Ali, Muhsin J Jweeg and Kadhim K Resan 2018 Improvement Fatigue Life and Strength of Isotropic Hyper Composite Materials by Reinforcement with Different Powder Materials (International Journal of Mechanical & Mechatronics Engineering) vol 18 no 02 [169] Lara E Yousif, Kadhim K Resan and Raad M Fenjan 2018 Temperature Effect on Mechanical Characteristics of A New Design Prosthetic Foot (International Journal of Mechanical Engineering and Technology) vol 09 no 13 pp 1431–1447 [170] Ragad Aziz Neama, Maher AR Sadiq Al-Baghdadi and Muhannad Al-Waily 2018 Effect of Blank Holder Force and Punch Number on the Forming Behavior of Conventional Dies (International Journal of Mechanical & Mechatronics Engineering) vol 18 no 04 [171] noor Dhia Yaseen, Jumaa S Chiad and Firas Mohammed Abdul Ghani 2018 The Study and Analysis of Stress Distribution Subjected on the Replacement Knee Joint Components using Photo-Elasticity and Numerical Methods (International Journal of Mechanical and Production Engineering Research and Development) vol 08 no 06 pp 449–464 [172] Ayad M Takhakh and Saif M Abbas 2018 Manufacturing and Analysis of Carbon Fiber Knee Ankle Foot Orthosis (International Journal of Engineering & Technology) vol 07 no 04 [173] H J Abbas, M J Jweeg, Muhannad Al-Waily and Abbas Ali Diwan 2019 Experimental Testing and Theoretical Prediction of Fiber Optical Cable for Fault Detection and Identification (Journal of Engineering and Applied Sciences) vol 14 no 02 pp 430–438 [174] Nada N Kadhim, Qahtan A Hamad and Jawad K Oleiwi 2020 Tensile and Morphological Properties of PMMA Composite Reinforced by Pistachio Shell Powder used in Denture Applications (2nd International Conference on Materials Engineering & Science, AIP Conference Proceedings) [175] Muhannad Al-Waily, Iman Q Al Saffar, Suhair G Hussein and Mohsin Abdullah Al-Shammari 2020 Life Enhancement of Partial Removable Denture made by Biomaterials Reinforced by Graphene Nanoplates and Hydroxyapatite with the Aid of Artificial Neural Network (Journal of Mechanical Engineering Research and Developments) vol 43 no 6 pp 269–285 [176] Fahad M Kadhim, Jumaa S Chiad and Maryam Abdul Salam Enad 2020 Evaluation and Analysis of Different Types of Prosthetic Knee Joint Used by above Knee Amputee (Defect and Diffusion Forum Journal) vol 398 pp 34–40 [177] S K Mahmood, S H Bakhy and M A Tawfik 2020 novel Wall-Climbing Robot Capable of Transitioning and Perching (IOP Conference Series: Materials Science and Engineering) [178] Najdat A Mahmood, Muhsin J Jweeg and Mumtaz Y Rajab 1989 Investigation of Partially Pressurized Thick Cylindrical Shells (Modelling, simulation & control B AMSE Press) vol 25 no 03 pp 47–64 [179] Ekhlas M Alfayyadh, Sadeq H Bakhy and Yasir M Shkara 2014 A New Multi-Objective Evolutionary Algorithm for Optimizing the Aerodynamic Design of HAWT Rotor (ASME 2014 12th Biennial Conference on Engineering Systems Design and Analysis, ESDA 2014) [180] Muhannad Al-Waily and Zaman Abud Almalik Abud Ali 2015 A Suggested Analytical Solution of Powder Reinforcement Effect on Buckling Load for Isotropic Mat and Short Hyper Composite Materials Plate (International Journal of Mechanical and Mechatronics Engineering) vol 15 no 4 [181] Rasha Hayder Al-Khayat, Maher A R Sadiq Al-Baghdadi, Ragad Aziz Neama and Muhannad Al-Waily 2018 Optimization CFD Study of Erosion in 3D Elbow During Transportation of Crude Oil Contaminated with Sand Particles (International Journal of Engineering and Technology) vol 07 no 03 pp 1420–1428 [182] Jawad K Oleiwi and Basim A Abass 2018 Thermal Properties of Polymeric Composites Reinforced by Nanoceramic Materials (International Journal of Mechanical and Production Engineering Research and Development) vol 8 no 6 pp 517–524 [183] Muhsin J Jweeg, Muhannad Al-Waily, Ahmed K Muhammad and Kadhim K Resan 2018 Effects of Temperature on the Characterisation of a New Design for a non–Articulated Prosthetic Foot (IOP Conference Series: Materials Science and Engineering) vol 433 2nd International Conference on Engineering Sciences, Kerbala, Iraq INTCSET 2020IOP Conf. Series: Materials Science and Engineering1094 (2021) 012081IOP Publishingdoi:10.1088/1757-899X/1094/1/01208125[184] Abeer R Abbas, Kadhim A Hebeatir and Kadhim K Resan 2018 Effect of Laser Energy on the Structure of Ni46–Ti50–Cu4 Shape–Memory Alloy (International Journal of Nanoelectronics and Materials) vol 11 no 04 pp 481–498 [185] Sihama I Salih, Jawad Kadhim Oleiwi and Sajid Abd Alkhidhir 2018 Comparative Study of Some Mechanical Properties of Hybrid Polymeric Composites Prepared by using Friction Stir Processing (Journal of Advanced Research in Dynamic and Control Systems) vol 10 no 02 [186] Ayad M Takhakh, Saif MAbbas and AseelK Ahmed 2018 A Study of the Mechanical Properties and Gait Cycle Parameter for a Below–Knee Prosthetic Socket (IOP Conference Series: Materials Science and Engineering) vol 433 2nd International Conference on Engineering Sciences [187] Muhannad Al-Waily, Emad Q Hussein and Nibras A Aziz Al-Roubaiee 2019 Numerical Modeling for Mechanical Characteristics Study of Different Materials Artificial Hip Joint with Inclination and Gait Cycle Angle Effect (Journal of Mechanical Engineering Research & Developments) vol 42 no 04 pp 79–93 [188] Muhsin J Jweeg, Abdulkareem Abdulrazzaq Ahumdany and Ali Faik Mohammed Jawad 2019 Dynamic Stresses and Deformations Investigation of the Below Knee Prosthesis using CT- Scan Modeling (International Journal of Mechanical & Mechatronics Engineering) vol 19 no 01 [189] Akeel Z Mahdi, Samir A Amin and Sadeq H Bakhy 2020 Influence of Refill Friction Stir Spot Welding Technique on the Mechanical Properties and Microstructure of Aluminum AA5052 and AA6061-T3 (3rd International Conference on Engineering Sciences) vol 671 IOP Conference Series: Materials Science and Engineering [190] Muhsin J Jweeg, Salah N Alnomani and Salah K Mohammad 2020 Dynamic Analysis of a Rotating Stepped Shaft with and without Defects (3rd International Conference on Engineering Sciences) vol 671 IOP Conference Series: Materials Science and Engineering [191] Fahad M Kadhim, Ayad M Takhakh and Jumaa S Chiad 2020 Modeling and Evaluation of Smart Economic Transfemral Prosthetic (Defect and Diffusion Forum Journal) vol 398 pp 48–53 [192] J N Reddy 1993 An Introduction to the Finite Element Method (McGraw-Hill, Inc) [193] Muhsin J Jweeg and S Z Said 1995 Effect of Rotational and Geometric Stiffness Matrices on Dynamic Stresses and Deformations of Rotating Blades (Journal of the Institution of Engineers (India): Mechanical Engineering Division) vol 76 pp 29–38 [194] S S Rao 2004 The Finite Element Method in Engineering (Elsevier Science and Technology Books, Miami) [195] Bashar A Bedaiwi and Jumaa S Chiad 2012 Vibration Analysis and Measurement in the Below Knee Prosthetic Limb Part I: Experimental Work (ASME 2012 International Mechanical Engineering Congress and Exposition, Proceedings) [196] S H Bakhy 2014 Modeling of Contact Pressure Distribution and Friction Limit Surfaces for Soft Fingers in Robotic Grasping (Robotica) [197] Muhannad Al-Waily, Alaa Abdulzahra Deli, Aziz Darweesh Al-Mawash and Zaman Abud Almalik Abud Ali 2017 Effect of Natural Sisal Fiber Reinforcement on the Composite Plate Buckling Behavior (International Journal of Mechanical & Mechatronics Engineering) vol 17 no 01 [198] Saif M Abbas, Ayad M Takhakh, Mohsin Abdullah Al-Shammari and Muhannad Al-Waily 2018 Manufacturing and Analysis of Ankle Disarticulation Prosthetic Socket (SYMES) (International Journal of Mechanical Engineering and Technology) vol 09 no 07 pp 560–569 [199] Mohsin Abdullah Al-Shammari, Lutfi Y Zedan and Akram M Al-Shammari 2018 FE Simulation of Multi-Stage Cold Forging Process for Metal Shell of Spark Plug Manufacturing (1st International Scientific Conference of Engineering Sciences) 3rd Scientific Conference of Engineering Science, ISCES 2018–Proceedings [200] Worood Hussein, Mohsin Abdullah Al-Shammari 2018 Fatigue and Fracture Behaviours of FSW and FSP Joints of AA5083-H111 Aluminium Alloy (IOP Conference Series: Materials INTCSET 2020IOP Conf. Series: Materials Science and Engineering1094 (2021) 012081IOP Publishingdoi:10.1088/1757-899X/1094/1/01208126 Science and Engineering) vol 454 International Conference on Materials Engineering and Science [201] Muhsin J Jweeg, Kadhim K Resan, Esraa A Abbod and Muhannad Al-Waily 2018 Dissimilar Aluminium Alloys Welding by Friction Stir Processing and Reverse Rotation Friction Stir Processing (IOP Conference Series: Materials Science and Engineering) vol 454 International Conference on Materials Engineering and Science, Istanbul, Turkey, 8 August [202] Yousuf Jamal Mahboba and Mohsin Abdullah Al-Shammari 2019 Enhancing Wear Rate of High-Density Polyethylene (HDPE) by Adding Ceramic Particles to Propose an Option for Artificial Hip Joint Liner (IOP Conference Series: Materials Science and Engineering, ICMSMT) vol 561 [203] Sihama I Salih, Jawad K Oleiwi and Hajir Mohammed Ali 2019 Investigation the Properties of Silicone Rubber Blend Reinforced by Natural Nanoparticles and UHMWPE Fiber (International Journal of Mechanical Engineering and Technology) vol 10 no 1 [204] Sadiq emad Sadiq, Sadeq Hussein Bakhy and Muhsin Jaber Jweeg 2020 Effects of Spot Welding Parameters on the Shear Characteristics of Aluminum Honeycomb Core Sandwich Panels in Aircraft structure (Test Engineering and Management) vol 83 pp 7244–7255 [205] Mohsin Abdullah Al-Shammari, Qasim H Bader, Muhannad Al-Waily and A M Hasson 2020 Fatigue Behavior of Steel Beam Coated with Nanoparticles under High Temperature (Journal of Mechanical Engineering Research and Developments) vol 43 no 4 pp 287–298 [206] Ghaith G Hameed, Muhsin J Jweeg and Ali Hussein 2009 Springback and Side Wall Curl of Metal Sheet in Plain Strain Deep Drawing (Research Journal of Applied Sciences) vol 04 no 05 [207] Mahmud Rasheed Ismail, Muhannad Al-Waily and Ameer A Kadhim 2018 Biomechanical Analysis and Gait Assessment for normal and Braced Legs (International Journal of Mechanical & Mechatronics Engineering) vol 18 no 03 [208] Marwah Mohammed Abdulridha, Nasreen Dakel Fahad, Muhannad Al-Waily and Kadhim K Resan 2018 Rubber Creep Behavior Investigation with Multi Wall Tube Carbon Nano Particle Material Effect (International Journal of Mechanical Engineering and Technology) vol 09 no 12 pp 729–746 [209] Sadeq Bakhy, Enass Flaieh and Mortada Jabbar 2018 An Experimental Study for Grasping and Pinching Controls for an Underactuated Robotic Finger using a PID Controller (2nd International Conference on Engineering Sciences), (IOP Conference Series: Materials Science and Engineering) vol 433 [210] Ahmed Khaleel Abdulameer and Mohsin Abdullah Al-Shammari 2018 Fatigue Analysis of Syme’s Prosthesis (International Review of Mechanical Engineering) vol 12 no 03 [211] Mohsin Abdullah Al-Shammari 2018 Experimental and FEA of the Crack Effects in a Vibrated Sandwich Plate (Journal of Engineering and Applied Sciences) vol 13 no 17 pp 7395–7400 [212] Abeer R Abbas, Kadhim A Hebeatir and Kadhim K Resan 2018 Effect of CO2 Laser on Some Properties of NI46TI50CU4 Shape Memory Alloy (International Journal of Mechanical and Production Engineering Research and Development) vol 08 no 02 pp 451–460 [213] Sihama Issa Salih, Jawad Kadhim Oleiwi and Arkan Saad Mohamed 2018 Investigation of Mechanical Properties of Pmma Composite Reinforced with Different Types of Natural Powders (ARPN Journal of Engineering and Applied Sciences) vol 13 no 22 [214] Fahad M Kadhim, Ayad M Takhakh and Asmaa M Abdullah 2019 Mechanical Properties of Polymer with Different Reinforcement Material Composite That used for Fabricates Prosthetic Socket (Journal of Mechanical Engineering Research and Developments) vol 42 no 4 [215] Ehab N Abbas, Muhsin J Jweeg and Muhannad Al-Waily 2020 Fatigue Characterization of Laminated Composites used in Prosthetic Sockets Manufacturing (Journal of Mechanical Engineering Research and Developments) vol 43 no 5 pp 384–399 [216] Suhair Ghazi Hussein, Mohsin Abdullah Al-Shammari, Ayad M Takhakh and Muhannad Al- Waily 2020 Effect of Heat Treatment on Mechanical and Vibration Properties for 6061 and 2024 Aluminum Alloys (Journal of Mechanical Engineering Research and Developments) vol INTCSET 2020IOP Conf. Series: Materials Science and Engineering1094 (2021) 012081IOP Publishingdoi:10.1088/1757-899X/1094/1/01208127 43 no 01 pp 48–66 [217] Mortada A Jabbar, Sadeq H Bakhy and Enass H Flaieh 2020 A New Multi-Objective Algorithm for Underactuated Robotic Finger During Grasping and Pinching Assignments (3rd International Conference on Engineering Sciences, IOP Conference Series: Materials Science and Engineering) vol 671 [218] Marwah Ali Husain and and Mohsin Abdullah Al-Shammari 2020 Analytical Solution of Free Vibration Characteristics of Partially Circumferential Cracked Cylindrical Shell (Journal of Mechanical Engineering Research and Developments) vol 43 no 3 pp 442–454 [219] Hussein I Mansoor, Mohsin Al-shammari and Amjad Al-Hamood 2020 Theoretical Analysis of the Vibrations in Gas Turbine Rotor (3rd International Conference on Engineering Sciences) , IOP Conference Series: Materials Science and Engineering) vol 671 [220] Muhannad Al-Waily, Maher AR Sadiq Al-Baghdadi and Rasha Hayder Al-Khayat 2017 Flow Velocity and Crack Angle Effect on Vibration and Flow Characterization for Pipe Induce Vibration (International Journal of Mechanical & Mechatronics Engineering) vol 17 no 05 [221] Fahad M Kadhim, Jumaa S Chiad and Ayad M Takhakh 2018 Design And Manufacturing Knee Joint for Smart Transfemoral Prosthetic (IOP Conference Series: Materials Science and Engineering, International Conference on Materials Engineering and Science) vol 454 [222] Kadhim K Resan, Abbas A Alasadi, Muhannad Al-Waily and Muhsin J Jweeg 2018 Influence of Temperature on Fatigue Life for Friction Stir Welding of Aluminum Alloy Materials (International Journal of Mechanical & Mechatronics Engineering) vol 18 no 02 [223] Saif M Abbas, Kadhim K Resan, Ahmed K Muhammad and Muhannad Al-Waily 2018 Mechanical and Fatigue Behaviors of Prosthetic for Partial Foot Amputation with Various Composite Materials Types Effect (International Journal of Mechanical Engineering and Technology) vol 09 no 09 pp 383–394 [224] Jawad K Oleiwi and Ahmed Namah Hadi 2018 Experimental and Numerical Investigation of Lower Limb Prosthetic Foot Made from Composite Polymer Blends (International Journal of Mechanical and Production Engineering Research and Development) vol 08 no 02 pp 1319– 1330 [225] Muhsin J Jweeg, Zaid S Hammoudi and Bassam A Alwan 2018 Optimised Analysis, Design, and Fabrication of Trans-Tibial Prosthetic Sockets (IOP Conference Series: Materials Science and Engineering) vol 433 2nd International Conference on Engineering Sciences [226] Mohsin Abdullah Al-Shammari and Sahar Emad Abdullah 2018 Stiffness to Weight Ratio of Various Mechanical and Thermal Loaded Hyper Composite Plate Structures (IOP Conference Series: Materials Science and Engineering, 2nd International Conference on Engineering Sciences) vol 433 [227] Ahmed A Taher, Ayad M Takhakh and Sabah M Thaha 2018 Experimental Study and Prediction the Mechanical Properties of Nano-Joining Composite Polymers (Journal of Engineering and Applied Sciences) vol 13 no 18 pp 7665 7669 [228] Ekhlas Edan Kader, Akram Mahdi Abed and Mohsin Abdullah Al-Shammari 2020 Al2O3 Reinforcement Effect on Structural Properties of Epoxy Polysulfide Copolymer (Journal of Mechanical Engineering Research and Developments) vol 43 no 4 pp 320–328 [229] S E Sadiq, S H Bakhy and M J Jweeg 2020 Crashworthiness Behavior of Aircraft Sandwich Structure With Honeycomb Core Under Bending Load (IOP Conference Series: Materials Science and Engineering) [230] Hussein I Mansoor, Mohsin Abdullah Al-shammari and Amjad Al-Hamood 2020 Experimental Analysis of Cracked Turbine Rotor Shaft using Vibration Measurements (Journal of Mechanical Engineering Research and Development) vol 43 no 2 pp 294–304 [231] Esraa A Abbod, Muhannad Al-Waily, Ziadoon M R Al-Hadrayi, Kadhim K Resan and Saif M Abbas 2020 Numerical and Experimental Analysis to Predict Life of Removable Partial Denture (IOP Conference Series: Materials Science and Engineering, 1st International Conference on Engineering and Advanced Technology, Egypt) vol 870 [232] Ahmed A Taher, Ayad M Takhakh and Sabah M Thahab 2020 Study and Optimization of the Mechanical Properties of PVP/PVA Polymer Nanocomposite as A Low Temperature INTCSET 2020IOP Conf. Series: Materials Science and Engineering1094 (2021) 012081IOP Publishingdoi:10.1088/1757-899X/1094/1/01208128 Adhesive in Nano-Joining (3rd International Conference on Engineering Sciences, IOP Conference Series: Materials Science and Engineering) vol 671 [233] J L Mantari and J C Monge 2016 Buckling, Free Vibration and Bending Analysis of Functionally Graded Sandwich Plates Based on an Optimized Hyperbolic Unified Formulation (International Journal of Mechanical Sciences) vol 119 pp 170–186 [234] A M Zenkour and M Sobhy 2010 Thermal Buckling of Various Types of FGM Sandwich Plates (Composite Structures journal) vol 93 no 1 pp 93–102 [235] A M Zenkour and M Sobhy 2011 Thermal Buckling of Functionally Graded Plates Resting on Elastic Foundations Using the Trigonometric Theory (Journal of Thermal Stresses) vol 34 no 11 pp 1119–1138 [236] Kiani Y and Eslami M R 2013 An Exact Solution for Thermal Buckling of Annular FGM Plates on an Elastic Medium (Composites Part B: Engineering) vol 45 no 1 pp 101–110 [237] K M Liew and J Yang 2004 Thermal Post-Buckling of Laminated Plates Comprising Functionally Graded Materials with Temperature-Dependent Properties (Journal of Applied Mechanics) vol 71 no 6 pp 839–850 [238] Chi S H and Chung Y L 2006 Mechanical Behavior of Functionally Graded Material Plates under Transverse Load, Part II: Numerical Results (International Journal of Solids and Structures) vol 43 no 13 pp 3675–3691 [239] S Pradyumna and J N Bandyopadhyay 2007 Static and Free Vibration Analyses of Laminated Shells Using A Higher-Order Theory (Journal of Reinforced Plastics and Composites) vol 27 no 2 [240] H Nguyen-Xuan, L V Tran, T Nguyen-Thoi and H C Vu-Do 2011 Analysis of Functionally Graded Plates using an Edge-Based Smoothed Finite Element Method (Composite Structures) vol 93 no 11 pp 3019–3039 [241] Khalili S M R, A Jafari and S A Eftekhari 2010 A Mixed Ritz-DQ Method for Forced Vibration of Functionally Graded Beams Carrying Moving Loads (Composite Structures) vol 92 no 10 [242] A Jafari and S A Eftekhari 2011 A New Mixed Finite Element–Differential Quadrature Formulation for Forced Vibration of Beams Carrying Moving Loads (Journal of Applied Mechanics) vol 78 no 1 [243] Zhao X, Lee Y Y and Liew K M 2009 Free Vibration Analysis of Functionally Graded Plates using the Element-Free Kp-Ritz Method (Journal of Sound and Vibration) vol 319 no 3–5 [244] Zhu P and Liew K M 2011 Free Vibration Analysis of Moderately Thick Functionally Graded Plates by Local Kriging Meshless Method (Composite Structures) vol 93 no 11 pp 2925– 2944 [245] Singha M K, Prakash T and Ganapathi M 2011 Finite Element Analysis of Functionally Graded Plates under Transverse Load (Finite Elements in Analysis and Design) vol 47 pp 453–460 [246] Dozio L 2013 Natural Frequencies of Sandwich Plates with FGM Core Via Variable Kinematic 2-D Ritz Models (Composite Structures) vol 96 pp 561–568 [247] P Gagnon, C Gosselin and L Cloutier 1997 A Finite Strip Element for the Analysis of Variable Thickness Rectangular Thick Plates (Computers & Structures) vol 63 no 2 pp 349–362 [248] K M Liew, L X Peng and S Kitipornchai 2009 Vibration Analysis of Corrugated Reissner- Mindlin Plates using a Mesh-Free Galerkin Method (International Journal of Mechanical Sciences) vol 51 no 9–10 pp 642–652 [249] Ö Civalek 2007 Three-Dimensional Vibration, Buckling and Bending Analyses of Thick Rectangular Plates Based on Discrete Singular Convolution Method (International Journal of Mechanical Sciences) vol 49 pp 752–765 [250] Song Xiang, Ke-ming Wang,Yan-ting Ai, Yun-dong Sha and Hong Shi 2009 Analysis of Isotropic, Sandwich and Laminated Plates by a Meshless Method and Various Shear Deformation Theories (Composite Structures) vol 91 no 1 pp 31–37 [251] S Chakraverty and K K Pradhan 2014 Free Vibration of Functionally Graded Thin Rectangular Plates Resting on Winkler Elastic Foundation with General Boundary Conditions using Rayleigh–Ritz Method (International Journal of Applied Mechanics) vol 06 no 04 [252] T Kant and R K Khare 1997 A Higher–Order Facet Quadrilateral Composite Shell Element INTCSET 2020IOP Conf. Series: Materials Science and Engineering1094 (2021) 012081IOP Publishingdoi:10.1088/1757-899X/1094/1/01208129 (International Journal for Numerical Methods in Engineering) vol 40 no 24 pp 4477–4499 [253] Mahmud Rasheed Ismail, Zaman Abud Almalik Abud Ali and Muhannad Al-Waily 2018 Delamination Damage Effect on Buckling Behavior of Woven Reinforcement Composite Materials Plate (International Journal of Mechanical & Mechatronics Engineering) vol 18 no 05 pp 83–93 [254] A J M Ferreira, R C Batra, C M C Roque, L F Qian and R M N Jorge 2006 Natural Frequencies of Functionally Graded Plates by a Meshless Method (Composite Structures) vol 75 no 1–4 pp 593–600 [255] Shufrin I and Eisenberger M 2005 Stability and Vibration of Shear Deformable Plates-First Order And Higher Order Analysis (International Journal of Solids and Structures) vol 42 no 3–4 [256] Reddy J N 1997 Mechanics of Laminated Composite Plates: Theory and Analysis (CRC Press– Taylor & Francis Group) [257] S A Pradyumna and J N Bandyopadhyay 2008 Free Vibration Analysis of Functionally Graded Curved Panels using a Higher–Order Finite Element Formulation (Journal of Sound and Vibration) vol 318 pp 176–192 [258] Young Soo Kim and Suong Van Hoa 1995 Bi–Axial Buckling Behavior of Composite Rectangular Plates (Composite Structures) vol 31 no 4 pp 247–252 [259] N Sundararajan, T Prakash and M Ganapathi 2005 Nonlinear Free Flexural Vibrations of Functionally Graded Rectangular and Skew Plates under Thermal Environments (Finite Elements in Analysis and Design) vol 42 no 2 [260] P Malekzadeh and S A Shojaee 2013 Dynamic Response of Functionally Graded Plates under Moving Heat Source (Composites Part B: Engineering) vol 44 no 1 pp 295–303 [261] Li S, Hao W and Liu W K 2000 Numerical Simulations of Large Deformation of Thin Shell Structures using Meshfree Method (Computational Mechanics) vol 25 pp 102–116 [262] Parandvar H and Farid M 2015 Nonlinear Reduced Order Modeling of Functionally Graded Plates Subjected to Random Load in Thermal Environment (Composite Structures) vol 126 pp 174–183 [263] Vuong Nguyen Van Do 2018 A Modified Kirchhoff Plate Theory for Free Vibration Analysis of Functionally Graded Material Plates using Meshfree Method (IOP Conference Series: Earth and Environmental Science, 2nd International Conference on Sustainable Development in Civil, Urban and Transportation Engineering) vol 143 [264] Qian L F, Batra R C and Chen L M 2003 Elastostatic Deformations of a Thick Plate by using a Higher–Order Shear and normal Deformable Plate Theory and Two Meshless Local Petrov– Galerkin (MLPG) Methods (Tech Science Press) vol 4 no 1 pp 161–175 [265] L F Qian, R C Batra and L M Chen 2003 Free and Forced Vibrations of Thick Rectangular Plates using Higher–Order Shear and normal Deformable Plate Theory and Meshless Petrov–Galerkin (MLPG) Method (Tech Science Press) vol 4 no 5 pp 519–534 [266] P Zhu and K M Liew 2012 A Local Kriging Meshless Method for Free Vibration Analysis of Functionally Graded Circular Plates in Thermal Environments (Procedia Engineering) vol 31 [267] H M Wang and D S Luo 2016 Exact Analysis of Radial Vibration of Functionally Graded Piezoelectric Ring Transducers Resting on Elastic Foundation (Applied Mathematical Modelling) vol 40 no 4 pp 2549–2559 [268] M H Yas and B S Aragh 2011 Elasticity Solution for Free Vibration Analysis of Four– Parameter Functionally Graded Fiber Orientation Cylindrical Panels using Differential Quadrature Method (European Journal of Mechanics–A/Solids) vol 30 no 5 pp 631–638 [269] Pandey S and Pradyumna S 2015 Free Vibration of Functionally Graded Sandwich Plates in Thermal Environment Using a Layerwise Theory (European Journal of Mechanics–A/Solids) vol 51 [270] S Hosseini-Hashemi, H Salehipour, S R Atashipour, R Sburlati 2013 On the Exact In Plane and Out–of–Plane Free Vibration Analysis of Thick Functionally Graded Rectangular Plates: Explicit 3D Elasticity Solutions (Composites Part B: Engineering) vol 46 pp 108–115 INTCSET 2020IOP Conf. Series: Materials Science and Engineering1094 (2021) 012081IOP Publishingdoi:10.1088/1757-899X/1094/1/01208130[271] Ehab N Abbas, Muhannad Al-Waily, Tariq M Hammza and Muhsin J Jweeg 2020 An Investigation to the Effects of Impact Strength on Laminated Notched Composites used in Prosthetic Sockets Manufacturing (IOP Conference Series: Materials Science and Engineering, 2nd International Scientific Conference of Al-Ayen University) vol 928 [272] Sadiq Emad Sadiq, Muhsin J Jweeg and Sadeq H Bakhy 2020 The Effects of Honeycomb Parameters on Transient Response of an Aircraft Sandwich Panel Structure (IOP Conference Series: Materials Science and Engineering) vol 928 [273] Muhannad Al-Waily, Moneer H Tolephih and Muhsin J Jweeg 2020 Fatigue Characterization for Composite Materials used in Artificial Socket Prostheses with the Adding of Nanoparticles (IOP Conference Series: Materials Science and Engineering, 2nd International Scientific Conference of Al-Ayen University) vol 928 View publication stats | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/364283963LAS NUEVAS FRONTERAS DEL DERECHO DE FAMILIA (II) 26 y 27 de OctubreVALENCIAPoster · October 2022CITATIONS0READS451 author:Some of the authors of this publication are also working on these related projects:The gestational surrogacy agreements View projectsurrogacy View projectAlfio Guido GrassoUniversity of Catania27 PUBLICATIONS 4 CITATIONS SEE PROFILEAll content following this page was uploaded by Alfio Guido Grasso on 09 October 2022.The user has requested enhancement of the downloaded file.Financiado por el Proyecto de Investigación AICO/2021/090 “La modernización del derecho de familia a través de la práctica jurisprudencial”, de la Conselleria d’Innovació, Universitats, Ciència i Societat DigitalCONGRESO INTERNACIONALLAS NUEVAS FRONTERASDEL DERECHO DE FAMILIA (II)26 y 27de OctubreVALENCIAOrganizado por:Entidades colaboradoras:DIRECTOR:José Ramón de Verda y BeamonteSECRETARIO:Víctor Moreno SolerCOMITÉ CIENTÍFICO:Josefina Alventosa del Río, Gabriele Carapezza Figlia, Pilar Mª Estellés Peralta, Giampaolo Frezza, Salvatore Aceto di Capriglia, María José Reyes López, Pietro VirgadamoCOMITÉ DE ORGANIZACIÓN:Belén Andrés Segovia, Adrián Arrébola Blanco, Alvaro Bueno Biot, Almudena Carrión Vidal, Pedro Chaparro Matamoros, Michele Ciancimino, Francesco Disalvo, Giuseppe Garofalo, Francesco La Fata, Marco Li Pomi, Gonzalo Muñoz Rodrigo, Manuel Ortiz Fernández, Carlo Petta, Isabel Rabanete, Valerio Rotondo, Calogero Alberto Valenza“La proposta per la riforma della disciplina italiana in tema di riserva a favore del coniuge superstite”.Giovanni Perlingieri, Università degli Studi di Roma “La Sapienza” (Italia).“Animales domésticos y crisis familiares”.Guillermo Cerdeira Bravo de Mansilla, Universidad de Sevilla (España).“Protección del menor y medidas coercitivas indirectas de los deberes parentales”.Gabriele Carapezza Figlia, LUMSA (Palermo) (Italia).“La modificación de la cuantía de la pensión compensatoria”.Marta Ordás Alonso, Universidad de León (España).“Responsabilidad civil por los daños causados por las personas con discapacidad”.Isabel Zurita Martínez, Universidad de Cádiz (España).26 DE OCTUBRE MESA 1ª 9:00 - 10:1510:15 - 10:45Preside: María José Reyes López, Universidad de ValenciaMESA 2ª10:45 - 12:15Preside: José Ramón de Verda y Beamonte, Universidad de ValenciaJosé Ramón de Verda y Beamonte, Universidad de ValenciaPausa café12:30 - 16:00Pausa comida12:15 - 12:30Conclusiones“Una nuova stagione del diritto di famiglia”.Mirzia Bianca, Università degli Studi di Roma “La Sapienza” (Italia).“Realidad, voluntad e interés en el Derecho actual de la filiación”.Roncesvalles Barber Cárcamo, Universidad de La Rioja (España).“La impugnación del reconocimiento de complacencia en la reciente jurisprudencia del Tribunal Supremo”.Carlos Martínez de Aguirre, Universidad de Zaragoza (España).“La maternidad subrogada en la experiencia jurídica italiana”.Vincenzo Barba, Università degli Studi di Roma “La Sapienza” (Italia).“Acuerdos prematrimoniales con previsiones de ruptura”.Dolores Cervilla Garzón, Universidad de Cádiz (España).Congreso internacional: Las nuevas fronteras del derecho de familia (II)- 01 -PONENCIAS MAGISTRALESSALÓN DEL REY26 DE OCTUBRE 16:00 - 17:00Preside: Alfredo Ferrante, Università di Pavia“Nuevas tipologías familiares: su urgente reconocimiento en los sistemas jurídicos”.Eduardo Oliva Gómez, Universidad Autónoma del Estado de Morelos (México).“Las familias monoparentales ab initio y su lento caminar hacia la consecución de derechos”.Julia Ammerman Yebra, Universidad de Santiago de Compostela (España).“La figura del allegado”.Francisco Pérez del Amo, Alma Mater Studiorum Università di Bologna (Italia)/Universidad de León (España). “Casos especiales de adopción y expansión del vínculo parental: nuevas perspectivas jurídicas”.Ivan Libero Nocera, Università di Palermo (Italia).“La pluriparentalidad, un reto del derecho de familia moderno”.Carlos Luis Martínez Artiga, Universidad Modular Abierta” (El Salvador). “Derechos de visita, comunicación y compañía: la onda expansiva de la Ley 8/2021”.Covandonga López Suárez, Universidad de Cádiz (España).“Sobre la justa causa para impedir las relaciones personales entre abuelos y nietos en la jurisprudencia”.Álvaro Bueno Biot, Alma Mater Studiorum Università di Bologna (Italia) / Universidad de Valencia (España). “La temporalidad de en la atribución del uso de la vivienda familiar”.Federico Arnau Moya, Universitat Jaume I de Castellón (España).“L’assegnazione della casa familiare tra interesse dei figli, conflittualità dei genitori ed esigenze economiche della famiglia”. Carlotta Ippoliti, Alma Mater Studiorum Università di Bologna (Italia).“Atribuciones gratuitas de bienes en el marco de las separaciones matrimoniales”.José María Cardós Elena, Universidad de Valencia (España).“Cuestiones relativas a la aplicación de la Ley 8/2021, de medidas de apoyo en el ámbito familiar”.Javier Badenas Boldó, Alma Mater Studiorum Università di Bologna (Italia) / Universitat Jaume I de Castellón (España).NUEVOS MODELOS DE FAMILIA17:00 - 18:00Preside: Prof. Enrico Al Mureden,Alma Mater Studiorum Università di BolognaCRISIS FAMILIARES18:00 - 18:30Pausa caféCongreso internacional: Las nuevas fronteras del derecho de familia (II)- 02 -SALÓN DEL REY26 DE OCTUBRE 18:30 - 19:30Preside: Ángel Sánchez Hernández,Universidad de la Rioja“La edad núbil en el ordenamiento español tras la reforma de 2015”.Alejandro Nieto Cruz, Universidad de Cádiz (España). “Alimentos de progenitores a los hijos mayores versus alimentos de los hijos mayores a los progenitores”.Raquel Pérez Día, Universidad de Oviedo (España).“Desafección paterno-filial como causa de extinción de la obligación de alimentos”.Pablo Tortajada Chardí, Universidad de València (España).“La pensión de alimentos de los hijos mayores de edad: estudio comparativo entre España e Italia”.Jésica Delgado Sáez, Universidad de Salamanca (España).“La pensión compensatoria en las situaciones de crisis matrimoniales: naturaleza jurídica y extinción por el “cese de la causa que lo motivó”.Manuel Ortiz Fernández, Universidad Miguel Hernández de Elche (España)“Dispense e privilegi: prospettive interculturali del diritto matrimoniale canonico”.Federico Gravino, Università degli Studi di Firenze (Italia).“Novedades en el nuevo Reglamento en materia de competencia judicial internacional”.Ana Sofía Gomes, Universidade Lusíada, Lisboa, miembro del CEJEA, Centro de Estudios Jurídicos, Económicos y Ambientales (Portugal).“Análisis de los procedimientos de nulidad, separación o divorcio. Especial referencia en el Código civil catalán”.María Irene Sánchez Fernández, Ca’ Foscari Venezia (Italia) .“La residenza abituale: nozione, ruolo e significato”. Maria Paola Francesca Bottoni, Unicam/Università degli Studi di Camerino (Italia).“Finalità di protezione e carattere inderogabile della disciplina italiana dell’impresa familiare nel contesto europeo”. Gaetano Guzzardi, Università degli Studi di Catania (Italia).19:30 - 20:30Preside: Prof. Giovanni di Rosa,Università degli Studi di CataniaINTERDISCIPLINARCRISIS FAMILIARESCongreso internacional: Las nuevas fronteras del derecho de familia (II)SALÓN DEL REY- 03 -26 DE OCTUBRE SALÓN DE LA REINA16:00 - 17:00Preside: Enrico Al Mureden,Alma Mater Studiorum Università di Bologna“Situación actual del derecho de corrección en el Derecho de familia español”. Andrés Marín Salmerón, Universidad de Murcia (España).“Esercizio della responsabilità genitoriale e poteri di intervento del giudice”.Nicolo Cevolani, Alma Mater Studiorum Università di Bologna (Italia).“La responsabilidad parental en los entornos digitales. Necesaria conciliación entre bienestar y privacidad de los menores de edad”.Lisandra Suárez Fernández, Università degli Studi dell’Aquila (Italia).“Derecho a la privacidad del menor en el entorno de las redes sociales y ejercicio de la patria potestad: análisis desde la perspectiva del ordenamiento italiano y español.Lagana Manon Eleonora, Universidad de Valencia (España).“La vacunación infantil contra la COVID-19: conflictos y respuestas”.Víctor Moreno Soler, Universidad de Valencia (España). “Retos y desigualdades en tema de reproducción asistida en Europa”.Giulia Giovaninni, Universitat Pompeu Fabra (España)/Universitá di Bologna (Italia). “Reclamación de la filiación: madre biológica vs. expareja”.Borja del Campo Álvarez, Universidad de Oviedo (España). “La reclamación de filiación no matrimonial sin posesión de estado”.Mónica Gallego Ramos, Universidad de Málaga (España). “La fecondazione eterologa e le coppie omoaffettive nel sistema italiano: prospettive di tutela del nato”.Carlo Attanasio, Università di Sassari (Italia).“De la posesión de estado a la adopción. Un camino iniciado por la STS (Pleno) 31 marzo 2022”.Gonzalo Muñoz Rodrigo, Universidad de Valencia (España).“Parto anonimo e diritto a conoscere le proprie origini: gli orientamenti giurisprudenziali in seguito all’intervento della Corte costituzionale”.Jessica Mineo, LUMSA (Palermo) (Italia).17:00 - 18:00Preside: Javier Barceló Doménech, Cátedra de Derecho Notarial de la Universidad de Alicante.FILIACIÓNRELACIONES PATERNO FILIALES18:00 - 18:30Pausa caféCongreso internacional: Las nuevas fronteras del derecho de familia (II)- 04 -26 DE OCTUBRE 18:30 - 19:30Preside: Javier Badenas Carpio,Universitat Jaume I de Castellón“Acerca del consentimiento para el tratamiento de datos personales de menores de edad por parte de los prestadores de servicios de internet”.Javier Martínez Calvo, Universidad de Zaragoza (España). “Il trattamento dei dati digitali del minore”.Daniela Marcello, Università degli Studi di Firenze (Italia).“La publicación de fotografías del menor en redes sociales por parte de alguno de los progenitores”.Alberto Miguel Primo Llácer, Universidad de Valencia (España).“L'educazione del minore al consumo digitale, consapevole e sostenibile: quale ruolo per i genitori?”.Gabriele Perfetti, Università Ca' Foscari Venezia (Italia).“Funciones implícitas y explícitas del principio del resguardo del interés superior de los niños, niñas y adolescentes en el ordenamiento jurídico chileno, con una especial referencia a la relación de consumo”.Erika Isler Soto, Universidad Autónoma de Chile (Chile).“A vueltas con la Ley 17/2021, de 15 de diciembre. Los animales de compañía en el Derecho de familia y sucesiones”.Tania Vázquez Muiña, Real Centro Universitario María Cristina (España).“El heredero forzoso en el siglo XXI”.Adrián Arrébola Blanco, Universidad Complutense de Madrid (España).“La sospensione dalla successione nella legge del 2018 a tutela degli orfani dei crimini domestici”.Giorgia Anna Parini, Università degli Studi di Verona (Italia).“Il diritto di abitazione del coniuge superstite: il superamento dei confini della successione testamentaria”.Giuliano Mattace, Universidad de Salamanca (España)/Universidad de Perugia (Italia).“Il patto di non pubblicazione del testamento olografo nell’ordinamento italiano”.Giovanna Di Benedetto, Università degli Studi di Camerino (Italia).19:30 - 20:30Preside: Josefina Alventosa del Río,Universidad de ValenciaFAMILIA Y SUCESIONESMENORESCongreso internacional: Las nuevas fronteras del derecho de familia (II)- 05 -SALÓN DE LA REINA27 DE OCTUBRE 9:30 - 10:30Preside: Luigi Nonne,Università degli Studi di Sassari“Contratto di convivenza e sopravvenienze”.Riccardo Mazzariol, Università di Padova (Italia).“Assegno divorzile e convivenza more uxorio: coerenza sistematica e dubbi di costituzionalità”.Giuseppe Garofalo, Università di Salerno (Italia).“Illecito endofamiliare e misure di coercizione indiretta”.Francesco Disalvo, LUMSA (Palermo) (Italia). “Attribuzione del doppio cognome: giusto passo verso eguaglianza e pari dignità?”.Susanna Marinelli, Università Politecnica delle Marche (Italia).“Le violenze nella famiglia in metamorfosi: i riflessi sostanziali della Riforma Cartabia”.Giuseppe Foti, Università degli Studi di Messina (Italia). “Límites a la autonomía de la voluntad en los pactos prematrimoniales”.Manuel García Mayo, Universidad de Sevilla (España).“Per una rilettura della funzione assistenziale dell’assegno divorzile”.Giulia Travan, Ca’ Foscari Venezia (Italia).“Diritto di famiglia e autonomia negoziale, un ossimoro apparente: la questione della ammissibilità dei patti in vista del divorzio”. Clarissa Gola, Università degli Studi Roma Tre (Italia).“Gli Accordi prematrimoniali”.Simona Scuderi, Università degli Studi di Messina (Italia).“Cassazione S.U. 18641/2022 e riflessioni sull’assegnazione della casa familiare”.Martina Rodovero, Università degli Studi di Firenze (Italia).10:30 - 11:30Preside: Luigi Nonne,Università degli Studi di SassariCRISIS FAMILIARESNUEVOS MODELOS DE FAMILIACongreso internacional: Las nuevas fronteras del derecho de familia (II)- 06 -11:30 - 12:00Pausa caféSALÓN DEL REY27 DE OCTUBRE 12:00 - 13:00Preside: Dario Scarpa,Università degli Studi di Milano-Bicocca.“La crisi del rapporto matrimonioale tra separazionee divorzio: evoluzione del sistema e prospettive di riforma”.Antonino Astone, Università degli Studi di Messina (Italia).“Mediazione familiare e negoziazione assistita: l'interesse superiore del minore”.Fabricio Cesareo, University of Milan-Bicocca (Italia).“Assegno divorzile e convivenza di fatto del coniuge beneficiario”.Chiara Sartoris, Università degli Studi di Firenze (Italia).“Crisi della coppia e revoca del consenso alla procreazione medicalmente assistita”.Maria Paola Nico, Università di Camerino (Italia).“Trasferimenti immobiliari in sede di separazione e divorzio: tra autonomia negoziale e certezza del traffico giuridico”.Matteo Florena, LUMSA (Palermo) (Italia).“La comparecencia ante el notario de las personas con discapacidad tras la Ley 8/2021”.Esther Alba Ferré, Eserp Business & Law School (España).“Nuevos intervinientes en procesos de Familia y provisión de medidas de apoyo”.Francisco de Asís González Campo, Universidad de Zaragoza (España).“L’amministrazione di sostegno e l’interesse preferito della persona con disabilità”.Ettore William Di Mauro, Università degli Studi di Roma “La Sapienza” (Italia).“Il curatore speciale del minore nella legge delega di riforma del processo civile”.Laura Durello, Università degli Studi di Ferrara (Italia).“Inquadramento giuridico e tutela del caregiver familiare”.Emanuela Morotti, Università di Padova (Italia).13:00 - 14:00Preside: Dario Scarpa, Università degli Studi di Milano-Bicocca.MEDIDAS DE APOYO EN FAVOR DE PERSONAS CON DISCAPACIDADCRISIS FAMILIARESCongreso internacional: Las nuevas fronteras del derecho de familia (II)- 07 -SALÓN DEL REY27 DE OCTUBRE 9:30 - 10:30Preside: Ettore Battelli,Università degli Studi Roma Tre “La responsabilità genitoriale: profili evolutivi e limiti operativi della disciplina codicistica”.María Cimmino, Università degli Studi di Napoli Parthenope (Italia).“La nuova vita dell’adozione in casi particolari, tra completa equiparazione dello status filiationis e legami parentali”.Federico Ruggeri, Università degli Studi di Palermo (Italia) .“Soggettività delle persone minori di età e allontanamento forzato dei figli”. Veronica Rota Miarelli, Università degli Studi di Camerino (Italia).“Procreazione medicalmente assistita post mortem, status filiationis e diritti successori”. Carlo Petta, LUMSA (Palermo) (Italia).“La procreazione medicalmente assistita post mortem. Un confronto italo-spagnolo”.Rosina Elsa de Rosa, Università degli Studi di Napoli “Federico II” (Italia). “Il tortuoso itinerario della giurisprudenza sulla maternità surrogata: tra ordine pubblico internazionale e interesse del minore”.Calogero Alberto Valenza, Ca’ Foscari Venezia (Italia).“La gestazione per altri tra diritto alla procreazione, dignità umana e superiore interesse del minore. Una riflessione comparatistica”. Tania Bortolu, Università Federico II di Napoli (Italia).“Gli accordi sulla sorte di embrioni e gameti crioconservati in caso di crisi familiare”. Marco Rizzuti, Università degli Studi di Firenze (Italia).“Il cognome dei figli nella giurisprudenza italiana tra identità familiare ed eguaglianza dei coniugi”. Monica Pucci, Università degli Studi di Perugia (Italia).“Reflexiones sobre el parto anónimo en el ordenamiento jurídico italiano”.Gabriele Toscano, Universidad de Perugia (Italia)/Universidad de Salamanca (España).“Omessa diagnosi della malformazione del feto: tra obbligo di assunzione del consenso informato e lesione del diritto all'autodeterminazione terapeutica della gestante”.Nicola Di Napoli, Università LUMSA (Palermo) (Italia).10:30 - 11:30Preside: Ettore Battelli,Università degli Studi Roma TreFILIACIÓNRELACIONES PATERNO FILIALES11:30 - 12:00Pausa caféCongreso internacional: Las nuevas fronteras del derecho de familia (II)- 08 -SALÓN DE LA REINA27 DE OCTUBRE 12:00 - 13:00Preside: Roberto Senigaglia,Ca’ Foscari Venezia. “Doppio cognome ai figli tra identità personale e principio di eguaglianza”.Rosita Lifrieri, Università degli Studi di Roma “La Sapienza” (Italia).“Il danno non patrimoniale da “disinteresse” del genitore nei confronti del figlio (anche non riconosciuto)”.. Claudio Corso, Università degli Studi di Napoli “Federico II” (Italia).“Minori e relazioni affettive”. María Teresa Nurra, Università degli Studi di Sassari (Italia).“Diritti del minore e interruzione della relazione con il genitore sociale”. Alfio Guido Grasso, Universidad de Catania (Italia). “La maternità surrogata alla luce della responsabilità genitoriale: alla ricerca di un delicato equilibrio tra genitorialità di intenzione e diritto del minore ad una famiglia”.Serena Cancellieri, Università degli Studi di Camerino (Italia). “Le sorti dell’azienda coniugale e dell’impresa familiare nel caso di morte di uno dei coniugi”.Francesco Angeli, Università degli Studi di Perugia (Italia).“Le posizioni creditorie dei coniugi durante e dopo la comunione legale”.Sara Scola, Università di Verona (Italia).“Sulla natura della comunione de residuo. Spunti a margine di una recente pronuncia delle Sezioni unite”.Michele Ciancimino, Università LUMSA (Palermo) (Italia).“Il fondo patrimoniale tra tutela della famiglia e interessi dei creditori”.Cristiana Boiti, Università degli Studi di Perugia (Italia).“Patto di famiglia quale strumento di trasmissione dei beni aziendali: stato dell’arte e prospettive di un istituto mai del tutto valorizzato”.Marco Li Pomi, Università LUMSA (Palermo) (Italia).13:00- 14:00Preside: Alfredo Ferrante,Università di PaviaECONOMÍA DE LA FAMILIAMENORESCongreso internacional: Las nuevas fronteras del derecho de familia (II)- 09 -SALÓN DE LA REINACongreso internacional: Las nuevas fronteras del derecho de familia (II)- 10 -NUEVOS MODELOS DE FAMILIA “El sistema legitimario español. Una nueva configuración en orden a los nuevos modelos familiares”. Maria Elena Cobas Cobiella, Universidad de Valencia (España).“El carácter excepcional de la cláusula rebus sic stantibus y su innecesaria articulación para la revisión de las obligaciones intrafamiliares”. Leire Imaz Zubiaur, UPV/EHU (España). “I nuovi confini della filiazione adottiva”. Angela Mendola, Università degli Studi di Salerno (Italia).Movimiento de fronteras en el derecho de familia: los desafíos de la maternidad subrogada”. Diana Sthefania Muñoz-Gómez, Universidad de La Sabana (Colombia).CRISIS FAMILIARES“Le mobili frontiere del diritto familiare. L’assegno divorzile”. Susanna Cappuccino, Università degli Studi di Messina (Italia).“El derecho y el deber de alimentos en el concurso de acreedores”. Sonia Martínez Uceda, Universidad de Jaén (España). “Los derechos afectados por la intervención de detectives en los procedimientos de familia. Comentario de la STS 851/2021, de 9 de diciembre”, Juan José Nevado Montero, UNED (España).“Cuando se confunde una guarda de hecho conflictiva con el derecho de visitas a favor del progenitor”. Celia Prados García, Universidad de Córdoba (España).“La pensión compensatoria a la luz de la STS 810/2021, de 25 de noviembre”. Nieves Rojano Martín, Universidad de Málaga (España).MEDIDAS DE APOYO EN FAVOR DE PERSONAS CON DISCAPACIDAD“La guarda de hecho judicial”. Davit Benavent Cuquerella, Universidad Miguel Hernández de Elche (España).“Aspectos procesales del nuevo sistema español de provisión de medidas judiciales de apoyo: dudas y posibles soluciones prácticas”. Rebeca Castillo Santamaría, Universidad Isabel I Burgos (España)/Universidad del Atlántico Medio (España).“Conducta pródiga y Ley 8/2021 de 2 de junio”. Almudena Carrión Vidal, Universidad de Valencia (España).“Medidas de apoyo en el testamento de la persona con discapacidad intelectual”. Beatriz Hermida Bellot, Universidad CEU-Cardenal Herrera (España).RELACIONES PATERNO FILIALES“Familias reconstituidas y ejercicio de funciones parentales. Una mirada desde las prerrogativas de infancia y adolescencia”. Rommy Álvarez Escudero, Universidad de Valparaíso (Chile).“Menores en internet: problemas del ejercicio de la patria potestad sobre los nativos digitales”. Silvia Durán Alonso, UCAM Universidad Católica San Antonio de Murcia (España).“Il ruolo della derivazione biologica nella genitorialità: il difficile bilanciamento tra divieti generali e diritti dei singoli”, Barbara Salvatore Università degli Studi di Napoli “Federico II” (Italia).FILIACIÓN“I termini giuridici della funzione educativa nell’attuale quadro delle relazioni tra genitori e figli”. Giovanni Di Rosa, Università degli Studi di Catania (Italia).“La renuncia a la maternidad: ¿hacia el parto anónimo?”. Jon Atxutegi Gutiérrez, UPV/EHU (España).“La (de)construcción de las relaciones de filiación y paternidad a través de la jurisprudencia del Tribunal Europeo de Derechos Humanos”. Jonatán Cruz Ángeles, Universidad de Jaén (España).“Los límites legislativos al ejercicio de las acciones de filiación”. Laura Fernández Echegaray, Universidad de Cantabria (España).“La maternidad subrogada y su regulación: ¿existe un divorcio entre la realidad social y su regulación?” María Paz Herrera Rodríguez, Universidad Pegaso.MENORES“Dal diritto ad una famiglia al diritto alle relazioni affettive del minore”. Concetta Parrinello, Università di Messina (Italia).“El interés del menor antes y después del proceso de reagrupación familiar”. Joel Harry Clavijo Suntura, Universidad Isabel I de Castilla (España).PONENCIAS ESCRITAS Congreso internacional: Las nuevas fronteras del derecho de familia (II)- 11 -“Violencia familiar hacia niñas, niños y adolescentes en México: coexistencia de modelos y control de constitucionalidad”. Mónica González Contró, Universidad Nacional Autónoma de México.“El beneficio afectivo derivado de las salidas al extranjero de niños, niñas y adolescentes”. Alexis Mondaca Miranda, Universidad de Talca (Chile), y Constanza Astudillo, Universidad Santo Tomás (Chile).“L’identità di genere dei minori a quarantianni dalla legge sul mutamento di sesso”. Alessia Valongo, Università degli Studi di Perugi (Italia).“Autonomía decisoria de menores de edad. Especial referencia al ejercicio del derecho a autodeterminación de género en el ámbito de la salud”. Miguel Vieito Villar, Universidad de Santiago de Compostela (España).“Statuto dei diritti del minore e diritto allo sport”. Ángela Busacca, Università degli Studi Mediterranea di Reggio Calabria (Italia).ECONOMÍA DE LA FAMILIA“Distribuzione patrimoniale nel regime della comunione legale tra coniugi: profili assiologici e tutela della famiglia”. Dario Scarpa, Università degli Studi di Milano-Bicocca.“Las consecuencias jurídicas del trabajo del cónyuge en el régimen económico matrimonial: especial atención al empresario y al trabajo en el hogar”. Héctor Buenosvinos González, Universidad de Vigo (España).“La prohibición de disponer impuesta en testamento como mecanismo de protección del patrimonio familiar”. Jesús Palomares Bravo, Universidad de Málaga (España).INTERDISCIPLINAR“La mera indiferencia o ausencia de trato no se considera maltrato psicológico y no constituye causa autónoma de desheredación”. María Begoña Ribera Blanes, Universidad de Alicante (España).“Concreción del maltrato psicológico como causa de desheredación: la falta de relación familiar continuada”. Alejandro Araque García, Universidad de Málaga (España).“Convencionalidad respecto al derecho de conciliación laboral y familiar en América Latina”. Hazel Lohengrin Cabrera Ávalos, Universidad de San Carlos (Guatemala).“Violencia de género en la relación de enamorados en los estudiantes de la Universidad”. Marco Antonio Carmona Brenis, Universidad César Vallejo, Piura (Perú).“El delito de sustracción de menores: análisis de la STS, Sala Penal 401/2022, de 22 de abril de 2022”. Montserrat Vila Roura, Universidad de Gerona (España).CONGRESO INTERNACIONALLAS NUEVAS FRONTERASDEL DERECHO DE FAMILIA (II)26 y 27de OctubreVALENCIAView publication stats | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/362965548Consumers-Brands Battle: Investigating The Impact of Cancel Culture OnSwitching Behaviour and Brand Cancellation: The Moderating Role ofNegative eWOMConference Paper · August 2022CITATIONS0READS1,0771 author:Some of the authors of this publication are also working on these related projects:Antecedents and Consequences of Adolescents’ Materialism in Egypt View projectResearch Paper View projectJasmine MohsenUniversity of Leeds5 PUBLICATIONS 19 CITATIONS SEE PROFILEAll content following this page was uploaded by Jasmine Mohsen on 26 August 2022.The user has requested enhancement of the downloaded file. 1 2 Track: Marketing and Retail Title: “Consumers-Brands Battle: Investigating The Impact of Cancel Culture On Switching Behaviour and Brand Cancellation: The Moderating Role of Negative eWOM” Author: Jasmine Mohsen PhD Candidate at Leeds University Business School, University of Leeds, United Kingdom. Assistant Lecturer at Cairo University, Department of Business Administration, Cairo, Egypt Abstract Cancel culture has become one of the most controversial topics in the last few years. Online cancel culture movement can be defined as the practice of any form of withdrawing customers’ support for a particular brand, company, or even a person. Customers can withdraw their support by engaging in boycotting acts and unfollowing the brands on various networking websites to express their outrage and disapproval to the company’s wrongdoings. In today’s current world, cancel culture debate is considered a crucial phenomenon due to its power to engender public pressure to boycott a particular brand by inducing others to switch to another competitor and thus, promoting brand cancellation. This paper sheds the light on the dangers of online cancel culture movement on brand’s image. It also aims to provide guidance to businesses to mitigate the effects of brand cancellation and avoid falling prey to cancel culture movements. To achieve the research objectives, a mixed-methods approach will be conducted using a qualitative content analysis and an online survey. Keywords: Cancel culture, eWOM, Customer Disloyalty, Brand cancellation, Content analysis Literature Review Cancel culture The whole world is now witnessing a new movement of a unique social media activism, which is known by cancel culture (Lentini, 2022). In the past few years, the term cancel culture has gained much attention from the researchers due to its importance in changing the transactional dynamics between customers and brands (Bakhtiari, 2020). Cancel culture has become a controversial topic in marketing because of its ability to increase the brand cancellations rate. Brand cancellation can be detrimental to any brand because it can destroy the brand equity and ruin its reputation in a blink of an eye. The term of cancel culture was originally coined in 1980s. However, in 2014, the sparkle of this term has renewed again after the #MeTooMovement, which was against sexual harassment, and Black Lives Matter movement, which was against racism, discrimination, and inequality (Mendes et al., 2018). 3 Cancel culture refers to the consumers practices of boycotting a certain brand, a person or a celebrity because of their wrongdoings or false claims. Thus, consumers use cancel culture to express their rage and disapproval online as a way to protest against the brand’s actions or views on social media (Norris, 2021). This movement exerts a greater social pressure to other consumers because they spreading boycotting behaviour, bad comments, and negative electronic word of mouth (negative eWOM), which induce others to follow their leads and imitate their behaviour to express their hostility (Kimmel, 2004), revenge (Richins, 1983) and their dissatisfaction of what happened (Jansen et al., 2009). The main aim of cancelling a particular brand, a company or even a person is to combat corruption and hold the other party accountable for their actions to live in a better world. Cancel culture faced a lot of critics as some researchers viewed the cancellation practices as an online form of mob rule, because they believed that cancel culture prevent the creation of open debate between the conflicting parties which is considered the fundamental for realizing democracy (Norris, 2021). They refer to cancel culture as a form of public shaming. Cancel culture was also accused of inducing various media platforms (e.g., HBO) in implementation a form of cultural censorship over its content (Romano, 2021). In contrast, supporters believe cancel culture is crucial in todays’ world to empower people and give voice to vulnerable people so they can speak up and share their stories with the whole world (Norris, 2021). Thus, cancel culture is considered an important tool for realizing social justice, eliminating racism and ethical malpractice. Brand cancellation Brand cancellation can be caused by the diffusion of online cancel culture movement (Novelli, 2021). It can be defined as any action or practice in which consumers decide to withdraw their support for a particular brand, company, or even for a service as a form to protest against the company’s behavioural discrepancies, or malpractice acts (Bakhtiari, 2020). Withdrawing the support activities (i.e., unliking, or unfollowing the brand) on various networking websites) (e.g., Facebook, Instagram, etc) as a way to express their disapproval towards the company’s wrongdoings. In other words, when a brand is cancelled, consumers decided to cease supporting a certain brand, and instead search for a better substitute that will satisfy their needs (KrisRuby, 2020). 4 Thus, brand cancellation can be considered a form of boycotting acts in which consumers decided to express their disengagement, dissatisfaction, and disloyalty with the brands by engaging in negative dysfunctional behaviours (e.g., participating in anti-brand actions and trash talking) (Japutra et al., 2018; Japutra et al., 2014). Moreover, Bacile et al., (2014) pointed out consumers may start acting like “Badvocates” by spreading negative word of mouth about the brand to prevent others from consuming or buying it. In addition, based on brand-hate literature, prior studies (e.g., Jain and Sharma, 2019; Grégoire and Fisher, 2006; Lee et al., 2013) posited that highly emotionally attached consumers to a certain brand can turn into dissatisfied angry customers when they confront service failure compared to low emotionally attached consumers. As consumers tend to feel betrayed, when brands fail to meet their expectations. Hence, their love and attachment towards this brand can become hatred when they facing service failure (Mattila, 2004), or as a result of the company’s transgression or violations to its promises (Jain and Sharma, 2019). In other words, the stronger the customer-brand relationship, the more hostile the anti-branding actions and behaviours once this relationship ceased to exist (Johnson et al., 2011; Japutra et al., 2014). Additionally, Grégoire and Fisher, (2006) came up with “love-becomes-hate” concept to express the degree of customers’ retaliation and their tendencies to engage in betrayal acts when they feel betrayed by their favourite brands. The role of social media and changes in consumer-brand power dynamics Cancel culture movement is driven primarily by two factors; firstly, the proliferation of social media has contributed to increased consumers activism behaviour in recent years, as these platforms have made consumers more socially aware, and they assist consumers in shedding the light on any irresponsible brand (Ng, 2020) Social media empowers consumers and enable them to stand up for their demands and express their anger when the companies do not keep their promises to them (da Silva, 2021; Ng, 2020). Because of the social media, consumers’ voices are getting heard and enable the consumers to shift the balance of power in their relationship with their favourite brands, when the behaviour of the company do not match with its promises (Jain & Sharma, 2019). Furthermore, these social networking websites provide a rich platform for the consumers to share their opinions honestly without any barrier, which also encourage others to engage in online negative WOM to express their anger and revenge against a particular company (Azemi et al., 2020). Online WOM can be defined as expressions of 5 negative opinions about products, brands, companies, or even a celebrity (Weitzl and Hutzinger, 2017). Second driver is that, the changes in the customer-brand power dynamics as consumers have gained more control of managing their relationships with their favourite brands (Gómez-Suárez et al., 2017; Labrecque et al., 2013). In other words, consumers are no longer the recipients of brand offerings, but have now gained the upper hand in their relationships with brands. Thus, this shift in the balance of power in the relationships between customers and brands can induce consumers to rebel against their favourite brands when they confront any service failure. The consequences of cancel culture Brands should exert greater efforts to eliminate and reduce customer dissatisfaction (KrisRuby, 2020). Dissatisfied customers can destroy the company’s reputation and demolish the brand equity in a flash. Cancel culture is based on boycotting certain brand, celebrity, or a company by revealing the company’s malpractices and bad actions, which induce consumers to spread online negative eWOM, bad comments, posts and tweets (Bakhtiari, 2020). Hence, it considered a form of consumer activism (Lentini, 2022). The outrage and bad publicity that exist on various networking websites can trigger other social media users to form negative perceptions towards this brand as other consumers can participate in sharing these complaints and inducing others to stop consuming this brand (Rosenmayer et al., 2018; Obeidat et al., 2017). Moreover, de Campos et al., (2018), and Jalilvand and Heidari (2017) posit that as a result of the negative eWOM, other users will have lower tendency to purchase or repurchase anything from this company, since the trust and loyalty are breached. Hence, brand cancellation, customer disloyalty, customer brand disengagement, and switching behaviour can be considered the right actions for the consumers to protest against company’s unethical actions, behavioural discrepancies, or malpractice acts. Consequently, consumers will use brand cancellation as a tool for exerting a greater pressure on brands to keep their promises to the consumer and be transparent in how they operate. For instance; millions of subscribers are now considering cancelling their Netflix subscriptions and switching to other platforms due to the increase in the subscription fees, as well as the greater restrictions imposed on sharing your account with family and friends (Nambiar, 2022). 6 Hypotheses development and research framework From the above mentioned, it is reasonable to expect that cancel culture movement that happened on various networking websites can encourage others to feel dissatisfied, disengaged and disloyal towards the brand as a mechanism to advocate for victims or anyone who has been harmed by a brand (Norris, 2021). As cancel culture is about telling others what exactly has happened to raise their awareness, and the most important is to force the evil brands to reveal themselves, which helps in protecting others and achieving social justice. Moreover, numerous studies stressed on the dangers of the negative word of mouth (e.g., Duana et al., 2008; Rosenmayer, McQuiken, Robertson, & Ogden, 2018; de Campos Ribeiro, Butori, & Nagrad, 2018; Umashankar, Ward, & Dahl, 2017) as they highlighted that negative electronic WOM can encourage others to participate in sharing the bad news and inducing others to stop buying this brand. This can be attributed to the power of negative eWOM, as it can reach to much wider audience across the whole world in just few seconds, which can be dangerous for any business firms and constitute a threat on a brand image and reputation (Zúñiga et al., 2014). In similar vein, Park & Lee (2009) argued that negative eWOM can have a greater effect compared to positive eWOM, as one study found that each 1,000 negative eWOM on internet can cause an accumulated loss up to $8.1 billion over 20 months. Hence, people always intervene and support the angry dissatisfied customers by decreasing their purchase intentions towards this brand and sharing the dissatisfied customers’ post or comments across various networking websites. Therefore, based on the previous literature, it is reasonable to expect that the movement of cancel culture and public backlash that happened online, can easily make other customers disloyal, and dissatisfied towards this brand. These feelings can be considered drivers for many consumers to increase their tendencies towards participating in anti-brand behaviours (e.g., trash talking, and schadenfreude) (Jain & Sharma, 2019; Japutra et al., 2018; Japutra et al., 2014). Thus, the aim of this study is to extend the previous literature by assuming that consumers may also participate in online cancel culture movement and consequently, promote brand cancellation to protect others from falling a prey towards the brand’s malpractices. 7 Research hypotheses H1 Online cancel culture movement positively affects customers disengagement H2 Online cancel culture movement positively affects customers disloyalty H3 Online cancel culture movement positively affects customers dissatisfaction H4 The relationship between online cancel culture movement and customers disengagement, customers disloyalty, and customers dissatisfaction are moderated by the spread of negative eWOM on various networking websites H5 Customers’ disengagement is positively associated with increased switching behaviour H6 Customers’ disloyalty is positively associated with increased switching behaviour H7 Customers’ dissatisfaction is positively associated with increased switching behaviour H8 Switching behaviour is positively associated with brand cancellation The research model shows the impact of the online cancel culture movement on other consumers’ intentions and behaviours towards the brand. This model portrays the dangers of cancel culture towards any brand. Figure 1: Conceptual Framework Methodology Stage one: a qualitative content analysis The first stage involves conducting a qualitative content analysis across various digital channels (e.g., Facebook, Twitter, and Instagram) to explore to what extent calling out others to boycott or cancel a certain brand can affect their actions and behaviours towards this brand. 8 In other words, to see whether the negative eWOM will impact people’s intentions to acquire this brand or not. By employing the content analysis, the researcher will be able to track all negative comments, posts, and shares on Twitter and Facebook against some particular brands (e.g., Heineken, Oatly, Netflix, etc). Then, the collected data will be analysed using NVivo. It considered a crucial tool, which help the researchers focusing on particular words such as boycotting, cancelled, backlash, etc. NVivo is also considered an excellent tool for organizing and categorizing the dataset and in determining the relevant theme for each group of words (Zakrzewski, 2008). NVivo assists the researchers in addressing particular research questions (Aureli, 2017). Thus, it helps in enhancing the reliability and the accuracy of the obtained data, which assists in achieving accurate results (Bae and Chi, 2022). Stage two: Online Survey The second phase includes conducting a quantitative online survey. This study will depend on Qualtrics to produce a self-administrated online survey. The questionnaire will be designed based on the existed and validated scales. Participants will be recruited through the various social networking websites to enable the researcher to statistically test the developed hypotheses as well as the proposed conceptual framework. Practical significance of the research Over the last few years, brand cancellation has increased rapidly, largely due to the changes in customer-brand power dynamics, as consumers have gained more control of managing their relationships with their favourite brands (Gómez-Suárez et al., 2017; Schultz and Bailey, 2000). Therefore, companies need to change and respond faster to avoid falling a victim towards brand cancellation movements. Thanks to social media, consumers nowadays have the power and the necessary information, which enable them to become picky and to set higher expectations from their brands. Bakhtiari, (2020) reported that a study conducted by the Edelman company revealed that 64% of the consumers are willing to buy or cancel a particular brand based on its views and opinions regarding a social or political issue. Since no brand has immunity towards brand cancellation, understanding the new demands of the customers and their updated expectations are a must for any company to achieve sustainable growth and maintain its reputation in the market (Novelli, 2021; Lentini, 2022). Hence, the purpose of this study is to emphasise the importance of customer satisfaction as well as provide 9 guidance for the businesses to mitigate any controversial issues that may arise online to prevent the brand cancellation, which in turn, help companies grow profitably by keeping their customers satisfied and delighted. Current stage At the current stage, an analysis of the literature review has been conducted. Furthermore, a preliminary analysis of publicly available content analysis on Twitter and Facebook has been carried out and analysed. 10 References Aureli, S., 2017. A comparison of content analysis usage and text mining in CSR corporate disclosure. International Journal of Digital Accounting Research, 17. Azemi, Y., Ozuem, W. and Howell, K.E., 2020. The effects of online negative word‐of‐mouth on dissatisfied customers: A frustration–aggression perspective. Psychology & Marketing, 37(4), pp.564-577. Bacile, T.J., Allen, A.M. and Hofacker, C.F., 2014. “Enter the badvocate: a unique consumer role emerging within social media complaint and recovery episodes”, in Paper Presented at the 2014 Marketing EDGE Direct/Interactive Research Summit, San DiegoCA (October 25-26). Bae, W. and Chi, J., 2022. Content Analysis of Passengers’ Perceptions of Airport Service Quality: The Case of Honolulu International Airport. Journal of Risk and Financial Management, 15(1), p.5. Bakhtiari, K., 2020. Why Brands Need To Pay Attention To Cancel Culture. [Online]. Available at https://www.forbes.com/sites/kianbakhtiari/2020/09/29/why-brands-need-to-pay-attention-to-cancel-culture/?sh=54276711645e. [Accessed 18 February, 2022]. da Silva, J.A.T., 2021. How to shape academic freedom in the digital age? Are the retractions of opinionated papers a prelude to “cancel culture” in academia?. Current Research in Behavioral Sciences, 2, p.100035. de Campos Ribeiro, G., Butori, R. and Le Nagard, E., 2018. The determinants of approval of online consumer revenge. Journal of Business Research, 88, pp.212-221. De Zúñiga, H.G., Copeland, L. and Bimber, B., 2014. Political consumerism: Civic engagement and the social media connection. New media & society, 16(3), pp.488-506. Gómez-Suárez, M., Martínez-Ruiz, M.P. and Martínez-Caraballo, N., 2017. Consumer-brand relationships under the marketing 3.0 paradigm: a literature review. Frontiers in psychology, 8, p.252. Vancouver Grégoire, Y., & Fisher, R. J., 2006. The effects of relationship quality on customer retaliation. Marketing Letters, 17(1), 31-46. Jain, K. and Sharma, I., 2019. Negative outcomes of positive brand relationships. Journal of Consumer Marketing. Jalilvand, M.R. and Heidari, A., 2017. Comparing face-to-face and electronic word-of-mouth in destination image formation: The case of Iran. Information Technology & People. Jansen, B.J., Zhang, M., Sobel, K. and Chowdury, A., 2009. Twitter power: Tweets as electronic word of mouth. Journal of the American society for information science and technology, 60(11), pp.2169-2188. 11 Japutra, A., Ekinci, Y., & Simkin, L., 2018. Positive and negative behaviours resulting from brand attachment: The moderating effects of attachment styles. European Journal of Marketing.https://www.independent.co.uk/life-style/food-and-drink/oatly-boycott-cancelled-twitter-blackstone-accusations-deforestation-trump-a9700656.html Japutra, A., Ekinci, Y., Simkin, L. and Nguyen, B., 2014. The dark side of brand attachment: a conceptual framework of brand attachment’s detrimental outcomes. The Marketing Review, Vol. 14 No. 3, pp. 245-264, available at: https://doi.org/10.1362/146934714X14024779061875 Johnson, A.R., Matear, M. and Thomson, M., 2011. “A coal in the heart: self-relevance as a post-exit predictor of consumer anti-brand actions”, Journal of Consumer Research, Vol. 38No. 1, pp. 108-125. Kimmel, A.J., 2013. Rumors and rumor control: A manager's guide to understanding and combatting rumors. Routledge. KrisRuby., 2020. Cancel Culture-Brand Management. {Online}. Available at https://rubymediagroup.com/cancel-culture-brand-management/ {Accessed on 18 April, 2022} Labrecque, L.I., Vor Dem Esche, J., Mathwick, C., Novak, T.P. and Hofacker, C.F., 2013. Consumer power: Evolution in the digital age. Journal of Interactive Marketing, 27(4), pp.257-269. Lee, J. S., Pan, S., & Tsai, H., 2013. Examining perceived betrayal, desire for revenge and avoidance, and the moderating effect of relational benefits. International Journal of Hospitality Management, 32, 80-90. Lentini, N., 2022. Cancel culture: Trouble for brands or just noise?. {Online}. Available at https://www.marketingdive.com/news/cancel-culture-trouble-for-brands-or-just-noise/616983/ {Accessed on 12 March., 2022} Mattila, A. S., 2004. The impact of service failures on customer loyalty: The moderating role of affective commitment. International journal of service industry management. Mendes, K., Ringrose, J. and Keller, J., 2018. # MeToo and the promise and pitfalls of challenging rape culture through digital feminist activism. European Journal of Women's Studies, 25(2), pp.236-246. Nambiar, P., 2022. Why Is Everyone Cancelling Netflix and Switching To Other Platforms?. https://www.hitc.com/en-gb/2022/04/21/why-is-everyone-cancelling-netflix/{Online} [Accessed on 16 June, 2022] Ng, E., 2020. No grand pronouncements here...: Reflections on cancel culture and digital media participation. Television & New Media, 21(6), pp.621-627. Norris, P., 2021. Cancel culture: Myth or reality?. Political Studies, p.00323217211037023. 12 Novelli, P., 2021. No Brand Is Immune to Cancel Culture, But Purpose Helps. {Online}. Available at https://sustainablebrands.com/read/marketing-and-comms/study-no-brand-is-immune-to-cancel-culture-but-purpose-helps {Accessed on 2 Feb., 2022} Obeidat, Z.M.I., Xiao, S.H., Iyer, G.R. and Nicholson, M., 2017. Consumer revenge using the internet and social media: An examination of the role of service failure types and cognitive appraisal processes. Psychology & Marketing, 34(4), pp.496-515. Richins, M.L., 1983. Negative word-of-mouth by dissatisfied consumers: A pilot study. Journal of marketing, 47(1), pp.68-78. Romano, A., 2021. The second wave of “cancel culture {Online}. Available at https://www.vox.com/22384308/cancel-culture-free-speech-accountability-debate {Accessed on 20 May, 2022} Rosenmayer, A., McQuilken, L., Robertson, N. and Ogden, S., 2018. Omni-channel service failures and recoveries: refined typologies using Facebook complaints. Journal of Services Marketing. Schultz, D.E. and Bailey, S.E., 2000. Customer/brand loyalty in an interactive marketplace. Journal of Advertising Research, 40(3), pp.41-52. Umashankar, N., Ward, M.K. and Dahl, D.W., 2017. The benefit of becoming friends: Complaining after service failures leads customers with strong ties to increase loyalty. Journal of Marketing, 81(6), pp.79-98. Weitzl, W. and Hutzinger, C., 2017. The effects of marketer-and advocate-initiated online service recovery responses on silent bystanders. Journal of Business Research, 80, pp.164-175. Zakrzewski, D., 2008. Assessing privatized airport performance from stakeholder viewpoints: a study of Sydney Airport. In Performance Measurement and Management Control: Measuring and Rewarding Performance. Emerald Group Publishing Limited. View publication stats | RESEARCH GATE |
Regio- and diastereoselective reactions of chiralsecondary alkylcopper reagents with propargylicphosphates: preparation of chiral allenes†Juri Skotnitzki, Alexander Kremsmair, Daniel Keefer,‡ Franziska Sch¨uppel,Brieuc Le Cacher de Bonneville, Regina de Vivie-Riedleand Paul Knochel*The diastereoselective SN20-substitution of secondary alkylcopper reagents with propargylic phosphatesenables the preparation of stereodefined alkylallenes. By using enantiomerically enriched alkylcopperreagents and enantioenriched propargylic phosphates as electrophiles anti-SN20-substitutions wereperformend leading to a-chiral allenes in good yields with excellent regioselectivity and retention ofconfiguration. DFT-calculations were performed to rationalize the structure of these alkylcopperreagents in various solvents, emphasizing their configurational stability in THF.IntroductionAllenes are common intermediates in organic synthesis andfound in natural products.1 They are typically prepared by thesubstitution reaction of propargylic electrophiles with nucleo-philes, such as organocopper reagents.2 Thereby, these prop-argylic reagents bear a good leaving group, such as acetates,ethers, epoxides, phosphates or halides.2–4 Axially chiral allenesaregenerallypreparedfromenantioenrichedpropargylicsubstrates3 or by the use of chiral ligands.4 The chirality transferfrom the chiral propargylic substrate to the allene depends onthe nature of the electrophile and nucleophile as well as on thesolvent and temperature.1e However, the enantioselective prep-aration of axially chiral allenes bearing a stereocenter in a-position (“a-chiral allenes”) is rather difficult and only a fewexamples have been reported.5 Thereby, the stereochemistry ofthe a-position results from an asymmetric synthesis usingchiral ligands.Recently, we reported a zinc-mediated anti-SN20-substitutionreactionofalkylcopperreagentsoftype1withallylicsubstrates (2) leading to chiral alkenes of type 3 with excellentregioselectivityandhighretentionofconguration(seeScheme 1(b and c)).6,7 These organocopper reagents werepreparedfromthecorrespondingalkyliodide4viaI/Li-exchange reaction leading to alkyllithium reagent 5.Subsequent transmetalation with CuBr$P(OEt)3 afforded alkyl-copper reagent 1.8 The regio-selectivity (SN20 : SN2 ratio) of thesubstitution reactions highly depended on the choice of allylicelectrophile 2 and the used organometallic species. The reac-tion of alkylcopper reagents 1 with allylic bromides 2a exclu-sively led to the SN2-product 3a (g : a < 1 : 99; see Scheme 1(a)).The addition of zinc chloride and the use of chiral allylicphosphates 2b as electrophiles exclusively led to the SN20-products 3b (g : a > 99 : 1; (b)).6 Furthermore, we reported anti-SN20-substitutions of secondary alkylcopper-zinc reagents withallylic epoxides 2c leading to chiral allylic alcohols of type 3c(g : a > 95 : 5; (c)).7 This method was used in the total synthesisof the natural product (3S,6R,7S)-zingiberenol.7Scheme 1Stereoretentive preparation of chiral secondary alkylcop-per reagents 1: (a–c): subsequent SN2- and zinc-mediated anti-SN20-substitution reactions with allylic substrates. (d): Anti-SN20-substitutionwith chiral propargylic phosphates leading to axially chiral allenes.Department of Chemistry, Ludwig-Maximilians-Universitaet, Butenandtstrasse 5-13,81377 M¨unchen, Germany. E-mail: paul.knochel@cup.uni-muenchen.de† Electronicsupplementaryinformation(ESI)available.SeeDOI:10.1039/c9sc05982b‡ Present address: Department of Chemistry, University of California Irvine,California 92697, United States.Cite this: Chem. Sci., 2020, 11, 5328All publication charges for this articlehave been paid for by the Royal Societyof ChemistryReceived 26th November 2019Accepted 7th April 2020DOI: 10.1039/c9sc05982brsc.li/chemical-science5328 | Chem. Sci., 2020, 11, 5328–5332This journal is © The Royal Society of Chemistry 2020ChemicalScienceEDGE ARTICLEHerein, we wish to report the anti-SN20-substitution ofsecondary alkylcopper reagents 1 with chiral propargylic phos-phates 6 leading to a-chiral allenes of type 7 with retention ofthe conguration (see Scheme 1(d)). Remarkably, this overallanti-SN20-substitution reaction proceeded directly with thealkylcopper reagent 1 with transfer of chirality from the prop-argylic substrate 6 to the allene 7.Results and discussionIn preliminary experiments, we examined the leaving groupof the propargylic electrophile for achieving the desired SN20-reaction. Thus, we prepared the secondary alkyllithiumreagent anti-5a via I/Li-exchange of the corresponding alkyliodide anti-4a at �100 �C in pentane/diethyl ether-mixture(3 : 2) using t-BuLi (2.2 equiv.) followed by subsequent treat-ment with CuBr$P(OEt)3 (2.0 equiv.) leading to alkylcopperreagent anti-1a (see Table 1). This alkylcopper reagent wascongurationally stable in THF up to �50 �C and thus, weperformed a solvent switch at this temperature.6 Subsequentaddition of the propargylic bromide9a (6a, 3.0 equiv.) fur-nished only traces of the desired allene anti-7a (see Table 1;entry 1) aer stirring for 1 h at �50 �C. The use of propargylicacetate (6b)9b showed a similar result (entry 2). Switching topentauorobenzoate (6c)9c or diphenylphosphate (6d)9d asleaving groups afforded anti-7a in good yields, but withmoderate stereoretention (48–50% yield, dr up to 93 : 7;entries 3 and 4). However, using the propargylic diethylphosphate 6e9e as electrophile signicantly increased thestereoretention of the secondary alkylcopper center (anti-7a,59% yield, dr ¼ 98 : 2). The same reaction afforded anti-7a inonly 40% yield and dr ¼ 92 : 8 when no solvent switch wasperformed, demonstrating the necessity of THF as solvent.With these results in hand, we performed stereoselective reac-tions with various diastereomerically pure alkyl iodides syn- or anti-4a–d and propargylic phosphates 6e–g leading to allenes 7a–e in42–65% yield and with dr higher than 95 : 5 (see Table 2).10,11 Inmost cases, a high retention of conguration was observed.However, using the TMS-substituted propargylic phosphate 6g aselectrophile led to allene anti-7c in 61% yield with moderate dia-stereoselectivity (dr ¼ 75 : 25; entry 4). The reaction of anti-1a withthe propargylic phosphate bearing a terminal methyl-group 6f ledto the methyl-substituted allene anti-7b in 65% yield anddr ¼ 97 : 3 (see Table 2; entry 3). Furthermore, the 1,2-substitutedsecondary alkylcopper reagents anti- and syn-1b reacted with 6e tothe corresponding allenes anti-7d (58% yield, dr ¼ 98 : 2; entry 5)and syn-7d (42% yield, dr ¼ 6 : 94; entry 6). The OTBS-substitutedallenes anti-7e (50% yield, dr ¼ 95 : 5; entry 7) and syn-7e (44%yield, dr ¼ 4 : 96; entry 8) were prepared with high retention ofconguration as well.Table 1Stereoretentive preparation of secondary alkylcopper reagentanti-1a and subsequent reaction with various propargylic substrates 6leading to the allene anti-7aEntryElectrophileYield of anti-7aa (%)dr of anti-7aa16a: X ] BrTraces—26b: X ] OAc590 : 1036c: X ] OCOC6F54891 : 946d: X ] OP(O)(OPh)25093 : 756e: X ] OP(O)(OEt)25998 : 2a The diastereoselectivity (dr; anti : syn ratio) was determined by GC-analysis using dodecane as internal standard.Table2Stereoselectivepreparationofdiastereomericallypureallenes 7a–e starting from alkyl iodides 4a–cEntryAlkylcopperElectrophile 6Product of type 7a,b126e3456e66e76e86ea The diastereoselectivity (dr; anti : syn ratio) was determined by 1H- or13C-NMR analysis. b The SN20 to SN2 ratio was higher than 99 : 1. c Theyield was determined by GC-analysis using dodecane as internalstandard.This journal is © The Royal Society of Chemistry 2020Chem. Sci., 2020, 11, 5328–5332 | 5329Edge ArticleChemical ScienceIn addition, this anti-selective substitution was extended tooptically enriched alkylcopper reagents 1d–e (see Table 3).Thus, the reaction of the secondary alkylcopper reagent (R)-1dwith propargylic phosphate 6e furnished (R)-7f in 41% yield ander ¼ 93 : 7 (see Table 3; entry 1). Analogously, the corresponding(S)-enantiomer (S)-7f was prepared in 48% yield and er ¼ 10 : 90(entry 2). To our delight, chiral alkylcopper reagents reacted alsowith higher substituted chiral propargylic phosphates 6h–ileading to axially chiral allenes bearing a stereocenter in thea-position (see Table 3; entries 3–8). Thus, the reaction of thealkylcopper (R)-1d with enantioenriched propargylic phosphate(R)-6h, prepared from the corresponding 3-butyn-2-ol,12 led tothe a-chiral disubstituted allene (R,S)-7g13 in 43% yield withhigh anti-SN20-substitution ratio (dr ¼ 92 : 8; er ¼ 99 : 1, entry3). Similarly, the allene (S,S)-7g was prepared from organo-copper (S)-1d and the chiral phosphate (R)-6h in 49% yield(dr ¼ 12 : 88; er ¼ 99 : 1;14 entry 4). Moreover, (R)-oct-3-yn-2-yldiethyl-phosphate (R)-6i was prepared according to literaturefrom the corresponding optically enriched propargylic alco-hol.3e,6,14 Subsequent reaction of alkylcopper (R)-1d with phos-phate (R)-6i furnished the a-chiral trisubstituted allene (R,S)-7hin 59% yield (dr ¼ 91 : 9, er ¼ 99 : 1; entry 5). It was also possibleto convert the methoxy-substituted secondary alkyl iodide (R)-and (S)-4e to the corresponding alkylcopper reagents (R)- and(S)-1e and aer reaction with (R)-6h the a-chiral disubstitutedallenes (R,S)-7i (52% yield, dr ¼ 93 : 7, er ¼ 99 : 1; entry 6) and(S,S)-7i (54% yield, dr ¼ 12 : 88, er ¼ 99 : 1; entry 7) were ob-tained. Furthermore, the reaction of (R)-1e with (R)-6i led to theTable 3Stereoretentive preparation of chiral allenes 7f–j via anti-SN20-substitution reaction of chiral alkylcopper reagents 1d–e with propargylicphosphates 6e, (R)-6h and (R)-6iEntryAlkylcopper of type 1Propargylic phosphate 6Product of type 7a,b,c126e34(R)-6h (er ¼ 99 : 1)567(R)-6h (er ¼ 99 : 1)8(R)-6i (er ¼ 99 : 1)a The diastereoselectivity (dr; anti : syn ratio) was determined by 1H- or 13C-NMR analysis. b The SN20 to SN2 ratio was higher than 99 : 1. c Theenantiomeric ratio (er) was determined by chiral GC-analysis.5330 | Chem. Sci., 2020, 11, 5328–5332This journal is © The Royal Society of Chemistry 2020Chemical ScienceEdge Articletrisubstituted allene (R,S)-7j in 51% yield and good diaster-eoselectivity (dr ¼ 92 : 8, er ¼ 99 : 1; entry 8). Unfortunately, thepreparation of tertiary propargylic phosphates was unsuccessfulalthough the subsequent preparation of axially chiral tetra-substituted allenes would be of high interest for organicsynthesis.To get a better understanding of the regioselectivity, we haveprepared the racemic phosphate 6j, which contains a prop-argylic moiety (see Scheme 2).15 The nucleophilic organocopperreagent rac-1d can undergo a substitution either in the a-posi-tion (SN2-substitution of the phosphate), the g-position (SN20-attack on the propargylic site) or g0-position (SN20-attack on theallylic site). Interestingly, the reaction of 1d with 6j afforded theallene 7k, the SN2-product 7l and the alkene 7m in 58% yield16with a ratio of 2.6 : 1.0 : 6.4 ¼ g : a : g0. This selectivity could beexplained by steric hindrance of the a-position and favoureddirect SN20-substitution of the allylic phosphate (g0-position)compared to the propargylic moiety (g-position).Computational calculationsFurthermore, DFT-calculations17 were performed to rationalizethe high congurational stability of these chiral secondaryalkylcopper reagents. Solvation effects were accounted for bythe Polarizable Continuum Model (PCM).18 First, we deter-mined the structure of secondary alkylcopper reagent anti-1a insolution. Thus, we calculated the free energies of anti-1a withcoordination to all possible ligands, namely triethyl phosphite(P(OEt)3; anti-8), tetrahydrofuran (THF; anti-9) and diethyl ether(Et2O; anti-10; see Scheme 3, (1–2)).19 Comparison of the freeenergies of anti-8 with the free energies of anti-9 showed thatthecoordinationtoP(OEt)3isthermodynamicallymorestable (DG ¼ +4.6 kcal mol�1; see Scheme 3, (1)). Similar resultswere obtained for the substitution of P(OEt)3 with Et2O(DG ¼ +6.8 kcal mol�1, (2)) showing again the high affinity ofphosphor to copper. These calculations emphasized that anti-8is the thermodynamically most stable structure. The directcomparison of anti-9 and anti-10 shows that the THF coordi-nated structure 9 is 3.9 kcal mol�1 more stable compared to theEt2O coordinated structure 10. In addition, the bond energiesand bond lengths of the carbon–copper bond for anti-8(53.9 kcal mol�1, 198.5 pm), anti-9 (51.3 kcal mol�1, 195.9 pm)and anti-10 (50.6 kcal mol�1, 195.8 pm) were determinedshowing that the carbon–copper bond is most stable when thecopper is coordinated to P(OEt)3. Comparison of the free ener-gies of anti-8 and syn-8 showed that the anti-isomer is thermo-dynamically more stable (DG ¼ +2.9 kcal mol�1; see Scheme 3).This result is in agreement with previous reported ndings.20Next, we investigated the epimerization of anti-8 to the cor-responding syn-isomer syn-8 via cleavage of the carbon–copperbond or a planar transition state ts-8 (see Scheme 3). The highcarbon–copper bond energy of 54.0 kcal mol�1 as well as thetransition state energy of 51.9 kcal mol�1 corroborate the highstability of anti-8 towards epimerization at �50 �C.21 However,the slight epimerization of the secondary alkylcopper reagents(1) may be due to polymolecular exchange reactions betweenthese copper reagents.22ConclusionsIn conclusion, we have reported the enantioselective prepara-tion of axially chiral allenes bearing a stereocontrolled a-chiralcenter via anti-SN20-substitution reaction of chiral secondaryalkylcopper reagents with enantioenriched propargylic phos-phates with retention of conguration. DFT-calculations wereperformed to determine the structure of these alkylcopperreagents and rationalize the high congurational stability inTHF. Further extensions are currently under investigation inour laboratories.Conflicts of interestThere are no conicts to declare.Scheme 2Regioselective addition of secondary alkylcopper reagent1d to allylic and propargylic moiety containing phosphate 6f.Scheme 3Theoretical calculations for the structure determination ofanti-1a and the epimerization of secondary alkylcopper reagent anti-8to syn-8.This journal is © The Royal Society of Chemistry 2020Chem. Sci., 2020, 11, 5328–5332 | 5331Edge ArticleChemical ScienceAcknowledgementsWe thank the Excellence Cluster “e-conversion” and theMunich-Centre for Advanced Photonics (MAP) for nancialsupport. We also thank Albemarle for the generous gi ofchemicals. J. S. thanks the FCI Foundation for a fellowship. D.K. and J. S. acknowledge nancial support by the Dr KlausRoemer Foundation through their PhD thesis award.Notes and references1 (a) Modern Allene Chemistry, ed. N. Krause and A. S. K.Hashmi, Wiley-VCH, Weinheim, 2004, vol. 1 and 2. Forreviews see: (b) A. Hoffmann-R¨oder and N. Krause, Angew.Chem., Int. Ed., 2004, 43, 1196–1216; (c) S. Yu and S. Ma,Angew. Chem., Int. Ed., 2012, 51, 3074–3112; (d) R. K. Neffand D. E. Frantz, ACS Catal., 2014, 4, 519–528; (e) J. Ye andS. Ma, Org. Chem. Front., 2014, 1, 1210–1224.2 (a) P. Rona and P. Crabbe, J. Am. Chem. Soc., 1968, 90, 4733–4734; (b) R. S. Brinkmeyer and T. L. Macdonald, J. Chem. Soc.,Chem. Commun., 1978, 876–877; (c) A. C. Oehlschlager andE. Czyzewska, Tetrahedron Lett., 1983, 24, 5587–5590; (d)A. Alexakis, I. Marek, P. Mangeney and J. F. Normant,Tetrahedron Lett., 1989, 30, 2387–2390; (e) A. Alexakis,I. Marek, P. Mangeney and J. F. Normant, Tetrahedron,1991, 47, 1677–1696; (f) J. A. Marshall and K. G. Pinney, J.Org. Chem., 1993, 58, 7180–7184; (g) J. P. Varghese,P. Knochel and I. Marek, Org. Lett., 2000, 2, 2849–2852.3 (a) I. Marek, P. Mangeney, A. Alexakis and J. F. Normant,Tetrahedron Lett., 1986, 27, 5499–5502; (b) A. Alexakis,I. Marek, P. Mangeney and J. F. Normant, J. Am. Chem.Soc., 1990, 112, 8042–8047; (c) M. T. Crimmins andK. A. Emmitte, J. Am. Chem. Soc., 2001, 123, 1533–1534; (d)M. Lecl`ere and A. G. Fallis, Angew. Chem., Int. Ed., 2008, 47,568–572; (e) H. Ohmiya, U. Yokobori, Y. Makida andM. Sawamura, Org. Lett., 2011, 13, 6312–6315.4 (a) R. K. Dieter, N. Chen and V. K. Gore, J. Org. Chem., 2006,71, 8755–8760; (b) H. Li, D. M¨uller, L. Gu´en´ee andA. Alexakis, Org. Lett., 2012, 14, 5880–5883; (c) D. Qian,L. Wu, Z. Lin and J. Sun, Nat. Commun., 2017, 8, 567.5 Extensive studies were done by S. Ma and others: (a)M. O. Frederick, R. P. Hsung, R. H. Lambeth, J. A. Mulderand M. R. Tracey, Org. Lett., 2003, 5, 2663–2666; (b)X. Jiang, C. Fu and S. Ma, Chem.–Eur. J., 2008, 14, 9656–9664; (c) Q. Li, C. Fu and S. Ma, Angew. Chem., Int. Ed.,2012, 51, 11783–11786; (d) Q. Li, C. Fu and S. Ma, Angew.Chem., Int. Ed., 2014, 53, 6511–6514; (e) J. Dai, X. Duan,J. Zhou, C. Fu and S. Ma, Chin. J. Chem., 2018, 36, 387–391;(f) B. Wang, X. Wang, X. Yin, W. Yu, Y. Liao, J. Ye,M. Wang and J. Liao, Org. Lett., 2019, 21, 3913–3917.6 (a)Thereactivityandcongurationalstabilityisconsiderably higher in THF. For details, see: ; J. Skotnitzki,L. Spessert and P. Knochel, Angew. Chem., Int. Ed., 2019,58, 1509–1514. (b) For a recent review, see: J. Skotnitzki,A. Kremsmair and P. Knochel, Synthesis, 2020, 52, 189–196.7 J. Skotnitzki, A. Kremsmair, B. Kicin, R. Saeb, V. Ruf andP. Knochel, Synthesis, 2020, 52, 873–881.8 (a) V. Morozova, J. Skotnitzki, K. Moriya, K. Karaghiosoff andP. Knochel, Angew. Chem., Int. Ed., 2018, 57, 5516–5519; (b)J. Skotnitzki, V. Morozova and P. Knochel, Org. Lett., 2018,20, 2365–2368.9 (a) Propargyl bromide is commercially available as a solution intoluene; (b) Propargyl acetate is commercially available (Sigma-Aldrich); (c) N. N. Solodukhin, N. E. Borisova, A. V. Churakovand K. V. Zaitsev, J. Fluorine Chem., 2016, 187, 15–23; (d)J. Eisenblaetter, M. Bruns, U. Fehrenbacher, L. Barner andC. Barner-Kowollik, Polym. Chem., 2013, 4, 2406–2413; (e)M. Hojo, R. Sakuragi, S. Okabe and A. Hosomi, Chem.Commun., 2001, 357–358. For details, see ESI.†10 The use of a phenyl group in a-position was unsuccessfulduetodimerisationofthecorrespondingbenzyl–alkylcopper reagent. Furthermore, we prepared racemicalkyl iodides bearing a n-butyl and cyclohexyl substituentin a-position, which could be used successfully for thepreparation of allenes. However, the preparation of thecorresponding chiral alkyl alcohols is more challengingand under investigation in our laboratories.11 The addition of ZnCl2 to the alkylcopper reagent syn-1a as inref. 6 and7 led to the corresponding alkylcopper-zincreagent.Aeradditionofpropargylicsubstrate6ecomparable regioselectivity was achieved leading to syn-7a,however in lower diastereomeric ratio and yield (dr ¼91 : 9 and 40% yield).12 (R)-(+)-3-Butyn-2-ol is commercially available (TCI; er >99 : 1).13 The enantiomeric ratio was determined by chiral GCanalysis or chiral HPLC analysis. For details, see ESI.†14 The enantiomeric ratio was determined by chiral GCanalysis. For details, see ref. 6.15 A. Czepa and T. Hofmann, J. Agric. Food Chem., 2004, 52,4508–4514.16 The yield was determined by GC-analysis using dodecane asinternal standard.17 A detailed description of the theoretical methodology, alongwith optimized structures and energies of all investigatedcompounds can be found in the ESI.†18 (a) S. Miertuˇs, E. Scrocco and J. Tomasi, Chem. Phys., 1981,55, 117–129; (b) Continuum Solvation Models in ChemicalPhysics, John Wiley & Sons, Ltd, 2007, DOI: 10.1002/9780470515235; (c) J. Tomasi, B. Mennucci and R. Cammi,Chem. Rev., 2005, 105, 2999–3094.19 Coordination of more than one solvent molecule decreasedthe free energy. For details, see ESI.†20 J. Skotnitzki, A. Kremsmair, D. Keefer, Y. Gong, R. de Vivie-Riedle and P. Knochel, Angew. Chem., Int. Ed., 2020, 59,320–324.21 We also performed DFT-calculations for the transition stateenergy with THF (ts-9) and diethyl ether (ts-10) as ligands.The energies are slightly higher (55.7 kcal mol�1 and57.4 kcal mol�1). For details, see ESI.†22 All attempts to investigate the bimolecular epimerizationpathway were unsuccessful due to inconclusive resultsfrom the DFT calculations.5332 | Chem. Sci., 2020, 11, 5328–5332This journal is © The Royal Society of Chemistry 2020Chemical ScienceEdge Article | RESEARCH GATE |
Jurnal Manajemen Indonesia (Vol. 21(1), pp. 17-29, 2021) Article info Received (07/02/2020) Revised (22/04/2020) Accepted (31/10/2020) Corresponding_author: irniyunita82@yahoo.com DOI: 10.25124/jmi.v21i1.2461 Copyright@2020. Published by School of Economics and Business – Telkom University This Journal is available in Telkom University online Journals Jurnal Manajemen Indonesia Journal homepage: journals.telkomuniversity.ac.id/ijm The Influence Of User-Generated Content To Consumer-Based Brand Equity Through Involvement In Indonesia's Top Brand Lipstic Consumer Sunu Puguh Hayu Triono1, Ratih Huriyati2 and Mokh Adib Sultan3 1, 2, 3Doktor Ilmu Manajemen, Sekolah Pascasarjana, Universitas Pendidikan Indonesia, Bandung, Indonesia Abstract Since 2000’s content marketing began to develop and become a buzzword lately. Nowadays, content marketing is a standard practice of business and one who do not use that will be left behind. The current democratization of information is expressed mainly through User-Generated Content (UGC), so that the media model becomes more user-centric. This study attempts to discuss the effect of UGC on consumer-based brand equity (CBBE) through involvement. UGC in this study is represented through variables that motivate it, namely co-creation, empowerment, community, and self-concept. The population in this study is Top Brand lipstick consumers who have watched the UGC of those brands. Data was collected using questionnaire that distributed online and 121 data were obtained. After analyzed using SEM PLS, the results are: 1) co-creation has positive effect on involvement; 2) self-concept has positive effect on involvement; 3) involvement has a positive effect on CBBE; 4) UGC has a substantial influence on involvement; 5) UGC and involvement have a moderate influence on CBBE. Keywords— brand equity; consumer-based brand equity; content marketing; involvement; user-generated content Abstrak Content marketing sejak tahun 2000 mulai berkembang dan menjadi buzzword akhir-akhir ini. Content marketing menjadi praktek standar perusahaan-perusahaan saat ini, dan perusahaan yang tidak menggunakannya akan tertinggal. Demokratisasi informasi yang terjadi saat ini utamanya diekspresikan melalui User-Generated Content (UGC), sehingga model media menjadi lebih user-centric. Penelitian ini bertujuan untuk memahami pengaruh UGC terhadap consumer-based brand equity (CBBE) melalui involvement. UGC pada penelitian ini direpresentasikan melalui variabel yang memotivasinya yaitu co-creation, empowerment, community, dan self-concept. Populasi pada penelitian ini yaitu konsumen lipstik Top Brand yang pernah menonton UGC brand tersebut. Data dikumpulkan melalui kuesioner yang disebarkan secara online dan diperoleh 121 data yang sesuai kriteria. Setelah dianalisis menggunakan SEM PLS, diperoleh hasil yaitu: 1) co-creation berpengaruh positif terhadap involvement; 2) self-concept berpengaruh positif terhadap involvement; 3) involvement berpengaruh positif terhadap CBBE; 4) UGC memiliki pengaruh substansial terhadap involvement; 5) UGC dan involvement memiliki pengaruh yang moderat terhadap CBBE. Kata kunci— brand equity; consumer-based brand equity; content marketing; involvement; user-generated content I. INTRODUCTION Marketing as an effort to understand the needs and human desires, need to combine many different fields, such as economics, business, psychology, sociology, or even anthropology. Basically, a marketer is required to be able to understand humans, both individually and in groups, then utilizing various factors, both internal and external, so that the sales can occur. From the pragmatic point of view, marketing as a method or technique in doing such engineering is related directly to the ongoing context. So that marketing develops from product-driven (1.0) to consumer-oriented (2.0), then develops to be human-centric (3.0) and then the development of digital technology Triono et al. Jurnal Manajemen Indonesia (Vol. 21(1), pp. 17-29, 2021) 18 is in turn changes economic conditions and digital economies emerge. All of those changes, accumulatively, makes marketing need to adapt, so marketing 4.0 then emerge. The key themes of marketing 4.0 lay in digital economics such as sharing-economy, the now-economy, omnichannel, social customer relationship management, content marketing, etc. Nowadays, content marketing is a standard practice of business and one who do not use that will be left behind (Ruffolo, 2017). Content marketing is an approach of marketing that use content (digital content) to creates some engagement through conversation with customer, so the content need to carefully selected, planned, deliver, and amplified, to meet the targeted customers style (Kotler et al., 2017). Content marketing is mostly driven by social media, where in the previous era consumers had little choice of information except advertisements from broadcast media. However, with the democratization of information through social media, anyone could create content. It turns out that the content which generated by consumers –user-generated content (UGC)– is preferred and trusted by the other consumers. Contents in YouTube that is most watched are UGC, so is the most subscribed channels (Kotler et al., 2017). Articles about content marketing can be traced since 2000 and since then the discussions have increased and become hottest buzzword lately (Mayer et al., 2000; Keyes, 2006; Pulizzi & Barrett, 2009; Halvorson, 2010; Handley & Chapman, 2010; Gunelius, 2011; Lieb, 2011; Nelli, 2012; Christodoulides et al., 2012; Jefferson & Tanton, 2013; Rahimia & Hassanzadeh, 2013; Slater, 2014; Ramos, 2014; Elisa & 2014 Gordini, 2014; Forrest, 2019). Content marketing is touted as a form of advertising in the era of the digital economy, ultimately aimed at generating curiosity about a brand (Kotler et al., 2017). With the democratization of information, starting with the Web 2.0 era, the role of internet users has increased, not only as consumers, but also producers of information. The role of users is becoming increasingly dominant with the presence of social media technology, so that online digital content is dominated by content generated by user (UGC). Christodoulides et al. (2012) obtained a model about content marketing that involving UGC and analyze the content creators. This study analyzing the relationship of UGC, involvement of user, and consumer-based brand equity (CBBE). However, they suggested to test the model with different subject, that is from content viewer. The unit of analysis used in this study is the consumer of Top Brand Indonesia lipstick in 2019, namely Wardah, Revlon, Maybelline, Pixy, and Viva. Lipstick products were chosen because they are cosmetic products that are commonly used by women, are easily found anywhere, and quite a lot of the consumer that creates a content, especially online reviews. II. LITERATURE REVIEW In this transition era, a new marketing approach is needed. Thus, Kotler et al. (2017) introduced marketing 4.0 as a continuity from marketing 3.0. The main cause of this evolution is the change of customers characteristics and nature in the face of digital economy. Marketing 4.0 need to be more human-centered, so the brands has such intimacy with the customer by humanizing the brands, have conversation through content, upscale the sales by integrating many form of channels (omnichannel), and involving customer more. Content marketing is an approach of marketing that use content (digital content) to creates some engagement through conversation with customer, so the content need to carefully selected, planned, deliver, and amplified, to meet the targeted customers style (Kotler et al., 2017). By using content marketing, the deeper connection or relationship with customer might be easier to achieve. The early UGC was marked by participatory inequality in which the parties could create only a few among others (Ochoa & Duval, 2008), so that from its origin, UGC was a niche activity and was driven by the accumulation of various things (Daugherty, Eastin, & Bright, 2008). The most frequently cited definition of UGC comes from the Organization for Economic Co-operation and Development (OECD) (2007), which defines UGC as having three main characteristics: content made publicly available via the internet; content that reflects a number of creative endeavors; and content created outside routine and professional practice. From personal characteristics, Daugherty et al. (2011) state that there are three things that motivate people to consume UGC, namely attitudes towards UGC, media experience, and the desire to control. To this, it can be observed that the factors driving the creation of UGC that proposed by Christodoulides et al. (2012) are still relevant to various factors that motivate people to consume UGC. Therefore, in this study the same factors as Christodoulides et al. (2012) are used to test the research models, namely co-creation, empowerment, community, and self-concept. Triono et al. Jurnal Manajemen Indonesia (Vol. 21(1), pp. 17-29, 2021) 19 Vargo & Lusch (2004) stated that co-creation is the participation of customer in creating value, at any points. Meanwhile, UGC is an online content that produced by consumer, involving collaboration with other consumer or the companies itself, and that adds value for companies (Humphreys & Grayson, 2008). In this democratize internet era, consumer has more power in determining their information consumption. The consumer perception of control can be achieved by accessing online content that other consumer creates (UGC) so it enabling them to have more connection (Harrison, Hunter & Waite, 2006) and able to define their own brand value (Christodoulides et al., 2009). The study of Wathieu et al. (2002) stated that there are three factors that influence consumer empowerment, those are the ability to determine choices, sense of progress in decision making process, and information of the other consumer. Social web allows communities that is not passive, rather interactive and involve in co-creation of brand value (de Chernatony & Christodoulides, 2004). It should be noted that UGC is primarily driven by digital-born or digital natives generation that characterized by young and digitally skilled, and behave differently from the previous generation (Burmann & Arnhold, 2008). Consumer can share their ideas or express themselves by involving in UGC. Consumer self-concepts influence the way they see the personality of a brand. For example, favored brands are formed, partly, by personalities that are preferred by consumers. By utilizing brands, UGC allows consumers to attract attention and to tell who they are from their own perspective (Berthon et al., 2008). Involvement influenced by three main factors, those factors are related to the characteristics of the person, the stimulus, and the situation (Zaichkowsky, 1986; Bloch & Richins, 1983). These factors can influence the level of involvement with product, advertising, or purchasing situation (Hupfer & Gardner, 1971; Krugman, 1965, 1967; Clarke & Belk 1978). Customer involvement is a significant factor in influencing customer behavior and higher involvement means higher loyalty (Jones & Sasser, 1995; Knox & Walker, 2003). Concept of brand equity is referred as more consumers view or based on consumer (CBBE) for some researchers (Cobb-Walgren et al., 1995; Sinha & Pappu, 1998; Yoo & Donthu, 2001, 2002; Yoo et al., 2000; Washburn & Plank, 2002). CBBE according to Aaker (1991) is "A set of brand assets and liabilities linked to a brand, its name and symbol that add to or subtract from the value provided by a product or service to a firm and / or to that firm’s customers". Schivinski & Dabrowski (2014) measures CBBE refers to four of Aaker's five core brand equity dimensions, namely: brand awareness, brand association, perceived quality, and brand loyalty. From those literature study, we propose five hypotheses and then arranged to the research model as follows: H1: Co-creation has positive effect on consumer involvement with the brand through UGC. H2: Consumer empowerment has positive effect on consumer involvement with the brand through UGC. H3: Community has positive effect on consumer involvement with the brand through UGC. H4: Self-concept has positive effect on consumer involvement with the brand through UGC. H5: Consumer involvement has positive effect on consumer-based brand equity. Triono et al. Jurnal Manajemen Indonesia (Vol. 21(1), pp. 17-29, 2021) 20 Fig. 1 Research Model III. RESEARCH METHODOLOGY The analysis technique used in this study is Partial Least Square (PLS) using SmartPLS 3.0 software. The population is consumers of Top Brand Indonesia's lipstick in 2019 (Wardah, Maybelline, Revlon, Pixy, Viva) who have watched UGC of those brands. The size of population is unknown because there is no data of exact number of Top Brand lipstick consumers and those who have watched the UGC. Even so, because this study uses PLS, then the minimum number of respondents can be concluded as five times of path (50 respondents). This study uses nonprobability sampling and using purposive or judgmental sampling method. Data collection was carried out using a questionnaire created in digital format using GoogleForm and distributed online. Out of 143 responses, 121 responses were chosen that fit the criteria. The questionnaire was arranged in two parts. The first part contains respondent’s characteristics and a question about whether or not they watch UGC Top Brand lipstick. If the respondent ever watched the UGC, then they proceed to the second part which contains questions about the variables. But if they has never watched the UGC, then the filling process is immediately stated as complete. Co-creation is measured using four question items. Empowerment was measured using three questions. Community is measured using three questions. Self-concept is measured using two questions. Question items for measuring those four variables were adopted from Christodoulides et al. (2012) based on interval scales ranging from 1 to 7 to state the level of agreement. Whereas to measure involvement, it was adopted from Zaichkowsky (1994) using seven semantic differential scale items that has range from 1 to 7. To measure CBBE, 13 question items were adopted from Schivinski & Dabrowski (2014) with interval scale ranging from 1 to 7. IV. RESULT OF STUDY AND DISCUSION There are five questions related to the respondent's characteristics, namely gender, age, occupation, income, and the UGC brand that they ever watched. Table 1. Sample Demography by Gender Gender Number Male 12 Female 109 Triono et al. Jurnal Manajemen Indonesia (Vol. 21(1), pp. 17-29, 2021) 21 Based on gender, 109 respondents were female and 12 male respondents. This is not surprising considering the product is lipstick, a product whose designed for women. Even so, it turns out there are male respondents who have watched the UGC of lipstick. Table 2. Sample Demography by Age Age (years old) Number 15 – 25 52 25 – 35 35 35 – 45 28 >45 6 Based on age, it can be seen from Table 2 that the composition of respondents is dominated by the range of 15 – 25 years old (generation Z) which is 52 respondents or 42.98% of the total respondents. The second most is the range of 25 – 35 years (generation Y) totaling 35 respondents or 28.93% of the total respondents. Respondents from generation Y and Z, or commonly referred as millennials, dominate the total respondents. This is probably because the millennials have higher technological skill, considering the generation Z is a generation that has been exposed to digital technology from birth (digital native), so does the generation Y that is growing up with digital technology. These technological skills then lead to different patterns of information seeking for each generation, where generation Y and Z prefer digital content (in this case UGC) to find information about the lipstick that they want. Table 3. Sample Demography by Occupation Occupation Number Student 17 Housewife 10 Employee 39 Entrepreneur 9 Civil servant 8 Professional 24 Others 14 Based on occupation, it can be seen from Table 3 that the most respondents work as employees that is as much as 32.23% of total respondent. The next most is of professionals and then the students. We can argue that this is because employees, professionals and students need more attractive appearance compared to housewives, entrepreneurs, civil servants and others. Table 4. Sample Demography by Income Income (million rupiahs per month in average) Number <5 88 5 – 10 26 10 – 15 3 15 – 20 3 >20 1 Respondents were also observed based on their income, in this case, average income per month. We categorized the income by interval of 5 million rupiah because generally those interval has quite a noticeable Triono et al. Jurnal Manajemen Indonesia (Vol. 21(1), pp. 17-29, 2021) 22 difference, from the lifestyle, cars, house, and so on. Based on Table 4, it can be seen that the most respondents have income below 5 million rupiah. The second largest is the range of income between 5 to 10 million rupiah. If those two are added together, a total of 114 (94.21%) is obtained. This means that most respondents are in the middle to lower social class. Even so, it can be seen that although respondents are not of high social class, the need for lipstick as a support for appearance or included in lifestyle is also needed. Another thing that can be seen is that the utilization of UGC is not only for higher social class, but in the lower too. This can be a further discussion regarding content marketing and its relationship with social class. Table 5. Sample Demography by Brand of UGC Brand of UGC Number Wardah 59 Revlon 4 Maybelline 29 Pixy 9 Viva 5 Others 15 In this study, we use five lipstick brands that achieve Indonesia’s Top Brand in 2019. The five brands and the brand index are: Wardah (33.2%), Revlon (9.2%), Maybelline (7.7%), Pixy (6.0%) and Viva (4.5%) (Frontier Groups, 2019). However, in this study we still offer "Other" option in order to see if there is a UGC that is not related to the five Top Brand lipsticks, or in other words, are there consumers from other brands that create or produce content. From Table 5, it can be seen that the most watched UGC is Wardah, with 59 respondents or 48.76% of the total respondents. This is relevant with the highest brand index. The second largest seen is Maybelline with 29 respondents or 23.97% of all respondents. This is different with the brand index achievement, where the second highest index brand is Revlon. Then the third largest is from the other brands as many as 15 people or 12.40% of all respondents. This is also interesting, considering the respondents in this study watched not only UGC about big brands, but also from brands that were not listed in Top Brand. This can be a subject for further discussion and research. The description or description of the variables in this study uses the mean and standard deviation based on the division of categories from Simamora (2004). The results of data acquisition for each variable are summarized in the following table. Table 6. Variable Means Brand of UGC Number Category Co-creation 5,195 High Empowerment 4,849 High Community 4,579 High Self-Concept 4,783 High Involvement 5,421 Very high CBBE 5,458 Very high SmartPLS analyzes the relationship between variables and indicators, namely measurement/outer model and structural/inner model. Measurement model is a characteristic specification that explains the relationship between latent variables and their indicators. While the structural/inner model is a characteristic specification that explains the relationship between latent variables. The first step is to do outer model testing that resulted as follow. Triono et al. Jurnal Manajemen Indonesia (Vol. 21(1), pp. 17-29, 2021) 23 Fig. 2 Outer Model Test Result Convergent validity is assessed based on outer loading. The criteria used for convergent validity is outer loading> 0.5. It can be seen from Figure 2 that there are 7 indicators that do not meet the convergent validity criteria, namely BE7, BE8, BE9, BE10, BE11, BE12, and BE13, so this indicator is not included in the next analysis. Then the outer loading was re-tested and found that all indicators met the convergent validity criteria. Discriminant validity test is assessed based on cross loading measurements with the construct. Another method used to assess discriminant validity is to compare AVE roots for each construct with the correlation of all constructs in the model. The model has sufficient discriminant validity if the root of AVE for each construct is greater than the correlation of all constructs in the model, and the biggest cross loading is in the construct that it formed. Table 7. Cross Loading Factor Indicator item Construct Co-creation Empowerment Community Self-Concept Involvement CBBE COCR1 0.858 0.553 0.432 0.595 0.733 0.719 COCR2 0.869 0.664 0.587 0.645 0.658 0.680 COCR3 0.887 0.641 0.545 0.575 0.671 0.669 COCR4 0.809 0.676 0.566 0.629 0.648 0.642 EMP1 0.770 0.891 0.621 0.671 0.651 0.640 EMP2 0.606 0.898 0.761 0.729 0.634 0.534 EMP3 0.567 0.866 0.744 0.608 0.537 0.600 COMM1 0.605 0.808 0.907 0.665 0.629 0.560 Triono et al. Jurnal Manajemen Indonesia (Vol. 21(1), pp. 17-29, 2021) 24 COMM2 0.473 0.678 0.900 0.626 0.514 0.463 COMM3 0.583 0.664 0.903 0.653 0.613 0.591 SC1 0.700 0.699 0.646 0.928 0.755 0.715 SC2 0.583 0.678 0.663 0.885 0.605 0.581 INV1 0.717 0.679 0.621 0.669 0.865 0.677 INV2 0.690 0.566 0.523 0.662 0.882 0.815 INV3 0.758 0.637 0.610 0.736 0.933 0.774 INV4 0.793 0.629 0.586 0.682 0.942 0.766 INV5 0.662 0.641 0.635 0.713 0.884 0.672 INV6 0.711 0.596 0.582 0.676 0.930 0.733 INV7 0.700 0.645 0.598 0.677 0.918 0.740 BE1 0.685 0.709 0.672 0.688 0.723 0.782 BE2 0.573 0.465 0.404 0.558 0.535 0.751 BE3 0.629 0.508 0.450 0.503 0.603 0.848 BE4 0.761 0.715 0.643 0.725 0.787 0.888 BE5 0.620 0.421 0.385 0.524 0.679 0.877 BE6 0.680 0.485 0.412 0.585 0.717 0.863 Table 8. Correlation and Root of AVE Variable AVE Root of AVE Co-creation Empowerment Community Self-Concept Involvement CBBE Co-creation 0.733 0.856 0.856 Empowerment 0.816 0.903 0.618 0.903 Community 0.700 0.837 0.793 0.600 0.836 Self-Concept 0.783 0.885 0.737 0.797 0.668 0.885 Involvement 0.825 0.908 0.793 0.653 0.816 0.690 0.908 CBBE 0.822 0.907 0.713 0.719 0.721 0.759 0.757 0.907 Based on Table 7 it can be seen that the largest cross loading is in the construct that it forms, so it can be concluded that the research indicators meet discriminant validity. Whereas based on Table 8, it can be seen that the root of AVE for each construct is greater than the correlation of all construct in the model so that it is declared valid in discriminant validity testing. In addition to the validity test, PLS also conducts a reliability test to measure the internal consistency of the measuring instrument. Reliability testing in PLS using composite reliability. Triono et al. Jurnal Manajemen Indonesia (Vol. 21(1), pp. 17-29, 2021) 25 Table 9. Composite reliability Construct Composite reliability’s score Rule of thumb Status Co-creation 0.917 0.7 Reliable Empowerment 0.916 0.7 Reliable Community 0.930 0.7 Reliable Self-Concept 0.902 0.7 Reliable Involvement 0.970 0.7 Reliable CBBE 0.933 0.7 Reliable Based on Table 9, it can be seen that the composite reliability value of all variables is greater than 0.7, so the entire construct is declared reliable. After passing the outer model testing stage, the next stage is the analysis of structural or inner models. Assessing based on the value of R-square is used to measure the level of variation in the changes of the independent variable to the dependent variable. The higher R-square means the better the prediction model of the research proposed. Table 10. R-square Endogenous Construct R-square Involvement 0.710 CBBE 0.661 From Table 10, it can be seen that the R-square for the Involvement is 0.710, meaning that Co-creation, Empowerment, Community, and Self-Concept able to explain 71.0% of Involvement, while 29.0% is explained by other variables outside this study. R-square of Involvement (0.710) is in the substantial category. R-square for the CBBE is 0.661, meaning that Involvement, Co-creation, Empowerment, Community, and Self-Concept able to explain CBBE by 66.1%, while 33.9% is explained by other variables outside this study. R-square of CBBE (0.661) is in the moderate category. Inter-construct causality test in PLS structural model is assessed by the path coefficient or t-values for each path to test the significance. The path coefficient or inner model score indicated by the t statistic, must be greater than 1.98 for one-tailed testing at alpha (α) 5 percent. Table 11. Hypotheses test Hypotheses Path Path coefficient t-statistic Decision H1 Co-creation Involvement 0,501 6,267 Accepted H2 Empowerment Involvement -0,039 0,329 Rejected H3 Community Involvement 0,135 1,612 Rejected H4 Self-Concept Involvement 0,333 3,195 Accepted H5 Involvement Consumer-Based Brand Equity 0,813 26,050 Accepted Of the five hypotheses proposed, there are two hypotheses rejected, that are H2 and H3. H2 (positive effect of empowerment on involvement) is rejected because the path coefficient is negative and the t statistic is less than 1.98. Whereas H3 (positive effect of community on involvement) was rejected because the t statistic was less than 1.98. In contrast, H1, H4, and H5 are accepted because they have positive path coefficients and the t statistic is greater than 1.98. Summary of the results of testing the research hypotheses is as follows: 1) H1 is accepted; it means that co-creation has positive effect on consumer involvement with the brand through UGC. The more consumer perceives that a brand is co-created, the higher their involvement. Triono et al. Jurnal Manajemen Indonesia (Vol. 21(1), pp. 17-29, 2021) 26 2) H2 is rejected; it means that consumer empowerment does not has positive effect on consumer involvement with the brand through UGC. The more consumer perceives that a brand can empower them, it does not necessarily increase their involvement with the brand through UGC. 3) H3 is rejected; it means that community does not has positive effect on consumer involvement with the brand through UGC. The more consumer perceives that a brand facilitates community, it does not necessarily increase their involvement with the brand through UGC. 4) H4 is accepted; it means that self-concept has positive effect on consumer involvement with the brand through UGC. The more consumer perceives that a brand expresses their self-concept, the higher their involvement with the brand through UGC. 5) H5 is accepted; it means that consumer involvement has positive effect on consumer-based brand equity. The higher the level of consumer involvement with UGC, the higher the consumer perceived brand equity. This finding might be a consideration for the five Top Brand lipstick to pay more attention in the implementation of content marketing strategies through UGC to increase their perceived brand equity in the eyes of consumer. Another thing in the content marketing implementation are related to the company's openness in accepting input from consumers so that it allow the innovation from co-creation. Also there is need to encourage consumers to create content that is appropriate or truly expresses themselves completely. Two of the four exogenous variables have been proved to have a significant positive effect on endogenous variables. While endogenous variables are also proved to have a positive and significant effect on other endogenous variables. Table 12. Rank of influence by magnitude Rank Path Path coefficient 1 Involvement Consumer-Based Brand Equity 0.813 2 Co-creation Involvement 0.501 3 Self-Concept Involvement 0.333 The path coefficient numbers in these variables have the following meanings: 1) Involvement Consumer-Based Brand Equity = 0.813, meaning that the influence of involvement on consumer-based brand equity of Top Brand lipstick consumers in Indonesia is 0.813. This means that the higher customer involved with the product through UGC, the higher their brand equity. It means that the customer perceive that the product is more valuable, has higher quality, they become more aware and associated their self with the product more. The effect of customer involvement to CBBE is substantial, means that the management need to generates more UGC to increase the involvement level of their customer so the brand would be more “valuable” in the eye of the customer and get more loyal customer. 2) Co-creation Involvement = 0.501, meaning that the influence of co-creation on involvement of Top Brand lipstick consumers in Indonesia is 0.501. This means that the more customer perceived that the brand is co-created through UGC, the higher their involvement with the product. The effect of co-creation to involvement is moderate. Management need to selectively response their customer inputs from UGC and applied those inputs to the brand. 3) Self-Concept Involvement = 0.333, meaning that the influence of self-concept on involvement of Top Brand lipstick consumers in Indonesia is 0.333. This means that the more customer perceives that the brand expresses their self-concept, the higher their involvement with the product through UGC. This effect is considered as weak, means that management should not concern a lot with how their brand talk about their customer’s self-concept in UGC, just keep an adequate level of self-concept; more general, not too personal. This is probably because most customer is more influenced rather than influencing, so the self-concept is something dynamic, influenced with other self-concept, mainly of influencer. Triono et al. Jurnal Manajemen Indonesia (Vol. 21(1), pp. 17-29, 2021) 27 V. CONCLUSIONS AND SUGGESTION This study aims to understand the relationship of UGC to Customer-Based Brand Equity, in the eye of the viewer, as it has bigger number than the creator. We tested a model from previous research but with different side of view. From this study, we find that UGC dimensions has effect on CBBE mediated by involvement level of customer. From the study of consumer of Indonesia Top Brand lipstick we able to draw some conclusions. First, the dimension of UGC that has significant influence on involvement are co-creation and self-concept. Second, the influence of co-creation on involvement is positive and considered as moderate with magnitude of 0.50. Third, the influence of self-concept on involvement is positive and considered as weak with magnitude of 0.333. Fourth, involvement has significant and positive effect on CBBE with magnitude of 0.813 which considered as substantial. Fifth, the contribution of UGC dimensions in explaining involvement is considered as substantial with R-square of 0.701. Sixth, the UGC dimensions and involvement altogether has moderate contribution in explaining CBBE with R-square of 0.661. This study has some limitations and need further research. First, we only tested the same model that has been proposed earlier from previous research, this model could be developed more with some motivational factors of viewer to watch UGC. Second, we only test the model in a narrow scope, that is consumer of product of lipstick. Future research could expand the scope to other industry to get more understanding, considering that nowadays many different industries, from cosmetics to smartphone, has content created by their customer. Third, we’re not test the loop path as proposed from the previous research of Christodoulides et al. (2012), future research could take this path into consideration. Fourth, we suggest that it might be necessary to examine the effect of age and social class on UGC and also the effect of its moderation on the model of this study. Finally, our research shows that consumer empowerment and community does not significantly influence consumer online involvement with UGC. This is quite interesting because in online world is related closely with high social interactions and able to form a particular community based on their interests. However, it was found that the community apparently had no significant effect on online consumer online involvement. Therefore, it needs to further investigated to confirm this effect. REFERENCES Aaker, J. L. (1997). Dimensions of Brand Personality. Journal of Marketing Research, 34(3), 347–356. Berthon, P., Pitt, L., & Campbell, C. (2008). Ad Lib: When Customers Create the Ad. California Management Review, 50(4), 6–30. Bloch, P. H., & Richins, M. L. (1983). A Theoretical Model for the Study of Product Importance Perceptions. Journal of Marketing, 47(3), 69. Bruns, A. (2007). Produsage: Towards a Broader Framework for User-Led Content Creation. Knowledge Management, 6(8), 99-106. Burmann, C., and Arnhold U. (2008). User Generated Branding: State of the Art Research. Berlin: LIT Verlag Berlin-Hamburg-Münster. Chaffey, D., Mayer, R., Johnston, K., & Ellis-Chadwick, F. (2000). Internet marketing. Edinburgh: Pearson Education. Christodoulides, G., Jevons, C., & Bonhomme, J. (2012). Memo to marketers: Quantitative evidence for change - how user-generated content really affects brands. Journal of Advertising Research, 52(1), 53–64. Christodoulides, G. (2009). Branding in the post-internet era. Marketing Theory, 9(1), 141–144. Clarke, K., & Belk, R. W. (1979). “The Effects of Product Involvement and Task Definition on Anticipated Consumer Effort” in NA - Advances in Consumer Research Volume 06, eds. William L. Wilkie, Ann Abor, MI : Association for Consumer Research, Pages: 313-3 Cobb-Walgren, C. J., Ruble, C. A., & Donthu, N. (1995). Brand equity, brand preference, and purchase intent. Journal of Advertising, 24(3), 25–40. Daugherty, T., Eastin, M. S., & Bright, L. (2008). Exploring Consumer Motivations for Creating User-Generated Content. Journal of Interactive Advertising 8(2), 16–25. Elisa, R., & Gordini, N. (2014). Content Marketing Metrics: Theoretical Aspects and Empirical Evidence. European Scientific Journal, 10(34), 1857–7881. Forrest, P. (2019). Content Marketing Today. Forrest, P. and Piper, W. S. (2018.) Increasing Business Dependence on Social Media, presented at Society for Business, Industry and Economics Conference, Destin, April 10—13, 2018: FL Gensler. Triono et al. Jurnal Manajemen Indonesia (Vol. 21(1), pp. 17-29, 2021) 28 Gardner, B., and Levy, S. (1955). The Product and the Brand. Harvard Business Review, 33(2), 33–39. Gunelius, S. (2011) Content Marketing for Dummies. New York: Wiley. Halvorson, K. (2010). Content strategy for the web. Barkeley: New Riders. Handley, A., & Chapman. (2010). C.C. Content Rules. New York: Wiley Harrison, T., Hunter, G. L., & Waite, K. (2006). The Internet, Information and Empowerment. European Journal of Marketing, 40(9), 972–993. Humphreys, A., & Grayson, K. (2008). The Intersecting Roles of Consumer and Producer: A Critical Per- spective on Co-Production, Co-Creation and Prosumption. Sociology Compass, 2(3), 963–980. Hupfer, N. T., & Gardner, D. M. (1971). Differential involvement in Products and Issues: An Exploratory Study. Advances in Consumer Research, 2, 262–270. Jefferson, S., & Tanton, S. (2013). Valuable Content Marketing. How to make quality content the key of your business success. London: Kogan Page. Jones, T. O., & Sasser, W. E. Jr. (1995). Why satisfied customers defect. Harvard Business Review, 73(4), 88-99. Keller, K.L. (1993). Conceptualizing, measuring and managing customer-based brand equity. Journal of Marketing, 57(1), 1-22 Keyes, J. (2006). Knowledge management, business intelligence and content management. Broken South Parken: Auerbach Publication Knox, S., & Walker, D. (2003). Empirical developments in the measurement of involvement, brand loyalty and their relationship in grocery markets. Journal of Strategic Marketing, 11, 271-286. Kotler, P., Kertajaya, H., & Setiawan, I. (2017). Marketing 4.0: Moving From Traditional to Digital. New Jersey: Wiley Kotler, P., & Keller, K. L. (2009). Manajemen Pemasaran (Jilid 13 Buku 2). Jakarta: Erlangga. Krishnamurthy, S., & Dou, W. (2008). Advertising with User-Generated Content: A Framework and Research Agenda. Journal of Interactive Advertising, 8(2), 1–4. Krugman, H. (1965). The Impact of Television Advertising: Learning without Involvement. Public Opinion Quarterly, 29, 349–56. Lieb, R. (2011). Content Marketing: think like a publisher. How to use content to market online and in social media. Indianapolis: Que Publishing Nelli, R.P. (2012). Branded Content Marketing. Vita e. Milan: Pensier. OECD. (2007). Participative Web and User-Created Content: Web 2.0, Wikis, and Social Networking. Paris: Organisation for Economic Co-operation and Development Pires, G. D., Rita, P., & Stanton J. (2006). The Internet, Consumer, Empowerment and Marketing Strategies. European Journal of Marketing, 40(9), 936-949. Pulizzi, J. & Barrett, N. (2009). Get content. Get customers. New York: McGraw Hill. Rahimia, F. & Hassanzadeh, J. (2013). The impact of website content dimension and e-trust on e-marketing effectiveness: the case of Iranian commercial saffron corporations. Information and Management, 50, 240-247. Ruffolo, B. (2017). What is Content Marketing? https;//www.impactbnd.com/blog/what-is-content-marketing Saunder, M., Lewis, P., & Thornhill, A. (2009). Research Methods for Business Students (5th Editon). Essex: Prentice Education Limited Sekaran, U. & Bougie. (2010). Research Methods for Business A Skill Building Approach. Sussex: Wiley Sekaran, U. (2006). Metodologi Penelitian Untuk Bisnis (Edisi 4 Buku 2). Jakarta: Salemba Empat. Sinha, A., & Pappu, R. (1998). “Parcelling of the sub components of consumer-based brand equity using factorial survey: an empirical investigation in the New Zealand consumer electronics sector”, Proceedings, Australia New Zealand Marketing Academy Conference (ANZMAC), University of Otago, Dunedin, December, pp. 2433-8. Slater, D. (2014) Content marketing: recycling and reuse. New York: Media Corporation. Vargo, S. L., & Lusch, R. F. (2004). Evolving to A New Dominant Logic for Marketing. Journal of Marketing 68(1), 1-17 Washburn, J. H. & Plank, R. E. (2002). Measuring brand equity: an evaluation of a consumer-based brand equity scale. Journal of Marketing Theory and Practice, 10(1), 46-62 Wathieu, L., Brenner, L., & Carmon, Z. (2002). Consumer Control and Empowerment: A Primer. Marketing Letters, 13(3), 297–305 Yoo, B. & Donthu, N. (2001). Developing and validating a multidimensional consumer-based brand equity scale. Journal of Business Research, 52, 1-14 Triono et al. Jurnal Manajemen Indonesia (Vol. 21(1), pp. 17-29, 2021) 29 Yoo, B. and Donthu, N. (2002). Testing cross-cultural invariance of the brand equity creation process. Journal of Product & Brand Management, 11(6), 380-98 Yoo, B., Donthu, N., & Lee, S. (2000). An examination of selected marketing mix elements and brand equity. Journal of the Academy of Marketing Science, 28(2), 195-211 Zaichkowsky, J. L. (1985). Measuring the Involvement Construct. Journal of Consumer Research, 12(3), 341–352 Zaichkowsky, J. L. (1994). The Personal Involvement Inventory: Reduction, Revision, and Application to Advertising. Journal of Advertising, 23(4), 59–70 | RESEARCH GATE |
Motivating Self-Efficacy in DiverseBiomedical SciencePost-baccalaureate and GraduateStudents Through ScientificConference ImplementationKasey R. Boehmer 1,2*, Suelen Lucio Boschen De Souza 3, Jason D. Doles 4,Nirusha Lachman 5, Dennis Mays 6, Karen E. Hedin 7, Cheryl A. Dornink 6, Louis J. Maher 4 andJ. Luis Lujan 31Knowledge and Evaluation Research Unit, Mayo Clinic, Rochester, MN, United States, 2Division of Health Care DeliveryResearch, Mayo Clinic, Rochester, MN, United States, 3Department of Neurologic Surgery, Mayo Clinic, Rochester, MN,United States, 4Department of Biochemistry and Molecular Biology, Mayo Clinic, Rochester, MN, United States, 5Department ofClinical Anatomy, Mayo Clinic, Rochester, MN, United States, 6Office for Education Diversity, Equity, and Inclusion, Mayo Clinic,Rochester, MN, United States, 7Department of Immunology, Mayo Clinic, Phoenix, AZ, United StatesTactics to increase the number of underrepresented (UR) students in biomedical researchPhD training programs have not yet translated to UR faculty numbers that reflect thediversity of the United States. Continued interventions are required to build skills beyondthose that result in placement into a PhD program. We hypothesize that successfulinterventions must build skills that give UR students foundations for confident self-efficacyin leadership. We seek interventions that allow UR students to envision themselves assuccessful faculty. We posit that development of such skills is difficult in the classroom orlaboratory alone. Therefore, novel interventions are required. As part of the NIH-fundedPost-baccalaureate Research Education Program (PREP) and Initiative for MaximizingStudent Development (IMSD) at the Mayo Clinic Graduate School of Biomedical Sciences,we designed and implemented a unique intervention to support development of studentleadership skills: a biannual student-organized and student-led national researchconference titled “Scientific Innovation Through Diverse Perspectives” (SITDP). Thisinitiative is based on the concept that students who actively live out realistic roles asscientific leaders will be encouraged to persist to scientific leadership as faculty. Here wedescribe the motivation for, design of, and outcomes from, the first three pilot conferencesof this series. We further discuss approaches needed to rigorously evaluate theeffectiveness of such interventions in the future.Keywords: underrepresented and minority groups, education - active learning, conference participation, self-efficacy, diversity & inclusionEdited by:Candice M. Etson,Wesleyan University, United StatesReviewed by:Marc Cox,The University of Texas at El Paso,United StatesCristin Welle,University of Colorado AnschutzMedical Campus, United States*Correspondence:Kasey R. Boehmerboehmer.kasey@mayo.eduSpecialty section:This article was submitted toSTEM Education,a section of the journalFrontiers in EducationReceived: 10 September 2021Accepted: 18 November 2021Published: 07 December 2021Citation:Boehmer KR, Boschen De Souza SL,Doles JD, Lachman N, Mays D,Hedin KE, Dornink CA, Maher LJ andLujan J (2021) Motivating Self-Efficacyin Diverse Biomedical Science Post-baccalaureate and Graduate StudentsThrough ScientificConference Implementation.Front. Educ. 6:774070.doi: 10.3389/feduc.2021.774070Frontiers in Education | www.frontiersin.orgDecember 2021 | Volume 6 | Article 7740701CURRICULUM, INSTRUCTION, AND PEDAGOGYpublished: 07 December 2021doi: 10.3389/feduc.2021.774070BACKGROUND AND RATIONALE FOR THEEDUCATIONAL ACTIVITY INNOVATIONThe National Institutes of Health define UR populations in theresearch enterprise of the Unites States as individuals 1) fromspecific racial or ethnic groups (e.g., Blacks or African Americans,Hispanics or Latinos), 2) with disabilities, 3) from certaindisadvantagedbackgrounds(e.g.,firstgenerationcollegegraduates, raised in rural areas), and 4) who are women,especially those in the aforementioned categories (NationalInstitutes of Health, 2019). These individuals are less likely tobe enrolled in graduate school programs (National Institutes ofHealth, 2020a), receive PhDs (National Institutes of Health,2020b), and receive research project grants (National Institutesof Health, 2020c; National Institutes of Health, 2020d; NationalInstitutes of Health, 2020e). Research indicates that there are keytargets for intervention to increase diversity amongst researchfaculty.First, early intervention to improve diversity in the researchworkforce appears critical. For example, in the case of women,those lacking enrichment support during their early-careerfaculty phases are less likely to achieve senior faculty ranks(Holliday et al., 2014; Lopez et al., 2014; Paulus et al., 2016)and have lower network reach, measures of productivity such ash-indices, and publications (Budden et al., 2008; John et al., 2016;Bernard, 2018; Schrouff et al., 2019). Others have noted theimportanceofinterventionforURstudentsduringundergraduate learning experiences as well as during thetransition from post-doc to tenure track faculty (Meyers et al.,2018). Second, undergraduate research experience, NIH-fundedprogramstoincreasediversity,conferenceparticipation,mentorship, and institutional cultures of commitment to URstudents are the activities deemed critical for enrichment of URlearners toward the goal of persistence in careers in biomedicalresearch (Martinez et al., 2018). Specific activities such asinternshipsincareerexplorationwherestudentsspendsignificant time in an on-the-job experience in a desired fieldand academic career coaching, which focuses on guiding learnersthrough career planning steps and professional developmentactivities have been used towards developing UR student self-efficacy (Williams et al., 2017; Schnoes et al., 2018).Furthermore, UR student success is correlated with identity asa career scientist, a self-perception that is more holistic thansimply reflecting competence to conduct experiments (Kim-Prietoetal., 2013). Asa result,wehypothesizedthatsuccessful interventions must build skills that give UR studentsfoundations for confident self-efficacy in leadership. Theseinterventions allow UR students to envision themselves assuccessful faculty. We posit that development of such skills isdifficult in the classroom or laboratory alone.Toward that end, our aim was to create a platform fordeveloping critical thinking and leadership skills outside ofthe classroom and laboratory settings. To accomplish this, thejoint Postbaccalaureate Research Education Program (PREP)and Initiative for Maximizing Student Development (IMSD)at the Mayo Clinic Graduate School of Biomedical Sciencespioneereda“ScientificInnovationsThroughDiversePerspectives(SITDP)”conferenceseries,organizedbystudents, to cultivate relevant scientific leadership skills nottaught in the classroom or laboratory. This effort wassupported with funding from the National Institute ofGeneral Medical Sciences (NIGMS) of the NIH, the MayoClinic College of Medicine and Science, the Mayo ClinicOffice for Education Diversity, Equity, and Inclusion, andthe Mayo Clinic Graduate School of Biomedical Sciences.To our knowledge, the SITDP conference is the firstintervention of its kind to specifically use the execution of astudent-led conference to develop leadership and scientific self-efficacy for UR students. While PREP and IMSD programs have along-standing history, to our knowledge their application hasfocused on traditional scientific development of scholars throughmentorship in the lab setting, writing and presentation skillspractice, and partaking in coursework (McGee et al., 2012;Remich et al., 2016). Literature reviews focused programcharacteristics of initiatives designed to increase UR studentparticipation in STEM and other health-related disciplinesappear to focus more on individual skill development, evenwhen interventions are conducted in a group or cohort setting(Kirui and McGee, 2021; Ureña et al., 2021). Other interventionsthat are conducted in group settings and centered on fieldworkactivities have also been described and may also increase URstudent persistence, self-efficacy, and science identity (Bowserand Cid, 2021).As such, our aim in this brief report is to describe theoverarching learning frameworks underpinning SITDP, specificactivities carried out to plan and execute the conference, earlyevaluation of the first three iterations of the activity, and futuredirections for the program to be considered by institutionswishing to adopt similar offerings for UR students.PEDAGOGICAL FRAMEWORKS,PEDAGOGICAL PRINCIPLES, ANDCOMPETENCIES UNDERLYING THEEDUCATIONAL ACTIVITYThe conceptual framework for SITDP can be described as one ofexperiential learning within the socio-cultural context (Yardley et al.,2012; Mukhalalati and Taylor, 2019). Specifically, we sought to createan experience in which students were explicitly given full leverage toorganize themselves as a scientific conference planning committeeand to collaboratively envision, plan, and implement all conferenceactivities within a given budget and time frame. The involved studentleadership team is offered guiding advice by a lead faculty coach onrequest, but ultimately students are given full latitude andresponsibility as decision-makers. This empowerment builds onother coaching strategies that have been designed to augmentmentoring for UR students by enhancing identity formation, self-efficacy, and the building of cultural capital (Williams et al., 2016).The socio-cultural context of the activity is both the graduate schoolwithin an academic medical center, as well as standard culturalnorms of academic conferences (e.g., various kinds of presentationsand networking activities).Frontiers in Education | www.frontiersin.orgDecember 2021 | Volume 6 | Article 7740702Boehmer et al.Motivating Self-EfficacyLEARNING ENVIRONMENT, LEARNINGOBJECTIVES, AND PEDAGOGICALFORMATThe SITDP Conference has convened every other year led by MayoClinic IMSD and PREP students in the summers of 2016, 2018, and2020. The 2016 and 2018 conferences occurred as traditional liveconferences involving travel, housing, and hospitality. The 2020conference was delivered virtually in the context of the COVID-19pandemic. While the 2020 conference could have been canceled, thestudent planning committee demonstrated significant resilience inchoosing to continue the conference and pivoting the fully in-persondesigntoavirtualformatusingthecommercialZoomteleconferencing platform with a lead time of only ∼4 months.Planning committee meetings were conducted in person during2016 and 2018. In 2020, these meetings began in-person and shiftedonline in March 2020. Conference planning commenced late fall ofthe years 2015, 2017, and 2019 respectively.Student organizational approach: Student leadership teams for theconferences are self-selected from participants in the PREP and IMSDprograms at Mayo Clinic. These programs are NIH-funded initiativesdesigned to increase UR student participation in the biomedicalsciences with PREP focusing on pre-professional training of post-baccalaureate students and IMSD focusing on students during theirfirst two years of graduate school training. Mayo Clinic’s joint PREP/IMSD program intentionally mingles both student groups with thegoal of mutual near-peer mentoring. Participants meet weekly to coverdiversity and resilience curriculum and to practice scientificpresentation and critique skills. During a regularly scheduledmeeting, IMSD and PREP faculty introduced the concept of theSITDP student-led conference, including proposed dates, actionsrequired for successful carry-out of the conference, and benefits ofparticipation to learners’ development and curriculum vitae. Studentsare asked to indicate their interest in participating in planningcommittee activities and those interested are invited to a separatemeeting led by a faculty coach to discuss timelines, responsibilities, anddirection more specifically. Topics covered in the initial two hourplanning include: planning tasks (e.g. venue selection, speakerinvitations), potential planning committee roles (e.g., chair/co-chair,budget director, sub-committee leads), selection of a conferencetheme, and a schedule for planning committee meetings. Studentcommittee members are provided the conference dates and thebudget, and otherwise given full control over other conferenceactivities. Students determine committee roles and sub-committeeorganization, and vote to establish their committee framework. Whilethe exact committee format was different from year to year basedupon student-organization, committees generally include two co-chairs that oversee planning sub-committees, which take onindividual tasks such as speaker invitations, venue and catering,social activities, travel, and budget. Committee scheduling isdictated by students but generally starts with weekly full committeemeetings in the early planning stages, followed by less frequent fullcommittee meetings as sub-committees are established and meet ontheir own schedule. As part of their responsibilities, planningcommitteesareexpectedtodocumenttheirprocessesandworkflowtodeveloptemplatesandbestpracticesforsubsequent teams.Faculty coaching: Faculty coaches and administrative staff offeringsupport and guidance are drawn from members of the PREP andIMSD leadership team, with the leadership team growing over the 6-year time frame to include more faculty from UR backgrounds toserve as faculty role models. Faculty coaches possess significantexperience in organizing scientific conferences and attend allcommittee meetings during early conference planning and thenapproximately monthly after planning roles and sub-committeesare established. Faculty coaches remain available by electroniccommunication and for ad-hoc meetings as needed. Attention isgiven to ensuring student empowerment, leadership, and negotiationso that faculty and administrators maintain only resource roles.Conference content: The conference theme and the agenda isorganized entirely by the student leadership, and as such differedbetween years. However, the conference maintains structuresfamiliar to scientific meetings, including plenary lectures andpanels given by invited speakers, oral and poster presentations bystudents attending the conference, workshop activities ondiversity and resilience topics, and social activities such as aconference dinners and networking lunches. Planning committeemembers are encouraged by faculty coaches to reflect on theirbest conferencing experiences and work towards designing aconference that aspires toward their unique vision of an idealscientific meeting. In-person conferences commenced lateafternoon/evening on a Friday and included a full agendamorning to evening on Saturday. For the virtual format, toprevent teleconference fatigue, the schedule was abbreviated tothe Friday evening and a half-day of sessions on Saturday.The specific learning objectives for SITDP conferenceactivities involving UR biomedical research students are:1. To foster a unique experience for UR students to build identity asleaders in a culturally relevant activity for their future as scientists.2. To increase the self-efficacy of UR students through self-directed problem-solving, negotiation, and planning tasksrequired to conduct a scientific conference3. To strengthen career-relevant social networking of URstudents both amongst themselves, with coaches on theleadershipcommittee,andwithstudentandfacultyconference participants from other institutions during bothplanning and conference phases.RESULTS TO DATE AND ASSESSMENTSIn all three instances the student-led SITDP conferences provedfeasibleandsuccessful.Evenunderthemostchallengingcircumstances in 2020, students were able to accomplish all threelearning objectives. The conferences attracted 65, 56, and 60attendees in 2016, 2018 and 2020, respectively. To date, studentsled the evaluation of the conference based upon attendee feedback.This included a different student-developed survey after each of theconferences. In 2016, all questions were formulated as open-endedquestions with qualitative responses from attendees. Promptsincluded items such as “Tell us what you thought about thestudent panel” and “what did you think about the diversitydiscussion?” In 2018, the post-course survey was shifted to aFrontiers in Education | www.frontiersin.orgDecember 2021 | Volume 6 | Article 7740703Boehmer et al.Motivating Self-Efficacyquantitative format, with only three qualitative questions to indicatethe most beneficial aspect of the conference, future sessionsuggestions, and additional comments. Quantitative questionsasked participants to rate their satisfaction with various aspects ofthe conference such as registration, conference content, and theconference’s usefulness to their current and future careers. The 2020survey similarly was mostly quantitative, asking attendees to expresstheir level of satisfaction with each of the sessions, as well as theopportunity to elaborate on opportunities for improvement in theconference in future years.Because of the differing questions and format across post-conference surveys, pooled analysis across years was not possible,and this presents an opportunity for future evaluation improvementof this program. Of what could be summarized through the analysisof the attendee surveys of content and format it is concluded thatmost conference attendees learned about the SITDP conference fromPREP and IMSD program directors of invited institutions. Feedbackwas strongly positive across all 3 years with approximately 90% ofattendees endorsing that the conference met their expectations for ascientific conference and that they would recommend the conferenceto colleagues.We recognize the need to formalize critical evaluation of this novelmethod for increasing student self-efficacy. Specifically, in futureconference years attendee survey questions must be standardized,using past student-created surveys as a guide, such that conferencescan be compared across years. Further, there is a need for evaluatingthe effect of the intervention on student leaders compared to those thatdo not participate in the SITDP conference process. Whilerandomizationisnotpossibleduetotheself-selectionofparticipants into the planning committee, administering a surveymeasure to all PREP/IMSD students both pre- and post-conference ispossible, and would offer insight into the unique effect of the SITDPconference leadership in isolation of participation in IMSD/PREPgenerally. We suggest a few key metrics that may be valuable bemeasured immediately pre- and post-conference. These includegeneralself-efficacy(Scholzetal.,2002),studentperceivedachievability and perceived desirability of an academic career asmeasured previously by Williams et al. (Williams et al., 2016),student sense of belonging within the academic community(Trujillo and Tanner, 2014), and science identity (Trujillo andTanner, 2014). Should programs wish to evaluate the skills of thefaculty coach, which is typically different from the students’ mainmentor(s), it is suggested to use a validated scale such as the MentoringCompetency Assessment (Fleming et al., 2013). Qualitative study ofstudent experience in the process through conducting individualinterviews and/or focus groups may also yield insights into theunique aspects of this intervention otherwise not captured throughsurveys. However, qualitative research is a time- and resource-intensive activity, so careful design of such study is required.Further, an in-depth qualitative evaluation of the program is likelyfeasible for a single conference, whereas quantitative evaluation isbetter suited for longitudinal study.Importantly, the Mayo Clinic SITDP conference developed anetwork of contributing allied programs that participated jointly.These allies included students and faculty from University ofChicago and Northwestern University. In fact, enthusiasm for thelearning objectives and outcomes was such that these institutionshosted a similar conference in 2019 and are considering andproposing comparable activities in complementary years. TheMayo Clinic IMSD/PREP plans to host a 2022 meeting andcontinues to seek expansion of the program through additionaloutreach.Theprogramremainscommittedtocontinuedcollaboration with other programs seeking to conduct similarinitiatives.DISCUSSION ON THE PRACTICALIMPLICATIONS, OBJECTIVES, ANDLESSONS LEARNEDIn the first three instances of the novel SITDP conference, we foundthat a completely student-led conference was both feasible andvaluable. In this sense, these conferences have served as pilots.Armed with these experiences and preliminary data, futureSITDP conferences hosted by the Mayo Clinic IMSD and PREPprograms will focus on coherent strategies for longitudinalevaluation of enrichment efficacy. Such evaluation should includecontrolled quantitative and qualitative assessment of impacts ofstudent planning committee participation, specifically focusing onadvanced leadership skill sets not otherwise gained throughtraditional mentored laboratory and classroom experiences. Assuch, our program recommends thoughtful collaboration withskilled education assessment professionals to develop instrumentsand an advanced plan for qualitative evaluation of studentexperiences and quantitative evaluation of student self-efficacy,perceived career cultural capital, and identity as scientists beforeand after the planning experience. It would be important to establishstandardizedevaluationapproachesthatcanbeappliedlongitudinally over multiple conferences across years, and withthe potential to review long-term impacts on biomedical sciencecareersofparticipants.Considerationshouldbegiventoidentification of control groups that do not experience the SITDPplanning enrichment to judge outcome impacts.CONCEPTUAL, METHODOLOGICAL,ENVIRONMENTAL, OR MATERIALCONSTRAINTSWe urge other programs seeking to enhance UR student learning topursue novel and experimental activities such as SITDP. It is worthnoting that the leadership team that proposed and provided coachingfor the student SITDP planning experience benefitted from manyyears of previous experience working with UR students, conferenceplanning, and material resources through NIH funding and generousinstitutional support of PREP and IMSD initiatives. Significantfunding is required to support in-person conference planning andexecution, especially when catering, student and speaker travel funds,and speaker fees are involved. Planning must include considerationof how many of the conference costs are to be covered fromregistration fees. The SITDP budgetary footprint was significantlysmaller for the virtual version of the conference. As is beingdiscovered for all modern science conferences, there are certainirreplaceable aspects of in-person conferences, such as impromptuFrontiers in Education | www.frontiersin.orgDecember 2021 | Volume 6 | Article 7740704Boehmer et al.Motivating Self-Efficacynetworking, but additional favorable cost factors must be consideredas mitigating. Virtual or hybrid models may also be more feasible fornew or young programs, as well as those with significant budgetaryconstraints for other reasons.In summary, the Mayo Clinic Graduate School of BiomedicalSciences scientific conference planning exercise as a leadershipenrichment for UR biomedical research students has been asuccess. When linked with an appropriate evaluation plan, thisactivity creates a unique and transferable template for developingself-efficacy.DATA AVAILABILITY STATEMENTThe raw data supporting the conclusions of this article will bemade available by the authors, without undue reservation.AUTHOR CONTRIBUTIONSKB is the first author of the manuscript. LL is the last, seniorauthor of the manuscript. All other authors contributed to theconduct of the SITDP conferences and preparation and approvalof the manuscript.FUNDINGThe SITDP conference was proposed as an aim of the NIGMSfunded award number R25GM 75148-15. Additional fundingfor conference expenses and personnel time were supported byMayo Clinic’s Office for Education Diversity, Equity, andInclusion, the Mayo Clinic Graduate School of BiomedicalSciences.REFERENCESBernard, C. (2018). Editorial: Gender Bias in Publishing: Double-Blind Reviewingas a Solution? eNeuro 5 (3). doi:10.1523/ENEURO.0225-18.2018Bowser, G., and Cid, C. R. (2021). Developing the Ecological Scientist MindsetAmong Underrepresented Students in Ecology fields. Ecol. Appl. 31 (6), e02348.doi:10.1002/eap.2348Budden, A. E., Tregenza, T., Aarssen, L. W., Koricheva, J., Leimu, R., and Lortie, C.J. (2008). Double-blind Review Favours Increased Representation of FemaleAuthors. Trends Ecol. Evol. 23 (1), 4–6. doi:10.1016/j.tree.2007.07.008Fleming, M., House, S., Hanson, V. S., Yu, L., Garbutt, J., McGee, R., et al. (2013).The Mentoring Competency Assessment: Validation of a New Instrument toEvaluate Skills of Research Mentors. Acad. Med. 88 (7), 1002–1008.doi:10.1097/ACM.0b013e318295e298Holliday, E. B., Jagsi, R., Wilson, L. D., Choi, M., Thomas, C. R., Jr., and Fuller, C. D.(2014). Gender Differences in Publication Productivity, Academic Position,Career Duration, and Funding Among U.S. Academic Radiation OncologyFaculty. Acad. Med. 89 (5), 767–773. doi:10.1097/ACM.0000000000000229John, A. M., Gupta, A. B., John, E. S., Lopez, S. A., and Lambert, W. C. (2016). AGender-Based Comparison of Promotion and Research Productivity inAcademic Dermatology. Dermatol. Online J. 22 (4). doi:10.5070/d3224030651Kim-Prieto, C., Copeland, H. L., Hopson, R., Simmons, T., and Leibowitz, M. J.(2013). The Role of Professional Identity in Graduate School Success for Under-represented Minority Students. Biochem. Mol. Biol. Educ. 41 (2), 70–75.doi:10.1002/bmb.20673Kirui, C., and McGee, J. (2021). Leveraging Resources for Educational Equity toPromote Academic Success Among Underrepresented Nursing Students: AnIntegrative Review. Nurs. Educ. Perspect. 42 (4), 212–215. doi:10.1097/01.NEP.0000000000000735Lopez, S. A., Svider, P. F., Misra, P., Bhagat, N., Langer, P. D., and Eloy, J. A. (2014).Gender Differences in Promotion and Scholarly Impact: An Analysis of 1460Academic Ophthalmologists. J. Surg. Educ. 71 (6), 851–859. doi:10.1016/j.jsurg.2014.03.015Martinez, L. R., Boucaud, D. W., Casadevall, A., and August, A. (2018). FactorsContributing to the Success of NIH-Designated Underrepresented Minoritiesin Academic and Nonacademic Research Positions. CBE Life Sci. Educ. 17 (2),ar32. doi:10.1187/cbe.16-09-0287McGee, R., Jr., Saran, S., and Krulwich, T. A. (2012). Diversity in the BiomedicalResearch Workforce: Developing Talent. Mt Sinai J. Med. 79 (3), 397–411.doi:10.1002/msj.21310Meyers, L. C., Brown, A. M., Moneta-Koehler, L., and Chalkley, R. (2018). Survey ofCheckpoints Along the Pathway to Diverse Biomedical Research Faculty. PLoSOne 13 (1), e0190606. doi:10.1371/journal.pone.0190606Mukhalalati, B. A., and Taylor, A. (2019). Adult Learning Theories in Context: AQuick Guide for Healthcare Professional Educators. J. Med. Educ. Curric Dev. 6,2382120519840332. doi:10.1177/2382120519840332National Institutes of Health (2020a). National Statistics on Graduate Students.Available at: https://report.nih.gov/nihdatabook/category/20 (Accessed August26, 2021).National Institutes of Health (2019). Notice of NIH’s Interest in Diversity.Availableat:https://grants.nih.gov/grants/guide/notice-files/NOT-OD-20-031.html (Accessed August 26, 2021).National Institutes of Health (2020d). Number of NIH Principal InvestigatorsFunded by Grant Mechanism and Ethnicity. Available at: https://report.nih.gov/nihdatabook/report/307 (Accessed August 26, 2021)National Institutes of Health (2020c). Number of NIH Principal InvestigatorsFunded by Grant Mechanism and Race. Available at: https://report.nih.gov/nihdatabook/category/25 (Accessed August 26, 2021).National Institutes of Health (2020e). Research Project Grants: Awards by Genderand Percentage to Women. Available at: https://report.nih.gov/nihdatabook/report/171 (Accessed August 26, 2021)National Institutes of Health (2020b). Trends in Race/Ethnicity of NIH-SupportedPh.D. Recipients. Available at: https://report.nih.gov/nihdatabook/category/19(Accessed August 26, 2021).Paulus, J. K., Switkowski, K. M., Allison, G. M., Connors, M., Buchsbaum, R. J.,Freund, K. M., et al. (2016). Where Is the Leak in the Pipeline? InvestigatingGender Differences in Academic Promotion at an Academic Medical centre.Perspect. Med. Educ. 5 (2), 125–128. doi:10.1007/s40037-016-0263-7Remich, R., Naffziger-Hirsch, M. E., Gazley, J. L., and McGee, R. (2016). ScientificGrowth and Identity Development During a Postbaccalaureate Program:Results from a Multisite Qualitative Study. CBE Life Sci. Educ. 15 (3).doi:10.1187/cbe.16-01-0035Schnoes, A. M., Caliendo, A., Morand, J., Dillinger, T., Naffziger-Hirsch, M.,Moses, B., et al. (2018). Internship Experiences Contribute to Confident CareerDecision Making for Doctoral Students in the Life Sciences. CBE Life Sci. Educ.17 (1), 1–14. doi:10.1187/cbe.17-08-0164Scholz, U., Gutiérrez Doña, B., Sud, S., and Schwarzer, R. (2002). Is General Self-Efficacy a Universal Construct?1. Eur. J. Psychol. Assess. 18 (3), 242–251.doi:10.1027//1015-5759.18.3.242Schrouff, J., Pischedda, D., Genon, S., Fryns, G., Pinho, A. L., Vassena, E., et al.(2019). Gender Bias in (Neuro)science: Facts, Consequences, and Solutions.Eur. J. Neurosci. 50 (7), 3094–3100. doi:10.1111/ejn.14397Trujillo, G., and Tanner, K. D. (2014). Considering the Role of Affect in Learning:Monitoring Students’ Self-Efficacy, Sense of Belonging, and Science Identity.CBE Life Sci. Educ. 13 (1), 6–15. doi:10.1187/cbe.13-12-0241Ureña, S., Ingram, L. A., Leith, K., Lohman, M. C., Resciniti, N., Rubin, L., et al.(2021). Mentorship and Training to Increase Diversity of Researchers andPractitioners in the Field of Aging and Alzheimer’s Disease: A Scoping Reviewof Program Characteristics. J. Aging Health 33 (1-2), 48–62. doi:10.1177/0898264320953345Williams, S. N., Thakore, B. K., and McGee, R. (2016). Coaching to AugmentMentoring to Achieve Faculty Diversity: A Randomized Controlled Trial. Acad.Med. 91 (8), 1128–1135. doi:10.1097/ACM.0000000000001026Frontiers in Education | www.frontiersin.orgDecember 2021 | Volume 6 | Article 7740705Boehmer et al.Motivating Self-EfficacyWilliams, S. N., Thakore, B. K., and McGee, R. (2017). Providing Social Support forUnderrepresented Racial and Ethnic Minority PhD Students in the BiomedicalSciences: A Career Coaching Model. CBE Life Sci. Educ. 16 (4), 1–12.doi:10.1187/cbe.17-01-0021Yardley, S., Teunissen, P. W., and Dornan, T. (2012). Experiential Learning: AMEEGuide No. 63. Med. Teach. 34 (2), e102–e115. doi:10.3109/0142159X.2012.650741Conflict of Interest: The handling editor declared a past co-authorship with oneof the authors JLL.The remaining authors declare that the research was conducted in the absence ofany commercial or financial relationships that could be construed as a potentialconflict of interest.Publisher’s Note: All claims expressed in this article are solely those of the authorsand do not necessarily represent those of their affiliated organizations, or those ofthe publisher, the editors and the reviewers. Any product that may be evaluated inthis article, or claim that may be made by its manufacturer, is not guaranteed orendorsed by the publisher.Copyright © 2021 Boehmer, Boschen De Souza, Doles, Lachman, Mays, Hedin,Dornink, Maher and Lujan. This is an open-access article distributed under theterms of the Creative Commons Attribution License (CC BY). The use, distributionor reproduction in other forums is permitted, provided the original author(s) and thecopyright owner(s) are credited and that the original publication in this journal iscited, in accordance with accepted academic practice. No use, distribution orreproduction is permitted which does not comply with these terms.Frontiers in Education | www.frontiersin.orgDecember 2021 | Volume 6 | Article 7740706Boehmer et al.Motivating Self-Efficacy | RESEARCH GATE |
DOI:10.5965/25944630622022e1806 VESTUÁRIO PEDAGÓGICO E INCLUSIVO COMO RECURSO DE APRENDIZAGEM INFANTILPedagogical and inclusive clothing as a child learning resourceLa indumentaria pedagógica e inclusiva como recurso de aprendizaje infantil Juliana Bononi1 Cássia Leticia Carrara Domiciano2Joedy Luciana Barros Marins Bamonte31 Doutoranda em Design de Produto na FAAC Universidade Estadual Paulista - UNESP. Lattes: http://lattes.cnpq.br/3722965298519701; ORCID: https://orcid.org/0000-0003-4386-9649; email: julianabs9@gmail.com2 Chefe do Departamento de Design Laboratório Inky Design Faculdade de Arquitetura, Ar-tes, Comunicação e Design - FAAC Universidade Estadual Paulista - UNESP. Lattes: http://lattes.cnpq.br/1534045638750216; ORCID: https://orcid.org/0000-0001-6497-2210, email: cassia.carrara@unesp.br3 Líder do grAVA - Grupo de Pesquisas Poéticas em Artes Visuais - CNPq / Leader of grAVA - Group of Poetic Research in Visual Arts – CNPq. Coordenadora e professora do Curso de Artes Visuais - bacharelado e licenciatura / Coordinator and Professor of the Visual Arts Course - Bachelor and Licentiate Degree. Depar-tamento de Artes e Representação Gráfica / Department of Arts and Graphic Representation, Faculdade de Arquitetura, Artes, Comunicação e Design / School of Architecture, Arts, Communication and Design Universi-dade Estadual Paulista / São Paulo State University. Lattes: http://lattes.cnpq.br/1534045638750216; ORCID: https://orcid.org/0000-0002-9519-624X; email: joedy.bamonte@unesp.br.REAMD, Florianópolis, v. 6, n. 2, e1806, p. 01-12, jun./ set. 20222RESUMOO presente trabalho, de revisão de literatura, possui como principal objetivo buscar cone-xões entre a inclusão escolar, os recursos didáticos de aprendizagem e o vestuário infantil, verificando se este último pode ser usado no desenvolvimento das relações, percepções e deduções que as crianças fazem durante a interação com os objetos, por meio da apren-dizagem criativa e significativa. Pode-se afirmar que o vestuário é um objeto pedagógico e lúdico, e seus elementos podem ser usados como recurso de aprendizagem pois auxiliam as crianças no desenvolvimento da percepção e interpretação por meio da exploração sen-sorial e interação social. Visto que o vestir é como uma construção pessoal, este auxilia na formação da autoimagem e oferece ao usuário infantil outro ponto de vista a respeito de si mesmo, fazendo-o sentir-se mais confiante, belo e até um “super-herói”. Contudo, é preciso capacitar professoras e professores para que estes profissionais estejam aptos a auxiliar as crianças neste processo.Palavras-chaves: Vestuário Infantil Pedagógico. Recurso Didático. Aprendizagem Signifi-cativa.AbstractThe present work, of literature review, has as main objective to seek connections between school inclusion, didactic learning resources and children’s clothing, verifying if the latter can be used in the development of relationships, perceptions and deductions that children make during interaction with objects, through creative and meaningful learning. It can be said that clothing is a pedagogical and playful object, and its elements can be used as a learning resource as they help children in the development of perception and interpretation through sensory exploration and social interaction. Since dressing is like a personal cons-truction, it helps in the formation of self-image and offers the child user another point of view about himself, making him feel more confident, beautiful and even a “superhero”. However, it is necessary to train teachers and teachers so that these professionals are able to help children in this process.Keywords: Educational Children’s Clothing; Didactic Resource; Meaningful Learning.ResumenEl presente trabajo, de revisión bibliográfica, tiene como principal objetivo buscar conexio-nes entre la inclusión escolar, los recursos didácticos de aprendizaje y la vestimenta infantil, verificando si esta última puede ser utilizada en el desarrollo de relaciones, percepciones y deducciones que los niños realizan durante la interacción con objetos, mediante un apren-dizaje creativo y significativo. Se puede decir que la ropa es un objeto pedagógico y lúdico, y sus elementos pueden ser utilizados como recurso de aprendizaje ya que ayudan a los niños en el desarrollo de la percepción e interpretación a través de la exploración sensorial y la interacción social. Dado que vestirse es como una construcción personal, ayuda en la formación de la imagen de sí mismo y ofrece al niño usuario otro punto de vista sobre sí mismo, haciéndolo sentir más seguro, hermoso e incluso un “superhéroe”. Sin embargo, es necesario capacitar a los docentes y maestras para que estos profesionales sean capaces de ayudar a los niños en este proceso.Contraseñas: Ropa Infantil Educativa; Recurso Didáctico. Aprendizaje significativo.REAMD, Florianópolis, v. 6, n. 2, e1806, p. 01-12, jun./ set. 20223Vestuário pedagógico e inclusivo como recurso de aprendizagem infantil1 INTRODUÇÃOAs crianças, nas brincadeiras de faz de conta, modificam espaços, criam instalações e intervenções, fazem sair sons de qualquer objeto e, principalmente, vivem experiências transformadoras através da imaginação. Por isso, os objetos e os recursos de aprendiza-gem infantil são considerados valiosas ferramentas pedagógicas, pois facilitam o processo de ensino-aprendizagem, além de torná-lo mais eficaz. Para atingir essa eficácia, Garofalo (2019) assegura que a aprendizagem deve ser tanto criativa quanto significativa.Vale ressaltar que a aprendizagem criativa é um método de ensino no qual os pro-fessores buscam desenvolver ligações pessoais e criar ambientes propícios que estimulem a criatividade, a imaginação e a colaboração, na intenção de despertar o interesse dos alunos. Já na aprendizagem significativa, o professor deve trabalhar como mediador e, por meio do conhecimento prévio dos alunos, desenvolver novos conhecimentos, ou seja, dar novos significados aos conhecimentos já existentes tornando a aprendizagem relevante e permanente. Sobre as roupas, para Bononi e Domiciano (2018), elas possuem forte apelo visual, porém uma peça de roupa pode trazer experiências sensoriais que estimulem outros sen-tidos, por meio do toque de determinados tecidos e texturas, como também pelo olfato, o que vem sendo estimulado por diversos produtos e marcas, como exemplo das sandálias Melissa, que são muito conhecidas por seu perfume, usado com a intenção de interagir e criar uma identidade com seus consumidores.Quando se fala em deficiência, ainda existe o estereótipo de crianças especiais, completamente dependentes dos pais ou que precisam de escolas especiais. Isso pode ser questionado diante do fato de todas as crianças, sem exceção, necessitarem de atenção. O que se deve ressaltar é que as crianças com deficiência precisam conviver com as demais, aprender a ter autonomia e a se socializar. Com isso, o principal objetivo desse trabalho de revisão de literatura é buscar co-nexões entre a inclusão escolar, recursos didáticos de aprendizagem e o vestuário lúdico e pedagógico, buscando respostas para uma questão: os mesmos podem ser usados no desenvolvimento das relações, percepções e deduções que as crianças fazem durante a interação com os objetos por meio da aprendizagem criativa e significativa? Para tanto, este artigo conceitua Objetos de Aprendizagem (OA) e os Recursos Edu-cacionais (RE), assim como, apresenta trabalhos relevantes sobre o vestuário infantil inclu-sivo e pedagógico que podem ser explorados pelos educadores em sala de aula.REAMD, Florianópolis, v. 6, n. 2, e1806, p. 01-12, jun./ set. 20224Vestuário pedagógico e inclusivo como recurso de aprendizagem infantil2 RECURSOS DIDÁTICOS Os recursos didáticos são todas as ferramentas indispensáveis no processo ensino--aprendizagem, e sua principal função é auxiliar os professores e facilitar a compreensão dos alunos sobre o assunto abordado, principalmente após a implantação dos recursos tecnológicos.Para Friso (2021), quando um recurso de aprendizagem é desenvolvido por um de-signer, esse objeto pode se relacionar de forma reciproca com o usuário, criando novas situações no ensino, auxiliar no processo ensino-aprendizagem e ser considerado como design educacional. Sem dúvida, o papel do Professor na escola é de extrema responsabi-lidade, pois o mesmo deve propor atividades que incentive o aluno a perceber o mundo, o espaço e seu próprio corpo. Como apoio aos Docentes, os livros podem ser considerados os mais antigos e importantes instrumentos didáticos e atualmente estão sendo lançados em versão digital, muitas vezes somente nesta, devido à importância da tecnologia nos dias atuais. Nesse sentido a educação, assim como os recursos, estão sendo modernizados. A definições encontradas para os OA são bastante amplas e abrangentes com relação à sua natureza e materialidade; no entanto a maioria dos textos encontrados voltam-se aos recursos digitais. Já os RE, caracterizam-se como objetos físicos, sendo classificados por Friso (2021, p. 31) como “Recursos Educacionais: livros, músicas, instrumentos musicais, vídeos, plantas ou qualquer outro meio de favorecimento da aprendizagem”. O conceito de OA é apresentado de várias formas, de acordo com cada autor, mas todos levam em consideração a capacidade de apoiar a educação, sua utilidade e impor-tância. Um Objeto de Aprendizagem é qualquer recurso, suplementar ao processo de aprendizagem, que pode ser reusado para apoiar a aprendizagem, termo geral-mente aplicado a materiais educacionais projetados e construídos em pequenos conjuntos visando a potencializar o processo de aprendizagem onde o recurso pode ser utilizado. (AGUIAR E FLORES, 2014, p.14). Ou seja, os OA podem ser criados em qualquer formato, seja slides, vídeos, simula-ções ou animações, desde que contenham linguagem apropriada, abordagem dos concei-tos e a confiança das informações. Além disso, softwares livres e gratuitos, também podem ser usados como recursos em áreas específicas. Devem ter o objetivo que pretendem al-cançar bem formulado, assim como a sua metodologia, pois esses são fatores-chave para levar o aluno ao desenvolvimento do pensamento crítico. REAMD, Florianópolis, v. 6, n. 2, e1806, p. 01-12, jun./ set. 20225Vestuário pedagógico e inclusivo como recurso de aprendizagem infantilVale ressaltar que, conforme assegura Garofalo (2019), a aprendizagem deve ser tanto criativa quanto significativa, portanto, o professor deve se especializar, pois essa ati-vidade exige grande experiência e responsabilidade. De acordo com a autora, a aprendi-zagem criativa considera o aprendizado mais efetivo quando há envolvimento do aluno, ou seja, quando esse estudo é significativo para ele, permitindo a construção a partir da experimentação concreta e ativa na prática, com erros e acertos.Portanto, os OA, assim como os RE, permitem que os estudantes explorem seus co-nhecimentos em várias situações. De acordo com Ausubel (1968, apud AGUIAR e FLÔRES, 2014), quando novas ideias e conceitos são apoiados por um processo de interação já exis-tente, ou seja, quando um objeto já é conhecido pelo aluno, o aprendizado se torna mais rico. A aprendizagem criativa, segundo Garofalo (2019), esta baseia-se em uma espiral que começa com a imaginação, passa pela criação, pelas brincadeiras, pelo compartilha-mento e reflexão. Para que a aprendizagem criativa seja eficaz, é preciso planejar e executar as ideias, estabelecendo um PROJETO que seja relevante e significativo, conectando ações e desejos, envolvendo PAIXÃO pelo que está sendo feito. O processo de construção torna-se rico e criativo à medida que o compartilhamos com mais PES-SOAS. O BRINCAR é essencial porque permite colocar em ação todas as etapas, dando corpo ao processo de experimentação – que pode ser revisto quantas vezes for necessário para estabelecer um processo de desenvolvimento contínuo. (GAROFALO, 2019). Contudo, a aprendizagem significativa é o processo mais importante na aprendi-zagem escolar e, por esse motivo, os requisitos básicos para que ela aconteça devem ser cumpridos: o conteúdo que deve ser organizado de forma lógica, estar relacionado com o conhecimento e possibilitar interação do aluno com o objeto. É normal que as crianças passem horas inventando coisas e, para isso, utilizem ma-teriais que tenham à disposição, como seus próprios brinquedos e até mesmo coisas que eventualmente seriam descartadas como lixo; essa imaginação deve ser estimulada. Nes-se sentido, o papel do Professor na escola é de extrema responsabilidade, pois o mesmo deve propor atividades que incentivem os alunos a perceberem o mundo, o espaço e seu próprio corpo, além de interagirem com seus colegas. Dessa forma, para Bononi e Domiciano (2018) o vestuário pode ser uma excelente ferramenta, pois é por meio da imaginação que as crianças entram no mundo da fantasia, por iniciativa própria, se permitindo viver qualquer coisa que queiram através de desenhos, estórias ou representações. Para tanto a sociabilização é de extrema importância, assim como o pertencimento a um grupo. Ainda assim, “[...] algumas crianças têm dificuldade em interagir, muitas vezes devido aos problemas de timidez ou de baixa autoestima” (BONONI, 2016, p. 67). REAMD, Florianópolis, v. 6, n. 2, e1806, p. 01-12, jun./ set. 20226Vestuário pedagógico e inclusivo como recurso de aprendizagem infantilDe acordo com Rebelo (2016), a escola e o sistema educacional buscam apoio para trabalhar as diferenças, sem tirar as crianças com deficiência do convívio social, na inten-ção de superar as situações de exclusão, reconhecendo os direitos da chamada diversida-de e estimulando sua participação social plena na sociedade. Quando o assunto é inclusão, infelizmente, ainda existem muitas barreiras no ensino: professores despreparados; invisibilidade na comunidade das crianças com deficiência que não frequentam a escola; falta de recursos financeiros na escola; acesso físico; dimensão das turmas; pobreza; discriminação por gênero e dependência (alto nível de dependência de algumas crianças com deficiência dos seus cuidadores).Há necessidade de muitas modificações, entre elas: qualificação do docente e das equipes de apoio especializado; estrutura curricular com métodos, técnicas e recursos edu-cativos; treinamento para lidar com a estruturação e organização curricular e com técnicas especializadas e instrumentalização das escolas para o uso de novos recursos educativos.3 RELAÇÕES ENTRE VESTUÁRIO INFANTIL, DESIGN E INCLUSÃO A chamada Escola Inclusiva vem a considerar as necessidades de todos os alunos, estruturando-se em função dessas necessidades, sendo que tal estrutura deve ser eficiente para atender a todos, nos seus diferentes níveis de ensino. Contudo, sob a perspectiva da inclusão, a escola tem que se preocupar não apenas com a convivência e com as trocas de experiências, mas também e, principalmente, com o aprendizado dos conteúdos neces-sários e valorizados pela sociedade atual, que tem acesso facilitado as informações, o que possibilita ter cada vez mais conhecimento e autonomia.Os objetos presentes em nosso cotidiano podem passar diferentes mensagens atra-vés de diferentes estímulos. Norman (2008) relata porque adoramos ou detestamos os objetos do dia-a-dia: temos uma relação emocional com os produtos. Quando são atrativos e funcionam melhor, produzem emoções positivas, fazendo com que o nosso processo mental seja mais criativo e ficamos mais benevolentes às imperfeições de uso. O vestuário é um dos produtos que mais mexe com nossas emoções, pois tem o poder de nos fazer sentir belos, autoconfiantes e inclusos. Esse objeto tão cotidiano, para Palma (2011), possui distintas funções e abordagens comunicacionais, de acordo com a cultura e o contexto em que é usado. Uma simples camiseta por exemplo, é uma peça básica e funcional, elemento de proteção corporal, produto histórico de uma cultura mas também um item altamente globalizado. Contudo, em nossa cultura a camiseta enquanto peça de vestuário, também é usada como uma ferramenta de comunicação visual podero-síssima, uma tela que através da aplicação de grafismos, elementos textuais, entre outros, REAMD, Florianópolis, v. 6, n. 2, e1806, p. 01-12, jun./ set. 20227Vestuário pedagógico e inclusivo como recurso de aprendizagem infantilpode conter todo tipo de mensagens associadas à identidade individual e/ou coletiva de uma sociedade, como explica Palma (2011).Quanto às roupas infantis, Pereira (2010), em sua pesquisa, assegura que as tex-turas variadas podem auxiliar no desenvolvimento da percepção e interpretação por meio da exploração sensorial, assim como a qualidade tátil também pode ser utilizada para fazer com que as crianças adquiram destreza nos movimentos e desperte a sua curiosidade. Ainda de acordo com a autora supracitada, o vestuário pode ser considerado lúdico e pedagógico, contribuir para o desenvolvimento da imaginação, além de ser uma rica fonte de estímulos para o sentido tátil, por meio de texturas introduzidas na modelagem através de pregas, recortes, dobras e bordados, assim como pela manipulação de diferentes teci-dos, das estampas em alto relevo e pingentes.Segundo Maximiliano e Tomasulo (2013), a moda proporciona experiências não ape-nas visuais, mas também táteis e sensoriais. É essencial manusear os tecidos e testar suas sensações, propriedades e como se comportam quando em contato com o corpo, pois o tato pode ser estimulado pelas diferentes texturas que os tecidos possuem e suas combinação diferentes. Assim o vestuário infantil, como recurso pedagógico, pode tornar a aprendizagem tanto criativa quanto significativa. Para a criança em desenvolvimento, de acordo com Bononi e Domiciano (2018) a relação entre o real e o imaginário é de grande importância, pois, quando veste a fantasia do homem aranha, incorpora o personagem se transformando em super-herói nas brinca-deiras com os amiguinhos, assim como, quando usa as roupas e sapatos de seus pais, estão comunicando sua admiração, o que querem ser quando crescer, ou seja ensaiando seu futuro. Para Munari (2007) nossa capacidade de fantasiar, depende das nossas vivências: quanto maior o nível cultural, provavelmente será maior a facilidade de fazer exercício mental e de relacionar distintos elementos, de diferentes áreas e culturas, o que conse-quentemente poderá levar um maior nível de fantasia. “O produto da fantasia, tal como a criatividade e da invenção, nasce de relações que o pensamento cria com o que conhe-ce...”. (MUNARI, 2007, p.31)Normalmente as crianças estão em constante atividade e raramente ficam quietas, vivem intensamente as brincadeiras, usam os sentidos, buscam conhecimento em tudo: cheiram, sentem, observam e às vezes até provam o material, pois elas possuem necessi-dade de movimento e contato físico.Por isso, para este público em questão, explorar o território de forma ampla e lúdica é uma necessidade, para expressar e exercitar todos os sentidos, sejam o olfato, a visão, o paladar ou o tato, pois utilizam seu corpo para explorar o mundo a sua volta de forma dinâmica e constante.REAMD, Florianópolis, v. 6, n. 2, e1806, p. 01-12, jun./ set. 20228Vestuário pedagógico e inclusivo como recurso de aprendizagem infantilAs crianças precisam de experiências enriquecedoras, que podem ser trabalhadas por meio da manipulação e experimentação de objetos reais, na intenção de conseguirem experimentar e/ou adquirir a maior consciência possível do mundo real. Neste cenário, deve ser possibilitado o desenvolvimento de atividades que promovam a exploração es-pontânea, tanto física como intelectual, não devendo ser minimizado o papel da mediação verbal na resolução de tarefas. Para Pereira (2011), o desenvolvimento do raciocínio pode ser estimulado no ves-tuário infantil por meio de imagens e textos inseridos nas peças, trazendo informações do interesse das crianças que podem ser sobre investigações, hipóteses, sequências, histó-rias entre outras.Em pesquisa sobre o lúdico no vestuário infantil, Pereira (2011) questionou profes-soras, psicólogas e pedagogas, quanto a percepção de interação com o vestuário, onde foi constatado que as crianças interagem com os personagens das roupas e se sentem confiantes ao mostrarem para as pessoas suas escolhas. Foi unânime a confirmação do interesse por elementos tais como: acessórios pendurados, imagens, estampas, texturas, botões, bordados, zíperes, personagens conhecidos e cores. Ou seja, as roupas foram apontadas como um grande recurso de interação e interes-se no universo infantil. As crianças apresentam sua relação com as vestimentas por meio da observação, das suas escolhas e mostrando seu traje como uma extensão de seu corpo. “Elementos inseridos nas peças, junto com imagens, estampas, texturas, botão e bordado, despertam o interesse das crianças e podem auxiliar no desenvolvimento de habilidades especificas e da criatividade” (PEREIRA, 2011, p. 121). Os participantes da pesquisa citada apontaram também algumas contribuições do vestuário como suporte para ensino infantil. De acordo com a pergunta “Em sua opinião quais as principais contribuições que a roupa pode dar para o desenvolvimento das crian-ças?”, pedagogas, psicólogas e professoras responderam: “postura, higiene e cuidado, percepções táteis e visuais, desenvolvimento da coordenação, imaginação, alfabetização, entre outros” (p.108).Ainda segundo Pereira, os participantes também indicaram que os materiais e tex-turas devem despertar o interesse e possibilitar sensações nas crianças, tais como; “liso, áspero, macio, duro, rugoso, emborrachados e matérias, como: velcro, zíper, babados. Desta forma, verifica-se que é necessário disponibilizar variedade de texturas e materiais, propiciando experiências diferenciadas às crianças” (2011, p. 123).Para Bononi (2016), a funcionalidade das roupas infantis é de grande importância, pois estas possuem aplicações, estampas e listras, que são desejadas pelo público (crian-ças e responsáveis), o qual tem sido exigente na busca de novas sensações e de criações bem elaboradas. Mas outro aspecto também deve ser levado em consideração: o estético, responsável pela satisfação emocional.REAMD, Florianópolis, v. 6, n. 2, e1806, p. 01-12, jun./ set. 20229Vestuário pedagógico e inclusivo como recurso de aprendizagem infantilEm pesquisa realizada com pais e educadoras de crianças com deficiência visual (DV), Bononi (2016) concluiu, por meio das respostas dos questionários, que os pais não sabem das potencialidades do vestuário, pois priorizam o conforto, os elementos estéticos e o preço no ato da compra. A presença de elementos táteis e a praticidade de uso acabam por ficar bem distantes da devida importância que deveriam ter para seus filhos. Mesmo afirmando que sabem e acreditam que o vestuário pode ser educativo e estimular a expe-riência tátil, não estimulam as próprias crianças quanto a isso. Quanto aos questionários aplicados por Bononi (2016) com os professores, a autora concluiu que; [...] a DV não é limitadora, principalmente quando há auxilio adequado para o apren-dizado e para a independência, em especial na infância, quando podemos educar para transformar adultos inseridos no mercado de trabalho e na sociedade de forma justa. Mas a principal constatação desse questionário foi que as próprias educado-ras não se dão conta do potencial da experiência tátil por meio do vestuário, por esse motivo nunca orientaram os pais a respeito. Portanto há um potencial inexplo-rado. (BONONI, 2016, p. 130).Ou seja, o vestuário pode contribuir para a formação e desenvolvimento infantil, pois está presente na maior parte do tempo como uma extensão do corpo, além de contribuir para a interação entre as crianças (BONONI, 2016).Vale ressaltar que o vestir é como uma construção, ao mesmo tempo que auxilia na formação da autoimagem, tem o poder de oferecer um outro ponto de vista a respeito de nós mesmos, portanto quando a criança recebe atenção devido a diferenciais estéticos pautados em sua realidade, essas se sentem especiais, e passam a se admirar, uma vez que a referência individual é construída a partir da interação com o ambiente.Especialista em design têxtil, Rüthschilling (2008), salienta que uma característica essencial do design é seu caráter social, uma vez que pode ser gerador de cultura, ferra-menta de educação e capacitação profissional, e não visto somente como uma solução estética em comunicação e produtos. Contudo, a moda é muito mais que um conjunto de roupas e acessórios, é um código visual que, entre outras coisas, expressa uma identidade, revela o estilo de vida e o status social, individualiza, agrupa e personifica as pessoas.Para Baptista (2006), quando há compreensão de que somos diferentes e que nos-sas diferenças que nos fazem humanos, estaremos prestes a resolver esse mistério da vida. “Há em cada um de nós qualidades, defeitos, potencialidades, surpresas que são infindáveis e imprevisíveis” (p. 13). E também:Quando nos deparamos com qualquer pessoa diferente de nós, sempre ocorre um sentimento ou sensação de estranheza. Isso ocorre por várias razões, mas a prin-cipal delas é que aquilo que difere de nós assusta, causa alarde. Em geral, esse susto fica mais destacado quando nos deparamos com alguém que tenha alguma deficiência. Mas o susto e o alarde diminuem, na medida em que passamos a con-REAMD, Florianópolis, v. 6, n. 2, e1806, p. 01-12, jun./ set. 202210Vestuário pedagógico e inclusivo como recurso de aprendizagem infantilviver com as pessoas e percebemos que todos têm habilidades e dificuldades, não importando aquilo que aparentamos (BAPTISTA, 2006, p. 13-14).Assim, o designer pode atuar no desenvolvimento de produtos de moda que con-tribuam para o bem-estar, inclusão e autonomia dos seus usuários, onde um universo de descobertas e estímulos pode ser explorado.4 CONSIDERAÇÕES FINAIS O principal objetivo desse trabalho foi buscar conexões entre os recursos didáticos de aprendizagem e o vestuário lúdico e pedagógico. Através das fontes pesquisadas, pode-mos concluir que, assim como acontece com a inclusão, há uma vasta lista de pesquisas, muitos recursos abordados, mas pouco investimento e treinamento. Fica então uma certeza constatada aqui, que os objetos de aprendizagem podem ser usados no desenvolvimento das relações, percepções e deduções, que as crianças fa-zem durante a interação com os mesmos por meio da aprendizagem criativa e significativa, o que leva o vestuário ser uma rica fonte de experiência.Podemos afirmar que as roupas possuem elementos que podem ser usados como recurso de aprendizagem, ou seja, como suporte para o ensino de postura, higiene, para o desenvolvimento de percepções táteis e visuais, da coordenação, da imaginação e da alfa-betização. Pois auxiliam no desenvolvimento da percepção e interpretação por meio da ex-ploração sensorial, da interação social, visto que o vestir é como uma construção individual, auxiliando na formação da autoimagem ajudando no sentimento de autoconfiança.Contudo, em pleno século XXI, ainda existem muitas dúvidas sobre como trabalhar a inclusão, muitas vezes os próprios professores não estão preparados para receber e conviver com crianças que apresentam qualquer tipo de deficiência. É preciso capacitar os docentes para que os mesmos sejam capazes de orientar os pais e auxiliar as crianças, pensando no futuro destas. REAMD, Florianópolis, v. 6, n. 2, e1806, p. 01-12, jun./ set. 202211Vestuário pedagógico e inclusivo como recurso de aprendizagem infantilREFERÊNCIAS AGUIAR, Eliane Vigneron Barreto; FLÔRES, Maria Lúcia Pozzatti. Objetos de Aprendi-zagem: conceitos básicos. In: TAROUCO, Liane Margarida Rockenbach; ÁVILA, Bárbara Gorziza; SANTOS, Edson Felix dos; BEZ, Marta Rosecler, COSTA, Valeria (orgs.). Obje-tos de Aprendizagem: teoria e prática. Porto Alegre: Evangraf/ CINTED/UFRGS, 2014. p. 12-28. Disponível em: https://lume.ufrgs.br/bitstream/handle/10183/102993/000937201.pdf?sequence=1&isAllowed=y. Acesso em: 23 set. 2021.BAPTISTA, Maria Isabel S. Dias. Convivendo com as diferenças. PUPO, Deise Tallarico; MELO, Amanda Meincke; FERRÉS, Sonia Pérez (orgs.). Acessibilidade: discurso e prá-tica no cotidiano das bibliotecas. Campinas: UNICAMP/Biblioteca Central Cesar Lattes, 2006. p. 13-16. Disponível em: http://www.santoandre.sp.gov.br/pesquisa/ebooks/363677.PDF#page=16. Acesso em: 28 dez. 2021. BONONI, Juliana. Design do Vestuário Infantil: As texturas como experiência tátil para crianças deficientes visuais. 2016. Dissertação (Mestrado em Design) - Faculdade de Ar-quitetura, Artes e Comunicação, Universidade Estadual Paulista “Júlio de Mesquita Filho”, Bauru, 2016. Disponível em: https://repositorio.unesp.br/bitstream/handle/11449/138569/bononi_j_me_bauru.pdf?sequence=3&isAllowed=y. Acesso em: 23 set. 2021.BONONI, J.; DOMICIANO, C. L. C. O uso do vestuário infantil como elemento de es-tímulo para o desenvolvimento tátil de crianças com deficiência visual: o papel dos pais e educadores. Revista de Ensino em Artes, Moda e Design, Florianópolis, v. 2, n. 2, p. 153 - 169, 2018. DOI: 10.5965/25944630222018153. Disponível em: https://www.revis-tas.udesc.br/index.php/ensinarmode/article/view/12467. Acesso em: 10 jan. 2022.FRISO, Valéria Ramos. Parâmetros de design para concepção de recursos pedagó-gicos auxiliares (RPAs) por professores das EMEIS de Bauru/SP. 2021. Dissertação (Mestrado em Design) - Faculdade de Arquitetura, Artes e Comunicação, Universidade Estadual Paulista “Júlio de Mesquita Filho”, Bauru, 2021. Disponível em: https://reposito-rio.unesp.br/bitstream/handle/11449/214304/friso_vr_dr_bauru.pdf?sequence=5&isAllowe-d=y. Acesso em: 10 jan. 2022.GAROFALO, Débora. Como a aprendizagem criativa pode alavancar o processo ensino e aprendizagem. Revista Nova Escola. 11 jun. 2019. Disponível em: https://novaescola.org.br/conteudo/17761/como-a-aprendizagem-criativa-pode-alavancar-o-processo-de-en-sino-e-aprendizagem. Acesso em: 28 out. 2021. REAMD, Florianópolis, v. 6, n. 2, e1806, p. 01-12, jun./ set. 202212Vestuário pedagógico e inclusivo como recurso de aprendizagem infantilMAXIMILIANO, C.; TOMASULO, S. B. O ensino de moda e a inclusão de deficientes visuais. E-Tech: Tecnologias para Competitividade Industrial, Florianópolis, v. 7, n. 2, p. 135-164, 2013.MUNARI, Bruno - Fantasia. Lisboa: Edições 70, 2007. 31-32, 37 p. ISBN 978-972-44-1357-0NORMAN, D. A. Design emocional: Por que adoramos (ou detestamos) os objetos do dia-a-dia. Rio de Janeiro: Rocco, 2008.PALMA, David Miguel Paixão. Vestir um Conto: Uma Experiência participativa entre o design de comunicação e o vestuário infantil. 2011. Dissertação (Mestrado em Design) - Departamento de Comunicação, Arquitectura, Artes e Tecnologias da Informação, Uni-versidade Lusófona de Humanidades e Tecnologias, Lisboa, 2011. Disponível em: https://recil.ensinolusofona.pt/bitstream/10437/5950/1/tese_david palma.pdf. Acesso em: 07 dez 2021. PEREIRA, Livia Marsari. Possibilidades de aprendizagem no vestuário infantil: um estudo exploratório. 2011. Dissertação (Mestrado em Design) - Faculdade de Arquitetura, Artes e Comunicação, Universidade Estadual Paulista “Júlio de Mesquita Filho”, Bauru, 2011. Disponível em: https://repositorio.unesp.br/bitstream/handle/11449/89732/pereira_lm_me_bauru.pdf?sequence=1&isAllowed=y. Acesso em: 23 set. 2021.RÜTHSCHILLING, Evelise Anicet. Design de superfície. Porto Alegre: UFRGS, 2008.Data de submissão: 10/03/2022Data de aceite: 27/05/2022Data de publicação: 01/06/2022 | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/348779361EVAPORATİF SOĞUTMA VE TERMOELEKTRİK SU SOĞUTUCU KOMBİNESİSTEM TASARIMIThesis · January 2021DOI: 10.13140/RG.2.2.23775.48802CITATIONS0READS3442 authors, including:Some of the authors of this publication are also working on these related projects:Evaporative air conditioning and Thermoelectric View projectOğulcan DemirPamukkale University1 PUBLICATION 0 CITATIONS SEE PROFILEAll content following this page was uploaded by Oğulcan Demir on 26 January 2021.The user has requested enhancement of the downloaded file.i PAMUKKALE ÜNİVERSİTESİ MÜHENDİSLİK FAKÜLTESİ MAKİNE MÜHENDİSLİĞİ BÖLÜMÜ LİSANS TEZİ I Oğulcan DEMİR / 16244032 Orhan Yiğit SEÇGİN /16244074 Ocak 2021 EVAPORATİF SOĞUTMA VE TERMOELEKTRİK SU SOĞUTUCU KOMBİNE SİSTEM TASARIMI Çalışmayı Yöneten: Prof. Dr. Mehmet Fevzi KÖSEOĞLU ii ÖNSÖZ / TEŞEKKÜR Bu çalışmanın hayata geçirilmesinde bizi doğru yönde yönlendiren ve ne kadar hata yaparsak yapalım bize olan inancını kaybetmeyen danışman hocamız Prof. Dr. Mehmet Fevzi KÖSEOĞLU’na, tez sürecinde bilgi ve tecrübelerinden faydalandığımız bizi sabırla dinleyen ve yardımlarını hiçbir zaman esirgemeyen Dr. Veli Doğan’a, manevi desteğinden dolayı Vemeks Mühendislik çalışanlarından Utku DELİOĞLU’na, sonsuz teşekkürlerimizi sunarız. Ayrıca çalışmamızda desteklerini hiçbir zaman eksik etmeyen, bize olan güvenlerini hiç kaybetmeyen ailelerimize teşekkür ederiz. OCAK 2021 Oğulcan DEMİR Orhan Yiğit SEÇGİN iii İÇİNDEKİLER ÖNSÖZ /TEŞEKKÜR ii İÇİNDEKİLER iii KISALTMALAR v TABLO LİSTESİ vi ŞEKİL LİSTESİ vii SEMBOL LİSTESİ xii ÖZET/ABSTRACT xiii 1.GİRİŞ 1 1.1. Literatür Özeti 1 2. EVAPORATİF SOĞUTMA 3 2.1. Çalışma Prensibi 5 2.2 Evaporatif Soğutma Teorisi 7 2.2.1 Psikrometri ve Psikrometri Tarihçesi 7 2.2.2 Psikrometri Özellikleri 7 2.3 Adyabatik Doyma İşleminin Deneysel Olarak İncelenmesi 13 2.3.1 Adyabatik Doyma İşlemi ve Girişte Havanın Özgül Neminin Saptanması 14 2.3.2 Adyabatik Doyma ve Yaş Termometre Sıcaklıkları 14 3.PSİKROMETRİK DİYAGRAM 15 3.1 Hava Karışımları 16 3.2 Duyulur Isıtma ve Soğutma 18 3.3 Nemli Havanın Neminin Sıkılması 22 3.4 Nemlendirme 24 4. TERMOELEKTRİK SOĞUTUCU 28 4.1 Termoelektrik Soğutma 29 iv 4.2 Termoelektrik Etkiler 31 4.2.1 Seebeck Etkisi 31 4.2.2 Peltier Etkisi 32 4.2.3 Thomson Etkisi 35 5. MATEMATİKSEL FORMÜLASYON 36 5.1 Evaporatif Soğutucu Tasarımında Kullanılacak Kabüller 36 5.2 Peltier Tasarımında Kullanılacak Kabüller 38 5.3 Sistem Dizaynı 39 5.3.1 Evaporatif Soğutucu Hesapları 40 5.3.1.1 Hava Debisi Hesabı 40 5.3.1.2 Hava Debisi ve Hava Basıç Kayıplarına Bağlı Olarak Fan Seçimi 40 5.3.1.3 Nemlendirme İşleminde Havaya Verilen Su Miktarı 41 5.3.1.4 Pompa Seçimi 43 5.3.2 Soğutucu Ped ve Cihaz Boyutlandırması 44 5.4 Peltier Su Soğutucu Dolap Sistem Dizaynı 49 5.4.1 Soğutulan Dolabın Malzeme Seçimi 49 5.4.2 Dolabın Tüm Yüzeylerinden Toplam Isı Transferi 49 5.4.3 Peltier Tipi ve Adet Seçimi 52 6. SONUÇ VE ÖNERİLER 54 7. KAYNAKLAR 55 8.EKLER v KISALTMALAR COP : Coefficent of Performance YT : Yaş Termometre sıcaklığı TCHU : Termoelektrik Soğutma-Isıtma Ünitesi DC : Doğru Akım TECT : Termoelektrik Soğutma Teknolojisi TE : Termoelektrik ekipman TEG : Termoelektrik Jenaratör KT : Kuru Termometre Sıcaklığı PPI : Birleşim Noktası vi TABLO LİSTESİ Tablo 4.1: Bazı maddelerin Seebeck katsayıları 32 vii ŞEKİL LİSTESİ Şekiller Şekil 2.1: Konfor Zonları 3 Şekil 2.2: Günümüzde Mobil Tip Evaporatif Soğutucu Cihaz 4 Şekil 2.3: Konya’da Bir Çiftlikte Evaporatif Soğutma Uygulaması 4 Şekil 2.4: Dış İskelet ve Su Haznesi İçi Açık Halde 6 Şekil 2.5: Delikli Kap ve Pompa 6 Şekil 2.6: Su Damlacıklarının Aşağı İnişi 6 Şekil 2.7: Açılan Panjur ve Fan 7 Şekil 2.8: Kuru-Yaş Termometre Sıcaklığı 8 Şekil 2.9: Çiğ Noktasına Erişen (2 noktası) Havadaki Su Buharının (T,s) Diyagramı 9 Şekil 2.10: Adyabatik Doyma İşlemi 13 Şekil 3.1: Psikrometrik Diyagram 13 Şekil 3.2: Psikrometrik Diyagram Üzerinde 1 ve 2. Konumundaki İki Nemli Hava Kütlesinin Karışım Noktası 15 Şekil 3.3: Örneğin DaikinPsikroemtrik Diyagramda Gösterimi 17 Şekil 3.4: Nemli Havanın Duyulur Isıtılması veya Soğutulması 18 Şekil 3.5: Örneğin DaikinPsikroemtrik Diyagramda Gösterimi 20 Şekil 3.6: Örneğin DaikinPsikroemtrik Diyagramda Gösterimi 21 Şekil 3.7: Havanın Sıkılması İşleminin Psikrometrik Diyagram Üzerinde Gösterimi 22 Şekil 3.8: Islak Hücre veya Yıkayıcılarda Havanın Nemlendirilmesi İşlemi 24 Şekil 3.9: Havanın su enjekte edilerek nemlendirilmesi 25 viii Şekil 3.10: Buhar Enjeksiyonu İle Hava Nemlendirmesi 26 Şekil 3.11: Örneğin DaikinPsikroemtrik Diyagramda Gösterimi 28 Şekil 4.1: Soğutma Çevrimi 29 Şekil 4.2: Seebeck Deneyi . 30 Şekil 4.4: Seebeck Voltajının Ölçümü 31 Şekil 4.5: Nem Alma Cihazında Peltier . 32 Şekil 4.6: Bilgisayar İşlemcisinde Peltier 33 Şekil 4.7: Peltier Görünümü 33 Şekil 4.8: TermoelementMödülünün İç Yapısı 34 Şekil 4.9: Termoelement bileşenleri 34 Şekil 4.10: P-tipi eleman çalışma şeması 35 Şekil 4.11: N-tipi eleman çalışma şeması 35 Şekil 4.12: Thomson Etkisi 36 Şekil 5.1: Soğutulacak Mahal ve Kesiti 37 Şekil 5.2: Psikrometrik Diyagram üzerinde üflenecek hava şartları 38 Şekil 5.3: Sistem Şeması 39 Şekil 5.4: 1 ve 2 nolu fan mahal içi gösterimi 40 Şekil 5.5: 1 numaralı fanın şematik gösterimi 40 Şekil. 5.6: 2 nolufanın şematik gösterimi 41 Şekil. 5.7: Psikrometrik Diyagramda Giriş ve çıkış şartları 42 ix Şekil 5.8: Malatya şartlarında ve Odaya Giriş Şartlarında Havanın Psikrometrik Özellikleri 42 Şekil 5.9: Cihaz Boyutu ve Su Tankı Boyutları 44 Şekil 5.10: Soğutucu Ped Modeli 44 Şekil 5.11: CELdek 7090-15 Soğutucu Pedin Oluk Açıları 45 Şekil 5.12: CELdek 7090-15 Soğutma Pedinin Havanın Hızlarına Oranla Doyma Verimi ve Basınç 45 Şekil 5.13: Soğutucu ped bilgileri 47 Şekil 5.14: Evaporatif Soğutucu Ped Boyutlandırması 48 Şekil 5.15: Dolap Yüzey Kesiti ve Numaralarıyla Yüzeyler 49 Şekil 5.16: Su Deposu Açık Hali 51 Şekil 5.17: Voltaj-Akım Grafiği 53 Şekil 5.18: Voltaj-Akım Grafiği 54 xi SEMBOL LİSTESİ ms Kuru havanın ağırlığı kg ma Su buharının ağırlığı kg x Özgül nem kg s.b./kg k.h. Va,Vs Su buharı hacmi m3 Rs Su buharının gaz sabiti J/kgK 𝑅𝑎 Kuru havanın gaz sabiti J/kgK Ps Su buharı basıncı Pa Pa Kuru havanın basıncı Pa 𝑇𝑠 Su buharının sıcaklığı °C 𝑇𝑎 Kuru havanın sıcaklığı °C 𝑉𝑠 Su buharının hacmi m3 𝑉𝑎 Kuru havanın hacmi m3 𝜙 Bağılnem % P Soğutma kapasitesi kW 𝜌 Yoğunluk kg/m3 L Soğutucu pedin genişliği m H Soğutucu pedin yüksekliği m Ps Hava içindeki kısmi buhar basıncı Pa Pd Doymuş hava içindeki kısmi buhar basıncı Pa h Entalpi kJ/kg Pv Akım enerjisi J U İç enerji J Cp Sabit basınçta özgül ısı J/kgK Qd Duyulur ısı kW Qg Gizli ısı kW Qt Toplam ısı kW Va Hava debisi m3/s v Özgül hacim m3/kg w Islak havanın özgül nemi kg s.b./kg k.h. gb Kuru havaya verilen buhar debisi kg/kg k.h. Tdb Kuru-Termometre Sıcaklığı °C Twb Islak (Yaş)-Termometre Sıcaklığı °C Tdp Çiğ Noktası (dewpoint) sıcaklığı °C TA A noktasının KT sıcaklığı °C TA’ A noktasının YT sıcaklığı °C h’A A noktasının entalpisi kJ/kg PA A noktasının kısmi buhar basıncı Pa φA A noktasının bağıl nemi % xii νA A noktasının özgül hacmi m3/kg TA” A noktasının çiğlenme sıcaklığı °C β Kontak faktörü - TD Giriş havasının çiğlenme noktası sıcaklığı °C α Seebeck Katsayısı µV/K 𝐺𝑠𝑢 Harcanan su miktarı kg/s ṁ ℎ𝑎𝑣𝑎 Havanın akış hızıdır kg/s V Hava akış hızı m/s 𝑘𝐴𝐵𝑆 ABS malzemesinin ısı iletim katsayısı W/m°C 𝐿𝑎𝑏𝑠 ABS malzemesinin kalınlığı m 𝑘𝑃𝑈 Poliüretan malzemesinin ısı iletim katsayısı W/m°C 𝐿𝑃𝑈 Poliüretan malzemesinin kalınlığı m ℎ𝑑𝚤ş Dış yüzeyin ortalama taşınım katsayısı W/m2°C 𝑅𝑥 Isı iletim direnci W/m2K 𝑄𝑥 Toplam ısı transferi W 𝑄𝑡𝑎ş𝚤𝑛𝚤𝑚 Taşınımla ısı transferi W 𝑄𝑖𝑙𝑒𝑡𝑖𝑚 İletimle ısı transferi W 𝐴𝑦ü𝑧𝑒𝑦 Isı transferine dik yüzey alanı 𝑚2 xiii ÖZET EVAPORATİF SOĞUTMA VE TERMOELEKTRİK SOĞUTUCU İLE SOĞUK SU ÜRETECİ KOMBİNE SİSTEM TASARIMI Birbirinden çok farklı yöntemle soğutma yapan iki farklı cihazın ve sistemin biraraya getirilerek enerji tasarrafu yapılıp yapılamayacağı veya bu iki cihazın bir sistemde birleştirilip, birleştirilemeyeceği analiz edilecektir. Bu her iki sistemin dayanmış olduğu termodinamik hesaplar ve fizik kanunları irdelenecektir. Evaporatif soğutucu ile soğutulması düşünülen bir oda içerisine konulacak evaporatif soğutma cihazı ile aynı oda içerisinde saatte belli miktarda soğuk su üretmesi düşünülen termoelektrik soğutucunun birlikte nasıl çalışabileceği araştırılacaktır. Anahtar Kelimeler: Evaporatif Soğutma, Termoelektrik Soğutma, Peltier, Soğuk Su Üreteci ABSTRACT This thesis studies whether two different devices and systems that are cooling with different methods can be combined to save energy in one system. Thermodynamic calculations and physics laws on which these two systems are based will be examined. The device that combines evaporative cooling and thermoelectric cooling methods will be placed in a room and how the system works will be analyzed. Key Words: Evaporative Cooling, Thermoelectric Cooling, Peltier, Cold Water Generator 1 1. GİRİŞ Özette belirtilen evaporatif mekân soğutma cihazı ile soğuk su üretecek termoelektrik soğutucunun beraber çalıştığı bir sistemi kurabilmek için öncelikle Bölüm 2’de evaporatif soğutma sistemi tanıtılacak, soğutmanın temel prensipleri anlatılacak, Bölüm 3’de havanın psikrometrik özelliklerine bağlı olarak evaporatif soğutma açıklanacaktır. Daha sonra Bölüm 4’de anlatılan termoelektrik soğutucunun, termoelektrik etkileri, soğutmanın temelini oluşturan fizik kanunları araştırılacaktır. Bu her iki araştırma ve tanıtımdan sonra birbirinden bağımsız olarak her iki sistem boyutlandırılacaktır. Boyutlandırılan sistemin birlikte çalışabilmesi için gerekli dizayn konseptleri ortaya konacaktır; Bu dizayn konseptlerine bağlı olarak her iki cihazın bağımsız devrelerindeki hava akışları ve su devreleri detaylı şekilde boyutlandırılacaktır. 1.1 Literatür Özeti: Evaporatif iklimlendirme, geçmişten günümüze pek çok alanda uygulamalarının varlığı bilinmektedir. Literatür araştırmasında da görülmüştürki evaporatif soğutma sistemi ile ilgili pek çok çalışma yapılmıştır. Yılmaz (1997) su veya buz depolama yöntemleriyle evaporatif soğutma sisteminin termodinamik analizini yapmıştır. Küçük sistemlerde (cami, düğün salonu, dans salonları, spor salonları, konferans salonları) genelde buzlu, büyük sistemlerde ise sulu sistemler tercih edilmesi gerektiğini vurgulamıştır. Bilge ve Bilge (1999) uygun dış hava koşullarında alternatif olarak önerilen evaporatif hava soğutma yöntemini “direk” ve “indirek” olmak üzere iki ana grupta incelemişlerdir. Bu iki yöntemi evaporatif soğutma sisteminde birlikte kullanmış ve klasik soğutma sistemi ile karşılaştırmışlardır. Klasik soğutma sistemi ile yapılan karşılaştırmada önerilen sistemin COP değerleri incelendiğinde, direk/indirek sistemin enerji tasarrufu açısından sağladığı olanakların kayda değer olduğuna varılmıştır. Sistem seçimi yapılırken gereken en önemli parametrenin dış hava YT sıcaklığı olduğu, bu değer yükseldikçe sistemin tüketeceği enerji miktarını arttırmaktadır. Bu nedenle çalışması yapılan soğutma sistemini kara iklimi tesirindeki bölgelerde kullanılması önerilmiştir. İdiz ve diğ. (2017) evaporatif soğutma sistemlerini kapsamlı tanımını, farklı çalışma sistemlerine göre termodinamik analizini yapmışlardır. Sistemleri karşılaştırmalı olarak incelemişlerdir. Direk Evaporatif Soğutma sistemi soğutma performansı açısından diğer sistemlere göre daha iyi sonuçlar verirken, sistemin çıkışında ortama verilecek havanın termodinamik özellikleri incelendiğinde kapalı ortamlarda gerekli konfor şartlarını sağlamadıklarını gözlemlemişlerdir. Endirek evaporatif soğutma sistemlerinin ise kapalı ortamları iklimlendirmek için daha iyi olduğunu gözlemlemişlerdir. Çelebi ve diğ. (2019) Evaporatif soğutucu sistem tasarımı yapmış ve kuru/yaş giriş termometre sıcaklıklarına göre performansını ölçmüştür. Üretilen sistemin verimi %45-%70 arasında değişiklik göstermiştir. Birbirine yakın kuru/yaş giriş sıcaklıklarında havanın debisi arttıkça sistemin 2 veriminin düştüğü gözlemlenmiştir. Tam tersi ise benzer debi değerlerinde giriş sıcaklıkları düştükçe sistemin verimi artmıştır. Termoelektrik modüllü soğutma uygulamaları ile ilgili ilk çalışma Rusya’da St. Petersburg’daki IoffeEnstitüsünde (1957), daha güncel ve yeni uygulamalar ise, Amerika Birleşik Devletlerinde ki Owens-Corning, Granville, OH (1997) ve Japonya’daki Mitsubishi elektrik tarafından yapılan çalışma sayılabilir (2020). Yılmazoğlu (2016) prototip termoelektrik ısıtma/soğutma ünitesinin performansını incelemiş, sayısal analizini yapmıştır. Sayısal sonuçları doğrulamak için deneysel analizler yapmış, hava hızının kanatçık yüzeylerinin sıcaklık dağılımı üzerindeki etkilerini ve termoelektrik soğutucu için voltaj farklarını ölçmüş. Uygun şekilde tasarlanmış ısı eşanjörleri ile havalandırma uygulamaları için termoelektrik soğutucuları alternatif bir yöntem olarak kullanmanın mümkün olduğunu gözler önüne sermiştir. Gang Tan (2014) termoelektrik malzemelerdeki son gelişmeleri, modelleme yaklaşımlarını ve uygulamalarını gözden geçirmiştir. Geleneksel soğutma cihazlarına göre avantajlarını karşılaştırmış. Termoelektrik modellemeyi iyileştirmek için modelleştirme tekniklerini karşılaştırmış, termoelektrik soğutma sistem tasarımı, simülasyonu ve analizi için faydalı olmasını beklemiştir. Wang ve diğ. (2020) Termoelektrik modülü, termal enerjiyi elektriğe dönüştürmek için bir jeneratör veya elektriği ısıya dönüştürmek için bir soğutucu olarak kullanılacağını belirtmiş, kapsamlı bir model önermiştir. Önerilen modelde, jeneratör ve soğutucu üzerindeki yükseklik, kesit alanı, seramik kaplama, çift sayısı ve ısı kayıplarının etkilerini araştırmıştır. Sonucunda: Düşük bacak yüksekliği, çıktı gücünü, soğutma gücünü ve COP değerini yükseltmiştir. Bu çalışmanın bulguları, hem enerji üretimi hem de soğutma için kullanılan termoelektrik modülün tasarımına ve parametrik optimizasyonuna rehberlik edebilir.Seramik plakaların optimizasyonu için, ısıl iletkenliğin iyileştirilmesine daha fazla dikkat edilmelidir.Sacristán-Fernández ve diğ. (2017) Termoelektrik Soğutma-Isıtma Ünitesi (TCHU) binalar için yenilikçi bir iklimlendirme sistemi olarak belirtmişlerdir. TCHU'da, bir doğru akım (DC) elektrik akımları, ısıyı akım akışına bağlı olarak bir yönde veya başka bir yönde aktarabilen bir Termoelektrik Ekipman (TE) ısı pompası sistemine güç sağlayacağını söylemişlerdir. Bu bağlamda, konut binaları için bir ısıtma ve soğutma sistemi olarak potansiyel uygulamaları için 16 TE modülünün dikey konfigürasyonu incelemişlerdir. Ayrıca hem soğutma hem de ısıtmada ünitenin en iyi Performans Katsayısını (COP) elde etmek için en uygun konfigürasyonu bulmak amacıyla bazı teorik analizler geliştirmişlerdir. Caia ve diğ. (2018) Bu makale, elektronik soğutmadaki termoelektrik uygulamalara ilişkin eksiksiz bir temel göstermek için kapsamlı bir termoelektrik soğutma teknolojisi (TECT) araştırması sunmuştur.Elektronik cihazların yüzey sıcaklıkları ve değerlendirme indeksleri olarak COP dikkate alınarak, soğutma yükü, hava sıcaklığı ve ısıtma tarafındaki tüm ısıl iletkenliklerin soğutma performansı üzerindeki etkileri denenmiştir. Yüksek ısı akısının elektronikleri için termal kontrolün, termoelektrik sistemin sıcak tarafındaki ısı transferini artırarak ve termoelektrik soğutucuların sayısını artırarak elde edilebileceğini ortaya koymaktadır.Zhou ve diğ. (2018)Bu makalede, yeni bir termoelektrik destekli dolaylı buharlaştırmalı soğutma sistemi önermektedirler.Termoelektrik soğutma (TEC) modülleri düz bir plaka çapraz akışının kanalları arasına sıkıştırılmıştır.Yeni sistemin matematiksel bir modeli geliştirilmiş ve ana işletim, geometrik parametrelerin sistemin performansına etkileri detaylı olarak incelemişlerdir.Önerilen sistemin çiğlenme noktası verimliliği ve COP'si, sırasıyla birincil hava, sıcaklık ve nemin giriş kütle akış hızı ile monoton bir 3 şekilde değiştiği görülmektedir.Analitik sonuçlardan, önerilen sistem birincil havayı YT sıcaklığından çok daha düşük bir sıcaklığa, hatta TEC modüllerinin sayısını ve çalışan elektrik akımını uygun şekilde seçerek çiğ noktası sıcaklığına kadar soğutulabileceği, TEC modüllerinin n sayısı ve elektrik akımı I'in artmasıyla, çiğ noktası etkinliği εdp artar ve COP düşeceği,Son olarak, maksimum COP'lerle sonuçlanan birincil hava kanalı ve ikincil hava kanalının optimum genişlikleri her zaman elde edilebilir. Maksimum COP ve optimum kanal genişliklerinin değerleri, TEC modüllerinin sayısı ve elektrik akımının artmasıyla azalır. Birincil hava kanalının optimum değeri, ikincil hava kanalınınkinden daha yüksek olduğunu kapsamlı bir şekilde araştırıp, sunmuşlardır. Zheng ve diğ. (2018) Direkt evaporatif soğutma teknolojisi ile soğutulan termoelektrik jeneratörün (TEG) teorik modeli geliştirmişlerdir. Teorik modelde, sıcaklık değişimine bağlı termoelektrik malzeme özellikleri ve ısı kaybının radyasyon, iletim ve Thomson fenomeni ile etkileri ele almışlardır. Evaporatif soğutmanın TEG verimliliğine etkisini araştırmak için simülasyon koşulları ayarlamışlardır. 2. EVAPORATİF SOĞUTMA: Doğada en yaygın ve en çok bilinen soğutma sistemi olan evaporatif soğutma, yaz aylarında havanın içerisindeki bağıl nemin ve YT sıcaklığının düşük olduğu bölgelerde kullanılır. Eğer yaz aylarında evaporatif soğutma yapılmak istenen bölgede bağıl nem ve YT sıcaklığı yüksek ise konfor için evaporatif soğutma yapmak doğru değildir. 15,51924,517%35%60%2026,5%70Baðýl NemOda Sýcaklýðý, °C22,5KonforluKýsmen KonforluKonforsuz Şekil 2.1 Konfor Zonları (R.H.-T. grafiği)(Doğan,2010) Ancak başka amaçlı evaporatif soğutma kısmen mümkündür. Menemen’de testi içerisine konulan su, özel olarak yapılmış testinin yüzeyinde ki kılcal deliklerden sızarak testi yüzeyinde buharlaşmakta ve buharlaşma esnasında testiden ısı çekerek testi içerisinde ki suyun kısmen soğumasın sağlamaktadır. Aynı şekilde karpuz üzerine ıslatılmış bir bez sarılarak güneşe konulması, sıcak ortam havasını zemindeki toprağa su tutmak suretiyle toprağın serinletilmesi, Anadolu’da yaygın olarak kullanılan soğutma yöntemleridir. İnsanların haricinde hayvanlarında konfor zonları içinde yer alması için ahır kümes hayvanları yetiştirilen büyük çiftliklerde, büyükbaş hayvan yetiştirilen ahırlarda, süt ineği sağım merkezlerinde de evaportif soğutma kullanılmaktadır. 4 Şekil 2.2 Günümüzde Mobil Tip Evaporatif Soğutucu Cihaz Şekil 2.3 Konya’da Bir Çiftlikte Evaporatif Soğutma Uygulaması Evaporatif Soğutucunun kısaca avantajları ve dezavantajlarından bahsedecek olursak: Avantajları: 5 1. Evaporatif soğutmada, soğutma kompresörü, kondenser, evaporatör vb. diğer soğutma çevrimi için gerekli olan bakır boru ve soğutma elemanları gerekmediği için maliyeti çok düşüktür. 2. İlk yatırım maliyetinin ucuz olmasının yanında diğer bir avantajı ise enerji tüketimi normal soğutma çevrimi içeren cihazlara göre kıyaslanamayacak kadar azdır. Küçük bir pompa ve fana ait enerji gideri söz konusudur. 3. Fan ve pomapnın toplam enerji sarfiyatı küçür bir odaya yerleştirelecek cihaz için 100 Watt’ın altındadır. Soğutma çevrimi içeren bir cihaz ise aynı odayı 1500 Watt civarında bir enerji tüketimiyle soğutur. 4. Bakım kolaylığı açısından küçük bir pompa ve fanın bakımı çok kolaydır. Soğutma çevrimi içeren bir cihazda gaz kaçakları, genleşme valfinde bozulmalar, kompresör arızaları, evaporatör ve kondenser fanlarının arızaları gibi. 5. Çevreci bir soğutucudur. Diğer cihazlar gibi herhangi bir soğutucu gaz ile çalışmaz. 6. Evaporatif soğutucularda soğutma çevrimi bir filtre gibi çalışır ve oluşturduğu nem sayesinde havanın içindeki kirliliği yoktur. 7. Evaporatif soğutma kurak iklimlerde havanın nemini arttırdığından solunan havayı yumuşatır. (Çalışır ve diğ. 2011) Dezavantajları: 1. En büyük dezavantajı ilerleyen bölümlerde bahsedileceği gibi komfor kliması için yaz aylarında bağıl nemin yüksek olduğu yerlerde kullanılması uygun değildir. Bağıl nemin yüksek olması komfor alanından uzaklaştırır. (Şekil 2.1) 2. Devamlı bir su kaynağına ihtiyaç duyar. 3. Bu soğutmadan elde edilen hava ortalama %80-90 oranında nem içerir. 4. Lejyoner hastalığı için en büyük riski oluşturur. Bu riski azaltmak için cihazların içerisindeki su haznelerien ultraviyole cihazlar yerleştirilmekte veya haznede bulnan su çok sık temiz suyla yenilenmelidir. 5. Ayrıca insan vücudundan ciğerler, deri, burun ve göz tarafından vücuttan uzaklaştırılacak nemin de dışarı atılamamasına neden olur. Bu da astım hastalarında olumsuz sonuçlar doğurur. Bu nedenle nem dengesinin kontrol edilmesi zorunludur. (Çalışır ve diğ. 2011) 2.1 Çalışma Prensibi: Şekil 2.3’de görüldüğü gibi profilden yapılmış bir kutunun tabanında, yukarıdan damlayarak gelen suları toplayacak bir su tankı konulmuştur. 6 Şekil 2.4 Dış İskelet ve Su Haznesi İçi Açık Halde Şekil 2.4’de kapta biriken su, küçük bir pompa vasıtasıyla bu profilden yapılmış cihazın üst kısmına aktarılmaktadır. Bu kabın tabanında küçük delikler açılmış olup, pompadan gelen suyu homojen bir şekilde aşağıya doğru damlatmaktadır. Şekil 2.5 Delikli Kap ve Pompa Şekil 2.5’te görüldüğü gibi yukarıdaki delikli kaptan aşağı doğru süzülen su damlacıkları bu soğutma cihazının içine yerleştirilen kamış dilimleri, plastik, kanatlı çubuk, kavak odun lifi vb. türevlerden yapılmış(su ve havanın daha çok temasını sağlayan) malzemelerin üzerinden geçerek alttaki kapta birikmektedir. Şekil 2.6 Su Damlacıklarının Aşağı İnişi Daha sonra bu kutunun karşılıklı iki tarafı kapatılmaktadır. Diğer iki yüzeyden birine dışarı su kaçırmayacak şekilde panjur yerleştirilmekte, tam karşısına bir fan takılmaktadır (Şekil 2.6.). Fanın emmiş olduğu hava oda içerisinden soğutma cihazına girerek su damlacıkları ile karşılaşmaktadır. Bağıl nemi düşük olan hava, su zerreciklerini buharlaştırarak, bağıl nem kazanmakta, duyulur ısısı düşmekte ve haliyle mutlak nemi artmaktadır. 7 Şekil 2.7 Açılan Panjur ve Fan 2.2 Evaporatif Soğutma Teorisi: Diyarbakır’da evaporatif soğutucu cihaz tasarlayan ve düşük maliyetinden ötürü ciddi paralar kazanan usta işini büyütmek, yeni bir tesis açmak amacıyla Antalya’ya gelir. Diyarbakır’da yaptığı makinenin aynısını sanayide anlaştığı ustalara yaptırır. İlk başta satışlar güzel gitse de aldığı dönütler hiç olumlu olmaz. Sattığı makinenin soğutma yapmadığını savunan müşteriler, Diyarbakırlı ustayı mahkemeye verir. Usta, yaptığı işte neyin yanlış olduğunu anlayamaz. Evaporatif soğutmanın teorisinin anlaşılabilmesi için soğutulacak havanın özelliklerinin iyi bilinmesi gerekir. Evaporatif soğutma işleminin anlaşılabilmesi için havanın psikrometrik özelliklerini incelemek gerekir. 2.2.1 Psikrometri ve Psikrometri Tarihçesi: Yunanca iki kelimenin, soğuk ve soğutma anlamına gelen “psuchron” ile ölçme anlamına gelen “meter” birleşiminden oluşmaktadır. Bu kelimeyi ilk kullanan Ernest Ferdinand August’tur. Kendi geliştirdiği ıslak termometreye “Psychrometer” adını vermiş, günümüzde ise psikrometreler kuru termometre ve YT’leri ayrı ayrı ölçmeye yarayan iki termometreden meydana gelmektedir. (Sevel, 2014) Havanın içerdiği nemi ölçme sanatı psikrometri olarak adlandırılır. Psikrometri, nemli havanın Termodinamik özelliklerini, havanın içerdiği nemin atmosferik koşullarda çeşitli maddeler üzerinde etkisini inceleyen ve konfor koşullarını sağlanmasına yarayan bir bilim dalıdır. (Demirci, 2004) Psikrometriyi net anlamak için temel gaz yasaları ve gaz karışımları konusunda bir temel bilgi gerekir. Bu bağlamda, bazı temel kavramlar ve tanımlamalar üzerinde durmak gereksinimi vardır. 2.2.2 Psikrometrik Özellikler: Kuru-Termometre Sıcaklığı, Tdb (°C KT): Herhangi bir alkollü ve civalı termometre ile ölçülen mahal veya çevre sıcaklığıdır. Termodinamik koordinatların arasında ki ilişki saptanmasında KT sıcaklığı kullanılır. 8 İleride de bahsedeceğimiz Psikrometrik Diyagramın altında yer alan KT, düşey hatlar ile gösterilir. 0 °C kuru hava sıcaklığı aynı zamanda entalpik değerlerin başlangıç, yani referans noktası olarak kabul edilir. Islak (Yaş)-Termometre Sıcaklığı, Twb (°C YT): Termometrenin ucu (haznesi) ıslak bir pamukla sarılıp ve nemli hava içinde hızla sallanarak sıcaklığı ölçülürse, termometrenin gösterdiği sıcaklığın düştüğü gözlenir (Şekil 2.7.). Düşüş sona erdiği zaman oluşan sıcaklık YT’dir. Psikrometrik Diyagramda birbirine paralel eğimli düz çizgiler olarak gösterilir. Bu çizgiler sabit entalpi çizgilerine de paraleldir. Kuru ve yaş termometre sıcaklıkları arasındaki fark Yaş Termometre Depressionu olarak adlandırılır. Şekil 2.8 Kuru-Yaş Termometre Sıcaklığı Çiğ Noktası (dewpoint) sıcaklığı Tdp(°C): Genel anlamda, bir gaz-buhar karışımının çiğ noktası, karışım sabit basınçta soğutulduğu zaman buharın yoğuşmaya ya da donmaya (katılaşmaya) başladığı sıcaklık değeridir. Benzeri bir tarif ise çiğ noktasını havanın doyma noktasına eriştiği, daha fazla buhar kabul edemeyeceği sıcaklık olarak da tanımlayabiliriz. Dolayısıyla doyma sıcaklığı ile çiğ noktası sıcaklığı birbirine eşit anlamdaki tabirlerdir. Havanın çiğlenme noktasına erişme süresinde, içerdiği su buharının Termodinamik işlemi aşağıdaki Sıcaklık-Entropi (T,s) diyagramında (Şekil 2.8) gösterilmiştir. 9 , Şekil 2.9 Çiğ Noktasına Erişen (2 noktası) Havadaki Su Buharının (T,s) Diyagramı Karışımın T sıcaklığını ve karışım içindeki buharın (başlangıçta 1 durumunda kızgın buhar olduğunu varsayarak) Ps kısmi basıncını göz önüne alalım; 1 durumunda karışım sabit basınçta soğutulursa, buharın kısmi basıncı 2 noktasına gelinceye değin sabit kalacaktır ve 2 noktasına gelince de yoğuşma başlayacaktır; karışımın sıcaklığı da T1 değerinden T2 değerine düşecektir. Yoğuşmanın başladığı 2 durumunda sıcaklık çiğ noktası sıcaklığıdır. Buhar, doyma basıncı ve sıcaklığında ise karışım bir doymuş karışım olarak göz önüne alınır, bu durumda ki hava+su buharı karışımı da doymuş hava olarak adlandırılır. 1 durumunda ve Ps kısmi basıncında kızgın buhar içeren karışım, sabit sıcaklıkta 4 noktasında yoğuşmaya başlar ve basınç p gibi bir değer alır. Doymuş karışımın sıcaklığı T ve basıncı p’dir. Çiğ noktası sıcaklığı buharı çizelgelerinden ya da su buharının MOLLIER (h,s) diyagramından, gerçek “kısmi basınca uyan doyma sıcaklığı” olarak anılır. 4 durum noktasındaki bir doymuş karışımın içerdiği buhar, doyma basıncında ve sıcaklığındadır. Bu durumdaki hava da doymuş havadır. P4 basıncı, T1 sıcaklığında ve ps kısmi basıncında su buharının doyma basıncıdır. Çiğ noktasında, Yaş ve Kuru Termometre sıcaklıkları birbirine eşit ve Ts=T2=Tdp=T(KT)=T(YT)’dir. Kuru termometre ve çiğ noktası sıcaklıkları arasındaki fark Çiğ Noktası Depressionu olarak adlandırılır. 10 Özgül Nem (nem oranı), x (kg su buharı/kg kuru hava): Pskrometride, birim ağırlıktaki kuru havanın içinde bulunan su buharı miktarına özgül nem adı verilir. Özgül neme bazı literatürde nem miktarı veya nemlik oranı da denmektedir. Tanım gereği özgül nem: 𝑥 = 𝑚𝑠𝑚𝑎 (𝑘𝑔𝑠. 𝑏./𝑘𝑔𝑘. ℎ. ) (2.1) olarak tanımlanır. Şimdi genel gaz kanunu ayrı ayrı su buharı ve kuru havaya uygulayalım. Kuru hava: 𝑃𝑎. 𝑉𝑎 = 𝑚𝑎. 𝑅𝑎. 𝑇𝑎 (2.2) 𝑚𝑎 = 𝑃𝑎𝑉𝑎𝑅𝑎𝑇𝑎 (2.3) Su buharı: 𝑃𝑠. 𝑉𝑠 = 𝑚𝑠. 𝑅𝑠. 𝑇𝑠 (2.4) 𝑚𝑠 = 𝑃𝑠𝑉𝑠𝑅𝑠𝑇𝑠 (2.5) Su buharı ve kuru havanın aynı hacimde eşdeğer sıcaklıkta olduğunu hatırlayarak (2.3) denklemini (2.4) denklemine taraf tarafa bölersek: 𝑥 = 𝑚𝑠𝑚𝑎= 𝑅𝑎𝑃𝑠𝑅𝑠𝑃𝑎 (2.6) olur. Öte yandan, 𝑅𝑎𝑅𝑠 =𝑚𝑠𝑚𝑎 = 18,0228,97 = 0,622 (2.7) olduğu hatırlanır ve 2.6 denkleminde yerine yazılırsa, 𝑥 = 𝑚𝑠𝑚𝑎= 0,622 𝑃𝑠𝑃𝑎 (2.8) veya 𝑥 = 0,622𝑃𝑠(𝑃𝑎−𝑃𝑠) . (𝑘𝑔𝑠.𝑏./𝑘𝑔𝑘.ℎ.) (2.9) olarak bulunur. Bağıl nem (relativehumidity), ϕ (%): Bağıl nem “bir karışımda buharın mol sayısının (aynı sıcaklıkta ve toplam basınçta) doymuş bir karışımdaki buharın mol sayısına oranı” olarak tanımlanır. 11 Diğer bir deyişle verilen (t) kuru termometre sıcaklığındaki nemli havanın içindeki kısmi buhar basıncının aynı (t) sıcaklığında doymuş havanın içindeki kısmi buhar basıncına oranına bağıl nem adı verilir. ϕ =𝑃𝑠𝑃𝑑 (% relative humidity) (2.10) Burada: Ps: (t) sıcaklığındaki havanın içindeki kısmi buhar basıncı Pd: (t) sıcaklığındaki doymuş havanın içinde ki kısmi buhar basıncını göstermektedir. Entalpi (ısı tutumu), h (J): Bir sürekli akım işleminde, “akım enerjisi (pV) ile için enerjinin (U) toplamını” gösteren bir bileşim enerjisidir ve bu toplam entalpi niceliği: 𝐻 = 𝑈 + 𝑝𝑉 (2.11) Bağıntısıyla tanımlanır. Entalpi diğer adıyla ısı tutumu olarak da adlandırılır. Özgül Entalpi ise m=1 (kg) kütle için tanımlanır. Aynı şekilde, özgül iç enerji (u) ve özgül akım enerjisi (pv) de (1 kg k.h.) kütlesi bazında tanımlanırsa, bu özellikler için aşağıdaki tanım eşitlikleri verilebilir: Özgül entalpi: ℎ = 𝐻𝑚𝑎 (2.12) Özgül iç enerji: 𝑢 = 𝑈𝑚𝑎 (2.13) Özgül akım enerjisi: 𝑝𝑣 = 𝑝( 𝑉𝑚𝑎) (2.14) olup bileşim enerjisi olarak (özgül) entalpi bağıntısı da, 12 ℎ = 𝑢 + 𝑝𝑣 (2.15) olacaktır. Özetle; entalpi, baz alınan bir noktanın üstünde havadaki ısı miktarını gösteren bir ısıl (termal) özelliktir. Nemli Özgül Isı, Cp: 1 kg kuru hava ile su buharı karışımının özgül ısısıdır. Su buharının nemli hava içindeki kütlesi: ms, kuru hava kütlesi: ma ve özgül ısıları sırasıyla Cp)s, Cp)a ise; Enerji korunumu ilkesini karışım için uygulayarak, T sıcaklığındaki karışım ve bileşenleri için entalpidengesi denklemlerini yazarsak; Toplam entalpi değerleri: Karışımın: 𝐻 = 𝐻𝑎 + 𝐻𝑠 (2.16) Kuru havanın: 𝐻𝑎 = 𝑚𝑎 + 𝐶𝑝)𝑎 + 𝑇 (2.17) Su buharının: 𝐻𝑠 = 𝑚𝑠 + 𝐶𝑝)𝑠 + 𝑇 (2.18) olacaktır. Özgül nem: 𝑥 = 𝑚𝑠𝑚𝑎 (2.19) olup, 1 (kg k.h.) bazında karışım için entalpi denklemi: ℎ = ℎ𝑎 + 𝑥ℎ𝑠 (2.20) ya da özgül ısılara bağlı olarak 𝐶𝑝𝑇 = (𝐶𝑝)𝑎 + 𝑥. 𝐶𝑝)𝑠)𝑇 (2.21) yazılabilir. Buradan karışımın özgül ısısı olarak tanımlanan nemli özgül ısı: 13 𝐶𝑝 = 𝐶𝑝)𝑎 + 𝑥. 𝐶𝑝)𝑠 (2.22) bağıntısıyla hesaplanır. Duyulur Isı, Qd: Bir maddeye ısı verildiği ya da ondan ısı çekildiği zaman maddenin sıcaklığında değişme oluyorsa onun sıcaklığını değiştiren bu ısı duyulur ısıdır. Havanın nemini değiştirmeden yalnız sıcaklığını değiştiren ısıdır. Gizli Isı,Qg: Bir maddeye verildiği ya da ondan ısı çekildiği zaman o maddenin faz durumunu değiştiren, ancak sıcaklığına etki etmeyen ısıdır. Toplam Isı, QT: Duyulur ve gizli ısıların toplamı olup: 𝑄𝑇 = 𝑄𝑑 + 𝑄𝑔 (2.23) şeklinde tanımlanır. 2.3. Adyabatik Doyma İşleminin Deneysel Olarak İncelenmesi: Bir hava su buharı karışımının doyması olayını içeren en önemli psikrometrik işlemlerden birisi Adyabatik doyma işlemidir. İyi yalıtılmış bir su kanalında suyu yalayarak akan bir hava su buharı karışımını ele alalım (Şekil 2.9): Şekil 2.10 Adyabatik Doyma İşlemi Havanın başlangıçtaki bağıl nemi %100’den az ise, bir miktar su buharlaşacak ve hava su buharı karışımının sıcaklığı doğal olarak düşecektir. Kanalın çıkışında karışım, doymuş ve geçirdiği işlem de adyabatik (ısı ve kütle kaybının veya 14 kazancının olmadığı haldeki süreç) ise çıkışta karışım sıcaklığı adyabatik doyma sıcaklığı olur. Bu işlemin bir sürekli akım işlemi içermesi için, buharlaşan su debisi kadar bir besleme suyu (adyabatik doyma sıcaklığında olan su) sisteme eklenmelidir. Bu işlem sürecinde basıncın sabit olduğu varsayılır. 2.3.1 Adyabatik Doyma İşlemi ve Girişte Havanın Özgül Neminin Saptanması: Adyabatik doyma işlemini, kararlı durum-sürekli akım için termodinamiğin 1. Yasasını uygulayarak açıklayalım. Kinetik ve potansiyel enerji değişimlerini yok sayarsak ısı dengesi içeren entalpi denklemini düzenleme yapıp kuru havayı da ideal gaz varsayarsak bu denklem: 𝑤1(ℎ𝑠1 − ℎ𝑙2) = 𝐶𝑝)𝑎(𝑇𝑎2 − 𝑇𝑎1) + 𝑤2(ℎ𝑠2 − ℎ𝑙2) (2.24) Burada: w: Islak havanın özgül nemi (1-girişteki ıslak havanın özgül nemi, 2- çıkıştaki özgül nem) h[kJ/kg]: Özgül entalpisi (a1: girişte kuru havanın özgül entalpisi, a2: çıkıştaki) Cp)a: Kuru havanın sabit basınçta özgül ısısı T [°C]: Sıcaklık (1: havanın girişteki sıcaklığı, 2: sıcaklığı) hl[kJ/kg]: Suyun özgül entalpisi (1: girişte havanın içerdiği su buharının entalpisi olup, aynı sıcaklıkta doymuş su buharının entalpisine eşittir kJ/kg ya da kcal/kg, 2: Yoğuşan sıvının ya da havaya karışan yani doymuş suyun özgül entalpisini) göstermektedir. 𝑇1 = 𝑇𝑎1 , 𝑇2 = 𝑇𝑎2 ve ℎ𝑠2 − ℎ𝑙2 = 𝑟2 İçin, doyurucuya giriş (başlangıç) koşulunda nemli havanın özgül nemi; (Denklem 2.24)’den (2.25) 𝑤1 = 𝐶𝑝)𝑎.𝑇𝑎2 − 𝑇𝑎1 + 𝑤2𝑟2ℎ𝑠2 − ℎ𝑙2 (𝑘𝑔 𝑠. 𝑏./𝑘𝑔. 𝑘. ℎ. ) (2.26) olarak saptanır. 2.3.2 Adyabatik Doyma ve Yaş Termometre Sıcaklıkları: Adyabatik doyma sıcaklığı, karışımın akım borusundan çıkış sıcaklığı olup, bu değer giren hava buhar karışımının, basıncının, sıcaklığının, bağıl neminin, çıkış basıncının bir fonksiyonudur. Böylece adyabatik doyurucu giriş ve çıkışında hava su buharının karışımının sıcaklığının ölçülmesiyle girişteki hava su buharı karışımının özgül nem ve bağıl nemi saptanmış olur. Bu yöntem, bir hava su buharı karışımının ıslaklığının yani neminin belirlenmesi yöntemlerinden en önemlisi olması ve diğer yöntemlere göre de oransal olarak daha kolay ölçümlerin yapılabilmesi nedeniyle en seçeneklisidir. 15 Yaş Termometre sıcaklığı ile adyabatik doyma sıcaklığı arasında ki fark iyi anlaşılmalıdır. Yaş Termometre sıcaklığı, ısı ve kütle transfer hızlarından etkilenmektedir. Buna rağmen, adyabatik doyma sıcaklığı yalnızca adyabatik doyma sıcaklığında ki su ile girişteki hava su buharı karışımı arasındaki dengeyi içermektedir. 3. PSİKROMETRİK DİYAGRAM: İklimlendirme işlemlerinde nemli havada meydana gelen fiziksel değişimlerin izlendiği bir tür durum diyagramıdır. İklimlendirme işlemlerini anlayabilmek için Psikrometrik Diyagramı iyice tanımak gerekir. Diyagram üzerinde her nokta nemli havaya ait bir durumu belirler. Psikrometrik Diyagram üzerinde bir durum noktasının belirlenebilmesi için nemli havaya ait en az iki özelliğin bilinmesi gerekir. Eğer nemli havanın en az iki özelliğin bilinmiyorsa, bu iki özelliğin özelliklerini tayin etmek mümkündür. Psikrometrik Diyagram üzerinde nemli havanın tanımlanmış fiziksel özelliklerinden (kuru, yaş termometre sıcaklığı bağıl,özgül nem vs.) bahsettik. Psikrometrik Diyagramlar genellikle sabit bir hava basıncı standardına göre hazırlanmışlardır. Bu standart hava basıncı 1013 mb (760 mm Hg)’dır. Dünyanın pek çok yerinde bu hava basıncı pek değişmesine karşın, eğer iklimlendirmenin yapılacağı ortamda hava basıncı standart basınçtan çok farklı ise, bu takdirde diyagram yerine kuru havanın kompozisyonu ve kanunlar kullanılmalıdır. Psikrometrik Diyagramı açıklamak için A noktasının özellikleri diyagramda göstertilmiştir. Şekil 3.1 Psikrometrik Diyagram Burada: TA [°C] : A noktasının KT sıcaklığı TA’ [°C] : A noktasının YT sıcaklığı h’A [kJ/kg] : A noktasının entalpisi xA[kg s.b./kg k.h.]: A noktasının özgül nemi 16 PA [ Pa] : A noktasının kısmi buhar basıncı φA [%] : A noktasının bağıl nemi νA [m3/kg]: A noktasının özgül hacmi TA” [°C]: A noktasının çiğlenme sıcaklığını göstermektedir. 3.1 Hava Karışımları: Şekilde de gösterildiği gibi A konumunda mA kütlesindeki hava ile B konumundaki mB kütlesindeki havanın karıştırıldığını varsayalım. Karışım durumunda, bir (C noktası) oluşturacak nemli havanın özellikleri A ve B noktasını birleştirecek doğru üzerinde olacaktır. Şekil 3.2 Psikrometrik Diyagram Üzerinde 1 ve 2. Konumundaki İki Nemli Hava Kütlesinin Karışım Noktası Karışım noktasının AB doğrusu üzerindeki mA ve mR hava kütleleri tayin edilecektir. C karışım noktası daima kütlesi fazla olan hava şartlarına yakın olacaktır. Karışım için denge denklemini yazacak olursak, duyulur ısıtma veya soğutma esnasında nemli havanın alacağı veya vereceği, ısı miktarı: 𝑄 = 𝑉𝐴1𝑣𝐴. (ℎ𝐴 − ℎ𝐵) (3.1) ifadesi ile hesaplanabilir. Burada: 𝑉𝐴[m3/s]: Hava debisi νA [m3/kg]: A noktasındaki havanın özgül hacmi Cp[kJ/kg°C]: Nemli havanın özgül ısısı (1,005) TA [°C]: Nemli havanın A noktasındaki duyulur (KT) sıcaklığı TB [°C]: Nemli havanın B noktasındaki duyulur (KT) sıcaklığı 17 hA [kJ/kg] :Nemli havanın A noktasında entalpisi hB [kJ/kg] : Nemli havanın B noktasında entalpisini göstermektedir. Örnek: 30 °C KT sıcaklığı ve %50 bağıl nemde bulunan 5000 m3/h hava ile 10 °C KT, %20 bağıl nemde bulunan 5000 m3/h hava, deniz seviyesi şartlarında karıştırılmaktadır. Karışım sonrası yeni oluşacak havanın KT’sini ve bağıl nemini hesaplayınız. Çözüm: Psikrometrik Diyagramda da okunacağı üzere; yeni karışım havasının KT sıcaklığı 22,2°C, bağıl nemi %51,9’dur. Debisi daha fazla olan 30 °C’de ki havaya karışım daha yakındır. Birleşim noktası olan PP1’in psikrometrik özellikleri Şekil 3.3’de gösterilmiştir. Şekil 3.3 Örneğin Daikin Psikroemtrik Diyagramda Gösterimi 18 3.2 Duyulur Isıtma ve Soğutma: İklimlendirme işlemlerinde en yaygın işlemlerden olan duyulur ısıtma veya soğutma nemli havanın ısıtıcı veya soğutucu bir serpantin üzerinden geçirilerek ısıtılması veya soğutulmasıdır. Böyle bir işlemde havanın özgül nemi sabit kalacaktır. Bu tip bir hal değiştirme Psikrometrik Diyagram üzerinde yataya paralel bir doğru olarak gösterilmektedir. Şekil 3.4’de A noktasından B noktasına duyulur ısıtılan veya soğutulan nemli bir hava kütlesinin hal değişimi gösterilmiştir. Şekil 3.4 Nemli Havanın Duyulur Isıtılması veya Soğutulması “Duyulur ısıtma” işlemi havaya ısının transferini içerir. Bu işlem klima santrallerinde ve ısıtma cihazlarında ısıtıcı bataryalar, ısı geri kazanım cihazlarında yalnız duyulur ısı transferi yapan ısı geri kazanım üniteleri, elektrikli ısıtıcılar gibi elemanlarca gerçekleştirilir. Bu işlem esnasında nem alış-verişi olmadığından mutlak nem oranı “𝑤” proses boyunca sabit kalır. Ancak ısıtma işlemi esnasında havanın KT sıcaklığı artmakta olduğundan bağıl nem devamlı düşüş gösterir. Bunun nedeni de yükselen KT sıcaklığının doyma noktasının daha yüksek mutlak nem oranlarına tekabül etmesidir. Bu da orantı neticesi bağıl nemin düşmesidir. Bu proses Psikrometrik Diyagramda yatay çizgi olarak gösterilir. İşlem yukarıdaki Psikrometrik Diyagramda görmektesiniz. (B-A arası) “Duyulur soğutma” işlemi ısıtma işleminin tam tersidir. Bu işlem esnasında havanın yalnız KT bazında entalpisi azaltılır, mutlak nem oranı sabit kalır. Bu işleme örnek olarak klima santralleri içindeki soğutucu bataryaları, fan-coil ve endüksiyon cihazlarındaki serpantinleri ve plakalı eşanjörleri gösterebiliriz. Bu işlem esnasında çiy noktası sıcaklığına erişilmediği için soğutulan nemli havanın içindeki mutlak nem oranında herhangi bir değişiklik söz konusu değildir. Duyulur soğutma prosesi yukarıdaki Psikrometrik Diyagramda gösterilmektedir (Şekil 3.4).Görüldüğü gibi proses boyunca mutlak nem oranı “𝑤” değişmemekte ve sabit kalmaktadır. (A-B arası). Duyulur ısıtma veya soğutma esnasında nemli havanın alacağı veya vereceği ısı miktarı: 19 𝑄𝑑𝑢𝑦𝑢𝑙𝑢𝑟 = 𝑉𝐴. 1𝑣𝐴. 𝑐𝑝. (𝑇𝐴 − 𝑇𝐵) (3.2) ifadesi veya 𝑄𝑑𝑢𝑦𝑢𝑙𝑢𝑟 = 𝑉𝐴. 1𝑣𝐴. (ℎ𝐴 − ℎ𝐵) (3.3) ifadesi ile hesaplanabilir. Burada: Q [kW]: Duyulur ısı debisini VA [m3/s] : Hava debisini vA [m3/kg] : A noktasındaki havanın özgül hacmini Cp [kJ/kg°C] : Nemli havanın özgül ısısını (1,005) TA [°C] : Nemli havanın A noktasındaki duyulur (KT) sıcaklığı TB [°C] : Nemli havanın B noktasındaki duyulur (KT) sıcaklığı hA [kJ/kg] : Nemli havanın A noktasındaki entalpisini hB [kJ/kg] : Nemli havanın B noktasındaki entalpisini göstermektir. Örnek: 15000m3/h, 10 °C (KT) sıcaklığında, %80 bağıl nemdeki havanın 30 °C‘ye (KT) kadar ısıtılması sonucu harcanması gereken duyulur ısıyı bulunuz. (İşlemler deniz seviyesindedir.) Çözüm: 10 °C KT sıcaklığında, %80 bağıl nemdeki havanın özgül hacmi 0,810 m3/kg’dır. Yukarıda duyulur ısıtma esnasında havanın alacağı ısı miktarını Denklem 3.2’de belirtmiştik. O halde Denklem 3.6’dan 𝑄𝑑𝑢𝑦𝑢𝑙𝑢𝑟 = 𝑉𝐴. 1𝑣𝐴. 𝑐𝑝. (𝑇𝐴 − 𝑇𝐵) 𝑄𝑑𝑢𝑦𝑢𝑙𝑢𝑟 = 150003600 𝑚3/𝑠.10,810𝑘𝑔/𝑚3. 1,005.(30 − 10) 𝑄𝑑𝑢𝑦𝑢𝑙𝑢𝑟 = 103,4 𝑘𝑊 olarak hesaplanır. Denklem 3.7’den 𝑄𝑔𝑖𝑧𝑙𝑖 = 𝑉.1𝑣 . 2501. (𝑥𝐴 − 𝑥𝐵) 𝑄𝑔𝑖𝑧𝑙𝑖 = 150003600 .10,810. 2501. (0,0061 − 0,0061) = 0 20 𝑄𝑇𝑜𝑝𝑙𝑎𝑚 = 103,4 𝑘𝑊 olarak hesaplanır. Şekil 3.5 Örneğin Daikin Psikroemtrik Diyagramda Gösterimi Örnek: 15000m3/h, 38 °C (KT) ve %60 bağıl nemdeki havanın 24 °C ve %50 bağıl neme kadar soğutulması sonucu havadan çekilecek ısı miktarını hesaplayınız. (İşlemler deniz seviyesindedir.) Çözüm: Psikrometrik Diyagramdan havanın 38 °C (KT) ve %60 bağıl nemdeki havanın özgül hacmi 0,918 m3/kg özgül nemi 0,0254 kg s.b./kg. k.h. olarak okunur. Aynı yöntemle 24 °C ve %50 bağıl nemdeki havanın özgül nemi 0,0093 kg s.b./kg. k.h. olarak belirlenir. Bu kabüller doğrultusunda Denklem 3.6’dan: 𝑄𝑑𝑢𝑦𝑢𝑙𝑢𝑟 = 𝑉𝐴. 1𝑣𝐴. 𝑐𝑝. (𝑇𝐴 − 𝑇𝐵) 𝑄𝑑𝑢𝑦𝑢𝑙𝑢𝑟 = 150003600 .10,918 . 1,005. (38 − 24) 21 𝑄𝑑𝑢𝑦𝑢𝑙𝑢𝑟 = 63,86 𝑘𝑊 Havanın sıcaklığı 38 °C KT’den 24 °C KT’ye düşerken bir miktar su buharlaşır. Bu buharlaşan su hesabı yapılabilmesi için Denklem 3.9 kullanılır. 𝐺𝑠𝑢 = 𝑉 1𝑣 . (𝑥𝐵 − 𝑥𝐴) 𝐺𝑠𝑢 =150003600 .10,918 . (0,0254 − 0,0093) = 0,073 𝑘𝑔/𝑠 su buharlaşır. Bunun için gereken gizli ısı miktarı Denklem 3.7’den: 𝑄𝑔𝑖𝑧𝑙𝑖 = 150003600 .10,918. 2501.(0,0254 − 0,0093) 𝑄𝑔𝑖𝑧𝑙𝑖 = 182.76 𝑘𝑊 Havayı 38 °C KT’den 24 °C KT’ye düşürebilmek için gereken toplam ısı miktarı: 𝑄𝑇𝑜𝑝𝑙𝑎𝑚 = 63,86 𝑘𝑊 + 182,76 𝑘𝑊 = 246,62 𝑘𝑊 olarak hesaplanır. Şekil 3.6 Örneğin Daikin Psikroemtrik Diyagramda Gösterimi 22 3.3 Nemli Havanın Neminin Sıkılması (Nem Alma) (De humidifaction): Nemli havanın neminin sıkılması veya duyulur sıcaklığının çiğlenme noktasının daha altında bir sıcaklığa düşürülmesi işlemine iklimlendirmede Dehümidifikasyon denir. Bu işlem iki yolla yapılabilir: 1. Nemli Hava, duyulur sıcaklığı havanın çiğlenme noktası sıcaklığından daha düşük bir yıkayıcıdan (sprey soğutucu) geçirilir 2. Nemli hava, yine ortalama yüzey sıcaklığı (cihaz çiğlenme noktası) havanın çiğlenme sıcaklığından daha düşük bir soğutucu serpantin üzerinden geçirilir. Her iki yöntemde de, yani gerek yıkayıcı da gerekse soğutucu serpantinle yapılan dehümidikiasyon işleminde; havanın çıkış şartları, Psikrometrik Diyagramda havanın konumunu belirten A noktası ile cihazın veya yıkayıcının çiğlenme noktasını gösteren C noktasını birleştiren AC doğrusu üzerinde olacaktır (Şekil 3.7). Teorik bir işlemde eğer yıkayıcıdan havanın tam olarak yıkanabilmiş veya soğutucu serpantinde tam olarak denge durumuna geldiği ve C noktası şartlarında yıkayıcıyı veya soğutucu serpantini terk ettiği söylenebilir. Ancak pratikte genellikle olaylar böyle gerçekleşmez. Hava yıkayıcıyı veya soğutucu serpantinini C noktasına yakın AC doğrusu üzerindeki bir B noktasında terk eder. Bir başka deyişle, bu tip yıkayıcı veya soğutucu serpantinlerin verimi 1’den küçüktür. İklimlendirme projesinde bu olayın etkililiği kontak faktörü olarak bilinen ve aşağıdaki şekilde gösterilen sabitle belirlenir. β = XA − XBXA − XC= hA − hBhA − hC (3.4) Şekil 3.7 Havanın Sıkılması İşleminin Psikrometrik Diyagram Üzerinde Gösterimi 23 Fiziksel olarak, kontak faktörü (β) yıkayıcıya veya soğutucu serpantine giren havanın, yıkayıcı veya soğutucu serpantin yüzeyleri ile temas eden kısmın yüzdesini göstermektedir. Bazı literatürde kontak faktörü tanımı yerine, βy-Pass faktörü tanımı kullanılmaktadır. Burada: TA [°C] : Giriş havanın KT sıcaklığı TB [°C] : Çıkış havasının KT sıcaklığı TD [°C] : Giriş havasının çiğlenme noktası sıcaklığı TC [°C] : Yıkayıcı su sıcaklığı veya soğutucu serpantin yüzey sıcaklığını göstermektedir. β = TA − TBTA − TC (3.5) yazılabilmektedir. Dehümüdifikasyon işlemine tabi tutulan havadan hem duyulur hem de gizli ısı alınmaktadır. 𝑄𝑑𝑢𝑦𝑢𝑙𝑢𝑟 = 𝑉. 1𝑣 . 𝐶𝑝. (𝑇𝐴 − 𝑇𝐵) (3.6) 𝑄𝑔𝑖𝑧𝑙𝑖 = 𝑉. 1𝑣 . 2501. (𝑥𝐴 − 𝑥𝐵) (3.7) ifadeleri yardımı ile hesaplanabilir. 2501 değeri kJ/kg olarak suyun ortalama buharlaşma ısısını göstermektedir. A noktasından B noktasına sıkılan havanın verdiği toplam ısıda, duyulur ve gizli ısıların toplamına eşit olacaktır. 𝑄𝑡𝑜𝑝𝑙𝑎𝑚 = 𝑄𝑑𝑢𝑦𝑢𝑙𝑢𝑟 + 𝑄𝑔𝑖𝑧𝑙𝑖 = 𝑉. 1𝑣 . (ℎ𝐴 − ℎ𝐵) (3.8) Pratik bir veri olarak endüstriyel amaçlı yapılan soğutucu serpantinlerde β (kontak faktörü) 0,80-0,90 değerleri oranında değişir. 24 3.4 Nemlendirme (Evaporatif Soğutma): Nemli havanın nem oranının arttırılması, bir başka deyişle özgül neminin arttırılması iklimlendirmede nemlendirme işlemi olarak adlandırılır. Nemli havanın özgül nemini arttırılması, birkaç değişik yöntemle yapılabilir bunlardan bazıları: 1. Nemli hava bir kanal içindeki uzun bir ıslak yüzey üzerinden geçirilebilir. 2. Nemli hava bir kanal içinde, sıralar halinde dizilmiş su püskürten veya damlatan damlaları arasından geçirilebilir 3. Hava bir kanalın içinden geçerken içine yüksek basınçta su pulverize bir halde enjekte edilebilir. 4. Son olarak, hava yine bir kanalın içinden geçerken içine buhar enjekte edilebilir. Yukarıda tanımlanan ilk üç metodun, Psikrometrik Diyagram üzerinde gösterimi birbirinin benzeri olacak ve nemli hava bir A noktasından B noktasına sabit YT doğrusu üzerinde gidecektir. Tabi böyle bir nemlendirme işleminde, şartlandırma hücresine su havanın ortalama YT sıcaklığında beslenecektir. Eğer su şartlandırma hücresine soğutulmuş olarak sevk edilecek olursa bu takdirde daha öncede bahsedildiği gibi hal değişimi A noktası ile soğutulmuş su sıcaklığını birleştiren doğru üzerinde olacak ancak bu durumda nemlendirme yerine havanın sıkılması işlemi (nem alma) gerçekleştirilecektir. İlk iki yöntem ile gerçekleştirilen bir nemlendirme işleminin Psikrometrik Diyagram üzerinde (Şekil 3.8) gösterilmiştir. Şekil 3.8 Islak Hücre veya Yıkayıcılarda Havanın Nemlendirilmesi İşlemi 25 Böyle bir işlemde havaya verilen su miktarı: 𝐺𝑠𝑢 = 𝑉. 1𝑣 . (𝑥𝐵 − 𝑥𝐴) (3.9) ifadesi ile hesaplanır. Burada: Gsu [kg/s] : Havaya verilen su miktarını V [m3/s] : Hava debisini 𝑣 [m3kg]: Nemli havanın özgül hacmi XA,XB , XC [kg s.b./kg k.h] : Havanın giriş ve çıkış konumlarında özgül nemi göstermektdir. Böyle bir nemlendirme meme tipi yıkayıcılarda yapılıyorsa yıkayıcı verimi; 𝐸 = XB − XAXC − XA≌ TA − TBTA − TC (3.10) şeklinde tanımlanabilir. Hava nemlendirmesinde bir diğer yöntem hava içine basınç altında pülverize halde su püskürtülmesidir. Bu yöntemde havaya püskürtülen suyun yaklaşık tamamı buharlaşarak havaya karışacaktır. Böyle bir işlemde havaya püskürtülen su sıcaklığının 0 °Cveya 100 °Colması halinde, hal değişimi olayının YT doğrusundan açısal kayması yaklaşık 7° civarında olacaktır (Şekil 3.9). Tasarımda böyle bir değişimin yaratacağı hata kabul edilebilir bir aralıktadır. Şekil 3.9 Havanın su enjekte edilerek nemlendirilmesi 26 Bu metotda havaya verilen su miktarı ile havanın gizli ısısındaki artma miktarı ilk iki metot için verilen denklemlerle hesaplanabilir. Ancak bu metot için bir yıkayıcı veya püskürtme verimi tarifi yapılamaz. Püskürtme memesini debisini ayarlamak sureti ile hava istenilen her hangi bir noktaya kadar nemlendirilebilir. Hava nemlendirmesinde son yöntem havaya buhar püskürtülmesidir. Böyle bir nemlendirme olayının Psikrometrik Diyagram üzerinde hal değişimi aşağıda gösterilmiştir (Şekil 3.10). Şekil 3.10 Buhar Enjeksiyonu İle Hava Nemlendirmesi A noktasında buharlı nemlendiriciye giren havanın nemlendirici çıkışında yeni entalpisi hB’dir. ℎ𝐵 = ℎ𝐴 + 𝑔𝑏ℎ𝑏 (3.11) kadar olacaktır. Burada: ℎ𝐵 [kJ/kg] : Nemli havanın çıkış noktasındaki entalpisi ℎ𝐴: [kJ/kg] : Nemli havanın giriş şartlarındaki entalpisi 𝑔𝑏: [kg/kg KH] : Birimi kuru havaya verilen buhar debisi ℎ𝑏: [kJ/kg] : Buharın entalpisini göstermektedir. Yine nemli havanın çıkış noktasındaki entalpisi, ℎ𝐵 = 𝐶𝑃[(1.007𝑇𝐵 − 0,026) + 𝑥𝐵(2501 + 1,84𝑇𝐵)] (3.12) ifadesi yardımıyla hesaplanabilir. 27 Burada: ℎ𝐵 [kJ/kg] : Nemli havanın çıkış noktasındaki entalpisi 𝑇𝐵: [°C] : Nemli havanın çıkış noktasındaki KT sıcaklığı 𝑥𝐵: [kg/kg KH] : Nemli havanın çıkış noktasındaki özgül nemini 𝐶𝑝: [kJ/kg°C] : Nemli havanın özgül ısısını göstermektedir. (1,005 kJ/kg°C) ℎ𝐵ve 𝑇𝐵 gibi havanın buharlı nemlendiriciden çıkış şartlarına ait iki özelliği tespit edildikten sonra Psikrometrik Diyagramda çıkış noktası bulunabilir (B). Bu nokta giriş noktası ile birleştirilirse hal değişimi doğrusunun düşey bir doğrudan yaklaşık 4° ‘lik bir açısal sapma yaptığı görülecektir (Şekil 3.10). Tasarımla ilgili iklimlendirme işlemlerinde buharla nemlendirme ile hal değişiminin genellikle (𝑥) apsisine dik yönde bir doğru üzerinde meydana geleceği varsayılır. Bu varsayımın meydana getireceği hata gene tasarım çalışmalarında kabul edilebilecek bir mertebededir. Örnek: 35 KT, %30 bağıl nemdeki 10000 m3/h’lik hava %80 bağıl nem olana kadar %80 verime sahip sulu püskürtme sistemi ile su verilmektedir. Nemlendirme sonucunda harcanan su miktarını, havanın nemlendirme sonucunda ulaştığı KT sıcaklığını hesaplayınız. Bu işlem sonucunda havadan ne kadar gizli ısı çekilmiştir? Çözüm: Psikrometrik Diyagramdan 35 KT, %30 bağıl nemdeki havanın 21,5°C YT sıcaklığı okunur. Yıkayıcı verimi Denklem 3.10’dan: 𝐸 = TA − TBTA − TC= 0,8 = 35 − TB35 − 21,5 TB = 24,2 °C Nemlendirme sonucu ulaşılan sıcaklıktır. Nemlendirme sırasında harcanan su miktarı Denklem 3.9’dan: 𝐺𝑠𝑢 = 𝑉 1𝑣 . (𝑥𝐵 − 𝑥𝐴) 𝐺𝑠𝑢 = 100003600 .10,888 . (0,0180 − 0,0105) 𝐺𝑠𝑢 = 0,024 𝑘𝑔/𝑠 su harcanmıştır. Bu işlemler sonucunda havadan çekilen duyulur ısı Denklem 3.6’dan: 𝑄𝑑𝑢𝑦𝑢𝑙𝑢𝑟 = 𝑉. 1𝑣 . 𝐶𝑝. (𝑇𝐵 − 𝑇𝐴) 28 𝑄𝑑𝑢𝑦𝑢𝑙𝑢𝑟 = 100003600 .10,888 . (35 − 21,5) = 42,23 𝑘𝑊 Şekil 3.11 Örneğin Daikin Psikroemtrik Diyagramda Gösterimi 4. TERMOELEKTRİK SOĞUTUCU (PELTİER): Bir maddenin veya bir ortamın soğutulması birkaç yöntem ile yapılabilmektedir: Termodinamik bir çevrime sahip olan (soğutma çevrimi) bir kompresör ile devrenin tamamlandığı sistemler. En yaygın olan soğutma sistemi bir çevrime sahip olan ve soğutucu bir gazın ısıyı bir kaynaktan diğerine taşıdığı sistemlerdir. Örneğin: Buzdolabı, klima sistemleri, büyük buz haneler, soğuk odalar vb. 29 Şekil 4.1 Soğutma Çevrimi Bölüm 2’de anlatılan evaporatif soğutma sistemleri Herhangi bir soğuk naktan faydalanarak yapılan soğutma işlemleri: Örneğin düşük bir sıcaklığa sahip akarsudan faydalanılarak yapılan soğutma. Termoelektrik soğutma: Bir doğru akım kaynağı ile beslenen yarı iletken malzemelerden faydalınalarak yapılan soğutma sistemidir. Aşağıda detaylı anlatılacaktır. 4.1 Termoelektrik Soğutma: Termoelektrik, ısı ve elektrik enerjisinin birbirine dönüşümüdür. Bu etki ilk olarak 1821 yılında Alman fizikçi Johann Seebeck tarafından keşfedilmiştir. Seebeck, iki farklı metal plakayı iki ucundan birbirine temas ettirmek suretiyle bir devre kurmuş ve bu devreyi bir ucundan ısıtmıştır. Isının etkisiyle yakındaki bir mıktnatısın hareket ettiğini fark etmiştir. Yani ısının etkisiyle bir elektrik akımı üretilmiş, üretilen elektrik akımı da bir elektrik alana neden olarak mıknatıs hareket etmiştir. 30 Şekil 4.2 Seebeck Deneyi Bu olayda ilginç olan, devrede uçlar arasında bir sıcaklık farklı oluşması ve dolayısıyla uçlar bir ısınma veya soğumanın meydana gelmesidir. Isıl eleman çiftlerinin (Termokapıl) temelini oluşturan ve 1821 yılında keşfedilen Seebeck etkisinin tersi olan Peltier etkisi, Fransız bilim adamı Jean Charles Athanese Peltier tarafından 1834 yılında keşfedilmiştir. 1950 yılından sonra yarıiletken teknolojisinin gelişimiş ile birlikte Peltier etkisi değişik alanlarda kullanılmıştır. (Godfrey, 1996) Termoelektrik soğutma sisteminde herhangi bir kompresör, ısının bir yerden bir yere transferi için bir sıvı veya gaz akışkan kullanımı söz konusu değildir. Basit olması nedeni ile küçük maddelerin ve küçük hacimlerin soğutulmasında yaygın olarak kullanılmaktadır. Temel olarak bir doğru akım kaynağından beslenen yarı iletken bir metal söz konusudur. Yarı iletken metale doğru akım verildiğinde metalin bir yüzeyinde sıcaklık düşümü olurken diğer yüzeyinde ısı artımı dolayısıyla sıcaklık yükselmesi söz konusudur. Şekil 4.3 Termoelektrik Soğutucu Çalışma Prensibi (url:1) Bu termoelektrik soğutucu sisteminin çalışma prensibini detaylı anlatabilmek için “Peltier” tarafından geliştirilen “Peltier etkisini” ve bu etkiye bağlı olarak sistemin çalışma prensibini aşağıda anlatılacaktır. Ayrıca termoelektrik soğutma sistemini 31 tanımlamamıza yarayan diğer iki yöntem de “Thomson Etkisi” ve “Seebeck Etkisi” kısaca bu konularda açıklanacaktır. 4.2. Termoelektrik Etkiler: Termoelektrik, sıvı ve katı maddelerde ısı ile elektrik enerjisinin birbirine dönüşümünü inceleyen bilim dalıdır. Elektronik aletleri şarj etmek için ve sıcaklık sensörlerinin yanı sıra ısıtma - soğutma amaçlı yaygın olarak kullanılır. 4.2.1 Seebeck Etkisi: İletkenlerin arasında bir ucun soğuk diğer ucun ise sıcak olduğunda arasında oluşan sıcaklık farkı sebebiyle elektirik oluşmasıdır. Yani sıcaklık farkından kaynaklıdır. Elektiriğin oluşmasının sebebi iletkenin üzerinde oluşan potansiyel farktır. Bu fark sayesinde peltier cihazı üzerinden elektrik üretimi yapılır. Bu elektrik üretiminde malzeme cinsi önemli yer alır. Malzeme cinsinin etkisini formülasyonda Seebeck Katsayısı olarak adlandırılır. Şekil 4.4 Seebeck Voltajının Ölçümü (Rowe ve Bhundar,1998) 𝛥𝐸 = 𝛼. 𝛥𝑇 (4.1) Burada: ΔE [Volt]: Potansiyel fark α [V/ oC] : Seebeck katsayısı ΔT [oC]: Sıcaklık farkını göstermektedir. 32 α' nın değeri; devreyi oluşturan malzemelerin özelliklerine bağlıdır. Madde Seebeck Katsayısı Madde Seebeck Katsayısı Madde Seebeck Katsayısı Alüminyum 3.5 Altın 6.5 Germanyum 300 Antimon 47 Demir 19 Potasyum -0.9 Bizmut -72 Kurşun 4 Selenyum 900 Kadmiyum 7.5 Cıva 0.6 Silikon 440 Karbon 3.0 Nikrım 25 Sodyum -2.0 Konstantan -35 Platin 0 Tellür 500 Tablo 4.1. Bazı maddelerin Seebeck katsayıları 4.2.2 Peltier Etkisi: Peltier etkisinde amaç elektrik akımı sayesinde bir cismi ısıtmak veya soğutmak için kullanılır. Peltierler günlük hayatımızda mini buzdolapları, bilgisayar işlemcisi, araba soğutucuları, uzay teleskopları gibi benzer yerlerde kullanılırlar. Bu bölgelerde kullanılmasının en büyük sebeplerinden birisi içerinden soğutucu akışkan geçmeden soğutma işlemi yapabilen bir cihazdır. Peltierler soğutucu akışkan motor vb. gibi cihazlar bulunmadığı için uzun ömürlülerdir. Şekil 4.5 Nem Alma Cihazında Peltier 33 Şekil 4.6 Bilgisayar İşlemcisinde Peltier Yukarıdaki fotoğraflarda peltieri günlük hayatımızda nerelerde kullacağımızı görmekteyiz. Şekil 4.5’de bir nem alma cihazında kullanılan peltier, Şekil 4.6’da bilgisayar işlemcisinin soğumasını sağlamak için görülmektedir. Genel olarak piyasada soğutma amaçlı kullanılan peltierlerin bu kadar yaygın olarak kullanılmasının sebebi montajı ve kullanımı kolaydır. Soğutma işleminde güvenlidir az yer kaplar ve soğutma işlemini iyi derece gerçekleştirebilir. Şekil 4.7 Peltier Görünümü Peltierler, hareketli parçası olmayan küçük ısı pompasıdır. Peltierler genelde alanın az olduğu, güvenliğin önemli olduğu ve soğutucu gazların kullanılmaması gereken ortamlar için uygundur. (Yavuz ve Ahıska, 2006). Peltier’in bir yüzeyindeki ısının diğer yüzeye transferiyle soğutma gerçekleşir. Şekil 4.8.' de bir Termoelement’in yapısı görülmektedir. 34 Şekil 4.8 Termoelement Mödülünün İç Yapısı (Thakkar,2016) Peltierler P ve N tipi yarı iletkenlerden oluşurlar. Yarı iletkenlerin P tipi olanları pozitif termoelement, N tipi olanlara negatif termolement isimleri verilir. Yarı iletkenler birbirine seri olacak şekilde bakır elektrik kablosu ile birbirlerine bağlanırlar. Bağlanmadan sonra elektrik akımı sayesinde P ve N tipi iletkenlerin üstünde bulunan elektronlar harekete geçerler. Hareketlenme sonucunda geçen elektron doğru akım sayesinde bir yüzeye doğru harekete geçer. Elektronlar belli bir enerjiye sahip oldukları için harekete geçtikleri yüzey ısınırken ayrıldıkları yüzey ise soğumaya başlar. Uygulanan akım yönü değiştirilse, elektronun hareket yönü değişeceği için sıcak yüzey soğuyacak, soğuk yüzey ise ısınacaktır (Şekil 4.9). Termoelektrik soğutma işleminde uygulanacak elektrik akımı doğru akım olmalıdır. Şekil 4.9Termoelement bileşenleri (Yavuz ve Ahıska,2006) “p” tipi elemanda (Şekil 4.10); • Elektrik akımı serbest "delikler" tarafından taşınır. • Isı akımı, delik hareketiyle aynı yöndedir. • “n”-tipi elemanla seri bağlanınca ısıyı ters yönde iletir. (Url:3). 35 Şekil 4.10 P-tipi eleman çalışma şeması (Thakkar,2016) “n” tipi elemanda (Şekil 4.11); • Bol miktardaki serbest elektronlar hareket ederken, ısıyı da beraber taşırlar. • Isı akımı, elektrik akımıyla aynı yöndedir (Url:3). Şekil 4.11 N-tipi eleman çalışma şeması(Thakkar,2016) 4.2.3 Thomson Etkisi: Termoelektrik etkilerden üçüncüsü olarak 1851 yılında keşfedilen Thomson Etkisidir. Termoelektrik etki, elektriksel gerilim ve sıcaklık farkı arasındaki doğrudan dönüşümdür. Bu etki en temelde Seebeck etkisi ile açıklanmıştır. Bu etkiye göre, farklı iki metalin bağlantı noktaları olan iki eklemde sıcaklık farkının olması akım oluşmasına yani gerilim indüklenmesine neden olmaktaydı. Bu farklı metallerin eklemlerinde açığa çıkan ya da emilen ısı miktarının ise bu eklem noktalarından geçen akımla ilgili olduğu Peltier ektisi ile açıklanmıştı. Thomson ektisi ise, homojen bir iletkenin faklı noktalarındaki farklı sıcaklık bölgeleri için açığa çıkan ya da emilen ısı miktarını açıklayan olayıdır. (Köysal, 2018) Bu etkide iletken uçların farklı sıcaklıkta tutulduğu bir durumda uygulanan akım yönüne bağlı olarak enerjinin iletken üzerinde absorde edilmesi veya soğurulması olayına denir.Thomson etkisinde bahsedilen iletkenler homojen iletkendir. Yani olayı biraz daha somutlaştırırsak sıcaklıkları farklı iki uçta akım taşındığında metal ya ısı 36 emer ya da ısı verir. Thomson etkisinde tüm olay, T1 ve T2 uç sıcaklıklarına, metalin cinsine ve akımın yönüne bağlı olarak değişir. Thomson etkisinin günümüzde doğrudan uygulama alanı bulunmamaktadır. Şekil 4.12 Thomson Etkisi (Tosun, 2014) 5. MATEMATİKSEL FORMÜLASYON: 5.1. Evaporatif Soğutucu Tasarımında Kullanılacak Kabüller: Soğutulması düşünülen örnek otel odası Şekil 5.1’de gösterilmiştir. Bu hesaplamada iç ortam şartları: 23 °C KT sıcaklığı, %40 bağıl nemdedir. Dış ortam Malatya şartları: 38 °C KT, 21 °C YT sıcaklığındadır. Hesaplamalar Malatya’da bir otel odasına göre yapılmıştır (Şekil 5.1)’de. Bu otel odasını oluşturan duvar, yalıtım, cam, kapı gibi malzemelerin ısı transfer katsayıları hesaplanmıştır. Bu ısı transfer katsayısı, soğutulması düşünülen mahalin saatlik ısı kazancını hesaplarken yardımcı olacaktır. 37 Şekil 5.1. Soğutulacak Mahal ve Kesiti Odada infiltrasyon yok sayılacaktır. Odanın saatlik ısı kazancı hesaplanırken CARRIER tarafından geliştirilen HAP programı kullanılacaktır. Programda gerekli parametreler girildiğinde soğutulması düşünülen mahalin saatlik ısı kazancı 2,34 kW’dır. Bu odanın soğutma yüküne bağlı olarak odaya gönderilmesi gereken hava debisi hesaplanacaktır. Odanın içi 23 °C KT sıcaklığında tutulacaktır. Odada ki ısı yükünü alabilmemiz için havayı bu sıcaklıktan daha düşük bir KT sıcaklığında üflemememiz gerekir. Ancak Malatya şartlarında Evaporatif Soğutucuyla üflenen havanın çıkış sıcaklığı minimum 23 °C KT sıcaklığında kabul edilmiştir. Havanın daha düşük bir sıcaklıkta üflenebilmesi demek, havayı 21 °C KT sıcaklığında üflemektir. Bu teoride gerçekleşebilse bile gerçeklik için mantıksızdır. 38 Şekil 5.2 Psikrometrik Diyagram üzerinde üflenecek hava şartları Evaporatif Soğutma yaparken havanın Psikrometrik Diyagram üzerinde işlemi h (entalpi)’ın sabit kalması yönündeydi. Birkaç derecelik kaymalar tasarımda kabul edilebilir düzeyde olduğu belirtilmişti. 21 °C KT sıcaklığında üflemek için havanın %100 doyma noktasında olması gerekir. Bu da önceki tanımlarımızda belirttiğimiz üzere çiğ noktasına tekabül etmektedir. Yine konu anlatımımızda da belirttiğimiz gibi %100 bağıl nemde üflenen hava, konfor şartlarının dışında sayılmaktadır (Şekil 2.1). O yüzden üflenecek hava 23 °C KT sıcaklığı kabul edilmiştir. Bu kabuller yapıldıktan sonra HAP programı ile odanın saatlik ısı kazancını 2,34 kW olarak belirtmiştik. 5.2. Peltier Tasarımında Kullanılacak Kabüller: Peltier soğutması, elektronlarla ısı transferi gerçekleşir. Doğru akım sayesinde elektronlar akım yönünde hareketlenmeye başlar, hareketlenen elektronların ayrıldıkları yüzey soğur ulaştıkları yüzey ise ısınır. Elektronlar P ve N iletkenlerinin üstünden geçerek ilerler. Peltier’in sıcak yüzeyin soğuması Peltier’in en önemli olaylarından biridir çünkü Peltier çok hassas bir cihazdır soğutma işlemi yapılırken sıcak yüzey soğutulmazsa cihaz bozulacaktır. Peltier verimi de bu olaya bağlıdır sıcak yüzey ne kadar soğutulabilirse,soğuk yüzey o kadar verimli olur. 39 Peltierin dış ortam sıcaklığı 27 °C İçme suyu 8 °C Kullanıcı tarafından doldurulan suyun sıcaklığı 28 °C olarak kabul edilmiştir. 5.3. SİSTEM DİZAYNI: Bu projede, evaporatif soğutucu ile yukarıda bahsi geçen otel odası 23 °C KT oda sıcaklığına kadar soğutulacaktır. Diğer tarafta Peltier soğutucu ile saatte 1,6 Litre su 28 °C ‘den 8 °C ‘ye kadar soğutulacaktır. Her iki sistemde de bir su sirkülasyonu söz konusudur. Evaporatif soğutucuda, soğutucunun alt haznesinden alınan su daha önce anlatıldığı gibi üst kısımdaki delikli kaba basılmaktadır. Diğer taraftan Peltierin sıcak yüzeyindeki ısıyı sistemden uzaklaştırmak için yine su kullanılacaktır. Sistem şeması Şekil 5.3 ‘de gösterilmiştir. Şekil 5.3 Sistem Şeması Bu projede amaç her iki cihaz için kullanılacak suyu bir devrede toplayıp kompakt bir cihaz yaratmaktır. Öncelikle evaporatif soğutucu boyutlandırılacak, gerekli olan su miktarına bağlı olarak evaporatif soğutucu pompa debisi hesaplanacaktır. Peltier soğutucu’da hesaplanacak olan gerekli su debisi evaporatif soğutucudakiyle kıyaslanacak, iki pompa yerine tek pompa seçilecektir. Tabi ki bu pompa hangi cihazın daha fazla suya ihtiyacı varsa minimum o kapasitede olacaktır. 40 5.3.1.Evaporatif Soğutucu Hesapları: 5.3.1.1. Hava Debisi Hesabı: Hava debisi hesabı için odadan çekilmesi gereken ısı miktarı belli olduğundan aşağıdaki formülü yazabiliriz: 𝑄 = 𝑚. 𝑐. 𝛥𝑇 = 𝑣. 𝜌. 𝐶. 𝛥𝑇 2,34 𝑘𝑊. 860𝑘𝑐𝑎𝑙ℎ= 𝑣. 1,2.0,24. (25 − 23) 𝑣 = 3493 𝑚3/ℎ ≌ 3500 𝑚3/ℎ 3500 𝑚3/ℎ hava debisini sağlayacak fanın seçimi için yapılmış deneyler ve testler incelenmiştir. Şekil 5.4 1 ve 2 nolu fan mahal içi gösterimi 5.3.1.2. Hava Debisi ve Hava Basınç Kayıplarına Bağlı Olarak Fan Seçimi: Şekil 5.5. 1 numaralı fanın şematik gösterimi Panjur+filtre+soğutucu ped için basınç kaybı 60 Pa’dır. Bu parametrelere göre fanın harcadığı güç: 41 𝑃𝑓𝑎𝑛−1 = 𝑄 𝑚3ℎ . ΔP𝑡𝑚𝑚𝑆𝑆3600.102 𝑃𝑓𝑎𝑛−1 =3500𝑚3ℎ⁄ . 6,12 𝑚𝑚𝑆𝑆3600 .102 𝑃𝑓𝑎𝑛−1 = 0,058 𝑘𝑊 = 60 𝑊𝑎𝑡𝑡 ‘dır. 2. fanın hava debisi aynı olmakla beraber fan basınç kaybı dışarıya konulan panjur ve fan bağlantı detayından kaynaklanır. 2 numaralı fanda filtre, soğutucu ped olmadığı için basınç farkı 25 Pascal’a eş değerdir. Aynı işlemi 2. fan için yaptığımızda: Şekil. 5.6. 2 nolufanın şematik gösterimi 𝑃𝑓𝑎𝑛−2 =3500𝑚3ℎ⁄ . 2,55 𝑚𝑚𝑆𝑆3600 . 102= 0,024 𝑘𝑊 = 24 𝑊𝑎𝑡𝑡′𝑑𝚤𝑟. 5.3.1.3 Nemlendirme İşleminde Havaya Verilen Su Miktarı: Evaporatif cihazın hava debisi belirlendiğine göre ve havanın evaporatif soğutucuya giriş ve çıkış sıcaklıkları da bilindiğine göre, havaya verilmesi gereken su miktarı hesaplanabilir. %85 nemlendirme verimine sahip sprey ile nemlendirme işlemine uğrayan hava 23 °CKT sıcaklığı, %85 bağıl nemde çıkmakta olduğu varsayılmaktadır. Hava aşağıda ki psikrometride görüldüğü gibi Malatya 38 °C KT, 21 YT sıcaklıklarında girip %85 bağıl nem ve 23 °C KT sıcaklığında cihazı terk etmektedir. 42 Şekil. 5.7 Psikrometrik Diyagramda Giriş ve çıkış şartları Nemlendirme işlemi sırasında harcanan su miktarını hesaplamak için Denklem 3.9 kullanılır. Burada dış ortam havası olan 38 °C KT sıcaklığı, 21 °C YT sıcaklığında Malatya havasının: mutlak nemi 0,0086 kg s.b./kg k.h., 23 °C KT, %85 bağıl nem olan odaya giriş havasının: mutlak nemi ise 0,015 kg s.b./kg k.h. olarak Psikrometrik Diyagramdan okunmuştur. Şekil 5.8. Malatya şartlarında ve Odaya Giriş Şartlarında Havanın Psikrometrik Özellikleri 43 Denklem 3.9’dan 𝐺𝑠𝑢 = 𝑉1𝑣 . (𝑥ç𝚤𝑘𝚤ş − 𝑥𝑔𝑖𝑟𝑖ş) 𝐺𝑠𝑢 = 3500 𝑚3ℎ⁄ .10,894 𝑚3𝑘𝑔⁄. (0,015 − 0,0086) 𝐺𝑠𝑢 = 25,06 𝑘𝑔ℎ ≌ 25 𝑘𝑔ℎ Nemlendirme verimi %85 olduğu kabulüne göre ise havayı nemlendirmek için 𝐺𝑠𝑢 0,85⁄= 29.47𝑘𝑔ℎ ≌ 30𝑘𝑔ℎ su harcanmaktadır. Sadece Evaporatif Soğutma işlemi için kullanılacak su miktarı 30 L’dir. Bu suyu depolayacak olan su tankının ve evaporatif soğutucu cihazın boyutları Şekilde gösterilmiştir. Yıkayıcı verimi Denklem 3.10’dan: 𝐸 = 𝑥𝐵 − 𝑥𝐴𝑥𝐶 − 𝑥𝐴= 𝑇𝐴 − 𝑇𝐵𝑇𝐴 − 𝑇𝐶 𝐸 = 0,0150 − 0,00860,0157 − 0,0086 = 0,901 ≌ %90 olarak hesaplanmıştır. 5.3.1.4. Pompa Seçimi: Bu suyu karşılayabilmek için hidrolik verimi %75 olan pompanın 0,04 m3/h ve basınç kaybı olarak 0.5 mSS statikden, 0,5 mSS evaporatif soğutmanın nozulu ve boru basınç kayıplarından da geldiği düşünülerek 1mSS kabul edilmiştir. Bu parametrelere göre pompanın gücü: 𝑃 =𝑄(𝑚3ℎ ) . 𝐻(𝑚𝑆𝑆) . 𝜌(𝑘𝑔𝑚3)367 . 𝜂 𝑃 = 0,04 . 1 . 1000367 .0,75 = 0,145 𝑘𝑊 = 145 𝑊𝑎𝑡𝑡 Pompanın mil gücü P ≤ 15kW olduğundan nominal gücü bulmak için 1,1 katsayısı ile çarpılır. 0,145 𝑘𝑊 . 1,1 = 160 𝑊𝑎𝑡𝑡 190 Watt enerji tüketen solenoid su tahliye pompası seçilebilir. 44 Şekil 5.9. Cihaz Boyutu ve Su Tankı Boyutları 5.3.2 Soğutucu Ped ve Cihaz Boyutlandırması: Obando ve diğ. (2020) kapalı bir süt üretim çiftliği için evaporatif soğutma sistemi kullanmak istemiş, bu sistemin soğutucu ped boyutlandırılmasını deneme yanılma yöntemi ile test etmektense matematiksel bir model geliştirmişlerdir. Soğutucu pedden geçen nemli havanın (Şekil 5.10) iki ideal gazın karışımı olduğu varsayılır: kuru hava ve su buharı. Nemli hava pede hızı 𝑢, KT sıcaklığı 𝑇𝑑𝑏𝑖ve spesifik nem𝑖ile girer,KT sıcaklığı 𝑇𝑑𝑏𝑜 ve spesifik nem𝑜 ile çıkar. Hava hızı 𝑢 ve nemli hava kütle akış hızı ṁ𝑎'nın soğutma pedi boyunca sabit olduğu varsayılır. Şekil 5.10 Soğutucu Ped Modeli (Obando ve diğ.) (2020) Kararlı durum koşulları dikkate alınırsa; ped soğutmanın iç kısmındaki sonsuz küçük ıslatılmış yüzey alanı (𝑑𝐴𝑠) üzerinde bir kütle ve enerji dengesi yazılırsa:Nemli 45 havanın kuru termometre sıcaklığı (𝑇𝑑𝑏) ve spesifik nemindeki () değişimi hesaplamak için diferansiyel denklem yazılır. Bu denklem, soğutma pedi ana çıkış değişkenlerinin davranışının belirlenmesini sağlar: KT sıcaklığı, bağıl nem ve soğutma verimliliği ve ayrıca su sıcaklığının bunlar üzerindeki etkisinin belirlenmesi gibi. Başlangıç koşulları:𝑇𝑑𝑏(0) = 𝑇𝑑𝑏𝑖𝑣𝑒(0) = 𝑖 0 ≤ x ≤ Liçin: 𝑑𝑇𝑑𝑏𝑑𝑥 =𝜍𝑊𝐻(𝑇𝑠 − 𝑇𝑑𝑏)(ℎ𝐻 + ℎ𝑀𝐶𝑝𝑢@0°C(ρvs − ρv))ṁ (𝐶𝑝𝑎 + 𝑝𝑢@0°C) 𝑑𝑑𝑥 =𝜍𝑊𝐻ℎ𝑀ρα(ρvs − ρv)m𝑎̇ Teorik olarak bu peltieri boyutlandırmak için geliştirilmiş formülasyon olmasına karşın daha düşük bir hacmi soğutmak nedeniyle belli şartlarda gerçekleşen deneysel sonuçlar kullanılacaktır. Soğutucu ped tasarımı yapılmadan önce piyasada var olan bir soğutucu ped (CELdek) incelenmiş ve tasarımında gerekli olan formülasyonlar belirtilmiştir. Su, bir pompa istasyonundan dolaştırılır ve bir dağıtım pedi aracılığıyla soğutma pedinin tepesine gider. Soğutma pedinin üst kısmındaki dağıtım pedi, eşit bir su dağılımı sağlar. Su, CELdek evaporatif soğutma pedinin oluklu yüzeyinden aşağı doğru akar. Suyun bir kısmı pedden geçen ılık ve kuru hava ile buharlaşır. Suyun geri kalanı pedin yıkanmasına yardımcı olur ve bir oluk sistemi aracılığıyla pompa istasyonuna geri boşaltılır. Yeşil şeritli ped, birbirine yapıştırılmış bir dik (60 °) ve bir düz (30 °) olmak üzere farklı oluk açılarına sahip özel olarak emprenye edilmiş ve oluklu selüloz kağıt tabakalardan oluşur. Bu tasarım, çok düşük bir basınç düşüşünde çalışmaya olanak sağlarken, yüksek buharlaşma verimliliğine sahip bir soğutma pedini yaratır. Ek olarak, ölçekleme minimumda tutulur ve suyun pedin hava giriş tarafına yönlendirilmesi nedeniyle su taşınması olmaz. Burası buharlaşmanın çoğunun gerçekleştiği yerdir. Şekil 5.11. CELdek 7090-15 Soğutucu Pedin Oluk Açıları 46 Buharlaşma için gerekli olan ısı, havanın kendisinden alınır. Bu nedenle pedden çıkan hava, buharlaştırma işlemi için herhangi bir harici enerji kaynağı olmaksızın aynı anda soğutulur ve nemlendirilir. Şekil 5.11.’de, D derinliğine sahip pedin hava akış hızlarına göre doyma verimi ve basınç düşümü gösterilmiştir. Şekil 5.12 CELdek 7090-15 Soğutma Pedinin Havanın Hızlarına Oranla Doyma Verimi ve Basınç Düşüşü Bu doğanın kendi soğutma sürecidir. Soğutucu pedlerde kalınlık ve ön hız arttıkça su tüketim oranı ve soğutma kapasitesinin arttığını, kalınlığın artması ve ön hızın düşürülmesi ile çalışabilirlik oranının arttığını görülmüştür. Öte yandan kalınlık ve ön hız arttıkça performans katsayısı azalmaktadır. 47 Şekil 5.13 Soğutucu ped bilgileri (Munters AB, 2011) Soğutucu pedin soğutma kapasitesi, giriş ve çıkıştaki sıcaklık farkı ile hesaplanır: 𝑃𝑝𝑒𝑑 𝑠𝑜ğ𝑢𝑡𝑚𝑎𝑠𝚤 = ṁℎ𝑎𝑣𝑎𝐶𝑝(𝑇ç𝚤𝑘𝚤ş − 𝑇𝑔𝑖𝑟𝑖ş) (5.1) ṁℎ𝑎𝑣𝑎 = ρ. V. L . H (5.2) Burada: 𝑃 (𝑘𝑊): Soğutma kapasitesi ṁ ℎ𝑎𝑣𝑎(𝑘𝑔𝑠 ): Havanın akış hızıdır 𝐶𝑝(Jg°C): Sabit basınçta havanın özgül ısısı (1,005) 𝑇𝑥 (°C): Havanın çıkış ve giriş sıcaklığı ρ (kgm3): Havanın yoğunluğu 𝑉 (m/s): Hava akış hızı 𝐿 (𝑚): Pedin genişliği 𝐻 (𝑚): Pedin yüksekliği göstermektedir. Soğutma etkisinin elektrik enerjisi tüketimine oranına eşit olan performans katsayısı aşağıdaki denklem ile hesaplanır: 𝐶𝑂𝑃 =𝑃𝑝𝑒𝑑𝑃𝑓𝑎𝑛 + 𝑃𝑝𝑜𝑚𝑝𝑎 (5.3) 48 Soğutulacak mahalin saatlik ısı kazancını 2,37 kW olarak hesaplamıştık. Emniyetli olması için 2,5 kW olarak kabulü ile işlemlere devam edersek: 𝑃𝑝𝑒𝑑 𝑠𝑜ğ𝑢𝑡𝑚𝑎𝑠𝚤 = 2,5 kW olması için havanın kütlesel debisi : 𝑃𝑝𝑒𝑑 𝑠𝑜ğ𝑢𝑡𝑚𝑎𝑠𝚤 = ṁ ℎ𝑎𝑣𝑎𝐶𝑝(𝑇ç𝚤𝑘𝚤ş − 𝑇𝑔𝑖𝑟𝑖ş) 2,5 𝑘𝑊 = ṁ ℎ𝑎𝑣𝑎(𝑘𝑔𝑠 ).1,005(kJkg °C). (38 − 23)(°C) ṁ ℎ𝑎𝑣𝑎 = 0,166 𝑘𝑔 𝑠⁄ olarak hesaplanır. Havanın kütlesel debisinin yoğunluğa, havadan pede aktarılan havanın hızına ve ısı transferi yapılan pedin dik yüzey alanına bağlı olduğu belirtmiştik. O halde: ṁℎ𝑎𝑣𝑎 = ρ. V. W. H 0,166 (𝑘𝑔 𝑠⁄ ) = 0,858(kgm3) . 1,5(ms ). W(m). H(m) W(m). H(m) = Isı Transferi Dik Pedin Y. A. (m2) = 0,1289 m2 = 1289 cm2 Yapılan işlemler sonucunda pedin yüksekliği 45 cm, genişliği 29 cm seçilmiştir. Şekil 5.14 Evaporatif Soğutucu Ped Boyutlandırması Hesaplarımızı Malatya’da bir otel odasına göre yapmıştık. Malatya’nın psikrometrik özellikleri 38 °C KT, 21 °C YT sıcaklığındadır (MMO, 1999). Havanın 38 °C KT sıcaklığından 23 °C KT sıcaklığına düşerken ki gizli duyulur ısısındaki değişim miktarı: 49 Denklem 3.7’den 𝑄𝑔𝑖𝑧𝑙𝑖 = 𝑉1𝑣 . (𝑥𝐵 − 𝑥𝐴) . 2501 𝑄𝑔𝑖𝑧𝑙𝑖 = 3500 𝑚33600.10,894 𝑘𝑔/𝑚3. (0,015 − 0,0086). 𝑘𝑔 𝑠. 𝑏.𝑘𝑔 𝑘. ℎ. .2501 𝑘𝐽/𝑘𝑔 𝑄𝑔𝑖𝑧𝑙𝑖 = 17,41 kW Denklem 3.8’den 𝑄𝑑𝑢𝑦𝑢𝑙𝑢𝑟 = 𝑉1𝑣 . 𝐶𝑝. (𝑇𝐴 − 𝑇𝐵) 𝑄𝑑𝑢𝑦𝑢𝑙𝑢𝑟 = 3500 𝑚33600.10,894𝑘𝑔/𝑚3 .1,005. (38 − 23)𝑘𝐽/𝑘𝑔°𝐶 𝑄𝑑𝑢𝑦𝑢𝑙𝑢𝑟 = 16,39 𝑘𝑊 𝑄𝑡𝑜𝑝𝑙𝑎𝑚 = 33,80 𝑘𝑊 5.4. Peltier Su Soğutucu Dolap Sistem Dizaynı: 5.4.1 Soğutulan Dolabın Malzeme Seçimi: Suyu soğutacağımız hazne seçimi yaparken ısı yalıtımı iyi olması için ısı iletimi az olan malzemeler kullanılacaktır . Bu malzemeler içten dışa doğru: ABS(3mm),Poliüretan(20mm),ABS(3mm) olarak belirlenmiştir. Şekil 5.15 Dolap Yüzey Kesiti ve Numaralarıyla Yüzeyler 1,2,4,5 yüzeylerinin arasında bir fark yoktur. Çünkü bu yüzeylerde Termoelektrik Soğutucu yok ve yüzeylerin alanları eşittir. Bu yüzeyler için ısı transferi aynı olacaktır fakat 3 yüzeyi farklıdır. Taşınım ve iletim ile ilgili olan ısı kayıpları toplanarak toplam ısı kaybı değeri elde edilmiştir. 5.4.2 Dolabın Tüm Yüzeylerinden Toplam Isı Transferi: 50 Dolabın yüzeylerinden dışarıya gerçekleşecek olan ısı transferini hesaplayabilmemiz için. taşınım ve iletim katsayısı kabülleri yapılmış, ısı yalıtım malzemelerinin iletim katsayıları belirlenmiştir. O halde 6 yüzeyden olan toplam ısı transferi: 𝑅1−2−4−5−6 = (1ℎ𝑑𝚤ş +𝐿𝑎𝑏𝑠𝑘𝐴𝐵𝑆 +𝐿𝑃𝑈𝑘𝑃𝑈 +𝐿𝑎𝑏𝑠𝑘𝐴𝐵𝑆) = ( 10,08 + 0,0030,17 + 0,020,024 + 0,0030,17 ) = 13,36 𝑊𝑚2𝐾⁄ Fourier ısı iletim kanunu, kesit alanı A, uzunluğu L olan bir ısıl temas yüzeyindeki ısı transfer miktarını gösteren formüldür. Sıcaklık farklı olan ısıl temas halindeki iki katı cisim arasında denge sıcaklığı sağlanana kadar sıcak cisimden soğuk cisime doğru iletim yoluyla ısı yayılımı olur. Taşınım, bir katı yüzey ile ona bitişik, hareket halindeki akışkan (sıvı ya da gaz) arasında ısı transfer türüdür. 𝑄𝑡𝑎ş𝚤𝑛𝚤𝑚 = ℎ . 𝐴 . 𝛥𝑇 (5.4) 𝑄𝑖𝑙𝑒𝑡𝑖𝑚 = 𝑘 . 𝐴. 𝛥𝑇. 1𝐿 (5.5) 𝑄1 = 13,36 ( 𝑊𝑚2. 𝐾) . (0,15𝑥0,075)(𝑚2). (23 − 8)(°𝐶) = 2,2𝑊 𝑄2 = 13,36 ( 𝑊𝑚2. 𝐾) . (0,15𝑥0,075)(𝑚2). (23 − 8)(°𝐶) = 2,2 𝑊 𝑄4 = 13,36 ( 𝑊𝑚2. 𝐾) . (0,15𝑥0,075)(𝑚2). (23 − 8)(°𝐶) = 2,2 𝑊 𝑄5 = 13,36( 𝑊𝑚2. 𝐾). (0,15𝑥0,075)(𝑚2). (23 − 8)(°𝐶) = 2,2𝑊 𝑄6 = 13,36 ( 𝑊𝑚2. 𝐾) . (0,15𝑥0,15)(𝑚2).(23 − 8)(°𝐶) = 4,5 𝑊 3 nolu yüzeyde peltierler olacağı için o yüzeyden bir ısı kaybı olmayacaktır. Bu yüzden hesaplara dahil edilmemiştir. 𝑄𝑡𝑜𝑝𝑙𝑎𝑚 𝑦 = 𝑄1 + 𝑄2 + 𝑄4 + 𝑄5 + 𝑄6 𝑄𝑇𝑜𝑝𝑙𝑎𝑚 = 13,3 𝑊 51 Şekil 5.16 Su Deposu Açık Hali Burada: 𝑅𝑥(𝑊 𝑚2𝐾⁄): Isı iletim direnci 𝑄𝑥 (𝑊): Toplam ısı transferi 𝐴𝑦ü𝑧𝑒𝑦 (𝑚2): Isı transferine dik yüzey alanı 𝐿𝑎𝑏𝑠 (𝑚): Dolabın ABS malzemesinin kalınlığı 𝑘𝐴𝐵𝑆(𝑊 𝑚°𝐶⁄): ABS malzemesinin ısı iletim katsayısı (0,17) 𝐿𝑃𝑈 (𝑚):Dolabın PU köpük malzemesinin kalınlığı 𝑘𝑃𝑈(𝑊 𝑚°𝐶⁄): PU köpük malzemesinin ısı iletim katsayısı (0,024) ℎ𝑑𝚤ş(𝑊 𝑚2°𝐶⁄): Dış yüzeyin ortalama taşınım katsayısını göstermektedir. Bizim amacımız su soğutmak olduğu için ilk olarak su kabının kendi başına kaybettiği ısı miktarını ve suyun soğumak için ne kadar enerjiye ihtiyacı olduğunu hesaplamamız gerekiyor.Bu veriler sonucunda peltier tipini ve adetini belirleyebiliriz. Su kabının kendi kaybettiği ısı miktarını 13,3 W olarak daha öncen belirlemiştik. İçme suyunun 8oC’de olması için ne kadar soğutma gücüne ihtiyacı var hesabı yapılacaktır: Su kabımızın boyutları 1500 mm x 150 mm x 75 mm olarak belirlemiştik. Buradan iç hacmin hesaplarak için kaç lt su alabileceğine göreceğiz; 𝑉 = 0,15𝑥0,15𝑥0,075 = 0,0016 𝑚3 V = Su tankımızın hacmi 52 0,0016m3 =1,6 lt su alabilecen bir kabımız bulunmaktadır. Suyun yoğunluğunu 1000 𝑘𝑔𝑚3 olduğu için kabımızın içindeki toplam su kütlesi 1.6 kg olur. 1,6 kg suyun 1 saate 28oC sıcaklıktan 8oC sıcaklığa soğutulması için ihtiyaç duyduğu ısı miktarı hesaplancaktır. Q = m . c . (T2 – T1) Q = İhtiyaç duyulan ısı miktarı (kW) 𝑚̇ = Suyun kütlesi (𝑘𝑔𝑠 ) c = Suyun özgül ısısı (𝑘𝐽𝑘𝑔.𝐶) T2-T1 =Suyun istenilen sıcaklığı ve olan sıcaklığı (K) Q = 1,63600(𝑘𝑔 𝑠⁄ ) . 4,18 (𝑘𝑗𝑘𝑔. 𝐾⁄). (28 − 10)(𝐾) 𝑄 = 0,037 𝑘𝑊 = 37 𝑊 suyu soğutmak için gerekli enerji miktarı olarak hesaplanır. Bu işlemler sonucunda s hesapladık.Hesaplarımız sonucu peltierden almamız gereken toplam enerji miktarını hesaplayabiliriz. 𝑄𝑡𝑜𝑝𝑙𝑎𝑚 = 𝑄 𝑠𝑢𝑦𝑢𝑛 𝑖ℎ𝑡𝑖𝑦𝑎𝑐𝚤 + 𝑄𝑘𝑎𝑝 𝑄 = 37 + 13,3 = 50,3𝑊 Bizim kabımızın içindeki suyu soğutmak için gereken enerji 50,3 W bu enerjiyi peltierden sağlayacağımız için peltier seçimi ve peltier adetini belirleyeceğiz. 5.4.3 Peltier Tipi ve Adet Seçimi: Peltier seçimi yapılırken ilk olarak peltier tipi belirlenmesi gerekir. Seçilen peltierin soğutma gücünü belirlemek için grafiklerden yararlanılır.Grafikte dikey kısımda ortam sıcaklığı ile peltier sıcak yüzeyinin sıcaklık farkı ve soğutmak için gerekn güç değerleri bulunmaktadır.Grafiğin yatay kısmında ise seçtiğimiz peltierin o gücü verebilmesi için ihtiyaç duyduğu akımı belirlemiş oluruz. Grafiklerin bağlı olduğu tek şey ortam sıcaklığı ile peltier sıcak yüzeyinin sıcaklık farkıdır. Bu sıcaklık farkı peltierin soğutma gücü ile doğru orantılıdır. 53 Peltier seçimini ve adetini tasarlayacağımız cihaz için yapalım. Peltierlerimizi ‘’ II-IV ’’ marka PL020-4-30 kodlu peltierimizi belirledik. Bu peltierden bizim için kaç adet yeterli olacak bunu grafikten okuyarak belirleyeceğiz. Şekil 5.17 Voltaj-Akım Grafiği (url:2) Peltier adetimizi seçerken bu grafiklerden üst kısımda olanı kullanacağız. Üst kısımdaki inceleğimiz grafike ΔT miktarı peltierin sıcak yüzeyi ile ortam sıcaklığı arasındaki farkıdır. Bu farka bakılarak tam karşısında bulunan soğutmak için verdiği enerji miktarına bulabiliriz.Gerekli enerji miktarının belirledikten sonra grafiğin en altından peltierimizin bize istenen performansını vermesi için ona vermemiz gerek akım miktarını belirlemiş oluruz. Tasarımımızda 50,3 W enerjiye ihtiyaç duyduğumuzu hesaplamıştık. 54 Şekil 5.18 Voltaj-Akım Grafiği (url:2) Bizim sıcaklık farkımız 11°C olduğu için 1 adet peltiere 4.0 amper akım vererek 17.50 W soğutma gücü elde elebiliyoruz. Bu grafik okuma işlemi sonucunda bizim için 3 adet peltierin yeterli olacağını görüyoruz. 6. SONUÇ VE ÖNERİLER: Bu çalışmamızda evaporatif soğutma sistemi ile termoelektrik soğutucu sistem kombine olarak dizayn edilmiş, termoelektrik soğutucunun verimi artması için fan veya bir kanatçık yardımıyla soğutmak yerine bir su bloğu konularak evaporatif soğutucu için kullanılan su sıcak yüzeyinde gezdirilmiştir. Böylece tek bir pompa vasıtasıyla hem termoelektrik soğutucu soğutulmuş hem de evaporatif soğutma yapılmıştır. Kombine tasarım sonucunda aynı elektrik kaynağından güç alarak çalışacağı tasarlanan cihaz diğer geleneksel soğutma yöntemlerine göre daha uygundur. Tek fan ve bir pompa yardımıyla havayı soğutan cihaz, termoelektrik modüller sayesinde soğuk içme suyu üretmiştir. Tasarlanan cihazın 25 m2 üzerindeki bir hacim için pek verimli olacağı söylenememektedir. 55 7. KAYNAKLAR: Ioffe, A F.,“Semiconductor thermoelements, andThermoelectriccooling”. (1957). Rowe, D. M., Bhandari C. M.,.“ModemThermoelectrics. Holt-Technology”, 168s, London,(1983). Godfrey, S., “An Introduction To Thermoelectric Coolers: Electronics Cooling” (1996). Yılmaz, T.,“Su veya buz depolama yöntemleriyle iklimlendirme”. III. Ulusal Tesisat Mühendisliği Kongresi. 137-149. (1997). Park, Brian V. (Austin, TX), Mcgrath, Ralph D. (Granville, OH)., “Enclosureforthermoelectricrefrigeratorandmethod”.No:5605047 dated 02.25.1997. Lau, P.G., Buist, R.J.,“Calculation Of Thermoelectric Power Generation Performance Usingfinite Element Analysis. Proceedings of the XVI Intemational Conference on Thermoelectrics“, 563 –566,(1997). Bilge, D. ve Bilge, M.,“İndirek ve direk evaporatif soğutma sistemleri kombinasyonu”. IV. Ulusal Tesisat Mühendisliği Kongresi. 197-204. (1999). Özkol, N., “Uygulamalı Soğutma Tekniği”, MMO: Yayın No:115(1999). Yılmaz, T., Bulut, H.,“Türkiye için yeni dış ortam sıcaklık tasarım değerleri”, (2004). Demirci, M., “Psikometri Nemli Havanın Termodinamiği”, (2004). Doğan, H.,“Uygulamalı havalandırma ve iklimlendirme tekniği”,Ankara : Seçkin,(2006). Yavuz, A. H., Ahıska, R., Hakim, M.,“Bulanık Mantık Kontrollü Termoelektrik Beyin Soğutucusu”, (2006). Ahıska, R., Dişlitaş S.”Microcontroller Based Thermoelectric Generator Application. G.U. Joumal of Science”.135-141.(2006). Ahıska, R., Ahıska, K., “Esnek İki Fazlı TE CPU Soğutucusu. Gazi Univ. Müh. Mim. Fak. Dergisi”.347-351.(2006) Heidarinejad, G., Bozorgmehr, M., Delfani, S., Esmaeelian, J., “Experimental investigation of two-stage indirect/direct evaporative cooling system in various climatic conditions”, (2007). Berber,T.S.,“Kendinden termal elektrik kaynaklı mikroişlemci soğutma sistemi”,(2008). Doğan, V.,“Isıtma”, (2010). 56 Çalışır, T., Alptekin, M., Yılmazoğlu M.Z., “Bir Direkt Evaporatif Soğutma Sisteminin Deneysel Ekonomik ve Çevresel İncelemesi” Ulusal İklimlendirme Kongresi (2011). Tosun, A.,“Termoelektrik Modül Kullanarak Soğutma Kabı Tasarımı Ve Deneysel Çalışması”,(2014). Gang Tan, D,.Z,.” A review of thermoelectriccooling: Materials, modelingandapplications”. (2014). Sevel, M.H., “Psikometri-2”, pp.10-73, (2014). Yılmazoğlu, Z. M., “Experimentalandnumericalinvestigation of a prototypethermoelectricheatingandcoolingunit”. (2015). Thakkar,M. P.“Peltier Cooling Module”, (2016). İdiz, A., Koçak, C., Özdemir, F., Akdemir, Ö., Güngör, A.,“İklimlendirme sistemlerinde evaporatif soğutma uygulamaları”.XIII. Ulusal Tesisat Mühendisliği Kongresi. 1761-1776.(2017) Ibañez-Puy, M., Bermejo-Busto, J.,Martín-Gómez, C.,Vidaurre-Arbizu, M., Sacristán-Fernández, J., A., “Thermoelectriccoolingheatingunitperformanceunderrealconditions”.(2017). Caia ,Y., Wangd,Y., Liud, D., Zhaoa, F-Y., “Thermoelectric cooling technology applied in thefield of electronic devices:Updated review on the parametric investigation sand model developments”. (2018). Zhou, Y., Zhang,T., Wang, F.,Yu, Y.,“Performance analysis of a novel thermoelectric assisted indirect evaporative cooling system”, (2018). Zheng, L., J., Kang, D., H., Kim, K.,N.,You, J., Y., Kang, H., W.,“Theoreticalanalysis of naturalevaporativecoolingtoenhancetheefficiency of thermoelectricdevices”. (2018). Köysal, Y., “Termoelektrik Sistemler” Ondokuz Mayıs Üniversitesi, (2018). Çelebi, T., Bedir, H., Sönmez, O,.“Evaporatif soğutucu tasarımı”.XIV Ulusal Tesisat Mühendisliği Kongresi. pp.349-354.(2019). Sun, D.,Shen L.,Chen H.,Jiang, B.,Jie, D.,Liu, H.,You, Y.,Tang, J.“Modeling and analysis of the influence of Thomson effect on micro-thermoelectric coolers considering interfacial and size effects” ,(2020) Mitsubishi Electric Corporation.,“Thermoelectric Cooler Built-in Stem”, No:20200313387 dated 01.10.2020. 57 Wang, R., Luo, D., Yu, Wei., Zhou, Weiqi.,“Parametric study of a thermoelectric module used for both power generation and cooling”. (2020). Obando, F., A., Montaya, P., A., Osorio, A., J., Damasceno, A., F., Norton, T.,“Evaporative pad cooling model validation in a closed dairy cattle building”. (2020). Url :1“https://www.z-max.jp/peltier_turk/experience/experience2.html” Url :2“https://ii-vi.com/wp-content/uploads/2020/07/PL054-6-40.pdf” Url :3”www.imder.org.tr/dosyalar/File/pptler/69 56.ppt” View publication stats | RESEARCH GATE |
1 C. elegans episodic swimming is driven by multifractal kinetics 1 Author Affiliation 2 Yusaku Ikeda1,2,†, Peter Jurica1,†, Hiroshi Kimura2, Hiroaki Takagi3, Struzik Zbigniew4,5,6, Ken Kiyono7, 3 Yukinobu Arata1, *, and Yasushi Sako1 4 1. Cellular Informatics Laboratory, RIKEN, 2-1 Hirosawa, Wako 351-0198, Saitama, Japan 5 2. Department of Mechanical Engineering, School of Engineering, Tokai University, 4-1-1 Kitakaname, 6 Hiratsuka, Kanagawa 259-1292, Japan 7 3. Department of Physics, Nara Medical University, 840 Shijocho, Kashihara, Nara 634-8521, Japan. 8 4. Graduate School of Education, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan 9 5. Advanced Center for Computing and Communication, RIKEN, 2-1 Hirosawa, Wako 351-0198, Saitama, 10 Japan 11 6. Faculty of Physics, University of Warsaw, Pasteur 5, 02-093 Warsaw, Poland 12 7. Graduate School of Engineering Science, Osaka University, 1-3 Machikaneyama-cho, Toyonaka, Osaka 560-13 8531, Japan. 14 † these authors contributed equally 15 * To whom correspondence may be addressed. Email arata@riken.jp 16 Author contributions 17 Conceptualization: Yukinobu Arata and Hiroaki Takagi 18 Formal analysis: Peter Jurica and Yukinobu Arata, 19 Funding acquisition: Yukinobu Arata and Yasushi Sako 20 Investigation: Yusaku Ikeda, Yukinobu Arata, and Peter Jurica 21 Methodology: Yusaku Ikeda and Hiroshi Kimura 22 Writing: original draft: Yukinobu Arata 23 Writing – review & editing: Hiroaki Takagi, Struzik Zbigniew, Ken Kiyono, and Yasushi Sako 24 25 The authors declare no conflict of interest. 26 This article contains supporting information online at 27 This open access article is distributed under Creative Commons Attribution License (CC BY). 28 29 Data deposition 30 .CC-BY-NC-ND 4.0 International licensewas not certified by peer review) is the author/funder. It is made available under aThe copyright holder for this preprint (whichthis version posted April 24, 2020. . https://doi.org/10.1101/2020.04.22.056606doi: bioRxiv preprint 2 The C. elegans swimming activity time series and movie data reported in this paper have been deposited in 31 Systems Science of Biological Dynamics (SSBD) database 1, http://ssbd.qbic.riken.jp/set/20190401/ 32 Keywords 33 microfluidic device, cGMP-dependent kinase (PKG), scale-free, power law, multifractal, temporal correlation 34 35 Abstract
36 Fractal scaling is a common property of temporal change in various modes of animal behavior. The molecular 37 mechanisms of fractal scaling in animal behaviors remain largely unexplored. The nematode C. elegans alternates 38 between swimming and resting states in a liquid solution. Here, we report that C. elegans episodic swimming is 39 characterized by scale-free kinetics with long-range temporal correlation and local temporal clusterization, which 40 is characterized as multifractal kinetics. Residence times in actively-moving and inactive states were distributed 41 in a power law-based scale-free manner. Multifractal analysis showed that temporal correlation and temporal 42 clusterization were distinct between the actively-moving state and the inactive state. These results indicate that C. 43 elegans episodic swimming is driven by transition between two behavioral states, in which each of two transition 44 kinetics follows distinct multifractal kinetics. We found that a conserved behavioral modulator, cyclic GMP 45 dependent kinase (PKG) may regulate the multifractal kinetics underlying an animal behavior. Our combinatorial 46 analysis approach involving molecular genetics and kinetics provides a platform for the molecular dissection of 47 the fractal nature of physiological and behavioral phenomena. 48 49 Abbreviations 50 cGMP-dependent kinase (PKG) 51 .CC-BY-NC-ND 4.0 International licensewas not certified by peer review) is the author/funder. It is made available under aThe copyright holder for this preprint (whichthis version posted April 24, 2020. . https://doi.org/10.1101/2020.04.22.056606doi: bioRxiv preprint 3 Introduction
52 Animal behaviors are organized over a broad range of time scales, ranging from seconds to years, including 53 expansive timescales over lifespan phases, such as infant, juvenile, adult, and elderly phases. Among the 54 temporally organized animal behaviors are rhythmic behaviors characterized by their frequencies. Molecular 55 regulators e.g. for daily rhythmic and oscillatory changes of animal behaviors i.e. for circadian rhythm have been 56 identified and are shown to be regulated by a feedback control 2,3. Contrastingly, many arrhythmic changes of 57 behavioral and physiological activities in a great variety of animal species, including humans are reported to show 58 self-similar and scale-free structures, which is an indicative for fractal scaling 4-6. Fractal geometry provides a 59 mathematical framework for characterizing scale-free and self-similar patterns 4-6. Fractal temporal patterns of 60 behavior and physiology have been reported in the locomotion of birds, mosquito larvae, and flies 7-9, crawling 61 of cultured C. elegans worms 10, clicking sounds produced by feeding sea horses 11, and swimming of zooplankton 62 12,13. In humans, temporal fractal patterns have been observed in wrist movements during habitual sleep/wake 63 cycles 14, gait 15,16, heartbeats 17,18, and brain activity 19,20. Altered fractal patterns of activity have been associated 64 with pathological conditions and aging 14-16,18,21,22, indicating that the fractal patterning in biological activity is a 65 critical measure for characterizing physiological status. Thus, fractal scaling is a clue to study molecular and 66 physiological basis of arrhythmic and complex animal behaviors, and there may be molecular and physiological 67 mechanisms that adhere to the basic principles of fractal geometry. However, the molecular and physiological 68 mechanisms remain unknown. 69 Owing to its simple anatomy and the availability of a range of genetic tools, C. elegans is a powerful 70 model organism for the study of the molecular bases of behavior. In solution, C. elegans worms alternate between 71 swimming and resting states on a minute to hour time scale 23,24. In the swimming state, they alternate between 72 continuously beating their bodies and resting with or without food 23-25. In the resting state, they maintain a 73 characteristic sharply-bent posture 23. Episodic swimming is conserved in nematodes cultured in a liquid solution 74 23. On a solid agar plate, C. elegans also move in an episodic manner, wherein they crawl actively and persistently 75 .CC-BY-NC-ND 4.0 International licensewas not certified by peer review) is the author/funder. It is made available under aThe copyright holder for this preprint (whichthis version posted April 24, 2020. . https://doi.org/10.1101/2020.04.22.056606doi: bioRxiv preprint 4 in one direction or crawl slowly and stay within a small area, behavioral states called roaming and 76 dwelling/quiescence, respectively 24,26,27. Individual Drosophila flies 9 and Leptothorax allardycei worker ants 28 77 also alternate between an actively-moving state and a resting state in an episodic manner. Thus, episodic behavior 78 is a conserved presentation in invertebrates that is thought to be adaptive for supporting food exploration, energy 79 conservation, and reproductive success 9,23,24,29. 80 In C. elegans, episodic swimming is regulated by egl-4, which encodes cGMP-dependent kinase (PKG). 81 PKG is a behavior modulating enzyme conserved across invertebrates and vertebrates 30,31. C. elegans egl-4/pkg 82 mutants roam continuously on a solid agar plate and swim when in a solution with less frequent resting than is 83 exhibited by wild-type animals 23,25-27. Because egl-4/pkg-dependent regulation is found in both medium 84 conditions 23,25-27, C. elegans episodic motions in both conditions are thought to be regulated by the same 85 molecular/physiological mechanism 24. In Drosophila, the foraging/pkg homolog of egl-4 had been discovered 86 from natural behavioral polymorphisms wherein flies tend to travel long distances (rover) or be sedentary (sitter) 87 32,33. The expression level of foraging/pkg differs between rover flies and sitter flies and Drosophila traveling 88 behavior can be switched by genetic manipulation of foraging/pkg 34. Foraging behaviors in social insects—89 including those exhibited by honey bees (Apis mellifera) and ants (Pheidole pallidula) that are determined by 90 developmental stage 35 and social caste 36, respectively—are also associated with PKG expression and activity. 91 Thus, PKG is considered to be a conserved modulator of animal behavior 30,31. 92 To study scaling of behavior across a broad range of time scales, it is necessary to obtain behavioral 93 activity time series of individual animals for an extended period of time at a high temporal resolution. In this 94 study, we recorded the swimming behavior of 108 individual C. elegans at a semi-video rate for a week-long 95 period by individually culturing the worms in a newly developed microfluidic device. The obtained time series 96 data encompassing 107 time points was submitted to multifractal analysis. We found that C. elegans episodic 97 swimming is a scale-free process and the transition between actively-moving and inactive states is driven in 98 multifractal kinetic. Additionally, we examined the sensitivity of the multifractal kinetics of the behavior to 99 .CC-BY-NC-ND 4.0 International licensewas not certified by peer review) is the author/funder. It is made available under aThe copyright holder for this preprint (whichthis version posted April 24, 2020. . https://doi.org/10.1101/2020.04.22.056606doi: bioRxiv preprint 5 modulation by PKG. A discussion of potential molecular and physiological mechanisms underlying 100 multifractality in behavior is provided on the basis of previously reported mathematical and physical models. 101 102 Results 103 C. elegans episodic swimming is a multi-time scale process 104 Swimming of multiple adult C. elegans individuals was monitored in a newly developed microfluidic device, 105 called WormFlo, which is equipped with an array of 108 disc-shaped chambers for culturing individual C. elegans 106 (Fig. 1A and B). C. elegans individuals were maintained alone in chambers under controlled chemical, 107 temperature, and light intensity conditions (Fig. S2, Fig. S3, and Methods). We recorded swimming at 107 time 108 points with a 50-ms interval (138 h ≈ 5.8 d), and quantified swimming activity with a pixel counting method (Fig. 109 2A, Fig. S4, Movie S1, Methods, and Supplemental Information). In the absence of an energy source, the animals’ 110 swimming activity diminished gradually over time (Fig. 2B and Movie S2). The swimming activity in WormFlo 111 chambers decayed with kinetics similar to that seen in animals cultured without an energy source in substantially 112 larger (2× diameter, 23× volume) 96-well plate wells (Movie S3); and the activity was sustained in animals 113 cultured in WormFlo with an energy source (glucose and cholesterol) (Fig. S5A and Movie S4). Therefore, the 114 activity decay observed can be attributed primarily to a physiological response to long-term cultivation without 115 an energy source rather than to physical damage or spatial restriction. 116 Comparing the activity decay kinetics over the 6-d observation period among individuals led us to 117 define three empirical activity classes: High, long-term high activity; Middle, intermediate-term high activity; 118 and Low, short-term high activity (Fig. 2B). The early and late culturing periods were defined as Pre-starved and 119 Starved time regimes, each of which has high and low swimming activities (for quantitative criteria, see Fig. 2B 120 legend, and Methods). Consistent with previous studies, we observed episodic swimming bouts on minute to hour 121 time scales (Fig. 2C; 8-min scale and 80-min scale), and a characteristic kinked posture during the resting state 122 in the Pre-starved regime (data not shown) 23,24. 123 .CC-BY-NC-ND 4.0 International licensewas not certified by peer review) is the author/funder. It is made available under aThe copyright holder for this preprint (whichthis version posted April 24, 2020. . https://doi.org/10.1101/2020.04.22.056606doi: bioRxiv preprint 6 Interestingly, we observed a swimming bout cluster on a 1 day-scale that could be divided in several 124 clusters in a magnified view of about 1 h (Fig. 2C, compare the red bracketed region in the 14-h scale vs. red 125 arrows in the 80-min scale). This nested temporal structure was observed repeatedly over a series of 126 magnifications from a 1-h scale to a 10-min scale (Fig. 2C, red region in 80-min scale vs. red arrows in 8-min 127 scale), and between the 1-min and 1-s scales (Fig. 2C, red bracket region in 8-min scale vs. red arrows in 50-s 128 and 5-s scales). In the 1-s scale, animals alternated between bending their bodies and beating their bodies for 129 swimming (Movie S1). Series of beating motions were interrupted with intermittent short resting periods, referred 130 as to “posing” (Movie S1). During posing, they were transiently motionless in a bent posture (Movie S1, see 131 legend). The temporal resolution at which continuous shape could be detected in H265 codec-compressed movies 132 was 0.25 s (Fig. S6 and Supplemental Information), which was sufficient to detect subsecond “posing” on a 1-s 133 scale. The temporally nested structure of C. elegans episodic motion was observed on time scales of a 134 magnification range of about 1,000 times. Thus, C. elegans episodic swimming has multi-timescale dynamics 135 with a self-similar temporal structure. 136 137 A scale-free property in C. elegans episodic swimming 138 Continuous swimming ceased and resumed suddenly (Fig. 2C). Consistently, we observed a bimodal probability 139 distribution of activity strength (Fig. 3A), indicating that C. elegans episodic swimming can be characterized by 140 a two-state transition model between actively-moving and inactive states. To reveal the kinetics of the state 141 transition, we studied the statistical distribution of residence times in active and inactive states. Active and inactive 142 states were defined as periods above or below an activity threshold, which was the value in the bimodal 143 distribution valley (Fig. 3A) of swimming activity time series (red horizontal line in Fig. 2C, 5-s scale). Residence 144 time in the two states were obtained alternately from swimming activity time series; eventually, we obtained 145 round series data for active and inactive states (Fig. 3B and C). Residence times in the two states were distributed 146 .CC-BY-NC-ND 4.0 International licensewas not certified by peer review) is the author/funder. It is made available under aThe copyright holder for this preprint (whichthis version posted April 24, 2020. . https://doi.org/10.1101/2020.04.22.056606doi: bioRxiv preprint 7 along a broad temporal range, ranging from subsecond to 10-s orders for the active state and ranging from 147 subsecond to 100-s orders for the inactive state (Fig. 3B and C). 148 The probability distributions of residence times in active and inactive states were linear lines in log-log 149 plots or power law distributions (Fig. 3D-F, and Table 1, fit parameter), indicating that the state residence times 150 lack a specific time scale and thus can be described as exhibiting a scale-free property. Comparing the power law 151 exponents between the two states in each activity class in Pre-starved and Starved time regime, we find that the 152 power law relationship slopes are significantly shallower for the inactive state than for the active state (a “shallow 153 slope” power law relationship indicates a relatively even appearance of longer and shorter residence times, Fig. 154 3D–F, Table 1 fit parameters and p values in Table 1). These results indicate that the mechanisms that regulate 155 the transition from the active to the inactive state and from the inactive to the active state have distinct scale-free 156 kinetics. 157 158 Multifractal analysis of numerically-generated round series data 159 Long and short bursts of state residence times in the round series data were clustered rather than random (Fig. 3B 160 and C). We employed multifractal detrended fluctuation analysis (MF-DFA, Methods) to study the temporal 161 structure of the residence-time round series data 37,38. First, time series of white, pink, and Brown noises were 162 numerically-generated by R software tools for fractional Gaussian noise generation 39,40, and were subjected to 163 MF-DFA. White noise time series is completely uncorrelated, whereas pink and Brown noise time series have 164 long-range correlation. Noise components of cumulatively summed original time series [𝑓(𝑣, 𝑠) in Eq. 1] was 165 estimated in MF-DFA by removing local trends in each scale, s, by linear fitting to various time scales (Methods). 166 As the temporal correlation in the time series became longer, the noise component (i.e. deviation from the trend 167 of cumulative sums of deviations from the average) became smoother (Fig. 4A; upper three series). The calculated 168 average noise component [𝐹(𝑞, 𝑠), Eq. 1] was plotted with temporal resolution for observation in s. In 𝐹(𝑞, 𝑠) 169 versus s plots on a log-log scale, 𝐹(𝑞, 𝑠) increased linearly with increasing scale s (Fig. 4A; left column of lower 170 .CC-BY-NC-ND 4.0 International licensewas not certified by peer review) is the author/funder. It is made available under aThe copyright holder for this preprint (whichthis version posted April 24, 2020. . https://doi.org/10.1101/2020.04.22.056606doi: bioRxiv preprint 8 three graphs), as is indicative of fractal scaling. Additionally, the slope of 𝐹(𝑞, 𝑠) versus s plot on a log-log scale 171 increased with an increasing length of temporal correlation of the original time series (Fig. 4A; left column of 172 lower three graphs). The 𝐹(𝑞, 𝑠) versus s plot gave rise to a multifractal spectrum (Methods and Supplemental 173 Information) showing a relationship between q-order (local) Hurst exponent (Hölder spectrum H(q)) vs. q-order 174 singularity dimension (singularity spectrum D(q)) 37,38. 175 The singularity dimension is a fractal dimension that represents time series sparsity; a D(q) < 1 176 represents sparsely distributed local structures of the entire time series, and a D(q) approaching 1 representing 177 broadening involvement of the time series. Thereby, the Hurst exponent at D(q) < 1 represents a Hurst exponent 178 of local structures of time series, which is the local Hurst (Hölder) exponent, whereas the Hurst exponent at D(q) 179 = 1 represents the Hurst exponent of the entire time series, which we call the global Hurst exponent (Hpeak). The 180 Global Hurst exponent Hpeak corresponds approximately to the conventional Hurst exponent, which is linked 181 theoretically with the scaling exponent of power spectrum density [𝛽, (𝑆(𝑓)~𝑓-.)], a measure of time series 182 autocorrelation [𝛽 = 2𝐻 + 1, (𝐻 > 0)] (Kiyono, 2015). Therefore, the global Hurst exponent Hpeak can be 183 interpreted as an index of behavioral memory in activity time series of behavior. The Hpeak values of the 184 cumulative sum of these noise time series were estimated to be approximately 0.5, 1, and 1.5 (Fig. 4A; right 185 column of lower three graphs), consistent with theoretical Hurst exponents of cumulative sums of these noise 186 time series. 187 Next, multifractal time series were numerically-generated by multiplicative cascading process 41,42, and 188 were subjected to MF-DFA. The multifractal time series were highly clusterized, such that the variety of 189 amplitudes and durations of temporal clusters became richer with greater multifractality (Fig. 4B; upper three 190 series). Because of the scale-free and self-similar properties of multifractal time series, such temporal clusters are 191 distributed in multiple temporal resolutions. As the variety of amplitudes and durations of temporal clusters 192 increased, the 𝐹(𝑞, 𝑠) versus s plot slope in a log-log scale varied widely with the exponentiation factor q (Fig. 193 4B; left column of lower three graphs), which enhances or suppresses the noise component from local trends at 194 .CC-BY-NC-ND 4.0 International licensewas not certified by peer review) is the author/funder. It is made available under aThe copyright holder for this preprint (whichthis version posted April 24, 2020. . https://doi.org/10.1101/2020.04.22.056606doi: bioRxiv preprint 9 scale s [𝑓(𝑣, 𝑠) in Eq. 1, and Methods]. The q-dependent change in slope in a log-log scale 𝐹(𝑞, 𝑠) versus s plot 195 is indicative of multifractal scaling. In accordance with this q-dependent variation, the width of the multifractal 196 spectrum (width) widens significantly without a substantial change in Hpeak (Fig. 4B; right column of lower three 197 graphs). A wide multifractal spectrum indicates that there are locally clustered structures with various local Hurst 198 exponents sparsely distributed throughout the time series. Therefore, multifractal spectrum width in animal 199 activity time series is an index of behavioral complexity. Note that global Hurst exponent Hpeak and multifractal 200 spectrum width are independently changeable in a scale-free time series with a given power law exponent. See 201 the Supplemental Information for detailed descriptions of numerical generation for noise time series and 202 multifractal analysis, and for an introduction into fractal scaling. 203 204 Episodic swimming is driven by two-state transition with two distinct multifractal kinetics 205 Global trend in the round series data of active- and inactive-state residence times became gradually shorter and 206 longer, respectively, after transition from the Pre-starved to the Starved time regime (~15,000 rounds; Fig. 3B 207 and C). The noise components of cumulative sums of the deviations from average residence times appeared to 208 differ qualitatively between active and inactive states; that of active states was smoothly curved, whereas that of 209 inactive states was locally straighter in the Pre-starved regime (Fig. 4C).), suggesting that noise properties of 210 active and inactive states may differ. In the log-log scale 𝐹(𝑞, 𝑠) versus s plot, 𝐹(𝑞, 𝑠) in active and inactive states 211 increased almost linearly with s and the slope varied with q (Fig. 4C; left column of lower two graphs), and the 212 multifractal spectra of active state and inactive states in the Pre-starved time regimes were widely distributed (Fig. 213 4C; right column of lower two graphs), indicating that the round series of active and inactive states had a 214 multifractal nature. 215 The Global Hurst exponents Hpeak in multifractal spectra of active and inactive states were located 216 around H(q) =1, indicating that the state residence times have a non-trivial long-range temporal correlation, as is 217 seen in pink noise time series. As shown in Figure 5A–C, before and after starvation in all activity classes, we 218 .CC-BY-NC-ND 4.0 International licensewas not certified by peer review) is the author/funder. It is made available under aThe copyright holder for this preprint (whichthis version posted April 24, 2020. . https://doi.org/10.1101/2020.04.22.056606doi: bioRxiv preprint 10 observed greater Hpeak values for the inactive state (average, 1.38) than for the active state (average, 0.85) as well 219 as greater width values in the inactive state (average, 1.80) than in the active state (average, 0.64), all p < 0.05). 220 Hence, residence-time round series revealed distinct multifractal properties between active and inactive states, 221 with a longer behavioral memory and greater behavioral complexity for the inactive state than for the active state. 222 We called kinetics that generate scale-free residence times with temporal correlation and temporal clusterization 223 as multifractal kinetics. Accordingly, we were able to explain C. elegans episodic swimming with a two-state 224 transition model in which opposite transitions between actively-moving and inactive states are driven by distinct 225 multifractal kinetics (Fig. 6A). 226 The power law exponent showed a more pronounced alteration between in Pre-starved and Starved time 227 regimes in residence-time round series for the inactive state (average, 22.4%) than for the active state (average, 228 7.0%). The change in the power law exponent was only statistically significant for the inactive state (p values and 229 fit parameters in Table 1). MF-DFA showed that the global Hurst exponent Hpeak was not altered significantly by 230 the transition to the Starved time regime in either the active state (average Pre-starved Hpeak , 0.86; average Starved 231 Hpeak, 0.84) or the inactive state (average Pre-starved Hpeak , 1.34; average Starved Hpeak, 1.42) (Hpeak values, and 232 p values in Table 2, all p > 0.05). However, the multifractal spectrum width was altered after starvation, becoming 233 narrower (-17.1%) in the active state and wider (+31.6%) in the inactive state (width values, and p values in Table 234 2, all p < 0.05 except active state in Low class, see Figure 5 legend). These results indicate that food signaling or 235 metabolic state regulates behavior by modulating multifractal kinetics in response to starvation. That is, in 236 response to food starvation, the animals do not simply decrease swimming activity, but show a selective 237 modulation of multifractal kinetics. 238 239 PKG may regulate the multifractal kinetics of C. elegans episodic swimming 240 For molecular dissection of the multifractal kinetics C. elegans episodic swimming, we studied the egl-4(n479) 241 mutant, in which the temperature sensitive n479 allele of egl-4 is associated with a defect in conventional episodic 242 .CC-BY-NC-ND 4.0 International licensewas not certified by peer review) is the author/funder. It is made available under aThe copyright holder for this preprint (whichthis version posted April 24, 2020. . https://doi.org/10.1101/2020.04.22.056606doi: bioRxiv preprint 11 swimming characterized by lengthened continuous swimming periods and a reduced frequency of resting 23,25. 243 We confirmed that egl-4(n479) mutants had reduced frequencies of a persistent inactive state in our quantitative 244 activity time series (Fig. S5B). The egl-4(n479) mutants cultured in WormFlo without an energy source were too 245 transparent for body detection in the later portion of the 6-d culture period. This transparency was likely caused 246 by the mutants’ unusually early lipid consumption due to their continuous swimming and lack of long resting; 247 lipids in worm bodies scatter illuminating light enabling imaging as a (relatively) dark object as shown in Figure 248 2A. Thus high-contrast images of the first tenth of the 6-d culturing period were used for analysis of the Pre-249 starved time regime of egl-4 mutants. 250 In egl-4 mutants, the power law relationships were maintained in active and inactive states, but the 251 power law exponent was preferentially changed in the inactive state (+21.0%) compared to that in active state 252 (+6.9%) (Fig. 3G and D). Power law exponents differed significantly between wild-type worms and egl-4 mutants 253 only in the inactive state (p-values and fit parameters in Table 1), indicating that the mutant has altered residence 254 times across time scales rather than a defect that is specific to a particular time scale. MF-DFA showed that the 255 shape of the multifractal spectrum in egl-4 mutants was largely altered in the inactive state (Figs. 4D and 5D). 256 Comparing the global Hurst exponent between egl-4 mutants and High-activity class wild-type worms, we found 257 that active-state Hpeak was similar across the groups (difference, 1.1%), whereas inactive-state Hpeak differed 258 significantly between the groups (difference, 18.7%). Meanwhile, spectrum width in egl-4 mutants was reduced 259 significantly relative to that obtained for the High activity wild-type group in both active (-20.9%) and inactive 260 (-14.1%) states (Hpeak and width values, p values in Table 2). Thus, egl-4 mutants showed a defect in both temporal 261 memory and temporal clusterization in the inactive state, but showed a defect in only temporal clusterization in 262 the active state. These results are consistent with the possibility that EGL-4/PKG may regulate the multifractal 263 kinetics of behavior, and, more specifically, suggest that EGL-4/PKG may regulate the active- and inactive-state 264 kinetics of C. elegans episodic swimming differently. 265 266 .CC-BY-NC-ND 4.0 International licensewas not certified by peer review) is the author/funder. It is made available under aThe copyright holder for this preprint (whichthis version posted April 24, 2020. . https://doi.org/10.1101/2020.04.22.056606doi: bioRxiv preprint 12 Discussion 267 C. elegans episodic swimming is driven by a multifractal transition cycle 268 Experimental measurements of temporal changes in biological system variables of interest and the statistical 269 analysis of those data can provide information about the hidden operating principles of that system 43. C. elegans 270 episodic swimming was characterized conventionally by the average residence time of active- and inactive-states 271 23-25,44. Here, we applied fractal analysis to study the molecular and genetic mechanisms regulating animal 272 behavior over a 6-d period recorded at a subsecond temporal resolution, showing that episodic swimming is a 273 scale-free process across a 1000-fold range of time scales (Fig. 2C). Because the experimentally-measured 274 average residence times often fail to be determined by values that greatly exceeded the average times, the scale-275 free process is characterized by a power law exponent in the relationship between appearance frequency and 276 residence time (Supplemental Information). Our multifractal analysis showed long-range memory and local 277 clusterized structures of episodic swimming residence times characterized by a multifractal process. Hence, we 278 referred to the swimming as “multifractal episodic swimming” (Fig. 6B). 279 A scale-free distribution of actively-moving and inactive residence times has been reported for episodic 280 Drosophila behavior 9; modeling suggested that the scale-free nature of these residence times contributes to 281 maximizing the food exploitation area while optimizing food intake time. Previous analyses have shown long-282 range temporal correlations in scale-free properties of the behaviors of many species, including fractal (but not 283 multifractal) analysis of C. elegans crawling 7,8,10-13. The fractal nature of C. elegans crawling in agar (Alves et 284 al., 2017) is consistent with our long-range memory and clusterization finding. In this study, we extended previous 285 findings by combining a two-state transition model with multifractal analysis. Employing multifractal analysis, 286 we found that scale-free actively-moving and inactive states had a long-range memory and complex local 287 temporal structures. Our analysis indicated that C. elegans episodic swimming is characterized by a two-state 288 transition between actively-moving and inactive states, wherein the two transitions are driven by distinct 289 multifractal kinetics. That is, the active-to-inactive transition is driven by a narrow multifractal kinetics (short 290 .CC-BY-NC-ND 4.0 International licensewas not certified by peer review) is the author/funder. It is made available under aThe copyright holder for this preprint (whichthis version posted April 24, 2020. . https://doi.org/10.1101/2020.04.22.056606doi: bioRxiv preprint 13 long-range temporal correlation and low complexity), whereas the inactive-to-active transition is driven by a wide 291 multifractal kinetics (extended long-range temporal correlation and relatively high complexity) (Fig. 6A). 292 Multifractal episodic motion of C. elegans in solid and liquid environments may be an adaptation to 293 food environments. The colonies of bacteria that C. elegans worms feed on grow in a fractal shape 45. In C. 294 elegans, the actively-moving state is likely to enable food foraging, whereas the inactive state is likely associated 295 with food intake, egg-laying, or resting to save energy or satiety 24,29-31,46,47. Temporal correlation in the inactive 296 state gives rise to a series of long and short periods for food intake that may follow the fractally-shaped bacterial 297 colony, whereas temporal correlation in active state round series gives rise to a series of long- and short-distance 298 foraging bouts that may follow the interbranch distances of a fractally-shaped bacterial colony. Thus, the scale-299 free and temporally structured residence times for food foraging and intake may be adaptive to the fractal shape 300 of bacterial colonies. C. elegans survival strategies are altered by food availability. Under starvation conditions, 301 C. elegans saves their energy for long-distance foraging and instead spend more effort for balancing resting and 302 food intake at a local area, e.g. through reuptake of their excrements. Food-dependent modulation of multifractal 303 kinetics of behaviors may improve food intake efficiency and reproductive success in natural environments. This 304 possibility should be tested in a modeling study. 305 Our quantitative studies showed that egl-4 mutants exhibited different alterations of multifractal 306 kinetics in the active versus the inactive state (Fig. 6C). This result suggests that EGL-4 may regulate the 307 multifractal kinetics of animal behaviors. However, egl-4 mutants’ defects in behavioral memory and behavioral 308 complexity were smaller in magnitude than the differences between the active and inactive states in wild-type C. 309 elegans, and also smaller than the index that characterizes qualitative differences of noise properties among white, 310 pink, and Brown noises. Differences between active and inactive states in wild-type C. elegans were comparable 311 to or greater than the index (compare Hpeak in Table 2: Hpeak of white, pink, and Brown noises differ by 0.5). 312 Therefore, multifractal kinetics in egl-4 mutants were within the range of kinetics qualitatively same as those in 313 wild-type. In addition, these defects may be caused by a behavioral variation of the egl-4 mutant strain rather than 314 .CC-BY-NC-ND 4.0 International licensewas not certified by peer review) is the author/funder. It is made available under aThe copyright holder for this preprint (whichthis version posted April 24, 2020. . https://doi.org/10.1101/2020.04.22.056606doi: bioRxiv preprint 14 by loss of function of egl-4/pkg per se. It is necessary to apply other fractal analyses and further molecular/genetic 315 studies to examine these possibilities. 316 317 PKG-modulated multifractal transition may be widely applicable to multi-time scale behaviors 318 Multifractal transition cycle and its PKG-dependent modulation may be shared among many invertebrates. In 319 short-term (sub-minute) observations, Drosophila 9 and Leptothorax allardycei worker ants 28 exhibit bimodal 320 behavioral mode switching 9,28 similar to that in C. elegans. Notably, Drosophila residence times in actively-321 moving and inactive states in a short-term observation period were distributed in a scale-free manner 9. 322 Interestingly, a bimodal behavioral choice between long-distance moving foragers and dwellers has been shown 323 to be regulated by PKG in C. elegans 26,27,48, Drosophila 32, bees 35, and ants 36,49. Both C. elegans mutant 26,27 324 and Drosophila polymorphism 32,50 local dwellers spend more time on local food intake/resting. Local dwellers 325 conduct brood care in bee hives 35 and ant nest defense 36,49. Forager and dweller behavioral phenotypes are 326 switched developmentally in bees, but maintained through the lifespan in the ant caste system. In C. elegans 26,27,48 327 and ants 36, long-distance foraging is associated with low PKG activity or expression, whereas local dwelling is 328 associated with high PKG expression. Conversely, long-distance foraging (or local dwelling) is associated with 329 high (low) activity or expression of PKG in flies 32 and bees 35. The time scale difference of PKG-active periods 330 between ants and honey bees, and the reversed function of PKG for long-distance foraging or local dwelling 331 between nematodes/ants and flies/bees may reflect a heterochronic evolutionary change in the developmental 332 control of PKG expression 35, and an evolutionary adaptation of the PKG signaling system in molecular drivers 333 of behavior, respectively. Despite evolutionary changes in time scale and the functional role of PKG, these results 334 indicate that appearance frequency for actively-moving and inactive states is modulated by PKG. Although the 335 role of PKG in the episodic motions of Drosophila and ants has not been studied directly, the aforementioned 336 findings suggest that at least episodic motions in Drosophila and ants may be regulated by multifractal kinetics 337 and its PKG-dependent modulation. 338 .CC-BY-NC-ND 4.0 International licensewas not certified by peer review) is the author/funder. It is made available under aThe copyright holder for this preprint (whichthis version posted April 24, 2020. . https://doi.org/10.1101/2020.04.22.056606doi: bioRxiv preprint 15 Surprisingly, PKG-dependent modulation of the transition kinetics in a multifractal transition cycle may 339 be involved in human heartbeat physiology. Electrocardiogram time series 51-53 of cardiac muscle depolarization 340 and repolarization are characterized by multifractality. The inter-beat (RR) interval is the period between peaks 341 of R waves, which reflect ventricular depolarization. The intra-beat (QT) interval is the period from the peak of 342 Q wave to the end of T wave, which corresponds to the depolarization in the left side of the intraventricular 343 septum and the repolarization of ventricular muscles, respectively. Round series of RR and QT intervals are 344 characterized by a multifractal structure 54. Interestingly, the shape of the multifractal spectrum for RR intervals 345 differs from that for QT intervals, raising the possibility that QT and RR intervals may be regulated by distinct 346 multifractal kinetics. PKG has been shown to regulate the heartbeat in mice and flies 55-58. PKG-knockout mice 347 56 or cardio-myocyte specific PKG-knockout mice 58 show hypertension. Although the mechanism is still in 348 dispute 57, the defects in these knockout mice indicate that PKG is involved in the relaxation phase of the cardiac 349 cycle. Thus, the heartbeat can be regulated by a multifractal transition cycle and is subject to PKG-dependent 350 modulation, similar to that we documented for C. elegans episodic swimming. Thus, PKG-modulated multifractal 351 transition cycles may occur in various animal species, including humans, and organs. 352 353 Molecular and system-level mechanisms for multifractal episodic swimming 354 Based on previously reported models designed to recapitulate scale-free and multifractal time series, we 355 developed hypotheses for molecular and system-level mechanisms underlying the multifractal nature of animal 356 behaviors, and discuss how PKG may be involved in the mechanism. First, the intermittent bursts of a single 357 neuron can be reproduced in chaos dynamical systems 59-65; owing to their non-Gaussian fluctuation with 358 intermittency, such systems may be associated with multifractality in animal behaviors. Chaotic dynamics 359 reproduced in these models are generated by interactions between fast inflows and slow outflows of ions to/from 360 a neuron. Thus, multifractal nature in behavior might be attributable to chaotic biochemical reactions in a single 361 neuron. Second, when a system is in a certain state, called the criticality, local interactions among system 362 .CC-BY-NC-ND 4.0 International licensewas not certified by peer review) is the author/funder. It is made available under aThe copyright holder for this preprint (whichthis version posted April 24, 2020. . https://doi.org/10.1101/2020.04.22.056606doi: bioRxiv preprint 16 components related to order and noise for randomization of the order reach a critical balance and, eventually, self-363 similar dynamics emerge spontaneously in the system 66,67; scale-free dynamics due to the criticality can be 364 associated with multifractality in animal behaviors 68,69. Experimentally observed scale-free neural activity in 365 intact brains was recapitulated in self-organized criticality-based models that assumed global interaction of 366 neuronal signals in a hierarchically structured neural network representative of brain structures in C. elegans and 367 humans 70,71. Therefore, multifractality in animal behaviors may be attributed to criticality in the hierarchal 368 structure of brains. Third, although implications of molecular and physiological mechanisms in the multifractal 369 nature of animal behaviors have not been reported, the multiplicative cascade model is a simple model that 370 generates multifractal time series (Fig. 4B, Supplemental Information) 41,42,72,73. Generally, multiplicative cascade 371 models generate multifractal time series via a series of multiplications of log-normal random noise in a cascade 372 manner, enhancing time series variance greatly as it progresses (see Supplemental Information). It has been shown 373 that waiting time distributions of neuronal action potentials exhibit a log-normal distribution consequent to the 374 nonlinearity of reaction systems in neurons 74-76, whereas hierarchal multiplication may correspond to a cascading 375 signal relay in the brain. Thus, multifractal nature in behaviors may be attributed to both non-linear reactions 376 within neurons and hierarchal structures of neural networks. To reveal molecular- and system-level mechanisms 377 of C. elegans multifractal episodic swimming, it will be critical to identify the operating principles that are 378 functioning in multifractal episodic swimming in C. elegans. With regard to PKG, selective ectopic expression 379 of PKG in R3 and R4d ring neurons is sufficient to restore behavioral defects in Drosophila with pkg/foraging 380 alleles 77,78. Additionally, differential PKG expression was identified in a set of five specific neurons between 381 different castes of ants 36. In C. elegans, PKG functions in a limited number of neurons 25. PKG has been shown 382 to regulate synaptic vesicle cycling, Ca2+ influx via G-protein signaling, and axon guidance in certain neuron 383 types 77,79-82. The endogenous and ectopic expression of PKG in specific neurons and functional analysis of PKG 384 suggest that PKG modulates multifractal kinetics of animal behaviors in single neurons or small numbers of 385 neurons. In this study, our experiments revealed a basic kinetic mechanism of C. elegans multifractal episodic 386 .CC-BY-NC-ND 4.0 International licensewas not certified by peer review) is the author/funder. It is made available under aThe copyright holder for this preprint (whichthis version posted April 24, 2020. . https://doi.org/10.1101/2020.04.22.056606doi: bioRxiv preprint 17 swimming. The findings are applicable to diverse fields of interest, including human heartbeat physiology. Due 387 to the wide variety of molecular and genetic tools available for C. elegans research, our observation and analysis 388 approach may be used to help reveal conserved mechanisms underlying the multifractal nature of animal 389 physiology and behavior. 390 391 Methods 392 C. elegans strains used and their maintenance 393 The Bristol N2 strain was used as wild-type C. elegans. N2 and egl-4(n479) mutant animals were maintained on 394 agar plates with E. coli OP50 strain at 15 °C 83. 395 396 Design, fabrication, and characterization of microfluidic device 397 To maximize the number of animals monitored in the recording area, our device employs a two-vertical-398 compartment structure with the array of culture chambers located over a flow path, whose boundary was 399 partitioned by a porous membrane (Whatman 111115, Nuclepore Hydrophilic Membrane, 10-µm pores, GE 400 Healthcare, USA) (Fig. 1A and B). The upper polydimethylsiloxane (PDMS) chip consisted of a 108-chamber 401 array (1 worm/chamber) with buffer-inlet and exchange solution-outlet ports in the chambers. The lower 402 PDMS chip consisted of a snake-shaped microchannel for supplying liquid buffer to the culture chambers. PDMS 403 chips were fabricated by conventional replica molding with the SU-8 epoxy-based photoresist 84,85. Animal-404 loading ports (100 µm wide) on the top of each of chamber in the upper PDMS chip were made with a Zing 16 405 laser (Epilog Laser, Japan). Before culturing and buffer exchanges, the loading ports were sealed with a PDMS 406 sheet. Each culture chamber (2-mm diameter and 0.3-mm height) in the upper PDMS chip was approximately 2 407 fold-longer and 3 fold-thicker than the ~1-mm-long and ~0.1-mm-wide C. elegans body (Fig. 1C). The serpentine 408 buffer supply microchannel in the lower PDMS chip had a 2.2-mm diameter, covering each upper-chip culture 409 .CC-BY-NC-ND 4.0 International licensewas not certified by peer review) is the author/funder. It is made available under aThe copyright holder for this preprint (whichthis version posted April 24, 2020. . https://doi.org/10.1101/2020.04.22.056606doi: bioRxiv preprint 18 chamber. The three parts (upper and lower PDMS chips, and the microporous membrane) were assembled by 410 covalent bonding with an aminosilane coupling agent and oxygen plasma treatment 86. 411 412 Culturing C. elegans in a microfluidic device 413 Animals at the young adult stage were collected from an agar plate. After an agar plate-culture period (2 d at 414 24 °C for wild-type and 3 d at 15 °C for egl-4 mutants), the animals were collected manually in room-temperature 415 M9 buffer and introduced into the microfluidic device by manual pipetting via chamber loading ports (Fig, 1C). 416 With previous devices, individual C. elegans were held in a clump structure and released into the chamber by 417 applying deforming high pressure 87-89. We instead used manual pipetting to avoid applying mechanical stress 418 during loading. To maintain a constant chemical environment in the culture chambers, M9 buffer (with or without 419 1 g/L glucose and 5 mg/L cholesterol) was perfused continuously with a peristatic pump (Fig. 1B and D) at a flow 420 rate at 5 ml/h. Flow rate was measured by monitoring weight changes in the water discharged from the outlet 421 (data not shown). Buffer exchange in chambers with the flow passing through the porous membrane was 422 confirmed in experiments with a fluorescent solution (Fig. S2). Air bubbles in the buffer supply tube were 423 removed with polytetrafluoroethylene membrane in an Omnifit bubble trap (006BT, Diba Industries Inc., USA) 424 (Fig. 1D). 425 426 Observation of C. elegans swimming 427 C. elegans behaviors were recorded at 20 frames/s through a macroscope with an apochromat objective lens (1×) 428 (Z16 APO, Leica, Germany) and a CCD camera (1940 × 1460 pixels, 2.8 Megapixel) with a USB3.0 connection 429 (MD028MU-SY, Ximea, Germany). The camera was controlled by Micromanager (https://micro-manager.org). 430 Movies were compressed (H265 codec in FFmpeg) every 10,000 frames (500 s) into mp4 files. C. elegans are 431 sensitive to temperature change of 4 ºC 90 and highly sensitive to blue-ultraviolet light, but not to green/yellow 432 light (>545 nm) 91,92. Temperature of culture chambers on WormFlo were maintained by submerging WormFlo 433 .CC-BY-NC-ND 4.0 International licensewas not certified by peer review) is the author/funder. It is made available under aThe copyright holder for this preprint (whichthis version posted April 24, 2020. . https://doi.org/10.1101/2020.04.22.056606doi: bioRxiv preprint 19 in M9 buffer in a 15-cm-diameter glass dish, whose temperature was maintained by temperature-controlled water 434 supplied from a high-precision water bath (HAAKE, Germany) (Fig. 1D and Fig. S3A and B). A temperature 435 logger (TC-08, Pico Technology, UK) confirmed that the temperature of the M9 buffer in the glass dish was 436 maintained within ±0.5 ºC during the 6-day recording period (Fig. S3C). The recording system was covered by a 437 light shield to prevent illumination changes from light fluctuations related to daily lab activities (Fig. 1D). Blue 438 light of illumination light from a halogen lamp was filtered out with a 0.5-mm-thick orange acryl plate (Fig. 1D 439 and S3A); spectrophotometry (USB400, Ocean Optics, USA) confirmed that wavelengths <500 nm were filtered 440 out (Fig. S3D). Standard deviation of light intensity change at the culture chambers during observation period 441 was kept within 10% of the average (Fig. S3E). This culturing and recording system allowed us to monitor 442 individually cultured C. elegans in an environment with minimized chemical, light, and temperature perturbations. 443 444 Quantification of C. elegans swimming activity 445 We measured C. elegans swimming activity by counting the number of pixels with an intensity over a certain 446 threshold in a matrix obtained from the difference between intensities obtained at t and t + 1 (we referred to it as 447 image difference) using the Open CV module in Python2.7. When an animal moves at a frame interval, the dark 448 pixels detecting its body in image[t] become brighter in image[t +1], yielding an increased image difference (= 449 image[t +1] – image[t]) (Fig. 2A). For pixel counting, the image difference matrix values were in the range of -450 256 to +256 at 1940 × 1460 pixels. The pixels with intensity differences greater than 12 value within this range 451 thus exhibit an image difference and are counted as “active” pixels (note that the pixel intensity threshold of 12 452 is different from the activity threshold (12 pixels/frame) to define residence time in active- and inactive states 453 (Fig. 3A)). The pixel intensity threshold of 12 was determined empirically so that actual movements of animals 454 are detected efficiently while avoiding artefact-based false hits due to thermal noise in a pixel on the camera 455 sensor. Based on pixel counts, the size of the area the animal moved through during each time frame (50 ms) was 456 measured (Fig. 2A). Image and data processing, including compensation for artifactual activity in the time series 457 .CC-BY-NC-ND 4.0 International licensewas not certified by peer review) is the author/funder. It is made available under aThe copyright holder for this preprint (whichthis version posted April 24, 2020. . https://doi.org/10.1101/2020.04.22.056606doi: bioRxiv preprint 20 are described in the Supplemental Information (Figs. S4 and S6). To classify swimming activity, we measured 458 average activities during the early half of the recording period (day 0 to day 3; abbreviated “AA0-3”) and the 459 ratio of AA0-3 to average activities during the latter half of the recording period (day 3 to day 6; abbreviated 460 “AA3-6”). Animals with high AA0-3 and a low AA0-3 to AA3-6 ratio were classified as High activity. Animals 461 with high AA0-3 and a high AA0-3 to AA3-6 ratio were classified as Middle activity. Animals with low AA0-3 462 and a low AA0-3 to AA3-6 ratio were classified as Low activity. The threshold of high AA0-3 was set empirically 463 at 0.75 and the threshold of a high ratio was set to 0.01. Note that 0.75 threshold values are low due to the 464 bursty/sparse nature of C. elegans swimming activity. 465 466 Data analysis 467 Active-state residence time was defined as the period after the start of the activity burst to the end of the burst; 468 inactive-state residence time was defined as the period starting immediately after the end of activity burst to the 469 next round of activity burst in the activity time series. Alternating-state round series data were obtained by 470 thresholding time series of swimming activity at 12 pixels/frame, which corresponded to the valley of a bimodal 471 activity distribution (Fig. 3A). 472 MF-DFA was performed in Python software 93 using 473 𝐹(𝑞, 𝑠) =6789 :∑[𝑓7(𝑣, 𝑠)]=/7789?@6A6/= (Eq. 1) 474 where 𝑓(𝑣, 𝑠) is the noise component or fluctuations from the local trend of cumulative sums of the deviation of 475 residence time from the average residence times in the vth segment at the temporal resolution for observation (𝑠). 476 MF-DFA was derived from detrended fluctuation analysis (DFA) 94; the equation used in DFA corresponds to the 477 equation used in MF-DFA when 𝑞 = 2 in Eq. 1). In operation 1, the entire cumulative sum series was segmented 478 into Ns segments at scale s, and the local trend in each vth segment at scale s was determined by piecewise fitting 479 with a linear function. In operation 2, the amplitude of fluctuations from the local linear trend at each segment 480 was enhanced (or suppressed) to a large (or small) amplitude of 𝑓(𝑣, 𝑠) by exponentiating with positive (or 481 .CC-BY-NC-ND 4.0 International licensewas not certified by peer review) is the author/funder. It is made available under aThe copyright holder for this preprint (whichthis version posted April 24, 2020. . https://doi.org/10.1101/2020.04.22.056606doi: bioRxiv preprint 21 negative) q-values; q value-exponentiated 𝑓(𝑣, 𝑠) was summed over all the vth segments. In operation 3. the 482 operations 1 and 2 above were performed bidirectionally, forward and backward, on the round series (in total 2 × 483 𝑁C segments) to obtain 𝐹(𝑞, 𝑠). 𝐹(𝑞, 𝑠) versus s plots were log-log plotted, and each q-value was fit with a linear 484 function to obtain a local Hurst (Hölder) exponent. Finally, multifractal spectrum [q-order (local) Hurst exponent, 485 or Hölder spectrum H(q) vs. q-order singularity dimension, or singularity spectrum D(q)] was obtained from 486 𝐹(𝑞, 𝑠) versus s at each q-value by Legendre transformation. Linear fitting to data in a log-log plot minimizes 487 relative error between the fit function (y_fit) and the data (y_data), which is log(y_fit/y_data) (= log(y_fit)-488 log(y_data)). Linear fitting to data in a log-log plot avoids biased fitting due to errors being weighted by high-489 value data points that happens when fitting is performed by minimizing absolute error (y_fit - y_data). Animals 490 whose chambers had long-term retained bubbles, and wild-type animals which were transparent at the final movie 491 frame were eliminated from the data analysis. 492 493 Statistical analysis 494 Student t-tests and chi-squared tests for independence were performed with scipy.stats.chi2_contingency and 495 scipy.stats.ttest_ind, respectively 95. 496 497 Funding 498 Y. Arata is supported by a research grant, Challenging Research (Pioneering), Grants-in-Aid for Scientific 499 Research, Ministry of Education, Culture, Sports, Science and Technology, Japan (18H05300). 500 501 Acknowledgements 502 We thank the Caenorhabditis Genetics Center (CGC) for C. elegans strains; the CGC is funded by NIH Office 503 of Research Infrastructure Programs (P40 OD010440). We are grateful to Yuki Shindo for providing valuable 504 help with Python programing, to Seiichi Uchida at Kyushu University for sharing computational power for 505 .CC-BY-NC-ND 4.0 International licensewas not certified by peer review) is the author/funder. It is made available under aThe copyright holder for this preprint (whichthis version posted April 24, 2020. . https://doi.org/10.1101/2020.04.22.056606doi: bioRxiv preprint 22 recording, and Hideitsu Hino at The Institute of Statistical Mathematics for advices for statistical tests. We are 506 grateful to RStudio Team (2018) for RStudio: Integrated Development for R. RStudio, Inc., Boston, MA URL 507 http://www.rstudio.com/. 508 509 References 510 1 Tohsato, Y., Ho, K. H., Kyoda, K. & Onami, S. SSBD: a database of quantitative data of spatiotemporal 511 dynamics of biological phenomena. Bioinformatics 32, 3471-3479, doi:10.1093/bioinformatics/btw417 512 (2016). 513 2 Reddy, A. B. & Rey, G. Metabolic and nontranscriptional circadian clocks: eukaryotes. Annu Rev Biochem 514 83, 165-189, doi:10.1146/annurev-biochem-060713-035623 (2014). 515 3 Mellor, J. The molecular basis of metabolic cycles and their relationship to circadian rhythms. Nat Struct 516 Mol Biol 23, 1035-1044 (2016). 517 4 Iannaccone, P. M. & Khokha, M. Fractal Geometry in Biological Systems: An Analytical Approach. 518 (CRC Press, 1996). 519 5 Mandelbrot, B. B. The fractal geometry of nature. Updated and augm. edn, (W.H. Freeman, 1983). 520 6 Bunde, A. & Havlin, S. Fractals in Science. (Springer-Verlag, 1994). 521 7 Kembro, J. M., Flesia, A. G., Gleiser, R. M., Perillo, M. A. & Marin, R. H. Assessment of long-range 522 correlation in animal behavior time series: The temporal pattern of locomotor activity of Japanese quail 523 (Coturnix coturnix) and mosquito larva (Culex quinquefasciatus). Physica A 392, 6400-6413 (2013). 524 8 Guzman, D. A. et al. The fractal organization of ultradian rhythms in avian behavior. Sci Rep 7, 684, 525 doi:10.1038/s41598-017-00743-2 (2017). 526 9 Cole, B. J. Fractal Time in Animal Behavior - the Movement Activity of Drosophila. Anim Behav 50, 527 1317-1324 (1995). 528 .CC-BY-NC-ND 4.0 International licensewas not certified by peer review) is the author/funder. It is made available under aThe copyright holder for this preprint (whichthis version posted April 24, 2020. . https://doi.org/10.1101/2020.04.22.056606doi: bioRxiv preprint 23 10 Alves, L. G. A. et al. Long-range correlations and fractal dynamics in C. elegans: Changes with aging and 529 stress. Phys Rev E 96 (2017). 530 11 Haris, K., Chakraborty, B., Menezes, A., Sreepada, R. A. & Fernandes, W. A. Multifractal detrended 531 fluctuation analysis to characterize phase couplings in seahorse (Hippocampus kuda) feeding clicks. J 532 Acoust Soc Am 136, 1972-1981, doi:10.1121/1.4895713 (2014). 533 12 Seuront L, M.C., B. & J.R., S. in Handbook of Scaling Methods in Aquatic Ecology: Measurement, 534 Analysis, Simulation (eds Laurent Seuront & Peter G. Strutton) 333 (CRC Press, 2003). 535 13 Seuront, L., Schmitt, F. G., Brewer, M. C., Strickler, J. R. & Souissi, S. From random walk to multifractal 536 random walk in zooplankton swimming behavior. Zool Stud 43, 498-510 (2004). 537 14 Hu, K., Van Someren, E. J., Shea, S. A. & Scheer, F. A. Reduction of scale invariance of activity 538 fluctuations with aging and Alzheimer's disease: Involvement of the circadian pacemaker. Proc Natl Acad 539 Sci U S A 106, 2490-2494, doi:10.1073/pnas.0806087106 (2009). 540 15 Hausdorff, J. M. et al. Altered fractal dynamics of gait: reduced stride-interval correlations with aging and 541 Huntington's disease. J Appl Physiol (1985) 82, 262-269, doi:10.1152/jappl.1997.82.1.262 (1997). 542 16 Hausdorff, J. M. Gait dynamics, fractals and falls: finding meaning in the stride-to-stride fluctuations of 543 human walking. Hum Mov Sci 26, 555-589, doi:10.1016/j.humov.2007.05.003 (2007). 544 17 Kobayashi, M. & Musha, T. 1/f fluctuation of heartbeat period. IEEE Trans Biomed Eng 29, 456-457, 545 doi:10.1109/TBME.1982.324972 (1982). 546 18 Ivanov, P. C. et al. Scaling behaviour of heartbeat intervals obtained by wavelet-based time-series analysis. 547 Nature 383, 323-327, doi:10.1038/383323a0 (1996). 548 19 Stam, C. J. & de Bruin, E. A. Scale-free dynamics of global functional connectivity in the human brain. 549 Hum Brain Mapp 22, 97-109, doi:10.1002/hbm.20016 (2004). 550 20 Linkenkaer-Hansen, K., Nikouline, V. V., Palva, J. M. & Ilmoniemi, R. J. Long-range temporal 551 correlations and scaling behavior in human brain oscillations. J Neurosci 21, 1370-1377 (2001). 552 .CC-BY-NC-ND 4.0 International licensewas not certified by peer review) is the author/funder. It is made available under aThe copyright holder for this preprint (whichthis version posted April 24, 2020. . https://doi.org/10.1101/2020.04.22.056606doi: bioRxiv preprint 24 21 Struzik, Z. R., Hayano, J., Sakata, S., Kwak, S. & Yamamoto, Y. 1/f scaling in heart rate requires 553 antagonistic autonomic control. Phys Rev E Stat Nonlin Soft Matter Phys 70, 050901, 554 doi:10.1103/PhysRevE.70.050901 (2004). 555 22 Goldberger, A. L. et al. Fractal dynamics in physiology: alterations with disease and aging. Proc Natl 556 Acad Sci U S A 99 Suppl 1, 2466-2472, doi:10.1073/pnas.012579499 (2002). 557 23 Ghosh, R. & Emmons, S. W. Episodic swimming behavior in the nematode C. elegans. J Exp Biol 211, 558 3703-3711, doi:10.1242/jeb.023606 (2008). 559 24 McCloskey, R. J., Fouad, A. D., Churgin, M. A. & Fang-Yen, C. Food responsiveness regulates episodic 560 behavioral states in Caenorhabditis elegans. J Neurophysiol 117, 1911-1934, doi:10.1152/jn.00555.2016 561 (2017). 562 25 Ghosh, R. & Emmons, S. W. Calcineurin and protein kinase G regulate C. elegans behavioral quiescence 563 during locomotion in liquid. BMC Genet 11, 7, doi:10.1186/1471-2156-11-7 (2010). 564 26 Fujiwara, M., Sengupta, P. & McIntire, S. L. Regulation of body size and behavioral state of C. elegans 565 by sensory perception and the EGL-4 cGMP-dependent protein kinase. Neuron 36, 1091-1102 (2002). 566 27 L'Etoile, N. D. et al. The cyclic GMP-dependent protein kinase EGL-4 regulates olfactory adaptation in 567 C. elegans. Neuron 36, 1079-1089 (2002). 568 28 Cole, B. J. Short-Term Activity Cycles in Ants - Generation of Periodicity by Worker Interaction. Am Nat 569 137, 244-259, doi:Doi 10.1086/285156 (1991). 570 29 McConnell, M. W. & Fitzpatrick, M. J. 'Foraging' for a place to lay eggs: A genetic link between foraging 571 behaviour and oviposition preferences. Plos One 12 (2017). 572 30 Sokolowski, M. B. Social interactions in "simple" model systems. Neuron 65, 780-794, 573 doi:10.1016/j.neuron.2010.03.007 (2010). 574 31 Reaume, C. J. & Sokolowski, M. B. cGMP-dependent protein kinase as a modifier of behaviour. Handb 575 Exp Pharmacol, 423-443, doi:10.1007/978-3-540-68964-5_18 (2009). 576 .CC-BY-NC-ND 4.0 International licensewas not certified by peer review) is the author/funder. It is made available under aThe copyright holder for this preprint (whichthis version posted April 24, 2020. . https://doi.org/10.1101/2020.04.22.056606doi: bioRxiv preprint 25 32 Osborne, K. A. et al. Natural behavior polymorphism due to a cGMP-dependent protein kinase of 577 Drosophila. Science 277, 834-836, doi:DOI 10.1126/science.277.5327.834 (1997). 578 33 Sokolowski, M. B. Foraging Strategies of Drosophila-Melanogaster - a Chromosomal Analysis. Behav 579 Genet 10, 291-302 (1980). 580 34 Kaun, K. R. et al. Natural variation in food acquisition mediated via a Drosophila cGMP-dependent 581 protein kinase. J Exp Biol 210, 3547-3558, doi:10.1242/jeb.006924 (2007). 582 35 Ben-Shahar, Y., Robichon, A., Sokolowski, M. B. & Robinson, G. E. Influence of gene action across 583 different time scales on behavior. Science 296, 741-744, doi:10.1126/science.1069911 (2002). 584 36 Lucas, C. & Sokolowski, M. B. Molecular basis for changes in behavioral state in ant social behaviors. 585 Proc Natl Acad Sci U S A 106, 6351-6356, doi:10.1073/pnas.0809463106 (2009). 586 37 Kantelhardt, J. W. et al. Multifractal detrended fluctuation analysis of nonstationary time series. Physica 587 A 316, 87-114 (2002). 588 38 Ihlen, E. A. Introduction to multifractal detrended fluctuation analysis in matlab. Front Physiol 3, 141, 589 doi:10.3389/fphys.2012.00141 (2012). 590 39 Beran, J., Whitcher, B. & Maechler, M. longmemo; Statistics for Long-Memory Processes (Book Jan 591 Beran), and Related Functionality, R package Version 1.1-1. (2018). 592 40 Ihaka, R. & Gentleman, R. R: A Language for Data Analysis and Graphics. Journal of Computational and 593 Graphical Statistics 5, 299-314, doi:10.2307/1390807 (1996). 594 41 Bacry, E., Delour, J. & Muzy, J. R. Multifractal random walk. Phys Rev E 64 (2001). 595 42 Kiyono, K., Struzik, Z. R. & Yamamoto, Y. Estimator of a non-Gaussian parameter in multiplicative log-596 normal models. Phys Rev E 76 (2007). 597 43 Arata, Y. & Takagi, H. Quantitative Studies for Cell-Division Cycle Control. Frontiers in Physiology 10, 598 doi:10.3389/fphys.2019.01022 (2019). 599 .CC-BY-NC-ND 4.0 International licensewas not certified by peer review) is the author/funder. It is made available under aThe copyright holder for this preprint (whichthis version posted April 24, 2020. . https://doi.org/10.1101/2020.04.22.056606doi: bioRxiv preprint 26 44 Gonzales, D. L., Zhou, J., Fan, B. & Robinson, J. T. A microfluidic-induced C. elegans sleep state. Nat 600 Commun 10, 5035, doi:10.1038/s41467-019-13008-5 (2019). 601 45 Matsuyama, T. & Matsushita, M. Fractal morphogenesis by a bacterial cell population. Crit Rev Microbiol 602 19, 117-135, doi:10.3109/10408419309113526 (1993). 603 46 You, Y. J., Kim, J., Raizen, D. M. & Avery, L. Insulin, cGMP, and TGF-beta signals regulate food intake 604 and quiescence in C-elegans: A model for satiety. Cell Metab 7, 249-257 (2008). 605 47 Raizen, D. M. et al. Lethargus is a Caenorhabditis elegans sleep-like state. Nature 451, 569-U566 (2008). 606 48 Raizen, D. M., Cullison, K. M., Pack, A. I. & Sundaram, M. V. A novel gain-of-function mutant of the 607 cyclic GMP-dependent protein kinase egl-4 affects multiple physiological processes in Caenorhabditis 608 elegans. Genetics 173, 177-187, doi:10.1534/genetics.106.057380 (2006). 609 49 Ingram, K. K., Oefner, P. & Gordon, D. M. Task-specific expression of the foraging gene in harvester 610 ants. Mol Ecol 14, 813-818, doi:10.1111/j.1365-294X.2005.02450.x (2005). 611 50 Osborne, K. A., de Belle, J. S. & Sokolowski, M. B. Foraging behaviour in Drosophila larvae: mushroom 612 body ablation. Chem Senses 26, 223-230, doi:10.1093/chemse/26.2.223 (2001). 613 51 Morhrman, D. E. & Heller, L. J. Cardiovascular Physiology. 7th edn, (McGraw-Hill Medical, 2010). 614 52 Lilly, L. S. Pathophysiology of Heart Disease: A Collaborative Project of Medical Students and Faculty. 615 (Wolters Kluwer Health, 2016). 616 53 Dubin, D. Rapid Interpretation of EKGs. 6th edn, (Cover Pubulishing Co, 1998). 617 54 Lewis, M. J., Short, A. L. & Suckling, J. Multifractal characterisation of electrocardiographic RR and QT 618 time-series before and after progressive exercise. Comput Methods Programs Biomed 108, 176-185, 619 doi:10.1016/j.cmpb.2012.02.014 (2012). 620 55 Johnson, E., Sherry, T., Ringo, J. & Dowse, H. Modulation of the cardiac pacemaker of Drosophila: 621 cellular mechanisms. J Comp Physiol B 172, 227-236, doi:10.1007/s00360-001-0246-8 (2002). 622 .CC-BY-NC-ND 4.0 International licensewas not certified by peer review) is the author/funder. It is made available under aThe copyright holder for this preprint (whichthis version posted April 24, 2020. . https://doi.org/10.1101/2020.04.22.056606doi: bioRxiv preprint 27 56 Pfeifer, A. et al. Defective smooth muscle regulation in cGMP kinase I-deficient mice. EMBO J 17, 3045-623 3051, doi:10.1093/emboj/17.11.3045 (1998). 624 57 Feil, R., Lohmann, S. M., de Jonge, H., Walter, U. & Hofmann, F. Cyclic GMP-dependent protein kinases 625 and the cardiovascular system: insights from genetically modified mice. Circ Res 93, 907-916, 626 doi:10.1161/01.RES.0000100390.68771.CC (2003). 627 58 Wegener, J. W. et al. cGMP-dependent protein kinase I mediates the negative inotropic effect of cGMP 628 in the murine myocardium. Circ Res 90, 18-20 (2002). 629 59 Fan, Y. S. & Holden, A. V. Bifurcations, Burstings, Chaos and Crises in the Rose-Hindmarsh Model for 630 Neuronal-Activity. Chaos Soliton Fract 3, 439-449 (1993). 631 60 Gu, H. G. & Xiao, W. W. Difference Between Intermittent Chaotic Bursting and Spiking of Neural Firing 632 Patterns. Int J Bifurcat Chaos 24 (2014). 633 61 Izhikevich, E. M. Neural excitability, spiking and bursting. Int J Bifurcat Chaos 10, 1171-1266 (2000). 634 62 Chay, T. R. Chaos in a 3-Variable Model of an Excitable Cell. Physica D 16, 233-242 (1985). 635 63 Fan, Y. S. & Chay, T. R. Generation of periodic and chaotic bursting in an excitable cell model. Biol 636 Cybern 71, 417-431 (1994). 637 64 Holden, A. V. & Fan, Y. From Simple to Complex Oscillatory Behaviour via Intermittent Chaos in the 638 Rose-Hindmarsh Model For Neuronal Activity. Chaos, Solitons & Fractals 2, 349-369, 639 doi:https://doi.org/10.1016/0960-0779(92)90012-C (1992). 640 65 Canavier, C. C., Clark, J. W. & Byrne, J. H. Routes to chaos in a model of a bursting neuron. Biophys J 641 57, 1245-1251, doi:10.1016/S0006-3495(90)82643-6 (1990). 642 66 Beggs, J. M. The criticality hypothesis: how local cortical networks might optimize information 643 processing. Philos T R Soc A 366, 329-343 (2008). 644 67 Bak, P., Tang, C. & Wiesenfeld, K. Self-organized criticality: An explanation of the 1/f noise. Phys Rev 645 Lett 59, 381-384, doi:10.1103/PhysRevLett.59.381 (1987). 646 .CC-BY-NC-ND 4.0 International licensewas not certified by peer review) is the author/funder. It is made available under aThe copyright holder for this preprint (whichthis version posted April 24, 2020. . https://doi.org/10.1101/2020.04.22.056606doi: bioRxiv preprint 28 68 Schertzer, D. & Lovejoy, S. Multifractal Generation of Self-Organized Criticality. Ifip Trans A 41, 325-647 339 (1994). 648 69 Tebaldi, C., De Menech, M. & Stella, A. L. Multifractal scaling in the Bak-Tang-Wiesenfeld sandpile and 649 edge events. Physical Review Letters 83, 3952-3955 (1999). 650 70 Rubinov, M., Sporns, O., Thivierge, J. P. & Breakspear, M. Neurobiologically realistic determinants of 651 self-organized criticality in networks of spiking neurons. PLoS Comput Biol 7, e1002038, 652 doi:10.1371/journal.pcbi.1002038 (2011). 653 71 Moretti, P. & Munoz, M. A. Griffiths phases and the stretching of criticality in brain networks. Nat 654 Commun 4, 2521, doi:10.1038/ncomms3521 (2013). 655 72 Arneodo, A., Manneville, S., Muzy, J. F. & Roux, S. G. Revealing a lognormal cascading process in 656 turbulent velocity statistics with wavelet analysis. Philos T R Soc A 357, 2415-2438 (1999). 657 73 Meneveau, C. & Sreenivasan, K. R. The Multifractal Nature of Turbulent Energy-Dissipation. J Fluid 658 Mech 224, 429-484 (1991). 659 74 Petersen, P. C. & Berg, R. W. Lognormal firing rate distribution reveals prominent fluctuation-driven 660 regime in spinal motor networks. Elife 5, doi:10.7554/eLife.18805 (2016). 661 75 Buzsaki, G. & Mizuseki, K. The log-dynamic brain: how skewed distributions affect network operations. 662 Nat Rev Neurosci 15, 264-278 (2014). 663 76 Roxin, A., Brunel, N., Hansel, D., Mongillo, G. & van Vreeswijk, C. On the distribution of firing rates in 664 networks of cortical neurons. J Neurosci 31, 16217-16226, doi:10.1523/JNEUROSCI.1677-11.2011 665 (2011). 666 77 Peng, Q. et al. cGMP-Dependent Protein Kinase Encoded by foraging Regulates Motor Axon Guidance 667 in Drosophila by Suppressing Lola Function. J Neurosci 36, 4635-4646, doi:10.1523/JNEUROSCI.3726-668 15.2016 (2016). 669 .CC-BY-NC-ND 4.0 International licensewas not certified by peer review) is the author/funder. It is made available under aThe copyright holder for this preprint (whichthis version posted April 24, 2020. . https://doi.org/10.1101/2020.04.22.056606doi: bioRxiv preprint 29 78 Kuntz, S., Poeck, B., Sokolowski, M. B. & Strauss, R. The visual orientation memory of Drosophila 670 requires Foraging (PKG) upstream of Ignorant (RSK2) in ring neurons of the central complex. Learn Mem 671 19, 337-340, doi:10.1101/lm.026369.112 (2012). 672 79 Eguchi, K., Nakanishi, S., Takagi, H., Taoufiq, Z. & Takahashi, T. Maturation of a PKG-dependent 673 retrograde mechanism for exoendocytic coupling of synaptic vesicles. Neuron 74, 517-529, 674 doi:10.1016/j.neuron.2012.03.028 (2012). 675 80 Wang, X. & Robinson, P. J. Cyclic GMP-dependent protein kinase and cellular signaling in the nervous 676 system. J Neurochem 68, 443-456, doi:10.1046/j.1471-4159.1997.68020443.x (1997). 677 81 Dason, J. S., Allen, A. M., Vasquez, O. E. & Sokolowski, M. B. Distinct functions of a cGMP-dependent 678 protein kinase in nerve terminal growth and synaptic vesicle cycling. J Cell Sci 132, 679 doi:10.1242/jcs.227165 (2019). 680 82 Krzyzanowski, M. C. et al. The C. elegans cGMP-dependent protein kinase EGL-4 regulates nociceptive 681 behavioral sensitivity. PLoS Genet 9, e1003619, doi:10.1371/journal.pgen.1003619 (2013). 682 83 Brenner, S. The genetics of Caenorhabditis elegans. Genetics 77, 71-94 (1974). 683 84 Lorenz, H. et al. High-aspect-ratio, ultrathick, negative-tone near-UV photoresist and its applications for 684 MEMS. Sensor Actuat a-Phys 64, 33-39 (1998). 685 85 Hosokawa, K., Fujii, T. & Endo, I. Handling of picoliter liquid samples in a poly(dimethylsiloxane)-based 686 microfluidic device. Anal Chem 71, 4781-4785 (1999). 687 86 Kimura, H., Yamamoto, T., Sakai, H., Sakai, Y. & Fujii, T. An integrated microfluidic system for long-688 term perfusion culture and on-line monitoring of intestinal tissue models. Lab on a Chip 8, 741-746 (2008). 689 87 Chung, K. et al. Microfluidic chamber arrays for whole-organism behavior-based chemical screening. Lab 690 Chip 11, 3689-3697, doi:10.1039/c1lc20400a (2011). 691 88 Li, S., Stone, H. A. & Murphy, C. T. A microfluidic device and automatic counting system for the study 692 of C. elegans reproductive aging. Lab Chip 15, 524-531, doi:10.1039/c4lc01028k (2015). 693 .CC-BY-NC-ND 4.0 International licensewas not certified by peer review) is the author/funder. It is made available under aThe copyright holder for this preprint (whichthis version posted April 24, 2020. . https://doi.org/10.1101/2020.04.22.056606doi: bioRxiv preprint 30 89 Hulme, S. E. et al. Lifespan-on-a-chip: microfluidic chambers for performing lifelong observation of C. 694 elegans. Lab Chip 10, 589-597, doi:10.1039/b919265d (2010). 695 90 Simonetta, S. H., Migliori, M. L., Romanowski, A. & Golombek, D. A. Timing of locomotor activity 696 circadian rhythms in Caenorhabditis elegans. PLoS One 4, e7571, doi:10.1371/journal.pone.0007571 697 (2009). 698 91 Edwards, S. L. et al. A novel molecular solution for ultraviolet light detection in Caenorhabditis elegans. 699 PLoS Biol 6, e198, doi:10.1371/journal.pbio.0060198 (2008). 700 92 Ward, A., Liu, J., Feng, Z. & Xu, X. Z. Light-sensitive neurons and channels mediate phototaxis in C. 701 elegans. Nat Neurosci 11, 916-922, doi:10.1038/nn.2155 (2008). 702 93 Jurica, P. Multifractal analysis for all. Frontiers in Physiology 6, doi:UNSP 27 703 10.3389/fphys.2015.00027 (2015). 704 94 Peng, C. K., Havlin, S., Stanley, H. E. & Goldberger, A. L. Quantification of scaling exponents and 705 crossover phenomena in nonstationary heartbeat time series. Chaos 5, 82-87, doi:10.1063/1.166141 706 (1995). 707 95 Jones, E., Oliphant, T., Peterson, P. & others. SciPy: Open Source Scientific Tools for Python [Online; 708 accessed 2019-04-09], <"http://www.scipy.org/"> (2001). 709 710 711 .CC-BY-NC-ND 4.0 International licensewas not certified by peer review) is the author/funder. It is made available under aThe copyright holder for this preprint (whichthis version posted April 24, 2020. . https://doi.org/10.1101/2020.04.22.056606doi: bioRxiv preprint 31 Figure legends 712 Figure 1 Culturing and recording system for individual C. elegans animals 713 (A and B) The WormFlo apparatus has a vertically two-compartment structure, wherein an array of 108 culture 714 chambers and loading ports of the upper PDMS chip are partitioned from the buffer flow path located in the lower 715 PDMS chip by a porous membrane (details in the Methods). M9 buffer was supplied from the inlet and withdrawn 716 from the outlet. (C) The diameter and height of each culture chamber in the upper PDMS chip were 2 mm and 717 0.3 mm, approximately 2-fold longer and 3-fold thicker than the ~1-mm-long and ~ 0.1-mm wide C. elegans 718 body, respectively. Animals are introduced in each chamber via a 0.1-mm-wide loading port on the roof of a disc-719 shaped culture chamber, which was closed with a thin PDMS sheet before M9 buffer was perfused. (D) C. elegans 720 animals cultured in the WormFlo are monitored by a macroscope with a CCD camera. Buffer perfusion is driven 721 by a peristatic pump. To avoid loss of water due to evaporation, the WormFlo was submerged in a 15-cm-diameter 722 glass dish (depth, 5.4 cm) filled with M9 buffer. The temperature of the M9-filled 15-cm glass dish was 723 maintained by circulating temperature-controlled water in a water flow path in the attached aluminum block under 724 the 15-cm glass dish (Fig. S3A). Animals cultured in the microfluidic device are illuminated by blue-light filtered 725 light from a halogen lamp in a light shielding box that buffers against light intensity changes due to daily lab 726 activity. 727 728 Figure 2 C. elegans episodic swimming exhibited a multi-time scale kinetics. 729 Swimming activity of animals cultured in individual chambers was quantified by a pixel counting method 730 (Methods). (A) Chambers in WormFlo are shown with row-column indexes (left figure). Active pixels (intensity 731 difference > 12 in the range of -256 to 256) between time point [t] and time point [t+1] are shown in red pixels in 732 the upper right image and in yellow pixels in the lower right image, respectively. The animal in chamber id 54 733 moved while the animal in culture chamber id 66 did not move from time [t] to [t+1]. (B) Active pixel numbers 734 .CC-BY-NC-ND 4.0 International licensewas not certified by peer review) is the author/funder. It is made available under aThe copyright holder for this preprint (whichthis version posted April 24, 2020. . https://doi.org/10.1101/2020.04.22.056606doi: bioRxiv preprint 32 are shown on the y-axis as an index of swimming activity with culture time is shown on the x-axis. Six-day 735 temporal activity decay patterns were classified into High, Middle, and Low activity classes by two criteria; 736 average activities during the early half of the recording period (first 3 days) and the ratio of average activity in 737 the early half to that in the late half of the recording period (see Methods). Pre-starved regime (red area) and 738 Starved regime (blue area) were defined for each activity class. For the High activity class, the Pre-starved and 739 Starved regimes were defined by early 50% period from the start of recording through the early half of the 740 recording period and the remainder of the period. For the Middle activity class, Pre-starved regime and Starved 741 regime were defined by the period from the start of recording to early 20% period of the recording period and the 742 remainder. For the Low activity class, the Pre-starved and Starved regimes were defined by the period from the 743 start of recording to early 10% of the recording period and the remainder. (C) Swimming activities of a 744 representative animal in various time scales; full recording time scale (107 timepoints) at the top. Each of the 745 lower panels are a 10× magnification of the first tenth of its upper panel (red area). Activity threshold at 12 746 pixels/frame is shown by a red horizontal line in the 5-s scale. Animal activities at 6 d-, 1 d-, 1 h-, 10 min-, 1 min-, 747 and 1 s-scales are shown. 748 749 Figure 3 A power law distribution of active/inactive state residence times in C. elegans 750 episodic swimming 751 (A) Appearance frequency of number of active pixels in a representative animal for a single image difference 752 frame. Bimodal distribution of active pixels in normalized probability density (NPD) was separated at 10~20 753 pixels/frame. The active state threshold was ≥12 active pixels for round series data analysis (details in Methods). 754 (B and C) Round series of residence times in the active (B) and inactive (C) states with the y-axis in log scale. 755 Activity periods above and below the active state threshold (separated by a red horizontal line on 5-s scale in Fig. 756 2C) were defined as active and inactive periods, respectively. (D-G) Mean and standard deviations of NPD of 757 residence times in active state (red) and in inactive state (blue) among individual animals in indicated time regime 758 .CC-BY-NC-ND 4.0 International licensewas not certified by peer review) is the author/funder. It is made available under aThe copyright holder for this preprint (whichthis version posted April 24, 2020. . https://doi.org/10.1101/2020.04.22.056606doi: bioRxiv preprint 33 were shown in log-log plot. Fitting was performed over 0.3–5.0 s for active states and over 1–50 s for inactive 759 states. The fit line is shown as a black line. Pre-starved vs. Starved time regimes are as described in Figure 2 760 (wild-type) and the main text (egl-4 mutants). 761 762 Figure 4 Multifractal analysis of numerically-generated round series and experimentally 763 obtained round series of active and inactive states in a representative C. elegans animal 764 (A-D) MF-DFA of numerically-generated round series with monofractal-like (A) and multifractal (B) properties. 765 Time series with small/middle/great multifractality were generated by the multiplicative cascading processes 766 using log-normal functions whose variances were determined by random noises with small/middle/large variances 767 (41,42; see Supplemental Information). Experimentally-obtained Pre-starved active/inactive state round series of 768 High activity wild-type (C) and egl-4 mutant (D) animals. Time series [blue, in (A) and (B), magnified by arbitrary 769 unit] and cumulative sum of the deviations from the average of values in the time series [red, in (A), (C), and (D)] 770 are shown in upper three (A and B) and two (C and D) graphs. 𝐹(𝑞, 𝑠) vs. scale s plot data (dots) and their fit 771 functions (lines) are shown at lower q values (cold colors) and higher q values (hotter colors) in a range of -10 < 772 q < 10 [left column in the lower three (A and B) and two (C and D) graphs]; corresponding multifractal spectra 773 are shown in right columns of each graph. Original white, pink, and Brown noise time series are shown, 774 respectively, after 10, 102, and 104-time magnifications (A). 775 776 Figure 5 Multifractal spectra averaged from multiple animals 777 Multifractal spectra means (and standard deviations) were calculated from round series data spectra of active 778 (left) or inactive (right) states from multiple wild-type animals in High (A), Middle (B), and Low (C) activity 779 classes in the Pre-starved (Pre-S, red circles) or Starved (S, blue circles) time regimes. For egl-4 mutants (D), 780 spectra represent only in Pre-S (red circles) time regime. Note that the spectrum width in Pre-starved time regime 781 .CC-BY-NC-ND 4.0 International licensewas not certified by peer review) is the author/funder. It is made available under aThe copyright holder for this preprint (whichthis version posted April 24, 2020. . https://doi.org/10.1101/2020.04.22.056606doi: bioRxiv preprint 34 was as narrow as in Starved time regime in Low activity classes, which was due to too short high activity period 782 during the defined Pre-starved regime among animals in Low activity classes (C). 783 784 Figure 6 Multifractal episodic swimming in C. elegans. 785 (A) Alternate state transitions are driven by distinct kinetics: active-to-inactive and inactive-to-active state 786 transitions follow narrow and wide multifractal kinetics, respectively. (B) Because C. elegans episodic swimming 787 is driven by alternating transitions between an actively moving state and an inactive (resting and posing) state 788 over a broad range of temporal scales (i.e. seconds to hours) the residence-time round series data of which are 789 characterized by a multifractal nature, we refer to this behavior as “multifractal episodic swimming”. (C) 790 Summary of our kinetic analyses. See Tables 1 and 2. 791 792 793 .CC-BY-NC-ND 4.0 International licensewas not certified by peer review) is the author/funder. It is made available under aThe copyright holder for this preprint (whichthis version posted April 24, 2020. . https://doi.org/10.1101/2020.04.22.056606doi: bioRxiv preprint (A)(B)(C)(D)Flow pathInletOutletChambers with introducing portsPorous membraneInletOutletCCD cameraBubble trapperTemperature-controlling water bathPeristatic pumpLight shieldWormFloBlue light-cut filterIllumination lightWaste reservoirSupply reservoirMacroscopeFigure 1_Ikeda et alAluminum boxIntroducing portC. elegans at adult stage (~1 mm)0.3 mm2.0 mm0.1 mm.CC-BY-NC-ND 4.0 International licensewas not certified by peer review) is the author/funder. It is made available under aThe copyright holder for this preprint (whichthis version posted April 24, 2020. . https://doi.org/10.1101/2020.04.22.056606doi: bioRxiv preprint �������������������������������������������������������������������������� ��������������������������������������������������������������������������������������������������������������������������������������������������������������(A)(B)(C)im[ t ]im[ t +1 ]High activity classMiddle activity classLow activity classFigure 2_Ikeda et al1 2 3 4 5 6 7 8 9 10 11 121 2 3 4 5 6 7 8 9id54id66.CC-BY-NC-ND 4.0 International licensewas not certified by peer review) is the author/funder. It is made available under aThe copyright holder for this preprint (whichthis version posted April 24, 2020. . https://doi.org/10.1101/2020.04.22.056606doi: bioRxiv preprint (C)(D)(E)(F)wild-type HighFigure 3 Ikeda et al(G)wild-type Middlewild-type Lowegl-4Pre-starvedStarved(A)(B)log10(NPD)log10(Time [s]).CC-BY-NC-ND 4.0 International licensewas not certified by peer review) is the author/funder. It is made available under aThe copyright holder for this preprint (whichthis version posted April 24, 2020. . https://doi.org/10.1101/2020.04.22.056606doi: bioRxiv preprint 0round−70040Time [sec]White noiseCummulative sum0round−4000300Pink noiseCumulative sum0round−5000000300000Brown noiseCumulative sum4 5 6 7 8 9 10 11 12-10123450123014 5 6 7 8 9 10 11 12-1012345log10(F(q, s))012301D(q)4 5 6 7 8 9 10 11 12log2(s)-10123450123H(q)010500010000 15000 20000 25000 30000round100102Time [sec]small0500010000 15000 20000 25000 30000round100102middle0500010000 15000 20000 25000 30000round100102great4 5 6 7 8 9 10 11 12-4-20240123014 5 6 7 8 9 10 11 12-4-2024log10(F(q, s))012301D(q)4 5 6 7 8 9 10 11 12log2(s)-4-20240123H(q)010500010000150002000025000round02000Integrated time [sec]0500010000150002000025000round−2000004 5 6 7 8 9 10 11 12log2(s)-1012345log10(F(q, s))0123H(q)0.000.250.500.751.00D(q)4 5 6 7 8 9 10 11 12log2(s)-1012345log10(F(q, s))0123H(q)0.000.250.500.751.00D(q)05000100001500020000round007000Integrated time [sec]05000100001500020000round−900000−04 5 6 7 8 9 10111213log2(s)-1012345log10(F(q, s))0123H(q)0.000.250.500.751.00D(q)4 5 6 7 8 9 10111213log2(s)-1012345log10(F(q, s))0123H(q)0.000.250.500.751.00D(q)A) monofractal-likeB) multifractalC) wild-type High Pre-starveD) egl-4 Pre-starveActiveInactiveActiveInactiveFigure 4 Ikeda et alActiveInactiveActiveInactiveActiveInactiveActiveInactive.CC-BY-NC-ND 4.0 International licensewas not certified by peer review) is the author/funder. It is made available under aThe copyright holder for this preprint (whichthis version posted April 24, 2020. . https://doi.org/10.1101/2020.04.22.056606doi: bioRxiv preprint 0123Holder spectrum, H(q)0.00.51.0Singularity spectrum, D(q)ActiveActivePre-SS01230.00.51.0InactiveInactive01230.00.51.0ActiveActivePre-SS01230.00.51.0InactiveInactive01230.00.51.0ActiveActivePre-SS01230.00.51.0InactiveInactive01230.00.51.0ActiveActivePre-S01230.00.51.0InactiveInactivePre-S(A) wild-type HighFigure 5 Ikeda et al(B) wild-type Middle(C) wild-type Low(D) egl-4 ActiveInactiveö.CC-BY-NC-ND 4.0 International licensewas not certified by peer review) is the author/funder. It is made available under aThe copyright holder for this preprint (whichthis version posted April 24, 2020. . https://doi.org/10.1101/2020.04.22.056606doi: bioRxiv preprint ActiveInactive(A)(B)Widemultifractal kineticsNarrow multifractal kineticsFigure 6_ Ikeda et alVariablesIndicator forMolecular mechanism forBehavioral statesPre-starvedStarvedegl-4 mutantsWild-typePower lawexponentHpeakWidthAppearance frequency of residence timeMemory in residence timeVariety of cluster structureScale-free residence timeBehavioral memoryBehavioral complexityActiveInactiveActiveInactiveActiveInactive-------shallow--PoorRich-steep-shortPoorPoorPre-starved(C)Multiractal episodic swimmingTime scale ranging from second to 1-hour.CC-BY-NC-ND 4.0 International licensewas not certified by peer review) is the author/funder. It is made available under aThe copyright holder for this preprint (whichthis version posted April 24, 2020. . https://doi.org/10.1101/2020.04.22.056606doi: bioRxiv preprint | RESEARCH GATE |
ORIGINAL RESEARCHpublished: 29 March 2021doi: 10.3389/fchem.2021.657028Frontiers in Chemistry | www.frontiersin.org1March 2021 | Volume 9 | Article 657028Edited by:Thierry Brigaud,Université de Cergy-Pontoise, FranceReviewed by:Ya-nan Yang,Institute of Materia Medica, ChinaPaul-Henri Ducrot,INRA UMR1318 Institut Jean PierreBourgin, France*Correspondence:Wenyi Kangkangweny@hotmail.comJiangmiao Huhujiangmiao@mail.kib.ac.cnSpecialty section:This article was submitted toOrganic Chemistry,a section of the journalFrontiers in ChemistryReceived: 22 January 2021Accepted: 04 March 2021Published: 29 March 2021Citation:Liu Z, Niu Y, Zhou L, Meng L, Chen S,Wang M, Hu J and Kang W (2021)Nine Unique Iridoids and IridoidGlycosides From Patriniascabiosaefolia.Front. Chem. 9:657028.doi: 10.3389/fchem.2021.657028Nine Unique Iridoids and IridoidGlycosides From PatriniascabiosaefoliaZhenhua Liu 1,2,3, Yun Niu 1,4, Li Zhou 1,4, Lijun Meng 1,4, Sitan Chen 1,4, Mengke Wang 1,4,Jiangmiao Hu 2,3* and Wenyi Kang 1,4*1 National Center for Research and Development of Edible Fungus Processing Technology, Henan University, Kaifeng, China,2 State Key Laboratory of Phytochemistry and Plant Resources in West China, Kunming Institute of Botany, ChineseAcademy of Sciences, Kunming, China, 3 College of Life Sciences, University of Chinese Academy of Sciences, Beijing,China, 4 Joint International Research Laboratory of Food & Medicine Resource Function, Henan University, Kaifeng, ChinaPatrinia scabiosaefolia is a medical and edible Chinese herb with high nutritional andmedicinal value. The continuing study of its chemical constituents led to the discoveryof nine unique iridoids and iridoid glycosides, including three new iridoids (1-3) and sixpreviously unknown irioid glycosides (5-10), and one known compound (4). Among them,compound 1 was a deformed iridoid, while compounds 3, 5-7, and 10 formed a new ringin their skeletons which was uncommon in this genus. For compound 3, the new ringexisted between C-3 and C-10, while a 1,3-dioxane appeared between C-7 and C-10in compounds 5-7 and 10. Moreover, compound 10 was a bis-iridoid glycoside, whichwas the first reported in P. scabiosaefolia. And the sugar of irioid glycosides (5-10) wasglucose at C-11, except in 9 which had a 5-deoxyglucose moiety. All their structures wereconfirmed based on the extensive spectroscopic analysis, including IR, UV, HR-ESI-MS,ECD, and 1D- and 2D-NMR experiments. Their cytotoxic activities against HL-60, A-549,SMMC-7721, MCF-7, SW480 were also tested.Keywords: Patrinia scabiosaefolia, iridoid, iridoid glycoside, bis-iridoid glycoside, cytotoxic activityINTRODUCTIONThere are about 400 species in 13 genera of Valerianaceae, mostly distributed in northern temperatezones and the subtropics or frigid zones. In China, this family comprises three genera, Nardostachys,Valeriana, and Patrinia, which can be found all over the country (Delectis Flora ReipublicaePopularis Sinicae Agendae Academiae Sinicae Edita., 1986). Plants from the genus Patrinia, such asP. scabiosaefolia, P. villosa, and P. heterophylla, have a long history of use as a traditional Chinesemedicine for detoxification, swelling, empyema, liver protection, and cholagogue,. Some species arealso edible as wild herbs, for example P. scabiosaefolia, P. villosa, P. punctiflora, and P. angustifolia(Xiao et al., 2007).Among them, P. scabiosaefolia is a medical and edible Chinese herb, first recorded in “ShengNong’s Herbal Classic.” It is of great nutritional value and is enriched with amino acids, vitamins,β-carotene, and trace elements. The content of vitamin C per 100 grams is 42.65 mg, which ishigher than that in some vegetables and fruits. Also, there is 14995.69 mg of amino acids in it,including eight kinds of essential amino acids which account for 36.06% of the total amountof amino acids (Zhong et al., 2001). It was also verified to have effects on the initial stagesof edema, appendicitis, endometriosis, and inflammation (Delectis Flora Reipublicae PopularisSinicae Agendae Academiae Sinicae Edita., 1986).Liu et al.Nine Unique Iridoids and GlycosidesGRAPHICAL ABSTRACT | Nine unique iridoids and iridoid glycosides from Patrinia scabiosaefolia.Phytochemical research showed that different kinds ofcompounds existed in the Patrinia genus, including triterpenes,iridoids, flavones, lignans, and their glycosides (Kim and Kang,2013). Among these, iridoids are considered as the maincomponents with diverse structures and various activities, whichattract our attention to further explore the plant P. scabiosaefolia.As a result, we found a series of iridoids from ethyl acetate extractof P. scabiosaefolia, including three bis-iridoids which were firstreported (Liu et al., 2017a,b; Liu et al., 2019).In the continuing study, chemical constituents on n-butanolextract of P. scabiosaefolia led to the discovery of nine uniqueiridoids and iridoid glycosides, including three new iridoids (1-3)and six novel irioid glycosides (5-10), and one known compound(4) (Figure 1, Graphical Abstract). Among them, compound 1was a deformed iridoid, compound 3 formed a cycle between C-3and C-10, compounds 5-7 with a 1,3-dioxane between C-7 andC-10, and compound 10 was a bis-iridoid glycoside, which wasthe first reported in P. scabiosaefolia. Herein, we discussed theirisolation, structure elucidation, and their cytotoxic activities.EXPERIMENTAL SECTIONGeneral Experimental ProceduresOptical rotation was obtained on a JASCO P-1020 digitalpolarimeter (Horiba, Tokyo, Japan). UV spectra were measuredby a Shimadzu UV-2401 PC spectrophotometer (Shimadzu,Kyoto, Japan). IR spectra were obtained on a Bruker Tensor27 infrared spectrophotometer (Bruker Optics GmbH, Ettlingen,Germany) with KBr pellets. Mass spectra were performed onan API QSTAR time-of-flight spectrometer (MDS Sciqaszex,Concord, Ontario, Canada) and LC/MS-IT-TOF (Shimadzu,Kyoto, Japan) spectrometer. NMR spectra were recorded onBruker DRX-500 and Av III-800 instruments with TMS asthe internal standard (Bruker, Bremerhaven, Germany). Thechemical shifts were given in δ (ppm) with reference tothe solvent signal. Column chromatography was performedon silica gel (200–300 and 300–400 mesh, Qingdao MarineChemical Inc., Qingdao, China), Lichroprep Rp-18 gel (40–63 µm, Merck, Darmstadt, Germany), MCI gel CHP-20P (75–150 µm, Mitsubishi Chemical Corp., Tokyo, Japan), SephadexLH-20 (20–150 µm, Amersham Biosciences, Uppsala, Sweden),and YMC∗GEL ODS-A-HG (50 µm, YMC Co. Ltd. Japan).Fractions were monitored by TLC, and spots were visualizedby UV light and sprayed with 10% H2SO4 in EtOH, followedby heating.Plant MaterialThe whole plants of Patrinia scabiosaefolia were collected inOctober 2010 from Shucheng county, Anhui Province, People’sRepublic of China, and were stored in a cool and dry place atroom temperature. The material was identified by Prof. Shou-Jin Liu in Anhui University of Chinese Medicine and a voucherspecimen (Wan1295) was deposited in Anhui University ofChinese Medicine. The plants of P. scabiosaefolia are common inthe local area and the collection was permitted. We also ensuredthat the local population of P. scabiosaefolia was not destroyedthrough the means of collection at different locations.Extraction and IsolationThe air-dried and powdered whole plants (29 kg) of P.scabiosaefoliawereextractedwith95%ethanol(3×75 L)underroomtemperatureandconcentratedunderreduced pressure. Then the residue (3 kg) was dissolvedinwaterandpartitionedsuccessivelywithn-butanolFrontiers in Chemistry | www.frontiersin.org2March 2021 | Volume 9 | Article 657028Liu et al.Nine Unique Iridoids and GlycosidesFIGURE 1 | The chemical structures of compounds 1-10 (*new compounds).toyieldn-butanolextract(0.85 kg)afterconcentration.The n-butanol extract was subjected to silica gel columnchromatography eluted with a gradient of CHCl3-MeOH(8:1→0:1, v/v) to obtain five fractions 1-5 by TLCplate analysis.Fraction 2 (22 g) was separated by silica gel columnchromatography eluted with a gradient system of CHCl3-MeOH(6:1→1:1, v/v) to afford 4 subfractions (Fr.2-1 to Fr.2-4).Fr.2-1 (5.9 g) was performed by Rp-18 column chromatography(MeOH-H2O, 50:50→ 100:0, v/v) to afford 6 subfractions (Fr.2-1-1 to Fr.2-1-6). Fraction 2-1-2 (1.0 g) was applied to SephadexLH-20 column chromatography (MeOH-H2O, 90:10, v/v) andthen purified by silica gel column chromatography repeatedly toobtain 1 (1.0 mg). Fraction 2-2-3 (1.3 g) was submitted to silicagel column chromatography eluted with a gradient of EtOAc-MeOH (4:1→ 1:1, v/v) to afford A (180 mg) and B (240 mg),and then purified by semi-prep. HPLC (MeOH-H2O, 65:35, v/v)to obtain 3 (7.0 mg) and 4 (16.0 mg) respectively. Fraction 2-2-4(2.1 g) was separated by the same methods as Fraction 2-2-3 andpurified by semi-prep. HPLC (MeOH-H2O, 55:45, v/v) to afford5 (1.0 mg) and 10 (16.0 mg).Fraction 3 (15 g) was subjected to silica gel columnchromatography eluted with a gradient system of CHCl3-MeOH(5:1→ 1:1, v/v) to afford 6 subfractions (Fr.3-1 to Fr.3-6). Fr.3-3(4.3 g) was separated by Rp-18 column chromatography (MeOH-H2O, 50:50→ 100:0, v/v) to get 5 subfractions (Fr.3-3-1 to Fr.3-3-5). Fraction 3-3-3 (560 mg) was applied to Sephadex LH-20column chromatography (MeOH-H2O, 90:10, v/v), and silica gelcolumn chromatography eluted with CHCl3- MeOH (4:1, v/v) toobtain 2 (3.0 mg).Fraction4(28 g)wassubmittedtoRp-18columnchromatography (MeOH-H2O, 30:70→ 100:0, v/v) to obtain6 subfractions (Fr.4-1 to Fr.4-6). Fr.4-2 (3.4 g) was applied tosilica gel column chromatography eluted with a gradient systemof EtOAc-MeOH (4:1→ 1:1, v/v), then isolated by SephadexLH-20 column chromatography (MeOH-H2O, 90:10, v/v), andfurther purified by semi-prep. HPLC (MeOH-H2O, 55:45, v/v)to afford 6 (3.0 mg) and 7 (10 mg). Fr.4-4 (5.6 g) was repeatedlyseparated by silica gel column chromatography and SephadexLH-20 column chromatography (MeOH-H2O, 90:10, v/v), andfinally purified by semi-prep. HPLC (MeOH-H2O, 55:45, v/v) toafford 8 (3.0 mg) and 9 (16.0 mg).Frontiers in Chemistry | www.frontiersin.org3March 2021 | Volume 9 | Article 657028Liu et al.Nine Unique Iridoids and GlycosidesTABLE 1 | The 1H and 13C NMR data of 1-4 (CD3OD, δ in ppm, J in Hz).Position123δHδCδHδCδHδC14.89 (s)111.2 d4.92 (d, 5.7)104.1 d5.19 (d, 2.8)99.5 d39.55 (s)196.3 d7.40 (br s)164.1 d5.49 (d, 1.7)94.7 d4-154.1 s-125.6 s2.00 (m)34.5 d52.91 (q, 7.5)40.2 d3.06 (q, 8.3)31.2 d3.51 (m)29.5 d6a1.85 (m)40.8 t2.24 (m)41.9 t1.97 (m)36.6 t6b1.78 (m)1.55 (m)2.02 (m)74.26 (q, 4.8)73.3 d4.27 (dt, 2.2, 5.1)73.5 d4.77 (t, 3.0)74.4 d82.86 (m)48.6 d1.93 (m)49.4 d2.57 (m)50.2 d92.47 (dd, 7.2, 8.8)57.1 d2.08 (m)42.7 d2.90 (m)46.1 d10a4.19 (dd, 1.6, 8.8)66.6 t3.78 (dd, 12.7, 5.6)62.3 t4.37 (dd, 6.7, 10.5)62.4 t10b3.75 (t, 8.0)3.65 (m)4.21 (m)11a6.40 (d, 1.6)134.5 t9.21 (s)193.3 d3.79 (d, 9.6)61.8 t11b6.14 (s)4.21 (m)OMe3.24 (s)54.5 qR1-1′a3.89 (m)70.7 t3.91 (m)67.5 t1′b3.61 (m)3.48 (m)2′a1.57 (m)32.7 t1.53 (m)32.2 t2′b1.39 (m)3′1.40 (m)20.3 t1.31 (m)19.7 t4′0.93 (t, 7.4)14.1 q0.81 (t, 7.4)14.0 qPatrirscabioin M (1): colorless oil; [α]23 D-32.2 (c0.03, MeOH);UV (MeOH) λmax (log ε): 204 (3.42) nm; IR (KBr) νmax 3,428,2,925, 2,855, 1,632, 1,384, 1,101 cm−1; 1H and 13C NMR data,see Table 1; positive ESIMS m/z 235 [M + Na]+; HREIMS m/z235.0943 [M + Na]+ (calcd for C11H16O4Na, 235.0941).Patrirscabioin N (2): light yellow oil; [α]24 D-42.6 (c0.05,MeOH); UV (MeOH) λmax (log ε): 248 (3.72) nm; 203 (3.69) nm;IR (KBr) νmax 3,425, 2,927, 2,874, 1,727, 1,625, 1,384, 1,161, 1,088cm−1; 1H and 13C NMR data, see Table 1; positive ESIMS m/z293 [M + Na]+; HREIMS m/z 293.1354 [M + Na]+ (calcd forC14H22O5Na, 293.1359).Patrirscabioin O (3): light yellow oil; [α]23 D-11.3 (c0.05,MeOH); UV (MeOH) λmax (log ε): 203 (3.03) nm; IR (KBr)νmax 3,444, 2,926, 2,870, 1,725, 1,634, 1,383, 1,088 cm−1; 1Hand 13C NMR data, see Table 1; positive ESIMS m/z 295[M + Na]+; HREIMS m/z 295.1515 [M + Na]+ (calcd forC14H24O5Na, 295.1516).Patrinoside B (5): light yellow oil; [α]23 D-54.6 (c0.02, MeOH);UV (MeOH) λmax (log ε): 218 (4.15) nm; IR (KBr) νmax 3,426,2,929, 2,873, 1,727, 1,642, 1,076 cm−1; 1H and 13C NMR data,see Table 2; positive ESIMS m/z 593 [M + Na]+; HREIMS m/z593.2931 [M + Na]+ (calcd for C29H46O11Na, 593.2932).Patrinoside C (6): light yellow oil; [α]24 D-61.4 (c0.14, MeOH);UV (MeOH) λmax (log ε): 209 (4.16) nm, 217 (4.18) nm; IR(KBr) νmax 3,431, 2,926, 2,876, 1,727, 1,637, 1,384, 1,075 cm−1;1H and 13C NMR data, see Table 2; positive ESIMS m/z 537[M + Na]+; HREIMS m/z 537.2302 [M + Na]+ (calcd forC25H38O11Na, 537.2306).Patrinoside D (7): light yellow oil; [α]24 D-54.2 (c0.05, MeOH);UV (MeOH) λmax (log ε): 209 (4.06) nm, 217 (4.06) nm; IR(KBr) νmax 3,429, 2,924, 2,855, 1,724, 1,635, 1,384, 1,075 cm−1;1H and 13C NMR data, see Table 2; positive ESIMS m/z 551[M + Na]+; HREIMS m/z 551.2467 [M + Na]+ (calcd forC26H40O11Na, 551.2463).Patrinoside E (8): light yellow oil; [α]23 D-23.7 (c0.08, MeOH);UV (MeOH) λmax (log ε): 205 (3.95) nm, 214 (3.94) nm; IR(KBr) νmax 3,429, 2,925, 2,857, 1,725, 1,635, 1,384, 1,077 cm−1;1H and 13C NMR data, see Table 2; positive ESIMS m/z 497[M + Na]+; HREIMS m/z 497.1988 [M + Na]+ (calcd forC22H34O11Na, 497.1993).Patrinoside F (9): light yellow oil; [α]24 D-74.0 (c0.32, MeOH);UV (MeOH) λmax (log ε): 207 (4.18) nm; IR (KBr) νmax 3,441,2,925, 2,860, 1,727, 1,639, 1,121, 1,065 cm−1; 1H and 13C NMRdata, see Table 2; positive ESIMS m/z 467 [M + Na]+; HREIMSm/z 467.1889 [M + Na]+ (calcd for C26H40O11Na, 467.1888).Patrirscabiobisin D (10): light yellow oil; [α]23 D-34.1 (c0.17,MeOH); UV (MeOH) λmax (log ε): 209 (4.24) nm, 217 (4.25)nm; IR (KBr) νmax 3,426, 2,926, 1,724, 1641, 1,077 cm−1;1H and 13C NMR data, see Table 3; positive ESIMS m/z 663[M + Na]+; HREIMS m/z 663.2627 [M + Na]+ (calcd forC31H44O14Na, 663.2623).Cytotoxicity AssaysThe following human tumor cell lines were used: HL-60, SMMC-7721, A-549, MCF-7, and SW-480. These were obtained fromATCC (Manassas, VA, USA). All the cells were cultured inRPMI-1640 or DMEM medium (Hyclone, Logan, UT, USA),supplemented with 10% fetal bovine serum (Hyclone) at 37◦C ina humidified atmosphere with 5% CO2. Cell viability was assessedby conducting colorimetric measurements of the amount ofFrontiers in Chemistry | www.frontiersin.org4March 2021 | Volume 9 | Article 657028Liu et al.Nine Unique Iridoids and GlycosidesTABLE 2 | The 1H and 13C NMR data of 5-9 (CD3OD, δ in ppm, J in Hz).Position56789δHδCδHδCδHδCδHδCδHδC15.90 (d, 5.0)93.0 d5.88 (d, 5.2)93.1 d5.90 (d, 5.0)93.1 d6.60 (d, 5.1)92.6 d5.91(d, 5.5)93.2 d36.35 (s)139.8 d6.30 (s)139.9 d6.36 (s)139.8 d6.62 (s)139.0 d6.36 (s)140.1 d4-117.3 s-117.2 s117.2 s-116.1 s-116.5 s53.10 (q, 8.6)34.8 d3.11 (q, 8.6)34.8 d3.12 (q, 8.5)34.7 d3.45 (q, 7.9)33.5 d3.02 (m)34.1 d6a2.18 (m)39.3 t2.17 (m)39.3 t2.18 (m)39.2 t2.42 (m)41.0 t2.06 (m)40.9 t6b1.76 (m)1.74 (m)1.76 (m)2.04 (m)1.82 (m)74.15 (t, 3.6)79.3 d4.18 (t, 3.8)79.2 d4.18 (t, 3.8)79.2 d4.74 (m)72.7 d4.32 (m)73.3 d82.56 (m)41.6 d2.54 (m)41.6 d2.56 (m)41.6 d2.36 (m)48.7 d1.95 (m)49.2 d91.66 (m)42.4 d1.66 (d, 9.2)42.3 d1.68 (m)42.3 d2.74 (m)42.2 d2.18 (m)42.7 d10a4.04 (d, 11.9)67.1 t4.02 (d, 3.7)67.1 t4.02 (m)67.1 t4.35 (m)62.1 t3.80 (dd, 7.6, 9.6)62.2 t10b3.98 (dd,12.1, 2.0)4.26 (d, 12.0)4.27 (d, 11.6)4.27 (m)3.73 (dd,5.6, 11.0)11a4.26 (d, 11.2)69.7 t4.09 (d, 11.6)69.7 t4.09 (d, 11.5)69.8 t4.53 (d, 11.9)69.2 t4.06 (d, 11.4)69.7 t11b4.10 (d, 11.7)4.28 (m)4.24 (d, 11.7)R1-1’-166.4 s166.4 s166.6 s165.5 s166.4 s2’5.71 (t, 1.2)116.1 d5.71 (t, 1.2)116.1 d5.68 (br s)114.6 d5.73 (q, 1.2)114.5 d5.71 (t, 1.4)116.2 d3’-161.2 s161.2 s166.2 s-164.2 s161.0 s4’1.94 (s)27.6 q1.93 (d, 1.0)27.6 q2.22 (q, 4.4)34.8 t1.92 (dq, 1.0, 7.5)33.8 t1.93 (d, 1.2)27.6 q5’2.17 (s)20.6 q2.17 (d, 1.0)20.6 q1.09 (t, 7.4)12.3 q0.80 (t, 7.4)11.9 q2.17 (d, 1.2)20.6 q3’-Me2.17 (s)19.2 q2.14 (d, 1.2)19.0 qR11-1”4.29 (d, 7.8)103.2 d4.27 (d, 7.9)103.2 d4.29 (d, 7.8)103.3 d4.88 (d, 7.8)104.2 d4.22(d, 7.8)103.8 d2”3.19 (dd, 9.1, 7.9)75.1 d3.18 (m)75.1 d3.19 (t, 7.9)75.1 d4.03 (m)75.3 d3.09 (dd, 7.8, 9.0)76.9 d3”3.26 (m)78.0 d3.26 (m)77.9 d3.26 (m)78.0 d4.24 (m)78.6 d3.59 (m)72.2 d4”3.26 (m)71.7 d3.26 (m)71.7 d3.27 (m)71.7 d4.24 (m)71.7 d3.52 (m)73.9 d5”a3.34 (t, 8.7)78.2 d3.34 (m)78.1 d3.34 (m)78.1 d3.94 (m)78.7 d1.91 (m)36.4 t5”b1.35 (m)6”a3.87 (dd, 12.1, 2.0)62.8 t3.86 (dd, 12.4, 1.9)62.8 t3.87 (dd, 12.0, 2.0)62.8 t4.55 (dd, 2.5, 12.5)62.8 t3.56 (m)65.6 t6”b3.65 (dd, 11.9, 5.5)3.65 (dd, 11.9 5.5,)3.65 (dd, 11.9, 5.5)4.37 (m)3.56 (m)R10-1”’4.45 (q, 1.9)103.8 d4.52 (t, 5.1)102.0 d4.52 (t, 5.1)102.0 d2”’1.38 (m)45.4 d1.51 (m)38.2 t1.52 (m)38.2 t3”’a1.50 (m)22.6 t1.39 (m)18.2 t1.40 (m)18.2 t3”’b1.32 (m)0.91 (t, 7.4)14.3 q4”’a1.46 (m)29.2 t0.90 (t, 7.4)14.3 q4”’b1.27 (m)5”’1.28 (m)30.6 t6”’1.28 (m)24.2 t7”’0.89 (d, 7.8)14.4 q2”’-Me0.88 (d, 7.4)12.0 qFrontiers in Chemistry | www.frontiersin.org5March 2021 | Volume 9 | Article 657028Liu et al.Nine Unique Iridoids and GlycosidesTABLE 3 | The 1H and 13C NMR data of 10 (CD3OD δ in ppm, J in Hz).PositionδHδC15.80 (d, 5.7)93.3 d36.35 (s)139.9 d4117.2 s53.11 (q, 8.0)34.9 d6a2.04 (m)39.3 t6b1.66 (m)73.26 (m)77.9 d82.45 (m)48.3 d92.46 (m)41.8 d10a3.93 (d, 11.9)66.9 t10b3.78 (dd, 12, 3.1)11a4.27 (dd, 7.8, 3.4)69.6 t11b4.06 (dd, 12.8, 4.2)R1-1’166.3 s2’5.70 (t, 1.3)116.0 d3’161.3 s4’1.93 (d, 1.0)27.6 q5’2.18 (d, 1.0)20.6 qR11-1”4.27 (d, 7.8)103.3 d2”3.19 (dd, 9.0, 7.9)75.1 d3”3.35 (m)78.1 d4”3.26 (m)71.7 d5”4.01 (t, 3.7)78.8 d6”a3.86 (dd, 12.6, 1.9)62.8 t6”b3.65 (11.9, 5.5)R10-1”’4.07 (d, 4.2)102.3 d3”’9.54 (s)196.9 d4”’151.7 d5”’3.45 (m)37.5 d6”’a2.09 (m)39.4 t6”’b1.66 (m)7”’4.38 (t, 4.2)73.2 d8”’2.20 (m)46.1 d9”’1.60 (m)42.1 d10”’3.73 (d, 7.0)63.5 t11”’a6.25 (s)134.1 t11”’b6.15 (s)insoluble formazan formed in living cells based on the reductionof 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfopheny)-2H-tetrazolium (MTS) (Sigma, St. Louis, MO,USA) (Monks et al., 1991). In brief, 100 µL of adherent cells wereseeded into each well of a 96-well cell culture plate and allowedto adhere for 12 h before drug addition, while suspended cellswere seeded just before drug addition, both with an initial densityof 1 × 105 cells/mL in 100 µL medium. Each tumor cell linewas exposed to the test compound at various concentrations intriplicate for 48 h, with cisplatin and paclitaxel (Sigma) as positivecontrols. After the incubation, MTS (100 µg) was added to eachwell, and the incubation continued for 4 h at 37 ◦C. The cells werelysed with 100 µL of 20% SDS-50% DMF after removal of 100µL medium. The optical density of the lysate was measured at490 nm in a 96-well microtiter plate reader (Bio-Rad 680). TheIC50 value of each compound was calculated by the Reed andMuench method (1938).Computational StudyThe CHARMM force field and DFT/TDDFT calculations wereperformed with Discovery Studio 4.0 and Gaussian09 programpackage, respectively. Conflex conformational search generatedlow-energy conformers within a 20 kcal/mol energy window andwere subjected to geometry optimization using DFT methodwithout imposing any symmetry constraints at the B3LYP/6-31G(d) level. Frequency calculations were carried out using thesame level to verify that the molecular structures were trueminimum. The calculated ECD spectra were generated by theprogram SpecDis2 using a Gaussian band shape with 0.3eVexponential half-width from dipole-length dipolar and rotationalstrengths (Bruhn et al., 2013).RESULTS AND DISCUSSIONCompound 1 was indicated as C11H16O4 by HR-ESI-MS[m/z 235.0943 [M + Na]+, calcd. for 235.0941]. Preliminaryinspection of the 1H and 13C NMR (Table 1) spectroscopicdata of compound 1 revealed the presence of one conjugatedaldehyde group [δH 9.55 (s), δC 196.3 (d)] and a pair of doublebond [δC 154.1 (s), δC 134.5 (t)], two oxygenated methinegroups [δH 4.89 (s), δC 111.2 (d); δH 4.26 (q), δC 73.3 (d)],and one oxygenated methylene at δC 66.6 (t). These datawere similar to those of 8,9-didehydro-7-hydroxydolichodial(Veith et al., 1986), except for the disappearance of conjugatedaldehyde and double bonds, and the appearance of two methinesand one oxygenated methane, especially the change of methylat C-10 to oxygenated methylene. Therefore, we speculatedthat the methyl at C-10 of 8,9-didehydro-7-hydroxydolichodialwas oxidized to oxygenated methylene, and then underwenta hemiacetal formation with the aldehyde group at C-1 toform the deformed iridoid—compound 1 with 5/5 rings, whichoccurred simultaneously to the hydrogenation at C-8 and C-9,and finally the C-1 was methylated (Figure 2). This conclusionwas confirmed by HMBC spectrum with the key correlationsfrom H-1 [δH 4.89 (s)] to C-10 (δC 66.6), C-8 (δC 48.6),and C-9 (δC 57.1), and OMe (δC 54.5) were observed, andthe correlations of H-1/H-9 and H-9/H-5 in the ROESYspectrum (Figure 3).The absolute configurations of H-5 and H-7 were bothS deduced from the comparison of chemical shift value totheknowncompound8,9-didehydro-7-hydroxydolichodial(Veithetal.,1986).Theα-orientationofH-8waselucidated by the correlation of H-7 with H-8 in ROESY(Figure 3).Theβ-orientationsofH-1andH-9wereconfirmed by the ROESY correlations from H-5 to H-1andH-9(Figure 2).AndtheR-configurationofthenew formed acetal methine at C-1 was further confirmedby the comparison of experimental and calculated ECDspectra,whichthepositiveCottoneffectnear200 nmincalculatedECDspectraofR-configurationatC-1Frontiers in Chemistry | www.frontiersin.org6March 2021 | Volume 9 | Article 657028Liu et al.Nine Unique Iridoids and GlycosidesFIGURE 2 | The possible pathway for the transformation of compound 1.FIGURE 3 | Key 1H-1H COSY, HMBC, and ROESY correlations of compound 1.agreedwiththeexperimentalECDspectra(Figure 4).Hence,thestructureofcompound1wasdefinedas(1R,5S,7S,8S,9S)-1,10-epoxy-7,10-dihydroxy-dolichodial, namedPatriscabioin M.Compound 2, a light yellow oil, had a molecular formulaof C14H22O5 based on HRESIMS ([M + Na]+, m/z 293.1354,calcd. for 293.1359). Detailed comparisons of 1H-NMR and 13C-NMR (Table 1) suggested it was a typical iridoid and includeda hemiketal methine at δH 4.92 (d, J = 5.7 Hz, H-1) and δC104.1 (d, C-1), a conjugated olefinic bond at δC 193.3 (d, C-11); δC 164.1 (d, C-3), δC 125.6 (s, C-4), and one oxygenatedmethylene at δC 62.3 (t, C-10). These data were similar topatriscabioin I (Liu et al., 2017), except for the substituentat C-1 (Figure 5). In compound 2 it was n-butanol group[δC 70.7 (t), 32.7 (t), 20.3 (t), 14.1 (q)], while it was methylin patriscabioin I. The correlation from H-1 [4.92 (d, 5.7)]to C-1’ [δC 70.7 (t)] in HMBC implied the location of n-butanol group at C-1 (Figure 5). The configuration of compound2 was the same as patriscabioin I. Therefore, the structureof compound 2 was defined as (1R,5S,7S,8S,9S)-1-n-butoxy-7,10-dihydroxy-11-aldehyde-5,6-dihydrovaltrate hydrin, namedpatriscabioin N.Compound 3 was found to have a molecular formula ofC14H24O5, determined by its positive HRESIMS data ([M+ Na]+, m/z 295.1515, calcd. for 295.1516) and 13C-NMRspectrum (Table 1), with 3 degrees of unsaturation. From theNMR data, compound 3 showed spectral characteristics ofiridoid at δH 5.19 (d, J = 2.8 Hz, H-1) and δC 99.5 (d, C-1),with a n-butanol group [δC 67.5 (t), 32.2 (t), 19.7 (t), 14.0(q)]) located in C-1, which was confirmed by the correlationfrom H-1 to C-1’ in HMBC (Figure 6). However, the biggestdifference from other iridoids’ skeletons (such as compounds2, 5-9) was the added 1 unsaturation, of which the former’sdegree was 2, and the latter was 3. Meanwhile, the NMR spectraexhibited the disappeared double bonds and the appearanceof two methines, among which the chemical shift was δC94.7 (d, C-3), but there was no other unsaturated group.Therefore, it was presumed that there was a ring that existedFrontiers in Chemistry | www.frontiersin.org7March 2021 | Volume 9 | Article 657028Liu et al.Nine Unique Iridoids and GlycosidesFIGURE 4 | Calculated and experimental ECD spectra of compound 1 at TDDFT/B3LYP/6-31G(d) level.FIGURE 5 | Key 1H-1H COSY and HMBC correlations of compound 2 and the structure of patriscabioin I.in the skeleton of compound 3. The new ring was formedby the ether bond between C-3 and C-10, which was furtherverified by the fragment of H-3—H-4—H-5—H-9—H-8—H-10in COSY, and the correlation from H-3 to C-10 in HMBC(Figure 7). According to the S-configuration of H-5 andH-9 and R-configuration of H-1, the configuration of H-4was determined as R-configuration and the configurationsof H-3 and H-8 were determined as S-configuration by thecorrelations of H-4/H-5, H-3/H-11, and H-7/H-8 in ROESY(Figure 6). There was good agreement between experimentaland calculated ECD spectra, which was further verifiedthese configurations (Figure 7). Hence, the structure of 3wasestablishedas(1R,3S,4R,5S,7S,8S,9S)-1-n-butoxy-3,10-epoxy-7-hydroxy-3,4,5,6-tetrahydrovaltratehydrinm,namedpatriscabioin O.Compound 5 was obtained as a slight yellow oil. Its molecularformula was assigned as C29H46O11 on the basis of positive-ionHR-ESI-MS at m/z 593.2931 [M + Na]+ (calcd for 593.2932)and 13C-NMR data (Table 2). Carefully analysis of 1H-NMRand 13C-NMR led to the discovery of a hemiketal methineat δH 5.90 (1H, d, J = 5.0 Hz, H-1) and δC 93.0 (d, C-1), atrisubstituted olefinic bond at δH 6.35 (1H, s, H-3), δC 139.8Frontiers in Chemistry | www.frontiersin.org8March 2021 | Volume 9 | Article 657028Liu et al.Nine Unique Iridoids and GlycosidesFIGURE 6 | Key 1H-1H COSY, HMBC, and ROESY correlations of compound 3.FIGURE 7 | Calculated and experimental ECD spectra of compound 3 at TDDFT/B3LYP/6-31G(d) level.(d, C-3), and δC117.3 (s, C-4), an oxymethine at δC 79.3 (C-7) and two oxygenated methylenes at δC 67.1 (t, C-10) and δC69.7 (t, C-11). The above data indicated the existence of 7,10,11-trihydroxy-3-en valtratehydrin. It was found that 5 had a 3-methylcrotonyl group at δC 166.4 (s), δC 116.1 (d), δC 161.2(s), δC 27.6 (q), δC 20.6 (q) and a glucopyranose unit at δC103.2 (d), δC 78.2 (d), δC 78.0 (d), δC 75.1(d), δC 71.7 (d), δC62.8 (t) through further detailed analysis of its 1D-NMR and2D-NMR. The 3-methylcrotonyl group and glucopyranose werearranged at C-1 and C-11 by the correlations from δH 5.90(H-1) to δC 166.4 (C-1’), and δH 4.29 (H-1”) to δC 69.7 (C-11”) in HMBC spectrum (Figure 8). The typical signal for β-configuration (δH 4.29, d, J = 7.8 Hz) of H-1” could also beobserved. Thus, eight carbons remained, including two methylgroups, four methylene groups, and two methine groups. The 1H-1H COSY correlations (Figure 8) showed the fragments: CH(H-1”’)—CH(H-2”’)—CH2 (H-3”’)—CH2(H-4”’)—CH (H-5”’, H-6”’)—CH3 (H-7”’) and CH(H-2”’)—CH3(2”’-Me). CombinedFrontiers in Chemistry | www.frontiersin.org9March 2021 | Volume 9 | Article 657028Liu et al.Nine Unique Iridoids and GlycosidesFIGURE 8 | Key 1H-1H COSY and HMBC correlations of compound 5.FIGURE 9 | Key 1H-1H COSY and HMBC correlations of compounds 9-10.with the correlations of H-2”’/2”’-Me, and H-1”’/ C-10, C-7 inHMBC, these revealed the presence of a 1,3-dioxane throughacetalation between OH-7, OH-10 and C-1.”’ It meant that 1,3-dioxane was formed by hemiacetal formation from hydroxylgroups of C-7 and C-10 and aldehyde of 2-methyl-heptaldehyde.The relative configuration of 5 was similar to patriscabioinsA-L and patriscabiobisins A and B (Liu et al., 2017a,b) from P.scabiosaefolia through comparison with their ROESY and CDspectrum (Figure 11). The ROESY cross peaks of H-1”’/H-8 and2”’-Me/H-1”’ suggested the α-orientation of H-1”’ and 2”’-Me.Thus, the structure of 5 was identified as shown and namedpatrinoside B.Compound 6 had the molecular formula C25H38O11 based onits quasimolecular ion peak at m/z 537.2302 [M + Na]+ (calcd for537.2306) in HR-ESI-MS spectrum and 13C-NMR data (Table 2).Comparison of the 13C NMR and DEPT spectra of 6 with thoseof 5 revealed that 6 shared the same aglycone, a glucose at C-11, and the 3-methylcrotonyl group at C-1 as compound 5, andthe significant difference between them was the disappearance oftwo methylenes, one methine and one methyl in the high fieldregion, and the molecular weight of 6 was 56 less than that of5. This indicated that the fragment of 2-methyl-heptaldehyde incompound 5 was changed to n-butyraldehyde in compound 6,which was further verified by the correlations from H-1”’ (δH4.52) to C-2”’ (δC 38.2), and H-4”’ (δH 0.90) to C-2”’ (δC 38.2), C-3”’ (δC 18.2) in HMBC. Therefore, the structure of 6 was deducedas shown and named patrinoside C.The molecular formula of 7 was identified as C26H40O11 fromHR-ESI-MS at m/z 551.2467 [M + Na]+ (calcd for 551.2463).Comparing 1H-NMR and 13C-NMR data of 7 with those of 6,it was found that they shared similar NMR data, and the obviousdistinction between them was the appearance of a methylene (δC34.8) and the change in chemical shift of methyl signal of C-4’ (δC27.6 → 12.3). It indicated that a 3-methylcrotonyl group at C-1in 6 was replaced by the 3,4-dimethylcrotonyl group in 7. Thiswas ultimately confirmed by 1D and 2D NMR spectra. In HMBCspectrum, the correlations between H-2’ (δH 5.68) and C-4’ (δC34.8) and C-5’ (δC 12.3) provided solid evidence for the existenceof the 3,4-dimethylcrotonyl group. Finally, the structure of 7 waselucidated as shown and named patrinoside D.Compound 8 was analyzed to have a molecular formulaof C22H34O11 by HR-ESI-MS at m/z 497.1988 [M + Na]+Frontiers in Chemistry | www.frontiersin.org10March 2021 | Volume 9 | Article 657028Liu et al.Nine Unique Iridoids and GlycosidesFIGURE 10 | The possible pathway for the transformation of compounds 1-10.(calcd for 497.1993), whose molecular weight is 54 less thanthat of 7. Careful analysis of its NMR data found that ithad the same aglycone as compound 7, as well as a glucoseat C-11 and a 3,4-dimethylcrotonyl group at C-1, but lackedan oxymethine at δC 102.0, two methylenes and a methyl,which suggested that it did not have a 1,3-dioxaneformed atC-7 and C-10. This could be further verified by the chemicalshifts at C-7 and C-10 which were shifted up field from δC79.2 to 72.7 and form δC 67.1 to 62.1, respectively, comparedwith compound 7. Finally, the structure of compound 8 wascharacterized as (1S,5S,7S,8S,9S)-1-O-(3,4-dimethylcrotonyl)-7,10-dihydroxy-11-β-D-glucose-5,6-dihydrovaltrate hydrin andnamed patrinoside E.Compound 9 was found to have a molecular formulaof C21H32O10 by HR-ESI-MS at m/z 467.1889 [M + Na]+(calcd for 467.1888). Careful analysis of its 1D and 2DNMR data led to finding that it was a 7,10,11-trihydroxy-3-envaltratehydrin iridoid with a 3-methylcrotonyl group atC-1 and a sugar. The 1H-1H COSY correlations (Figure 9)showed a fragment: CH (H-1”, δH 4.22)—CH (H-2”, δH3.09)—CH(H-3”,δH3.59)—CH(H-4”,δH3.52)—CH2(H-5”, δH1.35)—CH2(H-6”, δH3.56). Combined withthe correlations between H-5” (δH1.35) and C-2” (δC76.9) and C-3” (δC 72.2) in the HMBC, all these signalshinted the sugar was 5-deoxyglucose, which was linkedto C-11 from the correlation of H-11 (δH 4.06, 4.24)/C-1”(δC103.8).Theβ-configurationwasdeterminedbythe coupled constant of H-1” (δH 4.22, d, J = 7.8 Hz).Ultimately, the structure of compound 9 was confirmed as(1S,5S,7S,8S,9S)-1-O-(3-methylcrotonyl)-7,10-dihydroxy-11-(5-deoxy-β-D-glucofuranose)-5,6-dihydrovaltrate hydrin andnamed patrinoside F.Compound 10 was formulated as C31H44O14 from HR-ESI-MS at m/z 663.2627 [M + Na]+ (calcd for 663.2623) and13C-NMR data (Table 3). The carbon resonances at δC 103.3(d), 75.1 (d), 78.1 (d), 71.7 (d), 78.8 (d), 62.8 (t) and theanomeric proton signals at δH 4.27 (d, J = 7.8 Hz) suggestedthat compound 10 contained a glucose at C-11 by the correlationfrom the anomeric proton (δH 4.27) to C-11 (δC 69.6) in HMBC(Figure 9). The remaining carbon signals showed two iridoidunits, which means it was a bis-iridoidglycoside. Detailed analysisof the NMR data at δC 93.3 (d), δC 139.9 (d), δC 117.2 (s),δC 77.9 (d), δC 66.9 (t), δC 69.6 (t), and δH 5.80 (d, 5.7),combined with the 3-methylcrotonyl [δC 166.3 (s), δC 161.3(s), δC 116.0 (d), δC 27.6 (q), δC 20.6 (q)] at C-1, found oneiridoid was patrinoside A (Liu et al., 2019). The other iridoidwas similar to 8,9-didehydro-7-hydroxydolichodial (Georg andJoerg, 1985) through the conjugated aldehyde and double bondat δC 196.9 (d), δC 151.7 (s), δC 134.1 (t), two methylene atδC 39.4 (t), and δC 63.5 (t), and four methynes at δC 37.5Frontiers in Chemistry | www.frontiersin.org11March 2021 | Volume 9 | Article 657028Liu et al.Nine Unique Iridoids and GlycosidesFIGURE 11 | Experimental ECD spectra of compounds 2, 4-10 at TDDFT/B3LYP/6-31G(d) level.FIGURE 12 | The cytotoxic activities against five human cancer cell lines of compounds 1-10.(t), δC 73.2 (d), δC 46.1 (d), δC 42.1 (d) in 13C NMR, exceptfor the absence of one methyl, one pair of double bonds, andan aldehyde group. However, it did include an oxygenatedmethylene, two methines, and an acetal at δC 102.3 (d). Thissuggested that the other iridoid was 7,10-dihydroxydolichodial.And two units were linked through the aldehyde group betweenthe aldehyde group at C-1’ of 7,10 -dihydroxydolichodial andhydroxyl groups at C-7 and C-10 of patrinoside A, which wasfurther verified by the correlations of H-1’/C-7, C-10 in theHMBC spectrum. The α-orientations of H-1”’, H-8”’, and H-9”’were determined by the ROESY correlations of H-7”’ with H-1”’, H-8”,’ and H-9”’. Additionally, these were further proven bythe good agreement of ECD spectra between compound 10 andpatriscabiobisin A and B which were similar to compound 10.Frontiers in Chemistry | www.frontiersin.org12March 2021 | Volume 9 | Article 657028Liu et al.Nine Unique Iridoids and GlycosidesThus, the structure of compound 10 was characterized as shown,namely patriscabiobisin C.In conclusion, compounds 1-10 were a series of 5,-dihydrovaltrate hydrins with characteristic substituents, suchas 3-methylcrotonyl group or 3,4-dimethylcrotonyl groupin the Valerianaceae family, belonging to iridoids derivedfrom iridodial. Iridoial was enoxylated and then subjectedto intramolecular hemiacetal formation to form iridoid. Andiridoid suffered a series of chemical changes to achieve diversecompounds (1-10) (Figure 10). Iridodial was oxidized and thenunderwent the hemiacetal formation between C-1 and C-10to form the deformed iridoid—compound 1 with 5/5 rings.Iridoid was oxidized and combined with different substituentsto constitute compounds 2, 4, 8-9, and when it was dehydrationcondensed between C-10 and C-3 to generate a new ring,like compound 3. However, when the hydroxyl of C-7 andC-10 of iridoid combined with the aldehyde compounds, a1,3-dioxane group would be formed, such as compounds 5-7.Furthermore, if the aldehyde group came from an iridoid, thenit would generate a bis-iridoid, like compound 10. And theconfigurations of compounds 1-10 were further confirmed byECD (Figure 11).Finally, the cytotoxic activities of these compounds wereevaluated against five human cancer cell lines (HL-60, A-549,SMMC-7721, MCF-7, and SW-480). Unfortunately, none of thecompounds showed significant cytotoxicities at 40 µM, exceptfor compound 5 which showed the cell inhibition of 102.42,95.13, 73.07, and 80.93% against HL-60, SMMC-7721, MCF-7,and SW-480 respectively (Figure 12).CONCLUSIONP. scabiosaefolia is a medical and edible Chinese herb withhigh nutritional value and a wide range of biological activities.Research showed plants in the Patrinia genus are rich in iridoidsand terpenoids. In our ongoing study, we found a series ofiridoids and iridoid glycosides, including three new iridoids(1-3) and six novel irioid glycosides (5-10), and one knowncompound (4). Among them, compound 1 was a deformediridoid, compound 3 formed a cycle between C-3 and C-10,compounds 5-7 with a 1,3-dioxane between C-7 and C-10,and compound 10 was a bis-iridoid glycoside, which was thefirst reported in P. scabiosaefolia. Cytotoxicity assays foundcompound 5 showed good cell inhibition against HL-60, SMMC-7721, MCF-7, and SW-480. All these results enriched the studyon the chemical constituents and activities of Patrinia genus.However, the activities of these compounds were not thorough,just the evaluation of the cytotoxicity assays, and the otheractivities and their mechanisms are worth further exploration.DATA AVAILABILITY STATEMENTThe original contributions presented in the study are includedin the article/Supplementary Material, further inquiries can bedirected to the corresponding author/s.AUTHOR CONTRIBUTIONSZL, JH, and WK conceived the research subject. ZL, YN,LZ, LM, SC, and MW contributed to the developmentand writing of the manuscript. WK and ZL contributedinvalidating,reviewing,andsupervisingtheproject.All authors contributed to the article and approved thesubmitted version.FUNDINGThis work was supported by grants from the NationalNatural Science Foundation of China (31900292), Scienceand Technology Development Program of Henan Province(202102110149, 212102110469), and the Science and TechnologyProject of Kaifeng (1908005).SUPPLEMENTARY MATERIALThe Supplementary Material for this article can be foundonlineat:https://www.frontiersin.org/articles/10.3389/fchem.2021.657028/full#supplementary-materialREFERENCESBruhn, T., Schaumlöffel, A., Hemberger, Y., and Bringmann, G. (2013). Specdis:quantifying the comparison of calculated and experimental electronic circulardichroism spectra. Chirality 25, 25243–25249. doi: 10.1002/chir.22138Delectis Flora Reipublicae Popularis Sinicae Agendae Academiae Sinicae Edita.(1986). Flora of China. Beijing: Science Press. 73:006.Georg, S., and Joerg, V. (1985). Valepotriat-artefakte aus centranthus ruber (L.) Dc.Arch. Pharm. 318, 515–519. doi: 10.1002/ardp.19853180607Kim, J. S., and Kang, S. S. (2013). Chemical constituents of plants from the genusPatrinia. Nat. Prod. Sci. 19, 77–119. Available online at: http://www.e-nps.or.kr/.Liu, Z. H., Hou, B., Yang, L., Ma, R. J., and Li, J. Y., Hu, et al. (2017a).Iridoids and bis-iridoids from Patrinia scabiosaefolia. RSC Adv. 7, 24940–24949.doi: 10.1039/C7RA03345ALiu, Z. H., Ma, R. J., Yang, L., Li, J. Y., Hu, J. M., and Zhou, J. (2017b).Triterpenoids and iridoids from Patrinia scabiosaefolia. Fitoterapia 119,130–135. doi: 10.1016/j.fitote.2017.04.011Liu, Z. H., Xu, L. T., Xu, X. Q., Niu, Y., Saadeldeen, F. S. A., and Kang, W. Y. (2019).Effects and mechanisms of iridoid glycosides from Patrinia scabiosaefolia.FoodChem.Toxicol.134,110806–110811.doi:10.1016/j.fct.2019.110806Monks, A., Scudiero, D., Skehan, P., Shoemaker, R., and Paull, K., Vistica,et al. (1991). Feasibility of a high-flux anticancer drug screen using a diversepanel of cultured human tumor cell lines. J. Natl. Cancer Insit. 83, 757–766.doi: 10.1093/jnci/83.11.757Reed, L. J., and Muench, H. (1938). A simple method of estimatingfiftypercentendo-points.Amer.J.Hygiene.27,493–797.doi: 10.1093/oxfordjournals.aje.a118408Veith, J., Schneider, G., and Lemmer, B. (1986). The influence of some degradationproducts of valepotriates on the motor activity of light-dark synchronizedmice.PlantaMed.3,179–183.doi:10.1243/PIME_PROC_1954_168_069_02Xiao, M., Zhu, S., and Zhang, A. (2007). Research advance in medicinal andedible utilization of plants of Parinia. J. Jinling Inst. Technol. 23, 83–86.doi: 10.3969/j.issn.1672-755X.2007.03.021Frontiers in Chemistry | www.frontiersin.org13March 2021 | Volume 9 | Article 657028Liu et al.Nine Unique Iridoids and GlycosidesZhong, S., Li, G., Lin, H., Jin, S., and Qian, X. (2001). An analysis of nutrientconstituents of wild Patrinia villosa. Chin. Wild Plant Resour. 20, 45–46.doi: 10.3969/j.issn.1006-9690.2001.01.019Conflict of Interest: The authors declare that the research was conducted in theabsence of any commercial or financial relationships that could be construed as apotential conflict of interest.Copyright © 2021 Liu, Niu, Zhou, Meng, Chen, Wang, Hu and Kang. This is anopen-access article distributed under the terms of the Creative Commons AttributionLicense (CC BY). The use, distribution or reproduction in other forums is permitted,provided the original author(s) and the copyright owner(s) are credited and that theoriginal publication in this journal is cited, in accordance with accepted academicpractice. No use, distribution or reproduction is permitted which does not complywith these terms.Frontiers in Chemistry | www.frontiersin.org14March 2021 | Volume 9 | Article 657028 | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/362269900Is Habitual Sleep Duration an Important Factor for Face Recognition Memoryin Young Adults? A Cross-sectional StudyArticle in Journal of Clinical and Diagnostic Research · July 2022DOI: 10.7860/JCDR/2022/55661.16654CITATIONS0READS65 authors, including:Some of the authors of this publication are also working on these related projects:ICMR - STS - Anemia and Obesity View projectEthical Challenges in Imparting Medical Education View projectPrerna AgarwalGovernment Institute of Medical Sciences23 PUBLICATIONS 33 CITATIONS SEE PROFILEAll content following this page was uploaded by Prerna Agarwal on 26 July 2022.The user has requested enhancement of the downloaded file.Journal of Clinical and Diagnostic Research. 2022 Jul, Vol-16(7): CC27-CC302727DOI: 10.7860/JCDR/2022/55661.16654Original ArticlePhysiology SectionIs Habitual Sleep Duration an Important Factor for Face Recognition Memory in Young Adults? A Cross-sectional StudyIntrOductIOnSleep is as important for our wellbeing as are exercise and nutrition. It is integral to the process of maintenance and restoration of physical as well as mental health encompassing intelligence, memory as well as emotions [1,2]. Much of the present knowledge about sleep comes from the study of effect of sleep deprivation on the body [3]. But there is wide inter individual variability of these effects of sleep deprivation [4]. That puts forth the question about how much sleep is adequate sleep for any person. Is there any deviation from the normal functions of the body and brain in individuals with different habitual sleep durations? The phenomenon of memory is intriguing. Past researches have both proved and disproved the association between sleep and memory [1,2]. Present consensus is that memories are indeed consolidated during sleep [5]. But is there any difference in this neural function in those with different habitual sleep durations? Now that lifestyles of young people have changed and there are odd sleeping patterns [6], it becomes more pertinent to delve deeper into the relationship.Face recognition memory is a special form of memory that gives the ability to remember people meaningfully and is instrumental in any individual’s meaningful coexistence in society. Our role in society becomes unimaginable if we do not have this ability. For the same reason, the neural processing for familiar and unfamiliar face recognition memories are different. More areas of the brain may be involved in familiar face recognition memory while unfamiliar face recognition memory may be a purer representative of the main neural sites involved in processing of this type of episodic memory [7]. Research has also highlighted that women may have a better face recognition memory than men [8].Against this background, present study explored if individuals with different habitual sleep durations have any difference in unfamiliar face recognition memory and if any gender differences may be identified in the association of these variables with each other; unfamiliar face recognition memory may be a better indicator of relationship between habitual sleep duration and face recognition memory, as it removes the confounding effect of involvement of some areas of the brain in the neural process of face recognition memory and thereby, presents a comparatively simpler picture of the same. The present study was done with an aim to assess if unfamiliar face recognition memory is different among habitually average-duration sleepers and long-duration sleepers among healthy young adults, and to identify gender differences in unfamiliar face recognition memory is different among habitually average-duration sleepers and long-duration sleepers among healthy young adults.MAterIAls And MethOdsThis was an observational analytic cross-sectional studyconducted in for a month July 2012 in B. J. Medical College, Ahmedabad, Gujarat, India, including students, aged 17-20 years, of 1st year of different professional courses. A written informed consent was obtained from all participants, confidentiality of their identity and data was maintained; the study was approved by the departmental scientific committee. It did not involve any kind of harm to the subjects/participants and strictly abided by the Helsinki Declaration of 1975 revised in 2000.Prerna agarwal1, geeta B nair2, Dalia BiSwaS3, avinaSh B takSanDe4, JayenDraSinh M JaDeJa5Keywords: Facial recognition, Mental recall, Sleep deprivationABstrActIntroduction: Sleep is integral to the process of maintenance and restoration of physical as well as mental health encompassing intelligence, memory as well as emotions. Face recognition memory is a special form of memory that gives the ability to remember people meaningfully and is instrumental in any individual’s meaningful coexistence in the society. Aim: To determine if different habitual sleep durations may be associated with difference in unfamiliar face recognition memory and if there are any gender differences in the same, among young healthy individuals.Materials and Methods: This was an observational analytic cross-sectional study conducted over a month in B.J. Medical College, Ahmedabad, Gujarat, India, including 98 healthy young volunteers (39 females and 59 males) among 1st year students of different professional courses. The participants were subjected to a questionnaire-based interview followed by a face recognition test based on Parkin’s recognition memory test (1995). Mean values and Standard Deviations (SD) were calculated for age, sleep duration, correct and wrong responses in the test and statistical significance was tested by applying t-test with a significance level of p<0.05.results: The participants were grouped as average-duration (6.6±0.5 hours/day) and long-duration (8.4±1.0 hours/day)sleepers, 53 and 45 in number, respectively, and there was a significant difference in the sleep duration of the two groups (p=8.6×10-17). No significant difference was seen in the number of correct (p=0.08795) and incorrect (p=0.42205) responses of average-duration sleepers and the long-duration sleepers. Gender differences in correct and incorrect responses of average-duration and long-duration sleepers were also not significant. But there was a significant difference in the number of correct responses by females and males, in general.conclusion: Habitual duration of sleep may not affect face recognition memory in the young ages.Prerna Agarwal et al., Habitual Sleep Duration and Face Recognition Memory www.jcdr.netJournal of Clinical and Diagnostic Research. 2022 Jul, Vol-16(7): CC27-CC302828A total of 120 healthy volunteers were interviewed about their habitual sleep duration with the help of a questionnaire. Data about the habitual duration was collected and duration of sleep at night and during day, sleep disturbance, daytime sleepiness, history of any present or past illness and family history of any serious health condition were collected and used for excluding participants to reduce bias/confounding factors.inclusion criteria: Healthy young adults aged 17-20 years and willing to participate in the study. exclusion criteria: Those having any health condition including sleep disturbance were excluded and also, the participants who reported having inadequate sleep on the day of test were excluded.The interview was followed by a face recognition test. The test was based on Parkin et al., Recognition Memory Test (RMT) [9,10]. The participants were given a slide show, projecting 14 colour photographs of unfamiliar female faces before them for two seconds; each photograph was followed by a blank slide with grey background. Thereafter, for the next 30 minutes, they were kept busy with the activity of filling a form for test to refrain them from actively rehearsing the photos in their memory. Then, they were given a test slide show and were asked to identify from it the photographs which appeared to them to have been seen in the previous slide show. The test slide show had 20 photographs that included 10 photographs from the main slideshow interspersed with 10 new photographs of unfamiliar female faces. Each photo was projected for one second. Two photographs in starting and at end of the main slideshow were excluded from the test slideshow to do away effects of the previous immediate memory effect and recent memory, respectively. The score was given according to the number of photographs identified correctly (maximum score-10, minimum score-0). Finally, data of 98 Characteristictotal (98) (mean±SD)Females (39) (mean±SD)Males (59) (mean±SD)p-value by t-test (*Significant p<0.05)Age (years)17.6±0.617.5±0.617.8±0.60.071Usual duration of sleep (hours/day)7.4±1.26.9±1.07.7±1.20.0007*Duration of sleep on test day (hours/day)7.2±1.46.8±0.97.5±1.50.01*habitual sleep duration and sleep duration on day of testgenderUsual duration of sleep (hours/day)Duration of sleep on test day (hours/day)p-value by t-test (*Significant p<0.05)Females (39) (mean±SD)6.9±1.06.8±0.90.6438Males (59) (mean±SD)7.7±1.27.5±1.50.4255Total (98) (mean±SD)7.4±1.27.2±1.40.2843habitual sleep durations of average-duration sleepers and long-duration sleepersgenderaverage- duration sleepers (5-7 hours/day) (53=28 females + 25 males)long-duration sleepers (>7 hours/day) (45=11 females + 35 males)p-value by t-test (*Significant p<0.05)Females (39) (mean±SD)6.4±0.58.2±0.55.6×10-8*Males (59) (mean±SD)6.7±0.48.4±1.19.0×10-11*Total (98) (mean±SD)6.6±0.58.4±1.08.6×10-17*[table/Fig-1]: Baseline characteristics of the participants.Participants5-7 hours’sleep/day (mean±SD)>7 hours’ sleep/day (mean±SD)p-value by t-test (*Significant p<0.05)Correct responsesFemales (39)8.5±1.4 (n=28)8.9±1.4 (n=11)0.41846Males (59)7.96±1.5 (n=25)7.2±2.0 (n=34)0.1019Total (females + males) (98)8.3±1.5 (n=53)7.6±2.0 (n=45)0.08795incorrect responsesFemales (39)2.2±1.7 (n=28)2.3±1.9 (n=11)0.93009Males (59)2.4±1.5 (n=25)2.7±1.8 (n=34)0.53777Total (females + males) (98)2.3±1.6 (n=53)2.6±1.8 (n=45)0.42205[table/Fig-2]: Comparison of responses of average-duration sleepers and long-duration sleepers.responses of female and male average-duration sleepersFemales (n=28)Males (n=25)p-value by t-test (*Significant p<0.05)Number of pictures identified correct8.5±1.47.96±1.50.1813participants were analysed after excluding data from those whose responses to the test were incomplete. About 5-7 hour of sleep per day is adequate for most people [11]. So, the participants were grouped as average-duration sleepers, those sleeping for 5-7 hours and long-duration sleepers, those sleeping for >7 hours. No upper limit as such- long duration sleepers more than 7 hours per day - maximum reported sleep time was 13 hours.stAtIstIcAl AnAlYsIsThe mean and SD were calculated for age, sleep duration, correct responses and wrong responses. The difference of means was tested for statistical significance by applying t-test at a significance level of p<0.05. Data was analysed using Microsoft Excel (version 2019). resultsThe baseline characteristics of the participants are given in [Table/Fig-1]. The 98 participants included 39 females (39.8%) and 59 males (60.2%). Among them, 53 were average-duration sleepers (54.1%) (28 females and 25 males) and 45 were long-duration sleepers (45.9%) (11 females and 34 males). There was a significant difference in the habitual sleep durations of average-duration sleepers (5-7 hours/day) and long-duration sleepers (>7 hours) [Table/Fig-1].No significant difference was seen in the number of correct and incorrect responses of average-duration sleepers and the long-duration sleepers [Table/Fig-2].Gender differences in correct and incorrect responses of average and long sleepers were also not significant [Table/Fig-3]. But there was a significant difference in the number of correct responses by females and males, in general. Yet the difference was insignificant for number of incorrect responses.www.jcdr.net Prerna Agarwal et al., Habitual Sleep Duration and Face Recognition MemoryJournal of Clinical and Diagnostic Research. 2022 Jul, Vol-16(7): CC27-CC302929dIscussIOnIn present study, there is no significant difference in the numbers of correct and wrong responses of average-duration sleepers and the long-duration sleepers; similar results were reported by Hermits and Rehnman; Nepal DB et al., and Sheth BR et al., in their studies [8,12,13].Rasch B and Born J reviewed the role of sleep in the process of consolidation of both declarative and implicit types of memory and found them to be positively associated with each other [14]. Ashton R reviewed the association between sleep duration and working memory to be inconclusive [15]. Beattie L et al., found poor sleep to be associated with perceptual impairment of face recognition [16]. Systematic review by Devore EE et al., concluded cognitive impairment to be associated with longer sleep duration more than with shorter sleep duration [17]. Van Oostrom SH et al., have indicated an inverted U-shaped curve for cognitive performance, and for other parameters so affected, with respect to habitual sleep duration among older adults: both long and short duration of sleep are associated with reduced neurocognitive function and alertness [18]. Khalsa S et al., found both gray matter and white matter changes in frontal and temporal areas of the brain among short sleepers in their study [19]. Perceptibly, these are the areas involved in the functions of attention, memory, including face recognition memory, and cognition [7,19]. If the result of present study are examined in the light of these findings and those of other such studies, it is found that it is rather indicative of inter-individual variability in quantitative requirement of sleep for healthy neurocognitive function. In that case, there is a need to explore how an individual spends her/his waking hours, what levels of mental and exhaustions are felt by the individual with that and how these relate to their sleep duration and neurocognitive function. Kronholm E et al., and Mograss MA et al., have reported a similar conclusion from their study [4,20]. Elaborate studies on these aspects, spending of waking hours and level of mental exhaustion, may shed more light on the phenomena of sleep and memory and help to understand the complexities of the functions of sleep and mechanisms of memory.Another aspect of present study that makes us reserve our deductions from it is that of the age of the study participants. They have been young adults, while in most of the earlier relevant studies have been done on older individuals [16-19, 21]. Ageing affects other cells of the body and neurons alike. With respect to these facts, the results of present study work imply that younger neurons being not as vulnerable, it is a cumulative effect of sleep pattern over the years that affects memory. However, here too contrary results were obtained by Richards A et al., they studied a wide range of age groups from 15 years to 89 years and found that longer duration of sleep was associated with poorer performance among the young [22].The above two deductions indirectly emphasise the effect of sleep debt in terms of both quality and quantity on neural functions, including memory. Study by Beattie L et al., supports the fact [16]. While in present study, the aim being to analyse face recognition memory variability with different habitual sleep durations, authors had excluded any individual that reported poor sleep from participating in it.There was no significant difference in the number of correct and number of incorrect responses of average-duration and long-duration female and male sleepers. Yet, there was a significant difference in the number of correct responses by females and males, in general; again, not in the number of incorrect responses by them. This indicates that females may have better face recognition memory than males. Similar results were also noted in the studies of Herlitz A and Rehnman J [8].Women are perhaps better at episodic memory tasks, verbal ability and also, they are seemingly more interested in faces than men that results in women performing better at face recognition tasks [8,23-25]. Also, studies have suggested that women have an own-gender bias [26,27]. This aspect too could have contributed to present study result because authors used unfamiliar female faces in the test.The strength of present study lies in the young age of the participants when the neuronal circuits involved in face recognition memory are mostly unaffected by ageing which thereby does away with the confounding effect of age on face recognition memory.limitation(s)The main limitation present study faced was of not having any short-duration sleepers. Having a sample that better represented the general population could have led to better results. Having a limited number of participants too reserves the generalisability of present study findings. Authors have used unfamiliar female faces in test; so, the interpretation of females having a better face recognition memory than men may be biased.cOnclusIOn(s)It was concluded that that habitual duration of sleep may not be relevant for face recognition in young age. Sleep debt, in terms of both quantity and quality, may be the determining factor in defining the relationship between sleep and face recognition memory, and memory in general. Also, how waking hours are spent and how that affects the individual differences in habitual healthy sleep durations needs to be explored to shed more light on neural health in terms of sleep and memory. Future studies that explore these aspects may bring insight into the relationship of these essential and intriguing variables of human biology.reFerences Zielinski MR, McKenna JT, McCarley RW. Functions and mechanisms of sleep. [1]AIMS Neurosci. 2016;3(1):67-74. Worley SL. The extraordinary importance of sleep: The detrimental effects of [2]inadequate sleep on health and public safety drive an explosion of sleep research. P T. 2018;43(12):758-63. Banks S, Dinges DF. Behavioral and physiological consequences of sleep [3]restriction. J Clin Sleep Med. 2007;3(5):519-28. Kronholm E, Sallinen M, Suutama T, Sulkava R, Era P, Partonen T. Self-reported [4]sleep duration and cognitive functioning in the general population. J Sleep Res. 2009;18(4):436-46. Born J, Wilhelm I. System consolidation of memory during sleep. Psychol Res. [5]2012;76(2):192-03. Wang F, Bíró É. Determinants of sleep quality in college students: A literature [6]review. Explore. 2021;17(2):170-77. Lopatina OL, Komleva YK, Gorina YV, Higashida H, Salmina AB. Neurobiological aspects [7]of face recognition: The role of oxytocin. Front Behav Neurosci. 2018;12:195. Number of pictures identified wrong2.2±1.72.4±1.50.6534responses of female and male long-duration sleepersFemales (n=11)Males (n=34)p-value by t-test (*Significant p<0.05)Number of pictures identified correct8.9±1.47.2±2.00.6417Number of pictures identified wrong2.3±1.92.7±1.80.9058responses of male and female participantstotal (98) (mean±SD)Females (39) (mean±SD)Males (59) (mean±SD)p-value by t-test (*Significant p<0.05)Number of pictures identified correct8.0±1.88.6±1.47.5±1.80.001*Number of pictures identified wrong2.5±1.72.2±1.72.6±1.70.305[table/Fig-3]: Gender differences in responses of average-duration and long-duration sleepers.Prerna Agarwal et al., Habitual Sleep Duration and Face Recognition Memory www.jcdr.netJournal of Clinical and Diagnostic Research. 2022 Jul, Vol-16(7): CC27-CC303030 Herlitz A, Rehnman J. Sex differences in episodic memory. Curr Dir Psychol Sci. [8]2008;17(1):52-56. Parkin AJ, Walter BM, Hunkin NM. Relationships between normal aging, frontal [9]lobe function, and memory for temporal and spatial information. Neuropsychology. 1995;9(3): 304-12. Harrison Y, Horne JA. Sleep loss and temporal memory. Q J Exp Psychol A. [10]2000;53(1):271-79. Watson NF, Badr MS, Belenky G, Bliwise DL, Buxton OM, Buysse D, [11]et al. Joint consensus statement of the American Academy of Sleep Medicine and Sleep Research Society on the recommended amount of sleep for a healthy adult: Methodology and discussion. J Clin Sleep Med. 2015;11(8):931-52. Nepal DB, Kumar T, Mandal MB, Deshpande SB. Sleep duration of an individual [12]has adverse influence on auditory episodic memory. Indian J Physiol Pharmacol. 2007;51(4):326-32. Sheth BR, Nguyen N, Janvelyan D. Does sleep really influence face recognition [13]memory? PLOS ONE. 2009;4(5):e5496. Rasch B, Born J. About sleep’s role in memory. Physiol Rev. 2013;93: 681-66. [14] Ashton R. The relationship between sleep duration and working memory in [15]children. University College of London. 2016. Beattie L, Walsh D, McLaren J, Biello SM, White D. Perceptual impairment in face [16]identification with poor sleep. R Soc Open Sci. 2016;3(10):160321. Devore EE, Grodstein F, Schernhammer ES. Sleep duration in relation to [17]cognitive function among older adults: A systematic review of observational studies. Neuroepidemiology.2016;46(1):57-78. Van Oostrom SH, Nooyens ACJ, vanBoxtel MPJ, Verschuren WMM. Long sleep [18]duration is associated with lower cognitive function among middle-age adults - the Doetinchem Cohort Study. Sleep Med. 2018;41:78-85. Khalsa S, Hale JR, Goldstone A, Wilson RS, Mayhew SD, Bagary M, et al. Habitual [19]sleep durations and subjective sleep quality predict white matter differences in the human brain. Neurobiol Sleep Circadian Rhythms. 2017;3:17-25. Mograss MA, Guillem F, Stickgold R. Individual differences in face recognition [20]memory: Comparison among habitual short, average, and long sleepers. Behav Brain Res. 2010;208(2):576-83. Lee HJ, Oh KS, Kim T, Lee CS, Jeong J, Youn SY, et al. Prevalence, risk factors, [21]and impact of excessive daytime sleepiness in an elderly Korean population. Sleep Med Res. 2014;5(2):54-61. Richards A, Inslicht SS, Metzler TJ, Mohlenhoff BS, Rao MN, O'Donovan A, et [22]al. Sleep and cognitive performance from teens to old age: More is not better. Sleep. 2017;40(1):zsw 029. Winters KM. Gender Differences in Facial Recognition. California State University, [23]Chico. CDR: Chico Digital Repository. 2010. Loven J. Gender differences in face recognition: the role of interest and friendship. [24]2006. https://www.diva-portal.org/smash/get/diva2:189430/FULLTEXT01.pdf. Lewin C, Wolgers G, Herlitz A. Sex differences favoring women in verbal but not [25]in visuospatial episodicmemory. Neuropsychology. 2001;15(2):165-73. Herlitz A, Loven J. Sex differences and the own-gender bias in face recognition: [26]A meta-analytic review. Visual Cognition. 2013. 21:9-10,1306-36. Lewin C, Herlitz A. Sex differences in face recognition--women's faces make the [27]difference. Brain Cogn.2002;50(1):121-28. Doi:10.1016/s0278-2626(02)00016-7.PartiCUlarS OF COntriBUtOrS:1. Assistant Professor, Department of Physiology, Government Institute of Medical Sciences, Greater Noida, Uttar Pradesh, India.2. Associate Professor, Department of Physiology, B. J. Medical College, Ahmedabad, Gujarat, India.3. Professor, Department of Physiology, Jawaharlal Nehru Medical College, Datta Meghe Institute of Medical Sciences, Sawangi (Meghe), Wardha, Maharashtra, India.4. Associate Professor, Department of Physiology, Jawaharlal Nehru Medical College, Datta Meghe Institute of Medical Sciences, Sawangi (Meghe), Wardha, Maharashtra, India.5. Professor, Department of Physiology, Dungarpur Medical College, Dungarpur, Rajasthan, India.PlagiariSM CheCking MethODS: [Jain H et al.]• Plagiarism X-checker: Feb 19, 2022• Manual Googling: May 07, 2022• iThenticate Software: Jun 30, 2022 (6%)etyMOlOgy: Author OriginnaMe, aDDreSS, e-Mail iD OF the COrreSPOnDing aUthOr:Dr. Prerna Agarwal,Department of Physiology, Government Institute of Medical Sciences, Greater Noida, Uttar Pradesh, India.E-mail: dr.preranaagarwal@gmail.comDate of Submission: Feb 13, 2022Date of Peer Review: Mar 25, 2022Date of Acceptance: May 23, 2022Date of Publishing: Jul 01, 2022aUthOr DeClaratiOn:• Financial or Other Competing Interests: None• Was Ethics Committee Approval obtained for this study? No• Was informed consent obtained from the subjects involved in the study? Yes• For any images presented appropriate consent has been obtained from the subjects. NAView publication stats | RESEARCH GATE |
Volume 6 | Issue 2 | December 2022 https://doi.org/10.3126/njz.v6i2.51877 Nepalese Journal of Zoology 6(2):1–16 1 ISSN: 2705-4403 (Print) & 2705-4411 (Online) www.cdztu.edu.np/njz Research Article Influence of environmental factors on bird diversity in and around Kahundanda Hillscape, Pokhara, Nepal Shubhas Chandra Bastola1,2# | Jagan Nath Adhikari1,3# | Hemanta Dhakal2 | Bishnu Prasad Bhattarai1* 1 Central Department of Zoology, Institute of Science and Technology, Tribhuvan University, Kathamandu, Nepal 2 Department of Zoology, Prithivi Narayan Campus, Pokhara, Kaski, Nepal 3 Department of Zoology, Birendra Multiple Campus, Bharatpur, Chitwan, Nepal # These two authors contributed equally in this work * Correspondence: bishnu.bhattarai@cdz.tu.edu.np Abstract Birds are the indicators of the ecosystem health. Several environment variables affect the diversity, distribution and migration of birds. Kahundanda of Pokhara Valley is one of the important hillscapes having different habitat types such as human settlements, forest, farmland, wetland and open areas. This study revealed the factors affecting the diversity and distribution of birds of Kahundanda. The direct observation by “point count” method was applied for the bird survey. A total of 49 points were fixed along the nine transects in every 200–250 m distance of the Kahundanda (elevation ranges from 700 m to 1424 m). In each point, we also recorded habitat types and disturbance factors. A total of 101 species of birds belonging to 18 orders and 41 families were reported from the study area. The species diversity was the highest in forest habitat (H' = 3.558) and least in open area (H' = 2.365). Among them, 58.42% of birds were resident and 41.58% were migratory including four globally threatened (one critically endangered, two endangered and one near threatened) species. Other factors such as the number of livestock present, distance from roads and settlement, environmental factors such as distance from the water resources, elevation and number of fruiting trees present significantly affected the distribution of the birds at Kahundanda. Thus, environmental factors such as habitat types, human disturbance and proximity to water resources are important for bird species conservation in the human-dominated landscape. Keywords: Disturbance; Habitat; Kahundanda; Threatened birds 1 | Introduction Nepal is rich in avifauna (891 species, 8.87% of the global species) due to physiography, geography and climatic differences (DNPWC & BCN 2019; BCN 2022). Elevation, land use and land cover patterns, settlements, anthropogenic disturbance, etc. determine the diversity and abundance of the birds. Small ranged species are habitat and topography specific than large ranged birds (Pan et al. 2016; Pan et al. 2019). Habitat heterogeneity hypothesis states that where there is a greater variety of habitat types per unit area with complexity of vegetation structures (Kim et al. 2018) that increases faunal diversity by providing good space for birds in an area (Kim et al. 2018; Iswandaru et al. 2020). Alpine habitat, forest and scrub, wetlands, grassland, agricultural land and human settlements are the key habitats found in Nepal (Basnet et al. 2016, Grimmett et al. 2016; Baral & Inskipp 2020). Diversity of habitats ranging from bare rock and scrub in the alpine zone to tropical forest in the lowlands, support the diverse bird species. In the agriculture land, excessive use of agrochemicals, intensive crop rotations, removal of non-crop plants from the agriculture field directly or indirectly affect the resource availability and habitat Suggested citation: Bastola, S. C., Adhikari, J. N., Dhakal, H. and Bhattarai, B. P. 2022. Influence of environmental factors on bird diversity in and around Kahundanda Hillscape, Pokhara, Nepal. Nepalese Journal of Zoology 6(2):1–16. https://doi.org/10.3126/njz.v6i2.51877 Article History: Received: 06 September 2022 Revised: 21 November 2022 Accepted: 24 November 2022 Publisher’s note: The statements, opinions and data contained in the publication are solely those of the individual author(s) and do not necessarily reflect those of the editorial board and the publisher of the NJZ. Copyright: © 2022 by the authors Licensee: Central Department of Zoology, Tribhuvan University, Kathmandu, Nepal. Nepalese Journal of Zoology 6(2) Bastola et al. 2 density in agroecosystem (Redlich et al. 2018). The food availability is an example of good habitat quality, for example, food availability in the wetland environment increases the diversity of wetland birds (Inskipp et al. 2017; Meer & Mirza 2017; Baral & Inskipp 2020) Birds and their diversity provide bio-indication of that ecosystem or landscape and also indicate the health and quality of ecosystem (Bregman et al. 2014). Anthropogenic drivers such as habitat destruction, fragmentation and degradation are responsible for the change in the bird’s community and diversity. These factors are directly affecting their breeding, feeding and nesting grounds that lead to local threats to the birds and their conservation (Bregman et al. 2014). Out of the 891 bird species recorded in Nepal, 19% (n = 168 species) have been assessed as nationally threatened of which 64 species are Critically Endangered, 44 Endangered, 64 Vulnerable and 22 are Data Deficient (Inskipp et al. 2016). A total of 43 species of birds are globally threatened; 43 are globally near threatened species followed by nine Critically Endangered (CR), nine Endangered (EN), 25 Vulnerable (VU) and eight are Regionally Extinct (RE) species: jungle bush quail (Perdicula asiatica), rufous-necked hornbill (Aceros nipalensis), silver-breasted broadbill (Serilophus lunatus), green cochoa (Cochoa viridis), black-breasted parrotbill (Paradoxornis flavirostris), pink-headed duck (Rhodonessa caryophyllacea) and white-bellied heron (Ardea insignis) in Nepal (Inskipp et al. 2016; DNPWC & BCN 2019; BCN 2022). Spiny babbler (Acanthoptila nipalensis) is the only endemic bird of Nepal. Nine species; Himalayan monal, cheer pheasant, satyr tragopan, Bengal florican, lesser florican, great hornbill, sarus crane, black stork and white stork are protected birds of Nepal according to NPWC act 1973. Currently, 71 species of birds are accepted as a vagrant bird species in Nepal (Inskipp et al. 2020). Over half (53%) of Nepal’s nationally threatened birds inhabit in forests, 27% inhabit in wetlands and 15% inhabit grasslands (DNPWC & BCN 2019). Various anthropogenic changes and developmental process in these days in Nepal are affecting the diversity/richness of the avifauna and their habitats. Several environmental variables change the diversity and distribution of the birds. Seasonal migration of the birds is one of the important evolutionary hypotheses that explains birth and death as well as life process (Ferreira & Perbiche-Neves 2021; Ghimire et al. 2021). Migratory birds can adapt with the residential birds and share the same habitat, food and breeding ground (Ghimire et al. 2021). Therefore, the interrelationship between migrant and resident birds plays as an important role in the breeding season. The heterospecific attraction hypothesis states that migrants use residents as a signal to identify suitable sites for breeding as the resident birds already occupy the higher quality sites (Kim et al. 2018). Hence, increasing migration may positively affect the species richness and density of the given sites (Ferreira & Perbiche-Neves 2021). Previous studies have shown that habitat of the birds such as forest, wetlands, grasslands, bush area have been threatened by various anthropogenic activities such as development of infrastructures, roads, dams in wetlands, pollutions, habitat encroachments, habitat degradation by invasive and alien species (Junk et al. 2013). Food and water scarcity, overgrazing and use of pesticides, pollution from households and industrial discharges and agricultural run-off is seriously degrading the habitat of birds which are posing serious threats to birds of Nepal (Inskipp et al. 2016). Though various studies have looked at diversity of birds but they have focused on lake system, habitats (Khatri et al. 2019). Studies on altitudinal gradients of the human-dominated landscape of the surrounding hills of Pokhara Valley are scarce. The study of the diversity of bird community in different habitats is essential to find out the community structure and composition. This type of study is also important for regional or local landscapes for avian conservation Kahundanda is the stepping stone habitat for the long-range migratory birds that migrate through the Annapurna Conservation Area (ACA) to the Chitwan National Park (CNP) in the lowlands. Besides, it also provides the refugium and nesting habitats for different species of birds of lake clusters of Pokhara Valley- a Ramsar site. It is also an important ecotourism destination of Pokhara Valley for nature and cultural tourists. The conservation status and distribution of birds in different environmental gradients in this area are still scarce. Thus, we aimed to explore (i) the community structure of birds of the Kauhdanda area (ii) environmental factors affecting the diversity and distribution of birds. 2 | Materials and methods 2.1 | Study area Pokhara is the largest metropolitan city of Nepal in terms of area and second-largest in terms of populations. This valley is situated in the western part of Nepal and one of the most popular tourist Nepalese Journal of Zoology 6(2) Bastola et al. 3 destinations in Nepal. Many tourists visit this valley to see the Himalayan range, biodiversity, caves and lakes. The study site Kahundanda is located in the north-east of the main city of Pokhara at 28.218800°N to 28.250745°N and 84.016812°E to 84.012913°E from 838m to 1424m asl and covers an area of 9.27 km2 (Fig. 1). Kamalpokhari, a part of the lake clusters of Pokhara Valley- Ramsar site is located at the southern base of the Kahundanda. The study area is drained by Kahun Khola (stream ~ Khola in Nepali), Kali Khola and Seti River. The motorable roads and human settlement fragment the southern part of the study area. The northern part of this area is covered with evergreen broad leaf forest that provide the shelter to mammals, birds, herpetofauna. The forest patches in the settlement area are important for many birds, herpetofauna and urban mammals. The study area has a humid subtropical climate; however, the elevation keeps temperatures moderate. Temperatures in summer average between 25 and 35 °C; in winter around −2 to 15 °C. Pokhara and nearby areas receive a high amount of precipitation (the annual rainfall 4851mm) (DHM 2019). Vegetation found in the southern and northern face area are Schima wallichi (Chilaune), Castanopsis indica (Katus), Alnus nepalensis (Uttis), Ficus bengalensis (Bar) Ficus religiosa (Peepal), Acacia catechu (Khair), Bombax ceiba (Simal), Emblica officinalis (Amala), Aegle marmelos (Bel), etc. and northern face have Rhododendrons and associated forest. Lantana camara (Banmara) is the dominant invasive species of this area. The fruiting plants such as Rubus idaeus (Ainselu), Choerospondias axillaris (Lapsi), Berberis sp. (Chutro), Myrica esculenta (Khafal), Ziziphus mauritiana (Bayar), Psidium guajava (Amba), Prunus cerasoides (Paiyun), Citrus jambhiri (Jyamir), Diospyros kaki (Tiju), Phyllanthus emblica (Amala), Morus sp. (Mulberry), Musa acuminata (Banana), Prunus persica (Peach), Limonia acidissima (Wood apple), Ficus religiosa (Peepal), Mangifera indica (Mango) scattered in the open area (settlements) and forest area attract the frugivorous and herbivore as well as insectivore birds. 2.2 | Methods The study of birds of in and around Kahundanda Hillscape was carried out for three months from August 2019 to October 2019. There were nine transects laid randomly that covered different habitat types, slopes, elevation and aspects for data collection. A total of 49 points were fixed along transects at the interval of 200–250m. The birds were reported within 25m radius and spent 20 minutes for each point (Five minutes for settlements and 15 minutes for observation) at 6.30–11.00 AM and 3.30–5.30 PM by using binoculars (Olympus, 20×50 magnification) and photographed the Figure 1. Map of the study area showing sampling locations Nepalese Journal of Zoology 6(2) Bastola et al. 4 birds by using Nikon P900 (80x) camera. The point count method helps to cover the range of bird species, maximize bird detection and increase identification accuracy (Ralph et al. 1995). Global positioning system (GPS) was used to record geographic location of the sampling points. We collected bird’s data along with number of individuals reported, activities of the birds, time, and weather conditions by developing standard data sheet. We also recorded the sound of the birds which were not visible but singing. We used the field guide book “Birds of Nepal” Grimmett et al. (2016) for bird identification. Photographs of unidentified species were identified with the help of bird experts. During bird survey, we also recorded different environmental factors such as distance to road, distance to settlements, distance to water sources, number of livestock present, number of fruiting trees including habitat and disturbance (Table 1). The scientific name, family, order, conservation and migratory status were identified with the help of IUCN Red List of Threatened Species and The Status of Nepal's Birds: The National Red List Series Birds of Nepal (https://www.iucnredlist.org/, https://www.hi-alyanature.org/page/red-data-birds). The recorded sound of the birds was identified by using the bird song database of Xeno-Canto (https://www.xeno-canto.org/). The feeding guilds of the birds was grouped into four groups as Carnivores, herbivores, omnivores and insectivores based on primary food they uptake as described by Grimmett et al. (2016). Environmental variables such as distance to road, distance to village or settlements, distance to nearest water sources were measured as Euclidian distance from sampling points to nearest road, settlement or water resources using ArcGIS 10.7. Other variables such as the number of livestock present, number of fruiting plants present were reported in the sampling points through the direct observation method. The collected data were analyzed by diversity index, Simpson index of diversity and evenness. We calculated the diversity indices of birds as their habitat types. Shannon Index (H) = - ∑𝑝𝑖 ln 𝑝𝑖𝑠𝑖=1. Where, pi is the proportion (n/N) of individuals of one particular species found (n) divided by the total number of individuals found (N), ln is the natural log, Σ is the sum of the calculations, and s is the number of species. Simpson index was used to show community diversity in relation to different habitat types (Simpson 1949). Simpson Index (1-D) = 1∑pi2si=1 Where, p is the proportion (n/N) of individuals of one particular species found (n) divided by the total number of individuals found (N), Σ is the sum of the calculations, and s is the number of species. Evenness (e) determined the distribution of individuals of a species in a community. Evenness = H'/Hmax Where H’ is the Shannon diversity index and Hmax is the maximum possible value. E is constrained between 0 and 1.0. As with H', evenness assumes that all species are represented within the sample. Species discovery curve was developed to test whether the sampling effort was enough for the detection of all the species that reported from the study area. This curve was plotted in between cumulative number of species reported and the sampling effort (Willott 2001). The cumulative number of species was placed in Y-axis and sampling unit placed in X-axis. We used species Rank Abundance Curve (RAC) or Whittaker plot to show the relative abundance of the birds (Izsák & Pavoine 2012; Avolio et al. 2019). For this curve, the species reported were ranked according to their abundance. The most abundant species is given rank 1, the second most abundant is 2 and so on. The abundant rank is placed Table 1. The variables and parameters recorded during field study SN Variables Descriptions 1 Species variables All the bird species recorded during field study 2 Habitat variables Types of habitats (forest area, bush, open (include grassland and open area or settlement area), wetlands) 3 Conservation status IUCN category of threatened status Threatened status as National Redlist Data Book (NRDB) 4 Migratory status Residential and migratory 5 Feeding guilds Carnivores, herbivores, omnivores, piscivores and insectivores 6 Environmental variables Distance to road (RV), distance to village or settlements (DV), distance to water resources (DW), number of livestock present (Nliv), number of fruiting trees present (NFT), elevation (Ele) Nepalese Journal of Zoology 6(2) Bastola et al. 5 in X-axis and the relative abundance (RA) is placed on Y-axis. The relative abundance was simply calculated, dividing number of individuals of species by the total a number of individuals reported. The Data were analyzed using CANOCO 4.56 to show the relation between birds and habitat types (Ter Braak & Šmilauer 2009). We used Canonical Correspondence Analysis (CCA) because the gradient length was greater than 3 in all cases with species abundance data. Univariate generalized linear model (GLM) with Poisson distribution (identity function) was used to find out the relation between the species richness and different environmental parameters using R 4.0.0 version (R Core Team 2020). 3 | Results 3.1 | Bird community structure in Kahundada A total of 589 individuals of birds under 101 species, belonging to 41 families and 19 orders were recorded in the study area. Order Passeriformes is the largest order (n=52 species) followed by Piciformes (n = 9 species) (Fig. 2). Among the reported species 41.58% (n = 41) were migrant and 58.42% (n = 59) were resident birds (Table S1). We reported one globally Critically Endangered (CR)- red-headed vulture (Sarcogyps calvus), two Endangered (EN)- steppe eagle (Aquila nipalensis) and saker falcon (Falco cherrug), and one Near Threatened (NT) species- river lapwing (Vanellus duvaucelii). Similarly, two species was categorized as nationally Endangered (E) species- red-headed vulture (Sarcogyps calvus) and saker Falcon (Falco cherrug), two Vulnerable (V)- steppe eagle (Aquila nipalensis), black stork (Ciconia nigra) and two nationally Near Threatened categories- river lapwing (Vanellus duvaucelii) and watercock (Gallicrex cinerea) (Figs. 3 and 4). The results showed that most of the birds were least concerned category. We also reported only one endemic bird spiny babbler (Acanthoptila nipalensis) from Kahundanda area. The Shannon index of diversity (H'), Simpson index (1-D) and evenness (e) was the highest in forest habitat (H' = 3.55, 1-D = 0.96, e = 0.83) followed by wetlands (H' = 2.81, 1-D= 0.92, e = 0.72), bush habitat (H' = 2.66, 1-D = 0.87, e = 0.69) and the least diversity and evenness was in open area (H' = 2.36, 1-D = 0.76, e = 0.65) (Table 2). But, species dominance index (D) was the highest in open area habitat followed by bush (D = 0.92), wetland (D = 0.79) and forest area (D = 0.031) (Table 2). The species discovery curve based on cumulative number of species present in the sampling points showed linear trends and a greater number of species were reported with increasing the number of sampling points (Fig. 5). A total of 101 species reported were ranked into 22 different ranks depending the number of individuals Figure 3. Order-wise species of birds in Kahundada area 0204060AnseriformesBucerotiformesCiconiiformesPodicipediformesPsittaciformesSuliformesCaprimulgiformesCoraciiformesGalliformesGruiformesCuculiformesFalconiformesPelecaniformesStrigiformesColumbiformesAccipitriformesCharadriiformesPiciformesPasseriformesNumber of speciesOrder Figure 3. Conservation status of birds (global and national status) as Redlist Data Book, here, CR = Critically Endangered, EN = Endangered, VU = Vulnerable, NT = Near Threatened (NT) Nepalese Journal of Zoology 6(2) Bastola et al. 6 reported of each species. The rank abundance curve (RAC) showed the steep gradient up to rank 6 and gentle slope after that (Fig. 6). The steep slope showed the low evenness as the high-ranking species have much higher abundances than the low-ranking species. A shallow gradient indicates high evenness as the abundances of the different species are similar. 3.2 | Influence of environmental factors to the birds Bird species richness in Kahundanda area was the highest in forest followed by open area, wetlands, bush and lowest in grassland habitat (Fig. 7). The Monte-Carlo permutation test of significance of all canonical axes (ordination diagram (biplot)) showed the omnivores birds were significantly reported more from the open areas (agriculture fields, residential areas and grassland) than wetlands, forest area and bush habitat (F = 1.546, P = 0.002, Trace = 0.909, with 499 permutations). Similarly, insectivore birds were significantly associated with open area and forest area Figure 4. Birds reported from Kahundanda and associated area, A. red-headed vulture (Sarcogyps calvus), B. steppe eagle (Aquila nipalensis), C. saker falcon (Falco cherrug) and D. black stork (Ciconia nigra) (Photo credit: Subash Bastola). Table 2. Diversity and dominance indices of birds in different habitats of Kahundanda Forest Wetland Open Bush Dominance index (D) 0.032 0.079 0.13 0.09 Simpson index (1-D) 0.97 0.92 0.76 0.87 Shannon index (H) 3.56 2.81 2.36 2.66 Evenness (e) 0.84 0.72 0.65 0.69 Figure 5. Species discovery curve of birds reported from Kahundanda area Nepalese Journal of Zoology 6(2) Bastola et al. 7 (F= 2.616, P=0.002, Trace= 2.206, with 499 permutations (Fig. 8). Likewise, CCA ordination diagram (biplot) of carnivores birds with different habitat types of Kahundanda significantly showed relation with forest, open area and wetland habitat (F = 2.616, P = 0.002, Trace=2.206, with 499 permutations) and herbivores were reported more in forest and open area mainly agriculture fields and grassland area (F = 1.478, P = 0.08, Trace=1.282, with 499 permutations) (Fig. 9). The results of univariate GLM shows that birds were significantly reported far from the roads (showed the negative association with roads, Z = 3.097, P = 0.001). Similarly, distance to village shows the significant association with distance to settlements or village. Likewise, number of livestock present in the study area was the major indicator of the habitat use of birds. They showed significantly negative relation with number of livestock in the study area (Z = -6.061, P < 0.000) (Table 3). Water resources played a vital role for the occurrence of wetland and wetland dependent birds (Z = -1.378, p=0.16). Number of fruiting trees present in the study area was also the best predictor for the occurrence of bird species. Most of the omnivores and herbivore birds depend upon the fruits and flowers of the fruiting plants and GLM showed the significantly positive association with number of fruiting trees (Z = -6.061, P < 0.000). Our study area ranges from 838 m to 1424 m altitude. The result shows the positive association of bird species richness with elevation of Kauhndanda (Table 3, Fig. 10). Figure 6. Rank abundance curve (RAC) of bird species recorded in Kahundanda, Pokhara 00.010.020.030.040.050102030Relative abundanceAbundance rank Figure 7. Bird species richness in different habitat types 01020304050ForestGrasslandOpen areaShrubland WetlandSpecies richnessHabitat type Figure 8. CCA ordination diagram (biplot) showing the bird species association with different habitats A) omnivore and B) insectivore birds of Kahundanda. Nepalese Journal of Zoology 6(2) Bastola et al. 8 4 | Discussion Environmental variables such as habitat types, number of fruiting trees present and other anthropogenic activities affect the distribution and species richness of the birds. Habitat types and number of fruiting trees provide the appropriate nesting and feeding sites to the birds. This study examined the status of birds in and around Kahundanda Hillscape, Pokhara Valley, Nepal. Our results indicated that the species richness of birds varied with the habitat types. The Kamalpokhari wetland and associated wetlands and streams supported a large number of wetland birds. The lake clusters of Pokhara Valley including Kamalpokhari are listed in Ramsar site. The Kahundanda supported a considerable bird species (n = 101). Overall, higher bird diversity was found in forests and open areas as these areas are very close to the city area and one of the important tourist areas. Agriculture lands and residential areas supported a large number of urban and open area dependent birds. The forest patches scattered in the Kahundanda also support diverse bird species. Similar types of studies in and around Phewa Lake of the Pokhara valley reported 148 species of birds belonging to 44 families and 11 orders (Khatri et al. 2019). Similarly, Dhakal et al. (2020) reported 101 species of birds in Khaste Lake Complex, Pokhara Valley. The species diversity of the birds was higher in the terrestrial ecosystem (H' = 3.27) including forest than wetland of Phewa Lake (H' = 3.07) (Khatri et al. 2019), a similar type of results was found in our study. Habitat heterogeneity of the Kahundanda supported a large number of bird species. Omnivore birds were reported more from the open and forest areas, likewise, insectivores and herbivores were reported from the agriculture landscape and in and around Kamalpokhari Lake complex. Wetlands of Kamalpokhari and associated areas supported 21 species of wetland birds that highlighted the importance of such small lakes for the conservation of wetland birds. Similar type of study Figure 9. CCA ordination diagram (biplot) showing the bird species association with different habitats A) carnivore and B) herbivore birds of Kahundanda Table 3. Univariate Generalized Linear Model (GLM) with Poisson distribution and identity link function test showing the effects of different environmental and disturbance factors on abundance of birds in Kauhndanda. Model Estimate Std. Error z value Pr(>|z|) AIC DR 0.003 0.001 3.097 0.001** 273.3 DW -0.002 0.001 -1.378 0.168 281.42 DV 0.003 0.001 2.604 0.009** 276.42 Ele 0.007 0.003 2.051 0.04* 279.19 NFT 1.144 0.187 6.118 <0.000*** 243.77 Nliv -0.681 0.112 -6.061 <0.000*** 249.91 Note: Significance codes: 0 '***', 0.001 '**', 0.01 '*'; DR= Distance to road, DW= Distance to water, DV= Distance to village or settlements, Ele= Elevation, NFT= Number of fruiting trees, Nliv=Number of livestock Nepalese Journal of Zoology 6(2) Bastola et al. 9 conducted by Khadka et al. (2017) from the wetlands of Chitwan National Park reported 46 species of wetland birds. Adhikari et al. (2018) reported 44 wetland birds from Beeshhazari Lake system, one of the Ramsar sites of Nepal. Khatri et al. (2019) reported 63 wetland dependent birds from Phewa Lake (one of the parts of Ramsar site: the lake cluster of Pokhara Valley) whereas Thapa and Saund (2012) reported 77 species of wetland birds in Jagadishpur Reservoir (one of the Ramsar sites from lowland Nepal). Our study reported six globally threatened birds and one Near-Threatened bird according to the IUCN Figure 10. Relationship between bird species richness and different environmental factors (A) elevation (R2 = 0.07), (B) number of fruiting plants (R2 = 0.63), (C) distance to water (R2 = 0.03), (D) distance to motorable roads (R2 = 0.16), (E) distance to village (R2 = 0.11) and (F) number of livestock present (R2 = 0.52) in Kahundanda Hillscape. Nepalese Journal of Zoology 6(2) Bastola et al. 10 category. Likewise, seven species of birds were listed in the nationally threatened category including three nationally threatened birds. Hence, this area played an important role in bird conservation. A study by Khatri et al. (2019) reported seven globally threatened birds from in and around the Phewa Lake. Dhakal et al. (2020) reported six globally threatened birds from Khaste Lake complex, Pokhara. Hence, conservation priorities should be given for such type of wetland habitat for the conservation of birds. Ingesting pesticides directly or indirectly by eating pesticide-contaminated grains, fruits, worms and insects may adversely affect the birds. Due to lack of public awareness about the importance of wetland resources, the area has been found misused by the local villagers. Lack of proper wetland management and overgrazing are the existing major problems. A species discovery curve is a quantitative analysis that evaluates the minimum number of sampling size necessary to define the number of species that represent a community (Willott 2001; Mendes et al. 2020). The species discovery curve for Kahundanda clearly showed the sufficient sampling efforts for the birds. Similar type of species discovery curve was also used in the study of birds by La Sorte & Somveille (2020) and Pandey et al. (2021). The relative species abundance is a major component of biodiversity and indicate how common and rare species are relative to other species. Relative species abundances indicate the specific patterns that provide the scenario and effects of the macro ecosystems on the species distribution (Izsák & Pavoine 2012; Avolio et al. 2019). A rank abundance curve or Whittaker plot is a graphical tool that help to visualize species richness and species evenness (Izsák & Pavoine 2012; Yin et al. 2018; Avolio et al. 2019; Pandey et al. 2021). In this study, we used RAC to show the relation between the relative abundance of the birds and their abundance rank in the Kahundanda area. Habitat heterogeneity of the Kahundanda supported a large number of bird species. Omnivore birds were reported more from the open and forest areas, while insectivore and herbivore birds were reported from the agriculture landscapes and in and around Kamal-pokhari Lake. Birds are confined to specific habitats that provide feeding and breeding grounds (Martin & Fahrig 2018; Adhikari et al. 2019). Abundance and richness of birds are limited by the food available to the habitat (Kim et al. 2018). The abundance of herbivore birds was higher in the open areas, agricultural fields and forests. Number of fruiting trees present in the forest or agriculture or human settlement areas added more food availability for the herbivore birds. Hence, specific fruiting seasons increased the abundance of herbivore birds (Pandey et al. 2021). The avian research conducted by Herzog et al. (2005) in the Swiss agricultural landscape reported the pastureland didn’t contribute to bird diversity but our study found that agriculture along with settlements and grassland supported higher omnivores and insectivore bird diversity. Insectivorous birds are the habitat specialists but they used agriculture fields and residential areas for foraging as the insect diversity is high in agriculture fields (Redhead et al. 2018; Schumm et al. 2020). Carnivore birds (birds of prey and scavengers) were significantly reported in open and wetland areas while piscivores birds were reported in wetland habitats. Similar types of results reported also reported from Kennya (Virani et al. 2011), India (Chettri et al. 2018; Mazumdar 2019; Kakati et al. 2021), China (Zhang et al. 2019), Bhutan (Mazumdar 2019), Indonesia (Krisanti et al. 2017). The distribution and abundance of birds depend upon different environmental and disturbance variables. Elevation of the landscape is one of the major factors that influence the distribution of the birds (Basnet et al. 2016; Katuwal et al. 2016; Neupane et al. 2020; Pandey et al. 2020; Ghimire et al. 2021). Distance to village or settlements, distance to roads, number of livestock present are the major disturbance determinant of the birds. The agriculture land and home garden also provide the nesting and feeding ground to the open area birds (Pandey et al. 2021). The fruiting trees also increases the richness of the birds (Pandey et al. 2021). Our results also showed a significant relation with distance to roads and settlements and a negative association with the number of livestock present. Similar type of results were also found on the study of Adhikari et al. (2019) in Chitwan National Park, Nepal; Basnet et al. (2016) in the hilly region of Central Himalaya, Nepal; Andradas et al. (2019) in Northern Iberian Peninsula; Altaf et al. (2018) in Panjab, Pakisthan; Panigrahi and Jins (2018) in Kerela, India. Hence, presence of different environmental factors including disturbance are the important driver for the occurrence of birds in and around Kahundanda Hillscape, Pokhara, Nepal. 5 | Conclusions Different types of habitats of the Kahundanda Hillscape supported a total of 101 terrestrial and wetland dependent birds. Most of the birds were reported from forest areas and wetlands. The habitats of human Nepalese Journal of Zoology 6(2) Bastola et al. 11 dominated mountainous landscape play an important role to the conservation of birds. Mosaics of habitat patches with agricultural landscape favored the high species richness within a small area. Presence of fruiting trees near the open area and settlement supported higher bird diversity. Habitat types, human disturbance and proximity to water resources are important for bird species conservation and ecotourism in human dominated landscape like Kahundanda. Acknowledgements Authors would like to thank everyone who helped us during this research work. We are thankful to Divison Forest Office, Kaski for the research permission. Authors, Bishnu Prasad Bhattarai and Jagan Nath Adhikari are thankful to IdeaWild for supporting field equipment. We are grateful to Salikram Adhikari, Dilip Subedi, Iswori Prasad Tiwari and Ram Chandra Lamichhane for their kind help and support during field study. Special thanks go to Mr. Iswori Prasad Tiwari for proof reading the draft of this article. Authors’ contributions SCB designed research and collected data; SCB and HD identified the birds, JNA and BPB analyzed the data, JNA and SCB wrote draft of the article, SCB, JNA, HD and BPB revised the manuscript. All authors approved the final manuscript for publication. Conflicts of interest Authors declare no conflict of interest.References Adhikari, J. N., Bhattarai, B. P. and Dhakal, D. N. 2018. Conservation value of Beeshhazari Lake: an insight into diversity and abundance of wetland birds. Our Nature 16:17–26. https://doi.org/10.3126/on.v16i1.21563 Adhikari, J. N., Bhattarai, B. P. and Thapa, T. B. 2019. Factors affecting diversity and distribution of threatened birds in Chitwan National Park, Nepal. Journal of Threatened Taxa 11:13511–13522. https://doi.org/10.11609/jott.4137.11.5.13511-13522 Altaf, M., Javid, A., Khan, A. M., Khan, M., Umair, M. and Ali, Z. 2018. Anthropogenic impact on the distribution of the birds in the tropical thorn forest, Punjab, Pakistan. Journal of Asia-Pacific Biodiversity 11:229–236. https://doi.org/10.1016/j.japb.2018.03.001 Andradas, M. X., Arizaga, J. and Rodríguez-Pérez, J. 2019. Species co-occurrence and environmental factors and their effect on the distribution of forest birds in mature forests. Forestry: An International Journal of Forest Research 92:568–576. https://doi.org/10.1093/forestry/cpz032 Avolio, M. L., Carroll, I. T., Collins, S. L., Houseman, G. R., Hallett, L. M., Isbell, F., et al. 2019. A comprehensive approach to analyzing community dynamics using rank abundance curves. Ecosphere 10:e02881. https://doi.org/10.1002/ecs2.2881 Baral, H. S. and Inskipp, C. 2020. Birds of Nepal: Their Status and Conservation Especially with Regards to Watershed Perspectives. Pages 435-458 Hindu Kush-Himalaya Watersheds Downhill: Landscape Ecology and Conservation Perspectives. Springer. https://doi.org/10.1007/978-3-030-36275-1_22 Basnet, T. B., Rokaya, M. B., Bhattarai, B. P. and Münzbergová, Z. 2016. Heterogeneous landscapes on steep slopes at low altitudes as hotspots of bird diversity in a Hilly Region of Nepal in the Central Himalayas. PLoS One 11:e0150498. https://doi.org/10.1371/journal.pone.0150498 BCN. 2022. Status of Birds of Nepal. Bird Conservation Nepal, Kathmandu. Bregman, T. P., Sekercioglu, C. H. and Tobias, J. A. 2014. Global patterns and predictors of bird species responses to forest fragmentation: implications for ecosystem function and conservation. Biological Conservation 169:372–383. https://doi.org/10.1016/j.biocon.2013.11.024 Chettri, A., Sharma, K., Dewan, S. and Acharya, B. K. 2018. Bird diversity of tea plantations in Darjeeling hills, Eastern Himalaya, India. Biodiversitas Journal of Biological Diversity 19:1066–1073. https://doi.org/10.13057/biodiv/d190339 Dhakal, H., Ghimire, M., Poudel, A. K., Ghimire, P. and Bhusal, K. P. 2020. Avian diversity of Khaste Lake Complex, Pokhara Valley, Nepal. Minivet 3:17–25. DHM. 2019. Data of temperature, rainfall and relative humidity from 1989 to 2018.in Department of Hydrology and Meteorology, K., editor. Department of Hydrology and Meteorology, Kathmandu. Nepalese Journal of Zoology 6(2) Bastola et al. 12 DNPWC and BCN. 2019. An official checklist of the birds. Department of National Parks and Wildlife Conservation, Bird Conservation Nepal. Ferreira, C. D. and Perbiche-Neves, G. 2021. The importance of the standardizing sampling methodology to detect altitudinal gradients in mountains: A study case for the resident bird community in a hotspot (Atlantic forest) and the Middle Domain Effect. Acta Oecologica 110:103677. https://doi.org/10.1016/j.actao.2020.103677 Ghimire, A., Rokaya, M. B., Timsina, B., Bílá, K., Shrestha, U. B., Chalise, M. K., et al. 2021. Diversity of birds recorded at different altitudes in central Nepal Himalayas. Ecological Indicators 127:107730. https://doi.org/10.1016/j.ecolind.2021.107730 Grimmett, R., Inskipp, C., Inskipp, T. and Baral, H. S. 2016. Birds of Nepal: Revised Edition. Bloomsbury Publishing. Herzog, F., Dreier, S., Hofer, G., Marfurt, C., Schüpbach, B., Spiess, M., et al. 2005. Effect of ecological compensation areas on floristic and breeding bird diversity in Swiss agricultural landscapes. Agriculture, Ecosystems and Environment 108:189–204. https://doi.org/10.1016/j.agee.2005.02.003 Inskipp, C., Baral, H. S., Acharya, S., Chaudhary, H., Ghimire, M. and Giri, D. 2020. Rare birds in Nepal. Nepalese Journal of Zoology 4:108–132. https://doi.org/10.3126/njz.v4i2.33894 Inskipp, C., Baral, H. S., Inskipp, T., Khatiwada, A. P., Khatiwada, M. P., Poudyal, L. P., et al. 2017. Nepal’s National red list of birds. Journal of Threatened Taxa 9:9700–9722. https://doi.org/10.11609/jott.2855.9.1.9700-9722 Inskipp, C., Baral, H. S., Phuyal, S., Bhatt, T. R., Khatiwada, M., Inskipp, T., et al. 2016. The Status of Nepal's Birds: The National Red List Series. Zoological Society of London, UK. Iswandaru, D., Novriyanti, N., Banuwa, I. S. and Harianto, S. P. 2020. The Distribution of bird communities in University of Lampung, Indonesia. Biodiversitas Journal of Biological Diversity 21:2629–2637. https://doi.org/10.13057/biodiv/d210634 Izsák, J. and Pavoine, S. 2012. Links between the species abundance distribution and the shape of the corresponding rank abundance curve. Ecological Indicators 14:1–6. https://doi.org/10.1016/j.ecolind.2011.06.030 Junk, W. J., An, S., Finlayson, C., Gopal, B., Květ, J., Mitchell, S. A., et al. 2013. Current state of knowledge regarding the world’s wetlands and their future under global climate change: a synthesis. Aquatic sciences 75:151–167. https://doi.org/10.1007/s00027-012-0278-z Kakati, R., Bhuyan, A. and Borah, D. 2021. Status of biodiversity in wetlands of Biswanath District of Assam, India. Biodiversitas Journal of Biological Diversity 22(1): 453–471. https://doi.org/10.13057/biodiv/d220156 Katuwal, H. B., Basnet, K., Khanal, B., Devkota, S., Rai, S. K., Gajurel, J. P., et al. 2016. Seasonal changes in bird species and feeding guilds along elevational gradients of the Central Himalayas, Nepal. PLoS One 11:e0158362. https://doi.org/10.1371/journal.pone.0158362 Khadka, B. B., Acharya, P. M. and Rajbhandari, S. L. 2017. Population status and species diversity of wetland birds in the Rapti and Narayani rivers and associated wetlands of Chitwan National Park, Nepal. Journal of Threatened Taxa 9:10297–10306. https://doi.org/10.11609/jott.2364.9.6.10297-10306 Khatri, N. D., Neupane, B., Timilsina, Y. P. and Ghimire, S. 2019. Assessment of avifaunal diversity and threats to them in Phewa Wetland, Nepal. Forestry: Journal of Institute of Forestry, Nepal 16:31–47. https://doi.org/10.3126/forestry.v16i0.28352 Kim, J.-Y., Lee, S., Shin, M.-S., Lee, C.-H., Seo, C. and Eo, S. H. 2018. Altitudinal patterns in breeding bird species richness and density in relation to climate, habitat heterogeneity, and migration influence in a temperate montane forest (South Korea). PeerJ 6:e4857. https://doi.org/10.7717/peerj.4857 Krisanti, A., Choirunnafi, A., Septiana, N., Pratama, F., Amelia, F., Manjaswari, A., et al. 2017. The diversity of diurnal bird species on western slope of Mount Lawu, Java, Indonesia. Biodiversitas Journal of Biological Diversity 18:1077–1083. La Sorte, F. A. and Somveille, M. 2020. Survey completeness of a global citizen‐science database of bird occurrence. Ecography 43:34–43. https://doi.org/10.13057/biodiv/d180327 Martin, A. E. and Fahrig, L. 2018. Habitat specialist birds disperse farther and are more migratory than habitat generalist birds. Ecology 99:2058–2066. https://doi.org/10.1002/ecy.2428 Mazumdar, S. 2019. Composition of avian communities in a human-modified wetland Okhla bird sanctuary, India: with notes on conservation initiatives. Proceedings of the Zoological Society. Springer. P. 319–333. Meer, I. and Mirza, Z. B. 2017. Abundance of birds and variations in physico-chemical characteristics of freshwater reservoir. University of Sindh Journal of Animal Sciences 1:14–20. Mendes, A. B., Silva, E. P. and Duarte, M. R. 2020. Can sambaquis (shell mounds) be used as records of the Holocene marine fish biodiversity? Biodiversity and Conservation 29:39–56. https://doi.org/10.1007/s10531-019-01868-8 Nepalese Journal of Zoology 6(2) Bastola et al. 13 Neupane, J., Khanal, L., Gyawali, B. and Chalise, M. K. 2020. Elevational pattern and seasonality of avian diversity in Kaligandaki River Basin, central Himalaya. Journal of Threatened Taxa 12(14):16927–16943. https://doi.org/10.11609/jott.5815.12.14.16927-16943 Pan, X., Ding, Z., Hu, Y., Liang, J., Wu, Y., Si, X., et al. 2016. Elevational pattern of bird species richness and its causes along a central Himalaya gradient, China. PeerJ 4:e2636. https://doi.org/10.7717/peerj.2636 Pan, X., Liang, D., Zeng, W., Hu, Y., Liang, J., Wang, X., et al. 2019. Climate, human disturbance and geometric constraints drive the elevational richness pattern of birds in a biodiversity hotspot in southwest China. Global Ecology and Conservation 18:e00630. https://doi.org/10.1016/j.gecco.2019.e00630 Pandey, N., Khanal, L. and Chalise, M. K. 2020. Correlates of avifaunal diversity along the elevational gradient of Mardi Himal in Annapurna Conservation Area, Central Nepal. Avian Research 11(1). https://doi.org/10.1186/s40657-020-00217-6 Pandey, N., Khanal, L., Chapagain, N., Singh, K. D., Bhattarai, B. P. and Chalise, M. K. 2021. Bird community structure as a function of habitat heterogeneity: A case of Mardi Himal, Central Nepal. Biodiversitas Journal of Biological Diversity 22:262–271. https://doi.org/10.13057/biodiv/d220132 Panigrahi, M. and Jins, V. 2018. Status of birds in Agasthyamalai Hills, Western Ghats, Kerala, India. Journal of Threatened Taxa 10:11166–11184. https://doi.org/10.11609/jott.10.1.11166-11184 R Development Core Team.. 2020. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. R Foundation for Statistical Computing, Vienna, Austria. Ralph, C. J., Sauer, J. R., Droege, S., technical editors. 1995. Monitoring bird populations by point counts. Gen. Tech. Rep. PSW-GTR-149. Albany, CA: U.S. Department of Agriculture, Forest Service, Pacific Southwest Research Station: p. 161–168 Redhead, J., Hinsley, S., Beckmann, B., Broughton, R. and Pywell, R. 2018. Effects of agri-environmental habitat provision on winter and breeding season abundance of farmland birds. Agriculture, Ecosystems and Environment 251:114–123. https://doi.org/10.1016/j.agee.2017.09.027 Redlich, S., Martin, E. A., Wende, B. and Steffan-Dewenter, I. 2018. Landscape heterogeneity rather than crop diversity mediates bird diversity in agricultural landscapes. PLoS One 13:e0200438. https://doi.org/10.1371/journal.pone.0200438 Schumm, M., White, A. E., Supriya, K. and Price, T. D. 2020. Ecological limits as the driver of bird species richness patterns along the east Himalayan elevational gradient. The American Naturalist 195:802–817. https://doi.org/10.5061/dryad.hmgqnk9bx Shannon, C. E. 1948. Mathematical theory of communication. The Bell System Technical Journal 27:379–424. https://doi.org/10.1002/j.1538-7305.1948.tb01338.x Simpson, E. H. 1949. Measurement of diversity. Nature 163:688–688. https://doi.org/10.1038/163688a0 Ter Braak, C. J. F. and Šmilauer, P. 2009. Canoco for Windows. Biometris-quantitative methods in the life and earth sciences. Plant Research International, Wageningen University Research Centre, the Netherlands. Thapa, J. B. and Saund, T. B. 2012. Water quality parameters and bird diversity in Jagdishpur Reservoir, Nepal. Nepal Journal of Science and Technology 13:143–155. https://doi.org/10.3126/njst.v13i1.7453 Virani, M. Z., Kendall, C., Njoroge, P. and Thomsett, S. 2011. Major declines in the abundance of vultures and other scavenging raptors in and around the Masai Mara ecosystem, Kenya. Biological Conservation 144:746–752. https://doi.org/10.1016/j.biocon.2010.10.024 Willott, S. 2001. Species accumulation curves and the measure of sampling effort. Journal of Applied Ecology 38:484–486. https://doi.org/10.1046/j.1365-2664.2001.00589.x Yin, Z.-Y., Zeng, L., Luo, S.-M., Chen, P., He, X., Guo, W., et al. 2018. Examining the patterns and dynamics of species abundance distributions in succession of forest communities by model selection. PLoS One 13:e0196898. https://doi.org/10.1371/journal.pone.0196898 Zhang, H., Wu, F., Zhang, Y., Han, S. and Liu, Y. 2019. Spatial and temporal changes of habitat quality in Jiangsu Yancheng Wetland National Nature Reserve-Rare birds of China. Applied Ecology Environtal Research 17:4807–4821. http://dx.doi.org/10.15666/aeer/1702_48074821 Supplementary Table S1. Checklist of the birds reported from Kahundada SN Common name Scientific name Codes used in Canacoo Order Family IUCN Redlist National Redlist Feeding guild 1 Himalayan buzzard Buteo refectus (Portenko, 1929) But.ref Accipitriformes Accipitridae LC LC Carnivore 2 Black kite Milvus migrans (Boddaert, 1783) Mil.mig Accipitriformes Accipitridae LC LC Carnivore 3 Red-headed vulture Sarcogyps calvus (Scopoli, 1786) Sar.cal Accipitriformes Accipitridae CR EN Carnivore 4 Osprey Pandion haliaetus (Linnaeus, 1758) Pan.hal Accipitriformes Pandionidae LC LC Pisivore 5 Little grebe Tachybaptus ruficollis (Pallas, 1764) Tac.ruf Podicipediformes Podicipedidae LC LC Pisivore 6 Lesser whistling-duck Dendrocygna javanica (Horsfield, 1821) Den.jav Anseriformes Anatidae LC LC Omnivore 7 Common sandpiper Actitis hypoleucos Linnaeus, 1758 Act.hyp Charadriiformes Scolopacidae LC LC Omnivore 8 Common hoopoe Upupa epops (Linnaeus, 1758) Upu.epo Bucerotiformes Upupidae LC LC Granivore 9 House swift Apus nipalensis (Hodgson, 1837) Apu.nip Caprimulgiformes Apodidae LC LC Insectivore 10 Alpine swift Tachymarptis melba (Linnaeus, 1758) Tac.mel Caprimulgiformes Apodidae LC LC Insectivore 11 River lapwing Vanellus duvaucelii (Lesson, 1826) Van.duv Charadriiformes Charadriidae NT NT Insectivore 12 Red-wattled lapwing Vanellus indicus (Boddaert, 1783) Van.ind Charadriiformes Charadriidae LC LC Omnivore 13 Bronze-winged jacana Metopidius indicus (Latham, 1790) Met.ind Charadriiformes Jacanidae LC LC Omnivore 14 Little-ringed plover Charadrius dubius (Scopoli, 1786) Cha.dub Charadriiformes Charadriidae LC LC Insectivore 15 Black stork Ciconia nigra (Linnaeus, 1758) Cic.nig Ciconiiformes Ciconiidae LC VU Carnivore 16 Rock dove Columba livia (Gmelin 1789) Col.liv Columbiformes Columbidae LC LC Omnivore 17 Spotted dove Spilopelia suratensis (Gmelin, 1789) Spi.sur Columbiformes Columbidae LC LC Granivore 18 Oriental turtle-dove Streptopelia orientalis (Latham, 1790) Str.ori Columbiformes Columbidae LC LC Herbivore 19 Yellow-footed green pigeon Treron phoenicopterus (Latham, 1790) Tre.pho Columbiformes Columbidae LC LC Omnivore 20 Common kingfisher Alcedo atthis (Linnaeus, 1758) Alc.att Coraciiformes Alcedinidae LC LC Carnivore 21 White-throated kingfisher Halcyon gularis (Kuhl, 1820) Hal.gul Coraciiformes Alcedinidae LC LC Omnivore 22 Steppe eagle Aquila nipalensis (Hodgson, 1833) Aqu.nip Falconiformes Accipitridae EN VU Carnivore 23 Greater coucal Centropus sinensis (Stephens, 1815) Cen.sin Cuculiformes Cuculidae LC LC Carnivore 24 Western koel Eudynamys scolopaceus (Linnaeus, 1758) Eud.sco Cuculiformes Cuculidae LC LC Omnivore 25 Common hawk-cuckoo Hierococcyx varius (Vahl, 1797) Hie.var Cuculiformes Cuculidae LC LC Insectivore 26 Shikra Accipiter badius (Gmelin, 1788) Acc.bad Accipitriformes Accipitridae LC LC Carnivore 27 Saker falcon Falco cherrug (Gray, 1834) Fal.che Falconiformes Falconidae EN EN Carnivore 28 Common kestrel Falco tinnunculus (Linnaeus, 1758) Fal.tin Falconiformes Falconidae LC LC Carnivore 29 Red junglefoul Gallus gallus (Linnaeus, 1758) Gal.gal Galliformes Phasianidae LC LC Omnivore 30 Kalij pheasant Lophura leucomelanos (Latham, 1790) Lop.leu Galliformes Phasianidae LC LC Omnivore 31 White-breasted waterhen Amaurornis phoenicurus (Pennant, 1769) Ama.pho Gruiformes Rallidae LC LC Omnivore 32 Brown shrike Lanius cristatu (Linnaeus, 1758) Lan.cri Passeriformes Laniidae LC LC Insectivore 33 Watercock Gallicrex cinerea (Gmelin, 1789) Gal.cin Gruiformes Rallidae LC NT Omnivore 34 Cattle egret Bubulcus ibis (Linnaeus, 1758) Bub.ibi Passeriformes Ardeidae LC LC Insectivore 35 Large cuckooshrike Coracina javensis (Horsfield, 1821) Cor.jav Passeriformes Campephagidae LC LC Omnivore 36 Scarlet minivet Pericrocotus flammeus (Forster, 1781) Per.fla Passeriformes Campephagidae LC LC Insectivore Nepalese Journal of Zoology 6(2) Bastola et al. 15 37 Common tailorbird Orthotomus sutorius (Pennant, 1769) Ort.sut Passeriformes Cisticolidae LC LC Omnivore 38 Striated prinia Prinia crinigera (Hodgson, 1836) Pri.cri Passeriformes Cisticolidae LC LC Insectivore 39 Common green magpie Cissa chinensis (Boddaert, 1783) Cis.chi Passeriformes Corvidae LC LC Omnivore 40 House crow Corvus splendens (Vieillot, 1817) Cor.spl Passeriformes Corvidae LC LC Omnivore 41 Grey treepie Dendrocitta formosae (Swinhoe, 1867) Den.for Passeriformes Corvidae LC LC Omnivore 42 Rufous treepie Dendrocitta vagabunda (Latham, 1790) Den.vag Passeriformes Corvidae LC LC Omnivore 43 Red-billed blue magpie Urocissa erythrorhyncha (Boddaert, 1783) Uro.ery Passeriformes Corvidae LC LC Omnivore 44 Spangled drongo Dicrurus bracteatus (Gould, 1842) Dic.bra Passeriformes Dicruridae LC LC Omnivore 45 Black drongo Dicrurus macrocercus (Vieillot, 1817) Dic.mac Passeriformes Dicruridae LC LC Carnivore 46 Scaly-breasted munia Lonchura punctulata (Linnaeus, 1758) Lon.pun Passeriformes Estrildidae LC LC Herbivore 47 White-rumped munia Lonchura striata (Linnaeus 1766) Lon.str Passeriformes Estrildidae LC LC Omnivore 48 Red-rumped swallow Cecropis daurica Linnaeus, 1771) Cec.dau Passeriformes Hirundinidae LC LC Insectivore 49 Long-tailed shrike Lanius schach (Linnaeus, 1758) Lan.sch Passeriformes Laniidae LC LC Omnivore 50 Grey-backed shrike Lanius tephronotus (Vigors, 1837) Lan.tep Passeriformes Laniidae LC LC Carnivore 51 White-crested laughingthrush Garrulax leucolophus (Hardwicke, 1815) Gar.leu Passeriformes Leiotrichidae LC LC Omnivore 52 Spiny babbler Turdoides nipalensis (Hodgson, 1836) Tur.nip Passeriformes Leiotrichidae LC LC Insectivore 53 Jungle babbler Turdoides striata (Dumont, 1823) Tur.str Passeriformes Leiotrichidae LC LC Omnivore 54 Olive-backed pipit Anthus hodgsoni (Richmond, 1907) Ant.hod Passeriformes Motacillidae LC LC Omnivore 55 Richard's pipit Anthus richardi (Vieillot, 1818) Ant.ric Passeriformes Motacillidae LC LC Insectivore 56 Paddyfield pipit Anthus rufulus (Vieillot, 1818) Ant.ruf Passeriformes Motacillidae LC LC Omnivore 57 Grey wagtail Motacilla cinerea (Tunstall, 1771) Mot.cin Passeriformes Motacillidae LC LC Carnivore 58 White-browed wagtail Motacilla maderaspatensis (Gmelin, 1789) Mot.mad Passeriformes Motacillidae LC LC Omnivore 59 Oriental magpie robin Copsychus saularis (Linnaeus, 1758) Cop.sau Passeriformes Muscicapidae LC LC Omnivore 60 Little forktail Enicurus scouleri (Vigors, 1832) Eni.sco Passeriformes Muscicapidae LC LC Insectivore 61 Verditer flycatcher Eumyias thalassinus (Swainson, 1838) Eum.tha Passeriformes Muscicapidae LC LC Insectivore 62 Blue whistling thrush Myophonus caeruleus (Scopoli, 1786) Myo.cae Passeriformes Muscicapidae LC LC Omnivore 63 Pied bushchat Saxicola caprata (Linnaeus, 1766) Sax.cap Passeriformes Muscicapidae LC LC Omnivore 64 Grey bushchat Saxicola ferreus (Gray 1846) Sax.fer Passeriformes Muscicapidae LC LC Omnivore 65 Taiga flycatcher Ficedula albicilla (Pallas, 1811) Fic.alb Passeriformes Muscicapidae LC LC Insectivore 66 Common stonechat Saxicola torquatus (Linnaeus, 1766) Sax.tor Passeriformes Muscicapidae LC LC Insectivore 67 Indian golden oriole Oriolus kundoo (Sykes, 1832) Aet.sip Passeriformes Oriolidae LC LC Omnivore 68 Crimson sunbird Aethopyga siparaja (Raffles, 1822) Aet.sip Passeriformes Nectariniidae LC LC Omnivore 69 Purple sunbird Cinnyris asiatica (Latham, 1790) Cin.asi Passeriformes Nectariniidae LC LC Omnivore 70 Maroon oriole Oriolus traillii (Vigors, 1832) Ori.tra Passeriformes Oriolidae LC LC Omnivore 71 Great tit Parus major (Linnaeus, 1757) Par.maj Passeriformes Paridae LC LC Omnivore 72 Green-backed tit Parus monticolus (Vigors, 1831) Par.mon Passeriformes Paridae LC LC Omnivore 73 Black-lored tit Parus xanthogenys (Vigors, 1831) Par.xan Passeriformes Paridae LC LC Omnivore 74 House sparrow Passer domesticus (Linnaeus, 1758) Pas.dom Passeriformes Passeridae LC LC Omnivore 75 Eurasian tree sparrow Passer montanus (Linnaeus, 1758) Pas.mon Passeriformes Passeridae LC LC Omnivore Nepalese Journal of Zoology 6(2) Bastola et al. 16 76 Grey-hooded warbler Phylloscopus xanthoschistos (Gray, 1846) Phy.xan Passeriformes Phylloscopidae LC LC Insectivore 77 Black bulbul Hypsipetes leucocephalus (Gmelin, 1789) Hyp.leu Passeriformes Pycnonotidaetaiga LC LC Omnivore 78 Red-vented bulbul Pycnonotus cafer (Linnaeus, 1766) Pyc.caf Passeriformes Pycnonotidae LC LC Omnivore 79 Himalayan bulbul Pycnonotus leucogenys (Gray, 1837) Pyc.leu Passeriformes Pycnonotidae LC LC Omnivore 80 Black-crested bulbul Rubigula flaviventris (Tickell, 1833) Rub.fla Passeriformes Pycnonotidae LC LC Omnivore 81 Jungle myna Acridotheres fuscus (Wagler, 1827) Acr.fus Passeriformes Sturnidae LC LC Omnivore 82 Bank myna Acridotheres ginginianus (Latham, 1790) Acr.gin Passeriformes Sturnidae LC LC Omnivore 83 Common myna Acridotheres tristis (Linnaeus,1766) Acr.tri Passeriformes Sturnidae LC LC Omnivore 84 Indian white-eye Zosterops palpebrosus (Temminck, 1824) Zos.pal Passeriformes Zosteropidae LC LC Omnivore 85 Great cormorant Phalacrocorax carbo (Linnaeus, 1758) Pha.car Suliformes Phalacrocoracidae LC LC Omnivore 86 Intermediate egret Ardea intermedia Wagler, 1829 Ard.int Pelecaniformes Ardeidae LC LC Omnivore 87 Indian pond-heron Ardeola grayi (Sykes, 1832) Ard.gra Pelecaniformes Ardeidae LC LC Pisivore 88 Little egret Egretta garzetta (Linnaeus, 1767) Egr.gar Pelecaniformes Ardeidae LC LC Carnivore 89 Blue-throated barbet Megalaima asiatica (Latham, 1790) Meg.asi Piciformes Megalaimidae LC LC Omnivore 90 Golden-throated barbet Psilopogon franklinii (Blyth, 1842) Psi.fra Piciformes Megalaimidae LC LC Omnivore 91 Coppersmith barbet Psilopogon haemacephalus (Müller, 1776) Psi.hae Piciformes Megalaimidae LC LC Omnivore 92 Great barbet Psilopogon virens (Boddaert, 1783) Psi.vir Piciformes Megalaimidae LC LC Omnivore 93 Greater yellownape Chrysophlegma flavinucha (Gould, 1834) Chr.fla Piciformes Picidae LC LC Hervibore 94 Fulvous-breasted woodpecker Dendrocopos macei (Veillot, 1818) Den.mac Piciformes Picidae LC LC Omnivore 95 Grey-headed woodpecker Dendropicos spodocephalus (Bonaparte, 1850) Den.spo Piciformes Picidae LC LC Omnivore 96 Lesser yellownape Picus chlorolophus (Vieillot, 1818) Pic.chl Piciformes Picidae LC LC Insectivore 97 Scaly-bellied woodpecker Picus squamatus (Vigors, 1831) Pic.squ Piciformes Picidae LC LC Insectivore 98 Rose-ringed parakeet Alexandrinus krameri (Scopoli, 1769) Ale.kra Psittaciformes Psittacidae LC LC Herbivore 99 Spotted owlet Athene brama (Temminck, 1821) Ath.bra Strigiformes Strigidae LC LC Carnivore 100 Asian barred owlet Glaucidium cuculoides (Vigors, 1831) Gla.cuc Strigiformes Strigidae LC LC Carnivore 101 Jungle owlet Glaucidium radiatum (Tickell, 1833) Gla.rad Strigiformes Strigidae LC LC Carnivore | RESEARCH GATE |
Earth Syst. Sci. Data, 12, 2635–2645, 2020https://doi.org/10.5194/essd-12-2635-2020© Author(s) 2020. This work is distributed underthe Creative Commons Attribution 4.0 License.A dataset of microclimate and radiation and energyfluxes from the Lake Taihu eddy flux networkZhen Zhang1,2, Mi Zhang1,2,4, Chang Cao1,2, Wei Wang1,2, Wei Xiao1,2,4, Chengyu Xie1,2, Haoran Chu1,2,Jiao Wang1,2, Jiayu Zhao1,2, Lei Jia1,2, Qiang Liu1,2, Wenjing Huang1,2, Wenqing Zhang1,2, Yang Lu1,2,Yanhong Xie1,2, Yi Wang1,2, Yini Pu1,2, Yongbo Hu1,2, Zheng Chen1,2, Zhihao Qin1,2, and Xuhui Lee31Yale-NUIST Center on Atmospheric Environment, International Joint Laboratory on Climate andEnvironment Change (ILCEC), Nanjing University of Information Science and Technology,Nanjing, Jiangsu Province, China2Key Laboratory of Meteorological Disaster, Ministry of Education and Collaborative Innovation Center onForecast and Evaluation of Meteorological Disasters, Nanjing University of Information Science andTechnology, Nanjing, Jiangsu Province, China3School of the Environment, Yale University, New Haven, CT 06511, USA4NUIST-Wuxi Research Institute, Wuxi, Jiangsu Province, ChinaCorrespondence: Xuhui Lee (xuhui.lee@yale.edu)Received: 12 March 2020 – Discussion started: 17 June 2020Revised: 30 August 2020 – Accepted: 17 September 2020 – Published: 31 October 2020Abstract. Eddy covariance data are widely used for the investigation of surface–air interactions. Althoughnumerous datasets exist in public depositories for land ecosystems, few research groups have released eddy co-variance data collected over lakes. In this paper, we describe a dataset from the Lake Taihu eddy flux network, anetwork consisting of seven lake sites and one land site. Lake Taihu is the third-largest freshwater lake (area of2400 km2) in China, under the influence of subtropical climate. The dataset spans the period from June 2010 toDecember 2018. Data variables are saved as half-hourly averages and include micrometeorology (air tempera-ture, humidity, wind speed, wind direction, rainfall, and water or soil temperature profile), the four componentsof surface radiation balance, friction velocity, and sensible and latent heat fluxes. Except for rainfall and wind di-rection, all other variables are gap-filled, with each data point marked by a quality flag. Several areas of researchcan potentially benefit from the publication of this dataset, including evaluation of mesoscale weather forecastmodels, development of lake–air flux parameterizations, investigation of climatic controls on lake evaporation,validation of remote-sensing surface data products and global synthesis on lake–air interactions. The datasetis publicly available at https://yncenter.sites.yale.edu/data-access (last access: 24 October 2020) and from theHarvard Dataverse (https://doi.org/10.7910/DVN/HEWCWM; Zhang et al., 2020).1IntroductionInland lakes and reservoirs are a vital freshwater resourcefor the society. Globally, there are more than 27 million wa-ter bodies with size greater than 0.01 km2, occupying a to-tal of 3.5 % of the earth’s land surface area (Downing et al.,2006; Verpoorter et al., 2014). Accurate observation of thelake microclimate and lake–air interactions will help to bet-ter manage this water resource and to better predict how itmay be affected by environmental changes. Towards that end,an increasing number of studies have employed the eddy co-variance (EC) methodology to monitor physical state (tem-perature, wind, humidity) and process variables (momentumflux and radiation and energy fluxes) in the lake environment(Vesala et al., 2006; Blanken et al., 2011; Nordbo et al., 2011;Wang et al., 2014; Li et al., 2015; Yusup and Liu, 2016; Duet al., 2018; Hamdani et al., 2018; Xiao et al., 2018; Wanget al., 2019). Unlike EC studies in land ecosystems, however,data from these lake studies are rarely published as data pa-Published by Copernicus Publications.2636Z. Zhang et al.: Dataset from the Lake Taihu eddy flux networkpers or are rarely archived in public data depositories acces-sible by the broader scientific community. For example, ofthe nearly 500 sites that have contributed EC and microm-eteorological data to AmeriFlux, a public data depository(https://ameriflux.lbl.gov/data/data-availability/, last access:24 October 2020), none are a lake site. Although a few sci-entific groups have provided data supplements to their scien-tific papers on lake–air fluxes (e.g., Charusombat et al., 2018;Franz et al., 2018; Zhao and Liu, 2018), we are not aware ofa data paper devoted to systematic description and archivalof EC lake observations.In this paper, we describe the dataset from the Lake Taihueddy flux network (Lee et al., 2014). Established in 2010,the network currently consists of six active lake sites, one in-active lake site and one active land site. Lake Taihu is thethird-largest freshwater lake (area of 2400 km2) in China.Data variables are recorded at half-hourly intervals, and themeasurement has continued for over 8 years. Several areasof research can potentially benefit from the publication ofthis dataset, including evaluation of mesoscale weather fore-cast models, development of lake–air flux parameterizations,investigation of climatic controls on lake evaporation, val-idation of remote-sensing surface data products and globalsynthesis on lake–air interactions.This paper is organized as follows. Section 2 is a briefoverview of the sites and the instruments used by the net-work. This is followed, in Sect. 3, with a description of dataquality measures employed during the field monitoring. Sec-tion 4 provides the essential information about the dataset,including data variables, gap-filling methods and data qual-ity flags. Results of postfield evaluation of the data qualityare given in Sect. 5.Users of this dataset may be interested in the relevantpapers published by our group. Lee et al. (2014) gave anoverview of the Lake Taihu eddy flux network. Using thedata collected at a subset of the sites and during the earlyphase of the network. Wang et al. (2014) investigated thespatial variability of energy and momentum fluxes acrossthe lake. Xiao et al. (2013) improved the bulk parameter-izations of heat, water and momentum fluxes for shallowlakes. Deng et al. (2013) and Hu et al. (2017) modifiedthe Community Land Model (CLM) lake simulator (Subinet al., 2012) to improve its prediction of the lake evapora-tion. Wang et al. (2017) and X. Zhang et al. (2019) evalu-ated the performance of two mesoscale models of the lake–land breeze. More recently, Xiao et al. (2020) investigateddrivers of the interannual variability of the lake evaporationobserved at one of the lake sites (Bifenggang). The valueof the dataset is enhanced by these peer-reviewed publica-tions because they have helped us to continuously improveour measurement and data-processing protocols. For exam-ple, we have used the locally calibrated bulk parameteriza-tions of Xiao et al. (2013) to gap-fill the flux variables.2Sites and instrumentation2.1Sites and data periodsTable 1 shows the basic site information, and Fig. 1 is amap that gives the relative position of Lake Taihu in Chinaand locations of the EC measurement sites. Also shown inFig. 1 are World Meteorological Organization (WMO) base-line weather stations around the lake, whose data can be ob-tained from the National Meteorological Information Cen-ter in China (http://data.cma.cn/site/index.html, last access:24 October 2020). The lake, located between the latitudinalrange of 30◦5′40′′ to 31◦32′58′′ N and longitudinal range of119◦52′32′′ to 120◦36′10′′ E, has a total area of 2400 km2and an average depth of 1.9 m. The climate is subtropicalmonsoon, with an annual mean temperature of 16.2 ◦C andannual total precipitation of 1122 mm. The lake is ice-freethroughout the year.The EC network consists of seven lake sites and one landsite. The lake sites (Meiliangwan, MLW; Dapukou, DPK;Bifenggang, BFG; Xiaoleishan, XLS; Pingtaishan, PTS;Dongtaihu, DTH; Meiliangwan2, MLW2) are distributed ac-cording to biological characteristics and across eutrophica-tion gradients of the lake. The MLW site, located in Meil-iangwan Bay near the northern shore of Lake Taihu, was thefirst site in operation; the measurement began in June 2010and was replaced by MLW2 in 2018, 10 km southwest ofMLW. Both MLW and MLW2 sites are located in the lakeeutrophic zone. BFG is located in the eastern part of LakeTaihu in relatively clean water inhabited by submerged veg-etation with a growth season from April to November. DTHis located in the shallow water (mean depth of 1.3 m) in thesoutheastern part of the lake. After more than 20 years of crabaquaculture, this zone was returned to an unmanaged state inDecember 2018 in order to improve water quality. The obser-vation at DTH enables the examination of lake–air exchangeprocesses in the transition from human management to a nat-ural state. PTS is situated in the middle of Lake Taihu, whereoccasional algal blooms occur, and no aquatic vegetation ispresent. DPK is located near the western shore, in a rela-tively deep (depth 2.5 m) super eutrophic zone due to heavyinfluence of agricultural and urban runoffs. XLS is locatedin the relatively clean and vegetation-free zone in the south-east. Finally, DS is a land site surrounded by rice agriculture,serving as a land reference for the lake sites. The MLW siteis situated at a distance of 200 m from the northern shore ofthe lake. All the other lake sites in the lake are at a distanceof more than 1 km away from the land.The lake water level is monitored daily by the Taihu BasinAuthority at five locations around the lake (http://www.tba.gov.cn/, last access: 24 October 2020). Using the water leveltime series, we have constructed the water depth for our eddycovariance sites (Fig. 2).Earth Syst. Sci. Data, 12, 2635–2645, 2020https://doi.org/10.5194/essd-12-2635-2020Z. Zhang et al.: Dataset from the Lake Taihu eddy flux network2637Table 1. A list of sites in the Lake Taihu eddy flux network.Site IDMLWDPKBFGXLSPTSMLW2DTHDSLat (◦ N)31.419731.266131.168530.997231.232331.381831.061131.0799Long (◦E)120.2139119.9312120.3972120.1344120.1086120.1608120.4704120.4346Start dateJun 2010Aug 2011Dec 2011Nov 2012Jun 2013Feb 2018Nov 2017Apr 2011BiologyEutrophicSuper eutrophicSubmergedmacrophyteTransitionalMesotrophicEutrophicAquacultureCropland/ruralresidenceMet height (m)3.58.08.59.48.56.04.510.0Tw/Ts depths (cm)20, 50, 100, 150,sediment20, 50, 100,150, sediment20, 50, 100,150, sediment20, 50, 100,150, sediment20, 50, 100,150, sediment20, 50, 100,150, sediment20, 50,sediment5, 10, 20Radiation height (m)1.53.02.52.52.52.52.01.5EC height (m)3.58.58.59.48.56.04.520Figure 1. Map showing locations of Lake Taihu, eddy covariance sites (red bubbles) and WMO weather stations (yellow triangles). Citynames are shown in blue. DS is a land site, and MLW, MLW2, DPK, PTS, XLS, BFG and DTH are lake sites. The background is a naturalcolor image from Landsat 8 without correction for atmospheric interference.2.2InstrumentationEach site is equipped with an EC system for long-term, con-tinuous monitoring of the surface momentum, sensible heat,latent heat and carbon dioxide fluxes. The EC system consistsof a sonic anemometer and thermometer (Model CSAT3A;Campbell Scientific, Logan, UT, USA) and a CO2/H2O in-frared gas analyzer (Model 7500A, LI-COR, Inc., Lincoln,NE, USA at DS, MLW, MLW2 and DPK; Model EC150,Campbell Scientific, at other sites). The EC system is at aheight of 3.5 to 9.4 m above the water surface at the lake sitesand at a height of 20 m above the ground at the land site.Other measurements include air humidity and air temper-ature (Model HMP45D/HMP155A; Vaisala, Inc, Helsinki,Finland) as well as wind speed and wind direction (Model03002; R. M. Young Company, Traverse City, MI, USA) andfour components of the net radiation (Model CNR4; Kipp &Zonen B.V., Delft, the Netherlands). At the lake sites, watertemperature profile was measured with temperature probes(Model 109-L; Campbell Scientific) at water depths of 20,50, 100 and 150 cm and in the sediment at about 5 cm belowthe bottom of the water column. The top four temperaturesensors were tied to a nylon rope hanging from a buoy toensure that they were at the designed depths regardless ofwater level fluctuations. At the DS land site, soil tempera-ture profile was measured with the same type of probes atdepths of 5, 10 and 20 cm. The MLW and the DS sites aresupported by AC power, and other sites are powered by bat-tery packs connected to solar panels. Measurements at thelake sites were made on fixed platforms. Readers are referredto Lee et al. (2014) and Xiao et al. (2017) for photographs ofthe platform and the instruments.All the variables are reported as 30 min averages. The ECdata are expressed in the natural coordinate system (Lee etal., 2004). In this coordinate system, the longitudinal coordi-nate axis is aligned with the 30 min mean velocity vector sothat the 30 min mean lateral and vertical velocity componentsare 0, the magnitude of the mean velocity is equal to the meanlongitudinal component, and the covariance between the lat-eral and the vertical velocity components is 0. Additionally, ahttps://doi.org/10.5194/essd-12-2635-2020Earth Syst. Sci. Data, 12, 2635–2645, 20202638Z. Zhang et al.: Dataset from the Lake Taihu eddy flux networkFigure 2. Water depth at the eddy covariance sites.Table 2. Percent of data coverage. The percentage represents the proportion of high-quality original measurement.Variable typeMLWDPKBFGXLSPTSDTHMLW2DSMicrometeorology93.381.197.697.097.598.190.391.7Radiation flux85.590.896.997.498.698.298.282.7Water or soil temperature83.481.394.091.190.387.722.498.4Eddy flux73.361.882.779.180.685.785.582.8Table 3. A list of data quality flagsFlagData quality description0Original data1Gap-filled with time interpolation2Gap-filled with spatial interpolation3Gap-filled with bulk relationship4NANsmall density correction has been applied to the water vaporflux according to Webb et al. (1980).3Data quality control during field monitoringEvery site in the Lake Taihu eddy flux network is equippedwith a wireless transmission module for real-time monitoringand for data transmission. Time series of all 30 min variablesare examined weekly, and abnormal behaviors are flaggedfor site operators. Each site is visited every 1 to 2 months toperform instrument repair and maintenance and to download10 Hz EC data. The data coverage rates are summarized inTable 2, where the percentage values represent the propor-tions of data with quality flag 0, which indicates high-qualityoriginal measurement (Table 3).The four-way net radiometers at MLW and XLS werecompared in the field against a laboratory standard of thesame type in the summer of 2018 to check their long-termstability (Fig. 3). These two sites were chosen because theyhave been in operation for more than 5 years. Addition-ally, the radiometer at MLW was relocated to MLW2 afterMLW had been discontinued. The laboratory standard, whichhad been calibrated at the manufacturer prior to this perfor-mance evaluation, was mounted next to the field instrumentfor about 10 d at each site, covering overcast to clear-sky con-ditions. The mean bias error was smaller than 1 W m−2 forall the radiation components. It was −0.81, −0.81, 0.79 and−0.44 W m−2 for the downward shortwave, upward short-wave, downward longwave and upward longwave radiationflux at MLW, respectively. The corresponding values were0.91, 0.40, 0.69 and 0.77 W m−2 for XLS. (Comparison ex-periments are being planned for the other sites.)The EC gas analyzers were calibrated every 1 to 2 years.The zero-point calibration was carried out with high-puritynitrogen gas, the CO2 span calibration was made with stan-dard carbon dioxide gases (in the concentration range of 389to 525 ppm) provided by the National Institute of Metrologyof China (NIM) and certified to an accuracy of 1 %, and theH2O span calibration was made with a portable dew pointgenerator (LI-610; LI-COR, Inc.).Earth Syst. Sci. Data, 12, 2635–2645, 2020https://doi.org/10.5194/essd-12-2635-2020Z. Zhang et al.: Dataset from the Lake Taihu eddy flux network2639Figure 3. Comparison of four components of the radiation balance between the original radiometer (horizontal axis) and a laboratorystandard (vertical axis) at MLW and XLS. Refer to Table 4 for variable definitions.4Gap-filling methods and data quality flagsWe use a five-point moving average to screen outliers. If thedeviation from the moving average is greater than 2 stan-dard deviations, the data point is discarded. If a gap length is30 min to 1 h, the gap is filled by linear interpolation. Largergaps in meteorological variables, radiation components andwater temperature are filled with linear regression involvingobservation of the same variable at another site. This spa-tial interpolation consists of three steps. First, linear correla-tion is calculated using the valid data at the target site andat all other sites for the month during which the data gapoccurred. Second, the observation at the site with the high-est linear correlation is used to establish a linear-regressionequation. Third, the gap at the target site is filled with thelinear regression and the observation at that site.Radiation data gaps at the DS land site require specialtreatment. The radiometer at the DS eddy flux site endedin January 2013. Subsequent measurements of the radiationcomponent are provided by a radiometer belonging to theDongshan WMO weather station at a distance of 50 m fromthe eddy covariance tower (Fig. 1). While large gaps in mete-orological variables (air temperature, relative humidity, windspeed and air pressure), downward solar radiation and down-ward longwave radiation are filled with the spatial interpola-tion method, large gaps in upward shortwave radiation andupward longwave radiation cannot be filled with data fromother lake sites even with linear regression. In the case of theupward shortwave radiation, the data gaps were filled usingthe relationship between downward shortwave radiation andthe monthly mean albedo. In the case of upward longwaveradiation, the data gaps were filled by a regression equationbetween the upward longwave radiation and the fourth powerof soil temperature at 5 cm depth. Compared to the originaldata, the gap-filled data do not capture the full diurnal vari-ations because the 5 cm soil temperature has smaller diurnalamplitudes than the soil surface temperature, but the dailymean upward longwave radiation flux seems reasonable.Large data gaps in the EC variables (sensible heat flux, la-tent heat flux and friction velocity) are filled with a hybridmethod. First, if observations exist for the relevant state vari-able, the gap is filled with the bulk transfer relationship usinga transfer coefficient tuned locally for each site (Xiao et al.,2013). For example, the relationship for filling gaps in thesensible heat flux isH = ρacpCHU(Ts − Ta),where ρa is air density, cp is specific heat of air at constantpressure, CH is the transfer coefficient for sensible heat, Tais air temperature, and Ts is water surface temperature. Thetransfer coefficient CH is determined from the observed Hand the state variables (U, Ta and Ts) outside gap periods.The missing data on H are then filled with the above rela-tionship using the tuned CH and the observed U, Ta and Ts.Second, if data for the state variables are missing, the spa-tial interpolation method is used to fill the gaps in these ECvariables.The spatial interpolation method described above occa-sionally causes a sudden jump at the beginning or end of adata gap. To harmonize the data, we apply a five-point mov-ing average to the gap-filled time series. If a data point de-viates by 2 times the standard deviation from the movingaverage, it is replaced by linear interpolation using the twoadjacent data points.Each data variable is assigned a quality flag to distinguishoriginal measurements and gap-filled values and gap-fillinghttps://doi.org/10.5194/essd-12-2635-2020Earth Syst. Sci. Data, 12, 2635–2645, 20202640Z. Zhang et al.: Dataset from the Lake Taihu eddy flux networkTable 4. A list of data columns and variable definitions.ColumnDescriptionVariable nameUnit1YearYear–2MonthMonth–3DayDay–4HourHH–5MinuteMM–6Day of YearDOY–7Air pressurePkPa8Quality flag of air pressureP _flag9Air temperatureTa◦C10Quality flag of air temperatureTa_flag11Relative humidityRH%12Quality flag of relative humidityRH_flag13Wind speedWSm s−114Quality flag of wind speedWS_flag15Wind directionWD◦16Quality flag of wind directionWD_flag17RainfallRmm18Quality flag of rainfallR_flag19Upward shortwave radiationURW m−220Quality flag of upward shortwave radiationUR_flag21Downward shortwave radiationDRW m−222Quality flag of downward shortwave radiationDR_flag23Upward longwave radiationULRW m−224Quality flag of upward longwave radiationULR_flag25Downward longwave radiationDLRW m−226Quality flag of downward longwave radiationDLR_flag27Water temperature at 0.2 mTw_20◦C28Quality flag of water temperature at 0.2 mTw_20_flag29Water temperature at 0.5 mTw_50◦C30Quality flag of water temperature at 0.5 mTw_50_flag31Water temperature at 1.0 mTw_100◦C32Quality flag of water temperature at 1.0 mTw_100_flag33Water temperature at 1.5 mTw_150◦C34Quality flag of water temperature at 1.5 mTw_150_flag35Sediment temperatureTw_bot◦C36Quality flag of sediment temperatureTw_bot_flag37Friction velocityU∗m s−138Quality flag of friction velocityU∗_flag39Sensible heat fluxHW m−240Quality flag of sensible heat fluxH_flag41Latent heat fluxLEW m−242Quality flag of latent heat fluxLE_flagNotes: (1) time marks end of a half-hourly observation in Beijing time (UTC+8); (2) at the DS site, columns 27,29 and 31 represent soil temperature at 5, 10 and 20 cm, respectively; column 33 represents soil heat flux G(W m−2) measured at 5 cm depth; and column 34 represents quality flag of soil heat flux.methods (Table 3). The data flags employed here should notbe confused with quality flags commonly assigned to the ECmethodology in the literature. Specifically, Flag 0 indicateshigh-quality original data. Other flag values indicate gap-filled data or missing values. Flag 1 indicates that the datawere filled by temporal interpolation. Flag 2 indicates thatthe data were filled by the spatial interpolation method. Flag3 for the EC variables indicates that the data were filled bythe bulk relationship. We also use Flag 3 to mark the upwardshortwave and longwave radiation data filled with the albedoand the surface temperature relationship, respectively, for theDS land site. Missing values occur in some situations, whichare marked with Flag 4. Figure 4 is an example showing thegap-filled time series of several variables at BFG along withthe flag status.Earth Syst. Sci. Data, 12, 2635–2645, 2020https://doi.org/10.5194/essd-12-2635-2020Z. Zhang et al.: Dataset from the Lake Taihu eddy flux network2641Figure 4. Complete gap-filled time series for selected variables observed at BFG. Blue, black, cyan and red dots represent quality flag 0, 1,2 and 3, respectively. Variable definitions are given in Table 4.Rainfall data have not been quantity-controlled or gap-filled. Because of the episodic nature of rainstorms and highspatial variability of rainfall, it is not appropriate to filldata gaps with the time interpolation or spatial-interpolationmethod. The total rain amount is likely biased low becauseno wind screens are used to protect the rain gages from theinfluence of wind which is much higher on the lake than onland (Fig. 5 below). On several site visits, the drain openingto the tipping bucket was found to be partially blocked by de-bris. Rain amount at a constant and low rate and excessivelylong rain duration are evidence of such blockage. The flagstatus of 0 for the rainfall variable simply indicates that thefield measurement is available, but it does not guarantee highdata quality.The data coverage begins from the start time of each site(Table 1) and ends in December 2018. The time resolu-tion is 30 min. The dataset includes microclimate variables(air pressure, air temperature, relative humidity, wind speed,wind direction and rainfall); radiation fluxes (upward anddownward shortwave radiation, upward and downward long-wave radiation); water temperature at depths of 0.2, 0.5, 1.0and 1.5 m and in the 5 cm sediment; and eddy fluxes (fric-tion velocity, sensible heat and latent heat fluxes; Table 4).The time stamp is Beijing time (UTC+8), given by datacolumns 1 to 5 as year, month, day, hour and minute, andmarks the end of the observation period. For example, timestamp “2012, 1, 1, 12, 00” indicates that the data acquisitionperiod is from 11:30 to 12:00 UTC+8 on 1 January 2012.Although the data table does not include the radiative sur-face temperature Ts, the user can easily calculate it from thetwo longwave radiation fluxes asTs =�L↑ − (1 − ε)L↓εσ� 14,where σ is the Stefan–Boltzmann constant; ε is emissivity;and L↑and L↓ are upward and downward longwave radiationflux, respectively. We use a value of 0.97 for lake surfaceemissivity in this calculation (Deng et al., 2013; Wang et al.,2014).5Data consistency evaluationFigure 5 compares the annual mean air temperature, rela-tive humidity and wind speed at the Taihu eddy flux siteswith those at the four WMO weather stations (Wuxi, Liyang,Huzhou and Dongshan) around the lake (Fig. 1). The errorbars represent the maximum and minimum values among thefour WMO stations, and the lines represent the mean val-ues of the four station measurements. The annual mean airtemperature at DTH is 0.3 ◦C higher than the station mean.https://doi.org/10.5194/essd-12-2635-2020Earth Syst. Sci. Data, 12, 2635–2645, 20202642Z. Zhang et al.: Dataset from the Lake Taihu eddy flux networkFigure 5. Annual mean air temperature (a), relative humidity (b) and wind speed (c) observed at the eddy flux sites (symbols) and at thefour WMO weather stations around the lake (line). Error bars represent the range of the annual means of the four WMO stations.At other sites, air temperature is in close agreement with theweather station data in terms of both magnitude and inter-annual variability. The annual mean wind speed at MLW, asite near the shoreline, is comparable with the station data.At other more exposed sites, the wind speed is much higherthan observed at the WMO stations. The annual mean rel-ative humidity (RH) shows a larger spread among the eddyflux sites than among the WMO stations partly because themeasurement height at the eddy flux sites is not standardized(Table 1). The upward trends in RH over time at DPK andXLS seem to be related more to aging of the sensor than toa real interannual variability. We have not fully investigatedthis aging problem, but it is possible to rectify it by doing adetailed regression analysis against the station data.Consistency of the energy flux variables can be evaluatedwith the energy balance closure. Using observations made ata subset of the sites in the earlier years of the flux network,Wang et al. (2014) reported a closure rate of 70 % to 110 %on a monthly basis, meaning that the sum of the measuredmonthly sensible and the latent heat flux H + λE is 70 % to110 % of the monthly available energy Rn − G, where Rn isnet radiation, and G is heat storage in the water column. Byselecting days without data gaps, we found that the daily en-ergy balance closure is in the range between 66 % and 78 %for all the lake sites and all the years. Such closure rates aretypical of eddy covariance observations (Tanny et al., 2008;Wilson et al., 2002).We have shown that the monthly latent heat flux at the lakesites MLW, BFG and DPK during July 2010 to August 2012follows the Priestley–Taylor (PT) model prediction with theFigure 6. Comparison of observed monthly latent heat flux withPriestley–Taylor model prediction using the original α coefficient of1.26 and a modified coefficient of 1.03. Here Rn is net radiation, Gis heat storage in the water column, � is the slope of the saturationvapor pressure curve, and γ is the psychrometric constant.original PT constant α of 1.26 and that at the DS land siteit is in agreement with the PT model if the constant is low-ered to 1.0 (Lee et al., 2014). Figure 6 demonstrates that thesame relationships hold for all the sites and all the observa-tional months, indicating the overall stability of our measure-ment systems and the robustness of our gap-filling procedure.Earth Syst. Sci. Data, 12, 2635–2645, 2020https://doi.org/10.5194/essd-12-2635-2020Z. Zhang et al.: Dataset from the Lake Taihu eddy flux network2643Figure 7. The relationship between changes in observed annual mean upward longwave radiation flux and annual mean air temperature(dots). Solid lines represent the prediction of the Stefan–Boltzmann law.The reader is reminded that the monthly latent heat flux inFig. 6 has been adjusted to force energy closure followingthe method recommended by Barr et al. (1994), Blanken etal. (1997) and Twine et al. (2000). (The half-hourly flux datain the data archive have not been adjusted for energy bal-ance.)The Stefan–Boltzmann law offers another way for check-ing data consistency. Because the lake surface emits long-wave radiation like a blackbody and because the annual meanair temperature and the surface water temperature are nearlyidentical at this lake (Wang et al., 2014), the change in theannual upward longwave radiation �L↑ can be expressed as�L↑ = 4σT 3a �Ta,where Ta is annual mean air temperature, and � is the differ-ence between the target year and the year with the lowest airtemperature observed at the site. All the five long-term lakesites show good consistency between the longwave radiationand the air temperature observations (Fig. 7).Table 5 is a summary of the uncertainty of key measure-ment variables at half-hourly intervals. The performance un-certainty is 1 standard deviation of difference in a variablemeasured by the field instrument and the same variable mea-sured by a validation instrument (the closed-path EC in thecase of eddy fluxes and the laboratory standard radiometerin the case of the radiation fluxes). The environmental uncer-tainty is 1 standard deviation of spatial variation in a variablemeasured at multiple lake sites.6Data availabilityAll data are open-access and are available online for down-load and use at https://yncenter.sites.yale.edu/data-access(last access: 24 October 2020) and from the Harvard Data-verse (https://doi.org/10.7910/DVN/HEWCWM; Zhang etal., 2020).7SummaryThe dataset described here consists of microclimate variables(air temperature, air humidity, wind speed, wind direction,Table 5. Uncertainty of key measurement variables at half-hourlyintervals. Instrument uncertainty is provided by the manufacturers.Performance uncertainty is 1 standard deviation of the differencebetween measurements made by the field instrument and the vali-dation instrument. Environmental uncertainty is the spatial standarddeviation of the variable measured at the lake sites.VariableUncertaintyPeriod of evaluationInstrument uncertaintyP±0.3 hPaTa±0.2 ◦CRH±2 %WS±0.3 m s−1WD±3◦UR/DR< 5 %ULR/DLR< 10 %Tw±0.6 ◦CPerformance uncertaintyUR±2.1 W m−229 June–8 July 2018,6–15 October 2018DR±8.0 W m−229 June–8 July 2018,6–15 October 2018ULR±0.5 W m−229 June–8 July 2018,6–15 October 2018DLR±1.3 W m−229 June–8 July 2018,6–15 October 2018U∗±0.06 m s−113 July–23 August 2020H±3.1 W m−213 July–23 August 2020LE±21.2 W m−213 July–23 August 2020Environmental uncertaintyWater depth±0.06 m1 September 2017–31 August 2018Ta±0.50 ◦C1–31 July 2018DR±36.3 W m−21–31 July 2018https://doi.org/10.5194/essd-12-2635-2020Earth Syst. Sci. Data, 12, 2635–2645, 20202644Z. Zhang et al.: Dataset from the Lake Taihu eddy flux networkwater or soil temperature profile, and rainfall), four compo-nents of the radiation balance, friction velocity, and sensibleand latent heat fluxes observed at seven lake sites and oneland site. The period of coverage is from June 2010 to De-cember 2018. The observation interval is 30 min. Except forrainfall and wind direction, all other variables have been gap-filled. Every data point is tagged with a data quality flag tohelp the user determine how to best use the data.Author contributions. XL, WX and MZ directed the field pro-gram; ZZ performed data gap-filling and prepared the data for pub-lic release; CC, WW, CX, HC, JW, JZ, LJ, QL, WH, WZ, YL, YX,YW, YP, YH, ZC and ZQ participated in field data collection; andZZ, XL and MZ wrote the manuscript.Competing interests. The authors declare that they have no con-flict of interest.Financial support. This research has been supported by the Na-tional Key R&D Program of China (grant no. 2019YFA0607202),the National Natural Science Foundation of China (grant nos.41575147, 41801093, 41475141), and the Priority Academic Pro-gram Development (PAPD) of Jiangsu Higher Education Institu-tions.Review statement. This paper was edited by David Carlson andreviewed by two anonymous referees.ReferencesBarr, A. G., King, K. M., Gillespie, T. J., Den Hartog, G., and Neu-mann, H. H.: A comparison of Bowen ratio and eddy correlationsensible and latent heat flux measurements above deciduous for-est, Bound.-Lay. Meteorol., 71, 21–41, 1994.Blanken, P. D., Black, T. A., Yang, P. C., Neumann, H. H., Nesic,Z., Staebler, R., Den Hartog, G., Novak, M. D., and Lee, X.:Energy balance and canopy conductance of a boreal aspen forest:partitioning overstory and understory components, J. Geophys.Res.-Atmos., 102, 28915–28927, 1997.Blanken,P.D.,Spence,C.,Hedstrom,N.,andLenters,J. D.: Evaporation from Lake Superior: 1. Physical con-trols and processes, J. Great Lakes Res., 37, 707–716,https://doi.org/10.1016/j.jglr.2011.08.009, 2011.Charusombat, U., Fujisaki-Manome, A., Gronewold, A. D., Lof-gren, B. M., Anderson, E. J., Blanken, P. D., Spence, C., Lenters,J. D., Xiao, C., Fitzpatrick, L. E., and Cutrell, G.: Evaluatingand improving modeled turbulent heat fluxes across the NorthAmerican Great Lakes, Hydrol. Earth Syst. Sci., 22, 5559–5578,https://doi.org/10.5194/hess-22-5559-2018, 2018.Deng, B., Liu, S., Xiao, W., Wang, W., Jin, J., and Lee,X.: Evaluation of the CLM4 lake model at a large andshallow freshwater lake, J. Hydrometeorol., 14, 636–649,https://doi.org/10.1175/JHM-D-12-067.1, 2013.Downing, J. A., Prairie, Y. T., Cole, J. J., Duarte, C. M.,Tranvik, L. J., Striegl, R. G., McDowell, W. H., Korte-lainen, P., Caraco, N. F., Melack, J. M., and Middelburg,J. J.: The global abundance and size distribution of lakes,ponds, and impoundments, Limnol. Oceanogr., 51, 2388–2397,https://doi.org/10.4319/lo.2006.51.5.2388, 2006.Du, Q., Liu, H., Xu, L., Liu, Y., and Wang, L.: The monsoon effecton energy and carbon exchange processes over a highland lake inthe southwest of China, Atmos. Chem. Phys., 18, 15087–15104,https://doi.org/10.5194/acp-18-15087-2018, 2018.Franz, D., Mammarella, I., Boike, J., Kirillin, G., Vesala, T., Borne-mann, N., Larmanou, E., Langer, M., and Sachs, T.: Lake-atmosphere heat flux dynamics of a thermokarst lake in arcticSiberia, J. Geophys. Res.-Atmos., 123, 5222–5239, 2018.Franz, D., Mammarella, I., Boike, J., Kirillin, G., Vesala, T., Borne-mann, N., Larmanou, E., Langer, M., and Sachs, T.: Lake-atmosphere heat flux dynamics of a thermokarst lake in arcticSiberia, J. Geophys. Res.-Atmos., 123, 5222–5239, 2018.Hamdani, I., Assouline, S., Tanny, J., Lensky, I. M., Gertman, I.,Mor, Z., and Lensky, N. G.: Seasonal and diurnal evaporationfrom a deep hypersaline lake: the Dead Sea as a case study, J.Hydrol., 562, 155–167, 2018.Hu, C., Wang, Y., Wang, W., Liu, S., Piao, M., Xiao, W., and Lee,X.: Trends in evaporation of a large subtropical lake, Theor.Appl. Climatol., 129, 159–170, https://doi.org/10.1007/s00704-016-1768-z, 2017.Lee, X., Massman, W., and Law, B.: Handbook of micrometeorol-ogy: a guide for surface flux measurement and analysis, SpringerScience & Business Media, 2004.Lee, X., Liu, S., Xiao, W., Wang, W., Gao, Z., Cao, C., Hu, C.,Hu, Z., Shen, S., Wang, Y., Wen, X., Xiao, Q., Xu, J., Yang,J., and Zhang, M.: The taihu eddy flux network: An observa-tional program on energy, water, and greenhouse gas fluxes ofa large freshwater lake, B. Am. Meteorol. Soc., 95, 1583–1594,https://doi.org/10.1175/BAMS-D-13-00136.1, 2014.Li, Z., Lyu, S., Ao, Y., Wen, L., Zhao, L., and Wang, S.: Long-term energy flux and radiation balance observations over LakeNgoring, Tibetan Plateau, Atmos. Res., 155, 13–25, 2015.Nordbo, A., Launiainen, S., Mammarella, I., Leppäranta, M., Huo-tari, J., Ojala, A., and Vesala, T.: Long-term energy flux measure-ments and energy balance over a small boreal lake using eddycovariance technique, J. Geophys. Res.-Atmos., 116, D02119,https://doi.org/10.1029/2010JD014542, 2011.Subin, Z. M., Riley, W. J., and Mironov, D.: An improved lakemodel for climate simulations: Model structure, evaluation, andsensitivity analyses in CESM1, J. Adv. Model. Earth Sy., 4, 1–27,https://doi.org/10.1029/2011MS000072, 2012.Tanny, J., Cohen, S., Assouline, S., Lange, F., Grava, A., Berger, D.,Teltch, B., and Parlange, M. B.: Evaporation from a small waterreservoir: Direct measurements and estimates, J. Hydrol., 351,218–229, 2008.Twine, T. E., Kustas, W. P., Norman, J. M., Cook, D. R., Houser,Pr., Meyers, T. P., Prueger, J. H., Starks, P. J., and Wesely, M. L.:Correcting eddy-covariance flux underestimates over a grassland,Agr. For. Meteorol., 103, 279–300, 2000.Verpoorter, C., Kutser, T., Seekell, D. A., and Tranvik, L.J.: A global inventory of lakes based on high-resolutionsatelliteimagery,Geophys.Res.Lett.,41,6396–6402,https://doi.org/10.1002/2014GL060641, 2014.Earth Syst. Sci. Data, 12, 2635–2645, 2020https://doi.org/10.5194/essd-12-2635-2020Z. Zhang et al.: Dataset from the Lake Taihu eddy flux network2645Vesala, T., Huotari, J., Rannik, Ü., Suni, T., Smolander, S., So-gachev, A., Launiainen, S., and Ojala, A.: Eddy covariance mea-surements of carbon exchange and latent and sensible heat fluxesover a boreal lake for a full open-water period, J. Geophys.Res.-Atmos., 111, 1–12, https://doi.org/10.1029/2005JD006365,2006.Wang, B., Ma, Y., Wang, Y., Su, Z., and Ma, W.: Significant dif-ferences exist in lake-atmosphere interactions and the evapora-tion rates of high-elevation small and large lakes, J. Hydrol., 573,220–234, 2019.Wang, W., Xiao, W., Cao, C., Gao, Z., Hu, Z., Liu, S., Shen, S.,Wang, L., Xiao, Q., Xu, J., Yang, D., and Lee, X.: Temporaland spatial variations in radiation and energy balance acrossa large freshwater lake in China, J. Hydrol., 511, 811–824,https://doi.org/10.1016/j.jhydrol.2014.02.012, 2014.Wang, Y., Gao, Y., Qin, H., Huang, J., Liu, C., Hu, C., Wang,W., Liu, S., and Lee, X.: Spatiotemporal Characteristics of LakeBreezes over Lake Taihu, China, J. Appl. Meteorol. Climatol.,56, 2053–2065, 2017.Webb, E. K., Pearman, G. I., and Leuning, R.: Correction of fluxmeasurements for density effects due to heat and water vapourtransfer, Q. J. R. Meteorol. Soc., 106, 85–100, 1980.Wilson, K., Goldstein, A., Falge, E., Aubinet, M., Baldocchi, D.,Berbigier, P., Bernhofer, C., Ceulemans, R., Dolman, H., andField, C.: Energy balance closure at FLUXNET sites, Agr. For.Meteorol., 113, 223–243, 2002.Xiao, K., Griffis, T. J., Baker, J. M., Bolstad, P. V, Erickson, M. D.,Lee, X., Wood, J. D., Hu, C., and Nieber, J. L.: Evaporation froma temperate closed-basin lake and its impact on present, past, andfuture water level, J. Hydrol., 561, 59–75, 2018.Xiao, W., Liu, S., Wang, W., Yang, D., Xu, J., Cao, C., Li,H., and Lee, X.: Transfer Coefficients of Momentum, Heatand Water Vapour in the Atmospheric Surface Layer of aLarge Freshwater Lake, Bound.-Lay. Meteorol., 148, 479–494,https://doi.org/10.1007/s10546-013-9827-9, 2013.Xiao, W., Zhang, Z., Wang, W., Zhang, M., Liu, Q., Hu,Y., Huang, W., Liu, S., and Lee, X.: Radiation Controlsthe Interannual Variability of Evaporation of a Subtropi-cal Lake, J. Geophys. Res.-Atmos., 125, p.e2019JD031264,https://doi.org/10.1029/2019JD031264, 2020.Yusup, Y. and Liu, H.: Effects of Atmospheric Surface Layer Sta-bility on Turbulent Fluxes of Heat and Water Vapor across theWater–Atmosphere Interface, J. Hydrometeorol., 17, 2835–2851,2016.Zhang, X., Huang, J., Li, G., Wang, Y., Liu, C., Zhao, K., Tao, X.,Hu, X.-M., and Lee, X.: Improving Lake-Breeze Simulation withWRF Nested LES and Lake Model over a Large Shallow Lake,J. Appl. Meteorol. Climatol., 58, 1689–1708, 2019.Zhang, Z., Zhang, M., Cao, C., Wang, W., Xiao, W., Xie, C., Chu,H., Wang, J., Zhao, jiayu, Jia, L., Liu, Q., Huang, W., Zhang,W., Lu, Y., Xie, Y., Wang, Y., Pu, Y., Hu, Y., Chen, Z., Qin, Z.,and Lee, X.: A dataset of microclimate and radiation and energyfluxes from the Lake Taihu Eddy Flux Network, Harvard Data-verse, https://doi.org/10.7910/DVN/HEWCWM, 2020.Zhao, X. and Liu, Y.: Variability of surface heat fluxes and its driv-ing forces at different time scales over a large ephemeral lake inChina, J. Geophys. Res.-Atmos., 123, 4939–4957, 2018.https://doi.org/10.5194/essd-12-2635-2020Earth Syst. Sci. Data, 12, 2635–2645, 2020 | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/354352413Paleoproterozoic ophiolitic mélanges and orogenesis in the northern YangtzeCraton: Evidence for the operation of modern-style plate tectonicsArticle in Precambrian Research · September 2021DOI: 10.1016/j.precamres.2021.106385CITATIONS4READS4426 authors, including:Some of the authors of this publication are also working on these related projects:Precambrian geodynamics of the North China and Yangtze cratons View projectCrustal Evelution of South China Craton View projectWenxiao ZhouChina University of Geosciences36 PUBLICATIONS 316 CITATIONS SEE PROFILEBo HuangChina University of Geosciences29 PUBLICATIONS 471 CITATIONS SEE PROFILEYunxu Wei 魏运许17 PUBLICATIONS 79 CITATIONS SEE PROFILEHaiquan LiChina University of Geosciences17 PUBLICATIONS 22 CITATIONS SEE PROFILEAll content following this page was uploaded by Haiquan Li on 07 December 2022.The user has requested enhancement of the downloaded file.Precambrian Research 383 (2022) 106905Available online 9 November 20220301-9268/© 2022 Elsevier B.V. All rights reserved.Mesoarchean banded iron-formation from the northern Yangtze Craton, South China and its geological and paleoenvironmental implications Hang Zhou a, Wenxiao Zhou a,b,*, Yunxu Wei c,d,**, Ernest Chi Fru b, Bo Huang e,f, Dong Fu f, Haiquan Li a, Mantang Tan c,d a Institute of Geological Survey, China University of Geosciences, Wuhan 430074, China b School of Earth and Ocean Sciences, Centre for Geobiology and Geochemistry, Cardiff University, Cardiff CF10 3AT, Wales, United Kingdom c Wuhan Center of China Geological Survey, Wuhan 430205, China d Central South China Innovation Center for Geosciences, Wuhan 430205, China e National Observation and Research Station for Geohazards, Three Gorges Research Center for Geohazards, China University of Geosciences, Wuhan 430074, China f Center for Global Tectonics, State Key Laboratory of Geological Processes and Mineral Resources, China University of Geosciences, Wuhan 430074, China A R T I C L E I N F O Keywords: Banded iron-formation Iron isotopes Paleoenvironment Mesoarchean Yangtze Craton A B S T R A C T Banded iron formations (BIFs) are important sedimentary rocks used for reconstructing the geochemical and environmental evolution of Precambrian Earth surface conditions. Here, we describe the oldest BIFs of Meso-archean age in the Yemadong Formation, and their implications for paleo-environmental evolution of the Mesoarchean Ocean in the northern Yangtze Craton, South China. U-Pb dating of metamorphic zircons obtained from metavolcanic rocks interbedded within the Yemadong Complex, record two major tecto-thermal events at 2901 ± 8 Ma, in agreement with a Mesoarchean origin of the Dujiagou BIFs, and as recently as 775 Ma. A high SiO2 + total Fe2O3 content of 90.95–94.62 wt%, moderate 0.54–1.42 wt% Al2O3 + TiO2, and broadly elevated levels of incompatible elements, point to a predominantly chemical origin of the BIF deposit, with a minor but influential terrigenous input. Rare Earth Element (REE) plus Yttrium patterns normalized against Post Archean Average Shale composition are characterized by the depletion of light REE relative to heavy REE. However, extremely weak Eu and Ce anomalies, together with low average chondrite Y/Ho ratios compared to average Mesoarchean BIF values, indicate that the primary REE composition of the rocks may have been diluted to some extent by terrigenous input, pointing to proximity or connection of the depositional setting to a continental margin environment. Despite potential terrigenous dilution of the syngenetic seawater-hydrothermal chemical composition, the preservation of an inherited hydrothermal REE signal in some samples, point to submarine hydrothermal provenance of the Dujiagou BIFs, as is the case for most Precambrian BIFs. Positive δ56Fe values, ranging from 0.25 to 0.45 ‰, characteristic of primary Archean BIFs, suggest partial oxidation of reduced Fe(II) to divalent (Fe(II)/Fe(III)) magnetite-rich BIFs in a broadly reducing Archean Ocean lacking a persistent and pervasive redoxcline before the rise of atmospheric oxygen. 1. Introduction Banded iron formations (BIFs) are marine chemical sedimentary rocks of predominantly Precambrian origin. They consist mainly of iron- rich layers containing hematite and magnetite alternating with silica- rich bands, composed principally of chert and jasper. Deposition of BIFs in Earth’s history is closely related to the secular changes in the atmosphere, hydrosphere, and lithosphere, and thus BIFs are excellent archives for reconstructing the composition and redox state of the ancient ocean-atmosphere system under which they formed (Bolhar et al., 2004; Klein, 2005; Polat and Frei, 2005; Bekker et al., 2010; Basta et al., 2011; Mloszewska et al., 2012; Zhang et al., 2012; Cui et al., 2014; Brando Soares et al., 2017; Ghosh and Baidya, 2017; Huang et al., 2019). Three principal types of BIFs are known, defined primarily by their depositional environment (Bekker et al., 2010; Li et al., 2014a; Hage-mann et al., 2016). Algoma-type BIFs deposited in deep-sea island arc environments, and near subduction trenches. They are generally the most common BIFs found in Archean greenstone belts associated with * Corresponding author at: Institute of Geological Survey, China University of Geosciences, Wuhan 430074, China. ** Corresponding author at: Wuhan Center of China Geological Survey, Wuhan 430205, China. E-mail addresses: zhouwenxiao@cug.edu.cn (W. Zhou), wyx1856@163.com (Y. Wei). Contents lists available at ScienceDirect Precambrian Research journal homepage: www.elsevier.com/locate/precamres https://doi.org/10.1016/j.precamres.2022.106905 Received 17 May 2022; Received in revised form 4 September 2022; Accepted 31 October 2022 Precambrian Research 383 (2022) 1069052old cratonic sequences (Goodwin, 1973; Klein, 2005)—are small in scale, with the oldest known deposit formed ~3.8 billion years ago (Ga) in the Isua greenstone belt, West Greenland (Myers, 2001; Bolhar et al., 2004; Polat and Frei, 2005; Czaja et al., 2013). Superior-type BIFs are larger in scale, compared to Algoma-type BIFs. They formed mainly in oxidized environments lacking syn-depositional volcanic assemblages, and are found deposited with transgressive successions that formed on shallow continental shelves located in intra-cratonic Basins (Klein, 2005; Li et al., 2012a; Li et al., 2014a). The Neoarchean and early Paleoproterozoic intervals witnessed prolific Superior-type BIF deposi-tion, with examples found in the Hamersley Basin of Western Australia and the Transvaal Supergroup, South Africa (Morris, 1993; Johnson et al., 2003; Bekker et al., 2010; Smith, 2018). Large-scale BIF deposition terminated at ~1.8 Ga, when they disappeared from the marine sedi-mentary rock record, before an abrupt but brief reappearance of Rapitan-type BIFs in conjunction with the Neoproterozoic snowball Earth glaciations, ~0.7–0.635 Ga (Hoffman et al., 1998). The long depositional history of BIFs, timing and coincidence with major redox, climatic and geochemical changes in the ocean-atmosphere system, has led to their wide use as reliable proxies for Earth oxygenation history and the evolution of the Precambrian crust (Crowley et al., 2002; Holland, 2006; Jenner et al., 2009; Poulton and Canfield, 2011; Kasting, 2013; Chi Fru et al., 2016; Chi Fru et al., 2018). Although the timing and extent of Earth surface oxidation before the Paleoproterozoic Great Oxidation Event (GOE), ~2.4–2.0 Ga remains debated (Czaja et al., 2013), most workers agree that atmospheric O2 contents were low in the Archean (e.g., Farquhar et al. (2000); Anbar et al. (2007); Guo et al. (2009)). In line with this observation, the chemical composition of Archean BIFs often contain both reduced and oxidized iron reservoirs, with the deposition of BIFs 3.8–1.8 Ga taken as compelling evidence for the prevalence of anoxic iron rich waters that enabled high iron mobility in the marine realm (Poulton and Canfield, 2011). The Early Paleoproterozoic GOE and the Late Neoproterozoic Oxidation Event, NOE (e.g., Och and Shields-Zhou (2012); Lyons et al. (2014)), are proposed as the principal reason for the eventual disap-pearance of BIFs from the marine sedimentary rock record (Poulton and Canfield, 2011). Because of their great antiquity, in many instances, the primary chemical composition of BIFs has been altered by post-depositional diagenetic and metamorphic processes. Nonetheless, a substantial body of work suggests that Fe isotopes are particularly resistant to these secondary post-depositional changes in bulk BIF composition (Dauphas and Rouxel, 2006; Frost et al., 2006; Dauphas et al., 2007a; Czaja et al., 2013). Consequently, iron isotopes often play an important role in the study of BIFs, and their distribution in BIFs used frequently to track global early Earth surface redox evolution and the workings of the early biogeochemical elemental cycles (Johnson et al., 2003; Dauphas et al., 2004; Dauphas et al., 2007a; Dauphas et al., 2007b; Hyslop et al., 2008; Johnson et al., 2008; Li et al., 2012b; Chi Fru et al., 2013; Czaja et al., 2013; Wang et al., 2017). In China, most Neoarchean-Paleoproterozoic BIFs are deposited in the North China Craton (Li et al., 2012b; Wang et al., 2014; Wang et al., 2017; Wu et al., 2020), while the presence of BIFs in the older Meso-archean South China Block, host to the Yangtze Craton, remains poorly explored. Consequently, there exists a considerable gap in our under-standing of the redox and biogeochemical development of the iron-rich Mesoarchean oceans in the Yangtze Craton. This has major implications for the global evolution of the early ocean-atmosphere system, with respect to the independent development of the South China Meso-archean Block. In this paper, we coupled new U–Pb LA-ICP-MS ages to elemental and iron isotope geochemistry, to constrain the depositional age and genesis of Dujiagou BIFs in the Mesoarchean Yangtze Craton of the South China Block, and the contemporaneous marine conditions in which they formed. 2. Regional geology The South China Block is one of the largest cratonic blocks in East Asia and is generally divided into the Yangtze Craton in the northwest and the Cathaysia Block in the southeast, separated by the intervening Jiangnan orogenic belt ((Li et al., 2003; Zhao and Guo, 2012; Zhao and Cawood, 2012; Charvet, 2013; Zhai, 2013; Cawood et al., 2018); Fig. 1a). Located west of the Qinghai Tibet Plateau and south of the Qinling Dabie Sulu orogenic belt, the Yangtze Craton is composed mainly of late Paleoproterozoic and Neoproterozoic rocks with sporadic outcrops of Archean rocks (Zhao and Cawood, 2012; Zhang and Zheng, 2013; Ji et al., 2014; Jiang et al., 2016; Han et al., 2017; Wei et al., 2020). The Huangling dome, located in the northern part of the Yangtze Craton, is the only area where both Archean and Paleoproterozoic metamorphic rocks are exposed in the region (Gao et al., 1999; Qiu et al., 2000; Zheng et al., 2006; Gao et al., 2011; Chen et al., 2013; Han et al., 2018). They provide a natural laboratory for studying the exten-sive tecto-magmatic and metamorphic processes that occurred during the tectonic evolution of the Yangtze Craton. The Kongling terrane, a major Archean relic exposed in the center of the Huangling dome, covers an area of ~360 km2 (Peng et al., 2009; Wei et al., 2019). These Archean rocks were intruded by the late Paleo-proterozoic ~1.85 Ga Quanqitang A-type granites and the Neo-proterozoic ~0.85 Ga Huangling granitoids in the north and south, respectively. The rocks are unconformably overlain by Neoproterozoic and Paleozoic sedimentary sequences that have experienced negligible metamorphic alteration (Li et al., 2003; Peng et al., 2009; Peng et al., 2012a; Peng et al., 2012b; Zhao et al., 2013; Han et al., 2018). Regionally, the Wuduhe Fault divides the the Kongling terrane into the North and South Konling Terranes, abbreviated as NKT and SKT, respectively (Fig. 1b). The NKT is considered metamorphosed to amphibolite facies (Jiang, 1986; Ling et al., 2001; Wei and Jing, 2013; Wei et al., 2020) and to be compose mainly of the Mesoarchean 2.9–3.0 Ga Dongchonghe tonalitic-trondhjemitic-granodioritic (TTG) gneiss, granitic gneiss, and locally preserved amphibolites. The protolith of most of the amphibolites is considered to be 3.2–3.0 Ga ultramafic-mafic rocks—the oldest known lithologies in the Huangling basement, the maintain constituent of the Huangling greenstone belt, HGB (Xiong et al., 2004; Wei et al., 2019; Wei et al., 2020; Wei, 2021). Based on chronological and petrological evidence, the Yemadong Complex is considered a relic of the HGB, named for its location near the Yemadong village (Fig. 1c) (Wei et al., 2019; Zhou et al., 2019; Wei et al., 2020; Wei, 2021). Therefore, the lithological association of the Yemadong Complex inherits the HGB (Zhou et al., 2019). Their original igneous texture is almost obscured and their primary mineral composition replaced by secondary metamorphic minerals (Wei, 2021). In the Huangling dome, the Yemadong Complex is divided into three portions based on lithology and structural position, the lowest part composed of biotite-tremolite and magnetite-bearing plagioclase- amphibole schist, considered to have originated from an ultrabasic-basic volcanic rock protolith. The middle portion is composed of biotite- plagioclase gneiss, feldspathic quartzite, and the banded magnetite quartzite discussed in this paper. The upper section is comprised mainly of leptynite, leptite metamorphic intermediate and acidic volcanic rocks (Zhou et al., 2019). The Dujiagou BIFs occur in the middle and upper part of the greenstone sequence of Yemadong Complex (HGB residue), consistent with the occurrence of most Precambrian BIFs in greenstone belt sequences (e.g., Ghosh and Baidya (2017) and Singh and Slabunov (2015)). In the field, they are exposed as four stratified BIF layers that are generally 20–50 cm thick, dominated by magnetite and quartzite with characteristic quartz and magnetite enrichment (Fig. 2a-c). The magnetite and quartzite layers present a black and white stripe appearance, characterized by interbedded light-colored quartz and dark-colored magnetite stripes (Fig. 2d, e). Magnetite is generally Subhedral or heteromorphic, with a grain size of 0.1–0.2 mm and a Fe H. Zhou et al. Precambrian Research 383 (2022) 1069053content of 45–50%. In the magnetite-rich stripes, the grains are larger and interconnected, forming aggregates and irregular bedding planes, while magnetite is sparsely disseminated among quartz grains in the magnetite-poor stripes to form small, separate grains. Overall, the mineral assemblage in the sample is simple, consisting predominantly of quartz and magnetite with lesser amounts of actinolite and other accessory minerals (Fig. 2f, g). 3. Analytical methods In order to ensure the validity of the data and exclude the influence of weathering conditions on the analysis, exposed weathered surfaces of 13 representative magnetite and quartzite sample were cut and only the fresh portion (Fig. 2e) used for the determination of major, trace and rare earth element contents. 3.1. Zircon U-Pb dating U-Pb dating and trace element analysis of zircon were simulta-neously conducted by LA-ICP-MS at Wuhan Sample Solution Analytical Technology Co., Ltd., Wuhan, China. Detailed operating conditions for the laser ablation system and the ICP-MS instrument and data reduction are the same as described by Zong et al. (2017). Laser sampling was performed using a GeolasPro laser ablation system consisting of a COMPexPro 102 ArF excimer laser with a wavelength of 193 nm, a maximum energy of 200 ml, and a MicroLas optical system. A “wire” signal smoothing device was included with the laser ablation system (Hu et al., 2015). An Agilent 7700e ICP-MS instrument was used to acquire ion-signal intensities. The sample was delivered using the make-up gas argon, mixed with the carrier gas helium via a T-connector before entry into the ICP. The spot size and laser frequency were set to 32 µm and 5 Hz, respectively. Zircon 91,500 and glass NIST610 were used as external standards for U-Pb dating and trace element calibration, respectively. Each analysis incorporated a background acquisition of approximately 20–30 s followed by 50 s of sample data acquisition. An Excel-based software ICPMSDataCal was used to perform off-line selection and integration of background and analyzed signals, time-drift correction and quantitative calibration for trace element analysis and U-Pb dating. Concordia diagrams and weighted mean calculations were made using Isoplot/Ex_ver4.15. 3.2. Bulk-rock major element Bulk-rock major element analysis was conducted with an XRF Primus II, Rigaku (Japan) instrument at Wuhan Sample Solution Analytical Technology Co., Ltd., Wuhan, China. The detailed sample-digestion procedure was as follows: (1) Sample powder (200 mesh) were placed in an oven set at 105 ◦C and dried for 12 h; (2) ~1.0 g dried sample was accurately weighted and placed in ceramic crucibles and heated in a muffle furnace at 1000 ◦C for 2 h. After cooling to 400 ◦C, samples were placed in a drying vessel and weighed again in order to calculate the loss on ignition (LOI); (3) 0.6 g sample powder was mixed with 6.0 g Li2B4O7:LiBO2:LiF = 9:2:1 cosolvent and 0.3 g NH4NO3 oxidant in a Pt crucible and heated in a furnace for l150 ◦C at 14 min. The melted samples were quenched with air for 1 min to produce flat discs on the fire brick for XRF analyses. 3.3. Bulk-rock trace element Bulk-rock trace element analysis was conducted on an Agilent 7700e ICP-MS at Wuhan Sample Solution Analytical Technology Co., Ltd., Wuhan, China. Sample powders were prepared and dried in an oven at Fig. 1. (a) Simplified tectonic map of major Precambrian blocks in South China, (b) Detailed tectonic map of the Kongling area on the northern margin of the Yangtze Craton, (c) Geological map showing the Huangling dome in Yichang City as well as the sampling locations, modified after Wei et al. (2019). H. Zhou et al. Precambrian Research 383 (2022) 1069054105 ◦C as described in section 3.2, after which 50 mg of the dried powder was accurately weighed and placed in Teflon bombs. This was followed by the slow addition of l ml HNO3 and 1 ml HF. The Teflon bombs were then put in a stainless steel pressure jacket and heated to 190 ◦C in an oven for >24 h. After cooling, the Teflon bombs were opened and placed on a hotplate set at 140 ◦C and evaporated, followed by the addition of l ml HNO3 and another round of evaporation to dryness. To the dried sample was added 1 ml HNO3, 1 ml MQ water and l ml 1 ppm internal standard solution. The Teflon bombs were resealed and placed in the oven at 190 ◦C for >12 h. The final solution was transferred to a poly-ethylene bottle and diluted to 100 g with 2% HNO3. 3.4. Bulk-rock Fe isotope Iron isotopic measurements were performed on a Nu Plasma 1700 MC-ICP-MS (Nu Instruments, Wrexham, UK) at the State Key Laboratory of Geological Processes and Mineral Resources, China University of Geosciences, Wuhan. Typically, ~1.0 g sample powders were weighed into individual Teflon beakers and dissolved in a mixture of HF-HNO3- HCl-HClO4. Iron was purified using the AG1X-8 (200–400 mesh chloride form, Bio-Rad, Hercules, CA, USA) resin in HCl media. Matrix elements were removed with 8 ml 6 mol/L HCl, and Fe collected by eluting with 9 ml 0.4 mol/L HCl. The same column procedure was repeated twice to ensure complete elimination of the matrices. The final Fe eluate was acidified with 100 μl of concentrated HNO3. The dried sample was then dissolved in 3% HNO3 for isotopic determination. Double distilled acids and OptimaTM ultrapure HClO4 were used. The whole procedure Fe blank was < 0.01% of the processed samples and therefore considered negligible. Samples were introduced into the Ar plasma using a Cetac ASX-100 auto sampler through a Thermo Scientific Stable Introduction System (SIS), which consists of quartz cyclonic and Scott-type spray chambers and a set of PFA Teflon self-aspirating micro-nebulisers with different uptake rates (45, 50, 100 μl min−1). The instrument mass bias was routinely corrected by calibrator sample bracketing. Data were processed offline and reported in the δ notation (δ xFe (‰) = [(xFe/54-Fe)sample / (xFe/54Fe)calibrator − 1]*1000, where × is 56 or 57 and (xFe/54Fe), considering the average of the two bracketing calibrators). Most tests carried out in this study used GSB Fe (an ultrapure single elemental standard solution from the China Iron and Steel Research Institute) as an in-house reference solution (He et al., 2015). 4. Results 4.1. Zircon U-Pb ages The mostly sub-angular to sub-rounded zircon grains from the metamorphic mafic rock in the Yemadong Complex are 80–150 μm in length, with length/width ratios of 1–1.5 (Fig. 3). Cathodoluminescence images show that the light gray, moderately rounded zircons commonly lack oscillating zonations, and that they possess core morphological characteristics suggestive of metamorphic zircons. The LA-ICP-MS ages for 45 zircon grains with single point analyses are displayed in Table 1. Ten with abnormally high Eu content and four with anomalously high U concentrations were omitted from the age calculation. Thirty-one of the forty-five grains were successfully dated and plotted on the concordia diagram shown in Fig. 3a. Their U zircon concentrations varied from Fig. 2. Representative photographs of the Dujiagou BIF. (a-c) Field photographs showing the outcrop of Dujiagou Algoma-type BIFs interbedded with felsic met-avolcanic rocks, (d) Banded magnetite and quartzite layers composed of bands of dark color magnetite alternating with white color quartz layers. The Ibr-1 and Ibr-2 layers were sampled for iron isotope analysis, (e) An example of a fresh surface cut from the magnetite and quartzite layers in Fig. 2d, (f-g) BIF mainly consisting of alternating magnetite Fe-oxide and quartz bands with other minor minerals (e-Polarized Light; f- Reflected Light), Qz stands for quartz, Mgm for magnetite and Amp for amphibole. H. Zhou et al. Precambrian Research 383 (2022) 1069055323 to 1094 ppm and Th from 38.1 to 140 ppm, with Th/U ratios doubling from 0.09 to 0.18 (Table 1). Seventeen analyses formed a coherent group on the concrodia line (Fig. 3b), yielding a weighted 206Pb/207Pb mean age of 2901 ± 8 Ma (MSWD = 0.77, Fig. 3c). Fourteen points formed a discord stretched between the seventeen previously mentioned points, with an upper and lower 206Pb/238U age intercept of 2909 ± 14 Ma and 775 Ma (MSWD = 1.5), respectively (Fig. 3a). The obtained 1063.6 ± 3.6 Ma concordia age for 20 samples, corresponded with the 207Pb/235U age of 1063.5 ± 7.3 Ma measured for the 91,500 zircon standard, is within analytical uncertainty, and agrees well with the more robust ID-TIMS 207Pb/235U age of 1065.4 ± 0.3 Ma (Jackson et al., 2004). 4.2. Bulk-rock geochemical compositions The silicon-rich and iron-rich layers display characteristics typical of Archean BIF, composed mainly of 39.75–63.67 wt% SiO2 and 28.30–55.37 wt% TFe2O3. Al2O3, MgO, TiO2 and CaO contents range from 0.50 to 1.31 wt%, 1.37–3.35 wt%, 0.03–0.12 wt% and 0.50–1.92 wt%, respectively, suggesting a degree of terrigenous detrital contami-nation (Table 2). The samples contain low K2O, Na2O, MnO and P2O5 concentrations averaging 0.11, 0.10, 0.25, and 0.28 wt%, respectively, being generally compatible with the major element content of Early Precambrian BIF (Table 2, Fig. 4). The samples display a wide range in incompatible element content, with Zr, Hf, Sc, and Th occurring in high concentrations of 3.82–16 ppm, 0.17–0.58 ppm, 1.5–5.32 ppm, and 0.51–3.61, respectively. Their high abundance not encountered in Archean and Paleoproterozoic detritus- free BIFs, support terrigenous contamination (e.g., Li et al. (2012b); Wang et al. (2017), Table 2). Total REE+Y average 94.74 ppm and range from 67.40 to 155.06 ppm (Table 2). Post Archean Average Shale (PAAS) normalized REY patterns are characterized by the depletion of light REE relative to heavy REE, (La/Yb)*= 0.19–0.46, with an average of 0.30. Calculated Eu/Eu* values varying from 0.87 to 1.12, point to positive and negative Eu anomalies across the sedimentary suite, while Ce/Ce* values of 0.85–1.22 suggest Ce anomalies are muted across the sequence. The La/La* values of 1.05–7.35 and averaging 1.85, indicate a sample set characterized by moderate positive La anomalies (Table 2). Thirteen samples are significantly enriched in Y relative to Ho, yielding chondritic or suprachondritic Y/Ho ratios 24.55–31.25, with an average of 27.41. The Dujiagou BIFs show remarkably weaker positive or negative Eu* anomalies and higher ∑REY, which are significantly different from the average of most Archean BIFs (Table 2). 4.3. Fe isotopic composition The δ56FeIRMM-014 of magnetite from seven samples are enriched in heavy Fe isotopes, with a range of 0.25–0.45 ‰ and an average 0.35 ‰. These values are similar to standard δ57FeIRMM-014 values of 0.37–0.68 ‰ that average 0.52 ‰ (Table 3). The δ56FeIRMM014 values of 0.110 ±0.048 ‰ and δ57FeIRMM014 of 0.166 ± 0.080 ‰ (n = 12) for AVG-2 measured here overlap with newly recommended Isoplot values (Lud-wig, 2000) reported by ten independent laboratories (δ56FeIRMM014 =0.110 ± 0.048 ‰, δ57FeIRMM014 = 0.141 ± 0.005 ‰ (He et al., 2015). 5. Discussion 5.1. Metamorphic transformation and age of the Dujiagou BIFs Previous compilations of zircon Th/U ratios led to the proposition of a general Th/U threshold value of 0.1 to differentiate between meta-morphic and igneous zircon (Rubatto, 2017). Based on the new compi-lation of zircon analyses from Western Australia, igneous zircons rarely possess Th/U ratios of < 0.1 while those of metamorphic zircons range between < 0.01 and >10. Therefore, zircons with Th/U ratios < 0.1 are more likely to be metamorphic, and those with Th/U ratios >0 of either igneous or metamorphic origin. However, it has been proposed that a Th/U ratio = 0.4 is more suitable for distinguishing metamorphic rocks from igneous rocks (Yakymchuk et al., 2018). Because Th4+ has a larger ionic radius than U4+, it exhibits a weaker stability than U in the zircon lattice, and is easier to expel from the zircon lattice during metamorphic recrystallization, resulting in a relatively lower Th/U ratio in recrys-tallized metamorphic zircons (Hoskin and Black, 2000). Among the thirty-two zircons successfully dated in this study, six revealed Th/U ≤0.1 and the remaining twenty-six had 0.1 < Th/U ≤ 0.18, which collectively suggests that the Th/U ratios of all the studied zircons fell within the expected metamorphic range. In addition, cath-odoluminescence imaging, such as metamorphic accretion edge (Fig. 3a, b), support the metamorphic genesis of the zircons. Previous studies have shown that the TTG suites in the Kongling terrane were formed at ~3.0–2.90 Ga (Qiu et al., 2000; Zhang et al., 2006; Zheng et al., 2006; Gao et al., 2011; Kang et al., 2013; Qiu et al., 2018; Wei et al., 2019; Wei, 2021). However, the xenoliths in the TTG of the Kongling terrane show an older age of 3.2–3.0 Ga (Table 4). Fig. 3. Zircon U-Pb concordia diagrams and their representative cathodoluminescence images from the Dujiagou BIF samples. Red circles indicate the location of the laser ablation spots. Each spot location is labeled by analysis number. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) H. Zhou et al. Precambrian Research 383 (2022) 1069056Table 1 U-Pb age data of zircons from the metamorphic mafic rocks interbedded in Dujiagou BIFs. Testing No. Concentrations (ppm) Isotopic ratios 207Pb/206Pb 207Pb/235U 206Pb/238U 208Pb/232Th Conments Pb Th U Th/ U 207Pb/206Pb 1σ 207Pb/235U 1σ 206Pb/238U 1σ 208Pb/232Th 1σ Age (Ma) 1σ Age (Ma) 1σ Age (Ma) 1σ Age (Ma) 1σ D5216- 01 1825 64.1 484 0.13 0.2072 0.0035 16.0880 0.2833 0.5637 0.0131 0.1380 0.0041 2884 18 2882 17 2882 54 2613 72 Concordant D5216- 02 2167 140 782 0.18 0.2078 0.0029 16.3513 0.2948 0.5681 0.0116 0.1465 0.0041 2888 15 2898 17 2900 48 2763 73 Concordant D5216- 03 2068 105 790 0.13 0.1871 0.0029 10.5673 0.2509 0.4092 0.0110 0.1154 0.0039 2717 20 2486 22 2211 50 2207 71 Disconcordant D5216- 04 374 82.6 714 0.12 0.1865 0.0032 9.8135 0.3361 0.3781 0.0116 0.1032 0.0035 2712 25 2417 32 2067 54 1985 63 Disconcordant D5216- 05 192 166 1609 0.1 0.1075 0.0030 1.5302 0.0317 0.1033 0.0019 0.0299 0.0006 1757 52 943 13 634 11 596 11 Positive Eu D5216- 06 633 117 993 0.12 0.1972 0.0035 13.1472 0.4356 0.4784 0.0138 0.1295 0.0054 2803 24 2690 31 2520 60 2460 97 Disconcordant D5216- 07 900 73.4 580 0.13 0.2084 0.0039 16.3973 0.3808 0.5698 0.0133 0.1521 0.0059 2893 17 2900 22 2907 55 2861 104 Concordance D5216- 08 638 83.7 798 0.1 0.2081 0.0033 16.0604 0.3027 0.5575 0.0108 0.1558 0.0051 2891 14 2880 18 2856 45 2927 89 Concordance D5216- 09 639 89.6 754 0.12 0.2101 0.0034 16.3341 0.3041 0.5629 0.0117 0.1666 0.0051 2906 15 2897 18 2878 48 3114 88 Concordance D5216- 10 675 90.4 751 0.12 0.2082 0.0028 16.0273 0.2622 0.5556 0.0108 0.1617 0.0044 2891 14 2878 16 2848 45 3029 76 Concordance D5216- 11 531 75.3 634 0.12 0.2099 0.0032 16.3240 0.2665 0.5627 0.0119 0.1604 0.0046 2905 16 2896 16 2878 49 3007 80 Concordance D5216- 12 2162 81 539 0.15 0.2096 0.0033 16.4806 0.2857 0.5686 0.0132 0.1666 0.0047 2903 18 2905 17 2902 54 3115 82 Concordance D5216- 13 984 63.1 529 0.12 0.2117 0.0033 16.7031 0.3893 0.5701 0.0159 0.1666 0.0051 2919 20 2918 22 2908 65 3114 88 Concordance D5216- 14 2173 130 1096 0.12 0.1485 0.0055 3.7124 0.1000 0.1814 0.0046 0.0509 0.0013 2328 65 1574 22 1074 25 1003 25 Positive Eu D5216- 15 505 85.5 732 0.12 0.2001 0.0032 13.3950 0.2264 0.4855 0.0116 0.1320 0.0038 2827 19 2708 16 2551 50 2506 67 Disconcordant D5216- 16 323 60.7 1068 0.06 0.1611 0.0053 6.3583 0.1352 0.2863 0.0071 0.0797 0.0020 2467 57 2027 19 1623 36 1550 38 Positive Eu D5216- 17 474 138 1094 0.13 0.1708 0.0026 7.4467 0.1464 0.3154 0.0078 0.0913 0.0025 2566 19 2167 18 1767 38 1765 45 Disconcordant D5216- 18 1981 195 1660 0.12 0.1218 0.0070 1.3469 0.0627 0.0802 0.0027 0.0230 0.0008 1983 105 866 27 497 16 459 15 Positive Eu D5216- 19 1355 53.4 429 0.12 0.2065 0.0031 16.1155 0.3200 0.5642 0.0141 0.1460 0.0043 2879 19 2884 19 2884 58 2755 77 Concordance D5216- 20 1152 69.8 629 0.11 0.2107 0.0034 16.3426 0.2878 0.5617 0.0132 0.1524 0.0047 2911 18 2897 17 2874 54 2866 82 Concordance D5216- 21 453 92.8 739 0.13 0.1931 0.0037 11.4320 0.3174 0.4276 0.0123 0.1185 0.0035 2768 21 2559 26 2295 55 2264 63 Disconcordant D5216- 22 460 165 1309 0.13 0.1778 0.0029 6.7369 0.3240 0.2717 0.0130 0.0797 0.0042 2632 36 2077 43 1549 66 1549 79 High U D5216- 23 2614.9 41.9 1297 0.03 0.1104 0.0047 1.4999 0.0479 0.0985 0.0028 0.0285 0.0009 1806 80 930 19 606 16 568 18 Positive Eu D5216- 24 2147 125 830 0.15 0.1843 0.0033 8.4064 0.2898 0.3281 0.0119 0.0926 0.0044 2692 27 2276 31 1829 58 1789 81 Disconcordant D5216- 25 2301 73.2 570 0.13 0.1977 0.0032 12.9525 0.2619 0.4732 0.0123 0.1335 0.0048 2807 20 2676 19 2498 54 2533 85 Disconcordant 922 196 1283 0.15 0.1661 0.0039 7.2453 0.1569 0.3034 0.0076 0.0907 0.0031 2518 19 2142 19 1708 37 1755 58 Positive Eu (continued on next page) H. Zhou et al. Precambrian Research 383 (2022) 1069057Table 1 (continued) Testing No. Concentrations (ppm) Isotopic ratios 207Pb/206Pb 207Pb/235U 206Pb/238U 208Pb/232Th Conments Pb Th U Th/ U 207Pb/206Pb 1σ 207Pb/235U 1σ 206Pb/238U 1σ 208Pb/232Th 1σ Age (Ma) 1σ Age (Ma) 1σ Age (Ma) 1σ Age (Ma) 1σ D5216- 26 D5216- 27 1673 67.2 539 0.12 0.2096 0.0069 14.9327 0.3300 0.5167 0.0127 0.1403 0.0035 2902 55 2811 21 2685 54 2654 63 Disconcordant D5216- 28 1765 92 781 0.12 0.1776 0.0032 8.2399 0.3728 0.3299 0.0135 0.0979 0.0043 2630 34 2258 41 1838 66 1888 80 Disconcordant D5216- 29 1977 38.1 323 0.12 0.2079 0.0039 16.4853 0.3260 0.5732 0.0152 0.1687 0.0054 2889 20 2905 19 2921 62 3151 94 Concordance D5216- 30 3160 163 1253 0.13 0.1045 0.0052 1.3829 0.0548 0.0960 0.0029 0.0279 0.0008 1705 94 882 23 591 17 556 17 Positive Eu D5216- 31 467 72.3 528 0.14 0.2101 0.0036 16.3015 0.3614 0.5588 0.0152 0.1649 0.0048 2907 20 2895 21 2862 63 3086 83 Concordance D5216- 32 2653 50.1 444 0.11 0.2079 0.0037 16.2793 0.3365 0.5689 0.0172 0.1623 0.0046 2889 24 2893 20 2903 71 3040 79 Concordance D5216- 33 298 313 1546 0.2 0.1256 0.0039 2.1096 0.0972 0.1151 0.0044 0.0359 0.0016 2037 37 1152 32 702 25 714 32 High U D5216- 34 1125 346 2521 0.14 0.0796 0.0046 0.6444 0.0296 0.0587 0.0021 0.0176 0.0006 1188 118 505 18 368 13 352 12 Positive Eu D5216- 35 2004 219 1649 0.13 0.1155 0.0055 1.3622 0.0500 0.0855 0.0026 0.0246 0.0007 1888 88 873 21 529 15 491 15 Positive Eu D5216- 36 1191 875 1362 0.64 0.1905 0.0034 6.6736 0.3390 0.2492 0.0115 0.0458 0.0024 2746 37 2069 45 1434 59 905 47 Positive Eu D5216- 37 693 70.9 673 0.11 0.2142 0.0032 16.7282 0.2881 0.5650 0.0121 0.1579 0.0051 2937 16 2919 17 2887 50 2964 90 Concordance D5216- 38 275 149 1816 0.08 0.1059 0.0039 1.3545 0.0355 0.0928 0.0024 0.0269 0.0007 1730 70 870 15 572 14 537 14 High U D5216- 39 734 93.8 764 0.12 0.2112 0.0037 14.3764 0.3260 0.4926 0.0128 0.1423 0.0053 2915 19 2775 22 2582 55 2689 94 Disconcordant D5216- 40 318 202 1704 0.12 0.1309 0.0135 1.9136 0.1589 0.1060 0.0065 0.0301 0.0018 2110 188 1086 55 650 38 600 36 High U D5216- 41 618 92.5 991 0.09 0.1835 0.0036 9.9423 0.2736 0.3916 0.0116 0.1125 0.0051 2685 22 2429 25 2130 54 2154 92 Concordance D5216- 42 436 65.4 634 0.1 0.1972 0.0052 10.5984 0.2424 0.3819 0.0103 0.1117 0.0051 2803 20 2489 21 2085 48 2139 93 Concordance D5216- 43 462 44.1 495 0.09 0.2110 0.0038 16.4915 0.3976 0.5712 0.0182 0.1614 0.0056 2913 24 2906 23 2913 75 3025 97 Concordance D5216- 44 789 95.1 700 0.14 0.2112 0.0030 16.7125 0.2901 0.5717 0.0123 0.1533 0.0046 2915 16 2918 17 2915 51 2883 80 Concordance D5216- 45 701 106 1042 0.1 0.1896 0.0050 11.0389 0.3262 0.4051 0.0133 0.1151 0.0043 2739 24 2526 28 2193 61 2201 78 Concordance The obtained concordia age (1063.6 ± 3.6 Ma, n = 20) and 207Pb/235U age (1063.5 ± 7.3 Ma, n = 20) of zircon standard 91,500 agrees well with the preferred ID-TIMS 207Pb/235U age of 1065.4 ± 0.3 Ma (Jackson et al., 2004) within analytical uncertainty. H. Zhou et al. Precambrian Research 383 (2022) 1069058Table 2 Major element (wt. %) and trace element (ppm) compositions of BIF from Dujiagou. Sample 51-1 51-2 61-1 61-2 71-1 71-2 81-1 81-2 2018-1-1 2018-1-2 2018-2 PM040-37-1 D5216-2 Sijiaying Hamersley Itilliarsuk Um Anab Nkout Magnetite quartzite Average SiO2 56.00 54.56 55.95 54.55 56.12 54.39 63.47 63.42 55.90 54.70 63.67 39.75 55.04 55.88 50.35 47.37 54.62 46.93 TiO2 0.09 0.03 0.08 0.03 0.09 0.04 0.12 0.12 0.08 0.03 0.11 0.04 0.09 0.01 0.09 0.18 0.20 0.11 Al2O3 0.89 0.82 0.88 0.83 0.87 0.82 1.28 1.28 0.93 0.87 1.31 0.50 0.88 0.47 2.14 3.78 3.60 3.38 TFe2O3 38.63 40.32 38.67 40.30 38.76 40.20 29.39 29.14 37.53 39.30 28.30 55.37 35.91 41.76 41.71 44.61 35.14 44.54 MnO 0.27 0.24 0.27 0.24 0.27 0.23 0.29 0.29 0.28 0.24 0.30 0.07 0.23 0.05 0.07 0.06 0.07 0.12 MgO 3.33 2.43 3.35 2.41 3.32 2.42 2.03 2.01 3.27 2.33 1.96 1.37 3.15 0.67 1.86 0.95 1.28 2.47 CaO 1.42 0.53 1.43 0.53 1.43 0.53 1.92 1.90 1.39 0.50 1.88 0.58 1.28 0.43 1.79 1.99 2.98 1.59 Na2O 0.15 0.06 0.13 0.05 0.17 0.08 0.14 0.09 0.16 0.05 0.09 0.02 0.11 0.13 0.14 0.57 0.25 0.24 K2O 0.13 0.12 0.12 0.13 0.12 0.12 0.14 0.13 0.16 0.05 0.09 0.03 0.09 0.06 0.18 0.25 0.12 0.96 P2O5 0.24 0.18 0.24 0.18 0.24 0.18 0.45 0.45 0.25 0.18 0.45 0.32 0.25 0.07 0.22 0.23 0.54 0.09 LOI −1.15 0.32 −1.13 0.25 −1.21 0.20 0.73 0.89 −0.96 0.55 1.12 1.08 1.84 0.47 2.22 – 0.53 −0.27 FeO 23.40 12.25 23.80 12.20 23.50 11.60 7.75 7.55 23.80 12.71 8.08 16.22 25.40 9.58 – – – – Sc 3.85 3.05 3.59 3.02 3.67 3.02 5.32 5.31 3.40 2.60 4.80 1.50 2.93 0.85 2.80 5.54 6.20 – V 19.69 10.88 18.83 10.96 19.19 11.41 45.19 44.70 24.00 13.00 54.00 16.50 26.10 12.74 19.78 – 77.80 1.10 Cr 56.28 33.62 53.88 34.57 53.54 34.17 104.13 103.59 70.00 50.00 130.00 24.10 80.30 647.14 203.47 25.04 26.60 6.46 Co 118.76 110.74 111.75 112.65 111.01 112.29 144.17 142.72 115.00 112.50 143.50 2.45 9.89 4.78 4.49 – 95.50 0.25 Ni 36.98 49.25 35.58 50.51 35.15 50.83 102.63 101.90 40.50 53.50 103.50 14.40 42.00 21.74 – 17.36 10.67 1.01 Ga 4.58 2.45 4.44 2.52 4.53 2.52 11.07 10.64 4.70 2.70 11.50 4.46 4.15 1.81 – – – – Rb 3.18 8.59 3.27 8.80 3.15 8.65 7.61 7.49 3.40 9.00 8.20 4.70 4.20 5.68 – 3.09 3.60 8.06 Sr 11.41 5.36 11.16 5.66 11.06 5.43 10.66 10.71 11.90 5.80 11.40 8.80 7.08 29.59 – 85.27 147.80 1.28 Y 25.34 15.28 24.97 15.74 24.73 15.62 39.65 39.44 25.00 15.60 38.60 17.80 19.90 4.78 12.41 3.26 26.60 0.27 Zr 7.31 5.09 7.17 5.25 6.62 5.52 11.70 14.43 7.00 6.00 16.00 3.82 7.37 2.5 32.94 29.63 44.50 0.00 Nb 3.67 2.03 3.52 2.02 3.50 2.04 7.23 7.06 3.70 2.00 7.50 2.88 2.38 0.36 – – 2.38 0.00 Ba 145.24 151.29 149.99 152.32 143.30 152.12 148.88 147.57 151.50 157.50 150.00 40.40 36.50 18.13 75.67 – 64.90 34.69 La 6.37 9.58 6.25 9.73 6.06 9.70 11.98 11.58 6.70 10.10 13.50 7.30 9.75 3.12 8.14 3.24 12.13 0.67 Ce 13.98 26.88 13.80 27.70 13.46 27.16 37.51 37.12 16.80 27.00 39.80 15.90 17.50 4.97 14.54 8.88 27.09 1.21 Pr 2.07 2.77 2.06 2.79 2.02 2.80 5.10 5.07 2.15 2.87 5.38 1.84 2.31 0.65 1.71 0.80 3.78 0.13 Nd 9.37 11.35 9.51 11.33 9.37 11.53 24.20 24.35 9.50 11.60 23.60 10.00 11.90 2.71 6.50 3.23 16.58 0.44 Sm 2.82 3.10 2.66 2.94 2.67 3.09 7.29 7.12 2.71 2.81 6.93 2.59 2.94 0.54 1.23 0.65 3.98 0.06 Eu 0.64 0.63 0.70 0.63 0.67 0.65 1.65 1.59 0.71 0.66 1.58 0.50 0.74 0.16 0.30 0.22 1.05 0.02 Gd 3.18 2.99 3.53 3.08 3.28 2.97 7.41 7.27 3.37 3.12 7.46 2.74 2.99 0.61 1.30 0.29 4.65 0.05 Tb 0.59 0.48 0.62 0.53 0.55 0.51 1.16 1.20 0.56 0.48 1.19 0.49 0.54 0.1 0.21 0.11 0.75 0.00 Dy 3.90 2.99 3.92 3.14 3.93 3.12 7.58 7.49 3.65 2.81 6.94 3.18 3.56 0.63 1.34 0.63 4.67 0.04 Ho 0.86 0.62 0.80 0.61 0.82 0.60 1.44 1.49 0.82 0.58 1.42 0.70 0.79 0.15 0.34 0.12 1.06 0.01 Er 2.36 1.63 2.39 1.68 2.35 1.63 3.91 3.91 2.36 1.59 3.77 1.90 2.22 0.43 1.05 0.35 3.19 0.02 Tm 0.35 0.25 0.36 0.26 0.36 0.24 0.56 0.55 0.36 0.25 0.59 0.30 0.37 0.06 0.16 0.05 0.47 0.00 Yb 2.30 1.68 2.35 1.71 2.34 1.70 3.63 3.51 2.36 1.63 3.71 1.90 2.54 0.42 1.13 0.31 3.19 0.02 Lu 0.36 0.27 0.39 0.27 0.37 0.27 0.57 0.55 0.38 0.27 0.59 0.26 0.36 0.07 0.19 0.05 0.50 0.00 Hf 0.26 0.18 0.33 0.17 0.26 0.19 0.52 0.58 0.20 0.20 0.50 0.18 0.34 0.09 0.85 0.82 0.68 0.00 Ta 0.23 0.14 0.25 0.15 0.25 0.14 0.64 0.62 0.21 0.12 0.54 0.07 0.12 0.05 – – 1.38 – Pb 1.18 8.27 1.33 8.29 1.19 7.91 7.87 7.96 1.10 6.70 7.30 42.60 14.30 3.55 8.91 2.69 7.00 0.18 Th 0.70 0.51 0.61 0.53 0.64 0.53 1.12 1.14 0.60 0.51 1.22 2.92 3.61 0.31 2.26 0.49 1.85 0.11 U 0.48 0.53 0.47 0.54 0.47 0.54 0.81 0.85 0.46 0.49 0.87 0.40 0.37 0.26 0.86 0.32 0.40 0.07 ∑REE þ Y 74.49 80.52 74.31 82.15 72.98 81.59 153.62 152.24 77.43 81.37 155.06 67.40 78.41 19.4 50.55 22.19 109.69 3.12 (La/Yb)PAAS 0.20 0.42 0.20 0.42 0.19 0.42 0.24 0.24 0.21 0.46 0.27 0.28 0.28 0.59 0.53 – 0.29 7.16 La/La* 1.48 1.11 1.67 1.11 1.65 1.14 1.52 1.63 1.35 1.10 1.05 −40.23 7.35 1.34 1.11 1.33 – – Pr/Pr* 1.00 0.90 0.98 0.89 0.99 0.90 0.94 0.93 0.95 0.92 0.98 0.81 0.88 1.04 1.00 – 1.03 0.94 Ce/Ce* 0.87 1.19 0.87 1.22 0.87 1.19 1.06 1.06 1.01 1.15 1.04 1.00 0.85 0.8 0.89 1.63 0.91 0.93 Eu/Eu* 0.99 0.97 1.04 0.97 1.03 1.00 1.04 1.03 1.07 1.03 1.02 0.87 1.16 1.41 2.48 1.63 1.16 1.81 Y/Ho 29.59 24.55 31.25 25.64 30.21 25.90 27.55 26.51 30.49 26.90 27.18 25.43 25.19 31.64 36.50 26.86 25.10 34.04 La/La*=LaPAAS/(3PrPAAS-2NdPAAS); Ce/Ce*=2CePAAS/(LaPAAS + PrPAAS); Eu/Eu*=2EuPAAS/(SmPAAS + GdPAAS); Pr/Pr*=2PrPAAS/(CePAAS + NdPAAS) (Bolhar et al., 2004). BIF data of Sijiaying are from Wang et al. (2017); BIF data of Hamersley are from Warchola et al. (2018); BIF data of Itilliarsuk are from Haug et al. (2013); BIF data of Um Anab are from Basta et al. (2011). BIF data of Nkout West are from Ndime et al. (2019). H. Zhou et al. Precambrian Research 383 (2022) 1069059Fig. 4. Mean values of major elements in the Dujiagou BIFs and early Precambrian BIFs from other areas. BIF data are from Table 2. Table 3 Fe isotope compositions for BIFs representatives from the Dujiagou Formation. Sample no. Object (Whole rock) δ56FeIRMM-014 (‰) 2σ δ57FeIRMM-014 (‰) 2σ δ57/56FeIRMM-014 (‰) 2σ Ibr 1-1a BIF 0.45 0.0252 0.65 0.0411 0.20 0.0349 Ibr 1-1b BIF 0.45 0.0306 0.69 0.0371 0.22 0.0268 Ibr 1-1c BIF 0.45 0.0281 0.65 0.0446 0.22 0.0332 Ibr 2-2a BIF 0.27 0.0254 0.44 0.0372 0.17 0.0262 Ibr 2-2b BIF 0.29 0.0262 0.44 0.0365 0.14 0.0327 Ibr 2a BIF 0.27 0.0276 0.43 0.0390 0.14 0.0274 Ibr 2b BIF 0.25 0.0227 0.37 0.0406 0.11 0.0360 Table 4 Compilation of isotopic ages of xenoliths in TTG suite of Kongling terrane in Huangling region. Sample location Lithology/xenoliths Protolith Sample number Rock/Zircon Age (Ma) Method Yemadong (this study) metavolcanic mafic rocks volcanic mafic rocks 31 (zircon) 2901 ± 8 Zircon U-Pb LA-ICP MS ①Yemadong Miotite–tremolite schist Basic igneous rock 41 (zircon) 3011 ± 27, 2933 ± 13 Zircon U-Pb LA-ICP MS ②Yemadong Massive amphibolite Tholeiitic basalt 15 (zircon) 3000 ± 24 Zircon U-Pb LA-ICP MS ③Wangjiatai Banded amphibolite Tholeiitic basalt 6 (zircon) 3242 ± 140 Zircon U-Pb LA-ICP MS ④Hetaoyuan Banded amphibolite Komatiitic/Tholeiitic basalt 4 (zircon) 3013 ± 23 Zircon U-Pb LA-ICP MS ⑤Hetaoyuan Amphibole schist Komatiitic/Tholeiitic basalt 7 (whole-rock) 2998.9 Sm-Nd isochronIsotope ⑥Dongping Banded amphibolite Tholeiitic basalt 13 (whole-rock) 3290 ± 170 Sm-Nd isochronIsotope The given ages are from ①Wei et al. (2019); ②Wei and Wang (2012); ③Wei et al. (2020); ④-⑤Wei and Jing (2013); ⑥Ma et al. (2002), respectively. H. Zhou et al. Precambrian Research 383 (2022) 10690510Significant tecto-thermal activity between 3.2 and 3.0 Ga in the Yangtze Craton, formed the HGB (Table 4)—a typical Archean greenstone belt (Wei and Wang, 2012; Wei and Jing, 2013; Wei et al., 2020; Wei, 2021). The 3.0–2.9 Ga TTG plutons intruded and split the greenstone belt into separate greenstone remnants or septa, leading to the cropping out of these HGB residues as enclaves of ~10 cm to 20 m thick within the younger TTG gneisses (Gao and Zhang, 1990; Xiong et al., 2004; Li et al., 2014b; Guo et al., 2015; Wei et al., 2020; Wei, 2021). In the Kongling area, >96% of the HGB is invaded and covered by the Mesoarchean TTG gneiss, with a total sporadic exposure area of ~6 km2 (Wei, 2021). In this study, we obtain an average weighted 207Pb/206Pb Mesoarchean age of 2901 ± 8 Ma from the metamorphic zircons recovered from the metamorphosed mafic rocks interbedded with the thin Dujiagou BIFs. The layer thickness of these rocks is uniform, being regularly repetitive in the vertical section (Fig. 2a-c), which may represent intermittent volcanic activity, and the BIFs moments quiescent volcanic activity. Therefore, the interlayered metavolcanic mafic rocks can be considered contemporaneous with the Dujiagou BIFs (Zhang et al., 2012), and the mafic volcanic rocks form the HGB (Wei et al., 2019; Wei et al., 2020; Wei, 2021). In this paper, since all the zircons underwent strong meta-morphism, we interpret the 207Pb/206Pb age of 2901 ± 8 Ma as the metamorphic age related to the same thermal tectonic event at ~2.95–2.90 Ga recorded by the aforementioned mafic volcanic rocks, while the deposition of the Dujiagou BIFs occurred prior to the inferred metamorphic event. Since it is difficult to observe the actual extent and full cross section of the HGB, the sampled outcrop probably represents only a limited exposure, true thickness and size of the Dujiagou BIFs. For example, in the Pitangui greenstone belt, exposed BIF outcrops consist of both thin and extremely thick layers with >10 times thickness difference along the cross section (Brando Soares et al., 2017). Regardless of actual size and thickness, the data suggest that the Dujiagou BIFs are the oldest known deposit in the Yangtze Craton. 5.2. Elemental geochemistry Castic contamination, minor amounts of crustal material are suffi-cient to obscure original marine signatures if not carefully constrained (Bau, 1993; Bolhar et al., 2004). For instance, contamination with terrigenous detritus induces elevated and falsely correlated abundance of incompatible elements (Th, Hf, Zr, Sc, etc.) (Bolhar et al., 2004). BIF samples yielding high abundance of immobile trace elements (for example Sc >2 ppm) should be excluded from a debris-free sample set (Bau, 1993) and high abundances of immobile trace elements such as 1.50–5.32 ppm Sc, 0.51–3.61 ppm Th, 3.82–16.0 ppm Zr, and 0.17–0.58 ppm Hf, confirm detrital incorporation. These values are about an order of magnitude higher than for the debris-free BIF samples (e.g., Li et al. (2012b) and Wang et al. (2017)). Crustal components are rich in Th (Zhao and Zhou, 2007), and therefore an R2 = 0.88 supported positive correlation between Th and Zr (Fig. 5a). This co-variation indicates that the Dujiagou BIFs were contaminated by detrital material. In addition, a strong positive correlation between Al2O3 and TiO2 is often interpreted to reflect the incorporation of detrital components in sedimentary rocks, and therefore widely used as a proxy for detrital contamination of chemical sediments (e.g., Basta et al. (2011).The moderate positive correlation between Al2O3 and TiO2 in the Dujiagou BIFs with r value of 0.76 (Fig. 5b) is therefore consistent with detrital input from the weathering of existing continental landmass. Although fairly low, the 0.54–1.42 wt% Al2O3 + TiO2 content of the Dujiagou BIFs, are largely higher than for pristine examples of Algoma-type BIFs (e.g., Wang et al. (2017)). The discovery of a small amount of amphibole and other accessory minerals in the Dujiagou BIFs provide further strong evidence of terrigenous debris contamination (Fig. 2f). 5.3. Sources of iron and silica The sources of iron in BIFs is debated, with hydrothermal fluids (Basta et al., 2011; Wang et al., 2017) or a mixture of both hydrothermal and terrigenous supply suggested (Pecoits et al., 2009; Li et al., 2015). Also, as a remnant of the HGB, the metamorphic amphibolite xenoliths believed to have originated from a tholeiitic protolith (Wei, 2021), show distinct positive correlation between Al2O3 and TFe2O3 (R2 = 0.84). To the contrary, the strong negative correlation (R2 = 0.91) between Al2O3 and TFe2O3 (Fig. 6a), precludes the sourcing of Fe from the continent or derived by detrital contribution. Generally, diagenesis and various post- depositional processes, including hydrothermal alteration, could strongly affect the mobility of redox-sensitive elements, e.g., when Eu anomalies arise from the reduction of Eu (III) to Eu (II) (Bau, 1993). Previous studies have documented that REEs are largely immobile during low-grade metamorphism, e.g., in amphibolite facies (Finger et al., 1998; Huang et al., 2014), and that the mineral reactions that occur during prograde metamorphism of BIF are essentially isochemical, except for decarbonation and dehydration (Klein, 2005). The amphib-olite metamorphism experienced by the Dujiagou BIF may therefore be insignificant with respect to impact on the migration and enrichment of REEs. The geochemical characteristics of ancient oceans and hydrothermal fluids are often inferred from the compositions of modern seawater and seafloor hydrothermal fluids (Bolhar et al., 2004; Bolhar and Vankra-nendonk, 2007). Typically, when normalized to PAAS, REY patterns for the modern seawater are enriched in HREE relative to LREE. They are characterized by no Eu anomalies and a strong negative Ce anomaly typifies modern oxygenated seawater (Fig. 6b). Further, high- temperature (>250 ◦C) hydrothermal fluids contain pronounced posi-tive Eu anomalies, whereas a weak or no Eu anomaly characterizes low- temperature (<250 ◦C) hydrothermal fluids (Fig. 6b). Therefore, the Eu anomaly has become a mainstream proxy for identifying the material Fig. 5. Discrimination diagrams for major and trace elements in the Dujiagou BIF, Hangling region, Yangtze Craton, (a) Zr vs. Th and (b) Al2O3 vs. TiO2. H. Zhou et al. Precambrian Research 383 (2022) 10690511source of Precambrian BIFs. When normalized to PAAS, the Dujiagou BIFs ΣREYPAAS patterns are different from typical Precambrian debris- free IFs. For example, they are much higher, with several samples hav-ing element/PAAS ratios close to or above 1. Also, they contain extremely faint Eu and positive Ce anomalies, and the curves are much flatter than observed for pristine BIFs (Fig. 6c). The lack of positive Eu anomalies in our samples is inconsistent with most Archean BIFs that tend to have strong positive Eu anomalies. Interestingly, the strong positive correlation (R2 = 0.83) between Al2O3 and ∑REY (Fig. 6d), suggest that the slightly positive Eu anomalies are indicative of the dilution of a high-temperature hydrothermal source of the BIFs by detrital material (Yang et al., 2022). Nonetheless, the positive Eu anomalies observed in the Dujiagou BIFs are consistent with those recorded for world-wide BIF deposits of the same age, attributed to a hydrothermal provenance. It is thus possible that terrigenous contami-nation did not completely erase the primary hydrothermal REE signal, in which case the observed positive Eu anomalies may in fact be pointing to the hydrothermal source of the Dujiagou BIFs. Moreover, the geochemical characteristics of biotite-tremolite schist enclaves from the Yemadong Complex are similar to those of island arc volcanic deposits (Wei et al., 2019), leading to the conclusion that the Dujiagou BIFs and their host volcanic rocks were developed in an island arc setting as opposed to a shelf environment, similar to Algoma-type BIFs. The Dujiagou BIFs are intimately associated with volcanic rocks, therefore, the hydrothermal activity that formed them is considered to be associ-ated with volcanism, suggesting that the insignificant Eu anomalies are unlikely to be of low-temperature hydrothermal origin. Furthermore, BIFs of low-temperature hydrothermal origin appear to be mainly a feature of the Neoproterozoic (e.g., Basta et al. (2011); Yang et al. (2022)). The Y/Ho ratio of BIFs are also widely used to interpret the source of ore-forming materials and sedimentary settings, e.g., Basta et al. (2011); Viehmann et al. (2014); Wang et al. (2017). The Blundy and Wood model (Blundy and Wood, 1994) suggests that in comparison to the degree of fractionation within the REE series, Y-Ho fractionation should be minor to negligible, and even more evolved melts do not readily fractionate Y from Ho (Bau and Dulski, 1996). Although Y and Ho share similar geochemical properties they are fractionated differentially in seawater (Nozaki et al., 1997). In the modern oxidized ocean, where hydrothermal fluids mix with seawater near vent orifices, Y concentra-tions are conservative. On the contrary, REEs from the hydrothermal fluids are partially adsorbed by precipitating Fe-oxyhydroxides when Fe2+ begins to rapidly oxidize to Fe3+ immediately after release from the vents, producing a large decrease in sedimentary Y/Ho ratios relative to seawater (Bau and Dulski, 1999). Chondritic Y/Ho ratios average ~28 (Sun and McDonough, 1989), while seawater is characterized by superchondritic ratios of 44–47 (Bau and Dulski, 1999). Although the Y/ Ho ratio for rivers and estuaries fall between seawater and continental rocks values, experiments have revealed that fractionation in river and estuarine systems is negligible, with little influence on the relative abundance of Y and Ho in the ocean (Nozaki et al., 1997). Y/Ho ratios in the Dujiagou BIFs range from 24.6 to 31.3, with an average of ~27.4 being closer to chondritic values, but far lower than modern seawater ratios. However, small admixtures of any contaminant would depress seawater-like Y/Ho ratios (Bau and Dulski, 1999). The high Zr content in BIFs samples has been confirmed as a source of terrigenous debris input, Fig. 6. Discrimination diagrams for major and trace elements (a) Relationship between Al2O3 and TFe2O3 in the Dujiagou BIF and metamorphic amphibolite xe-noliths (Wei, 2021), Hangling region, Yangtze Craton, (b) South Pacific ocean oxygenated modern seawater, the modern Black Sea anoxic seawater, and Eu anomaly of modern high and low-temperature marine hydrothermal fluids (Bolhar and Vankranendonk (2007) and references therein), (c) PAAS-normalized REE + Y dia-grams of the relationships between REEs and different BIF types. Yuanjiacun Superior-type BIF data from Wang et al. (2014); An-ben Algoma-type BIF data from Li et al. (2012b), and (d) Relationship between Al2O3 and ∑REY in the Dujiagou BIF. H. Zhou et al. Precambrian Research 383 (2022) 10690512and the Y/Ho ratios show a negative correlation with Zr (R2 = 0.18, Fig. 7). Therefore, the real Y/Ho ratio is considered affected by terrig-enous debris, thus presenting the illusion of a relatively low average Y/ Ho ratio. Although terrigenous debris reduces the Y/Ho ratio, four of the BIF samples contained superchondritic ratios, consistent with a hydro-thermal origin. In summary, the provenance of the Dujiagou BIFs is considered to be more closely related to a hydrothermal source rather than terrigenous input. 5.4. Sedimentary environment 5.4.1. Rare Earth elements (REE) In general, anoxic and suboxic waters lack large negative Ce anom-alies when normalized to PAAS, while oxygenated seawater displays a strong negative Ce anomaly (Fig. 6b). In some cases, abnormal La enrichment may cause non-redox-dependent Ce depletion. As a conse-quence, Bau and Dulski (1996) proposed the use of Ce/Ce*vs. Pr/Pr* cross plots to identify La and Ce anomalies in BIFs (Fig. 8a). In Fig. 8a, six BIF samples plot within positive Ce anomalies, which is generally not considered a predominant feature of chemical sediments precipitated from Archean seawater (Bolhar et al., 2004). The most likely cause is debris contamination, just as the Eu anomalies in the same batch of samples suggest. We are therefore inclined to consider the Ce anomalies as potentially unreliable tracers of the redox conditions in which they formed. Nonetheless, the lack of negative and true Ce anomalies is consistent with the expected anoxic waters of the Archean Ocean. 5.4.2. Iron isotopes 5.4.2.1. Iron isotopes and metamorphic influences. Given the fact that the Dujiagou BIFs have been metamorphosed, as shown by different lines of evidence, to amphibolite grade metamorphism, it is necessary to eval-uate the influence of metamorphism on the studied iron isotopes before they are used for geochemical interpretative purposes. Studies on the Paleoproterozoic Biwabik iron formation, Hyslop et al. (2008) showed that metamorphism can alter the primary isotopic composition of O atoms in quartz and magnetite and iron in magnetite and iron silicate minerals. The self-diffusion coefficient of iron in magnetite at a meta-morphic temperature of 500 ◦C depends on the activity of O2, but is greater than ~10−16cm2/s (Atkinson et al., 1983), with a corresponding diffusion length of ~1 mm (Dauphas et al., 2007a). In this case, ~ 3 Ma is required for complete diffusion of iron isotopes over an average length of 2 mm (Czaja et al., 2013). However, the diffusion of iron isotopes can only occur in individual iron layers, because magnetite bands are separated by quartz bands, which would be an effective barrier to iron diffusion. In the hand samples, it seems unlikely that large-scale transport of iron occurred during metamorphism and the overall average δ56FeMt values are likely to be reflective of those of the original BIF sediment (Czaja et al., 2013). The independent study of the ~1.8 Ga Biwabik BIFs shows that metamorphism could slightly reduce the δ56FeMt in BIFs, but that metamorphism cannot explain the positive δ56FeMt values (Hyslop et al., 2008). Consequently, Fe isotopic compo-sition of magnetite is much less affected by metamorphism compared to other mineral phases (Dauphas et al., 2007a; Hyslop et al., 2008). Moreover, it is has been proposed that the iron isotopic composition of BIFs are not involved in amphibolite facies metamorphism after Fig. 7. Plots of a Zr vs. Y/Ho in the Dujiagou BIF, Hangling region, Yang-tze Craton. Fig. 8. (a) Ce/Ce* vs. (Pr/Pr*) plot according to Bau and Dulski (1996), (b) Scatter plot of Fe isotopic composition of igneous rocks, clastic sedimentary rocks and chemically precipitated sedimentary rocks through Earth history. Igneous rock data from Beard et al. (2003b), other data from (Beard et al., 2003a), (c) whole-rock δ56Fe values for iron formations from this study and literature (Dauphas et al., 2007b; Johnson et al., 2008; Steinhoefel et al., 2009; Li et al., 2012b; Planavsky et al., 2012; Ye et al., 2017), and δ56Fe values for sedimentary pyrites from Rouxel et al. (2005) and references therein. H. Zhou et al. Precambrian Research 383 (2022) 10690513formation (Dauphas and Rouxel, 2006; Frost et al., 2006; Dauphas et al., 2007a; Czaja et al., 2013). We therefore assume that the iron isotopic composition of the Dujiagou BIFs is likely representative of the original primary composition. 5.4.2.2. Iron isotopes and sedimentary environment. Recently, iron iso-topes were proposed as a redox indicator of ancient oceans, resulting from the enrichment of heavy iron isotopes in Fe(III) minerals during the chemical oxidation of dissolved Fe (II) to Fe (III) precipitates (Rouxel et al., 2005; Anbar and Rouxel, 2007). The enrichment of the heavy iron isotopes in the poorly mobile Fe (III)-rich minerals results in the transfer of the light isotopes to the highly mobile Fe (II) phase dissolved in anoxic waters (Dauphas and Rouxel, 2006). The seven whole rock δ56Fe data from the Ibr 1 and Ibr 2 horizons (Fig. 2d) are divided into ~0.45 ‰ for sample set Ibr 1(n = 3) and ~0.25–0.27 ‰ for Ibr 2 (n = 4). A study highlighted that the residual iron isotopic composition in magnetite and seawater gradually becomes lighter as the proportion of precipitate increases, implying early magnetite acquires heavier iron isotopes than late precipitates (Li et al., 2012b). Therefore, the variation of iron isotope composition of the Dujiagou BIFs can be explained by different precipitation stages. However, magnetite as a major BIF min-eral, is not an authigenic phase but was probably formed diagenetically from ferro-ferric hydroxide precursors (Ewers, 1983; Klein, 2005; Dau-phas et al., 2007a). For example, green rust was considered to be an important magnetite precursor in the Archean ocean (Sumoondur et al., 2008; Zegeye et al., 2012; Sun et al., 2022). Conversion to magnetite with inherited primary δ56Fe values of precursor ferric oxide/hydroxide minerals, probably occur through interaction with hydrothermal Fe2+ in the deep marine environments of BIF deposition (Johnson et al., 2008). Hence, the iron isotopes in the BIF indicate changes in the ocean iron cycle as well as changes in atmospheric redox, preserving a chemical record of Earth’s most primitive oceans (Rouxel et al., 2005; Planavsky et al., 2012). The heavy iron isotope compositions demonstrate the opposite behavior expected for modern weathered products and hydrothermal fluids (Fig. 8b). They indicate that the valence state of iron changed from divalent to trivalent state during the deposition of magnetite. Precambrian BIFs generally exhibit a wide range of δ56Fe values, span-ning −1.6 to +1.5 ‰ (Fig. 8c), with >0 values consistent with the partial oxidation of iron, while the <0 values are closer to those produced by microbial dissimilatory Fe(III) reduction, DIR (Lovley et al., 2004). DIR can generate large iron isotope fractionation by reducing trivalent Fe (III) to highly soluble divalent Fe(II) (Beard et al., 1999; Crosby et al., 2007). Previous studies have shown that DIR may have evolved by 3.8 Ga (Craddock and Dauphas, 2011), but the broad absence of negative δ56Fe values in ＞2.5 Ga BIFs, has been taken to indicate that DIR may not have left a strong impact on the Archean iron sedimentary record (Johnson et al., 2008; Czaja et al., 2013). The variable negative δ56Fe values in the Hamersley BIF in Western Australia may thus represent the earliest record of expansive DIR expression in Earth history (Fig. 8c) (Johnson et al., 2008). Therefore, the consistency of the data with numerous observations for various Archean BIFs, indicate that δ56Fe＞ 0 is a common characteristic of Archean IFs, regardless of paleo-geographyin (Fig. 8c). This may be a reflection of the unique nature of biogeochemical cycling of iron during the Archean (Hyslop et al., 2008). Archean black shale pyrites are enriched with light iron isotopes, with the most negative iron isotope values being as low as −3.5‰ (Fig. 8c). Because BIFs are a sink for isotopically heavy iron, black shale pyrite might therefore represent a corresponding reservoir for isotopically light iron in the sedimentary record (Planavsky et al., 2012). Briefly, divalent Fe(II) is oxidized to trivalent Fe(III), enriching the heavy iron isotopes in the trivalent Fe(III) to form iron oxide-hydroxide. The residual dissolved divalent Fe(II) with its light iron isotopes would have been captured by the pyrite and enriched in pyrite-rich Archean sedimentary rocks. The rich iron oxide content of BIFs indicates the presence of significant Fe(II) oxidants in the Archean upper water column, much of which is debated, considering that the oceans are believed to have been strongly reducing for most of the Archean Eon (Bekker et al., 2010). The three processes most commonly advocated for iron oxidation are; (1) oxidation with O2 generated by oxygenic photosynthesis, (2) Oxidation by anoxic photo-synthesis using Fe (II) as an electron donor rather than H2O, and (3) Photooxidation (Dauphas et al., 2007a). Experimental results show that the enrichment of heavy isotopes of iron in Fe(III) minerals occurs during biotic and abiotic oxidation of Fe (II) to Fe(III) under varying redox conditions (Bullen et al., 2001; Johnson et al., 2002; Welch et al., 2003; Balci et al., 2006), while the possibility of ultraviolet photooxi-dation in natural seawater compositions may be less likely than previ-ously thought (Konhauser et al., 2007). The Great Oxidation Event (GOE) saw the permanent rise of oxygen in the atmosphere, but there is increasing evidence that free oxygen may have predated the GOE (Brocks et al., 1999; Anbar et al., 2007; Godfrey and Falkowski, 2009; Ostrander et al., 2021). We therefore interpret the prevalence of positive δ56Fe values in the Dujiagou BIFs as representative of either abiotic and biotic partial oxidation of Fe(II) to Fe(III) minerals, or a combination of both processes. Thus, Archean marine shales and BIFs record iron iso-topic composition different from those of modern marine sediments with typical δ56Fe values of ~0‰ (Fig. 8b). Collectively, these observations indicate that the Archean oceans were vastly anoxic/hypoxic before the rise of atmospheric oxygen, which as explained above would have influenced the rates and quantitative abiotic and biotic oxidation of mobile Fe(II) in the Archean seawater column to particulate Fe oxide- oxyhydroxides, and therefore exerting a strong control on the δ56Fe values of particulate and dissolved iron. 6. Conclusions The Dujiagou BIFs formed in an oceanic setting, probably near a subduction zone, similar to the suggested location where the Algoma- type BIFs that are generally prevalent in the Archean sedimentary rock record, deposited. With a predicted depositional age estimated at >2.90 Ga, the Dujiagou BIFs are likely the oldest BIFs in the Yangtze Craton. Their geochemical composition indicate that they are chemical sediments, marked by a minor detrital component that diluted and masked their primary REE composition. The data suggest that iron was sourced from submarine hydrothermal fluids, consistent with the origin of iron in most Precambrian BIFs. Moreover, their positive iron isotopic composition is characteristic of primary Archean BIF deposits, indi-cating that partial oxidation of iron led to the deposition of the divalent magnetite-rich BIFs in an Archean Ocean that was broadly reducing. Credit authorship contribution statement Hang Zhou: Writing-Original draft preparation, Data curation. Wenxiao Zhou: Supervision, Formal analysis, Data curation, Reviewing and Editing, Funding acquisition. Yunxu Wei: Conceptualization, Data curation, Funding acquisition. Ernest Chi Fru: Reviewing and Editing. Haiquan Li: Data curation, Reviewing and Editing. Bo Huang: Data curation, Reviewing and Editing. Dong Fu: Reviewing and Editing. Mantang Tan: Visualization, Funding acquisition. Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Data availability Data will be made available on request. H. Zhou et al. Precambrian Research 383 (2022) 10690514Acknowledgments We thank the editor Prof. Frances Westall and two anonymous re-viewers for constructive comments, which significantly improved the quality of the manuscript. We also appreciate Prof. Ming Li for help with iron isotopic composition testing. This study was supported by the Na-tional Natural Science Foundation of China (grant numbers: 41703024), the China Geological Survey (grant numbers: DD20190374 and DD20160029-02), and the Fundamental Research Funds for National Universities from the China University of Geosciences, Wuhan (grant number: 2020098). References Anbar, A.D., Duan, Y., Lyons, T.W., Arnold, G.L., Kendall, B., Creaser, R.A., Kaufman, A. J., Gordon, G.W., Scott, C., Garvin, J., Buick, R., 2007. A whiff of oxygen before the Great Oxidation Event? Science 317 (5846), 1903–1906. Anbar, A.D., Rouxel, O., 2007. Metal Stable Isotopes in Paleoceanography. Annu. Rev. Earth Planet. Sci. 35 (1), 717–746. Atkinson, A., O’Dwyer, M.L., Taylor, R.I., 1983. 55Fe Diffusion in magnetite crystals at 500◦ and its relevance to the oxidation of fe. Radiation Effects 75 (1–4), 169–172. Balci, N., Bullen, T.D., Witte-Lien, K., Shanks, W.C., Motelica, M., Mandernack, K.W., 2006. Iron isotope fractionation during microbially stimulated Fe(II) oxidation and Fe(III) precipitation. Geochim. Cosmochim. Acta 70 (3), 622–639. Basta, F.F., Maurice, A.E., Fontbot´e, L., Favarger, P.Y., 2011. Petrology and geochemistry of the banded iron formation (BIF) of Wadi Karim and Um Anab, Eastern Desert, Egypt: Implications for the origin of Neoproterozoic BIF. Precambr. Res. 187 (3–4), 277–292. Bau, M., 1993. Effects of syn- and post-depositional processes on the rare-earth element distribution in Precambrian iron-formations. Eur. J. Mineral. 5 (2), 257–267. Bau, M., Dulski, P., 1996. Distribution of yttrium and rare-earth elements in the Penge and Kuruman iron-formations, Transvaal Supergroup, South Africa. Precambr. Res. 79 (1–2). Bau, M., Dulski, P., 1999. Comparing yttrium and rare earths in hydrothermal fluids from the Mid-Atlantic Ridge: implications for Y and REE behaviour during near-vent mixing and for the Y/Ho ratio of Proterozoic seawater. Chem. Geol. 155 (1), 77–90. Beard, B.L., Johnson, C.M., Cox, L., 1999. Iron Isotope Biosignatures. Science 285 (5435), 1889–1892. Beard, B.L., Johnson, C.M., Damm, K.V., Poulson, R.L., 2003a. Iron isotope constraints on Fe cycling and mass balance in oxygenated Earth oceans. Geology 31 (7), 629–632. Beard, B.L., Johnson, C.M., Skulan, J.L., Nealson, K.H., Cox, L., Sun, H., 2003b. Application of Fe isotopes to tracing the geochemical and biological cycling of Fe. Chem. Geol. 195 (1), 87–117. Bekker, A., Slack, J.F., Planavsky, N., Krapez, B., Hofmann, A., Konhauser, K.O., Rouxel, O.J., 2010. Iron Formation: The Sedimentary Product of a Complex Interplay among Mantle, Tectonic, Oceanic, and Biospheric Processes. Econ. Geol. 105 (3), 467–508. Blundy, J., Wood, B., 1994. Prediction of crystal-melt partition coefficients from elastic moduli. Nature 372 (6505), 452–454. Bolhar, R., Kamber, B.S., Moorbath, S., Fedo, C.M., Whitehouse, M.J., 2004. Characterisation of early Archaean chemical sediments by trace element signatures. Earth Planet. Sci. Lett. 222 (1), 43–60. Bolhar, R., Vankranendonk, M., 2007. A non-marine depositional setting for the northern Fortescue Group, Pilbara Craton, inferred from trace element geochemistry of stromatolitic carbonates. Precambr. Res. 155 (3–4), 229–250. Brando Soares, M., Corrˆea Neto, A.V., Zeh, A., Cabral, A.R., Pereira, L.F., Prado, M.G.B. d., Almeida, A.M.d., Manduca, L.G., Silva, P.H.M.d., Mabub, R.O.d.A., Schlichta, T. M., 2017. Geology of the Pitangui greenstone belt, Minas Gerais, Brazil: Stratigraphy, geochronology and BIF geochemistry. Precambr. Res. 291, 17–41. Brocks, J., Logan, G., Buick, R., Summons, R., 1999. Archean molecular fossils and the early rise of eukaryotes. Science 285, 1033–1036. Bullen, T.D., White, A.F., Childs, C.W., Vivit, D.V., Schulz, M.S., 2001. Demonstration of significant abiotic iron isotope fractionation in nature. Geology 29 (8), 699–702. Cawood, P.A., Zhao, G., Yao, J., Wang, W., Xu, Y., Wang, Y., 2018. Reconstructing South China in Phanerozoic and Precambrian supercontinents. Earth Sci. Rev. 186, 173–194. Charvet, J., 2013. The Neoproterozoic-Early Paleozoic tectonic evolution of the South China Block: An overview. J. Asian Earth Sci. 74, 198–209. Chen, K., Gao, S., Wu, Y., Guo, J., Hu, Z., Liu, Y., Zong, K., Liang, Z., Geng, X., 2013. 2.6–2.7Ga crustal growth in Yangtze craton, South China. Precambr. Res. 224, 472–490. Chi Fru, E., Ivarsson, M., Kilias, S.P., Bengtson, S., Belivanova, V., Marone, F., Fortin, D., Broman, C., Stampanoni, M., 2013. Fossilized iron bacteria reveal a pathway to the biological origin of banded iron formation. Nat. Commun. 4, 2050. Chi Fru, E., Kilias, S., Ivarsson, M., Rattray, J., Gkika, K., McDonald, I., He, Q., Broman, C., 2018. Sedimentary mechanisms of a modern banded iron formation on Milos Island, Greece-Supplement. Solid Earth 9, 573–598. Chi Fru, E., Rodriguez, N.P., Partin, C.A., Lalonde, S.V., Andersson, P., Weiss, D.J., El Albani, A., Rodushkin, I., Konhauser, K.O., 2016. Cu isotopes in marine black shales record the Great Oxidation Event. Proc. Natl. Acad. Sci. U.S.A. 113 (18), 4941–4946. Craddock, P.R., Dauphas, N., 2011. Iron and carbon isotope evidence for microbial iron respiration throughout the Archean. Earth Planet. Sci. Lett. 303 (1), 121–132. Crosby, H.A., Roden, E.E., Johnson, C.M., Beard, B.L., 2007. The mechanisms of iron isotope fractionation produced during dissimilatory Fe(III) reduction by Shewanella putrefaciens and Geobacter sulfurreducens. Geobiology 5 (2), 169–189. Crowley, J.L., Myers, J.S., Dunning, G.R., 2002. Timing and nature of multiple 3700–3600 Ma tectonic events in intrusive rocks north of the Isua greenstone belt, southern West Greenland. Geol. Soc. Am. Bull. 114 (10), 1311–1325. Cui, M.L., Zhang, L.C., Wu, H.Y., Xu, Y.X., Li, W.J., 2014. Timing and tectonic setting of the Sijiaying banded iron deposit in the eastern Hebei province, North China Craton: Constraints from geochemistry and SIMS zircon U-Pb dating. J. Asian Earth Sci. 94, 240–251. Czaja, A.D., Johnson, C.M., Beard, B.L., Roden, E.E., Li, W., Moorbath, S., 2013. Biological Fe oxidation controlled deposition of banded iron formation in the ca. 3770Ma Isua Supracrustal Belt (West Greenland). Earth Planet. Sci. Lett. 363, 192–203. Dauphas, N., Cates, N.L., Mojzsis, S.J., Busigny, V., 2007a. Identification of chemical sedimentary protoliths using iron isotopes in the >3750 Ma Nuvvuagittuq supracrustal belt. Canada. Earth and Planetary Science Letters 254 (3–4), 358–376. Dauphas, N., Rouxel, O., 2006. Mass spectrometry and natural variations of iron isotopes. Mass Spectrom. Rev. 25 (4), 515–550. Dauphas, N., van Zuilen, M., Wadhwa, M., Davis, A.M., Marty, B., Janney, P.E., 2004. Clues from Fe isotope variations on the origin of early Archean BIFs from Greenland. Science 306 (5704), 2077–2080. Dauphas, N., Zuilen, M.V., Busigny, V., Lepland, A., Wadhwa, M., Janney, P.E., 2007b. Iron isotope, major and trace element characterization of early Archean supracrustal rocks from SW Greenland: Protolith identification and metamorphic overprint. Geochim. Cosmochim. Acta 71 (19), 4745–4770. Ewers, W.E., 1983. Chapter 13 Chemical Factors in the Deposition and Diagenesis of Banded Iron-Formation. Developments in Precambrian Geology 6, 491–512. Farquhar, J., Bao, H.M., Thiemens, M., 2000. Atmospheric influence of Earth’s earliest sulfur cycle. Science 289 (5480), 756–758. Finger, F., Broska, I., Roberts, M.P., Schermaier, A., 1998. Replacement of primary monazite by apatite-allanite-epidote coronas in an amphibolite facies granite gneiss from the eastern Alps. Am. Mineral. 83 (3–4), 248–258. Frost, C.D., von Blanckenburg, F., Schoenberg, R., Frost, B.R., Swapp, S.M., 2006. Preservation of Fe isotope heterogeneities during diagenesis and metamorphism of banded iron formation. Contrib. Miner. Petrol. 153 (2), 211. Gao, S., Ling, W., Qiu, Y., Lian, Z., Hartmann, G., Simon, K., 1999. Contrasting geochemical and Sm-Nd isotopic compositions of Archean metasediments from the Kongling high-grade terrain of the Yangtze craton: Evidence for cratonic evolution and redistribution of REE during crustal anatexis. Geochim. Cosmochim. Acta 63 (13), 2071–2088. Gao, S., Yang, J., Zhou, L., Li, M., Hu, Z., Guo, J., Yuan, H., Gong, H., Xiao, G., Wei, J., 2011. Age and growth of the Archean Kongling terrain, South China, with emphasis on 3.3 Ga granitoid gneisses. Am. J. Sci. 311 (2), 153–182. Gao, S., Zhang, B.R., 1990. The discovery of Archean TTG gneisses in northern Yangtze craton and their implications. Earth Sci. 15, 675–679. Ghosh, R., Baidya, T.K., 2017. Mesoarchean BIF and iron ores of the Badampahar greenstone belt, Iron Ore Group, East Indian Shield. J. Asian Earth Sci. 150, 25–44. Godfrey, L.V., Falkowski, P.G., 2009. The cycling and redox state of nitrogen in the Archaean ocean. Nat. Geosci. 2 (10), 725–729. Goodwin, A., 1973. Archean Iron-Formations and Tectonic basins of the Canadian Shield. Econ. Geol. 68, 915–933. Guo, Q.J., Strauss, H., Kaufman, A.J., Schroder, S., Gutzmer, J., Wing, B., Baker, M.A., Bekker, A., Jin, Q.S., Kim, S.T., Farquhar, J., 2009. Reconstructing Earth’s surface oxidation across the Archean-Proterozoic transition. Geology 37 (5), 399–402. Guo, J.L., Wu, Y.B., Gao, S., Jin, Z.M., Zong, K.Q., Hu, Z.C., Chen, K., Chen, H.H., Liu, Y. S., 2015. Episodic Paleoarchean-Paleoproterozoic (3.3–2.0Ga) granitoid magmatism in Yangtze Craton, South China: Implications for late Archean tectonics. Precambr. Res. 270, 246–266. Hagemann, S.G., Angerer, T., Duuring, P., Rosi`ere, C.A., Figueiredo e Silva, R.C., Lobato, L., Hensler, A.S., Walde, D.H.G., 2016. BIF-hosted iron mineral system: A review. Ore Geol. Rev. 76, 317–359. Han, Q.S., Peng, S.B., Kusky, T., Polat, A., Jiang, X.F., Cen, Y., Liu, S.F., Deng, H., 2017. A Paleoproterozoic ophiolitic melange, Yangtze craton, South China: Evidence for Paleoproterozoic suturing and microcontinent amalgamation. Precambr. Res. 293, 13–38. Han, Q.S., Peng, S.B., Polat, A., Kusky, T., Deng, H., Wu, T.Y., 2018. A ca.2.1 Ga Andean- type margin built on metasomatized lithosphere in the northern Yangtze craton, China: Evidence from high-Mg basalts and andesites. Precambr. Res. 309, 309–324. Haug, A.R.R., Frei, R., Stendal, H., Konhauser, K., 2013. Petrology and geochemistry of the 2.9 Ga Itilliarsuk banded iron formation and associated supracrustal rocks, West Greenland: Source characteristics and depositional environment. Precambr. Res. 229, 150–176. He, Y.S., Ke, S., Teng, F.Z., Wang, T.T., Wu, H.J., Lu, Y.H., Li, S.G., 2015. High-Precision Iron Isotope Analysis of Geological Reference Materials by High-Resolution MC-ICP- MS. Geostand. Geoanal. Res. 39 (3), 341–356. Hoffman, P.F., Kaufman, A.J., Halverson, G.P., Schrag, D.P., 1998. A Neoproterozoic snowball earth. Science 281 (5381), 1342–1346. Holland, H.D., 2006. The oxygenation of the atmosphere and oceans. Philos. Trans. Royal Society B-Biol. Sci. 361 (1470), 903–915. Hoskin, P.W.O., Black, L.P., 2000. Metamorphic zircon formation by solid-state recrystallization of protolith igneous zircon. J. Metamorph. Geol. 18 (4), 423–439. Hu, Z., Zhang, W., Liu, Y., Gao, S., Li, M., Zong, K., Chen, H., Hu, S., 2015. “Wave” Signal-Smoothing and Mercury-Removing Device for Laser Ablation Quadrupole and Multiple Collector ICPMS Analysis: Application to Lead Isotope Analysis. Anal. Chem. 87 (2), 1152–1157. H. Zhou et al. Precambrian Research 383 (2022) 10690515Huang, H., Fryer, B.J., Polat, A., Pan, Y., 2014. Amphibole, plagioclase and clinopyroxene geochemistry of the Archean Fiskenæsset Complex at Majorqap qˆava, southwestern Greenland: Implications for Archean petrogenetic and geodynamic processes. Precambr. Res. 247, 64–91. Huang, B., Kusky, T.M., Wang, L., Deng, H., Wang, J.P., Fu, D., Peng, H.T., Ning, W.B., 2019. Age and genesis of the Neoarchean Algoma-type banded iron formations from the Dengfeng greenstone belt, southern North China Craton: Geochronological, geochemical and Sm-Nd isotopic constraints. Precambr. Res. 333. Hyslop, E.V., Valley, J.W., Johnson, C.M., Beard, B.L., 2008. The effects of metamorphism on O and Fe isotope compositions in the Biwabik Iron Formation, northern Minnesota. Contrib. Miner. Petrol. 155 (3), 313–328. Jackson, S.E., Pearson, N.J., Griffin, W.L., Belousova, E.A., 2004. The application of laser ablation-inductively coupled plasma-mass spectrometry to in situ U-Pb zircon geochronology. Chem. Geol. 211 (1), 47–69. Jenner, F.E., Bennett, V.C., Nutman, A.P., Friend, C.R.L., Norman, M.D., Yaxley, G., 2009. Evidence for subduction at 3.8 Ga: Geochemistry of arc-like metabasalts from the southern edge of the Isua Supracrustal Belt. Chem. Geol. 261 (1–2), 82–97. Ji, W.B., Lin, W., Faure, M., Chu, Y., Wu, L., Wang, F., Wang, J., Wang, Q.C., 2014. Origin and tectonic significance of the Huangling massif within the Yangtze craton, South China. J. Asian Earth Sci. 86, 59–75. Jiang, J., 1986. Isotopic geochronology and crustal evolution of the Huangling metamorphic terrain. Journal of Jiling University (Earth Science Edition) 3, 1–11. Jiang, X.F., Peng, S.B., Polat, A., Kusky, T., Wang, L., Wu, T.Y., Lin, M.S., Han, Q.S., 2016. Geochemistry and geochronology of mylonitic metasedimentary rocks associated with the Proterozoic Miaowan Ophiolite Complex, Yangtze craton, China: Implications for geodynamic events. Precambr. Res. 279, 37–56. Johnson, C.M., Beard, B.L., Beukes, N.J., Klein, C., O’Leary, J.M., 2003. Ancient geochemical cycling in the Earth as inferred from Fe isotope studies of banded iron formations from the Transvaal Craton. Contrib. Miner. Petrol. 144 (5), 523–547. Johnson, C.M., Beard, B.L., Klein, C., Beukes, N.J., Roden, E.E., 2008. Iron isotopes constrain biologic and abiologic processes in banded iron formation genesis. Geochim. Cosmochim. Acta 72 (1), 151–169. Johnson, C.M., Skulan, J.L., Beard, B.L., Sun, H., Nealson, K.H., Braterman, P.S., 2002. Isotopic fractionation between Fe(III) and Fe(II) in aqueous solutions. Earth Planet. Sci. Lett. 195 (1–2), 141–153. Kang, C., Shan, G., Wu, Y., Guo, J., Hu, Z., Liu, Y., Zong, K., Liang, Z., Geng, X., 2013. 2.6–2.7 Ga crustal growth in Yangtze craton, South China. Precambr. Res. 224, 472–490. Kasting, J.F., 2013. What caused the rise of atmospheric O2? Chem. Geol. 362 (362), 13–25. Klein, C., 2005. Some Precambrian banded iron-formations (BIFs) from around the world: Their age, geologic setting, mineralogy, metamorphism, geochemistry, and origins. Am. Mineral. 90 (10), 1473–1499. Li, H.M., Zhang, Z.J., Li, L.X., Zhang, Z.C., Chen, J., Yao, T., 2014a. Types and general characteristics of the BIF-related iron deposits in China. Ore Geol. Rev. 57, 264–287. Li, X.H., Li, Z.X., Ge, W., Zhou, H., Li, W., Liu, Y., Wingate, M.T., 2003. Neoproterozoic granitoids in South China: crustal melting above a mantle plume at ca. 825 Ma? Precambr. Res. 122, 45–83. Li, L.M., Lin, S.F., Davis, D.W., Xiao, W.J., Xing, G.F., Yin, C.Q., 2014b. Geochronology and geochemistry of igneous rocks from the Kongling terrane: Implications for Mesoarchean to Paleoproterozoic crustal evolution of the Yangtze Block. Precambr. Res. 255, 30–47. Konhauser, K.O., Amskold, L., Lalonde, S.V., Posth, N.R., Kappler, A., Anbar, A.D., 2007. Decoupling photochemical Fe(II) oxidation from shallow-water BIF deposition. Earth Planet. Sci. Lett. 258 (1–2), 87–100. Li, W., Beard, B.L., Johnson, C.M., 2015. Biologically recycled continental iron is a major component in banded iron formations. Proc. Natl. Acad. Sci. U.S.A. 112 (27), 8193–8198. Li, Y.H., Hou, K.J., Wan, D.F., Zhang, Z.J., 2012a. A compare geochemistry study for Algoma- and Superior-type banded iron formations. Acta Petrologica Sinica 28 (11), 3513–3519. Li, Z.H., Zhu, X.K., Tang, S.H., 2012b. Fe isotope compositions of banded iron formation from Anshan-Benxi area: Constraints on the formation mechanism and Archean ocean environment. Acta Petrologica Sinica 28 (11), 3545–3558. Ling, W.L., Gao, S., Zhang, B.R., Zhou, L., Xu, D.L., 2001. The recognizing of ca. 1.95 Ga tectono-thermal eventin Kongling nucleus and its significance for the evolution of Yangtze Block, South China. Chinese Science Bulletin 46, 326–329. Zong, K.Q., Klemd, R., Yuan, Y., He, Z.Y., Guo, J.L., Shi, X.L., Liu, Y.S., Hu, Z.C., Zhang, Z.M., 2017. The assembly of Rodinia: The correlation of early Neoproterozoic (ca. 900 Ma) high-grade metamorphism and continental arc formation in the southern Beishan Orogen, southern Central Asian Orogenic Belt (CAOB). Precambr. Res. 290, 32–48. Lovley, D.R., Holmes, D.E., Nevin, K.P., 2004. Dissimilatory Fe(III) and Mn(IV) reduction. In: Poole, R.K. (Ed.), Advances in Microbial Physiology, Vol. 49. Advances in Microbial Physiology, pp. 219-286. Ludwig, K., 2000. Users manual for Isoplot/Ex: a geochronological toolkit for Microsoft Excel. Berkeley Geochronology Center Spec. Publ. 1a, 53. Lyons, T.W., Reinhard, C.T., Planavsky, N.J., 2014. The rise of oxygen in Earth’s early ocean and atmosphere. Nature 506 (7488), 307–315. Ma, D.Q., Du, S.H., Xiao, Z.F., 2002. The origin of Huangling granite batholith. Acta Petrol. Mineral. 21, 151–161. Mloszewska, A.M., Pecoits, E., Cates, N.L., Mojzsis, S.J., O’Neil, J., Robbins, L.J., Konhauser, K.O., 2012. The composition of Earth’s oldest iron formations: The Nuvvuagittuq Supracrustal Belt (Quebec, Canada). Earth Planet. Sci. Lett. 317, 331–342. Morris, R.C., 1993. Genetic modelling for banded iron-formation of the Hamersley Group, Pilbara Craton, Western Australia. Precambr. Res. 60 (1), 243–286. Myers, J.S., 2001. Protoliths of the 3.8–3.7 Ga Isua greenstone belt, West Greenland. Precambr. Res. 105 (2), 129–141. Ndime, E.N., Ganno, S., Nzenti, J.P., 2019. Geochemistry and Pb-Pb geochronology of the Neoarchean Nkout West metamorphosed banded iron formation, southern Cameroon. Int. J. Earth Sci. 108 (5), 1551–1570. Nozaki, Y., Zhang, J., Amakawa, H., 1997. The fractionation between Y and Ho in the marine environment. Earth Planet. Sci. Lett. 148 (1–2), 329–340. Och, L.M., Shields-Zhou, G.A., 2012. The Neoproterozoic oxygenation event: Environmental perturbations and biogeochemical cycling. Earth Sci. Rev. 110 (1), 26–57. Ostrander, C., Johnson, A., Anbar, A., 2021. Earth’s First Redox Revolution. Annu. Rev. Earth Planet. Sci. 49, 337–366. Peng, M., Wu, Y., Gao, S., Zhang, H., Wang, J., Liu, X., Gong, H., Zhou, L., Hu, Z., Liu, Y., Yuan, H., 2012a. Geochemistry, zircon U-Pb age and Hf isotope compositions of Paleoproterozoic aluminous A-type granites from the Kongling terrain, Yangtze Block: Constraints on petrogenesis and geologic implications. Gondwana Res. 22 (1), 140–151. Pecoits, E., Gingras, M.K., Barley, M.E., Kappler, A., Posth, N.R., Konhauser, K.O., 2009. Petrography and geochemistry of the Dales Gorge banded iron formation: Paragenetic sequence, source and implications for palaeo-ocean chemistry. Precambr. Res. 172 (1), 163–187. Peng, S.B., Kusky, T.M., Jiang, X.F., Wang, L., Wang, J.P., Deng, H., 2012b. Geology, geochemistry, and geochronology of the Miaowan ophiolite, Yangtze craton: Implications for South China’s amalgamation history with the Rodinian supercontinent. Gondwana Res. 21 (2–3), 577–594. Peng, M., Wu, Y.B., Wang, J., Jiao, W.F., Liu, X.C., Yang, S.H., 2009. Paleoproterozoic mafic dyke from Kongling terrain in the Yangtze Craton and its implication. Chin. Sci. Bull. 54 (6), 1098–1104. Planavsky, N., Rouxel, O.J., Bekker, A., Hofmann, A., Little, C., Lyons, T.W., 2012. Iron isotope composition of some Archean and Proterozoic iron formations. Geochim. Cosmochim. Acta 80, 158–169. Polat, A., Frei, R., 2005. The origin of early Archean banded iron formations and of continental crust, Isua, southern West Greenland. Precambr. Res. 138 (1–2), 151–175. Poulton, S.W., Canfield, D.E., 2011. Ferruginous Conditions: A Dominant Feature of the Ocean through Earth’s History. Elements 7 (2), 107–112. Qiu, Y.M., Gao, S., McNaughton, N.J., Groves, D.I., Ling, W., 2000. First evidence of ＞ 3.2 Ga continental crust in the Yangtze craton of south China and its implications for Archean crustal evolution and Phanerozoic tectonics. Geology 28 (1), 11–14. Qiu, X.F., Ling, W.L., Liu, X.M., Lu, S.S., Jiang, T., Wei, Y.X., Peng, L.H., Tan, J.J., 2018. Evolution of the Archean continental crust in the nucleus of the Yangtze block: Evidence from geochemistry of 3.0 Ga TTG gneisses in the Kongling high-grade metamorphic terrane, South China. J. Asian Earth Sci. 154, 149–161. Rouxel, O.J., Bekker, A., Edwards, K.J., 2005. Iron isotope constraints on the Archean and Paleoproterozoic ocean redox state. Science 307 (5712), 1088–1091. Rubatto, D., 2017. Zircon: The Metamorphic Mineral. Rev. Mineral. Geochem. 83 (1), 261–296. Singh, V.K., Slabunov, A., 2015. The Central Bundelkhand Archaean greenstone complex, Bundelkhand craton, central India: geology, composition, and geochronology of supracrustal rocks. Int. Geol. Rev. 57 (11–12), 1349–1364. Smith, A.B., 2018. The Iron Formations of Southern Africa. Geology of Southwest Gondwana. Steinhoefel, G., Horn, I., Blanckenburg, F.V., 2009. Micro-scale tracing of Fe and Si isotope signatures in banded iron formation using femtosecond laser ablation. Geochim. Cosmochim. Acta 73 (18), 5343–5360. Sumoondur, A., Shaw, S., Ahmed, I., Benning, L.G., 2008. Green rust as a precursor for magnetite: an in situ synchrotron based study. Mineral. Mag. 72 (1), 201–204. Sun, McDonough, 1989. Chemical and isotopic systematics of oceanic basalts: Implications for mantle composition and processes. Geological Society London Special Publications 42 (1). Sun, L., Lechte, M., Shi, X., Zhou, X., Zhou, L., Fang, H., Xie, B., Wu, M., Tang, D., 2022. Hexagonal magnetite in Algoma-type banded iron formations of the ca. 2.52 Ga Baizhiyan Formation, North China: Evidence for a green rust precursor? Am. Mineral. 107 (5), 970–984. Viehmann, S., Hoffmann, J.E., Münker, C., Bau, M., 2014. Decoupled Hf-Nd isotopes in Neoarchean seawater reveal weathering of emerged continents. Geology 42 (2), 115–118. Wang, C., Wu, H., Li, W., Peng, Z., Zhang, L., Zhai, M., 2017. Changes of Ge/Si, REE+Y and SmNd isotopes in alternating Fe- and Si-rich mesobands reveal source heterogeneity of the ~2.54Ga Sijiaying banded iron formation in Eastern Hebei, China. Ore Geol. Rev. 80, 363–376. Wang, C.L., Zhang, L.C., Lan, C.Y., Dai, Y.P., 2014. Rare earth element and yttrium compositions of the Paleoproterozoic Yuanjiacun BIF in the Lüliang area and their implications for the Great Oxidation Event (GOE). Science China: Earth Sciences 57, 2469–2485. Warchola, T., Lalonde, S.V., Pecoits, E., von Gunten, K., Robbins, L.J., Alessi, D.S., Philippot, P., Konhauser, K.O., 2018. Petrology and geochemistry of the Boolgeeda Iron Formation, Hamersley Basin, Western Australia. Precambr. Res. 316, 155–173. Wei, J., 2021. Petrology and geochemistry of the Archean Huangling greenstone belt in the Yangtze Craton, South China. Precambr. Res. 364, 106340. Wei, J.Q., Wang, J.X., 2012. Zircon age and Hf isotope compositions of amphibolite enclaves from the Kongling complex. Geological J. China Universities 18 (4), 589–600. H. Zhou et al. Precambrian Research 383 (2022) 10690516Wei, J., Wei, Y., Wang, J., Wang, X., 2020. Geochronological constraints on the formation and evolution of the Huangling basement in the Yangtze craton, South China. Precambr. Res. 342, 105707. Wei, J.Q., Jing, M.M., 2013. Chronology and geochemistry of amphibolites from the Kongling complex. Chinese J. Gology 48 (4), 970–983. Wei, Y.X., Zhou, W.X., Hu, Z.X., Li, H.Q., Huang, X.X., Zhao, X.M., Xu, D.L., 2019. Geochronology and Geochemistry of Archean TTG and Tremolite Schist Xenoliths in Yemadong Complex: Evidence for >= 3.0 Ga Archean Continental Crust in Kongling High-Grade Metamorphic Terrane, Yangtze Craton, China. Minerals 9 (11). Welch, S.A., Beard, B.L., Johnson, C.M., Braterman, P.S., 2003. Kinetic and equilibrium Fe isotope fractionation between aqueous Fe(II) and Fe(III). Geochim. Cosmochim. Acta 67 (22), 4231–4250. Wu, C., Yang, T., Shields, G.A., Bian, X., Gao, B., Ye, H., Li, W., 2020. Termination of Cryogenian ironstone deposition by deep ocean euxinia. Geochem. Perspect. Lett. 15, 1–5. Xiong, C.Y., Wei, C.S., Jin, G.F., Tan, W.Q., Li, W.X., 2004. Pre-sinian paleostructural framework and major geological events in the huangling anticline, western hubei. J. Geomech. 10 (2), 97–112. Yakymchuk, C., Kirkland, C.L., Clark, C., 2018. Th/U ratios in metamorphic zircon. J. Metamorph. Geol. 36 (6), 715–737. Yang, X., Mao, J., Santosh, M., Li, R., Wu, C., Gao, J., Zhang, X., Wang, Z., 2022. Ediacaran iron formations from the North Qilian Orogenic Belt, China: Age, geochemistry, Sm–Nd isotopes and link with submarine volcanism. Precambr. Res. 368, 106498. Ye, H., Wu, C.Z., Yang, T., Santosh, M., Yao, X.Z., Gao, B.F., Wang, X.L., Li, W.Q., 2017. Updating the Geologic Barcodes for South China: Discovery of Late Archean Banded Iron Formations in the Yangtze Craton. Sci. Rep. 7. Zegeye, A., Bonneville, S.C., Benning, L., Sturm, A., Fowle, D., Jones, C., Canfield, D., Ruby, C., Maclean, L., Nomosatryo, S., Crowe, S., Poulton, S., 2012. Green rust formation controls nutrient availability in a ferruginous water column. Geology 40 (7), 599–602. Zhai, M.G., 2013. The main old lands in China and assembly of Chinese unified continent. Science China Earth Sciences 56 (11), 1829–1852. Zhang, L.C., Zhai, M.G., Wan, Y.S., Guo, J.H., Dai, Y.P., Wang, C.L., Liu, L., 2012. Study of the Precambrian BIF-iron deposits in the North China Craton: Progresses and questions. Acta Petrologica Sinica 28 (11), 3431–3445. Zhang, S.B., Zheng, Y.F., 2013. Formation and evolution of Precambrian continental lithosphere in South China. Gondwana Res. 23 (4), 1241–1260. Zhang, S.B., Zheng, Y.F., Wu, Y.B., Zhao, Z.F., Gao, S., Wu, F.Y., 2006. Zircon isotope evidence for ≥3.5Ga continental crust in the Yangtze craton of China. Precambr. Res. 146 (1), 16–34. Zhao, G.C., Cawood, P.A., 2012. Precambrian geology of China. Precambr. Res. 222, 13–54. Zhao, G., Guo, J., 2012. Precambrian geology of China: Preface. Precambr. Res. 222–223, 1–12. Zhao, J.H., Zhou, M.F., 2007. Geochemistry of Neoproterozoic mafic intrusions in the Panzhihua district (Sichuan Province, SW China): Implications for subduction- related metasomatism in the upper mantle. Precambr. Res. 152 (1–2), 27–47. Zhao, J.H., Zhou, M.F., Zheng, J.P., 2013. Neoproterozoic high-K granites produced by melting of newly formed mafic crust in the Huangling region, South China. Precambr. Res. 233, 93–107. Zheng, J., Griffin, W.L., O’Reilly, S.Y., Zhang, M., Pearson, N., Pan, Y.J.G., 2006. Widespread Archean basement beneath the Yangtze craton. Geology 34 (6), 417. Zhou, Z., Wei, X.Y., Mou, Z.Y., Li, F.H., D, Q.F., 2019. New Progresses in Geological Survey of Huangling Basement, Yangtze Craton. Resour. Environ. Eng., 33(S1), 1-6. H. Zhou et al. View publication stats | RESEARCH GATE |
REVIEWpublished: 28 April 2022doi: 10.3389/fnut.2022.800901Frontiers in Nutrition | www.frontiersin.org1April 2022 | Volume 9 | Article 800901Edited by:Andrew Scholey,Swinburne University ofTechnology, AustraliaReviewed by:Romana Stark,Monash University, AustraliaGeorge A. Brooks,University of California, Berkeley,United States*Correspondence:Wanju Sunsunwanju2021@163.comJingyun Hujingyunhu121@126.com†These authors have contributedequally to this work and share firstauthorshipSpecialty section:This article was submitted toNutrition, Psychology and BrainHealth,a section of the journalFrontiers in NutritionReceived: 24 October 2021Accepted: 18 March 2022Published: 28 April 2022Citation:Cai M, Wang H, Song H, Yang R,Wang L, Xue X, Sun W and Hu J(2022) Lactate Is Answerable for BrainFunction and Treating Brain Diseases:Energy Substrates and SignalMolecule. Front. Nutr. 9:800901.doi: 10.3389/fnut.2022.800901Lactate Is Answerable for BrainFunction and Treating BrainDiseases: Energy Substrates andSignal MoleculeMing Cai 1,2†, Hongbiao Wang 3†, Haihan Song 4, Ruoyu Yang 5, Liyan Wang 5, Xiangli Xue 6,Wanju Sun 4* and Jingyun Hu 4*1 Department of Rehabilitation Medicine, Shanghai University of Medicine and Health Sciences Affiliated Zhoupu Hospital,Shanghai, China, 2 Bio-X Institutes, Shanghai Jiao Tong University, Shanghai, China, 3 Department of Physical Education,Shanghai University of Medicine and Health Sciences, Shanghai, China, 4 Central Lab, Shanghai Pudong New Area People’sHospital, Shanghai, China, 5 College of Rehabilitation Sciences, Shanghai University of Medicine and Health Sciences,Shanghai, China, 6 Key Laboratory of Exercise and Health Sciences of Ministry of Education, Shanghai University of Sport,Shanghai, ChinaResearch to date has provided novel insights into lactate’s positive role in multiple brainfunctions and several brain diseases. Although notable controversies and discrepanciesremain, the neurobiological role and the metabolic mechanisms of brain lactate havenow been described. A theoretical framework on the relevance between lactate and brainfunction and brain diseases is presented. This review begins with the source and route oflactate formation in the brain and food; goes on to uncover the regulatory effect of lactateon brain function; and progresses to gathering the application and concentration variationof lactate in several brain diseases (diabetic encephalopathy, Alzheimer’s disease, stroke,traumatic brain injury, and epilepsy) treatment. Finally, the dual role of lactate in the brainis discussed. This review highlights the biological effect of lactate, especially L-lactate, inbrain function and disease studies and amplifies our understanding of past research.Keywords: lactate, brain function, brain diseases, energy substrates, signal moleculeINTRODUCTIONBrain lactate, as a well-known metabolite, primarily roots in astrocytic glycolysis from bloodglucose, glycogen, and blood lactate. Recently, the role of the “good guy” has gradually supersededthe traditional concept of metabolic waste in medical literature in neuroscience (1). The mostinteresting dimension of this is the physiological character of lactate’s role in mediating brainfunction (2). These canonical function involves learning and memory (3), cerebral blood flow (4),neurogenesis (5, 6) and cerebral microangiogenesis (7), energy metabolism (8), neuronal activity(9–11), and neuroprotection (12–15). Therefore, lactate is competent to be a potential therapyfor ameliorating the pathological process of some brain diseases associated with impaired brainfunction. In mammals, lactate exists as two enantiomers. The structure of asymmetrical C2 carbonleads to the two stereoisomers of lactate that are designated as L-lactate and D-lactate (16). L-lactateis the major enantiomer found in the brain and blood whereas D-lactate is normally present in veryCai et al.Lactate Role in Brain Functionlow concentrations under healthy physiological conditions (17).D-lactate is also considered as the rivalrous inhibitor of L-lactate since it competitively inhibits L-lactate transport (18). Indifferent brain disease patterns, L- and D-lactate is reported toexert a similar or distinct effect on brain function. The involvingmechanisms are far more complex than originally thought. Forthe most part, L-lactate can be utilized as a preferred energysubstrate of neurons for meeting the energy demand (19, 20) oract as the novel hormone-like effect called lactormone (21, 22).But current research about D-lactate’s role in brain function andbrain-related disease is sparse and debatable. D-lactate-mediatedmechanisms are also unclear.In this narrative review, we aim to provide a comprehensiveand profound summary of the role of lactate in brain functionand related diseases. Consequently, we expound the food sourceof lactate intake, discuss the lactate enantiomers and theirmetabolism manner in the brain, compare the influence of L-and D-lactate on brain functions, expound on the effect of L- andD-lactate replenishment on several common brain diseases, andsummarize the mechanisms of L-lactate.LACTATE ENANTIOMERSLactate in mammals exists as two enantiomers: the most commonform of L-isomers (known as L-lactate) is produced duringmammalian glucose metabolism, and a quite small quantity ofD-isomers (known as D-lactate) is generated from carbohydratesby bacterial metabolism (23, 24). The L-isomer of lactate isbelieved to have biological metabolic activity, while the D-isomeris too low in the body to activate the relevant enzymes forcatabolism (25).The Formation of Lactate in the BrainL-LactateIn the brain, the main source of L-lactate is astrocytic glycolysisfrom blood glucose, glycogen, and blood lactate (26, 27). Inastrocytes, glucose can be converted directly to L-lactate byglycolysis or be stored in the form of glycogen (28). Glycogen isalmost exclusively localized in astrocytes (20). As the neuronalactivity intensifies, astrocytic glycogen is mobilized to supplyneurons in case of neuronal glucose dissatisfying the energeticdemand (29). In consequence, besides maintaining the astrocytesitself energy demand, L-lactate also supports neuronal activityby providing ATP (30, 31). Furthermore, the elevated blood L-lactate can also cross the blood-brain barrier into the brain viathe monocarboxylate transporter 1 (MCT1) in some conditions(32, 33), such as vigorous exercise (34, 35) and fermentative (36)or fiber-containing foods (37).D-LactateD-lactate, as the stereoisomer of L-lactate, is barely found inthe brain and does not participate in energy production (23). Atiny amount of methylglyoxal (MG), the metabolic intermediaryproduct, can be produced during glycolysis (38). The glyoxalasesystem, mostly located in astrocytes, allows bulk MG to convertinto endogenic D-lactate or glutathione (GSH) (39).The Metabolism of Lactate in the BrainL-LactateThe hypothesis of astrocyte-neuron lactate shuttle (ANLS)describing L-lactate shuttling between astrocytes and neuronsis linked to glutamatergic signaling by Pellerin and Magistretti(40). A model opens the new insight for the L-lactate role inthe brain and perfectly elaborates the mechanism of L-lactatehow to serve as energy substrates. The ANLS switches onglutamate released by neuronal terminals and then taken upby astrocytes via the excitatory amino acid transporters 1 and2 (EAAT1 and 2) to convert into glutamine or glutathione,which is activated by a gradient. This process stimulates theNa+-K+ pump to favor astrocytic mitochondrial acidificationfor launching glycolysis (21). Phosphofructokinase-2/fructose-2,6-bisphosphatase 3 (Pfkfb3) is a key positive modulator ofglycolysis. Its expression and activity are high in astrocytesbut not in neurons. These cell-specific expression and activityprofiles result in the capacity of glycolysis being more activein astrocytes than neurons (26). The stored glycogen firstdecomposes into glucose-6-Phosphate (glucose-6P) catalyzed byglycogen phosphorylase. Then the glucose-6P gives rise to twomolecules of pyruvate. Finally, pyruvate can be either convertedto acetyl coenzyme A (acetyl-CoA) to be used directly by theastrocytic TCA cycle or transformed into L-lactate by lactatedehydrogenase 5 (LDH5) (26, 39).Monocarboxylate transporters (MCTs) belong to the SLC16gene family and facilitate the transmembrane H+-linkedmonocarboxylate, like lactate, pyruvate, and ketone bodies,to shuttle between astrocytes and neurons for the brainmetabolic demands (26, 41). MCT1 is mainly located at themembrane of astrocytes, oligodendrocytes, and endothelial cellsof blood vessels (42–44). The Km value of MCT1 for L-lactate approximates 3.5 mM. MCT4 is widely expressed at themembrane of astrocytes to cooperate with MCT1 in L-lactateefflux transport, in which the Km value for L-lactate is 34 mM.Compared with MCT1 and MCT4, MCT2 is specific in themembrane of neurons (45, 46) and has a higher affinity for L-lactate with the Km value of 0.7 mM (20, 47). The L-lactatefrom astrocytes or blood will be released into intercellularsubstances through MCT1 and/or MCT4, transported by MCT2into neurons (30, 48), and then oxidized at the Krebs cycle (21).The L-lactate oxidation in neuronal cytoplasm into pyruvateby LDH1 is the first step to metabolic removal (1, 38). Thenewly generated pyruvate later crosses the mitochondrial innermembrane in favor of presumptive pyruvate transporters MPC(MPC1 and MPC2) and/or MCT1 (16, 49), to transform intoacetyl-CoA for entering the tricarboxylic acid cycle (TCA)and therefore producing 14–17 ATPs per lactate molecule formaintaining neuronal activity (26). But the remarkable thing isthat the MPC is only identified in yeast, drosophila, and humans(49). The molecular masses of MPC1 (15 kD) and MPC2 (14 kD)are akin to the findings in the experiment performed in rat liverand heart (50). That is to say, deficient evidence shows that theMPC exists in the neurons so far. The localization of MCT1 inmitochondrial inner membrane is not only verified in rat heartand muscle (51–53) but also in the rat cortical, hippocampal, andthalamic neurons (54). Besides deductive pyruvate transporters,Frontiers in Nutrition | www.frontiersin.org2April 2022 | Volume 9 | Article 800901Cai et al.Lactate Role in Brain FunctionL-lactate may be directly oxidized by mitochondria (in heart,muscle, liver, spermatozoa, and brain) (22, 55) in the wakeof the concept “cytosol-mitochondria lactate shuttle” proposedby Brooks’ team (53, 56–58). In this model, L-lactate entersinto the mitochondrial intermembrane space through porins inthe outer mitochondrial membrane. Then it will be oxidizedby putative mitochondrial LDH (mLDH) (located at the innermitochondrial membrane) into pyruvate (59). Several studieshave confirmed that L-lactate can be taken up and metabolizedby the mLDH in both astrocytes (human astrocytoma cells)(60) and neurons (rat cerebellar granule cells, cortex neurons,and hippocampal neurons) (54, 61). The evidence implies thatcytosol-mitochondria lactate shuttle may occur in the brain(Figure 1).D-LactateD-lactate exerts metabolically inert in mammals (62) becauseof the absence of specific cytosolic lactate dehydrogenase (D-LDH) to metabolize D-lactate into pyruvate (63, 64). This enzymebarely exists in mammals (24). Some evidence shows that MCTscan transport D-lactate to inhibit L-lactate out of astrocytes orinto neurons despite the Km value of MCTs for D-lactate beingfar higher than L-lactate (65, 66). The rate of oxidation of D-lactate in the brain is considerably slower than that of L-lactatedue to the rather low expression levels of D-LDH in the brain(36, 67). Since interfering with more efficient energy substrates(pyruvate and L-lactate) for mitochondrial use, the physiologicalfunction of D-lactate is now known as the competitive inhibitorof L-lactate (18). Moreover, D-lactate is confirmed to interferewith the pyruvate metabolism in the brain, which eventuallyimpairs in state 3 and state 4 of mitochondrial respiration (63), itsaccumulation may have an adverse impact on energy metabolismand thus lead to toxicity (Figure 1).LACTATE INTAKES FROM THE FOODLactate, as the predominant end-product of lactose fermentationin the food (68), is produced by the lactic acid bacteria,namely, gram-positive and catalase-negative microorganisms.Such microorganisms are involved in the fermentation of arange of foods and beverages, such as dairy products, meat, fish,vegetable, sourdough, wine, and cider (68), which creates thespecific flavors and aromas for the food and benefits humanhealth (69). People can uptake lactate from cheese, yogurt, wine,fermented vegetables (69), and fermented oyster extract (70).Thereby, lactate will appear in the gut via the consumption offermented foods. In addition, prebiotic, fiber-containing foodsalso are the way of the lactate intake, such as broccoli, brusselssprouts, cabbage, cauliflower, collard greens, kale, radish, andrutabaga (37). Since gut microbiota is likely to produce aracemic mixture, it is not surprising that L- and D-lactate aregenerated simultaneously following the above food intake (37).The disposal of lactate in the lower gut can be converted toacetate, butyrate, propionate, and succinate. Alternatively, theunverified idea of “gut-soma lactate shuttle” is another wayfor the disposal of lactate in the gut, in which the lactateproduction in the gut releases into the systemic circulation(37). The discovery of the sodium-dependent monocarboxylatetransporter (SMCT) (including SMCT1 and SMCT2) is favorableevidence for supporting this hypothesis, which is located inthe mouse digestive tract and involved in the transport offood-derived monocarboxylates, such as lactate (71). The otherimplicit clue is the phenomenon of the rise in the blood L-lactate after a carbohydrate diet (37, 72). Tappy’s team finds thatdietary fructose or the co-mixture (fructose and glucose) facilitatethe L-lactate release into the systemic circulation (73, 74).Furthermore, glucose rooted in the oxidation of carbohydrateonly provides 10–20% energy, while other carbohydrate energysources like glycogen and L-lactate accounts for 70–80% in thecondition of exercise (75). That is to say when blood glucoseis supported by hepatic glycogenolysis and gluconeogenesis, L-lactate plays important role in carbohydrate energy substratedistribution (72). As blood L-lactate can be transported intothe brain via the MCT1 (32, 33), net L-lactate uptake directlyprovides 12% of brain fuel (37, 75, 76). Besides, gluconeogenesisprovides 45% of brain fuel. Thereby, L-lactate comprises 57%of the total brain energy source (75). Regarding D-lactate, itcan be excreted in urine by renal (64) or feces by gut (37) andcannot be detected in the blood under normal physiologicalconditions (64). Its excessive accumulation can result in acidosisand irritation of the lower bowel. Furthermore, the release ofD-lactate into circulation also cause neurotoxic effect (37). Theclinical presentation of D-lactic acidosis is characterized byepisodes of encephalopathy and metabolic acidosis (24).THE ROLE OF LACTATE IN REGULATINGBRAIN FUNCTIONSLactate is reported to participate in the regulation of various brainfunctions in the terms of learning and memory (3), cerebral bloodflow (4), neurogenesis (5, 6) and cerebral microangiogenesis(7), energy metabolism (8), neuronal activity (9–11), andneuroprotection (12–15). L-lactate and D-lactate are reported toexert similar or distinct effects on those functions.The Distinction Between L-Lactate andD-Lactate for Influencing Learning andMemoryThe Effect of L-Lactate on Learning and MemoryTheglycogenisnecessaryforlong-termpotentiationmaintenance in the mouse Shaffer collateral-CA1 synapse(77). Shima et al. find that 4-week exercise can increase glycogenreserve, along with the elevated transport rate of L-lactate into theneurons, to ameliorate memory dysfunction in diabetic rats (78).Extracellular L-lactate rapidly increases in the rat hippocampusduringspontaneousalternation.Intrahippocampal50 nML-lactate can not only enhance the memory in this task butwill rescue impaired memory in the condition of glycogenolysisinhibition (27). This evidence shows that the use of L-lactatemetabolically coupled astrocytes and neurons depends on thecharacter of high-energy demands for memory consolidation andstorage (30). The provision of L-lactate by astrocytes is proved tobe a more generally important element of learning and memoryFrontiers in Nutrition | www.frontiersin.org3April 2022 | Volume 9 | Article 800901Cai et al.Lactate Role in Brain FunctionFIGURE 1 | The formation and metabolism of lactate in the brain. The action of the glutamate-glutamine cycle switches on the glycolysis in the astrocytes, a processof glutamate released by neuronal terminals and then is taken up by astrocytes via the EAATs to convert into glutamine or GSH. Pfkfb3 (a key positive modulator ofglycolysis) is a high expression in astrocytes and low in neurons, while the activity of PDH is low in astrocytes and low in neurons. These characteristics determine thatastrocytes have strong glycolysis ability and neurons have strong aerobic oxidation ability. Astrocytic glycogen first decomposes into glucose-6P catalyzed by GP andfinally into pyruvate through a series of catalytic reactions. During this process, a tiny amount of MG generates and then converts into endogenic D-lactate or GSH. Thenewly formed pyruvate can be either converted to acetyl-CoA for the astrocytic TCA cycle or transformed into L-lactate by LDH5. The L-lactate in the astrocytes andin the brain blood vessel is transported into the neurons through MCTs (MCT1, MCT2, and MCT4) and then oxidized at the Krebs cycle. EAATs, excitatory amino acidtransporters; GSH, glutathione; Pfkfb3, phosphofructokinase-2/fructose-2,6-bisphosphatase 3; PDH, pyruvate dehydrogenase; Glucose-6P, glucose-6-Phosphate;GP, glycogen phosphorylase; MG, methylglyoxal; acetyl-CoA, acetyl coenzyme A; lactate dehydrogenase 5, LDH5; Monocarboxylic acid transporters, MCTs.processing than memory consolidation (29). Harris et al. findthat the L-lactate produced by glycolysis, is required for memoryacquisition but not for established memory in mice at the ageof 9 months (79), suggesting that the main function of L-lactateis regulating the process of learning (80). Recent studies revealthe fact that L-lactate transporters are necessary for L-lactatemediating the memory process, especially MCT2. Inhibition ofMCT1 or MCT2 can impair the rat reconsolidation of cocainememory (81) or the mouse long-term memory formation (82).L-lactate replenishment will reverse the memory impairmentin MCT1 or MCT4 knockdown of rat hippocampus (77). Bycontrast, L-lactate replenishment or even glucose supplementwill not rescue the impaired memory once blocking MCT2activity, which transports L-lactate into neurons (27, 77). Inthese studies, L-lactate in neurons is seen as the pyruvate donor,which produces ATP for neuronal energy demand (see section“Serves as an energy substrate for neurons”). Therefore, thetransfer of L-lactate from astrocytes into neurons to supportneuronal functions is necessary for memory consolidation(77, 83). The mechanisms may involve sustaining synaptictransmission and function (84–88) and regulating synapticplasticity-related genes and proteins expression (89, 90). Infacilitating the synaptic transmission and function, Tang et al.find that injections of 2 mM L-lactate into the locus coeruleuscan activate an unidentified Gs receptor to evoke the NEergicneuronal excitability for norepinephrine release (85). Schurr etal. first find that L-lactate can replace glucose as a sole energysubstrate for sustaining the normal synaptic function in rathippocampal slices for hours (87). Lucas et al. declare thatthe physiological concentration of L-lactate can be utilized bypresynaptic terminals as the energy for meeting the demand ofmaintaining functional presynaptic release sites (84). Herrera-Lo’pez et al. find that 1–2 mM L-lactate can induce glutamatergicsynapsepotentiationtopromotethememoryformationprocess in rat hippocampal CA3 pyramidal cells (86). As formodulating the synaptic plasticity genes and proteins expression,Margineanu et al. find that 20 mM L-lactate treatment canpromote various plastic plasticity and plastic activity-relatedgenes in mouse cortical neurons through RNA sequence (89).Yang et al. find that L-lactate from 2.5 to 20 mM will potentiatethe NMDAR activity to improve the immediate early genes(IEGs) expression in mouse cortical neurons, which benefitsthe form of learning and memory (90). Hayek et al. find thatintraperitoneal injection physiological concentration of L-lactate(117 or 180 mg/Kg) for 1 month can promote the proteins andgenes expression of brain-derived neurotrophic factor (BDNF)and IEGs to improve the mouse learning and memory ability(3) (Figure 2).Frontiers in Nutrition | www.frontiersin.org4April 2022 | Volume 9 | Article 800901Cai et al.Lactate Role in Brain FunctionFIGURE 2 | The effect of L-lactate and its possible mechanism on the brain. L-lactate is capable of regulating brain function in the terms of learning and memory,cerebral blood flow, neurogenesis and cerebral microangiogenesis, energy metabolism, neuronal activity, and neuroprotection via different pathways.The Effect of D-Lactate on Learning and MemoryD-lactate is proved to impair the neonatal chick memoryprocess in the way of intracranial injection (23, 91). Thereasons ascribe inhibition of L-lactate uptake into neuronsand interference with the astrocytic metabolism. Baker andEdwards find that bilateral administration of D-lactate in thelower dosage range of 1.75–2.25 mM inhibits memory retentionafter discrimination avoidance task from 40 min and onward,and the inhibitory effect sustains 140 min. And the scope ofthe validity period is from 10 min before to 20 min after thetask (91). This time window is similar to the dosage of 10 nM(23). The observation indicates that the inhibitory effect of D-lactate on memory is valid for a certain period (23). Furtherto support this idea is the study of rodents. Michael et al.find that intracranial injection of a higher concentration of D-lactate (18 mM) at 30 min before the Y maze task seems to notaffect memory retention (92). Scavuzzo et al. even find thatsubcutaneous injections of D-lactate (1 g/kg) impair memoryat 15 min before the inhibitory avoidance (IA) task, whereasit significantly enhances the rat memory at 2 min after thetraining (80) (Table 1). Thereby, it is still debatable that thedifferent concentration and time-point of D-lactate usage leadto the distinct results for memory up to date. The conflictresults are puzzling what is the physiological role of D-lactatein learning and memory. In short, it will be interesting tofigure out the mechanism of D-lactate on the memory atthe level of molecules and physiology. More interestingly, theresults of intracranial unilateral injection by Gibbs and Hertzalso show that D-lactate injection into left intermediate medialmesopallium inhibits memory formation from 10 min beforetask, whereas injection into the right hemisphere works from10 min after task (23), implying that the time window of D-lactate on inhibiting memory formation varies from the part ofthe brain hemisphere.L-Lactate Increases Cerebral Blood Flow(Vasodilation)For the last few years, the function of L-lactate regulatingcerebral blood flow (CBF) has been proposed in severalstudies (4, 95–98). First, a hypoxia-induced increase of L-lactatecan improve the prostaglandin transporter (PGT) efficacy forvasodilator prostaglandin E2 (PGE2) intake in the brain (4).Second, exercise is reported to trigger the N-methyl D-aspartate(NMDA) receptor/nitric oxide (NO) signaling for realizing thehippocampal functional hyperemia during neural activation (99).Previous evidence indicates the effect of L-lactate on enhancingthe NMDAR activity in cortex neurons (90). Moreover, whenthe porcine second-order retinal arterioles are pressurized to noflow, 10 mM L-lactate can induce the release of NO synthaseto activate the downstream guanylyl cyclase/cGMP signaling,which opens the KATP channels for vasodilation in retinalarterioles (100). Contemplating all this, it is possible for L-lactate to dilate the brain microvessels for increasing CBF viaNO biological effect. During hypoglycemia, the CBF increasecan be considered neuroprotective since it is an attempt toincrease capillary glucose concentration for improved glucosesupply to the brain (101). Third, L-lactate has the autoregulatoryvascular function in switching the pericyte response fromcontraction to dilation in the retinal vasculature. In the normalphysiological state, L-lactate activates the Na+/H+ exchangers,inhibits Na+/Ca2+ exchangers, and excites Na+-K+ pumpsto trigger pericyte vasoconstriction regardless of the L-lactateconcentration (Figure 2). While under the hypoxic condition,20 mM L-lactate induces mural cell relaxation throughoutthe retinal vasculature, from the arterioles to the capillaries(102). Exploring the L-lactate-induced vascular reaction whendealing with the physiological and pathological situation, maybe crucial for aiding in our understanding of neurovascularcoupling, enhancing brain functions, and even treating multiplebrain diseases.Frontiers in Nutrition | www.frontiersin.org5April 2022 | Volume 9 | Article 800901Cai et al.Lactate Role in Brain FunctionTABLE 1 | The effect of D-lactate on the brain function.Brain functionModelMethod ofadministrationConcentrationIntervention timeEffectLearning and memoryNeonatal chick (91)Tail intravenousinjection1.75–2.25 mMThe scope of validity periodfrom 10 min before to 20 minafter the discriminationavoidance taskImpair memoryNeonatal chick (23)Intracranialinjection10 nMThe scope of validity periodfrom 10 min before to 20 minafter the discriminationavoidance taskInjection at 20 min before or25 min after the taskImpair memoryNo adverseeffecton memory9∼10-week-old rats (92)Intracranialinjection18 mMInjection at 30 min before theY maze taskNo adverseeffect onmemoryretention6∼11-week-old rats (80)Subcutaneousinjection1 g/kgInjection within 2 min followingIA taskInjection at 15 min prior to IAtaskEnhance memoryImpair memoryNeuronal activityGlutamatergic neuronsand GABAergic neurons(9)5 mMDecrease thespontaneouscalcium spikingfrequencyNeuroprotectiontMCAO modelHippocampal slices inOGD model (12)Intravenousinjection1 µmol/g4 mMInjection at 45 min aftertMCAO48 h after OGDDecrease theinfarct volumeReduce neuronalcell deathNMDAR-inducedneuronal excitotoxicity(93)10 mM22 min before addingglutamate.No alteration inexcitotoxicityGlutamate-inducedneurotoxicity model (94)Infusion6 mMCombination with containingglutamate perfusion mediumAggravate thecortex lesionareaL-Lactate Benefits the Process ofNeurogenesis and CerebralMicroangiogenesisNeurogenesis is important for memory and learning. Lambertuset al. find that the 7-week exercise or 18 mM L-lactate can induceneurogenesis in the mouse ventricular-subventricular zone butnot in the hippocampus (5). Notably, the author mentioned thatneurogenesis in this zone contributes to the olfactory memory,which may provide contextual clues to the spatial-visual memorycontrolled by the hippocampus (5). Controversially, Vachnishet al. find that 6-week 13–17 mM L-lactate can promote micehippocampal neurogenesis (6). The distinct results are probablyrelated to the selection of neurogenic markers. The formerrefers to DCX and Ki-67 to represent neurogenesis, whichshows immature granular neurons and intermediate neuralprogenitor cells, respectively. The latter chooses BrdU as theneurogenic marker, which shows nerve cells in proliferation anddivision. The different markers are on behalf of different stagesof neurogenesis (103). The above evidence suggests that theneurogenic effect of L-lactate is tissue-specific. For example, theeffect of L-lactate on the ventricular-subventricular zone dependson the GPR81 action but the hippocampus depends on the MCT2metabolic action. Since neurogenesis is highly correlated withcerebral microangiogenesis, there is no doubt that angiogenesisalso may be influenced by L-lactate in the brain. The directevidence is supported by Moland’s team. The study declares that7-week 18 mM L-lactate is capable of increasing the density ofmicrovessels in the dentate gyrus of the hippocampus along witha higher VEGFA expression. Furthermore, this effect depends onthe GPR81 action (7) (Figure 2). In the further study, figuringout the approaches of L-lactate how to deal with neurogenesisand cerebral microangiogenesis (metabolic pathway, lactormoneaction, or both), may be worthiness for understanding thephysiological role of L-lactate in brain function regulation.L-Lactate Influences Brain EnergyMetabolismMitochondria are best known for their role in the generationof ATP that supplies eukaryotes with energy to serve theircellular needs (104). Recent studies report that L-lactate can alsomediate various mitochondria-related genes. A novel discoveryis first reported by Brooks’s team in L6 cells (105). Theyfind that 20 mM L-lactate can upregulate 79 genes involvedin mitochondrial metabolism (MFN1, MFN2, PGC-1α, NRF2,Frontiers in Nutrition | www.frontiersin.org6April 2022 | Volume 9 | Article 800901Cai et al.Lactate Role in Brain FunctionLDHb, ATP5g1, NADH-dh, SDH, and TIM) and oxidative stress(GPX1, Glrx2, Glrx5, Prdx2, and Txndc12) (105). The evidence isalso verified in the brain. For example, intraperitoneal injectionof 18 mM L-lactate for 14 consecutive days will promote thePRC mRNA expression and the mtDNA levels (8). 20 mML-lactate pretreats SY5Y cells can reverse high-concentrationhydrogen peroxide (H2O2)-induced oxidative stress injury,including NRF2 expression improvement and mitochondrialmembrane potential potentiation (106). The biological effect ofL-lactate mediating mitochondrial metabolism and function isalso in the primary mouse neurons. A total of 15–20 mM L-lactate can improve mitochondrial fusion (OPA1, MFN1, andMFN2), inhibit mitochondrial fission (DRP1 and FIS1), andpromote biogenesis (PGC-1α, NRF2, TFAM, and mtDNA) (3,107) (Figure 2). Besides, “cytosol-mitochondria lactate shuttle”also elaborates the role of L-lactate as the substrate sourceof mitochondrial ATP production (53, 56–58). Thus, L-lactateought to have a close relationship with mitochondria in the brain.Considering the fact that multiple brain diseases in a large partare also associated with the energy crisis, the application of L-lactate in an animal-related experiment to shed light on the effectand mitochondria mechanism on brain energy metabolism asearly as possible should have a great significance of the noveltreatment in clinic brain diseases.L-Lactate and D-Lactate InfluenceNeuronal ActivityThe previous study demonstrates that lactate interacting with itsreceptor GPR81 operates negative feedback on neuronal activity.Application of 5 mM L- or D-lactate reversibly diminishes thecalcium transient frequency by more than 50% in both principaland GABAergic neurons (Table 1). Moreover, the activationof GPR81 can also mimic a similar potency as the lactate(9). In fact, lactate can bind to GPR81 to couple with theGiα subunit to inhibit the intracellular adenylate cyclase (AC)-cyclic adenosine monophosphate (cAMP) cascade signal thatcontributes to the reduction of exocytosis (10, 11). In addition,the GPR81 activation can bind to the Giβ subunit to regulatethe activity of phospholipase C (PLC) and thus inhibit neuronalexcitability (10) by inducing a hyperpolarization for potassium(K+) indrawal or activating the GABA receptor for exocytosisreduction (9). Interestingly, 5 mM L-lactate and 0.56 mM 3,5-dihydroxybenzoic acid (3,5-DHBA) (one of the GPR81 agonists)reduce the firing frequency of CA1 pyramidal cells, whereasa higher level of L-lactate (30 mM) and 3,5-DHBA (3.1 mM)increase the firing frequency (11). In the locus coeruleus, 2 mML-lactate has rather an excitatory effect on the NEergic neuronsby activating an unidentified Gs receptor, which will be abolishedby the D-lactate (85) (Figure 2). Given these, we speculate thatthe effect of L-lactate on neuronal activity may depend on theL-lactate concentration, the types of neurons, and the types ofreceptors. Knowledge of the physiological effects of differentconcentrations and lactate isomers on the activity of differentneurons, application of lactate is a potential therapeutic wayto improve the complex neurological symptoms of epilepsy,which disease is characterized by neuronal hyperexcitability andsudden, synchronized electrical discharges (47, 108). In fact,some studies have focused on the application of L-lactate intreating epilepsy (see section Epilepsy).The Neuroprotective Role of L-Lactate andD-LactateLactate has been reported the role of neuroprotection in multiplebrain disease models, however, the effect of the isomer of L-and D-lactate is split depending on the disease patterns. Asknown, transient middle cerebral artery occlusion (tMCAO) invivo and oxygen and glucose deprivation (OGD) in vitro arethe desired models for stroke. In the tMCAO model, L-lactatetreatment can ameliorate neurological deficits after ischemiaby inducing sustained neuroprotection for up to two weeks(14, 15). And as for the D-lactate treatment, it is also reportedto exert neuroprotection by decreasing the infarct volume andameliorating neurological deficits by working as the energysubstrate or activating the receptor GPR81 (12). In the OGDmodel, the effect of ameliorating neuronal injury is possessed byboth L-lactate and D-lactate (12–14). Hence, L- and D-lactateseem to exert a neuroprotective role in stroke.In the human and rodent traumatic brain injury (TBI)model, evidence shows that endogenous produced brain L-lactateor exogenous L-lactate replenishment can facilitate neurologicrecovery by improving synaptic plasticity (109) and regulatingthe brain metabolic state (76, 110–114) (see section “TBI”).However, the level of blood D-lactate is adopted as the biomarkerof TBI degree. This is because that D-lactate is one of themicrobe-dependent metabolites which will pass through thedestroyed intestinal barrier into the blood when TBI causes gutdysbiosis (115, 116).In addition, L-lactate but not D-lactate can reduce theexcessive NMDAR-induced neuronal excitotoxicity, which iscommon in some acute brain pathologies. 10 mM L-lactatecan improve the intracellular ATP production to stimulatemetabotropic purinergic receptor P2Y for activating the PI3Kpathway and opening the KATP channels, as a result ofdecreasing the intracellular Ca2+ concentration (93) (Figure 2).Furthermore, in the high concentration of glutamate-inducedneurotoxicity model, infusion of 6 mM L-lactate can obviouslyreduce the rat cortex lesion from 6.05 ± 0.64 mm3 to 4.16 ±0.43 mm3. On the contrary, the same concentration of D-lactateaggravates the lesion area from 2.7 ± 0.4 mm3 to 4.4 ± 0.5 mm3(18) (Table 1). With a view to the metabolism of lactate and theuniversality of neuroprotective effects in the brain, L-lactate is apreferred research direction and practical value of the applicationand treatment in neuroprotection.LACTATE IS THE POTENTIAL THERAPYFOR MULTIPLE BRAIN DISEASESIn the preceding decades, lactate was long considered as a wasteproduct or even a sign of cerebral harm. As the proposal of thenovel opinion such as “lactate shuttles” (37) and “lactormone”(22), these theories transform our stereotype and graduallyunlock the mystery of lactate in physiological and pathologicalFrontiers in Nutrition | www.frontiersin.org7April 2022 | Volume 9 | Article 800901Cai et al.Lactate Role in Brain Functionstates. Recently, lactate (especially L-lactate) is postulated toprotect the brain in several pathologic conditions, such asdiabetic encephalopathy, Alzheimer’s disease (AD), stroke, TBI,and epilepsy.Diabetic EncephalopathyDiabetes mellitus often hurts brain health, which accompaniesthe symptoms of brain functional decline (117–119), especiallycognitive impairment and vascular dementia (120). Thissymptom is called “diabetic encephalopathy” (121). Evidenceshows that the brain lactate falls in response to hypoglycemiain diabetes (122, 123). Intravenous injection of L-lactate iscapable of reducing the symptomatic and adrenaline responsesto hypoglycemia and simultaneously alleviating hypoglycemia-induced cognitive dysfunction in diabetic patients (124–127).In addition, intracerebroventricular administration of 200 nML-lactate can also mitigate the deterioration of memoryconsolidation via recovering the amplitude of sharp-waveripples (SWRs) in diabetic mice (128). These results directlydemonstrate that the L-lactate supplement benefits diabetic brainfunction. Another indirect evidence is that an increase of L-lactate transporter MCTs in the cortex and/or the hippocampus,especially MCT2 (transport L-lactate into neurons), is observedto be capable of ameliorating the cognitive impairment indiabetic encephalopathy (32, 78, 129). It is conceivable that L-lactate uptake enhancement supported by MCTs may improvediabetic encephalopathy. However, some studies show thephenomenon that the L-lactate increases in the brain regions ofcorpus callosum and hippocampus in diabetic encephalopathy(130–132). One possibility is the low utilization rate of L-lactateduring diabetic encephalopathy (133, 134). The convincingevidence listed is that the L-lactate is secreted to increase inastrocytes but not in neurons in the condition of a high-glucoseenvironment (134, 135), implying that the transport of L-lactateinto neurons is hampered and thus the elevated brain L-lactate isobserved in diabetic encephalopathy.ADAlzheimer’s disease (AD) is one of the most common forms ofbrain disease (136), characterized by a series of neuropathologicalchanges, such as cognitive dysfunction. Cognitive impairmentin AD is associated with recessionary energy metabolism (137).Whereas, L-lactate seems like an important energy substrate formemory formation in connection with energy supply (19, 27). Infact, decreased L-lactate content is observed in the cerebral cortexand hippocampus in AD rodent models (138–140). Nevertheless,elevated L-lactate levels occur in aging models (94, 141, 142) andAD patients (143). The increased lactate dehydrogenase A (LDH-A)/lactate dehydrogenase B (LDH-B) gene activity ratios (94) anddownregulated MCTs (MCT1, MCT2, and MCT4) are due forthe opposite results (139, 140). This is because that increasedbrain L-lactate production owing to the LDH-A/LDH-B ratio butblockage of L-lactate transport from glia to neurons, resultingin L-lactate deficit in neurons. Inferior L-lactate utilization rateleads to energy deficiency in neurons and thereby exacerbates theprogression of neuronal injury, including cognitive impairmentduring the pathological process of AD or aging.Regarding D-lactate, a study reports that D-LDH, locatedattheinnersideoftheratlivermitochondrialinnermembrane (MIM), oxidizes D-lactate to pyruvate in themitochondrial matrix (144). This process will improve oxidativephosphorylation (OXPHOS) efficiency, synthesis, and effluxof biosynthetic precursors from mitochondria, balance thecytosolic and mitochondrial GSH pools, and enhance theNADPH production in the cytosol. All of these can contributeto mitochondrial energy production (145). Nevertheless, theD-LDH isoforms in both human and mouse brains show arather weak signal (67). Based on this evidence, we speculatethat moderate D-lactate combined with D-LDH supplementmay target a novel mitochondrial energetic mechanism forAD treatment.StrokeStroke, one of the important factors causing neurologicalmorbidity and mortality, is most characterized by ischemia andinfarcts involving the territory of the middle cerebral artery(146). Berthet et al. adopt different concentrations of L-lactatetreatment after OGD in the rat hippocampal slices. The resultsshow that, immediately after OGD, the low dose treatment ofL-lactate (4 mM) significantly inhibits the hippocampal neurondeath, while L-lactate at the 20 mM dose aggravates the neuronalinjury (14). Whereas, during OGD, only a high dose of L-lactateat 20 mM is the capacity of reversing the ATP decline to preventcell death in primary cultured neurons and N2A cells, but notthe low dose at 1 mM (13). Hence, the neuroprotective effectof L-lactate on stroke may be related to the time points forintervention or the types of neuron cells. Further study showsthat L-lactate can reduce the size of the lesion area to promoteneurological rehabilitation in the way of intracerebroventricularinjection of 100 mM L-lactate (14) or tail vein injection of200 mM L-lactate (15) immediately after tMCAO. They speculatethat the ischemia-induced lactate increase may initially have aneuroprotective effect on the brain, but also may cause lacticacidosis when the lactate concentration reaches higher levels(14). Therefore, different concentrations of L-lactate may havedifferent regulatory effects on brain metabolism and functionunder different physiological metabolic states. The complexmetabolic characteristics of L-lactate bring immense challengesfor revealing its role in brain function (14). Soon afterward,their team finds that similar neuroprotection occurs in D-lactate.The results suggest that intravenous injection of 1 µmol/gD-lactate will decrease about half of the infarct volume intMCAO of the mice. Moreover, in the OGD model, 4 mMD-lactate treatment can also reduce about 50% neuronal celldeath in the rat hippocampal slices (12). Future studies needfurther excellently clarify the mechanism of how lactate performsneuroprotection although it has fully considered the dual roleof lactate as the energy substrate (pyruvate replacement) orsignal molecular (receptor agonist replacement). So far, tworesults are responsible for explaining the neuroprotective effectof lactate on stroke. First, the accumulated lactate is transportedby MCTs into the neurons and utilized as the alternative energysubstrate immediately postischemia (147) in favor of sustainingneuronal integrity (148) and delaying the neuronal damage (149).Frontiers in Nutrition | www.frontiersin.org8April 2022 | Volume 9 | Article 800901Cai et al.Lactate Role in Brain FunctionSecond, lactate may interact with the known receptor-GPR81 orunknown putative Gs-type receptors to resist ischemic injury.It is worth noting that the evidence and characterization ofGPR81 are still disputed in stroke. For example, Castillo et al.find the 3,5-DHBA treatment, one of the GPR81 agonists, canreduce neuronal cell death in OGD (12). In contrast with thestudy, Shen et al. find that inhibition of the GPR81 activitycan promote the ERK1/2 signal to copy with apoptosis inneurons (13). Moreover, a newly published study also reportsthat the GPR81 agonists (3Cl-5OH-BA and 3,5-DHBA) do exertno neuroprotective effect on stroke mice whether intravenousadministration or intracerebroventricular administration (150).Applying the genetic tools for GPR81 knockout may be apreferred protocol to definite the role and the mechanism of thisreceptor in the stroke.TBITraumatic brain injury (TBI) is a type of brain injury acquiredfrom an external force that inflicts devastating effects on thebrain vasculature and neurons (116). It is always reported toface an energy crisis in the brain in the fact of glucose uptakesuppression following cerebral injury. Although glucose is thepreferred fuel of the brain, it may result in odious insults, such ashyperglycemia and infection or mortality events increment (151–153). As in previous studies, many TBI patients have observedthe increased net L-lactate uptake after-hours postinjury (76,154) without hyperglycemia (110), implying the protectivefunction of L-lactate in TBI. Increasing blood arterial L-lactate tosupraphysiologic range (4–5 mmol/l) in the early phase of braininjury can accelerate the recovery of neural function throughcerebral perfusion and brain glucose availability improvement(155). Brooks and Martin find that endogenous L-lactate, inducedby TBI, can support brain metabolism indirectly or directly (76).On the one hand, endogenous lactate is reported to be related toaccelerating hepatic and renal gluconeogenesis to make glucoseavailable for essential organs like the brain. On the other hand,elevated blood L-lactate may provide substrate directly to theinjured brain in link with the L-lactate shuttle (114). Throughisotope tracer, ∼70% carbohydrate (direct lactate uptake andindirect glucose uptake from lactate) is provided for the TBIbrain (113), suggesting the importance of endogenous L-lactategeneration as the energy supply in TBI. In addition, exogenousintraperitoneal injection of 500 mg/kg L-lactate before 30 min ofTBI is reported to alleviate the brain injury-induced neurologicaldeficits by promoting the neuronal plasticity proteins (PSD95,GAP43, and BDNF) (109). Intravenous infusion of 5 mM L-lactate is also verified to utilize with sparing of cerebral glucosealong with a reduction of brain glutamate and intracranialpressure after TBI, which implies the benefit of cerebral metabolicand hemodynamic effects (110). A total of 100 mM L-lactateinfusion is shown reduced lesion volume at day 2 after cerebralcortical impact through slightly regulating cerebral blood flow(111). Hence, L-lactate, as the alternative source of glucose,has dream prospects to improve the outcome of posttraumaticbrain damage.Recently, the lactate receptor GPR81 is also reported topotentially participate in the pathological process of TBI. Theevidence is that the increased GPR81 gene expression lasts atleast 28 days after TBI injury in lesion areas of the cerebralcortex and hippocampus (156). Coincidentally, L-lactate furtherenhances the expression of GPR81 in the ipsilateral cortex andhippocampus 24 h after TBI (109). These results show that L-lactate may not only be the brain fuel, but also the possibleupstream molecule that interacts with the receptor to activate orinhibit the downstream molecular signals for neuroprotection inthe TBI.EpilepsyEpilepsy is a common neurological emergency with considerableassociated healthcare costs, morbidity, and mortality, whichis often causing neuronal and glial damage (157). Duringepilepsy, L-lactate is observed to rise in the brain regionsof cerebral gray and white matter (158), cerebrospinal fluid(159), cerebral cortex (160), and hippocampus (161). Previously,one research has proved that the elevated L-lactate level intissues can persist for approximately 1 h (161). In some cases,the L-lactate level can even reach 6 mM (162). Jorwal andSikdar find that 6 mM L-lactate application reduces the spikefrequency and hyperpolarizes the subicular neurons in rathippocampal slices, suggesting an anticonvulsant effect of L-lactate. Further electrophysiological recordings reveal that L-lactate induces the Giβγ subunit activity to expedite the inwardlyrectifying potassium (GIRK) channel. They also point out aninteresting phenomenon that normal brain concentration ofL-lactate (2 mM) and even up to 6 mM have few effects onrat hippocampus neural activity in the absence of epileptiformactivity. This fact reminds us that the important role of elevatedbrain L-lactate in epilepsy (163). It is worth thinking thatsince there is no negative effect of excessive brain L-lactateon the mean spike frequency and membrane potential ofsubicular pyramidal neurons in the normal physical condition,what physiological effects can this L-lactate have after beingmetabolized or working as the neurotransmitter? Consideringthe fact that and concentration of 6 mM L-lactate is sufficientto activate its receptor GPR81, which is also known to coupleto Giβγ subunit to inhibit the neuronal activity (10). Anotherfunction of GPR81 is to couple to the Giα subunit to inhibit theAC-cAMP cascade signal, as a result, the decrease of neuronalexcitement (11). However, unlike L-lactate, 3,5-DHBA does notresult in hyperpolarizing subicular neurons (10). Therefore, it islikely that L-lactate may interact with other unknown Gi-typeproteins. L-lactate-receptor signals may be the novel target forantiepileptic treatment.THE NEUROBIOLOGICAL ROLE OFL-LACTATE IN THE BRAINServes as an Energy Substrate for NeuronsAstrocytic derived-L-lactate has been proposed to serve as anenergy pool for neurons (40). In the presence or absenceof oxygen, glycogen in the astrocytes can be broken downinto L-lactate (30). An interesting phenomenon that manifeststhe role of L-lactate in neuronal energy metabolism is thatthe levels of glycogen in various brain regions (cortices,Frontiers in Nutrition | www.frontiersin.org9April 2022 | Volume 9 | Article 800901Cai et al.Lactate Role in Brain Functionhippocampus, brainstem, epencephalon, and hypothalamus) areconstant during the first hour of moderate-intensity exercise,while the levels of glycogen will decrease approximately 50% inthe several hours that follow. Accompanied with the decreasedglycogen levels are the elevated L-lactate levels in these regions(164). A similar phenomenon exists in the exhaustive swimmingmodel, especially in the hippocampus (165). L-lactate oxidationmetabolism is preferentially located in neurons by using the3-[13C] lactate label (166). Furthermore, when increasing theplasma L-lactate levels, the cerebrum tends to reduce glucoseutilization and prefers to utilize L-lactate by using the 1-[11C]L-lactate and [18F]fluorodeoxyglucose (FDG) for local cerebralmetabolism measurement (148). The formed L-lactate in thebrain is a crucial aerobic energy substrate that enables neuronsto endure activation (167). Another evidence for L-lactate is theenergy substrate of neurons is that the capacity of sustainingthe presynaptic transmission function in the lack of glucose.The interesting note is that the transport of L-lactate intoneurons is significant for ATP production to satisfy the energydemand under physiological or hypoxic conditions (28, 84, 168)(Figure 3).Signal Pathways Regulation as aLactormoneInadditiontobeinganenergysubstrate,L-lactateisalso suggested its role as a signal regulator molecule orneurotransmitter since the proposal of “lactormone” (22) andthe confirmation of its endogenous receptor GPR81 in micebrain (169).Influences the Intracellular NADH/NAD+ RatioThe catalytic action of LDH1 in neurons allows L-lactate toconvert into pyruvate. This process of reducing NAD+ to NADHis supposed to influence NMDAR excitability (16, 93, 170). Itis L-lactate, but not pyruvate, that generates NADH to enhancethe NMDAR activity, which increases the intracellular Ca2+and then promotes the expression of plasticity-associated genes(90). Along with the changes of redox state (NADH/NAD+ratio), L-lactate is permissible to activate the NAD+-dependenthistone deacetylase Sirtuin1 (SIRT1) (3, 171). SIRT1 deacetylatesand amplifies the activity of the peroxisome proliferator-activated receptor-gamma coactivator-1 alpha (PGC-1α), whichis known to mediate the BDNF expression (3) and mitochondrialbiogenesis (172, 173). For example, Hayek et al. find that L-lactatecan induce the expression of SITRT1 to activate the PGC1-1α,which will contribute to BDNF signals and some immediate earlygene (IEGs) as a result in the mouse neurons (3). Our experimentalso finds that a high concentration of L-lactate promotes thePGC-1-NRF2 signal axis and subsequently boosts mitochondrialbiogenesis in mouse primary hippocampal cells (107) (Figure 3).However, Lezi et al. find that 14 consecutive-day-intraperitonealinjections of 18 mM L-lactate do not affect the PCG-1α-mediatedmitochondrial biogenesis in the mouse brain (8). This may relateto the short administration time or the examination of thewhole brain.Activates Reactive Oxygen Species-Related SignalsIt is well-known that excessive reactive oxygen species (ROS)production is considered a cause of several pathologicalconditions, which are harmful to the cells. Yet, low ROSproductionmayberelevanttocellularactivityunderphysiological conditions (174). Moreover, the moderate elevationof ROS can even protect against oxidative damage through theinduction of antioxidant and detoxification enzymes (175).Hashimoto et al. first report that L-lactate is able to stimulatesome of the ROS-sensitive transcription factors’ activity in L6cells. Their research also reports that the PGC1α, known as amaster coordinator of mitochondrial biogenesis, is activated ina high concentration of L-lactate treatment which is believedto be regulated by ROS and H2O2. Then the activated PGC1αinteracts with transcription factors for mitochondrial geneexpression, including cAMP-response element-binding protein(CREB) and nuclear respiratory factor (NRF)-2 in L6 cells (105).L-lactate metabolism also potentiates the phosphorylation ofadenosine 5’-monophosphate (AMP)-activated protein kinase(AMPK) to activate PCG-1α signals for mitochondrial biogenesisand inhibits the phosphorylation of mTOR for mitophagy viamild production of ROS in skin fibroblasts (175). Recently,a similar phenomenon occurs in neurons. In SY5Y cells,20 mM L-lactate is proved to induce moderate ROS forantioxidant stress and cellular homeostasis improvement, forinstance, activating mTOR signaling for cell growth and survival,phosphatidylinositol 3-kinase (PI3-K)/protein kinase B (AKT)signaling for neuroprotection, mild endoplasmic reticulumstress (ERS) for protein homeostasis, and NRF2 signaling forROS detoxification (Figure 3). Interestingly, it is an extremelyhigh concentration of L-lactate (100 mM) but not lower orphysiological concentration of L-lactate (50 and 10 mM) exertsantiaging phenotypes when Caenorhabditis elegans suffer fromoxidative stress. Furthermore, 10 mM L-lactate suffice to resiststress and increase their lifespan (106).Noteworthy, it is not known how L-lactate is ready-wittedenough to mediate and maintain moderate ROS productionfor safeguarding cells when a high concentration of H2O2provokes excessive ROS. One of the reasons may be associatedwith the potential ability of L-lactate sensing and decreasingthe mitochondrial H2O2 production accurately to ensuremoderate ROS production (175). Another hypothesis of L-lactate mitochondrial metabolism may explain the ROS signalsactivation. Bari et al. find that L-lactation can generate theH2O2 production via the inner membrane of putative L-lactate oxidase (LOX) in the pure rat liver mitochondria(176). Hence, L-lactate metabolism in mitochondria is likelyto produce partial ROS and then activate its related signalsif there exists analogous oxidase located at the mitochondrialinner membrane in the neurons. The evidence supports thishypothesis is that a mitochondrial lactate oxidation complex(mLOC) (including MCT1 or MCT2, LDH, and COX) ispresent in the neurons discovered by Hashimoto et al. (54).However, how L-lactate metabolism in neurons producesand regulates ROS in the subsequent process remains to befurther investigated.Frontiers in Nutrition | www.frontiersin.org10April 2022 | Volume 9 | Article 800901Cai et al.Lactate Role in Brain FunctionFIGURE 3 | Graphical summary of the neurobiological role of L-lactate in the brain. On the one hand, L-lactate serves as an energy pool for neurons to facilitatemitochondrial energy production through converting into pyruvate. On the other hand, L-lactate may act as the role like the hormone, defined as lactormone. Ittriggers multiple downstream signaling cascades via influencing the intracellular NADH/NAD+ ratio, mediating the ROS generation, and activating the receptorGPR81. NADH, nicotinamide adenine dinucleotide; ROS, reactive oxygen species; GPR81, Gi-protein-coupled receptor 81.Interacts With the Receptor GPR81GPR81 is a type of G protein (Gi) coupled receptor and involvesthe metabolic process of some tissues and cells (177–179). So far,recent research has revealed that GPR81 regulates multiple signalpathways in tissues and cells, such as the extracellular regulatedprotein kinases (ERK1/2) (7, 180–182), nod-like receptor familypyrin domain-containing 3 (NLRP3)/ nuclear factor kappa-B (NF-κB) inflammation (183, 184), peroxisome proliferator-activated receptor γ (PPARγ) (185, 186) and Wnt signalingpathway (187, 188).In the brain, it expresses in the regions of the pituitary(177), hippocampus, cerebellum, and brain stem (169). Celllocalization indicates that it widely distributes at neuronalsynaptic membranes (169), implying that GPR81 may havean important role in regulating brain function. It is knownas the endogenic lactate receptor in 2008 (189, 190). Theefficiency of lactate on GPR81 has a wide range from 0.2 to30 mM (191). Furthermore, partial activation of GPR81 by lactaterequires only 0.2–1 mM (192). Therefore, lactate absolutely hasthe ability to activate the GPR81 in the physiological stateFrontiers in Nutrition | www.frontiersin.org11April 2022 | Volume 9 | Article 800901Cai et al.Lactate Role in Brain Function(human blood lactate concentration range is 0.5–2 mM) (193).Needless to say, in the conditions of intensive exercise andhypoxia, the blood lactate level can even reach 20–30 mM (194,195). In neuronal activity, the activation of GPR81 inhibits theexcitability of neurons (9, 10). In cellular energy, the astrocyticglycolysis rates accelerate ATP production to meet the energyrequirements of neurons when the GPR81 agonists stimulatethe cells (196). In terms of cerebral microangiogenesis andneurogenesis, L-lactate and exercise enhance the PI3K/AKTand ERK1/2 signaling pathway to promote VEGFA proteinexpression and vascular density in the dentate gyrus of micevia GPR81 (7). In addition, the neurogenesis effect of bothL-lactate and exercise on the ventriculosubventricular regionis dependent on the GPR81 in the mice (5). As for thelearning and memory, intraperitoneal injection of GPR81 agonist3,5-DHBA before the inhibitory avoidance task will impairthe memory of rats, while the injection after the trainingcontributes to memory consolidation. The distinct effects maybe attributed to the reasons that the activation of GPR81 inthe forebrain after training, can promote slow-wave activity(SWA) and enhance the consolidation of previous experience,but it also impairs the ongoing learning process at thesame time (80). The results suggest that GPR81 activation atdifferent stages of training may have different functions ofinfluencing learning and memory. Synaptic plasticity, inducedby appropriate synapses, is both necessary and sufficient forthe information storage of memory (197). Previous studiesrevealed that L-lactate could promote plastic plasticity-relatedproteins and genes expression via enhancing the NMDAR-mediated phosphorylation of ERK1/2 (89, 90). The interestingthing is the fact that GPR81 is upstream of the ERK1/2 (7)(Figure 3). Hence, GPR81 is likely to act as a sensor for L-lactatemetabolism in neurons and participate in L-lactate-regulatedsynaptic plasticity (198).CONCLUSIONTo summarize the current evidence, this review shows that lactateparticipates in coordinating various brain functions in both thehealthy and diseased states. The involving mechanisms are farmore complex than originally thought and further knowledgeof the lactate operating, as outlined in this article, is a dropin the bucket to our understanding of brain physiologicaland pathological mysteries. L-lactate not only serves as theenergy substrate but also acts as lactormone in the regulationof downstream cascaded signaling pathways. Considering theconsensus that lactate is a common metabolite and readilyavailable (food intake or exercise), knowledge of lactate’s role inthe brain may provide precise tactics for brain diseases involvinglactate metabolism. It also may set new insight and ideas forstudying the relationship between lactate metabolism and brainfunction improvement.AUTHOR CONTRIBUTIONSMC and HW drafted the manuscript. HS, RY, and LW assistedwith drafting the tables and figures. XX suggested valuableadvice for the article. WS and JH conceptualized the article andrevised the final version. All authors read and approved thefinal manuscript.FUNDINGThis work was sponsored by Shanghai Sailing Program(22YF1441600), the grant of the funding of Youth FundProject of Research Planning Foundation on Humanities andSocial Sciences of the Ministry of Education (20YJCZH001),China Postdoctoral Science Foundation (2021M702128), and theScientific Research Foundation of SUMHS (SSF-21-03-008).REFERENCES1. Mason S. Lactate shuttles in neuroenergetics-homeostasis, allostasis andbeyond. Front Neurosci. (2017) 11:43. doi: 10.3389/fnins.2017.000432. ToddJJ.Lactate:valuableforphysicalperformanceandmaintenanceofbrainfunctionduringexercise.BiosciHoriz.(2014). doi: 10.1093/biohorizons/hzu0013. El Hayek L, Khalifeh M, Zibara V, Abi Assaad R, Emmanuel N, Karnib N, etal. Lactate mediates the effects of exercise on learning and memory throughSIRT1-dependent activation of hippocampal brain-derived neurotrophicfactor (BDNF). J Neurosci. (2019) 39:2369–82. doi: 10.1523/JNEUROSCI.1661-18.20194. GordonGR,ChoiHB,RungtaRL,Ellis-DaviesGC,MacVicarBA.Brainmetabolismdictatesthepolarityofastrocytecontroloverarterioles.Nature.(2008)456:745–9.doi:10.1038/nature075255. Lambertus M, Overberg LT, Andersson KA, Hjelden MS, Hadzic A,Haugen OP, et al. L-lactate induces neurogenesis in the mouse ventricular-subventricular zone via the lactate receptor HCA1. Acta Physiol. (2021)231:e13587. doi: 10.1111/apha.135876. Lev-Vachnish Y, Cadury S, Rotter-Maskowitz A, Feldman N, Roichman A,Illouz T, et al. L-Lactate promotes adult hippocampal neurogenesis. FrontNeurosci. (2019) 13:403. doi: 10.3389/fnins.2019.004037. Morland C, Andersson KA, Haugen OP, Hadzic A, Kleppa L, Gille A, et al.Exercise induces cerebral VEGF and angiogenesis via the lactate receptorHCAR1. Nat Commun. (2017) 8:15557. doi: 10.1038/ncomms155578. E L, Lu J, Selfridge JE, Burns JM, Swerdlow RH. Lactate administrationreproduces specific brain and liver exercise-related changes. J Neurochem.(2013) 127:91–100. doi: 10.1111/jnc.123949. Bozzo L, Puyal J, Chatton JY. Lactate modulates the activity of primarycortical neurons through a receptor-mediated pathway. PLoS ONE. (2013)8:e71721. doi: 10.1371/journal.pone.007172110. de Castro Abrantes H, Briquet M, Schmuziger C, Restivo L, Puyal J,Rosenberg N, et al. The lactate receptor HCAR1 modulates neuronal networkactivity through the activation of galpha and gbetagamma subunits. JNeurosci. (2019) 39:4422–33. doi: 10.1523/JNEUROSCI.2092-18.201911. Herrera-Lopez G, Galvan EJ. Modulation of hippocampal excitabilityvia the hydroxycarboxylic acid receptor 1. Hippocampus. (2018) 28:557–67. doi: 10.1002/hipo.2295812. Castillo X, Rosafio K, Wyss MT, Drandarov K, Buck A, Pellerin L,et al. A probable dual mode of action for both L- and D-lactateneuroprotection in cerebral ischemia. J Cereb Blood Flow Metab. (2015)5:1561–9. doi: 10.1038/jcbfm.2015.11513. Shen Z, Jiang L, Yuan Y, Deng T, Zheng YR, Zhao YY, et al. Inhibition of Gprotein-coupled receptor 81 (GPR81) protects against ischemic brain injury.CNS Neurosci Ther. (2015) 21:271–9. doi: 10.1111/cns.12362Frontiers in Nutrition | www.frontiersin.org12April 2022 | Volume 9 | Article 800901Cai et al.Lactate Role in Brain Function14. Berthet C, Lei H, Thevenet J, Gruetter R, Magistretti PJ, Hirt L.Neuroprotective role of lactate after cerebral ischemia. J Cereb Blood FlowMetab. (2009) 29:1780–9. doi: 10.1038/jcbfm.2009.9715. Berthet C, Castillo X, Magistretti PJ, Hirt L. New evidence of neuroprotectionby lactate after transient focal cerebral ischaemia: extended benefit afterintracerebroventricular injection and efficacy of intravenous administration.Cerebrovasc Dis. (2012) 34:329–35. doi: 10.1159/00034365716. FergusonBS,RogatzkiMJ,GoodwinML,KaneDA,RightmireZ,GladdenLB.Lactatemetabolism:historicalcontext,priormisinterpretations, and current understanding. Eur J Appl Physiol. (2018)118:691–728. doi: 10.1007/s00421-017-3795-617. Kondoh Y, Kawase M, Kawakami Y, Ohmori S. Concentrations of D-lactateand its related metabolic intermediates in liver, blood, and muscle of diabeticand starved rats. Res Exp Med. (1992) 192:407–14. doi: 10.1007/BF0257629818. Ros J, Pecinska N, Beat Alessandri, Landolt H, Fillenz M. Lactate reducesglutamate-induced neurotoxicity in rat cortex. J Neurosci Res. (2001) 66:790–4. doi: 10.1002/jnr.1004319. Bouzier-SoreAK,VoisinP,CanioniP,MagistrettiPJ,PellerinL.Lactateisapreferentialoxidativeenergysubstrateoverglucose for neurons in culture. J Cereb Blood Flow Metab. (2003)23:1298–306. doi: 10.1097/01.WCB.0000091761.61714.2520. Pellerin L, Magistretti PJ. Sweet sixteen for ANLS. J Cereb Blood Flow Metab.(2012) 32:1152–66. doi: 10.1038/jcbfm.2011.14921. Sobral-Monteiro-JuniorR,MaillotP,Gatica-RojasV,AvilaWRM,de Paula AMB, Guimaraes ALS, et al. Is the “lactormone” a key-factor for exercise-related neuroplasticity? A hypothesis based on analternative lactate neurobiological pathway. Med Hypotheses. (2019)123:63–6. doi: 10.1016/j.mehy.2018.12.01322. Brooks GA. Cell-cell and intracellular lactate shuttles. J Physiol. (2009)587:5591–600. doi: 10.1113/jphysiol.2009.17835023. Gibbs ME, Hertz L. Inhibition of astrocytic energy metabolism byd-lactate exposure impairs memory. Neurochem Int. (2008) 52:1012–8. doi: 10.1016/j.neuint.2007.10.01424. Uribarri J, Oh MS, Carroll HJ. D-lactic acidosis. A review of clinicalpresentation, biochemical features, and pathophysiologic mechanism.Medicine. (1998) 77:73–82. doi: 10.1097/00005792-199803000-0000125. Robergs RA, Ghiasvand F, Parker D. Biochemistry of exercise-inducedmetabolic acidosis. Am J Physiol Regul Integr Comp Physiol. (2004)287:R502–16. doi: 10.1152/ajpregu.00114.200426. MagistrettiPJ,AllamanI.Acellularperspectiveonbrainenergymetabolismandfunctionalimaging.Neuron.(2015)86:883–901. doi: 10.1016/j.neuron.2015.03.03527. Newman LA, Korol DL, Gold PE. Lactate produced by glycogenolysisinastrocytesregulatesmemoryprocessing.PLoSONE.(2011)6:e28427. doi: 10.1371/journal.pone.002842728. Sickmann HM, Walls AB, Schousboe A, Bouman SD, WaagepetersenHS.Functionalsignificanceofbrainglycogeninsustainingglutamatergicneurotransmission.JNeurochem.(2009)109(Suppl.1):80–6. doi: 10.1111/j.1471-4159.2009.05915.x29. Gold PE, Newman LA, Scavuzzo CJ, Korol DL. Modulation of multiplememorysystems:fromneurotransmitterstometabolicsubstrates.Hippocampus. (2013) 23:1053–65. doi: 10.1002/hipo.2218230. Steinman MQ, Gao V, Alberini CM. The role of lactate-mediated metaboliccoupling between astrocytes and neurons in long-term memory formation.Front Integr Neurosci. (2016) 10:10. doi: 10.3389/fnint.2016.0001031. Matsui T, Omuro H, Liu YF, Soya M, Shima T, McEwen BS, et al.Astrocytic glycogen-derived lactate fuels the brain during exhaustive exerciseto maintain endurance capacity. Proc Natl Acad Sci. (2017) 114:6358–63. doi: 10.1073/pnas.170273911432. Aveseh M, Nikooie R, Sheibani V, Esmaeili-Mahani S. Endurance trainingincreases brain lactate uptake during hypoglycemia by up regulationof brain lactate transporters. MolCellEndocrinol. (2014) 394:29–36. doi: 10.1016/j.mce.2014.06.01933. Masaki Takimoto TH. Acute exercise increases brain region-specificexpression of MCT1, MCT2, MCT4, GLUT1, and COX IV proteins. J ApplPhysiol. (2014) 116:1238–50. doi: 10.1152/japplphysiol.01288.201334. Hashimoto T, Tsukamoto H, Takenaka S, Olesen ND, Petersen LG,Sorensen H, et al. Maintained exercise-enhanced brain executive functionrelated to cerebral lactate metabolism in men. Faseb J. (2018) 32:1417–27. doi: 10.1096/fj.201700381RR35. van Hall G, Stromstad M, Rasmussen P, Jans O, Zaar M, Gam C, et al. Bloodlactate is an important energy source for the human brain. J Cereb Blood FlowMetab. (2009) 29:1121–9. doi: 10.1038/jcbfm.2009.3536. Tubbs PK. The metabolism of D-alpha-hydroxy acids in animal tissues. AnnN Y Acad Sci. (1965) 119:920–6. doi: 10.1111/j.1749-6632.1965.tb47452.x37. Brooks GA. The science and translation of lactate shuttle theory. Cell Metab.(2018) 27:757–85. doi: 10.1016/j.cmet.2018.03.00838. Adeva-Andany M, López-Ojén M, Funcasta-Calderón R, Ameneiros-Rodríguez E, Donapetry-García C, Vila-Altesor M, et al. Comprehensivereview on lactate metabolism in human health. Mitochondrion. (2014) 17:76–100. doi: 10.1016/j.mito.2014.05.00739. Finsterwald C, Magistretti PJ, Lengacher S. Astrocytes: new targets for thetreatment of neurodegenerative diseases. Curr Pharm Des. (2015) 21:3570–81. doi: 10.2174/138161282166615071014450240. PellerinL,MagistrettiPJ.Glutamateuptakeintoastrocytesstimulatesaerobicglycolysis:Amechanismcouplingneuronalactivitytoglucoseutilization.ProcNatAcadSci.(1994)9:10625–9. doi: 10.1073/pnas.91.22.1062541. Baltazar F, Afonso J, Costa M, Granja S. Lactate beyond a wastemetabolite: metabolic affairs and signaling in malignancy. Front Oncol.(2020) 10:231. doi: 10.3389/fonc.2020.0023142. Pierre K, Pellerin L, Debernardi R, Riederer BM, Magistretti PJ. Cell-specificlocalization of monocarboxylate transporters, MCT1 and MCT2, in the adultmouse brain revealed by double immunohistochemical. Neuroscience Vol.(2000) 100:617–27. doi: 10.1016/S0306-4522(00)00294-343. Gerhart DZ, Enerson BE, Zhdankina OY, Leino RL, Drewes LR.Expression of monocarboxylate transporter MCT1 by brain endotheliumand glia in adult and suckling rats. Am J Physiol. (1997) 273:E207–213. doi: 10.1152/ajpendo.1997.273.1.E20744. HanuR,McKennaM,O’NeillA,ResneckWG,BlochRJ.Monocarboxylicacidtransporters,MCT1andMCT2,incorticalastrocytes in vitro and in vivo. Am J Physiol Cell Physiol. (2000)278:C921–30. doi: 10.1152/ajpcell.2000.278.5.C92145. Rafiki A, Boulland JL, Halestrap AP, Ottersen OP, Bergersen L. Highlydifferential expression of the monocarboxylate transporters MCT2 andMCT4 in the developing rat brain. Neuroscience. (2003) 122:677–88. doi: 10.1016/j.neuroscience.2003.08.04046. Bergersen LH. Is lactate food for neurons? Comparison of monocarboxylatetransporter subtypes in brain and muscle. Neuroscience. (2007) 145:11–9. doi: 10.1016/j.neuroscience.2006.11.06247. Dienel GA. The metabolic trinity, glucose–glycogen–lactate, links astrocytesand neurons in brain energetics, signaling, memory, and gene expression.Neurosci Lett. (2017) 637:18–25. doi: 10.1016/j.neulet.2015.02.05248. Pierre K, Pellerin L. Monocarboxylate transporters in the central nervoussystem: distribution, regulation and function. J Neurochem. (2005) 94:1–14. doi: 10.1111/j.1471-4159.2005.03168.x49. BrickerDK,TaylorEB,SchellJC,OrsakT,BoutronA,ChenYC, et al. A mitochondrial pyruvate carrier required for pyruvateuptakeinyeast,Drosophila,andhumans.Science.(2012)337:96–100. doi: 10.1126/science.121809950. Thomas AP, Halestrap AP. Identification of the protein responsible forpyruvate transport into rat liver and heart mitochondria by specificlabellingwith[3H]N-phenylmaleimide.BiochemJ.(1981)196:471–9. doi: 10.1042/bj196047151. Brooks GA, Brown MA, Butz CE, Sicurello JP, Dubouchaud H. Cardiac andskeletal muscle mitochondria have a monocarboxylate transporter MCT1. JAppl Physiol (1985). (1999) 87:1713–8. doi: 10.1152/jappl.1999.87.5.171352. ButzCE,McClellandGB,BrooksGA.MCT1confirmedinratstriatedmusclemitochondria.JApplPhysiol(1985).(2004)97:1059–66. doi: 10.1152/japplphysiol.00009.200453. Hashimoto T, Hussien R, Brooks GA. Colocalization of MCT1, CD147, andLDH in mitochondrial inner membrane of L6 muscle cells: evidence of amitochondrial lactate oxidation complex. Am J Physiol Endocrinol Metab.(2006) 290:E1237–44. doi: 10.1152/ajpendo.00594.200554. Hashimoto T, Hussien R, Cho HS, Kaufer D, Brooks GA. Evidence forthe mitochondrial lactate oxidation complex in rat neurons: demonstrationFrontiers in Nutrition | www.frontiersin.org13April 2022 | Volume 9 | Article 800901Cai et al.Lactate Role in Brain Functionof an essential component of brain lactate shuttles. PLoS ONE. (2008)3:e2915. doi: 10.1371/journal.pone.000291555. Passarella S, de Bari L, Valenti D, Pizzuto R, Paventi G, Atlante A.Mitochondria and L-lactate metabolism. FEBS Lett. (2008) 582:3569–76. doi: 10.1016/j.febslet.2008.09.04256. Stainsby WN, Brooks GA. Control of lactic acid metabolism in contractingmusclesandduringexercise.ExercSportSciRev.(1990)18:29–63. doi: 10.1249/00003677-199001000-0000557. Hashimoto T, Brooks GA. Mitochondrial lactate oxidation complex and anadaptive role for lactate production. Med Sci Sports Exerc. (2008) 40:486–94. doi: 10.1249/MSS.0b013e31815fcb0458. Brooks GA. Intra- and extra-cellular lactate shuttles. Med Sci Sports Exerc.(2000) 32:790–9. doi: 10.1097/00005768-200004000-0001159. RogatzkiMJ,FergusonBS,GoodwinML,GladdenLB.Lactateisalwaystheendproductofglycolysis.FrontNeurosci.(2015)9:22. doi: 10.3389/fnins.2015.0002260. Lemire J, Mailloux RJ, Appanna VD. Mitochondrial lactate dehydrogenase isinvolved in oxidative-energy metabolism in human astrocytoma cells (CCF-STTG1). PLoS ONE. (2008) 3:e1550. doi: 10.1371/journal.pone.000155061. Atlante A, de Bari L, Bobba A, Marra E, Passarella S. TransportandmetabolismofL-lactateoccurinmitochondriafromcerebellar granule cells and are modified in cells undergoing lowpotassiumdependentapoptosis.BiochimBiophysActa.(2007)1767:1285–99. doi: 10.1016/j.bbabio.2007.08.00362. Tekkök SB, Brown AM, Westenbroek R, Pellerin L, Ransom BR.Transfer of glycogen-derived lactate from astrocytes to axons via specificmonocarboxylate transporters supports mouse optic nerve activity. JNeurosci Res. (2005) 81:644–52. doi: 10.1002/jnr.2057363. Ling B, Peng F, Alcorn J, Lohmann K, Bandy B, Zello GA. D-Lactate alteredmitochondrial energy production in rat brain and heart but not liver. NutrMetab. (2012) 9:6. doi: 10.1186/1743-7075-9-664. Connor H, Woods HF, Ledingham JGG. Comparison of the kinetics andutilisation of D(-)-and L(+)-sodium lactate in normal man. Ann Nutr Metab.(1983) 27:481–7. doi: 10.1159/00017672365. JacksonVN,HalestrapAP.Thekinetics,substrate,andinhibitorspecificity of the monocarboxylate (Lactate) transporter of rat livercellsdeterminedusingthefluorescentintracellularpHindicator,2′,7′-Bis(carboxyethyl)-5(6)-carboxyfluorescein.JBiolChem.(1996)271:861–8. doi: 10.1074/jbc.271.2.86166. Halestrap AP. The monocarboxylate transporter family–Structure andfunctional characterization. IUBMB Life. (2012) 64:1–9. doi: 10.1002/iub.57367. Flick MJ, Konieczny SF. Identification of putative mammalian D-lactatedehydrogenase enzymes. Biochem Biophys Res Commun. (2002) 295:910–6. doi: 10.1016/S0006-291X(02)00768-468. Liu S. Practical implications of lactate and pyruvate metabolism by lactic acidbacteria in food and beverage fermentations. Int J Food Microbiol. (2003)83:115–31. doi: 10.1016/S0168-1605(02)00366-569. Sousa MA, Rama GR, Volken de Souza CF, Granada CE. Acid lacticlactobacilli as a biotechnological toll to improve food quality and humanhealth. Biotechnol Prog. (2020) 36:e2937. doi: 10.1002/btpr.293770. Reid SNS, Park JH, Kim Y, Kwak YS, Jeon BH. In vitro and in vivo effects offermented oyster-derived lactate on exercise endurance indicators in mice.Int J Environ Res Public Health. (2020). 17:8811. doi: 10.3390/ijerph1723881171. Teramae H, Yoshikawa T, Inoue R, Ushida K, Takebe K, Nio-KobayashiJ, et al. The cellular expression of SMCT2 and its comparison with othertransporters for monocarboxylates in the mouse digestive tract. Biomed Res.(2010) 31:239–49. doi: 10.2220/biomedres.31.23972. Brooks GA, Arevalo JA, Osmond AD, Leija RG, Curl CC, Tovar AP.Lactate in contemporary biology: a phoenix risen. J Physiol. (2022) 600:1229–51. doi: 10.1113/JP28095573. Lecoultre V, Benoit R, Carrel G, Schutz Y, Millet GP, Tappy L, et al.Fructose and glucose co-ingestion during prolonged exercise increaseslactate and glucose fluxes and oxidation compared with an equimolarintake of glucose. Am J Clin Nutr. (2010) 92:1071–9. doi: 10.3945/ajcn.2010.2956674. Theytaz F, de Giorgi S, Hodson L, Stefanoni N, Rey V, Schneiter P, et al.Metabolic fate of fructose ingested with and without glucose in a mixed meal.Nutrients. (2014) 6:2632–49. doi: 10.3390/nu607263275. Brooks GA. The precious few grams of glucose during exercise. Int J Mol Sci.(2020) 21:5733. doi: 10.3390/ijms2116573376. Brooks GA, Martin NA. Cerebral metabolism following traumatic braininjury: new discoveries with implications for treatment. Front Neurosci.(2014) 8:408. doi: 10.3389/fnins.2014.0040877. Suzuki A, Stern SA, Bozdagi O, Huntley GW, Walker RH, Magistretti PJ,et al. Astrocyte-neuron lactate transport is required for long-term memoryformation. Cell. (2011) 144:810–23. doi: 10.1016/j.cell.2011.02.01878. Shima T, Matsui T, Jesmin S, Okamoto M, Soya M, Inoue K, et al.Moderate exercise ameliorates dysregulated hippocampal glycometabolismand memory function in a rat model of type 2 diabetes. Diabetologia. (2017)60:597–606. doi: 10.1007/s00125-016-4164-479. Harris RA, Lone A, Lim H, Martinez F, Frame AK, Scholl TJ, et al. Aerobicglycolysis is required for spatial memory acquisition but not memoryretrieval in mice. eNeuro. (2019). doi: 10.1523/ENEURO.0389-18.201980. Scavuzzo CJ, Rakotovao I, Dickson CT. Differential effects of L- and D-lactateon memory encoding and consolidation: Potential role of HCAR1 signaling.Neurobiol Learn Mem. (2020) 168:107151. doi: 10.1016/j.nlm.2019.10715181. Zhang Y, Xue Y, Meng S, Luo Y, Liang J, Li J, et al. Inhibition of lactatetransport erases drug memory and prevents drug relapse. Biol Psychiatry.(2016) 79:928–9. doi: 10.1016/j.biopsych.2015.07.00782. Tadi M, Allaman I, Lengacher S, Grenningloh G, Magistretti PJ. Learning-induced gene expression in the hippocampus reveals a role of neuron -astrocyte metabolic coupling in long term memory. PLoS ONE. (2015)10:e0141568. doi: 10.1371/journal.pone.014156883. Alberini CM, Cruz E, Descalzi G, Bessières B, Gao V. Astrocyte glycogen andlactate: New insights into learning and memory mechanisms. Glia. (2018)66:1244–62. doi: 10.1002/glia.2325084. Lucas SJ, Michel CB, Marra V, Smalley JL, Hennig MH, Graham BP, et al.Glucose and lactate as metabolic constraints on presynaptic transmission atan excitatory synapse. J Physiol. (2018) 596:1699–721. doi: 10.1113/JP27510785. Tang F, Lane S, Korsak A, Paton JF, Gourine AV, Kasparov S, et al. Lactate-mediated glia-neuronal signalling in the mammalian brain. Nat Commun.(2014) 5:3284. doi: 10.1038/ncomms428486. Herrera-Lo’pez G, Griego E, Galva’n EJ. Lactate induces synapse-specificpotentiation on CA3 pyramidal cells of rat hippocampus. PLoS ONE. (2020)15:e0242309. doi: 10.1371/journal.pone.024230987. Schurr A, West CA, Rigor BM. Lactate-supported synaptic functionin the rat hippocampal slice preparation. Science. (1988) 240: 1326–8. doi: 10.1126/science.337581788. BrownAM,TekkökSB,RansomBR.Glycogenregulationandfunctionalroleinmousewhitematter.JPhysiol.(2003)549:501–12. doi: 10.1113/jphysiol.2003.04241689. Margineanu MB, Mahmood H, Fiumelli H, Magistretti PJ. L-lactateregulates the expression of synaptic plasticity and neuroprotection genesin cortical neurons: a transcriptome analysis. Front Mol Neurosci. (2018)11:375. doi: 10.3389/fnmol.2018.0037590. Yang Y, Ruchti E, Petit JM, Jourdain P, Magistretti PJ. Lactate promotesplasticity gene expression by potentiating NMDA signaling in neurons. ProcNatl Acad Sci U S A. (2014) 111:12228–33. doi: 10.1073/pnas.132291211191. Baker KD, Edwards TM. D-Lactate inhibition of memory in a single trialdiscrimination avoidance task in the young chick. Neurobiol Learn Mem.(2007) 88:269–76. doi: 10.1016/j.nlm.2007.06.00492. Ragozzino ME, Hellems K, Robert C, Lennartz, Gold PE. Pyruvate infusionsinto the septal area attenuate spontaneous alternation impairments inducedby intraseptal morphine injections. Behav Neurosci. (1995) 109:T074–1080. doi: 10.1037/0735-7044.109.6.107493. Jourdain P, Allaman I, Rothenfusser K, Fiumelli H, Marquet P, MagistrettiPJ. L-Lactate protects neurons against excitotoxicity: implication of an ATP-mediated signaling cascade. Sci Rep. (2016) 6:21250. doi: 10.1038/srep2125094. Ross JM, Oberg J, Brene S, Coppotelli G, Terzioglu M, PernoldK, et al. High brain lactate is a hallmark of aging and causedbyashiftinthelactatedehydrogenaseA/Bratio.ProcNatlAcadSciUSA.(2010)107:20087–92.doi:10.1073/pnas.100818910795. langer MB, Allaman I, Magistretti PJ. Brain energy metabolism_ focuson astrocyte neuron metabolic cooperation. Cell Metab. (2011) 14:724–38. doi: 10.1016/j.cmet.2011.08.016Frontiers in Nutrition | www.frontiersin.org14April 2022 | Volume 9 | Article 800901Cai et al.Lactate Role in Brain Function96. Dienel GA, McKenna MC. A dogma-breaking concept: glutamate oxidationin astrocytes is the source of lactate during aerobic glycolysis in restingsubjects. J Neurochem. (2014) 131:395–8. doi: 10.1111/jnc.1283597. Dienel GA. Brain lactate metabolism: the discoveries and the controversies.J Cereb Blood Flow Metab. (2011) 32:1107–38. doi: 10.1038/jcbfm.2011.17598. Vafaee MS, Vang K, Bergersen LH, Gjedde A. Oxygen consumption andblood flow coupling in human motor cortex during intense finger tapping:implication for a role of lactate. J Cereb Blood Flow Metab. (2012) 32:1859–68. doi: 10.1038/jcbfm.2012.8999. Nishijima T, Okamoto M, Matsui T, Kita I, Soya H. HippocampalfunctionalhyperemiamediatedbyNMDAreceptor/NOsignalinginratsduringmildexercise.JApplPhysiol(1985).(2012)54:197–203. doi: 10.1152/japplphysiol.00763.2011100. Hein TW, Xu W, Kuo L. Dilation of retinal arterioles in response to lactate:role of nitric oxide, guanylyl cyclase, and ATP-sensitive potassium channels.Invest Ophthalmol Vis Sci. (2006) 47:693–9. doi: 10.1167/iovs.05-1224101. Falkowska A, Gutowska I, Goschorska M, Nowacki P, Chlubek D,Baranowska-Bosiacka I. Energy metabolism of the brain, includingthe cooperation between astrocytes and neurons, especially in thecontext of glycogen metabolism. Int J Mol Sci. (2015) 16:25959–81. doi: 10.3390/ijms161125939102. Yamanishi S, Katsumura K, Kobayashi T, Puro DG. Extracellular lactate as adynamic vasoactive signal in the rat retinal microvasculature. Am J PhysiolHeart Circ Physiol. (2006) 290:H925–934. doi: 10.1152/ajpheart.01012.2005103. Wojtowicz JM, Kee N. BrdU assay for neurogenesis in rodents. Nat Protoc.(2006) 1:1399–405. doi: 10.1038/nprot.2006.224104. Roger AJ, Munoz-Gomez SA, Kamikawa R. The origin and diversification ofmitochondria. Curr Biol. (2017) 27:R1177–92. doi: 10.1016/j.cub.2017.09.015105. HashimotoT,HussienR,OommenS,GohilK,BrooksGA.LactatesensitivetranscriptionfactornetworkinL6cells:activation of MCT1 and mitochondrial biogenesis. FASEB J. (2007)21:2602–12. doi: 10.1096/fj.07-8174com106. Tauffenberger A, Fiumelli H, Almustafa S, Magistretti PJ. Lactate andpyruvate promote oxidative stress resistance through hormetic ROSsignaling. Cell Death Dis. (2019) 10:653. doi: 10.1038/s41419-019-1877-6107. Hu J, Cai M, Shang Q, Li Z, Feng Y, Liu B, et al. Elevated lactate byhigh-intensity interval training regulates the hippocampal bdnf expressionand the mitochondrial quality control system. Front Physiol. (2021)12:629914. doi: 10.3389/fphys.2021.629914108. Lauritzen F, Eid T, Bergersen LH. Monocarboxylate transporters in temporallobe epilepsy: roles of lactate and ketogenic diet. Brain Struct Funct. (2015)220:1–12. doi: 10.1007/s00429-013-0672-x109. Zhai X, Li J, Li L, Sun Y, Zhang X, Xue Y, et al. L-lactate preconditioningpromotes plasticity-related proteins expression and reduces neurologicaldeficits by potentiating GPR81 signaling in rat traumatic brain injury model.Brain Res. (2020) 1746:146945. doi: 10.1016/j.brainres.2020.146945110. Bouzat P, Sala N, Suys T, Zerlauth JB, Marques-Vidal P, Feihl F,et al. Cerebral metabolic effects of exogenous lactate supplementationon the injured human brain. Intensive Care Med. (2014) 40:412–21. doi: 10.1007/s00134-013-3203-6111. Alessandri B, Schwandt E, Kamada Y, Nagata M, Heimann A, Kempski O.The neuroprotective effect of lactate is not due to improved glutamate uptakeafter controlled cortical impact in rats. J Neurotrauma. (2012) 29:2181–91. doi: 10.1089/neu.2011.2067112. Dienel GA. Lactate shuttling and lactate use as fuel after traumatic braininjury: metabolic considerations. J Cereb Blood Flow Metab. (2014) 34:1736–48. doi: 10.1038/jcbfm.2014.153113. Glenn TC, Martin NA, Horning MA, McArthur DL, Hovda DA, Vespa P,et al. Lactate: brain fuel in human traumatic brain injury: a comparisonwith normal healthy control subjects. J Neurotrauma. (2015) 32:820–32. doi: 10.1089/neu.2014.3483114. GlennTC,MartinNA,McArthurDL,HovdaDA,VespaP,JohnsonML,etal.Endogenousnutritivesupportaftertraumaticbraininjury:peripherallactateproductionforglucosesupplyviagluconeogenesis. J Neurotrauma. (2015) 32:811–19. doi: 10.1089/neu.2014.3482115. Ma Y, Liu T, Fu J, Fu S, Hu C, Sun B, et al. Lactobacillus acidophilus exertsneuroprotective effects in mice with traumatic brain injury. J Nutr. (2019)149:1543–52. doi: 10.1093/jn/nxz105116. Li H, Sun J, Du J, Wang F, Fang R, Yu C, et al. Clostridiumbutyricum exerts a neuroprotective effect in a mouse model of traumaticbrain injury via the gut-brain axis. Neurogastroenterol Motil. (2018)30:e13260. doi: 10.1111/nmo.13260117. Mijnhout GS, Scheltens P, Diamant M, Biessels GJ, Wessels AM, Simsek S, etal. Diabetic encephalopathy: a concept in need of a definition. Diabetologia.(2006) 49:1447–8. doi: 10.1007/s00125-006-0221-8118. McCrimmon RJ, Ryan CM, Frier BM. Diabetes and cognitive dysfunction.Lancet. (2012) 379:2291–9. doi: 10.1016/S0140-6736(12)60360-2119. Biessels GJ, van der Heide LP, Kamal A, Bleys RL, Gispen WH. Ageing anddiabetes implications for brain function. Eur J Pharmacol. (2002) 441:1–14. doi: 10.1016/S0014-2999(02)01486-3120. Yi SS. Effects of exercise on brain functions in diabetic animal models. WorldJ Diabetes. (2015) 6:583–97. doi: 10.4239/wjd.v6.i4.583121. Fazeli SA. Neuroprotection in diabetic encephalopathy. Neurodegener Dis.(2009) 6:213–8. doi: 10.1159/000278694122. Wiegers EC, Rooijackers HM, Tack CJ, Heerschap A, de Galan BE, vander Graaf M. Brain lactate concentration falls in response to hypoglycemiain patients with type 1 diabetes and impaired awareness of hypoglycemia.Diabetes. (2016) 65:1601–5. doi: 10.2337/db16-0068123. Wiegers EC, Rooijackers HM, Tack2 CJ, Groenewoud HJMM, HeerschapA, Galan BEd, et al. Effect of exercise-induced lactate elevation onbrain lactate levels during hypoglycemia in patients with type 1 diabetesand impaired awareness of hypoglycemia. Diabetes. (2017) 66:3105–10. doi: 10.2337/db17-0794124. King P, Kong MF, Parkin H, MacDonald IA, Barber C, TattersallRB.Intravenouslactatepreventscerebraldysfunctionduringhypoglycaemia in insulin-dependent diabetes mellitus. Clin Sci. (1998)94:157–63. doi: 10.1042/cs0940157125. Maran A, Crepaldi C, Trupiani S, Lucca T, Jori E, Macdonald IA, et al. Brainfunction rescue effect of lactate following hypoglycaemia is not an adaptationprocess in both normal and type I diabetic subjects. Diabetologia. (2000)43:733–41. doi: 10.1007/s001250051371126. Wiegers EC, Rooijackers HM, Tack CJ, Philips BW, Heerschap A, van derGraaf M, et al. Effect of lactate administration on brain lactate levels duringhypoglycemia in patients with type 1 diabetes. J Cereb Blood Flow Metab.(2019) 39:1974–82. doi: 10.1177/0271678X18775884127. Chan O, Paranjape S, Horblitt A, Zhu W, Sherwin RS. Lactate-inducedrelease of GABA in the ventromedial hypothalamus contributes tocounterregulatory failure in recurrent hypoglycemia and diabetes. Diabetes.(2013) 62:4239–46. doi: 10.2337/db13-0770128. Kobayashi R, Maruoka J, Norimoto H, Ikegaya Y, Kume K, Ohsawa M.Involvement of l-lactate in hippocampal dysfunction of type I diabetes. JPharmacol Sci. (2019) 141:79–82. doi: 10.1016/j.jphs.2019.09.004129. Mason GF, Petersen KF, Lebon V, Rothman DL, Shulman GI. increased brainmonocarboxylic acid transport and utilization in type 1 diabetes. Diabetes.(2006) 55:929–34. doi: 10.2337/diabetes.55.04.06.db05-1325130. Hackett MJ, Hollings A, Majimbi M, Brook E, Cochran B, Giles C, etal. Multimodal imaging analyses of brain hippocampal formation revealreduced Cu and lipid content and increased lactate content in non-insulin-dependent diabetic mice. ACS Chem Neurosci. (2019) 10:2533–40. doi: 10.1021/acschemneuro.9b00039131. Li J, Liu B, Cai M, Lin X, Lou S. Glucose metabolic alterations inhippocampus of diabetes mellitus rats and the regulation of aerobicexercise. Behav Brain Res. (2019) 364:447–56. doi: 10.1016/j.bbr.2017.11.001132. Choi Y-S, Song JE, Lee JE, Kim E, Kim CH, Kim D-H, et al.Hyperpolarized [1-13C] lactate flux increased in the hippocampal regionin diabetic mice. Molecular Brain. (2019) 12:88. doi: 10.1186/s13041-019-0505-9133. De Feyter HM, Mason GF, Shulman GI, Rothman DL, Petersen KF. Increasedbrain lactate concentrations without increased lactate oxidation duringhypoglycemia in type 1 diabetic individuals. Diabetes. (2013) 62:3075–80. doi: 10.2337/db13-0313Frontiers in Nutrition | www.frontiersin.org15April 2022 | Volume 9 | Article 800901Cai et al.Lactate Role in Brain Function134. Zhao L, Dong M, Ren M, Li C, Zheng H, Gao H. Metabolomicanalysis identifies lactate as an important pathogenic factor in diabetes-associated cognitive decline rats. Mol Cell Proteomics. (2018) 17:2335–46. doi: 10.1074/mcp.RA118.000690135. Wang D, Zhao L, Zheng H, Dong M, Pan L, Zhang X, et al. Time-dependent lactate production and amino acid utilization in culturedastrocytesunderhighglucoseexposure.MolNeurobiol.(2018)55:1112. doi: 10.1007/s12035-016-0360-y136. Tapeinos C, Battaglini M, Ciofani G. Advances in the design of solid lipidnanoparticles and nanostructured lipid carriers for targeting brain diseases.J Control Release. (2017) 264:306–32. doi: 10.1016/j.jconrel.2017.08.033137. Yin F, Sancheti H, Patil I, Cadenas E. Energy metabolism and inflammationin brain aging and Alzheimer’s disease. Free Radic Biol Med. (2016) 100:108–22. doi: 10.1016/j.freeradbiomed.2016.04.200138. Lu WT, Sun SQ, Li Y, Xu SY, Gan SW, Xu J, et al. Curcumin amelioratesmemory deficits by enhancing lactate content and MCT2 expression inAPP/PS1 transgenic mouse model of Alzheimer’s disease. Anat Rec. (2019)302:332–8. doi: 10.1002/ar.23969139. Lu W, Huang J, Sun S, Huang S, Gan S, Xu J, et al. Changes in lactatecontent and monocarboxylate transporter 2 expression in Abeta(2)(5)(-)(3)(5)-treated rat model of Alzheimer’s disease. Neurol Sci. (2015) 36:871–6. doi: 10.1007/s10072-015-2087-3140. Zhang M, Cheng X, Dang R, Zhang W, Zhang J, Yao Z. Lactate deficit inan alzheimer disease mouse model: the relationship with neuronal damage. JNeuropathol Exp Neurol. (2018) 77:1163–76. doi: 10.1093/jnen/nly102141. Datta S, Chakrabarti N. Age related rise in lactate and its correlation withlactate dehydrogenase (LDH) status in post-mitochondrial fractions isolatedfrom different regions of brain in mice. Neurochem Int. (2018) 118:23–33. doi: 10.1016/j.neuint.2018.04.007142. Long DM, Frame AK, Reardon PN, Cumming RC, Hendrix DA,Kretzschmar D, et al. Lactate dehydrogenase expression modulateslongevity and neurodegeneration in Drosophila melanogaster. Aging. (2020)12:10041–58. doi: 10.18632/aging.103373143. Mullins R, Reiter D, Kapogiannis D. Magnetic resonance spectroscopyreveals abnormalities of glucose metabolism in the Alzheimer’s brain. AnnClin Transl Neurol. (2018) 5:262–72. doi: 10.1002/acn3.530144. de Bari L, Atlante A, Guaragnella N, Principato G, Passarella S. D-Lactatetransport and metabolism in rat liver mitochondria. Biochem J. (2002)365:391–403. doi: 10.1042/bj20020139145. de Bari L, Atlante A, Armeni T, Kalapos MP. Synthesis and metabolismof methylglyoxal, S-D-lactoylglutathione and D-lactate in cancer andAlzheimer’s disease. Exploring the crossroad of eternal youth and prematureaging. Ageing Res Rev. (2019) 53:100915. doi: 10.1016/j.arr.2019.100915146. MaidaCD,NorritoRL,DaidoneM,TuttolomondoA,PintoA.Neuroinflammatory mechanisms in ischemic stroke: focus on cardioembolicstroke, background, and therapeutic approaches. Int J Mol Sci. (2020)21:6454. doi: 10.3390/ijms21186454147. Hyacinthe JN, Buscemi L, Le TP, Lepore M, Hirt L, MishkovskyM. Evaluating the potential of hyperpolarised [1-(13)C] L-lactate asa neuroprotectant metabolic biosensor for stroke. Sci Rep. (2020)10:5507. doi: 10.1038/s41598-020-62319-x148. Wyss MT, Jolivet R, Buck A, Magistretti PJ, Weber B. In vivo evidencefor lactate as a neuronal energy source. J Neurosci. (2011) 31:7477–85. doi: 10.1523/JNEUROSCI.0415-11.2011149. SchurrA,PayneRS,MillerJJ,TsengMT,RigorBM.Rigor.Blockadeoflactatetransportexacerbatesdelayedneuronaldamagein__aratmodelofcerebralischemia.BrainRes.(2001)895:268–72. doi: 10.1016/S0006-8993(01)02082-0150. Buscemi L, Blochet C, Magistretti PJ, Hirt L. Hydroxycarboxylic acidreceptor 1 and neuroprotection in a mouse model of cerebral ischemia-reperfusion. Front Physiol. (2021) 12:689239. doi: 10.3389/fphys.2021.689239151. Moro N, Ghavim S, Harris NG, Hovda DA, Sutton RL. Glucoseadministration after traumatic brain injury improves cerebral metabolismand reduces secondary neuronal injury. Brain Res. (2013) 1535:124–36. doi: 10.1016/j.brainres.2013.08.044152. BergsneiderM,HovdaDA,ShalmonE,KellyDF,VespaPM,MartinNA,etal.Cerebralhyperglycolysisfollowingseveretraumatic brain injury in humans: a positron emission tomographystudy.JNeurosurg.(1997)86:241–51.doi:10.3171/jns.1997.86.2.0241153. Glenn TC, Kelly DF, Boscardin WJ, McArthur DL, Vespa P, Oertel M, et al.Energy dysfunction as a predictor of outcome after moderate or severe headinjury: indices of oxygen, glucose, and lactate metabolism. J Cereb Blood FlowMetab. (2003) 23:1239–50. doi: 10.1097/01.WCB.0000089833.23606.7F154. Jalloh I, Helmy A, Shannon RJ, Gallagher CN, Menon DK, CarpenterKL, et al. Lactate uptake by the injured human brain: evidence from anarteriovenous gradient and cerebral microdialysis study. J Neurotrauma.(2013) 30:2031–7. doi: 10.1089/neu.2013.2947155. CarteronL,SolariD,PatetC,QuintardH,MirozJP,BlochJ,et al. Hypertonic lactate to improve cerebral perfusion and glucoseavailability after acute brain injury. Crit Care Med. (2018) 46:1649–55. doi: 10.1097/CCM.0000000000003274156. Zhou J, Burns MP, Huynh L, Villapol S, Taub DD, Saavedra JM,etal.Temporalchangesincorticalandhippocampalexpressionofgenesimportantforbrainglucosemetabolismfollowingcontrolled cortical impact injury in mice. Front Endocrinol. (2017)8:231. doi: 10.3389/fendo.2017.00231157. Betjemann JP, Lowenstein DH. Status epilepticus in adults. Lancet Neurol.(2015) 14:615–24. doi: 10.1016/S1474-4422(15)00042-3158. Thoresen M, Hallstrom A, Whitelaw A, Puka-Sundvall M, Loberg EM,Satas S, et al. Lactate and pyruvate changes in the cerebral gray and whitematter during posthypoxic seizures in newborn pigs. Pediatr Res. (1998)44:746–54. doi: 10.1203/00006450-199811000-00018159. Mariani CL, Nye CJ, Ruterbories L, Tokarz DA, Green L, Lau J, et al.Cerebrospinal fluid lactate concentrations in dogs with seizure disorders. JVet Intern Med. (2020) 34:2562–70. doi: 10.1111/jvim.15953160. Slais K, Vorisek I, Zoremba N, Homola A, Dmytrenko L, SykovaE. Brain metabolism and diffusion in the rat cerebral cortex duringpilocarpine-induced status epilepticus. Exp Neurol. (2008) 209:145–54. doi: 10.1016/j.expneurol.2007.09.008161. DuringMJ,FriedI,LeoneP,KatzA,SpencerDD.Directmeasurementofextracellularlactateinthehumanhippocampusduringspontaneousseizures.JNeurochem.(1994)62:2356–61. doi: 10.1046/j.1471-4159.1994.62062356.x162. Cavus I, Kasoff WS, Cassaday MP, Jacob R, Gueorguieva R, Sherwin RS, et al.Extracellular metabolites in the cortex and hippocampus of epileptic patients.Ann Neurol. (2005) 57:226–35. doi: 10.1002/ana.20380163. Jorwal P, Sikdar SK. Lactate reduces epileptiform activity through HCA1and GIRK channel activation in rat subicular neurons in an in vitro model.Epilepsia. (2019) 60:2370–85. doi: 10.1111/epi.16389164. Matsui T, Soya S, Okamoto M, Ichitani Y, Kawanaka K, Soya H. Brainglycogen decreases during prolonged exercise. J Physiol. (2011) 589:3383–93. doi: 10.1113/jphysiol.2011.203570165. MatsuiT,SoyaS,KawanakaK,SoyaH.Brainglycogendecreasesduringintenseexercisewithouthypoglycemia:thepossibleinvolvementofserotonin.NeurochemRes.(2015)40:1333–40. doi: 10.1007/s11064-015-1594-1166. Bouzier AK, Thiaudiere E, Biran M, Rouland R, Canioni P, Merle M.The metabolism of [3-(13)C]lactate in the rat brain is specific of apyruvate carboxylase-deprived compartment. J Neurochem. (2000) 75:480–6. doi: 10.1046/j.1471-4159.2000.0750480.x167. Schurr A, Miller JJ, Payne RS, Rigor BM. An increase in lactate outputby brain tissue serves to meet the energy needs of glutamateactivatedneurons. J Neurosci. (1999) 19:34–9. doi: 10.1523/JNEUROSCI.19-01-00034.1999168. Genc S, Kurnaz IA, Ozilgen M. Astrocyte-neuron lactate shuttle mayboost more ATP supply to the neuron under hypoxic conditions–insilico study supported by in vitro expression data. BMC Syst Biol. (2011)5:162. doi: 10.1186/1752-0509-5-162169. Lauritzen KH, Morland C, Puchades M, Holm-Hansen S, HagelinEM, Lauritzen F, et al. Lactate receptor sites link neurotransmission,neurovascular coupling, and brain energy metabolism. Cereb Cortex. (2014)24:2784–95. doi: 10.1093/cercor/bht136170. Magistretti PJ, Allaman I. Lactate in the brain: from metabolic end-product to signalling molecule. NatRevNeurosci. (2018) 19:235–49. doi: 10.1038/nrn.2018.19Frontiers in Nutrition | www.frontiersin.org16April 2022 | Volume 9 | Article 800901Cai et al.Lactate Role in Brain Function171. Brooks GA. Lactate as a fulcrum of metabolism. Redox Biol. (2020)35:101454. doi: 10.1016/j.redox.2020.101454172. Fernandez-MarcosPJ,AuwerxJ.RegulationofPGC-1α,anodalregulatorofmitochondrialbiogenesis.AmJClinNutr.(2011)93:884S−90S. doi: 10.3945/ajcn.110.001917173. Gureev AP, Shaforostova EA, Popov VN. Regulation of mitochondrialbiogenesis as a way for active longevity: interaction between theNrf2andPGC-1alphasignalingpathways.FrontGenet.(2019)10:435. doi: 10.3389/fgene.2019.00435174. Galardo MaN, Regueira M, Riera MaF, Pellizzari EH, Cigorraga SB, MeroniSB. Lactate regulates rat male germ cell function through reactive oxygenspecies. PLoS ONE. (2014) 9:e88024. doi: 10.1371/journal.pone.0088024175. Zelenka J, Dvorak A, Alan L. L-lactate protects skin fibroblasts againstaging-associated mitochondrial dysfunction via mitohormesis. Oxid MedCell Longev. (2015) 2015:351698. doi: 10.1155/2015/351698176. Bari LD, Valenti D, Atlante A, Passarella S. L-Lactate generates hydrogenperoxide in purified rat liver mitochondria due to the putative L-lactate oxidase localized in the intermembrane space. FEBS Letters. (2010)584:2285–90. doi: 10.1016/j.febslet.2010.03.038177. Ge H, Weiszmann J, Reagan JD, Gupte J, Baribault H, Gyuris T, et al.Elucidation of signaling and functional activities of an orphan GPCR,GPR81. J Lipid Res. (2008) 49:797–803. doi: 10.1194/jlr.M700513-JLR200178. AhmedK,TunaruS,TangC,MullerM,GilleA,SassmannA,etal.Anautocrinelactateloopmediatesinsulin-dependentinhibitionoflipolysisthroughGPR81.CellMetab.(2010)11:311–19. doi: 10.1016/j.cmet.2010.02.012179. Brooks GA. Role of the heart in lactate shuttling. Front Nutr. (2021)8:663560. doi: 10.3389/fnut.2021.663560180. Li G, Wang HQ, Wang LH, Chen RP, Liu JP. Distinct pathways ofERK1/2 activation by hydroxycarboxylic acid receptor-1. PLoS ONE. (2014)9:e93041. doi: 10.1371/journal.pone.0093041181. Ohno Y, Oyama A, Kaneko H, Egawa T, Yokoyama S, Sugiura T, et al. Lactateincreases myotube diameter via activation of MEK/ERK pathway in C2C12cells. Acta Physiol. (2018) 223:e13042. doi: 10.1111/apha.13042182. Wu Y, Wang M, Zhang K, Li Y, Xu M, Tang S, et al. Lactate enhancedthe effect of parathyroid hormone on osteoblast differentiation via GPR81-PKC-Akt signaling. Biochem Biophys Res Commun. (2018) 503:737–43. doi: 10.1016/j.bbrc.2018.06.069183. Hoque R, Farooq A, Ghani A, Gorelick F, Mehal WZ. Lactate reduces liverand pancreatic injury in Toll-like receptor- and inflammasome-mediatedinflammation via GPR81-mediated suppression of innate immunity.Gastroenterology. (2014) 146:1763–74. doi: 10.1053/j.gastro.2014.03.014184. Harun-Or-RashidM,InmanDM.ReducedAMPKactivationandincreasedHCARactivationdriveanti-inflammatoryresponseandneuroprotectioninglaucoma.JNeuroinflammation.(2018)15:313. doi: 10.1186/s12974-018-1346-7185. Jeninga EH, Bugge A, Nielsen R, Kersten S, Hamers N, Dani C, et al.Peroxisome proliferator-activated receptor gamma regulates expression ofthe anti-lipolytic G-protein-coupled receptor 81 (GPR81/Gpr81). J BiolChem. (2009) 284:26385–93. doi: 10.1074/jbc.M109.040741186. Wanders D, Graff EC, Judd RL. Effects of high fat diet on GPR109A andGPR81 gene expression. Biochem Biophys Res Commun. (2012) 425:278–83. doi: 10.1016/j.bbrc.2012.07.082187. Madaan A, Chaudhari P, Nadeau-Vallée M, Hamel D, Zhu T, Mitchell G, etal. Müller cell–localized G-protein–coupled receptor 81 (hydroxycarboxylicacid receptor 1) regulates inner retinal vasculature via Norrin/Wnt pathways.Am J Pathol. (2019) 189:1878–96. doi: 10.1016/j.ajpath.2019.05.016188. Lee Y-S, Kim T-Y, Kim Y, Lee S-H, Kim S, Kang SW, et al. Microbiota-derivedlactate accelerates intestinal stem-cell-mediated epithelial development. CellHost Microbe. (2018) 24:833–46.e6. doi: 10.1016/j.chom.2018.11.002189. Cai TQ, Ren N, Jin L, Cheng K, Kash S, Chen R, et al. Role of GPR81in lactate-mediated reduction of adipose lipolysis. Biochem Biophys ResCommun. (2008) 377:987–91. doi: 10.1016/j.bbrc.2008.10.088190. Liu C, Wu J, Zhu J, Kuei C, Yu J, Shelton J, et al. Lactate inhibits lipolysis in fatcells through activation of an orphan G-protein-coupled receptor, GPR81. JBiol Chem. (2009) 284:2811–22. doi: 10.1074/jbc.M806409200191. Alarcon P, Manosalva C, Carretta MD, Hidalgo AI, Figueroa CD, TaubertA, et al. Fatty and hydroxycarboxylic acid receptors: The missing link ofimmune response and metabolism in cattle. Vet Immunol Immunopathol.(2018) 201:77–87. doi: 10.1016/j.vetimm.2018.05.009192. Morland C, Lauritzen KH, Puchades M, Holm-Hansen S, AnderssonK, Gjedde A, et al. The lactate receptor, G-protein-coupled receptor81/hydroxycarboxylic acid receptor 1: expression and action in brain. J.Neurosci Res. (2015) 93:1045–55. doi: 10.1002/jnr.23593193. Huckabee WE. Relationship of pyruvate and lactate during anaerobicmetabolism.Vcoronaryadequacy.AmJPhysiolPhysiol.(1961)200:1169. doi: 10.1152/ajplegacy.1961.200.6.1169194. Osnes JB, Hermansen L. Acid-base balance after maximal exercise of shortduration. J Appl Physiol. (1972) 32:59–63. doi: 10.1152/jappl.1972.32.1.59195. Kreisberg RA. Lactate homeostasis and lactic acidosis. Ann Intern Med.(1980) 92:227. doi: 10.7326/0003-4819-92-2-227196. Vardjan N, Chowdhury HH, Horvat A, Velebit J, Malnar M, MuhicM, et al. Enhancement of astroglial aerobic glycolysis by extracellularlactate-mediatedincreaseincAMP.FrontMolNeurosci.(2018)11:148. doi: 10.3389/fnmol.2018.00148197. Martin SJ, Grimwood PD, Morris RG. Synaptic plasticity and memory:an evaluation of the hypothesis. Annu Rev Neurosci. (2000) 23:649–711. doi: 10.1146/annurev.neuro.23.1.649198. Sun S, Li H, Chen J, Qian Q. Lactic acid: no longer an inert and end-product of glycolysis. Physiology. (2017) 32:453–63. doi: 10.1152/physiol.00016.2017Conflict of Interest: The authors declare that the research was conducted in theabsence of any commercial or financial relationships that could be construed as apotential conflict of interest.Publisher’s Note: All claims expressed in this article are solely those of the authorsand do not necessarily represent those of their affiliated organizations, or those ofthe publisher, the editors and the reviewers. Any product that may be evaluated inthis article, or claim that may be made by its manufacturer, is not guaranteed orendorsed by the publisher.Copyright © 2022 Cai, Wang, Song, Yang, Wang, Xue, Sun and Hu. This is anopen-access article distributed under the terms of the Creative Commons AttributionLicense (CC BY). The use, distribution or reproduction in other forums is permitted,provided the original author(s) and the copyright owner(s) are credited and that theoriginal publication in this journal is cited, in accordance with accepted academicpractice. No use, distribution or reproduction is permitted which does not complywith these terms.Frontiers in Nutrition | www.frontiersin.org17April 2022 | Volume 9 | Article 800901 | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/353807609An emerging municipalist culture in Curitiba. Fromdirect action to citizen-led governance, 2007 – 2018.Poster · August 2021DOI: 10.13140/RG.2.2.24930.25281CITATIONS0READS431 author:Some of the authors of this publication are also working on these related projects:PhD in Architecture of Contemporary Metropolitan Territories (ISCTE, Lisbon, Portugal) View projectBotanic Spine (Dublin, Ireland) View projectDuncan CrowleyISCTE-Instituto Universitário de Lisboa30 PUBLICATIONS 9 CITATIONS SEE PROFILEAll content following this page was uploaded by Duncan Crowley on 10 August 2021.The user has requested enhancement of the downloaded file.- Common agreement is that Curitiba is not an Ecocity today, nor a Sustainable City. The city showedinnovation in previous years, but the wider city struggles today with a quickly growing metropolis.- Since 2007, Curitiba’s urban landscape has been transformed by direct action focused socialmovements based on radical ecology, citizen participation, inclusivity and a vibrant political culture.- In 2016 Goura was elected as a Curitiba city councillor, representing radical eco activist socialmovements. Many proposals and laws were implemented. In 2018 he was elected as a Deputy inParana State and as chairman of the Commission on Ecology, Environment and Animal Protection.- The Curitiba Ecocity festival lasted 5 days in July 2017. It had 5 themes, 14 events and 160participants. More than 30 different groups helped co create or participate in the programme.- Curitiba city consists of 6 river basins. The basins and their fractal like sub basins can form base foreco neighbourhoods, ecobairros. A micro basin of the Rio Belem, the Pilarzinho, has a size of 5km2.There, the Jacu community garden could become a catalyst for the first ecobairro of Curitiba ecocity.AnemergingmunicipalistcultureinCuritiba.FromdirectactiontoCitizen-ledGovernance,2007– 2018.DuncanCrowley@fuspeyduncan_crowley@iscte-iul.ptHorta JacuParque GommPraça CiclistaJardim deSofiaView publication stats | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/363647901Tests, modelling and design of cold-formed steel moment resisting jointswith bolts in the web and outer flangeArticle in Thin-Walled Structures · April 2023DOI: 10.1016/j.tws.2022.110176CITATIONS0READS3084 authors:Some of the authors of this publication are also working on these related projects:Rapid post-earthquake bridge inspection through 3D reconstruction View projectDevelopment of light gauge steel truss portal framing system View projectAndrzej M WrzesienUniversity of the West of Scotland28 PUBLICATIONS 234 CITATIONS SEE PROFILEKrishanu RoyUniversity of Auckland142 PUBLICATIONS 2,243 CITATIONS SEE PROFILEZhiyuan Fang37 PUBLICATIONS 560 CITATIONS SEE PROFILEJames B.P. LimUniversity of Auckland287 PUBLICATIONS 4,720 CITATIONS SEE PROFILEAll content following this page was uploaded by Krishanu Roy on 18 September 2022.The user has requested enhancement of the downloaded file.1 Tests, modelling and design of cold-formed steel moment resisting 1 joints with bolts in the web and outer flange 2 Andrzej M. Wrzesien a*, Krishanu Roy b*, Zhiyuan Fangb,c and James B.P. Lim b,c 3 a School of Computing, Engineering and Physical Sciences, The University of the West of Scotland, 4 Paisley, High Street, PA1 2BE, United Kingdom 5 b School of Engineering, The University of Waikato, New Zealand 6 c Department of Civil and Environmental Engineering, The University of Auckland, New Zealand 7 *Corresponding Author Contact Details: 8 Andrzej M. Wrzesien, Krishanu Roy 9 E: andrzej.wrzesien@uws.ac.uk, School of Engineering and Computing, University of the West of 10 Scotland, Paisley, High Street, PA1 2BE, United Kingdom 11 E: Krishanu.roy@waikato.ac.nz, School of Engineering, The University of Waikato, New Zealand 12 13 Abstract: This paper presents an experimental and numerical investigation into cold-formed 14 steel portal frame moment-resisting joints, having bolts in both the web and outer flange. The 15 bearing resistance and stiffness of threaded bolt shanks in single and double shear lapped 16 connections are determined experimentally. A total of seven four-point bending tests on jointed 17 back-to-back channel beams, and one on a continuous beam, are then reported. In these tests, 18 three types of joint arrangements were considered. A finite element model is then described 19 and validated against the experimental test results in terms of ultimate moment capacity, failure 20 mode and initial stiffness of the moment-displacement curve. The accuracy of existing design 21 recommendations presented in the codes of practice (i.e., BS 5950-5; EN 1993-1-3; AS/NZS 22 4600; AISI S100-16) were compared against test results. Existing models for calculating the 23 moment resistance and the rotational stiffness of the moment connections were updated for the 24 different joint configurations tested. From this comparison, it was found that the models from 25 2 the design standards can provide safe predictions of the joint capacity, albeit only if the bolt 26 group length-to-width ratio is less than 3.5. 27 Keywords: Bending test, Proposed equation, Finite element analysis, Cold-formed steel, Bolt 28 connector 29 3 1 Introduction 30 Cold-formed steel (CFS) portal frame buildings are one of the most popular and efficient 31 types of steel structures and have been widely used in such sectors as retail, leisure, transport, 32 distribution and manufacturing [1]. CFS portal frames generally use bolted joints composed of 33 steel gusset plates (also known as brackets) with folded edge stiffeners. In recent times, the use 34 of CFS for portal frames as an alternative to hot-rolled steel is increasing, as they provide some 35 of the following benefits: higher strength-to-weight ratio; lower erection costs; lower joint 36 production costs [2]. 37 Several previous studies have demonstrated the viability of using cold-formed sections 38 to construct portal frames with moderate spans [3-6]. The key features of all these studies were 39 that eaves and apex joints were designed to function as rigid. These joints, however, were either 40 expensive to manufacture (i.e. the necessity of using thick welded brackets; access to 41 specialised forming equipment to form complex geometry and swages), or difficult to erect on-42 site (i.e. swages/additional stiffeners limit manoeuvrability of connected parts in the assembly; 43 welding of galvanised steel is not recommended; installation of self-drilling self-tapping 44 screws is time-consuming). 45 The earliest tests reported in the literature on cold-formed steel portal frame joints are 46 those by Baigent and Hancock [7]. The study focused on developing an analytical method of 47 predicting the strength and stiffness of portal frames using thin-walled asymmetric cross-48 sections under combined compression, major and minor bending and warping torsion. The 49 work of Baigent and Hancock [7] was followed by tests reported by Kirk [4] on the award-50 winning Swagebeam portal framing system (Design Council Award for innovative British 51 Design). The extensive experimental investigation was undertaken including tests on eaves and 52 apex joints as well as full-scale tests on portal frames at Salford University. The primary 53 4 innovation was that the joints could form through the swages rolled in the brackets which 54 interlocked with matching swages in the webs of the channel sections. Such an interlock system 55 allowed a reduction in the number of fixings by 50% and formed full-strength joints that can 56 be assumed as ‘rigid’ in the frame analysis. Another early contribution to knowledge on the 57 behaviour of bolted moment connections was the work produced by Zadanfarrokh [8-9]. The 58 authors focused on investigating the strength and stiffness of single-fixing lapped joints. 59 Mäkeläinen and Kankaanpää [10] reported tests on portal frame joints constructed from 60 back-to-back sigma sections connected through the web via brackets. Chung and Lau [11] 61 reported tests on an arrangement, where the joint was formed through back-to-back brackets, 62 bolted between the webs of the channel sections. Lim and Nethercot [12] proposed an 63 inexpensive cold-formed steel folded bracket to form easy-to-erect, semi-rigid joints for portal 64 frames of moderate spans. The connection brackets could be easily slid between widely-65 available standard lipped channel sections and were connected via the web bolts only, reducing 66 the installation time and cost [13]. Mills and LaBoube [6] conducted an experimental study on 67 different joint arrangements from the Australian cold-formed steel shed industry. Dubina et al. 68 [14] investigated eaves and apex joints formed by three different types of welded brackets. 69 Zaharia and Dubina [15] also conducted an experimental study on the flexibility of a single 70 fixing lap joint similar to that reported by Zadanfarrokh and Bryan [9]. The authors however 71 introduced an additional variable into the equation previously proposed by Zadanfarrokh and 72 Bryan [9] in the form of the bolt diameter. Dundu and Kemp [16] conducted research on single 73 channels connected back-to-back without a bracket. Such an arrangement is similar to that of 74 Mills and LaBoube [6], the main difference being that self-drilling, self-taping screws were 75 replaced by bolts. Kwon et al. [17] reported research on applications of closed sections 76 produced by a combination of cold-rolling and clinching techniques. Rhodes and Burns [18] 77 conducted extensive tests on joints of a cold-formed steel portal framing system. Yang and Liu 78 5 [19] reported tests on 10 different Sigma sections connected through cold-formed sleeves of 79 the same thickness. Santos and Simoes da Silva [20] presented a conceptual study of the 80 potential use of back-to-back cold-formed steel members of 450mm depth and 4mm thickness 81 in light industrial portal frames with crane girders. Zhang et al. [21] investigated the behaviour 82 of hot-rolled steel brackets using experimental testing and explicit finite element analysis to 83 investigate the effect of localized bending and failure in these thin-walled sections on the 84 ultimate capacity of the frame. Blum and Rasmussen [22] investigated experimentally and 85 numerically a set of connections for large-span portal frames including cold-formed steel 86 ‘knee-brace’ connection for eaves, bracket connections for apex, and mild steel welded 87 connection for the base. Bučmys et al. [23] developed a component technique for cold-formed 88 steel beam-to-column bolted gusset plate joints, which treats each junction as a set of 89 independent basic components and has been shown to predict joint stiffness accurately. Ye et 90 al. [24-25] investigated the seismic performance of cold-formed steel bolted moment 91 connections, and developed simplified bolt model for simulating friction-slip mechanism in 92 cold-formed steel connections. Rinchen et al. [26-29] investigated the behaviour of cold-93 formed steel portal frames with connections using experimental tests, numerical simulations 94 and direct strength method. Mojtabaei et al. [30-32] developed and explored the numerical and 95 analytical methods on studying the behaviour of cold-formed steel sections in bolted moment 96 connections. 97 Typical hot-rolled steel welded eaves and apex joints can be assumed as rigid and full-98 strength whereas cold-formed steel joints are often flexible (semi-rigid) and of partial-strength. 99 Joint rotation in cold-formed steel portal frames is associated with a combination of slip and 100 bearing of the mechanical fasteners (generally bolts) acting in shear. Typical CFS bolted 101 moment-resisting frames employ bolts placed in holes, with diameters 2mm larger than the bolt 102 6 diameter, to ensure the required construction tolerance, as seen in frames investigated by 103 McCrum et al. [33]. 104 When oversized holes are used only in the web of the channel, the rotational stiffness of 105 the joint is different during the slip and bearing stages of the structural response. Such joint 106 stiffness can lead to excessive lateral frame deflections under wind and earthquake loading, 107 which in turn may result in greater load transfer between frames via external sheeting (stressed 108 skin action). The unwanted consequences of stressed skin action, such as damage to the 109 cladding, purlins and gable frames, resulting from the use of joints with oversized holes have 110 been highlighted in the literature [34]. 111 Bolt slip in oversized holes can occur during frame erection and may result in changes 112 in the slope of the rafters or the verticality of the columns, presenting further alignment 113 problems. To prevent sagging or misalignment due to bolt slip, frame erectors often use 114 self-drilling, self-tapping screws (also known as tek screws) in conjunction with standard 115 bolts, according to the study by Pouladi et al. [35] 116 Although the effect of joint rotational stiffness on the structural modelling was 117 acknowledged by BS EN 1993-1-8 [36], the code does not present clear rules on how the 118 strength and rotational stiffness of the cold-formed steel bolted joint should be determined. Due 119 to a lack of worked examples, designers in the UK often follow dated joint design procedures 120 presented in SCI P125 [37] in which the moment resistance of connections between cold-121 formed sections is assumed to be governed by the bearing resistance of fasteners. With several 122 pieces of research published on CFS moment-resisting joints, a conclusive design procedure 123 has not yet been agreed but it is understood that bolt bearing rarely governs the design. The 124 most recent CFS design manual by Dubina et al. [38] offers guidance on design CFS columns 125 but design checks for joints’ resistance and flexibility are not presented. 126 7 This paper considers cold-formed steel moment-resisting joints with bolts in the web and 127 outer flange though both experimental tests and numerical models. A total of 8 bending tests 128 were conducted to produce experimental results showing the mode of failure of moment-129 resisting bolted joints. The analytical methods of predicting the moment capacity of the bolted 130 joints for cold-formed steel portal frames were presented. To validate common design 131 assumptions, structural analysis methods and design codes were compared against the 132 experimental data and existing codes of practice. Finally, a non-linear elasto-plastic finite 133 element (FE) model was developed to validate the experimental results in terms of ultimate 134 moment capacity, failure mode and initial stiffness of the moment-displacement curve. 135 2 Current design rules 136 2.1 BS 5950-5 [39] 137 The bending moment resistance, Mc,BS of the beam according to BS 5950-5 [39] is 138 calculated as: 139 (),0min,c BSyrMp E p E= (1) 140 Where, E is the elastic modulus; Er is the reduced elastic modulus; p0 is the reduced design 141 strength, 001.130.0019,280ywypbpppt=−; bw is the depth of the beam; t is the section 142 thickness; py is the design yield strength. 143 2.2 EN 1993-1-3 [40] 144 The cross-section strength and stability provisions are provided in EN 1993-1-3 [40] for 145 both local and distortional buckling modes. 146 147 8 2.2.1 Local buckling 148 In EN 1993-1-3 [40], the effect of local buckling is considered through the effective width 149 concept. The cross section is idealised as an “effective cross-section”. The effective width of a 150 (doubly-supported) plate is determined by: 151 10.055(3)1ellbb+=− (2) 152 With =ylcrf (3) 153 2222=12(1)crkEtb− (4) 154 Where, b and be are the total and the effective width of the plate, respectively; 21=/ff is the 155 stress ratio of the plate; k is the buckling factor; E is the Young’s modulus. 156 2.2.2 Distortional buckling 157 Distortional buckling of CFS members is linked to any buckling mode causing a 158 distortion of the shape of the cross-section, but excludes those deformations related to local 159 buckling. The distortional buckling behaviour can be studied by considering an equivalent strut 160 on the elastic foundation. The elastic critical buckling stress of the strut ,cr s is: 161 ,2scr ssKEIA= (5) 162 Where Is is the second moment; K is the spring stiffness; As is the effective cross-sectional area. 163 9 2.3 AS/NZS 4600 [41] and AISI S100-16 [42] 164 The unfactored design moment capacity of channels can be predicted from the AS/NZS 165 4600 [41] and AISI S100-16 [42]. The nominal moment capacity (Mnl) for local buckling shall 166 be calculated by using the following Equations 6 and 7: 167 For, l ≤ 0.776 nlneMM= (6) For, l > 0.776 0.40.4nl1 0.15crlcrlneneneMMMMMM=− (7) Where Mne is the nominal flexural strength for lateral-torsional buckling; Mcrl is the critical 168 elastic local moment. 169 The nominal moment capacity (Mnd) for distortional buckling shall be calculated using 170 the following Equations 8 and 9: 171 For, d ≤ 0.673 ndyMM= (8) For, d > 0.673 0.50.5nd10.22crdcrdyyyMMMMMM=− (9) Where Mcrd is the critical elastic distortional buckling moment; My is the member yield moment 172 (My= Sf×Fy). 173 3 Analytical methods 174 3.1 Analytical methods of predicting the rotational stiffness of the joints 175 For joints types A, B and C (see Fig.1), the elastic bolt group design method was used in 176 order to establish the centre of the rotation. The fundamental assumption of the method is that 177 the level of force in the bolt is proportional to the distance between the bolt and the centre of 178 the rotation. The centre of the rotation is therefore established based on this assumption and 179 static equilibrium of bending moment, longitudinal forces (Fx) and transverse forces (Fy). 180 10 a) Type A - 2x1 web bolt array b) Type B - 2x2 web bolt array c) Type C-3x3 web bolt array Fig.1 Joint types selected for testing In order to develop an analytical model representing the behaviour of a joint with bolts 181 in the web and bolts in the flange, the following assumptions were made: 182 1) The linear stiffness of the springs representing the web bolt (kb,w) are equal in 183 longitudinal and transverse directions. 184 2) Once the bolt is in contact with the steel plate, the linear stiffness of the spring 185 representing the flange bolt (single lap joint) is assumed to be half of that representing a web 186 bolt (double lap joint); thus if a single channel is considered the kb,f = kb,w = kb. 187 3) The bracket stiffener is in contact with the flange of the member. This limits a free 188 rotation at the joint therefore the linear slip allowed was assumed as 1mm (half of the hole 189 tolerance). 190 4) In Model 1, the flange bolts are engaged in carrying the loads and the transverse 191 stiffness of the sections flange of 2mm thickness is neglected in the analysis; thus flange bolts 192 11 are only able to carry longitudinal forces (see Fig.2). This model represents an upper bound of 193 joint stiffness. 194 5) In Model 2, the flange bolts are not included in carrying the loads; this model 195 represents a lower bound of joint stiffness. 196 6) The value of linear stiffness (kb) was assumed as 11.06 kN/mm according to Section 197 5.3 and Table 13. 198 7) The elastic method was used for calculating the reaction forces in the bolts subject to 199 concentrated moment also employed by Bučmys et al. [23], Rinchen et al. [28], and Wong and 200 Chung [43]. The symmetry of the joint about the axis along the direction y (see Fig.2) was 201 considered, thus, the x coordinate of the centre of the rotation (cor) was assumed as half of the 202 bolt group length. 203 Based on these assumptions, the analytical models including the contribution of the 204 flange bolts (Model 1) are shown in Fig.2. When the flange bolts are not engaged in carrying 205 the load (Model 2, Ff=0), the centre of the rotation is shifted down and the centre of rotation 206 coincides with the centre of gravity of the bolt group. 207 12 a) Joint type A -2x1 bolt array b) Joint type B – 2x2 bolt array c) Joint type C – 3x3 bolt array Fig.2 Joints behaviour analytical models including contribution of the flange bolts In order to determine the location of the centre of rotation (cor) and forces in each row 208 of bolts a set of equations was assembled according to Fig.2: 209 ()()corbgfbgMbyFyFaxFxF=−++−+321 (10) 210 Where, 211 021=+ FF (11) 212 03=+fFF (12) 213 )(21bgaxFxF−= (13) 214 )(223bgaxFyF−= (14) 215 13 )(3bgfbyFyF−= (15) 216 Where F1,2,3, Ff – forces in the relevant lines of bolts according to Fig.2, Table 3 and Table 4; 217 x,y – distances to the centre of the rotation bolts according to Fig.2; abg, bbg – length and width 218 of the bolt group according to Fig.2; bes – distance between flange a web line of bolts according 219 to Fig.2; kb – linear bearing stiffness of the single bolt in shear according to Table 13. 220 The solutions including locations of centres of rotations (cor) and forces in each line of 221 bolts are presented in Table 3 and Table 4 respectively for the upper and lower bound models. 222 The rotational stiffness formulas were also worked out for each bolt array according to 223 Equation (16) and are presented in Table 2. 224 ,/j inicorbgSM= (16) 225 Where Mcor – bending moment about the centre of the rotation; θbg=δi/ri – angle of the rotation 226 according to Fig.2. 227 Table 2 Model 1 and Model 2 – rotational stiffness formulas for all joints types Joint configuration Rotational stiffness according to Model 1 Sj,ini,1 Rotational stiffness according to Model 2 Sj,ini,2 2x1 bolt array ()2222bkbgbbga+ 22bkbga 2x2 bolt array ()32442423bkesbesbbgbbgbbga+++ ()bkbgbbga22+ 3x3 bolt array ()1621212211283bkesbesbbgbbgbbga+++ bkbgbbga2223+ 14 Table 3 Joints types A,B and C with flange bolts contribution (Model 1) – design formulas Notation (Fig.2) 2x1 bolt array 2x2 bolt array 3x3 bolt array X 2bga 2bga 2bga Y 2bgb 332esbbgb+ 485esbbgb+ F1 222bgbbgabgacorM+ +++224233esbesbbgbbgbbgabgacorM +++212211288esbesbbgbbgbbgabgacorM F2 222bgbbgabgacorM+ +++224233esbesbbgbbgbbgabgacorM 0 F3 2222bgbbgabgbcorM+ ()++++2242322esbesbbgbbgbbgaesbbgbcorM +++212211288esbesbbgbbgbbgabgacorM F4 2222bgbbgabgbcorM+ ()+++−224232esbesbbgbbgbbgaesbbgbcorM ()++++212211284.010esbesbbgbbgbbgaesbbgbcorM F5 - - ()++++2122112822esbesbbgbbgbbgaesbbgbcorM F6 - - ()+++−21221128232esbesbbgbbgbbgaesbbgbcorM Ff 2222bgbbgabgbcorM+ ()++++2242322esbesbbgbbgbbgaesbbgbcorM ()++++2122112826esbesbbgbbgbbgaesbbgbcorM 15 Table 4 Joints type A,B and C without flange bolts contribution (Model 2) – design formulas Notation (Fig.2) 2x1 bolt array 2x2 bolt array 3x3 bolt array X 2bga 2bga 2bga Y 2bgb 2bgb 2bgb F1 bgacorM 22bgbbgabgacorM+ 22bgbbgabgacorM+ F2 bgacorM 22bgbbgabgacorM+ 0 F3 - 22bgbbgabgbcorM+ 22bgbbgabgacorM+ F4 - 22bgbbgabgbcorM+ 22bgbbgabgbcorM+ F5 - - 0 F6 - - 22bgbbgabgbcorM+ Ff 0 0 0 3.2 Analytical methods of predicting ultimate strength of the joints 228 Two main criteria have to be considered when designing cold-formed steel joints. First 229 of all, the bearing capacity of the thin plate cannot be reached and local buckling of the plate 230 under concentrated forces from bolts must be prevented. 231 The design equations for predicting the bearing capacity of the plate Fb,Rd can be found 232 either in BS 5950-5 [39] or in EN 1993-1-3 [40] and the maximum force Fi,max in the farthest 233 bolt must be less than the calculated bearing capacity of the plate to satisfy Equation (17): 234 ,max,/1ib RdFF (17) 235 The design codes also offer little guidance regarding an analysis method for preventing 236 local web buckling mode of failure. Lim and Nethercot [44-45] conducted a parametric 237 numerical study using a validated finite element shell model and calibrated it against four tests 238 16 on a 3x3 bolt array. The design equation was also formulated. This numerical study was later 239 extended by Lim et al. [46] and authors proposed the use of the DSM design method in 240 estimating the reduction in bending resistance of the connection resulting from combined 241 bending and bimoment. Further refinements to design recommendations based on combined 242 bending and bimoment were proposed by Phan Duoc et al. [47] and in this paper, the design 243 principles published in EN 1993-1-3 [40] were employed to estimate reduction in moment 244 resistance. Concurrently, Mojtabaei et al. [48] used experimental results published by Lim and 245 Nethercot [44-45] to produce an extensive numerical study and proposed a resistance reduction 246 factor for 2x2, 3x3, 4x4 bolted connections. The same numerical methodology was used by 247 Mojtabaei et al. [49] to develop design formulas for portal frame bolted connections subject to 248 combined actions. The accuracy of the proposed interaction equations was not validated against 249 full-scale tests. All of the aforementioned researches focused on connections in which bolts 250 were located only in the webs of back-to-back channels. The research presented by Rinchen 251 et al. 2019 [26 -28] on thy other hand, proposed a design method for single-channel portal 252 frames subject to combined actions. Again, Vlasov’s theory was used in the calculation of the 253 bimoment stresses resulting from longitudinal and transverse bearing forces in fasteners. 254 Unlike in previous research, this method is not limited to a particular geometry of the bolt 255 group (fasteners in both web and flanges can be accounted for) and bolts combined with self-256 drilling, self-tapping screws were used in tested joints. The proposed method to calculate the 257 magnitude of the bimoment stresses was based on the full torsional restraint assumption and 258 was not validated experimentally. 259 In this paper, two analytical methods that are based on interaction equations published in 260 European design codes were used to initially estimate the resistance of the proposed joints: 261 17 1) Yu et al. [50] proposed a design method based on using a quadratic interaction 262 equation for a combination of shear and bending. The Effective Width Method was used for 263 predicting the bending capacity of the channel section Mcy,BS according to BS 5950-5 [39] and 264 a modified equation for critical shear buckling stress was proposed: 265 22,,(/)(/)1Edcy BSEdc RdMMVV+ (18) 266 Where: 267 ,,c Rdcr ChungVqAv= (19) 268 22,2=6.9 12(1)cr ChungEtqvDt−− (20) 269 0()vvADn d t=− (21) 270 In equations (18) to (21) remaining symbols are explained as follows: MEd - applied major axis 271 bending moment; VEd - applied shear force in the direction of channel’s web; FEd – applied 272 local transverse force in the direction of channel’s web; Vc,Rd - shear capacity of the channel; 273 qcr,Chung – critical buckling stress under shear; 6.9 – shear buckling coefficient, Av – net area of 274 the connected section in shear; E – Young’s Modulus; t – net thickness of the steel; ν – Poisson 275 ratio; D – overall web depth; nν – number of bolt holes in the critical shear section; d0 – nominal 276 diameter of the hole; Mcy,Bs and Mcy,Rd – design moment resistance of the cross-section in 277 bending according to BS 5950-5 [39] and EN 1993-1-3 [40] respectively; Vw,Bs and Vw,Rd – 278 design local transverse resistance of the web according to BS 5950-5 [39] and EN 1993-1-3 279 [40] respectively. 280 The force distribution was assumed to be based on a beam idealisation and elastic joint 281 design as presented in Fig.3. The critical cross-section lies on the right-hand side of the first 282 line of bolts (cross-section 4 shown in Fig.3). The effective cross-section area of the web in 283 shear Av was reduced by the bolt holes at the critical cross-section; 284 18 2) Dubina et al. [51] proposed an analytical method of predicting the ultimate strength of 285 the bolted joint based on an interaction equation for a combination of bending and web 286 crippling (EN 1993-1-3 [40]): 287 ,,//1.25Edcy RdEdw RdMMFV+ (22) 288 The interaction equation describing the same phenomenon was previously presented in a 289 slightly modified form in BS 5950-5 [39]: 290 ,,/1.1/1.50Edcy BSEdw BSMMFV+ (23) 291 In methods 1) and 2), the critical cross-section coincided with the line of bolts closest to 292 the end of the bracket. The maximum bending moment and the maximum transverse force 293 acting on the member in the critical cross-section can be calculated according to the principles 294 presented in Fig.3. At each bolt, the horizontal component of the force (FxM) resulting from the 295 pure moment applied at the centre of the bolt group is represented on the bending moment 296 diagram (see Fig.3) as a concentred moment. The maximum bending moment and the 297 maximum transverse force acting on the member in the critical cross-section can be calculated 298 from (e.g. Joint type C and Model 1): 299 𝑀𝐸𝑑 = 𝑀4𝐿 − 𝑀6𝐿 = 𝑀𝑐𝑜𝑟 − ⌊𝐹4𝑦+𝐹5(−0.5𝑏𝑏𝑔+𝑦)+ 𝐹6(𝑏𝑏𝑔−𝑦)+ 𝐹𝑓(𝑏𝑏𝑔+𝑏𝑒𝑠−𝑦) 3⌋ (24) 300 Where Mcor is bending moment about the centre of the rotation (see Fig.2(c)); F4 and Ff 301 are forces in the respective line of bolts according to Fig.2(c) and Table 3. 302 1EdEdVFF== (25) 303 Where F1 is the force in the first line of bolts according to Fig.2(c) and Table 3. 304 19 Fig.3 Force distribution within the half of the jointed beam considering single channel section (Model 2) The analytical joint capacities were determined using limit state criteria and design rules 305 according to British Standard (BS) and Eurocode 3 (EC). The relevant limit state design 306 equations are quoted in Table 5 and Table 6. The ultimate bending capacity of the continuous 307 member (Mcy) is also computed for comparison. The ultimate bending capacity of the tested 308 joint was taken as a minimum of three moment capacities representing the bearing capacity of 309 a plate (Mb), combined bending and shear force (MBSF) and combined bending and web 310 crippling (MBWC). The ultimate bending capacity according to Yu et al. [50] has been also 311 computed (MBSF,Chung) and presented in Table 5 as it contains a slight modification of the critical 312 shear buckling stress as compared with that adopted by the British Standard. The measured 313 mechanical properties of steel, as presented in Section 4.3, were used in the calculations. 314 In all tests, the combined bending and web crippling design criterion governed the 315 capacity and the values computed to EC3 were significantly lower than those computed to BS 316 standard. 317 20 Table 5 Upper bound analytical resistances according to BS 5950-5 [39] ultimate limit state design criterions Test designation Mcy,BS Mb,BS MBSF,BS MBSF,Chung MBWC,BS Eq. (17) Eq. (19) Eq. (19 to 22) Eq.(24) kNm kNm kNm kNm kNm T2/A/2x1/d 21.8 17.5 20.62 20.4 18.26 T3/B/2x2/d 22.43 30.74 19.2 18.64 16.84 T4/B/2x2/u 22.35 30.79 19.06 18.56 16.8 T5/C1/3x3/d 21.78 35.51 17.94 16.67 14.81 T6/C1/3x3/u 21.6 35.17 17.84 16.6 14.67 T7/C2/3x3/d 21.98 47.92 18.76 17.71 16.94 T8/C3/3x3/d 22.25 54.04 19.6 18.7 18.12 Table 6 Upper bound analytical resistances according to EN 1993-1-3 [40] ultimate limit state design criterions Test designation Mcy,EC Mb,EC MBSF,EC MBWC,EC Eq. (17) Eq. (19) Eq. (23) kNm kNm kNm kNm T2/A/2x1/d 21.17 17.96 20 16.39 T3/B/2x2/d 21.71 30.79 19 15.16 T4/B/2x2/u 21.65 30.79 18.86 15.14 T5/C1/3x3/d 21.22 36.02 17.63 12.9 T6/C1/3x3/u 21.07 35.34 17.53 12.8 T7/C2/3x3/d 21.44 47.61 18.46 15.23 T8/C3/3x3/d 21.56 54.23 19.05 16.15 4 Experimental study 318 4.1 Test specimens 319 CFS joints were proposed with stiffeners folded on both sides of the bracket, ensuring 320 that the bending capacity of the bracket is the same regardless of the load direction as shown 321 in Fig.1. The joints type A, B and C can be manufactured in the cold-forming, automated 322 process, in order to keep costs low. Stiffeners also allow the introduction of additional bolts at 323 the flange of the section, which are usually required for attaching secondary members. The 324 effect of these bolts on joint strength and stiffness is also investigated. The objective of the 325 experimental study was to: 326 1) Develop a simple analytical method for predicting the strength and stiffness of the 327 proposed type of connection; 328 21 2) Validate a beam idealisation of a structure that incorporates the proposed connections. 329 Since the experimental investigation contained not only component tests but also full-330 scale tests on portal frame buildings, the capacity of members under investigation had to be 331 reduced to match the capacity of available equipment and laboratory facilities. 332 4.2 Specimen labelling 333 The test specimens were labelled so that the joint type, bolt array in the web, and 334 direction of loading (see Fig.1). For example, the label “T2/A/2x1/d” can be explained as 335 follows: “T2” means the test series; “A/2x1” indicates the joint Type A with web bolts located 336 in 1 row; “d” identifies the loading condition (d: downward loading condition; u: upward 337 loading condition). 338 4.3 Material Testing 339 Tensile coupon tests were carried out to determine the material properties of the channel 340 sections. The tensile coupons were taken from the centre of the flange plate in the longitudinal 341 direction of the untested specimens. The tensile coupons were prepared and tested according 342 to the BS EN 10002-1:2001 [52]. Table 7 summarises the yield strength (fy) taken as the 343 measured static 0.2% proof stress and the ultimate static tensile strength (fu) for the channel 344 section which first showed signs of failure. The respective nominal values of proof stress and 345 ultimate tensile strength were 350MPa and 420MPa. 346 22 Table 7 Mechanical properties of steel based on coupon tests Test designation fy (MPa) fu(MPa) T2/A/2x1/d 386 502 T3/B/2x2/d 395 501 T4/B/2x2/u 398 504 T5/C1/3x3/d 397 508 T6/C1/3x3/u 393 499 T7/C2/3x3/d 403 506 T8/C3/3x3/d 395 501 4.4 Lap bolted joints 347 The test arrangement was similar to this proposed by Zadanfarrokh and Bryan [9] and is 348 shown in Fig.4 for both single and double lap joints. The Denison universal tensile machine 349 was used to apply the load at an approximate rate of 1kN/min. The displacement was measured 350 by two displacement transducers located on both sides of a bolt, connected with the data logger, 351 saving the displacement at 0.5kN load increments. The loading was stopped when the value of 352 the tension force reached 40kN for the single lap, and 70kN for the double lap joint 353 respectively. These values are approximately equal to the bearing resistance of investigated 354 joints calculated to BS EN 1993-1-3 [40]. Only two thicknesses of the plates were considered 355 to emulate the joint between the 2mm thick section and the 3mm thick bracket in both single 356 and double lap joints. A summary of the tested components is presented in Table 8. The test 357 specimens were manufactured from the same batch of steel as the components of the joint tests; 358 thus the mechanical properties established in tensile coupon tests were used in all further 359 analyses. 360 23 a) single lap joint b) double lap joint c) transducers arrangement Figure 4 Test arrangement for single and double lap joint Table 8 Summary of tested components Steel pieces Fastener Test series No. of tests Grade of steel – thicker piece t2 Grade of steel – thinner piece t1 Type dw tw mm mm mm mm S1/3.0/2.0 3 S350GD +Z275 3.0 S350GD +Z275 2.0 M16 30 2.7 D2/3.0/2.0 3 S350GD +Z275 3.0 S350GD +Z275 2.0 M16 30 2.7 dw − diameter of the washer, tw – thickness of the washer 4.5 Test rig and loading procedure 361 Laboratory tests were conducted on seven joints, as well as back-to-back continuous 362 members. Details of the laboratory test setup are shown in Fig.5. As can be seen, each joint test 363 comprises two identical bolt group arrangements, one on either side of the vertical axis of 364 symmetry, that are tested under four-point bending. Since the pitch of the roof should not have 365 a significant effect on the joint’s behaviour, a ‘zero degree’ pitch was assumed so the existing 366 test rig can be used. For all joint tests, the total length of the test specimen was 3 m and the 367 distance from the end support to the load point was 1 m. To prevent lateral-torsional buckling, 368 lateral restraints were provided at the supports, load points and the mid-span (see Fig.5b). At 369 support and load point, heavy T- sections were used (manufactured from RSJ 203x152x52), 370 24 and the test specimen was sandwiched between pair of square hollow (SHS 100x10x8.0) at the 371 mid-span. 372 δ - displacement transducers; FT – jacking force a) Front view of the tested beam b) Lateral-torsional restraints at loading points and mid-span Fig.5 Details of the joint test arrangement The bolts in each test were only finger-tightened without pre-torquing. Load cycles to 373 remove the bolt slip from the joint were not conducted, as any bolt-hole elongation caused, 374 would not be recoverable. However, before each test, the bracket was pulled as far as possible 375 in the direction of the applied load. At this position, the displacement transducers used to 376 measure deflections were set to zero. Bedding down change of angle due to slip and bolt-holes 377 tolerance was also measured before each test by a digital inclinometer. 378 25 4.5 Joints with web and flange bolts 379 A total of 8 joint tests were conducted on continuous member and joint configurations 380 type A, B and C according to Fig.1, cross-referenced with Table 9. A summary of the 381 components used for each test is presented in Table 10. 382 Table 9 Joint tests - connection geometries Test designation Bolt array web Bolts flange abg bbg ab aeb bebb bebt aes bes (mm) (mm) (mm) (mm) (mm) (mm) (mm) T2/A/2x1/d 2x1 2 240 106 680 35 69 49 35 - T3/B/2x2/d & T4/B/2x2/u 2x2 2 240 60 680 35 69 49 35 46 T5/C1/3x3/d & T6/C1/3x3/u 3x3 3 160 80 520 35 59 39 35 36 T7/C2/3x3/d 3x3 3 240 80 680 35 59 39 35 36 T8/C3/3x3/d 3x3 3 280 80 760 35 59 39 35 36 26 Table 10 Joint tests - summary of component geometries Test designation Average component dimensions (mm) Beam Section Bracket type B D B C tcor Db Bb tcor,b Nominal 152 64 20 1.96 178 67 2.96 T1 152.3 64.1 19.9 1.97 - - - T2/A/2x1/d 152.9 65.1 19.9 2.01 179 64.8 2.98 T3/B/2x2/d 152.8 65.5 19.8 2.02 179.6 66 3 T4/B/2x2/u 152.6 65 19.9 2.01 179.5 64.4 2.99 T5/C1/3x3/d 152.2 64.6 20.3 1.98 179.4 65.4 2.98 T6/C1/3x3/u 152.6 64.4 20.2 1.98 179.9 65.8 2.98 T7/C2/3x3/d 152.2 64.7 20.1 1.98 179.5 65.7 3.01 T8/C3/3x3/d 152.7 65.2 19.9 2.01 179.6 65.4 2.98 In order to measure the rotational stiffness of the bolt group, the flexural stiffness of the 383 back-to-back channel beam was also established, since it would contribute to the overall 384 deflection. For this reason, one test on a continuous beam was conducted. 385 The nominal thickness of each bracket was 3mm, and the nominal diameters of the bolts 386 and bolt-holes were 16mm and 18mm, respectively; all the bolts used were M16 Grade 8.8 and 387 had fully threaded bolt-shanks. 388 In test T2/A, the compression forces were carried only through the flange bolts (see 389 Fig.1(a)). Although such a joint would not be considered in full-scale tests, it was designed 390 purely to validate the simple physical model used to estimate the strength and stiffness of joints 391 type A. This joint used only two bolts in the web and four in the flange and the bolt group 392 27 length of 240mm. Tests T3/B and T4/B used the same length of the bolt group but an additional 393 row of bolts in the web was introduced as presented in Fig.1(b). 394 In portal frame applications, the connection brackets require a haunch shape to increase 395 their strength and stiffness without the necessity of using very thick plates. If such brackets are 396 to be manufactured by press break, they can only be bolted to the channel sections via the outer 397 flanges. This introduces an asymmetry and for this reason, in tests 3/B and 4/B, the same joint 398 was tested under the downward and uplift load. In the case of tests T5/C1 to T8/C3 the 3x3 399 bolt array in the web was used with an additional 6 bolts in the flanges (see Fig.1(c)). Due to 400 asymmetry of the connection in tests T5/C1 and T6/C1, the same joint was subjected to 401 downward and uplift loading. 402 5 Experimental results 403 5.1 Continuous beam test 404 The nominal and average section dimensions are shown in Table 10. The cold-formed 405 steel channels were manufactured from S350GD grade of steel according to BS EN 10326:2004 406 [53]. Tensile coupons were collected from each member and the mechanical properties of the 407 steel were established according to BS EN 10002-1:2001 [52]. The average measured proof 408 strength and average ultimate tensile strength are presented in Table 11. It should be mentioned 409 that there was no bedding-down load applied to the continuous beam as a bedding-down test 410 was not conducted for joint tests to prevent bolt hole ovalisation. 411 Three different analytical methods were used to evaluate the bending capacity of the 412 investigated beam: 413 28 1) Effective Width Method (Mc,BS - according to BS 5950-5 [39] ) 414 2) Effective Width and Effective Thickness Method (Mc,EC – according to EN 1993-1-3 415 [40]) 416 3) Direct Strength Method (Mc,DSM according to AS/NZS 4600 [41] and AISI S100-16 417 [42]) 418 For each method the following assumptions were made: 419 1) Yield strength according to Table 11 420 2) Modulus of elasticity E=210GPa 421 3) Effective length between lateral and torsional restraints of 500mm 422 The calculated capacities based on listed standard methods are summarised in Table 11. 423 Each of the listed codes also offers analytical rules for predicting the reduction in flexural 424 stiffness by reducing the gross second moment of are to the effective value Ific. This stiffness 425 reduction is described by the factor Qb defined in Equation (26) 426 ,/bficy grQII= (26) 427 Where Ific is the effective second moment of area for deflection; Iy,gr is the gross second moment 428 of area. 429 Table 11 Summary of analytical versus experimental bending moment of continuous beam Test designation fy fu Mcy,BS Mcy,EC Mcy,DSM Mexp Mcy,BS Mcy,EC Mcy,DSM / Mexp / Mexp / Mexp N/mm2 N/mm2 kNm kNm kNm kNm T1 392 512 21.23 20.73 20.12 21.85 0.97 0.95 0.92 Table 12 Summary of analytical versus experimental flexural stiffness of continuous beam Test designation Qb,BS Qb,EC Qb,DSM Qb,Exp T1 0.998 0.962 0.901 0.772 29 The values of the flexural stiffness reduction along with the reference are presented in 430 Table 12. The stiffness reduction based on the measured deflection for test T1 is also presented 431 in the same table. The experimental stiffness was evaluated by drawing a straight line through 432 point (0;0) and (t;0.6Mexp). A comparison between the analytical and the experimental 433 deflections is presented in Fig.6. The vertical displacement is plotted as an average of 434 displacements recorded by transducers 3a and 4a (see Fig.5a). 435 Fig.6 Experimental versus analytical moment-displacement relationship for continuous beam 5.2 Continuous beam – test results and concluding remarks 436 The comparison between experimental and analytical bending capacity is presented in 437 Table 11. The beam failed in distortional buckling of the compression flange as shown in Fig.6. 438 This mode of failure is not recognised by BS 5950-5 [39] which uses the Effective Width 439 Method to consider the local buckling of the cross-section. 440 Although the local buckling mode of failure was not observed, the estimated bending 441 capacity was still conservative and the closest to the experimental value. The distortional 442 buckling phenomenon is accounted for by the analytical method presented in Eurocode 3 and 443 the Direct Strength Method (DSM) adopted by AS/NZS 4600 [41] or North American 444 Specification for the Design of Cold-formed Steel Structural Members AISI S100-16 [42]. 445 T1 BS EC3 DSM 30 Both these analytical methods produced conservative predictions in terms of bending capacity 446 by 5% and 8% respectively as compared with the experimental results. The calculations were 447 based on measured rather than nominal cross-sectional geometries of channels at the mid-448 length. It was confirmed that all of the analytical methods can account for the effect of initial 449 imperfections with a degree of safety. 450 In terms of beam deflection, all of the analytical methods underpredicted the flexural 451 deformation (see Fig.6). The DSM method was the closest in predicting the beam deflections 452 as can be seen in Table 12 and Fig.6. The flexural stiffness of cold-formed steel members 453 received little attention in the literature since, in practical design, deflection limits are 454 discretional. In this work, however, accurate flexural stiffness is needed, so that the deflection 455 due to rotational stiffness of the joint, can be accounted for. 456 5.3 Results for lap bolted joints 457 In the case of bolted joints with oversized holes, Zadanfarrokh and Bryan [9] highlighted 458 the importance of separating the initial slip resulting from a rigid body movement of the bolt 459 in the oversized hole, from the working extension due to the bearing of the bolt into the steel 460 plate. This fact is clearly seen in Fig.7 where the experimental load-displacement relationships 461 are presented. The shear flexibility of the joint (sh) is derived at the point of the maximum 462 service load: 463 sh = mean (s1/0.6Fk, si/0.6Fk) (27) Where: 464 Fk - characteristic tearing resistance of a fastener 465 s1i – the extension measure at 0.6Fk for each individual test 466 The experimental shear stiffness (kb,exp) values for both single and double lap joints are 467 presented in Table 13. The experimental stiffness (kb,exp) was calculated as the mean gradient 468 of the load-extension relationships shown in Fig.7. The average displacement due to slip (ssl) 469 31 was recorded in Table 13 and the equivalent stiffness of the joint in the slip stage (kb,sl) was 470 also evaluated and presented in the same table. Table 13 also presents the typical modes of 471 failure observed during the tests for the single and double lap joints. 472 32 Table 13 Experimental shear stiffness of the joint Test series Average from 3 repeated tests Slip Bearing Mode of failure ssl kb,sl kb,exp mm kN/mm kN/mm S1/3.0/2.0 2.70 0.19 5.58 D2/3.0/2.0 1.61 3.10 11.06 a) Single lap joint 33 b) Double lap joint Fig.7 Load - extension relationship for single-bolt shear connections 5.4 Beam idealisation of the joint tests 473 A simple beam model was employed to represent the behaviour of a cold-formed steel 474 beam connected via bolted joints (see Fig.8). The flexural stiffness of sections was modelled 475 assuming gross cross-sectional properties reduced by the experimental factor of 0.772 (see 476 Table 12). The flexural stiffness of the bracket section was not reduced as no experimental data 477 was available. The length of the bracket is also significantly smaller than the length of the 478 channel, thus its exact stiffness is less important. The position of the centre of rotation from the 479 centre of the bracket (lb) was established according to Section 3 and was presented in Table 14. 480 The rotational stiffness of the bolt group was modelled as a rotational spring element (Sj,ini) 481 calculated in accordance with Table 2. The bracket was modelled as a beam element joining 482 two rotational springs. 483 34 Fig.8 Beam idealisation of the test arrangement In the case of bolted joints with clearance holes, the effect of slip due to the rigid body 484 movement of bolts in clearance holes must be considered. This problem is particularly 485 important in moment-resisting joints subject to load in opposite directions. Theoretical 486 maximum slip (ssl) between two sheets of metal is twice the hole clearances. The average linear 487 displacement due to slip in double lap joints was recorded as 1.61mm but for some tests reached 488 nearly 2mm (see Table 13 and Fig.7). It was assumed, however, that the slip displacement (δi,th) 489 can be reduced by half for joints with flange bolts, according to Section 3. The maximum 490 theoretical slip rotation sl,anl was calculated based on the length of the bolt group according to 491 Equation (28): 492 ,, /sl anli thir= (28) 493 Where δi,th is free displacement before the furthest bolt in contact with the plate; ri is distance 494 from the centre of the rotation to the furthest bolt. 495 Before the test specimen was loaded, the angle (sl) was recorded so that the analytical 496 (sl,anl) and measured values of joint rotation can be compared. 497 As presented in section 3, two stiffness models were considered: Model 1 considered the 498 contribution of the flange bolts and Model 2 neglected this contribution. The distance from the 499 centre of the rotation to the outer bolt was calculated based on data provided in Table 3, Table 500 4 and Fig.2. 501 The experimental data presented in Fig.7, shows a clear bilinear load-displacement 502 relationship. The initial shear stiffness recorded during the experiment was therefore classified 503 35 as slip. Once the bolt started bearing onto the steel plate, the joint shear stiffness increased 504 nearly four times (see Table 13 and Fig.7). Due to the complexity of the problem, the 505 assumption was made that the shear stiffness of the joint in the slip stage (kb,sl) is close to 0. 506 The initial rotational stiffness due to the slip in the joint Sj,ini,sl was calculated from 507 Equation (29): 508 ,,,/j ini slSWsl anlSM= (29) 509 Where MSW is bending moment from the self-weight of the structure; sl,anl – analytical slip 510 rotation of the joint. 511 The analytical rotational stiffness (Sj,ini.sl) was calculated for Model 2 as the location of 512 the centre of the rotation is different for each model. Analogously the upper and lower bound 513 analytical rotational stiffness (Sj,ini) due to bolt bearing were calculated according to section 3 514 and Table 2. 515 The rotational stiffness of the joint was represented as a bilinear moment-rotation spring 516 (see Fig.9) and the values of rotational stiffness were presented in Table 14. All the remaining 517 parameters were also listed in the same table. 518 Fig.9 Bilinear rotational stiffness model 5.5 Experimental results versus analytical models 519 This section presents the results of seven experiments conducted on joints with different 520 bolt arrays (Type A, B and C) and lengths of the bolt group (abg) between 160mm and 280mm. 521 36 The analytical methods for predicting rotational stiffness summarised in section 3 were 522 compared against test results in Table 15. 523 The initial sag of the tested beam under the dead load was represented by the initial slip 524 angle (sl) measured by the digital inclinometer of 0.1° accuracy before the loading was 525 initiated (see Table 15). The maximum initial analytical rotation (sl,anl,2) based on Model 2 is 526 also presented in Table 15 along with the joint experimental rotational capacity (Xd,exp). 527 As can be seen in the experimental moment-rotation relationships (Fig.10 to Fig.14), 528 some joints presented clear bilinear stiffness behaviour with the initial stiffness stage due to 529 slip, and final stiffness stage due to bearing of the bolts into the steel plate. Only the final 530 experimental rotational stiffness values (Sj,ini,exp) are presented in Table 15. The analytical over 531 experimental rotational stiffness ratios are also summarised in Table 15 for upper (Sj,ini,1) and 532 lower bound (Sj,ini,2) of stiffness. 533 Table 14 Analytical beam idealisation characteristic of tests Test designation lb Iy,gr,s Iy,gr,b abg bbg sl,anl,1 sl,anl,2 Sj,ini,1 Sj,ini,2 (Fig.8) (Fig.1) (Eq. 28) (Table 2) mm cm2 cm2 mm mm rad Rad kNm/rad kNm/rad T2/A/2x1/d 185 450.85 839.89 240 60 0.008 0.008 443 319 T3/B/2x2/d 185 453.94 858.76 240 60 0.008 0.008 762 677 T4/B/2x2/u 185 448.74 843.79 240 60 0.008 0.008 762 677 T5/C1/3x3/d 145 438.27 847.51 160 80 0.01 0.011 675 531 T6/C1/3x3/u 145 440.17 854.91 160 80 0.01 0.011 675 531 T7/C2/3x3/d 185 436.96 859.05 240 80 0.007 0.008 1205 1062 T8/C3/3x3/d 205 450.27 837.63 280 80 0.007 0.007 1551 1407 Since that analytical models presented here are not capable of considering the asymmetry 534 of the joint, the experimental stiffness and strength of the joint can vary depending on the load 535 direction. Under the downward load, the compressed part of the section prone to buckling is 536 restrained by the bracket stiffener which can increase the capacity. For this reason, one of each 537 37 joint consisting 2x2 and 3x3 bolt array was tested twice under both downward ("+" moment) 538 and uplift load ("-" moment) as can be seen in Fig.11 and Fig.12. 539 Table 15 Experimental beam idealisation characteristic of tests Test designation abg bbg sl sl,anl,2 Xd,exp Sj,ini,exp Sj,ini,1 Sj,ini,2 / Sj,ini,exp / Sj,ini,exp mm mm rad rad rad kNm/rad T2/A/2x1/d 240 60 0.003 0.008 0.066 534 0.83 0.6 T3/B/2x2/d 240 60 0.002 0.008 0.061 797 0.96 0.85 T4/B/2x2/u 240 60 0.005 0.008 0.052 822 0.93 0.82 Avr 0.9 0.76 T5/C1/3x3/d 160 80 0.01 0.011 0.075 601 1.12 0.88 T6/C1/3x3/u 160 80 0.003 0.011 0.066 591 1.14 0.9 T7/C2/3x3/d 240 60 0.003 0.008 0.061 1055 1.14 1.01 T8/C3/3x3/d 280 60 0.002 0.007 0.08 1229 1.26 1.14 Average 1.17 0.98 38 The experimental moment-rotation relationships are presented in Fig.10 to Fig.14. Each 540 graph contains: 541 1) Experimental moment-rotation relationship for the continuous beam test T1 as 542 described in Section 5.2; 543 2) Idealised stiffness for the continuous beam Model (T1) based on Section 5.2; 544 3) Experimental moment-rotation relationship of the jointed beam including the offset 545 due to measured initial slip rotation (sl); 546 4) Upper bound rotational stiffness joint model excluding slip due to hole tolerance 547 (Model 1) according to section 3. The BS 5950-5 [39] combined bending and web crippling 548 design criteria were chosen for establishing the ultimate moment capacity of the joints as they 549 offered the closest correlation with the experimental capacities (see Section 3). The analytical 550 rotation due to slip was not considered in the upper bound model to define a conservative value 551 of rotational capacity; 552 5) Lower bound rotational stiffness joint model including slip due to hole tolerance 553 (Model 2) according to Section 3, and the strength criterion as above; 554 6) Experimental rotation capacity (Xd,exp) as the limiting value in terms of joint ductility. 555 Fig.10 Moment-rotation relationship for experiment T2 39 Fig.11 Moment-rotation relationship for experiments T3 & T4 Fig.12 Moment-rotation relationship for experiments T5 & T6 Fig.13 Moment-rotation relationship for experiment T7 40 Fig.14 Moment-rotation relationship for experiment T8 A similar mode of failure was observed in every test where the buckling initiated in the 556 web of the section, propagated further to the flange as shown in Fig.15. This figure compares 557 the mode of failure of the 2x2 bolted joint of the same bolt group length under downward and 558 uplift load. Generally, both failure mechanisms are similar, however, when the compression 559 flange was not in contact with the bracket (see Fig.15(b)), the buckle in the flange developed 560 closer to the first line of bolts. 561 a) Downward load b) Uplift load Fig.15 Typical mode of failure of joint Type B 41 5.6 Joint tests – concluding remarks 562 It is highly recommended that the design of cold-formed steel moment-resisting joints is 563 assisted by testing as the current design codes do not offer specific guidance. In section 3, a 564 simple analytical model for designing not yet investigated joints is presented. The model was 565 carefully calibrated against seven experiments in terms of both stiffness and strength 566 prediction. 567 It is common practice for engineers to design such joints based on the following 568 assumptions: 569 1) full-strength assumption as long as bearing capacity check is satisfied (Rhodes and 570 Burns [18], SCI P125 [37]), 571 2) full rigidity assumption, 572 3) disregard of initial slip due to holes tolerance. 573 As the moment resistance based on bearing capacity is a function of number of bolts and 574 bolt group length (as can be seen in Table 16 and Table 17 for test T8/B/3x3/d), the assumption 575 1) is extremely dangerous as it will overestimate the joints capacity by nearly a factor of 3. 576 Only in one out of seven cases, the bearing check produced a safe prediction of the joint 577 capacity. 578 Table 16 Upper bound analytical versus experimental resistance according to BS 5950-5 [39] and Model 1 Test designation Mj,exp Mcy,BS Mj,exp/ Mcy,BS Mb,BS/ Mj,exp MBSF,BS MBWC,BS / Mj,exp / Mj,exp kNm kNm T2/A/2x1/d 18.39 21.8 0.84 0.95 1.12 0.99 T3/B/2x2/d 20.19 22.43 0.9 1.52 0.95 0.83 T4/B/2x2/u 17.22 22.35 0.77 1.79 1.11 0.98 T5/C1/3x3/d 18.17 21.78 0.83 1.95 0.99 0.81 T6/C1/3x3/u 16.23 21.6 0.75 2.17 1.1 0.9 T7/C2/3x3/d 20.62 21.98 0.94 2.32 0.91 0.82 T8/C3/3x3/d 20.31 22.25 0.91 2.66 0.96 0.89 42 The second assumption of the full-rigidity of the joint is also grossly incorrect as 579 highlighted by the experimental moment-rotation relationships for continuous and jointed 580 beams (see Fig.10 to Fig.14). In fact, the normalised stiffness ratio (Kj,nor) of the tested joints 581 ranged from 0.56 to 1.3 (see Fig.16). When such joints are used in a portal frame of 6m span 582 their stiffness ratios (Kj) would range from 3.4 to 6.7 (i.e.0.56x6=3.4) meaning that they should 583 be classified as semi-rigid joints (see Fig.16). 584 Fig.16 Classification of joints by strength and stiffness according to BS EN 1993-1-8 [36] Ignoring slip in moment-resisting connections is also very problematic as when the free 585 slip rotation occurs in the joint, the stiffness of the frame is low causing engagement of other 586 components (e.g. building envelope) in carrying the load. When the critical load is reached then 587 tearing or local buckling of the cladding can cause leakage and other serious serviceability 588 problems. It is therefore recommended that 3D stressed skin analysis is employed twice and 589 both upper and lower bound joint stiffness are modelled as described in Section 3. The upper 590 bound joint stiffness should be modelled in order to establish the maximum load in the internal 591 frame whereas the lower bound joint stiffness should be modelled to estimate the maximum 592 load acting on the cladding and gable frames [34]. Fig.10 to Fig.14 highlight the conservatism 593 of the analytical method summarised in Section 3, as every experimental curve falls between 594 the analytical representation lines in terms of rotational stiffness and rotational capacity. 595 43 Table 17 Lower bound analytical versus experimental resistance according to BS 5950-5 [39] and Model 2 Test designation Mj,exp Mcy,BS Mj,exp/ Mcy,BS Mb,BS/ Mj,exp MBSF,BS MBWC,BS / Mj,exp / Mj,exp kNm kNm T2/A/2x1/d 18.39 21.8 0.84 0.75 0.9 0.77 T3/B/2x2/d 20.19 22.43 0.9 1.45 0.88 0.76 T4/B/2x2/u 17.22 22.35 0.77 1.7 1.02 0.89 T5/C1/3x3/d 18.17 21.78 0.83 1.71 0.85 0.69 T6/C1/3x3/u 16.23 21.6 0.75 1.9 0.94 0.76 T7/C2/3x3/d 20.62 21.98 0.94 2.16 0.85 0.75 T8/C3/3x3/d 20.31 22.25 0.91 2.52 0.92 0.84 To discuss the effect of bolt group length on the moment capacity of the bolted joint, the 596 experimental data needs to be normalised against nominal yield strength and nominal 597 dimensions. This was done for tests T5, T7 and T8 by dividing the test results and analytical 598 results by the coefficient (Rs). The coefficient (Rs) was calculated according to the equation: 599 ,,,,/scy BS actcy BS nomRMM= (30) 600 Where Mcy,BS,act is the bending capacity of the channel section assessed based on actual yield 601 strength and actual dimensions; Mcy,BS,nom is the bending capacity of the channel section 602 assessed based on nominal yield strength and nominal dimensions. 603 Fig.16 Effect of bolt group length on the capacity of the joint The comparison of experimental and analytical relationships between bolt group size 604 (abg/bbg) and nominal moment capacity (Mnom) is presented in Fig.17. It can be seen that none 605 44 of the investigated joints reached the full capacity of the continuous member (highlighted on 606 the graph by the dashed line). As observed from experiments (see Fig.17), the capacity of the 607 joint increased linearly with the bolt group length until the ratio of 3 was reached, then a 608 decrease in the capacity was observed. The trends of experimental and analytical relationships 609 diverge, when the ratio of bolt group length-to-width passed the value of 3. The analytical 610 method may not be conservative in designing joints in which the bolt group length-to-width 611 ratio is greater than 3.5 612 Analogously, the relationship between the bolt group length and the joint rotational 613 stiffness was drawn in Fig.18, using normalised stiffness ratio Kj. As can be seen in Fig.18 the 614 investigated joints are very flexible falling just above pin classification marked on the plot by 615 the dashed line. 616 Fig.18 Effect of bolt group length on the rotational stiffness of the joint As expected, the analytical method was also unable to predict the increase of strength 617 when the compressed flange was bolted with the bracket. When comparing the normalised 618 experimental bending capacity, 17% and 11% increases in strength were observed, when the 619 joint was subject to downward load, for 2x2 and 3x3 bolt arrays respectively. 620 45 It is shown in this section that the analytical methods based on the work of Chung and 621 Lau [43], Yu et al. [50], Dubina et al. [51] can produce safe predictions of the joint capacity. 622 These are, however, not free of inaccuracies and they do not describe the ‘true’ mode of failure. 623 It was recently demonstrated by McCrum et al. [33] that the failure of the back-to-back 624 channels can be triggered by the additional compression stress due to warping torsion, as it was 625 first described by Baigent and Hancock [3,7] (see Fig.19(b)) in their research on single-channel 626 portal frames. Although one could say that this theory should not apply to back-to-back double 627 symmetric channel beams, the evidence of the bimoment action was visually detected in tested 628 joints as shown in Fig.19(a). Theoretically, in the symmetric cross-section under pure bending, 629 the bottom flange should be in tension and thus should not show evidence of compression 630 buckling as presented in Fig.19(a). The bimoment stress distribution (see Fig.19(b)) shows the 631 concentration of compression stresses at the exact location where buckling was observed. 632 Similar conclusions were also presented by McCrum et al. [33] where the magnitudes of 633 bimoment stresses in back-to-back channel columns were measured using the strain gauges. 634 a) local buckling of the bottom flange and lip close to outer line of bolts b) Stress distribution due to bimoment in single channel section Fig.19 Evidence of the additional stress due to bimoment The simple analytical model presented in Section 3, with its limitations highlighted here, 635 can offer a viable alternative to more complex and time-consuming analysis methods. Because 636 the model is based on well-known and well-understood analysis methods, it may be used by 637 practising steel designers to avoid unsafe designs. 638 46 6 Finite element analysis 639 6.1 General 640 The FE program ABAQUS [54] was used to develop a numerical model in this study. 641 The measured cross-section dimensions and the material properties obtained from the coupon 642 tensile tests were included in the FE model. 643 6.2 Material model 644 The true stress-strain relationship was obtained from the engineering stress-strain curves 645 by using the equations given next. 646 (1)true=+ (31) 647 ()ln(1)truetrue plE=+− (32) 648 Where E is the Young’s modulus, σ true and εtrue(pl) is the true stress and strain used in numerical 649 modeling, σ and are the engineering stress and strain, respectively. 650 6.3 Element type and mesh size 651 S4R shell elements were used to model the CFS channel beams, as the S4R elements are 652 linear 4-noded quadrilateral thick shell elements which have six degrees of freedom per node. 653 The mesh sensitivity analysis was conducted, and the results suggested that a mesh size of 5 654 mm×5mm (length by width) was suitable for CFS channel beams. Also, for the load transfer 655 brackets, a mesh size of 5mm×5mm (length by width) was used. Finer mesh sizes were used 656 near the rounded corners (Fig.20(a)). 657 47 (a) FE mesh (b) Boundary conditions and loading procedure Fig.20 FE modelling 6.4 Boundary conditions and loading procedure 658 In the bending tests, the concentrated loads and reaction forces were transferred from the 659 T-shaped load transfer plates to the specimens. Therefore, in the FE model, the Static (General) 660 analysis was used to apply the vertical load on the beams, and contact pairs were defined 661 between the webs of each CFS channel beam and load transfer plates using a surface-to-surface 662 interaction. In the normal direction of the contact pairs, a “hard” surface was used while in the 663 tangent direction between the two channels, a “frictionless” surface was assigned. The out-of-664 plane displacement of channel brackets is restrained. A coupling constraint was used to 665 simulate the connection between the beams and load transfer blocks. The multi-point constraint 666 (MPC) beam connector element was utilised in ABAQUS [54] to determine the stiffness of the 667 5mm x 5mm S4R mesh 5mm x 5mm S4R mesh Finer mesh around corner between flange and web: A total of five elements were set at the corner. Vertical displacement through two reference loading points Load transfer bracket with thickness at 6mm: Uy=0 Bolt connector Bracket: Uy=0 Boundary conditions applied to the support: Ux=Uy=0 48 fasteners, which was estimated based on the bolt diameter and section thickness. Similar 668 modelling technique has been employed by Roy et al. [55-67]. Using these connection element 669 stresses, a reasonable match between test and FEA findings was attained, lending confidence 670 to the connector modelling approach. 671 To simulate the experimental boundary conditions, the supported boundary condition 672 was modelled by releasing both the in-plane rotation and axial displacement along with the 673 specimen at the roller support. Two reference points were established at the top of the load 674 transfer plates and the translations of U1 and U3 were fixed at these reference points. The 675 loading was applied via a displacement control method by specifying a vertical displacement 676 at the two reference loading points (see Fig.20(b)). 677 6.5 FE validation 678 The FE model was validated against the test results. In Table 18, a comparison of the test 679 results (Mj,EXP) with the numerical results (Mj,FEA) is shown for all investigated specimens. The 680 mean value of the Mj,EXP/Mj,FEA ratio is 1.00 with the corresponding coefficient of variation 681 (COV) of 0.01. Fig.21 shows the failure mode obtained from FEA and tests for all test 682 specimens. Fig.22 depicts the moment versus displacement relationship obtained from both the 683 FEA and the test. The moment-displacement curves from experimental tests and FEA agreed 684 well with each other, as shown in Fig.22. The little disparity in the curves is caused by the 685 loading position used for the actual test. In experimental tests, it is difficult to apply the vertical 686 load from the loading cell to the midpoint of the beams. Minimal eccentric loading will result 687 in a difference between experimental and FEA curves. As can be observed, the FE results agree 688 fairly well with the experimental results in terms of moment capacity, deformed forms, and 689 initial stiffness of the moment-displacement curve. A comparison of several methods for 690 predicting moment capacity was carried out. When determining the moment capacity, the 691 developed FE models outperformed the other techniques. 692 49 Table 18 Comparisons of moment capacity of different methods Test designation Mj,FEA Mj,exp/ Mcy,BS Mb,BS/ Mj,exp MBSF,BS MBWC,BS Mj,exp/ Mj,FEA / Mj,exp / Mj,exp kNm T2/A/2x1/d 18.25 0.84 0.95 1.12 0.99 1.01 T3/B/2x2/d 20.15 0.9 1.52 0.95 0.83 1.00 T4/B/2x2/u 17.09 0.77 1.79 1.11 0.98 1.01 T5/C1/3x3/d 18.32 0.83 1.95 0.99 0.81 0.99 T6/C1/3x3/u 16.08 0.75 2.17 1.1 0.9 1.01 T7/C2/3x3/d 20.78 0.94 2.32 0.91 0.82 0.99 T8/C3/3x3/d 20.48 0.91 2.66 0.96 0.89 0.99 Average 0.85 1.91 1.02 0.89 1.00 COV 0.07 0.52 0.08 0.07 0.01 50 (i) Experimental deformation (ii) FEA deformation (a) T2/A/2x1/d (i) Experimental deformation (ii) FEA deformation (b) T3/B/2x2/d (i) Experimental deformation (ii) FEA deformation (c) T5/C1/3x3/d (i) Experimental deformation (ii) FEA deformation (d) T7/C2/3x3/d Fig.21 Deformed shapes at failure from experiments and FEA 51 (a) T2/A/2x1/d (b) T3/B/2x2/d (c) T4/B/2x2/u (d) T5/C1/3x3/d Fig.22 Comparison of FEA and experimental moment versus displacement curves 7 Conclusions 693 This paper presents an experimental, analytical and numerical investigation into the cold-694 formed steel moment-resisting joints with bolts in the web and outer flange. A total of 8 four-695 point bending tests were reported. Before bending tests on connections, a continuous channel 696 beam was also tested to confirm the accuracy of current design methods in predicting bending 697 resistance and deflection of the CFS beam. The moment resistance obtained from the tests was 698 also compared against the calculated moment resistances determined from different analytical 699 methods (Effective Width Method according to BS 5950-5 [39]; Effective Width and Effective 700 Thickness Method according to EN 1993-1-3 [40]; Direct Strength Method according to 701 AS/NZS 4600 [41] and AISI S100-16 [42]). It was demonstrated that the Effective Width 702 Method produced the least conservative and safe prediction of the channel’s resistance. 703 52 The experimental data also showed that all tested joints should be classified as partial-704 strength with the strength reduction factor ranging from 0.77 to 0.94. In order for the analytical 705 model to estimate joint moment resistance correctly, additional design criteria for combined 706 bending and shear and bending and web crippling should be considered (see Section 3, Table 707 16 and Table 17). It is the combined bending and web crippling design criterion that governed 708 the design in 6 out of 7 design cases. This design criterion alone produced conservative 709 predictions in both lower and upper bound models that respectively ignore (Model 2) and 710 consider (Model 1) the contribution of the flange bolts. 711 As expected, disregarding the contribution of flange bolts increased the level of 712 conservatism in predicting joint resistance between 5 and 20%. Such simplification will 713 however result in greater accuracy in predicting the joint rotational stiffness. From the 714 comparison in Fig.18, it was found that the analytical model including the contribution of the 715 flange bolts (Model 1) overestimated rotational stiffness in all tested joints. Model 2, ignoring 716 the flange bolts, offered safer predictions of the deformation albeit only if the bolt group length-717 to-width ratio was less than 3.0. 718 The FE models of connections were then presented, with non-linear material properties 719 incorporated into the models. All FE models were validated against the corresponding 720 experimental results, and they showed very good agreement both in terms of ultimate moment 721 capacity and failure modes. It is therefore evident that for best accuracy and the economy of 722 the design, such joints are analysed by more advanced FE models that are validated against test 723 results. 724 Acknowledgements 725 The work presented here was funded by a CASE Award (EPSRC) in association with 726 Capital Steel Buildings. 727 53 References [1] Target Zero 2011. Guidance on the design and construction of sustainable, low carbon warehouse buildings. Tata Steel and the British Constructional Steelwork Association. [2] R. P. D. Johnston, T. McGrath, S. Nanukuttan, J. B. P. Lim, et al. Sustainability of Cold-formed Steel Portal Frames in Developing Countries in the Context of Life Cycle Assessment and Life Cycle Costs. Struct. 13 (2018) 79–87. [3] A. H. Baigent, G. J. Hancock, 1982. The strength of cold-formed portal frames. 6th International Specialty Conference in Cold-Formed Steel Structures. St Louis, Missouri, USA. [4] P. Kirk, 1986. Design of a cold formed section portal frame building system. 8th International Specialty Conference on Cold-Formed Steel Structures. St. Louis, Missouri, USA. [5] B. W. J. Rensburg, G. P. De Vos, 1996. Lower cost lightweight cold-formed portal frames. 13th International Specialty Conference in Cold-Formed Steel Structures. St Louis, Missouri, USA. [6] J. Mills, R. Laboube, Self-drilling screw joints for cold-formed channel portal frames. J. Struct. Eng. 130 (2004) 1799-1806. [7] A. H. Baigent, G. J. Hancock, 1978. The behaviour of portal frames composed of cold-formed members. Sydney: University of Sydney, School of Civil Engineering. [8] F. Zadanfarrokh, 1991. Analysis and design of bolted connections in cold formed steel memebers. PhD, University of Salford. 54 [9] F. Zadanfarrokh, E. R. Bryan, Testing and design of bolted connections in cold-formed steel sections. 11th International Specialty Conference on Cold-Formed Steel Structures, 1992 St. Louis, Missouri, USA. 625-662. [10] P. Mäkeläinen, J. Kankaanpää, 1996. Structural design study on a light-gauge steel portal frame with cold-formed sigma sections. 13th International Specialty Conference on Cold-Formed Steel Structures. St. Louis, Missouri, USA. [11] K. F. Chung, L. Lau, Experimental investigation on bolted moment connections among cold formed steel members. Eng. Struct. 21 (1999) 898-911. [12] J. B. P. Lim, D. A. Nethercot, F. E.-assisted design of the eaves bracket of a cold-formed steel portal frame. Steel Compos. Struct. 2 (2002) 411-428. [13] A. M. Wrzesien, 2016. Effect of stressed skin action on the behaviour of cold-formed steel portal frames with non-linear flexible joints and top-hat purlins. Ph.D. Thesis, The University of Strathclyde. [14] D. Dubina, A. Stratan, A. Ciutina, L. Fulop, Zsolt, N. 2004. Monotonic and cyclic performance of joints of cold formed steel portal frames. In: LOUGHLAN, J. (ed.) 4th International Conference on Thin-walled structures. Loughborough, UK. [15] R. Zaharia, D. Dubina, Stiffness of joints in bolted connected cold-formed steel trusses. J. Constr. Steel Res. 62 (2006) 240-249. [16] M. Dundu, A. R. Kemp, Plastic and lateral-torsional buckling behaviour of single cold-formed channels connected back-to-back. J. Struct. Eng. 132 (2006) 1223-1233. [17] Y. B. Kwon, H. S. Chung, G. D. Kim, Experiments of cold-formed steel connections and portal frames. J. Struct. Eng. 132 (2006) 600-607. 55 [18] J. Rhodes, R. Burns, 2006. Development of a portal frame system on the basis of component testing. 18th International Specialty Conference in Cold-Formed Steel Structures. Orlando, Florida, USA. [19] J. Yang, Q. Liu, 2011. An experimental study of cold-formed steel sigma purlins with sleeve connections. 6th International Conference on Thin-Walled Struct. Timisoara, Romania. [20] S. Santos, D.A. Simoes, L. Silva, 2011. Connections of cold formed profiles in industrial buildings. 6th International Conference on Thin-Walled Struct. Timisoara, Romania. [21] X. Zhang, K. J. R. Rasmussen, H. Zhang, Structural modeling of cold-formed steel portal frames. Struct. 4 (2015) 58–68. [22] H. B. Blum, K. J. R. Rasmussen, Experimental investigation of long-span cold-formed steel double channel portal frames. J. Constr. Steel Res. 155 (2019) 316-330. [23] Ž. Bučmys, A. Daniūnas, J.-P. Jaspart, J.-F. Demonceau, A component method for cold-formed steel beam-to-column bolted gusset plate joints. Thin-Walled Struct. 123 (2018) 520-527. [24] J. Ye, S. M. Mojtabaei, I. Hajirasouliha, Seismic performance of cold-formed steel bolted moment connections with bolting friction-slip mechanism. J. Constr. Steel Res. 156 (2019) 122-136. [25] J. Ye, S. M. Mojtabaei, I. Hajirasouliha, K. Pilakoutas, Efficient design of cold-formed steel bolted-moment connections for earthquake resistant frames. Thin-Walled Struct. 150 (2020). [26] R. Rinchen, K. J. R. Rasmussen, Behaviour and modelling of connections in cold-formed steel single C-section portal frames. Thin-Walled Struct. 143 (2019) 106233. 56 [27] R. Rinchen, K. J. R. Rasmussen, Numerical modelling of cold-formed steel single C-section portal frames. J. Constr. Steel Res. 158 (2019) 143-155. [28] R. Rinchen,, K. J. R. Rasmussen, H. Zhang, Design of cold-formed steel single C-section portal frames. J. Constr. Steel Res. 162 (2019) 105722. [29] R. Rinchen, K. J. R. Rasmussen, Experiments on Long-Span Cold-Formed Steel Single C-Section Portal Frames. J. Struct. Eng. 146 (2020) 04019187. [30] S.M. Mojtabaei, I. Hajirasouliha, J. Ye, Optimisation of cold-formed steel beams for best seismic performance in bolted moment connections. J. Constr. Steel Res. 181 (2021), 106621 [31] D.T. Phan, S.M. Mojtabaei, I. Hajirasouliha, T.L. Lau, J.B.P. Lim, Design and Optimization of Cold-Formed Steel Sections in Bolted Moment Connections Considering Bimoment. J. Struct. Eng. 146 (2020) 04020153. [32] I. Papargyriou, S.M. Mojtabaei, I. Hajirasouliha, J. Becque, K. Pilakoutas. Cold-formed steel beam-to-column bolted connections for seismic applications, Thin-Walled Struct. 172 (2022), 108876. [33] D. P. McCrum, J. Simon, M. Grimes, B. M. Broderick, J. B. P. Lim, A. M. Wrzesien, Experimental cyclic performance of cold-formed steel bolted moment resisting frames. Eng. Struct. 181 (2019) 1-14. [34] A. M. Wrzesien, J. B. P. Lim, Y. Xu, I. A. Macleod, R. M. Lawson, Effect of stressed skin action on the behaviour of cold-formed steel portal frames. Eng. Struct. 105 (2015) 123-136. 57 [35] P. Pouladi, J. Ronaldson, G. C. Clifton, J. M. Ingham, A. M. Wrzesien, J. B. P. Lim, Finite-element assisted design of eaves joint of cold-formed steel portal frames having single channel-sections. Structures, 2019. 20, 452-464 [36] BS EN 1993-1-8 2005. Eurocode 3 - Design of steel structures. Part 1-8: Design of joints. Brussels: European Committee for Standardization. [37] SCI P125 1993. Building Design using Cold Formed Steel Sections: Worked Examples to BS 5950: Part 5: 1987. Ascot, UK,: The Steel Construction Institute. [38] D. Dubina, V. Ungureanu, R. Landolfo, 2012. Design of Cold-formed Steel Structures: Eurocode 3: Design of Steel Structures. Part 1-3 Design of cold-formed Steel Structures, Wiley. [39] BS 5950-5 1998. Structural use of steelwork in building Part 5: Code of practice for design of cold formed thin gauge sections. London: British Standards Institution. [40] BS EN 1993-1-3: Eurocode 3: Design of steel structures – Part 1-3: General rules – Supplementary rules for cold-formed members and sheeting. CEN, Brussels, 2006. [41] AS/NZS 4600:2018 2018. Cold-formed steel structures. Sydney / Wellington: Standards Australia / Standards New Zealand. [42] AISI (American Iron and Steel Institute). 2016. North American specification for the design of cold-formed steel structural members. AISI S100-16. Washington, DC: AISI. [43] M. F. Wong, K. F. Chung, Structural behaviour of bolted moment connections in cold-formed steel beam-column sub-frames. J. Constr. Steel Res. 58 (2002) 253-274 [44] J. B. P. Lim, D. A. Nethercot, Serviceability design of a cold-formed steel portal frame having semi-rigid joints. Steel Compos. Struct. 3 (2003) 451-474. 58 [45] J. B. P. Lim, D. A. Nethercot, Ultimate strength of bolted moment-connections between cold-formed steel members. Thin-Walled Struct. 41 (2003) 1019-1039. [46] J. B. P. Lim, G. J. Hancock, C. G. Charles, C. H. Pham, R. Das, DSM for ultimate strength of bolted moment-connections between cold-formed steel channel members. J. Constr. Steel Res. 117 (2016) 196–203. [47] D.T. Phan, S. M. Mojtabaei, I. Hajirasouliha, T.L. Lau, J. B. P. Lim, Design and Optimization of Cold-Formed Steel Sections in Bolted Moment Connections Considering Bimoment. J. Struct. Eng. 146 (2020) 04020153. [48] Mojtabaei, S. M., Becque, J. and Hajirasouliha, I. (2020). "Local Buckling in Cold-Formed Steel Moment-Resisting Bolted Connections: Behavior, Capacity, and Design". Journal of Structural Engineering, 146, 04020167. [49] S. M. Mojtabaei, J. Becque, I. Hajirasouliha, Behavior and Design of Cold-Formed Steel Bolted Connections Subjected to Combined Actions. J. Struct. Eng. 147 (2021) 04021013. [50] W. K. Yu, K. F. Chung, M. F. Wong, Analysis of bolted moment connections in cold-formed steel beam–column sub-frames. J. Constr. Steel Res. 61 (2005) 1332-1352. [51] D. Dubina, V. Ungureanu, 2008. Behaviour of multi-span purlins of bolted lapped cold-formed Z-sections. In: RASMUSSEN, K. & WILKINSON, T. (eds.) 5th International Conference on Coupled Instabilities in Metal Structures. Sydney, Australia. [52] BS EN 10002-1:2001 2001. Metallic materials - Tensile testing. Part 1: Method of test at ambient temperature. Brussels: European Committee for Standardization. [53] BS EN 10326:2004 2004. Continuously hot-dip coated strip and sheet of structural steels - Technical delivery conditions. Brussels: European Committee for Standardization. 59 [54] ABAQUS Analysis User’s Manual-Version 2021. ABAQUS Inc., USA, 2021. [55] B. Chen, K. Roy, Z.Y. Fang, A. Uzzaman, Y.H. Chi, J.B.P. Lim, Web crippling capacity of fastened cold-formed steel channels with edge-stiffened web holes, un-stiffened web holes and plain webs under two-flange loading, Thin-Walled Struct. (2021) 163. [56] B. Chen, K. Roy, Z. Fang, A. Uzzaman, C. H. Pham, G. Raftery, J. B.P. Lim. Shear behaviour and design cold-formed steel channels with edge -stiffened hole, un-stiffened hole, and plain web, ASCE J. Struct. Eng. 148 (2022). [57] B. Chen, K. Roy, Z. Fang, A. Uzzaman, G.M. Raftery, J. B. P. Lim, Moment capacity of back-to-back cold-formed steel channels with edge-stiffened holes, un-stiffened holes, and plain webs, Eng. Struct. 235 (2021) 112042. [58] K Roy, B Chen, Z Fang, A Uzzaman, X Chen, JBP Lim, Local and distortional buckling behaviour of back-to-back built-up aluminium alloy channel section columns, Thin-walled struct. 163 (2021) 107713. [59] K. Roy, H.H. Lau, Z. Fang, A.M. Ahmed, J.B.P. Lim, Axial capacity of back-to-back built-up cold-formed stainless steel unlipped channels-Numerical investigation and parametric study, Steel Compos. Struct. 40 (5) (2021) 761-780. [60] Z.Y. Fang, K. Roy, Q. Ma, A. Uzzaman, J.B.P. Lim, Application of deep learning method in web crippling strength prediction of cold-formed stainless steel channel sections under end-two-flange loading, Struct. 33 (2021) 2903-2942. [61] Z. Fang, K. Roy, Uzzaman. A, J.B.P Lim, Numerical simulation and proposed design rules of cold-formed stainless steel channels with web holes under interior-one-flange loading, Eng. Struct. 252 (2022) 113566. 60 [62] Z. Fang, K. Roy, H. Liang, K. Poologanathan, K. Ghosh, A.M. Mohamed, J. B. P. Lim, Numerical Simulation and Design Recommendations for Web Crippling Strength of Cold-Formed Steel Channels with Web Holes under Interior-One-Flange Loading at Elevated Temperatures, Buildings 11 (12) (2022) 666. [63] Z. Fang, K. Roy, D. Lakshmanan, P. Pranomrum, F. Li, H.H. Lau, J. B. P. Lim, Structural behaviour of back-to-back cold-formed steel channel sections with web openings under axial compression at elevated temperatures, J. Build. Eng. 54 (2022) 104512. [64] Z. Fang, K. Roy, B. Chen, Z. Xie, J. Ingham, J. B. P. Lim, Effect of the web hole size on the axial capacity of back-to-back aluminium alloy channel section columns, Eng. Struct. 260 (2022) 114238. [65] Z. Fang, K. Roy, Y. Dai, J.B.P. Lim, Effect of web perforations on end-two-flange web crippling behaviour of roll-formed aluminium alloy unlipped channels through experimental test, numerical simulation and deep learning, Thin-walled Struct. 179 (2022), 109489. [66] Y. Chi, K. Roy, B. Chen, Z. Fang, A. Uzzaman, G.B.G. Ananthi, J.B.P. Lim, Effect of web hole spacing on axial capacity of back-to-back cold-formed steel channels with edge-stiffened holes, Steel Compos. Struct. 40(2): 287-305. [67] Y. Dai, K. Roy, Z. Fang, B. Chen, G.M. Raftery, J. B. P. Lim, A novel machine learning model to predict the moment capacity of cold-formed steel channel beams with edge-stiffened and un-stiffened web holes, J. Build. Eng. 53 (2022) 104592. View publication stats | RESEARCH GATE |
Treatment of Hypophonia in Parkinson’s Disease ThroughBiofeedback in Daily Life Administered with A Portable VoiceAccumulator*,†Joakim K€orner Gustafsson, *,†Maria S€odersten, ‡Sten Ternstr€om, and *,†,§Ellika Schalling, *yzStockholm, andxSwedenSummary: Objectives. The purpose of this study was to assess the outcome following continuous tactile bio-feedback of voice sound level administered, with a portable voice accumulator to individuals with Parkinson’sdisease (PD).Method. Nine out of 16 participants with PD completed a 4-week intervention program where biofeedback ofvoice sound level was administered with the portable voice accumulator VoxLog during speech in daily life. Thefeedback, a tactile vibration signal from the device, was activated when the wearer used a voice sound level belowan individually predetermined threshold level, reminding the wearer to increase voice sound level during speech.Voice use was registered in daily life with the VoxLog during the intervention period as well as during one base-line week, one follow-up week post intervention and 1 week 3 months post intervention. Self-to-other ratio(SOR), which is the difference between voice sound level and environmental noise, was studied in multiple noiseranges.Results. A significant increase in SOR across all noise ranges of 2.28 dB (SD: 0.55) was seen for participantswith scores above the cut-off for normal function (>26 points) on the cognitive screening test Montreal CognitiveAssessment (MoCA) (n = 5). No significant increase was seen for the group of participants with MoCA scoresbelow 26 (n = 4). Forty-four percent ended their participation early, all which scored below 26 on MoCA (n = 7).Conclusions. Biofeedback administered in daily life regarding voice level may help individuals with PD toincrease their voice sound level in relation to environmental noise in daily life, but only for a limited subset. Onlyparticipants with normal cognitive function as screened by MoCA improved their voice sound level in relation toenvironmental noise.Key Words: Voice use— Parkinson’s disease— Portable voice accumulator— Biofeedback.INTRODUCTIONMost patients with Parkinson’s disease (PD) experiencesymptoms related to their speech and voice function. Up to90% show symptoms of hypokinetic dysarthria as the dis-ease progresses.1-3 The most prominent symptom is areduced voice intensity leading to difficulties making oneselfheard and communicating efficiently. Other speech andvoice symptoms common in PD include imprecise articula-tion, monotonous pitch and loudness, variable speech rate,and harsh and breathy voice quality, all which can be detri-mental to intelligibility.4The motor and non-motor symptoms associated with PDmay have psychosocial consequences such as restrictions incommunicative participation and social interactions. Theimpaired motor function may limit the individual’s abilityto move, travel and take part in social activities, thusdecreasing the opportunities for interaction.5 Both the cog-nitive and the physical effort related to speaking mayincrease following language changes such as impaired wordretrieval, sentence formulation and comprehension as wellas speech motor deficits that impact voice intensity andintelligibility.6,7 This can lead to social withdrawal both as aresult of an inability to keep up in a conversation and as acoping strategy to preserve energy.5,6Speech and voice function are traditionally assessedthrough standardized clinical assessment tools and per-ceptual and acoustic analysis of audio recordings made incontrolled environments.4 Standardized recordings in acontrolled setting are used to ensure a reliable and repli-cable assessment, paramount in both research and clinicalevaluation of treatment effects. Voice use in such con-trolled conditions may however not be representative ofvoice use during habitual speech in daily life, where manyexternal factors can affect the person�s speech production.Varying levels of environmental noise, movement andposture, stress, emotional state and the cognitive load ofparticipating in a conversation can all be challenging forthe speaker. These factors can be expected to have aneven larger impact for individuals with PD, since environ-mental noise will affect a soft voice more, posture is oftenaffected in PD, stress sensitivity increases, and cognitiveAccepted for publication October 10, 2021.From the *Department of Clinical Science, Intervention and Technology, Divisionof Speech and Language Pathology, Karolinska Institutet, Stockholm, Sweden;yMedical Unit Speech and Language Pathology, Karolinska University Hospital,Stockholm, Sweden; zSchool of Electrical Engineering and Computer Science, Divi-sion of Speech, Music and Hearing, Royal Institute of Technology (KTH), Stock-holm, Sweden; and the xDepartment of Neuroscience/Speech and LanguagePathology, Uppsala University, Uppsala, Sweden.Address correspondence and reprint Joakim K€orner Gustafsson, Department ofClinical Science, Intervention and Technology/Division of Speech and LanguagePathology, Karolinska Institutet, Karolinska University Hospital F67, 141 86, Stock-holm, Sweden. E-mail: joakim.gustafsson@ki.seJournal of Voice, Vol. &&, No. &&, pp. &&−&&0892-1997© 2021 The Authors. Published by Elsevier Inc. on behalf of The Voice Founda-tion. This is an open access article under the CC BY license(http://creativecommons.org/licenses/by/4.0/)https://doi.org/10.1016/j.jvoice.2021.10.009ARTICLE IN PRESSloading can have more substantial impact following cog-nitive changes.8Effective therapies focusing on speech and voice functionare of great importance to lessen the risk of negative psycho-social consequences PD. The diverse pharmacological andsurgical treatment options available today have a limitedand variable effect on speech and voice symptoms.9-12Behavioral treatment focusing on speech and voice functionis the primary option to alleviate symptoms related to hypo-kinetic dysarthria. Lee Silverman Voice Treatment (LSVTLOUD) is the leading voice therapy option today.13-16LSVT LOUD is a high-intensity treatment program focus-ing on a high phonatory effort to increase voice intensityand intelligibility. The program includes 16 hour-long treat-ment sessions distributed over 4 weeks, 4 times a week aswell as daily home exercises. The tasks are hierarchicallystructured with an increasing level of difficulty and includesustained phonation, pitch glides and production of words,phrases, sentences and conversational speech with higheffort and a loud voice. Positive changes regarding voiceintensity have been reported up to two years after treatment,when studied through controlled audio recordings of struc-tured speech and voice tasks.13Many individuals with PD who undergo traditional orhigh intensity voice treatment show positive changes invoice function in controlled conditions, such as during aclinical evaluation or a studio recording of speech.14,16,17Still, some individuals with PD may struggle with the gener-alization of their new voice capabilities to habitual speech indaily life. This could be attributed to deficits in internal cue-ing, which may lead to difficulties adjusting voice intensitywillfully in daily life. During treatment, cues are providedby the clinician and the setting, but such support is missingin daily life.17 Changes in the sensory perception of theirown voice production have been reported for individualswith PD, which could lead to an underestimation of therequired effort when speaking.18,19The difficulties couldalso be explained by an impaired ability to regulate andscale the intensity and range of motor functions.20-21Different approaches to further increase accessibility andimprove maintenance of treatment effects for voice therapyfor individuals with PD have been used, including teleprac-tice and digital training software.22-23 Another way to facili-tate a more effective generalization of treatment effectsoutside a clinical setting could be to provide an external cuein daily life, eg through biofeedback of voice sound level.Feedback can be used during speech and voice treatment tohelp the individual become aware of their performance andfunction. It can be in the form of external verbal cues or ofthe clinician guiding and showing the patient what theyshould do. The clinician can also prompt the patient to payattention to internal cues and to react to his/her own kines-thetic or auditory biofeedback. Enhanced external biofeed-back can also be provided when working with individualswith PD and a variety of methods have been used. Theseinclude visual biofeedback,24-25 auditory biofeedback,26combined auditory and visual biofeedback27 as well astactile biofeedback28-30 and results have generally been posi-tive in terms of changing the studied voice production.One method for providing continuous biofeedback onvoice intensity in daily life is the use of portable voice accu-mulators (PVAs). A voice accumulator is a technical devicethat can perform long-term registration of different aspectsof voice function during the wearer’s daily activities. Threedifferent PVAs have been commercially available to dateand they share many similar functions, the VoxLog (SonvoxAB, Umeå, Sweden), the Ambulatory Phonation Monitor(APM [KayPENTAX, Montvale, NJ]) and the Vocalog(Griffin Laboratories, Temecula, CA). The three devices allallow for registration of voice SPL (dB) and phonation time(time spent phonating during the registration period). TheVoxLog and the APM can register phonation frequency (inHz) and the VoxLog can also register the level of environ-mental noise SPL (in dB). All three devices can deliver tac-tile biofeedback on voice intensity in form of a vibration ifthe wearer uses a voice intensity below or above a set thresh-old level.The VoxLog was used by Schalling and colleagues29 withthe aim to increase voice sound level in daily life in individu-als with PD, by providing continuous biofeedback of voiceintensity. 6 participants wore the VoxLog for a total of threeweeks, which included 1 intervention week where biofeed-back was administered from morning to evening for thewhole week. The participants also wore the VoxLog 1 weekbefore and 1 week after the intervention period without bio-feedback activated to assess the effects. Mean voice intensitywas 1.5 dB higher during the intervention week when com-pared to the baseline week, which represented a statisticallysignificant difference at the group level, although there werelarge individual differences. There was however no sign of astrong retention when the biofeedback was removed theweek after intervention. Even though it is hard to determinewhether the change was clinically significant and meaning-ful during the speakers’ daily communication, the resultssuggest that tactile biofeedback on voice intensity deliveredin daily life can help individuals with Parkinson’s diseaseincrease their voice sound level. Other practice and feedbackschedules, based on principles for motor learning, designedto improve the effect and facilitate retention need to be eval-uated. A follow-up study evaluated how the VoxLog couldbe configured to better control the frequency with which thewearer receives feedback.30In the field of motor learning, different structures of prac-tice and feedback have been experimentally tested forimproved acquisition or improved retention of a desiredmotor function. Acquisition relates to improved perfor-mance during training and retention relates to improved per-formance after training. Several studies have investigatedhow these principles of motor learning relate to speechmotor learning. It has been found that, in many cases,knowledge based on motor function in limbs and fingerscan be applied to speech motor function as well31-33 By pro-viding biofeedback on habitual voice use in daily life, manyof the principles of motor learning that facilitate a strongARTICLE IN PRESS2Journal of Voice, Vol. &&, No. &&, 2021retention of an improved motor function are fulfilled.31-32 Itallows for a distributed practice over long periods of time,and also leads to a high number of trials. Different environ-ments with different external factors such as varying levelsof environmental noise and different communication part-ners lead to varied practice and training with habitualspeech, as the target task provides randomized practice withdifferent speech targets with each utterance. Furthermore,the tactile biofeedback from the VoxLog also provides feed-back in the form of knowledge of results, also potentiallyfacilitating retention.32Providing biofeedback on voice sound level during habit-ual speech in daily life could be an effective complement tovoice therapy to help with transfer, generalization, andmaintenance of treatment effects. If effective, it could alsobe used as a standalone tool to help individuals with PD toincrease their voice sound level in daily life, without concur-rent voice therapy. The aim of the present study was to eval-uate continuous tactile biofeedback regarding voice soundlevel during speech in daily life as a standalone intervention.The intervention program was extended in comparison to aprevious pilot study.29 As managing a technical device andresponding to a tactile biofeedback signal in daily life couldbe expected to be cognitively demanding, a secondary aimwas to assess whether cognitive level impacted the interven-tion outcome.MATERIAL AND METHODParticipants16 participants were recruited through their clinical contactat the neurology or speech and language pathology clinic ata university hospital, or through local patient organizations.Inclusion criteria were: diagnosed idiopathic PD, and sub-jective and documented speech and voice symptoms. Exclu-sion criteria were: severe hearing impairment and otherneurological disorders or communication deficits. All par-ticipants underwent strictly pharmaceutical treatment fortheir PD and the treatment dosage was constant during theparticipation period. None of the participants had takenpart in the pilot study by Schalling and colleagues.29 Noneof the participants had prior experience of voice therapy.All were instructed to refrain from other forms of voice ther-apy during the study.MaterialsThe VoxLog (firmware 2.2.3) is a portable voice accumula-tor, a device that enables long-term registration of voice usein daily life. The VoxLog can register fundamental fre-quency, phonation ratio (percent time spent phonating dur-ing registration), voice sound level (in dB) and the level ofthe background noise level (in dB), without recording theactual speech signal. The device uses an accelerometer incombination with a microphone mounted in a plastic neckcollar to detect phonation and register the voice use andbackground noise parameters. The neck collar is connectedto the device that can be worn in a pocket. The phonationratio is derived from the accelerometer signal: phonation isregistered when the energy level registered by the accelerom-eter signal exceeds a factory-set threshold level. The signalfrom the pre-calibrated microphone provides SPL of thevoice and environmental noise. The classification of voice vsenvironmental noise is made as follows: (1) When the accel-erometer detects phonation (skin acceleration at the neckgenerated by the activity in the vocal folds), the SPL regis-tered by the microphone is classified as the wearer’s voicesound level. (2) When there is no activity detected by theaccelerometer, the SPL registered by the microphone is clas-sified as level of environmental noise. This detection hasalso a timed gate of 200 ms that prevents unvoiced pho-nemes and room reverberation of the wearer’s voice frombeing classified as ambient noise.34-35 The device has a bio-feedback function that can provide continuous biofeedbackon voice sound level or fundamental frequency if thespeaker phonates above or below a preset threshold level.The biofeedback signal is delivered in the form of a tactilevibration generated in the box which the collar is connectedto. Data from up to one week can be collected continuouslybefore it must be downloaded to a computer, with the soft-ware VoxLog Connect (version 3.1.18).Dysarthria Assessment (in Swedish: Dysartribed€omning).It is divided into different sub-scales focusing on assessmentof respiration, phonation, oral motor function, articulation,prosody and intelligibility. It uses a 4-point rating scaleranging from 0 to 3 (0: normal function to 3: severe devia-tion or no function) for each item. The mean score of all thesub-scales is calculated in a compound rating of dysarthriaseverity.36 The dysarthria assessment was performed by thelead author.TheQuestionnaireonAcquiredSpeechDisorders(QASD, in Swedish: Sj€alvsvarsformul€ar Om F€orv€arvadeTalst€orningar, SOFT) is a self-report questionnaire withitems covering the speaker’s subjective symptoms related toliving with an acquired speech disorder. It includes threesub-scales; “my speech and language”, “speech and lan-guage in social interaction” and “personal and environmen-tal factors”. It uses a four-point rating scale (0: definitelyfalse to 3: definitely true) with a higher score representingmore severe symptoms. QASD is a clinical tool developedin Sweden and has been shown to have a moderate-to-strong correlation with similar but more extensive instru-ments that are used internationally.36-37TheVoiceHandicapIndex(VHI,inSwedish:R€osthandikappindex) is a self-report questionnaire contain-ing 30 items, rated on a 5-point scale (0: never to 4: always)covering three different dimensions of subjective voicesymptoms. The sub-scales include physical, functional andemotional symptoms, each represented by ten items. A sub-score for each dimension is calculated as well as a total score(0−120).38-39The Montreal Cognitive Assessment (MoCA) is a screen-ing instrument for mild cognitive dysfunction. It consists ofseveral tasks designed to assess different cognitive domains,ARTICLE IN PRESSJoakim K€orner Gustafsson, et alTreatment of Hypophonia in Parkinson’s Disease3including attention and concentrations, executive functions,memory, language, visuo-constructional skills, conceptualthinking, calculations and orientation. The maximum scoreis 30 points and a result of ≥ 26 is considered to reflect nor-mal cognitive function.40ProcedureThe participants’ voice use was monitored for a total ofseven weeks with the VoxLog. This included 1 week of base-line registrations, four weeks of biofeedback intervention,one week for evaluation directly after treatment and one fol-low-up week three months post treatment. (Figure 1) showsa flowchart of the whole participation period. The portablevoice accumulator VoxLog was used for the ambulatoryvoice monitoring. Voice use was monitored each day fromthat the participants woke up in the morning until they wentto bed in the evening.Baseline periodParticipants visited the clinic 3 times during the baselineweek. Initial assessment of speech and voice symptoms weremade with the standardized clinical tool called DysarthriaAssessment. The participants also rated their subjectiveexperiences of speech and communication with Question-naire on Acquired Speech Disorders (QASD) and VoiceHandicap Index (VHI). Cognitive screening was performedwith Montreal Cognitive Assessment (MoCA). The firstvisit was made for inclusion purposes, initial assessment andto initiate VoxLog registrations. The second visit was per-formed to ensure correct field registration and managementof the device. The third visit was performed to downloadbaseline data and to activate the feedback function, therebystarting the intervention period. Registrations of monologuespeech for a minimum of three minutes on a neutral topicwith the VoxLog were performed in a recording studio dur-ing each visit to the clinic.Intervention periodThe 4 week-long intervention period began after the base-line week. During this period, the participants wore theVoxLog with its biofeedback function activated. They wereinformed that the biofeedback signal would be activated iftheir voice level descended below a set threshold level. Theywere instructed to treat the biofeedback signal as a reminderto increase their voice level during conversation. Thebiofeedback settings were based on recommendations froma previous study by Gustafsson and colleagues30 whichsought to optimize the voice level biofeedback with regardto retention of improved motor function. The goal of thatstudy was to evaluate systematically how different configu-rations of the adjustable biofeedback settings in the VoxLogchanged the resulting feedback frequency. For the presentstudy, settings that resulted in a low feedback frequencyclose to 20% were chosen, based on the assumption that thebiofeedback would help the participants increase their voicesound level and the low feedback frequency would promotea stronger retention of improved voice function after theintervention. The starting threshold level used in this studywas individually adjusted based on the speaker’s voicesound level used during the baseline week. A threshold level3 dB below their mean voice sound level from the long-termregistration during the baseline week and an activation timeof 500 ms was used, following the recommendations byGustafsson and colleagues.30 The activation time is the timeperiod that the speaker needs to phonate below the thresh-old level before the biofeedback is activated. The partici-pantsvisitedthecliniconceperweekduringthebiofeedback intervention period to upload registered data.If the participants had increased their mean voice soundlevel during each week, the threshold level was increased aswell, to maintain a threshold level 3 dB below their meanvoicesoundlevelfollowingtherecommendationsdescribed.30Follow-up periodVoice use without biofeedback was registered during twoweek-long periods for follow-up: 1 week directly followingthe intervention period, and 1 week 3 months post interven-tion. Follow-up ratings of subjective experiences of speech,voice use and communication were made using the QASDand VHI instruments. The participants were also asked todescribe freely their subjective experiences from the inter-vention after the second follow-up period. 2 controlledregistrations of monologue speech in a recording studiowere performed with the VoxLog during each period.Data analysisThe primary outcome variable used to study the effect of theintervention was the Self-to-Other Ratio (SOR) registered indaily life with the VoxLog. SOR41-42 was defined asdescribed by Szabo Portela and colleagues.43 The SOR iscalculated as the difference between the voice sound leveland the level of environmental noise, for each time framecollected by the VoxLog in which both voice and noise havebeen registered. The VoxLog continuously registers the dif-ferent parameters with an observation every 4 ms, each ofwhich is classified as noise or voice, or inhibited by theaforementioned transition gate. For the present study, thedevice was set to generate a data point every 5 secondsreporting both the voice and noise level averages within thattime frame. SOR provides an indication of how well theFIGURE 1. Flowchart of data collected during the participationperiod.ARTICLE IN PRESS4Journal of Voice, Vol. &&, No. &&, 2021speakers can make themselves heard in a given context, spe-cifically how much their voice sound level exceeds the envi-ronmental noise level. The SOR was calculated in threedifferent noise ranges based on the Swedish Work Environ-ment Agency’s categorization regarding speech in noisedefined as: low levels of environmental noise (below 55 dB),normal levels of environmental noise (5 −570 dB), and highlevels of environmental noise (above 70 dB).44 The SORwas divided into these noise ranges as the SOR can beexpected to vary greatly in different levels of environmentalnoise. Arlinger44 states that a speakers’ voice level should be15 dB above the noise level to be comfortably heard. TheVoxLog Connect software was used to access the voice leveland noise level data. These data were exported to Matlaband a custom script was used to calculate SOR. To be ableto compare SOR across different levels of noise, SOR valueswas centered within each noise range (ie, the mean SOR fora certain noise range was subtracted from all SOR values inthat noise range so that the centered variables had a meanof 0 independent of which noise range they originatedfrom). A secondary outcome variable, voice sound levelduring the monologue recordings in a studio setting, isreported as well to enable comparisons to previous research,since monitoring of voice use in daily life is not commonlyperformed. Mean noise sound level during speech, ie noisesound level within frames with speech registered, was calcu-lated to control for changes in SOR with regards to varia-tions in noise sound level. Repeated measures ANOVA wasperformed to study changes in mean voice sound levelacross the different registration periods for both groups.The participants were divided into 2 groups; participantswith MoCA scores ≥ 26, reflecting normal cognitive func-tion,40 and participants with MoCA scores < 26.Ethical considerationsThe study was approved by the Regional Ethics Committeein Stockholm, Sweden (Dnr 2010/1023 −31/1) and was per-formed in accordance with the ethical standards laid downin the 1964 declaration at Helsinki.RESULTSGeneralA total of 16 participants enrolled in the study, of whom 9(56 %) completed the full program were included in theanalysis. The seven participants who terminated the inter-vention program did so for a variety of reasons. (Figure 2)shows the number of participants that completed each par-ticipation period and the reasons for the participants whodropped out. All seven participants who ended their partici-pation early had a MoCA result below normal levels (<26).(Table 1) shows participant information for the 9 partici-pants who completed the full program. Eight participantswere male and 1 female (ID 9). Each participant registeredvoice use for a total of seven weeks. The mean number ofhours acquired from weekly registrations for the group was54.7 hours (SD = 20.3). Mean individual recording lengthper week can be seen in (Table 1). The participants’ subjec-tive reports were positive following the intervention period.FIGURE 2. Flow diagram of the study and number of partici-pants who completed or dropped out of the study.ARTICLE IN PRESSJoakim K€orner Gustafsson, et alTreatment of Hypophonia in Parkinson’s Disease5All of them reported that the biofeedback had made themmore aware of what voice level they were using and how dif-ferent environments posed different challenges for them.Several also reported that their spouses perceived them asmore intelligible. A mean decrease in subjective speech andvoice symptoms rated with VHI and QASD could be seenfor both groups after the intervention. The mean VHI scoredecreased 10.2 points directly after the intervention com-pared to during baseline and the QASD score decreased 0.1points during the same period for the group with MoCAscores above 26. The corresponding change for the groupwith MoCA scores below 26 was a 6.5 points decrease inmean VHI score and a 0.4 points decrease in mean QASDscore. There was no significant difference in registrationlength between the group with MoCA scores above 26 andthe group with MoCA scores below 26, t(7) = -0.279,P = 0.789.Self-to-other ratio in daily life(Figure 3) shows each participant’s SOR in the differentnoise ranges during the 3 time periods; baseline (T1), postintervention (T2) and 3-month follow-up (T3). The partici-pants were divided into 2 groups, MoCA scores above cut-off for normal cognitive function (≥26 points) and belowcut-off on MoCA (<26 points). Different patterns can beseen for the 2 groups. The group with MoCA scores abovecut-off showed an increased SOR during T2 compared toT1 in all cases. There was a positive difference when thefollow-up time period (T3) is compared to the baseline(T1). A different pattern can be seen for the group withMoCA-score <26. In a few cases a positive change wasfound when comparing the T2 period to T1, but not to thesame degree as the group with MoCA scores ≥26. None ofthe participants in the group with MoCA scores >26showed any retention of the increased SOR at the three-month follow-up.(Figure 4) shows centered SOR values for all partici-pants in all noise ranges, presented for each time periods(T1-T3). Each participant, represented by the same colorineachnoiserange,showedagenerallyconsistentchange in magnitude across all noise ranges. This can beseen as the lines for the different noise ranges aregrouped together for each participant.Considering the consistent change in magnitude across allnoise ranges, the mean increase in SOR across all noiseranges was calculated. When comparing baseline with thepost period for the group with MoCA scores ≥ 26 points,the mean increase in SOR was 1.62 dB (95% CI [0.95, 2.3];SD = 0.54). The corresponding increase for the group withMoCA scores < 26 points was 0.05 dB (95% CI [-0.84,0.94]; SD = 0.56). The mean increase in SOR for the groupwith MoCA scores > 26 when comparing baseline with fol-low-up was 2.28 dB (95% CI [1.6, 2.95]; SD = 0.55). Thecorresponding change for the group with MoCA scores< 26 was -1.35 dB (95 % CI [-2.59, -0.12]; SD = 0.78). Thelack of overlap in the confidence intervals between theTABLE 1.Participant InformationParticipantAgeYs sincediagnosisDysarthriascore (0 − 4)MoCA(0 − 30)VHI Pre(0 − 120)VHI Post(0 − 120)VHI FU(0 − 120)QASD Pre(0 − 3)QASD Post(0 − 3)QASD FU(0 − 3)Mean registration(hs per wk/SD)ID 169100.4282223210.70.40.455.4/21.8ID 27260.42858435111.21.137.7/ 18.7ID 37450.82239293920.91.272.9/17.7ID 46760.4244139360.70.70.773.9/10.2ID 581101225540581.10.8136.1/11.6ID 67120.3243536150.50.30.242.8/19.1ID 76930.4303221330.60.50.646/9.9ID 87310.4293319250.40.40.464/16.1ID97430.5294230330.70.50.564.4/12.1Mean725.10.526.239.731.134.60.90.60.754.7/20.3Grey rows highlight participants with MoCA scores > 26.ARTICLE IN PRESS6Journal of Voice, Vol. &&, No. &&, 2021groups showed a significant difference in mean SORincrease between the groups.Noise sound level during speech(Figure 5) shows mean noise sound level during speech forall participants during the 3 time periods. No clear patternscan be seen between the 2 groups across the different timeperiods. Both positive and negative changes in mean noisesound level can be seen within both groups across the differ-ent time periods.Voice sound level in controlled environmentRecordings of monologue speech were made during each ofthe visits to the clinic during the participation period.(Figure 6) shows individual values for all participants acrossall registration points as well as means for each registrationperiod. An increase in mean voice sound level was seen forboth groups post intervention, 3.1 dB for the group withMoCA scores ≥ 26 and 1.4 dB for the group with MoCAscores below 26. At follow-up, the group with normalMoCA scores used a voice sound level 2.4 dB higher thanduring baseline. The group with MoCA scores below 26used a voice sound level 2.0 dB lower than during baseline.A repeated measures ANOVA showed a significant effect oftime period for the group with MoCA scores > 26, F(2,4) = 4.459, P = 0.003. Post hoc testing of simple effectswith the Bonferroni correction showed a significant increasein voice sound level when comparing the baseline with thepost period, t(4)=2.775, P = 0.023, and when comparingbaseline with the follow-up period, t(4)=2.776, P = 0.002. Arepeated measures ANOVA showed a non-significant effectof time period for the group with MoCA scores < 26, F(2,3)=5.143, P = 0.068.DISCUSSIONThis study examined the effects of a 4-week long interven-tion study in which continuous biofeedback on voice levelwas administered in daily life for participants with PD.Treatment focusing on speech and voice in PD is often verytime consuming and demanding both for the patient andcaregiver, as effective treatment programs such as LSVTLOUD are highly intensive. An intervention method bywhich the practice and training are performed in thepatient’s daily life could therefore be a valuable comple-ment. This study showed improvements in SOR followingFIGURE 3. Line plots presenting SOR, presented for different ranges of background noise, for each participant during baseline (T1), postintervention (T2) and at three-month follow-up (T3). SOR in low noise ranges are shown in solid lines (top), normal noise ranges are showndashed lines (middle) and high noise ranges are shown in dotted lines (bottom). Each participant is represented by the same color in all threenoise ranges. The left figure shows participants with a MoCA score ≥ 26 and the right figure shows participants with a MoCA score < 26.ARTICLE IN PRESSJoakim K€orner Gustafsson, et alTreatment of Hypophonia in Parkinson’s Disease7the biofeedback intervention, but only for participants witha MoCA scores of 26 or above.Difference in SOR and impact of cognitive functionThe Self-to-Other Ratio was chosen as the primary outcomemeasure for this study as voice sound level are highly depen-dent on the environmental noise around the speaker,following the Lombard effect.45-47 It is important to takethe environmental noise into consideration when studyingchanges in voice sound level in daily life. Otherwise, achange in voice sound level could be related primarily to anincrease in noise level, rather than to a changed behaviordue to intervention. Changes in SOR were studied in differ-ent levels of environmental noise. Different noise rangeswere calculated for the SOR parameter as it can be expectedFIGURE 4. Line plots presenting centered SOR divided by different noise ranges, for each participant during baseline (T1), post interven-tion (T2) and at three-month follow-up (T3). SOR in low noise ranges are shown in solid lines, normal noise ranges are shown in dashed linesand high noise ranges are shown in dotted lines. Each participant is represented by the same color in all three noise ranges. The left figureshows participants with a MoCA score ≥ 26 and the right figure shows participants with a MoCA score < 26.FIGURE 5. Mean noise sound level during speech for all participants during the three time periods; baseline (T1), post intervention (T2)and at three-month follow-up (T3). The left figure shows participants with a MoCA score ≥ 26 (n = 5) and the right figure shows participantswith a MoCA score < 26 (n = 4).ARTICLE IN PRESS8Journal of Voice, Vol. &&, No. &&, 2021to change as a result of variations in the noise level, consid-ering that it is harder to make oneself heard in loud noisethan in a quiet setting. Mean noise sound levels duringspeech were also calculated to further control for the varia-tions in noise sound level. This was important in order toinvestigate that the changes in SOR was a result of the inter-vention and not dependent on the level of environmentalnoise. No clear trends in the variation in noise sound levelacross the different time periods could be seen for eithergroup. If the observed changes in SOR were a consequenceof changes in noise sound level rather than of the interven-tion, a pattern should be seen with variations in SOR fol-lowing the changes in noise sound level. This does notappear to have been the case; the outcome for the groupwith MoCA scores > 26 can not be attributed to the Lom-bard effect. A clear distinction was seen in that the partici-pants who scored 26 points or above on the cognitivescreening test MoCA increased their SOR in a positive man-ner following the intervention and at follow-up, while theopposite pattern was seen for the group who scored below26 points on MoCA. The importance of cognitive functionis further exemplified when considering the participantswho dropped out from the study early and did not completethe whole program. 44 percent (7/16) participants endedtheir participation early for varying reasons, but all theseparticipants scored below 26 points on MoCA, suggestingthat the intervention method might have been too cogni-tively taxing in some cases. The mean increase in SORacross all noise ranges for the group with MoCA scores >26 was 1.62 dB when comparing the period post interven-tion with baseline. It is not an easy task to determine howbig a change in SOR might be considered clinically signifi-cant. A small change in SOR could make all the differenceif it helps the speaker to exceed the surrounding noise levelsenough to be intelligible. Arlinger44 states that a speaker’svoice level should be 15 dB above the noise level to be com-fortably heard, which could be used as a guiding value.However, environmental noise in daily life often fluctuatesquickly, and the SOR is calculated for a time frame of 5 sec-onds in the current study, making the comparison not opti-mal. With this in mind, all participants exceeded a SOR of15 dB in low noise ranges before the intervention, and allparticipants did not exceed a SOR of 15 dB in high noiseranges after intervention. Only 2 participants (ID8 and ID9)increased their SOR from a value below 15 to a value above15 in normal noise ranges after the intervention, and at fol-low-up. One previous study has reported SOR for healthyspeakers within the same age range48. The group of healthyspeakers used a mean SOR of 27.5 and 28.6 dB (mean forthe male group, n = 11, followed by the female group,n = 10) in low noise ranges, 16.9 and 16.5 dB in normalnoise ranges, as well as 8.8 and 9.2 dB in high noise ranges.Considering this, the changes in SOR in this study showedthat 3 of the 5 participants (ID 7, 8 & 9) with MoCA scores≥ 26 reached “normal” SOR values post intervention and atfollow-up which could be motivated as a clinically meaning-ful change.Difference in voice and noise levelsThe mean voice sound level during monologue speech in acontrolled setting, registered with the VoxLog in a sound-proofed recording booth, was collected to enable compari-sons with previous intervention studies. Ramig and col-leagues16 reported an increase in voice sound level duringmonologue speech after LSVT LOUD of 5.2 dB (SD = 3.2).The corresponding increase in this study post interventionwas 3.1 dB for the group with MoCA scores ≥ 26, whichhad decreased to 2.4 at 3-month follow-up. This smallerchange shows a smaller treatment effect when studied in acontrolled setting. One possible explanation for this couldbe the fact that the intervention in the current study is aimedat variable speech during daily life in an uncontrolled settingand not controlled speech tasks, which are an importantpart of more traditional voice therapy. There have beenattempts to provide other forms of external cues, with thegoal to increase voice sound levels for individuals with PD.The SpeechVive is a tool that triggers multi-talker babble inan earpiece when the wearer speaks, thereby using the Lom-bard effect to trigger a nonvolitional increase in voice soundlevel. Individuals with PD showed a mean increase in voiceFIGURE 6. Line plots presenting mean voice sound level during monologue speech registered with the VoxLog in a recording booth. Theleft figure shows participants with a MoCA score ≥ 26 and the right figure shows participants with a MoCA score < 26.ARTICLE IN PRESSJoakim K€orner Gustafsson, et alTreatment of Hypophonia in Parkinson’s Disease9sound level of 2.0 dB with the noise cue activated comparedto when speaking in a quiet, controlled, environment.49Changes in subjective ratingsA decrease in subjective symptoms related to voice functioncollected through subjective ratings with VHI was seen forboth groups. A 10.5 points decrease of mean VHI for thegroup with MoCa scores > 26 and a 6.5 points decrease forthe group with MoCA scores < 26 was seen post interven-tion. Both groups reported a decrease in subjective voicesymptoms even though positive changes in objective out-come measures were only seen for the group with MoCAscores above 26. This could reflect positive functional oremotional changes from an increased awareness of voiceuse. The corresponding change in QASD ratings was 0.1points for the group with MoCA scores above 26 and 0.4for the group with MoCA scores below 26. The differencebetween the groups was highly influenced by a single partici-pant, ID3. He reported the most severe subjective symptomswith QASD at the start of the study and subsequently expe-rienced the biggest decrease in subjective symptoms as ratedwith the QASD even though positive changes in objectiveoutcome measures were limited. Detailed analyses showedthat the majority of the changes for participant 3 came fromthe item categories “speech and language in social interac-tion” and “personal and environmental factors” with itemssuch as “I avoid situations where I am expected to talk”(decreased from 3 to 0) and “I am worried about my speechproblems” (decreased from 2 to 0).Methodological and clinical considerationsThe ability of the VoxLog and similar devices to administerbiofeedback on voice use in daily life presents an intriguingopportunity to provide intervention, with increased avail-ability for patients and practice, which is tailored toimprove transfer and generalization of treatment effects tohabitual voice use. The leading treatment option today is ahigh-intensity program which requires the patient to visit aclinic 4 days a week during a four-week period for individ-ual treatment sessions and accompanying home practice.15Providing biofeedback intervention in the patient’s daily lifewould allow more patients access to treatment if, eg, theyhave problems combining the high intensity program withwork, or do not have the energy to make many trips to andfrom the clinic. It could also reduce the caregiver’s cost,since the bulk of the training is performed by the patients athome during their daily conversations. It would be of valueto study the effect of biofeedback on voice sound level incombination with voice therapy to investigate the potentialof this combination as a future treatment option. The effectof the biofeedback alone was studied as a first step to mini-mize potential confounding factors influencing the results. Ifvoice therapy and biofeedback were studied in combination,it would be harder to parse out how the respective compo-nents of the intervention contributed to possible treatmenteffects. In addition, not all people with PD are able toparticipate in traditional voice treatment or LSVT for vari-ous reasons such as fatigue, distance to the clinic or limitedresources for service delivery, and for this group it is ofinterest to investigate the effect of biofeedback alone as analternative form of treatment.Placing the intervention in the patient’s daily life is alsopositive from a motor learning perspective. Many individu-als with PD struggle with carryover of treatment effects todaily life. This could be related to deficits in internal cue-ing, changes in sensory perception and an impaired abilityto regulate and scale intensity and range of motor func-tions.19-21 The feedback frequency, ie, how often the bio-feedback should be administered, is key in motor learning.Too high a feedback frequency can hinder the retention ofan improved function; it can lead to an over-reliance onthe feedback resulting in a deteriorating performance whenthe feedback no longer is administered.50 The feedback fre-quency is the main factor that can be adjusted in the Vox-Log software, by configuring the threshold level (the voicelevel the speaker needs to go below before the biofeedbackis activated), and the activation time (how long the speakerneeds to speak below the threshold level before the biofeed-back is activated). The settings used in this study werethose found earlier37 to produce a low feedback-frequencywhich has been shown to facilitate retention when learninga speech task.51-54 Recommendations in the Gustafssonet al30 study were to use a threshold level 3 dB below thespeaker’s mean voice level during baseline and an activa-tion time of 500 ms, which was used in the present study.This is however only the best option available with the dif-ferent configurations that can be made in the VoxLog soft-ware. It is in no way a guarantee that it will provide asteady feedback frequency. Varying degrees of environ-mental noise during the different periods compared to thebaseline registration will result in varying results. A morestable and reliable biofeedback function could be achievedif for example the biofeedback function were adaptive andbased on the SOR instead of on a fixed threshold level. Agoal SOR could then be set, eg a voice level 15 dB higherthan the noise level as recommended by Arlinger,44 andthe biofeedback activation could be based on the meanSOR for the previous minute and continuously adapt tothe environmental noise. While the ability to administer abiofeedback cue for to remind the wearer to increase theirvocal intensity is a promising feature, the tactile cue pro-vided by the VoxLog is not necessarily ideal. It introducesa factor taxing the cognitive system as a tactile vibration isnot a natural cue to increase vocal intensity and thereforerequiring the wearer to translate the meaning of the cue.The VoxLogs tactile cue is similar to the ones available inother devices such as the APM and the VocaLog, but theVoxLogs unique ability to register environmental noisemade it the best choice for the current study. This allowedthe analysis to include the confounding factor of variationsin environmental noise.The results from this study are limited in scope as datafrom only nine participants are reported. There was a highARTICLE IN PRESS10Journal of Voice, Vol. &&, No. &&, 2021percent of dropouts (44%) of the 16 participants originallyrecruited and who entered the study. All the participantswho dropped out early scored below cut-off for normal per-formance on MoCA and even though it was seldom explic-itly described except in 2 cases, the interpretation was inmany cases that they felt the treatment program was hard tofollow: they had to put on and take off the device each day,charge it in-between and constantly be aware of how theywere supposed to react to the biofeedback. In addition,results regarding changes in SOR show that cognitive level,as screened by MoCA, impacts how well the participantsbenefitted from the intervention. This treatment approachcould therefore be expected to be more suitable for individu-als with PD who perhaps have a hard time participating inconventional voice therapy due to still being active at work,rather than individuals declining voice therapy because offatigue or limited cognitive resources.Monitoring voice use in daily life during a long period oftime presents many challenges, as previously described byNix and colleagues.55 The challenges could be exacerbatedby the cognitive symptoms related to PD.4,56 All theincluded participants in the present study wore the VoxLogduring all seven weeks, but the amount of data (hours regis-tered) varied depending on compliance and technical diffi-culties. eg, in some cases a technical malfunction in thecollar led to lost data; while in others, the participants for-got to switch on the device in the morning. In a few cases,participants reported that they chose to not turn on thedevice for a period of time, eg during physical exercise at agym, or that they knew they were going to be at home aloneduring the day and therefore wouldn’t speak much, eventhough they were explicitly instructed to use the device any-way.CONCLUSIONSBiofeedback administered in daily life regarding voice levelmay help individuals with PD to increase their voice soundlevel in relation to environmental noise in daily life for alimited subset of participants. Only participants with nor-mal cognitive function as screened by MoCA improvedtheir voice sound level in relation to environmental noise,and not all the participants who increased their SORreached normal levels, when compared to previouslyreported data on healthy speakers within the same agerange. Furthermore, many participants with MoCA scoresbelow cut-off values for normal function exhibited difficul-ties managing the device during such a long period, andsome found it challenging to respond to the tactile cue, asan isolated intervention method. Considering this, themethod in its current form could be more suitable in combi-nation with traditional voice therapy, to help with mainte-nance and transfer of treatment effects to daily voice useand maintenance during or after well-established methods.Further studies are needed to determine the efficacy of suchpractice. All seven of sixteen participants who did not com-plete the study had MoCA scores below cut-off for normalcognitive function. Careful screening of cognitive functionis therefore recommended in future studies.AcknowledgmentsThis study was supported by the Promobilia Foundationand NEURO Sweden grants.REFERENCES1. Hartelius K, Svensson P. Speech and swallowing changes associatedwith Parkinson’s disease and multiple sclerosis: a survey. Folia Pho-niatr Logo. 2012;46:9–17.2. M€uller J, Wenning GK, Verny M, et al. Progression of dysarthria anddysphagia in postmortem-confirmed parkinsonian disorders. ArchNeurol. 2001;58:259–264.3. Schalling E, Johansson K, Hartelius L. Speech and communicationchanges reported by people with Parkinson’s disease. Folia PhoniatrLogo. 2017;69:131–141.4. Duffy JR. Motor Speech Disorders: Substrates, Differential Diagnosisand Management. 4. ed. Missouri: Elsevier Mosby; 2019. St Louis.5. Sj€odal Hammarlund C, Westergren A, A� str€om I, et al. The impact ofliving with Parkinson’s disease: balancing within a web of needs anddemands. Parkinsons Dis. 2018. https://doi.org/10.1155/2018/4598651.6. Miller N, Noble E, Jones D, et al. Life with communication changes inParkinson’s disease. Age Ageing. 2006;35:235–239.7. Yorkston K, Baylor C, Britton D. Speech vs speaking: the experiencesof people with Parkinson’s disease and implications for intervention.Am J Speech-Lang Pat. 2017;26:561–568.8. K€orner Gustafsson J, S€odersten M, Ternstr€om S, et al. Long-termeffects of lee silverman voice treatment on daily voice use in Parkin-son’s disease as measured with a portable voice accumulator. LogopPhoniatr Voco. 2019;44:124–133.9. De Letter M, Santens P, De Bodt MS, et al. The effect of levodopa onrespiration and word intelligibility in people with advanced Parkin-son’s disease. Clin Neurol Neurosur. 2007;109:495–500.10. Ho AK, Bradshaw JL, Iansek R. For better or worse: the effect oflevodopa on speech in Parkinson’s disease. Mov Disord. 2008;23:574–580.11. Plowman-Prine EK, Okun MS, Sapienza CM, et al. Perceptual charac-teristics of Parkinsonian speech: a comparison of the pharmacologicaleffects of levodopa across speech and non-speech motor systems. Neu-roRehabilitation. 2019;24:131–144.12. Sandstr€om L, H€agglund P, Johansson L, et al. Speech intelligibility inParkinson’s disease patients with zona incerta deep brain stimulation.Brain Behav. 2016;5:e00394.13. Ramig LO, Sapir S, Countryman S, et al. Intensive voice treatment(LSVT) for patients with Parkinsons disease: a 2 year follow-up. J Neu-rol Neurosur Ps. 2001;71:493–498.14. Sapir S, Ramig LO, Fox C. Intensive voice treatment in Parkinson’sdisease:leesilvermanvoicetreatment.ExpertRevNeurother.2011;11:815–830.15. Fox C, Ebersbach G, Ramig LO, et al. LSVT LOUD and LSVT BIG:behaviroal treatment programs for speech and body movement in Par-kinson’s disease. Parkinsons Dis. 2012:1–12.16. Ramig LO, Halpern A, Spielman J, et al. Speech treatment in Parkin-son’s disease: randomized controlled trial (RCT). Mov Disord.2018;33:1777–1791.17. Ramig LO, Fox C, Sapir S. Speech treatment for Parkinson’s disease.Expert Rev Neurother. 2008;8:297–311.18. Fox C, Ramig LO. Vocal sound pressure level and self-perception ofspeech and voice in men and women with idiopathic Parkinson disease.Am J Speech-Lang Pat.. 1997;6:85–94.19. Ho AK, Bradshaw JL, Iansek R. Volume perception in parkinsonianspeech. Mov Disord. 2000;15:1125–1131.20. Klockgether T, Borutta M, Rapp H, et al. A defect of kinesthesia inParkinson’s disease. Mov Disord. 1995;10:460–465.ARTICLE IN PRESSJoakim K€orner Gustafsson, et alTreatment of Hypophonia in Parkinson’s Disease1121. Demirci M, Grill S, McShane L, et al. A mismatch between kinesthes-ticandvisualperceptioninParkinson’s disease.Ann Neurol.1997;41:781–788.22.. Halpern AE, Ramig LO, Matos CEC, et al. Innovative technology forthe assisted delivery of intensive voice treatment (LSVT LOUD) forParkinson disease. Am J Speech Lang Pathol. 2012;21:354–367.23. Theodoros DG, Hill AJ, Russel TG. Clinical and quality of life out-comes of speech treatment for Parkinson’s disease delivered to thehome via telehabilitation: a noninferiority randomized controlled trial.Am J Speech Lang Pathol. 2016;25:214–232.24. Scott S, Caird FI. Speech therapy for Parkinson’s disease. J NeurolNeurosur Ps. 1983;46:140–144.25. Norrlinder K, Olsson L. Phonetogram as visual feedback in intensivevoice therapy for 3 patients with Parkinson’s disease. Unpublished mas-ter’s thesis. Stockholm, Sweden: Department of Clinical Science, Inter-vention and Technology: Division of Speech and Language Pathology,Karolinska Institutet; 2009.26. Sadagopan N, Huber JE. Effects of loudness cues on respiration inindividuals with Parkinson’s disease. MovDisord. 2007;22:651–659.27. Le Dorze G, Dionne L, Ryalls J, et al. The effects of speech and lan-guage therapy for a case of dysarthria associated with Parkinson’s dis-ease. Eur J Disorder Comm. 1992;27:313–324.28. Hauser CW. Use of ambulatory biofeedback to supplement Lee Silver-man Voice Treatment in patients with idiopathic Parkinson’s disease.Boston, MA: MGH Institute for Health Professions; 2005. Unpub-lished Master thesis.29. Schalling E, Gustafsson J, Ternstr€om S, et al. Effects of tactile biofeed-back by a portable voice accumulator on voice sound level in speakerswith Parkinson’s disease. J Voice. 2013;27:729–737.30. Gustafsson J, Ternstr€om S, S€odersten M, et al. Motor-learning-basedadjustment of ambulatory feedback on vocal loudness for patientswith Parkinson’s disease. J Voice. 2015;30:407–415.31. Maas E, Robin DA, Austerman Hula SN, et al. Principles of motorlearning in treatment of motor speech disorders. Am J Speech-LangPat. 2008;17:277–298.32. Bislick LP, Weir PC, Spencer K, et al. Do principles of motor learningenhance retention and transfer of speech skills? a systematic review.Aphasiology. 2012;26:709–728.33. Friedman I, Hancock AB, Bamdad MJ, et al. Using motor learning totreat speech: the modified motor learning guided approach. J MedSpeech-Lang Pa. 2010;18:13–34.34. Wirebrand M. Real-time monitoring of voice characteristics using accel-erometer and microphone measurements. Unpublished Master thesis.Stockholm, Sweden: Department of Electrical Engineering, Link-€opings University; 2011.35. Wirebrand M. VoxLog Report. Stockholm, Sweden. Sweden’s Innova-tion Agency; 2012.36. Hartelius L. Dysartribed€omning [Dysarthria assessment]. Stockholm,Sweden: Studentlitteratur; 2015.37. Hartelius L, Elmberg M, Holm R, et al. Living with dysarthria: evalua-tion of a self-report questionnaire. Folia Phoniatr Logo. 2008;60:11–19.38. Ohlsson AC, Dotevall H. Voice handicap index in Swedish. Logo Pho-niatr Voco. 2009;34:60–66.39. Jacobsen BH, Johnson A, Grywalski C, et al. The voice handicapindex (VHI): development and validation. Am J Speech-Lang Pat.1997;6:66–70.40. Nasreddine ZS, Phillips NA, Bedirian V, et al. The montreal cognitiveassessment,MoCA:abriefscreeningtoolformildcognitiveimpairment. J Am Geriatr Soc. 2005;53:695–699.41. Ternstr€om S. Preferred self-other-ratios in choir singing. J Acoust SocAm. 1999;105:3563–3574.42. Granqvist S. The self-to-other ratio applied as a phonation detector forvoice accumulation. Logop Phoniatr Voco. 2003;28:71–80.43. Szabo Portela A, Granqvist S, Ternstr€om S, et al. Vocal behavior inenvironmental noise: comparisons between work and leisure condi-tions in women with work-related voice disorders and matched con-trols. J Voice. 2018;32:126.e23–126.e38.44. Arlinger S. St€orande buller: Kunskaps€oversikt f€or kriteriedokumentation[Disturbing noise: systematic review for criteria documentation]. Stock-holm, Sweden: Arbetslivsinstitutet, F€orlagstj€anst; 1999.45. Lombard E. (1911) Le signe de l’élévation de la voix. Annales des Mal-adies L’Oreille et du Larynx. 1911;37:101–109.46. Lane H, Tranel B. The Lombard sign and the role of hearing in speech.J Speech Hear Res.. 1971;14:677–709.47. Garnier M, Henrich N, Dubois D. Influence of sound immersion andcommunicative interaction on the Lombard effect. J Speech LangHear Res. 2010;54:588–608.48. K€orner Gustafsson J, S€odersten M, Ternstr€om S, et al. Voice use indaily life studied with a portable voice accumulator in individuals withParkinson’s disease and matched healthy controls. J Speech Lang HearRes. 2019;62:4324–4334.49. Stathopoulos ET, Huber JE, Richardson K, et al. Increased vocal intensitydue to the Lombard effect in speakers with Parkinson’s disease: simulta-neous laryngeal and respiratory strategies. J Commun Disord. 2014;48:1–17.50. Salmoni AW, Schmidt RA, Walter CB. Knowledge of results andmotor learning: a review and critical reappraisal. Psychol Bull.1984;95:355–386.51. Adams S, Page A. Effects of selected practice and feedback variableson speech motor learning. J Med Speech-Lang Pat. 2000;8:215–220.52. Steinhauer K, Grayhack JP. The role of knowledge of results in perfor-mance and learning of a voice motor task. J Voice. 2000;14:137–145.53. Adams S, Page A, Jog M. Summary feedback schedules and speechmotor learning in Parkinson’s disease. J Med Speech-Lang Pat.2002;10:215–220.54. Kim IS, Lapointe LL, Stierwalt JA. The effect of feedback and practiceon the acquisition of novel speech behaviors. Am J Med Speech-LangPat. 2012;21:89–100.55. Nix J, Sveg JG, Laukkanen AM, et al. Protocol challenges for on-the-job voice dosimetry of teachers in the United states and Finland. JVoice. 2007;21:385–396.56. Chaudhuri KR, Odin O. The challenge of non-motor symptoms inParkinson’s disease. Prog Brain Res. 2010;184:325–341.ARTICLE IN PRESS12Journal of Voice, Vol. &&, No. &&, 2021 | RESEARCH GATE |
Studia Ceranea 10, 2020, p. 63–82 DOI: 10.18778/2084-140X.10.04ISSN: 2084-140Xe-ISSN: 2449-8378Aneta Dimitrova (Sofia)https://orcid.org/0000-0003-1973-8462Translation and Transformation of John Chrysostom’s Urban Imagery into Old Church SlavonicSt John Chrysostom preached for 20 years in the two major cities of the East-ern Roman empire – Antioch (386–397) and Constantinople (398–404). He delivered hundreds of sermons, some of these were written down at the time of preaching, others were edited and published later. More than 800 are consid-ered genuine, another thousand texts bear his name as the author1. His works were widely popular in all the neighbouring cultures from the 5th century onwards. In the 9th century, the Slavonic tradition joined this trend.The title of my present research suggests mainly a survey on literary and cul-tural reception, but it also allows a discussion on several broader questions, such as the history of rhetoric and preaching, the use of literary sources as historical evidence, late antique and medieval architecture, city planning, and everyday life. John Chrysostom’s extensive and diverse body of work provides a large number of examples and theoretical models in various study domains. Here I will focus on his depictions of the city and urban life and their rendition in the early Slavonic tradition.John Chrysostom was a preacher and a writer – many of his sermons were both oral performances in an actual reality, and written works meant for reading in posterity, outside of their immediate context. His fellow-citizens in Antioch and Constantinople were his audience, but they were also subjects of his sermons. He drew material from contemporary events, natural disasters, political and social turmoil, local landmarks, the neighbourhood, etc. It is not surprising then, that his works served as documental sources in academic research. The written texts, which have come down to us in numerous manuscripts, still keep traces of direct communication2. The preacher addresses his audience from time to time, points 1 S. Voicu, Une nomenclature pour les anonymes du corpus pseudo-chrysostomien, B 51, 1981, p. 297–305.2 Some aspects of this question are discussed also by other researchers, Preacher and Audience. Stud-ies in Early Christian and Byzantine Homiletics, ed. P. Allen, M. Cunningham, Leiden 1998, p. 18: Aneta Dimitrova64to the surroundings, refers to the previous day (“yesterday”), names particular persons, and local suburbs. One way of looking at his preaching, as Wendy May-er points out, is as a liturgical act which takes place within a liturgical setting3. In this line of reasoning, Mayer poses a number of questions concerning the actual moment of delivery and the interaction between the preacher and his congrega-tion, such as: “What behaviour does he expect of the audience during the homily”, “Where is his audience situated?”, “Can John project his voice adequately?”, “Who is sitting and who is standing?”, and so on4. Another strain of questions refers to the preacher’s surroundings: “In which city are John and the audience in question situated?”, “In what part of the urban or suburban landscape are they positioned?”, “In which building are they located?”, etc.5 None of these questions, which repre-sent the liturgical, topographical, social, or personal perspective on Chrysostom’s preaching, is relevant to the afterlife of his homilies. Later copies and translations take the homily away from the initial moment of delivery and bring it to a new readership (or audience), into another era and another cultural and topographical context. This loss of actuality is typical for all oral sermons put into writing6. The double nature of the homily – oral and written – creates an artificial, rhetorical reality, cf. W. Mayer again:Even if we can confirm that the homily that survives was delivered before a live audience and is identical to the original, and we can demonstrate that John individualised the content in response to his audience, we must still deal with the fact that the information itself is pre-sented within a rhetorical medium and represents a constructed reality7.Some elements of the live delivery, such as improvised dialogues with the audi-ence, exempla, deictic expressions and other references to the context, are often preserved in written sermons. However, they are not only remnants from a single past event, but also rhetorical devices aimed at attaining more convincing moral and instructive power8. The homily – be it exegetical, panegyrical, polemical or Homilies which were preached ex tempore obviously represent the best sources for this type of infor-mation; those which were prepared beforehand or edited after the event rarely indicate the dynamics of a particular occasion.3 W. Mayer, John Chrysostom: Extraordinary Preacher, Ordinary Audience, [in:] Preacher and Audi-ence…, p. 115.4 Ibidem, p. 115–116.5 Ibidem, p. 126. W. Mayer gives a detailed account on the geographical, topographical, urban and architectural data in Chrysostom’s homilies, ibidem, p. 126–129; see also W. Mayer, The Homilies of St John Chrysostom. Provenance. Reshaping the Foundations, Rome 2005, p. 289–302.6 The medieval sermon both as oral and literary genre in the Western tradition is examined in: The Sermon, ed. B.M. Kienzle, Turnhout 2000, esp. p. 159–174; the signs of orality in written sermons and the tension between the written text and the oral discourse are summarized on p. 965–978. The volume gives also an extensive bibliography on general and specific questions.7 W. Mayer, John Chrysostom: Extraordinary Preacher…, p. 108.8 See, e.g. Preacher and Audience…, p. 13: By employing an informal and conversational method of dis-course, frequently inventing imaginary interlocutors, preachers may be inventing a dialogue which did 65Translation and Transformation of John Chrysostom’s Urban Imagery…ethical – is a rhetorical genre and John Chrysostom is famously one of its best champions. His eloquence, acquired through classical education, applies some methods of the second sophistic in Christian context9. He uses metaphors, com-parisons, ecphraseis and other vivid figures of speech in a wide range of topics. For instance, he borrows images from athletic games, medicine, sea and navigation, the hyppodrome, the theater, etc. – sources typical for the sophists10, but always directed by the preacher at moral or religious instruction.John Chrysostom’s urban imagery also falls into these two categories. On the one hand, his descriptions of the cityscape and the urban life give a snapshot of the era and of the actual moment of delivery. On the other hand, they are topoi that transcend the particular space and time and, as written literature, fit into oth-er contexts. By comparing some of these images with their translations into Old Church Slavonic I will try to determine how much of Chrysostom’s urban imag-ery was preserved, what was adapted to the new audience, and what remained unchanged and detached from the actual reality. Some aspects of this cultural transfer were addressed in previous (predominantly lexical) studies on Greek and Roman realia and their rendition in Old Church Slavonic11. Terms, names, and places from the classical and late antique world were not entirely unfamiliar to the edu-cated Slavic audience. Personal and geographical names, exotic food, and other objects are frequently mentioned in many genres of translated literature, such as biblical translations, historiography, hagiography, rhetoric, juridical texts, etc. At the same time, the abundant scribal errors suggest that many realia were mis-understood or entirely incomprehensible to the scribes.not really exist, and again: rhetorical devices such as dialogue and diatribal interjections to the audi-ence, the use of everyday imagery or exempla, and familiar topoi all must have helped to engage an audience which was expecting to some extent to be entertained, ibidem, p. 18.9 Cf. T. Ameringer, The Stylistic Influence of the Second Sophistic on the Panegyrical Sermons of St. John Chrysostom. A Study in Greek Rhetoric, Washington 1921 [= PSt, 5]; M.A. Burns, Saint John Chrysostom’s Homilies on the Statues. A Study of Their Rhetorical Qualities and Form, Washington 1930 [= PSt, 22].10 Examples from 4th century pagan orators, such as Himerius, Themistius and others, see in T. Am-eringer, The Stylistic Influence…, p. 17–19. Special chapters are dedicated to the praise of a country and of a city in Menander Rhetor, cf. Menander Rhetor, ed. et trans. D.A. Russell, N.G. Wilson, Oxford 1981, p. 28–43, 46–75.11 On this topic see e.g. the following research papers and the references therein: А.-М. ТОТОМА-НОВА, Сведенията за гръко-римския свят в славянския ексцерпт от Хрониката на Юлий Африкан. Проблеми на рецепцията, [in:] ΠΟΛΥΙΣΤΩΡ. Scripta slavica Mario Capaldo dicata, ed. К. ДИДДИ, Москва 2015, p. 316–327; Т. ИЛИЕВА, Античната култура през призмата на средновековния български книжовен език, Дзяло, 10, 2017, http://www.abcdar.com/magazine/X/ T.Ilieva_1314–9067_X.pdf [3 IV 2020]; Т. СЛАВОВА, Византийски реалии в преводаческата практика на старобългарските книжовници, [in:] Laudator temporis acti. Studia in memoriam Ioannis A. Božilov, vol. II, Ius, imperium, potestas litterae ars et archaeologia, ed. I.A. Biliarsky, Sofia 2018, p. 242–253, and many others.Aneta Dimitrova66The descriptions of the urban life in the late antique city do not always con-tain specific vocabulary and therefore may remain undetected and unexamined in lexical research. Some of the examples cited below depict scenes set in an urban environment, and the present study investigates not only how specific objects were named but also how ordinary situations were described. The selection of the examples is based on several criteria. The study is focused on genuine Chrysosto-mian homilies12 translated into Old Church Slavonic in the 9th–10th century. John Chrysostom’s authorship is an important criterium, because it gives a reliable point of reference in terms of time and place of origin of the homilies. The translations, however, belong to various anonymous Old Bulgarian translators and are mostly preserved in late manuscripts, some of which are not published13. Therefore, the manuscript tradition and the reception in the following centuries should also be kept in mind.The passages I will discuss below refer to the urban environment, public and private buildings, and the everyday life of the citizens. The examples are excerpted from the following Old Church Slavonic collections and manuscripts: the Chrys-orrhoas collection (Zlatostruy)14, Chrysostom’s On the Statues15, Codex Suprasl- 12 Cf. W. Mayer, The Homilies of St John Chrysostom…, p. 26–27; S. Voicu, Pseudo-Giovanni Crisos-tomo: i confini del corpus, JAC 39, 1996, p. 105–115.13 The Greek text of the examples below is cited according to the edition in Patrologia Graeca. The Old Church Slavonic translations are cited according to their respective editions, or according to the earliest accessible manuscripts, if they are unedited.14 Translated in the 10th century, its various versions are preserved in 12th–17th century manuscripts, see Я. МИЛТЕНОВ, Златоструй: старобългарски хомилетичен свод, създаден по инициатива на българския цар Симеон. Текстологическо и извороведско изследване, София 2013. The so-called Longer Zlatostruy, which is preserved almost only in Russian manuscripts from 15th century onwards, is unedited. Here it is cited after the earliest complete copy, a Russian manuscript from the Moscow Theological Academy (Russian State Library 173/I, No 43, 1474), cf. AРХИМ. ЛЕОНИД, Све-дение о славянских рукописях, поступивших из книгохранилища Свято-Троицкой Сергиевой лавры в библиотеку Троицкой духовной семинарии в 1747 г. (ныне находящихся в библиотеке Московской духовной академии), Вып. 2, Москва 1887, p. 66–68. It is available online http://lib-fond.ru/lib-rgb/173-i/f-173-i-43/ [11 VII 2020].15 De statuis (Ad populum Antiochenum homiliae 1–21), CPG 4330. The scholars are not unanimous about the date and the number of the Old Church Slavonic translations, cf. А.А. ТУРИЛОВ, Андри-анты, [in:] Православная энциклопедия, vol. II, Москва 2001, p. 410, http://www.pravenc.ru/text/115376.html [26 V 2020]; М.С. МУШИНСКАЯ, Адрианты Иоанна Златоуста в южнославян-ских и русских памятниках, [in:] Лингвистическое источниковедение и история русского язы-ка (2002–2003), Москва 2003, p. 27–74, http://www.ruslang.ru/istochnik_2003 [26 V 2020], espe-cially p. 28 – no data support a complete early Old Church Slavonic (Old Bulgarian) translation, only fragments are extant. But according to D. Bulanin, there was a nearly complete 10th-century transla-tion that was revised and preserved in later copies, cf. Д. БУЛАНИН, Текстологические и библиогра-фические арабески. Приложение V. Андрианты в старшем славянском переводе, [in:] Ката-лог памятников древнерусской письменности XI–XIV вв., Санкт-Петербург 2014, p. 489–510. It is unedited, here it is cited after a 16th-century Russian manuscript from the Russian State Library 304/I, No 151 (1597), cf. ИЕРОМ. ИЛАРИЙ, ИЕРОМ. АРСЕНИЙ, Описание славянских рукописей 67Translation and Transformation of John Chrysostom’s Urban Imagery…iensis (10th century)16, and Uspenskij codex (12th century)17. Other important Sla-vonic manuscripts, such as the Symeon florilegium (Izbornik 1073, 11th century), the Troickij codex (12th century), the Mihanović homiliary (13th century), and the German codex (14th century)18, did not provide any more examples. Although John Chrysostom is the most translated author in the medieval Slavonic litera-ture, one of the reasons for the scarcity of examples is the fact that only selected works and fragments were translated into Slavonic in the early period (9th–11th century) and the selection was based on their topic and function. As a result, catechetical, festal, and panegyrical homilies in the homiliaries are less likely to contain urban descriptions, compared to the ethical and even exegetical ser-mons, collected in instructive miscelanies such as Zlatostruy. Some brilliant ref- erences of John Chrysostom to the life of his fellow-citizens in Antioch and Con-stantinople were simply left out of the Slavonic selection. The translations of the later period (from 14th century onwards) are not taken into account, because they represent a different cultural context and principles of translation. Neverthe-less, the available instances are sufficient for drawing some conclusions about the way the Slavonic audience saw the 4th-century Byzantine city.City streets and buildingsThe first group of examples describes spacious streets, squares, and buildings. The two biggest cities of the Eastern Roman Empire in the 4th–5th century were impres-sive in terms of infrastructure and population even by today’s standards. They shared some features, e.g. busy streets and markets, big churches and tall buildings, a hippodrome, noise, nightlife, streetlights, baths, dense and stratified population. Reference to any of these features in Chrysostom’s homilies could pertain to either city. Aside from that, Antioch was famous for its porticoed streets, the proximity to the Orontes river and the mountain, and the luxurious suburb Daphne19, whereas библиотеки Свято-Троицкой Сергиевой лавры, Москва 1878, p. 125–128. It is available online http://lib-fond.ru/lib-rgb/304-i/f304i-151/ [11 VII 2020].16 Супрасълски или Ретков сборник, vol. I–II, ed. Й. ЗАИМОВ, М. КАПАЛДО, София 1982.17 Успенский сборник XII–XIII вв., ed. С.И. КОТКОВ, Москва 1971.18 These manuscripts contain Old Church Slavonic original and translated texts from the 9th–10th century, including Chrysostomian homilies, cf. Симеонов сборник (по Светославовия препис от 1073 г.), vol. I, Изследвания и текст, София 1991; vol. II, Речник-индекс, София 1993; vol. III, Гръцки извори, София 2015; J. Popovski, F.J. Thomson, W.R. Veder, The Troickij Sbornik (Cod. Moskva, GBL, F.304 (Troice-Sergieva Lavra) N 12). Text in Transcription, ПК 21–22, 1988; Mihano-vić Homiliar, ed. R. Aitzetmüller, Graz 1957; Е. МИРЧЕВА, Германов сборник от 1358/1359 г. Изследване и издание на текста, София 2006.19 For a detailed study on the topography of Antioch in John Chrysostom’s works see W. Mayer, The Topography of Antioch Described in the Writings of John Chrysostom, [in:] Les sources de l’histoire du paysage urbain d’Antioche sur l’Oronte. Actes des journées d’études des 20 et 21 septembre 2010. Colloques de l’université Paris 8, ed. C. Saliou, Paris 2012, p. 81–100, with an exhaustive list of topo-graphic data on p. 89–100.Aneta Dimitrova68Constantinople was surrounded on three sides by the sea and was distinguished by the emperial palace and the occasional presence of the emperor20.The following example mentions not only the noise in the (unspecified) city, but also suburbs and houses with golden roofs and triclinia:[1] Τοῦτο γὰρ μέγιστον ἐγκώμιόν ἐστι τῆς ἡμετέρας πόλεως, οὐ τὸ θορύβους ἔχειν καὶ προάστεια, οὐδὲ χρυσορόφους οἴκους καὶ τρικλίνους, ἀλλὰ τὸ ἔχειν δῆμον σπουδαῖον καὶ διεγηγερμένον (De paenitentia homilia 3, CPG 4333; PG, vol. XLIX, col. 291).This is the biggest praise to our city, not its noise and suburbs, nor its golden-roofed houses furnished with couches, but its devoted and diligent people.се бо сть похвала вашего града. ꙗкоже н плща мѣт н хлѣвьць. н ꙁлатокровьныхъ домовъ. н полатъ. нъ же мѣт народъ поспѣшвъ въставленъ (Uspenskij codex, 12th c., f. 180v, ed. С.И. КОТКОВ, Успенский сборник…, p. 305).There are several issues in the Slavonic translation that should be addressed. First of all, it is the overall meaning of the phrase. According to John Chrysostom, the noise, the rich houses and the suburbs (where, supposedly, the wealthy citi-zens could retreat in their villas) are inherent to his city, but it is the people who are more praiseworthy21. However, the Slavonic translation suggests that this town has nothing but its people. The meaning is further adjusted by the vocabulary. Chrysostom’s “our city” (τῆς ἡμετέρας πόλεως) has become “your city” (вашего града, *ὑμετέρας) – a common itacistic error, but also a lost connection between the preacher and the audience. The Slavonic хлѣвьць ‘field, farm’ hardly implied luxury22, and the big Roman house with many rooms, including the typical din-ing-room triclinium furnished with three couches, where the guests reclined for dinner, was rendered as домы полаты, ‘houses and palaces’.The following passage refers undoubtedly to Antioch:20 Constantinople is well studied, see e.g. C. Mango, Le développement urbain de Constantinople (IVe–VIIe siècles), Paris 1985; Byzantine Constantinople. Monuments, Topography and Everyday Life, ed. N. Necipoglu, Leiden 2001 [= MMe, 33]; P. Magdalino, Studies on the History and Topography of Byzantine Constantinople, Aldershot 2007.21 The reference to the suburbs in this homily was one of the reasons for it to be located in Antioch because of its famous suburb Daphne. However, W. Mayer questions the validity of this criterion and comments on the meaning of the plural προάστεια: Although in its singular form προάστειον is genuinely used by Chrysostom to indicate a physical suburb, it is possible that when the term appears in its plural form without a definite article, as in the instance adduced, it is being employed by him to describe not a physical area beyond the confines of the city but the dwellings or estates situated in those areas, W. Mayer, The Homilies of St John Chrysostom…, p. 389.22 Cf. the next example below. There are also other instances where хлѣвьць, хлѣвца, хлѣвьнца correspond to προάστεια, cf. И. СРЕЗНЕВСКИЙ, Материалы для словаря древнерусского языка по письменным памятникам, vol. III, Санкт-Петербург 1912, col. 1376.69Translation and Transformation of John Chrysostom’s Urban Imagery…[2] Ὅταν ἐθέλῃς τῆς πόλεως εἰπεῖν ἐγκώμιον, μή μοι τὴν Δάφνην εἴπῃς τὸ προάστειον, μηδὲ τὸ πλῆθος καὶ μῆκος τῶν κυπαρίσσων, μηδὲ τὰς πηγὰς τῶν ὑδάτων, μηδὲ τὸ πολλοὺς τὴν πόλιν οἰκεῖν ἀνθρώπους, μηδὲ τὸ μέχρι βαθυτάτης ἑσπέρας ἐπὶ τῆς ἀγορᾶς διατρίβειν μετὰ ἀδείας πολλῆς, μηδὲ τῶν ὠνίων τὴν ἀφθονίαν (Ad populum Antiochenum homilia 17, CPG 4330; PG, vol. XLIX, col. 179).Whenever you want to praise the city, do not tell me about the suburb of Daphne, the mul-titude and magnitude of the cypresses, and the water fountains, nor that many people live in the city and one can walk around the agora deep into the night without any fear, nor about the abundance of goods.егда хощеш градꙋ повѣдат хвалꙋ, не гл͠ м даѳн хлѣвець, н мноⷤства н высост кпарсныѧ нї стоьнкъ водныⷯ, н жвꙋщъ многъ л͠вкъ по градѣ не еже до веера темна на торѕѣхъ ходт съ пространьствоⷨ мноꙁѣмъ, н кꙋпован ѡбїлѧ (Russian State Library 304/I, No 151, 1597, f. 183v).Unlike the first example, this translation is faithful and unadapted, including the mention of the Daphne suburb and the cypresses in Antioch. Since the entire homiletical series Ad populum Antiochenum is famously dedicated to this particu-lar city and its people, it is not surprising that the references to the topography remained unchanged. In the translation, it constructs a “rhetorical” reality, which is consistent within its own context and is not necessarily connected to the sur-roundings of the Slavonic reader23. Another passage from the same homily also mentions the columns and stoas in the city and adds some information about the lexical variety of the translation:[3] Οὐ τὸ μητρόπολιν εἶναι, οὐδὲ τὸ μέγεθος ἔχειν καὶ κάλλος οἰκοδομημάτων, οὐδὲ τὸ πολλοὺς κίονας, καὶ στοὰς εὐρείας καὶ περιπάτους (Ad populum Antiochenum homilia 17, CPG 4330; PG, vol. XLIX, col. 176).Not because it is a capital city, nor because of its big and beautiful buildings, numerous columns, broad colonnades and covered walks.не еже мтропол быⷮ нї еже велества мѣт добротꙋ ꙁданї, нї еже л многы столпы пркрылы ѹтелнїца (Russian State Library 304/I, No 151, 1597, f. 181r)24.23 Such a “constructed reality” is present not only in translation, but also in Chrysostom’s original: The information that is supplied is largely allusive rather than specific. In addition, the way in which John refers collectively to “the baths”, “the agora”, and “the theatre” leads one to suspect that for peda-gogical and polemical purposes he operates largely within a symbolic topography, W. Mayer, The Topo- graphy of Antioch…, p. 86.24 In the 15th-century translation (or revision, cf. note 15 above) of the homilies On the Statues, this sentence is as follows: не еже мтрополїа быт. н же еже велество мѣт. добротѹ ꙁданїѡмь, н же еже мнѡгы стлъпы пртвор мѣт ѿвод (Vladislav the Grammarian’s manuscript Rila 3/6, 1473, f. 254r). The Slavonic пртворъ is a standard rendition of the Greek word στοά, two more instances are listed below. The word отъводъ is not an exact match for the Greek περίπατος (‘walk’), but is etymologically closer than the word ѹтельнца in the example cited above.Aneta Dimitrova70The translation in this example (especially пркрылы ѹтелнїца) does not convey the exact meaning of the Greek στοὰς εὐρείας καὶ περιπάτους. The contex-tual synonyms “columns”, “colonnades” and “(covered) walks” allude to the famous covered streets in Antioch. The Slavonic ѹтельнца ‘school’ is an unusual coun-terpart to περίπατος ‘walk, place for walking’. However, it corresponds to one of the secondary meanings of this word ‘philosophical school’25 and reveals either a very well educated translator, or a random and inexplicable mistake.The next two examples are from a Constantinopolitan homily, In sancto hiero-martyre Phoca (CPG 4364). The selected passages are just a small sample of all the references to the imperial city. They describe the surroundings and refer to a par-ticular event at the time of the delivery of the sermon:[4] Λαμπρὰ γέγονεν ἡμῖν χθὲς ἡ πόλις, λαμπρὰ καὶ περιφανὴς, οὐκ ἐπειδὴ κίονας εἶχεν, ἀλλ’ ἐπειδὴ μάρτυρα πομπεύοντα ἀπὸ Πόντου πρὸς ἡμᾶς παραγενόμενον (In sancto hieromar-tyre Phoca, CPG 4364; PG, vol. L, col. 699).The city was bright yesterday, bright and prominent, not because it has columns, but because of the martyr who came to us in a procession from the sea.Свѣтелъ намъ граⷣ свѣтелъ ⷭ҇тенъ. не мⸯже мраморѧн, маⷮ стлъп стоѧща. нъ елмаже м͠нїка одолѣвⸯша. ѿ морѧ к наⷨ прведеⷮ (Longer Zlatostruy, homily No 6, ed. Я. МИЛТЕНОВ, Златоструй…, p. 264).[5] Ἀπελείφθης χθές; παραγενοῦ κἂν σήμερον, ἵνα ἴδῃς αὐτὸν εἰς τὸν οἰκεῖον χῶρον ἀπαγό-μενον. Εἶδες αὐτὸν διὰ τῆς ἀγορᾶς ἀγόμενον; βλέπε αὐτὸν καὶ διὰ τοῦ πελάγους πλέοντα (In sancto hieromartyre Phoca, CPG 4364; PG, vol. L, col. 699).Did you miss it yesterday? Then be here today and see him being brought back to his own place. Did you see how he was carried across the agora? Watch him cross the sea, as well.оста л вера. прїд поне въ вⸯторы д͠нь. да вдш. на свое мѣсто несома. вдѣ л ресъ торгъ несома. вжⷣь ресъ пѹнѹ пловѹща (Longer Zlatostruy, homily No 6, ed. Я. МИЛТЕНОВ, Златоструй…, p. 264).Chrysostom’s homily celebrates the two-day procession carrying the saint’s relics through the city and across the sea – a single event, which connects the preacher and his audience with their shared actual reality26. To the Slavonic read-er (and, indeed, to every member of an audience other than the one present at 25 Cf. LSJ, p. 1382, s.v. περιπατέω – one of the meanings is ‘walk about while teaching, discourse’ and ‘dispute, argue’, and for περίπατος cf. ‘school of philosophy, first used of the Academy’, and ‘generally, any school of philosophy’.26 On the date and provenance of the homily see Saint John Chrysostom, The Cult of the Saints. Select Homilies and Letters, praef. et trans. W. Mayer, B. Neil, New York 2006, p. 75–76.71Translation and Transformation of John Chrysostom’s Urban Imagery…the delivery of the sermon on this day in Constantinople) the deictic χθές ‘yesterday’ was already anachronic, hence it was omitted in the translation of example 4. The word вьера in example 5, together with the sea and the “marble” columns of the city, create a new “rhetorical” reality, which includes also the images of other exempla, ecphraseis, and comparisons.Chrysostom’s cities were lively and dynamic, and their social and economic centres were the city squares and marketplaces, the agorai27. Chrysostom trans-fers the realistic image of the agora into the constructed reality of the rhetorical figures of speech. In the following exemplum he compares the troubled soul of an angry man to an agora and it is difficult to differentiate between the 4th-century reality and the topos. The description is very realistic, but its function in the text is purely rhetorical:[6] ἡ τούτου δὲ (sc. ψυχή) ἀγορᾷ καὶ θορύβῳ καὶ ταῖς μέσαις τῶν πόλεων, ἔνθα πολλὴ ἡ κραυγὴ τῶν ἀπιόντων, τῶν ἐπανιόντων, καμήλων, καὶ ἡμιόνων, καὶ ὄνων, μεγάλα τοῖς προσιοῦσιν ἐμβοώντων, ὥστε μὴ καταπατηθῆναι, καὶ πάλιν ἀργυροκόπων, χαλκοτύπων ἑκατέρωθεν ἐκκρουόντων, καὶ τῶν μὲν ἐπηρεαζομένων, τῶν δὲ ἐπηρεαζόντων (In Acta apostolorum homilia 6, CPG 4426; Oxford, New College No 75, p. 143)28.The soul of a troubled man] is similar to the clamour at the marketplace and the city streets, there is great noise from people coming and going, camels, mules, and donkeys, people shouting to the passers-by, so that they do not get trampled; and silversmiths and black-smiths hammering from both sides; and people either bullying, or being bullied.а гнѣвлваго подобна тръжщѹ. дѣже всѧкь плщь есть. бесѣдаⷨ (v.l. стьгнѣмъ) граⷣныⷨ. дѣже многъ кль (v.l. плщь) мѧтѣжь. сходѧщⷯ ꙁ града входѧщⷯ. вѣ-льблѹⷣ мьщатъ. женѹщⷯ велм крат по нⷯ. да л͠кь (v.l. къ предънїмъ) не поперѹть. пакы подобна есть къ ꙁлатареⷨ. къ желѣꙁоковцеⷨ (v.l. кърꙗмъ), обо-юдѹ клекьтанїе (v.l. клюкан) творѧщⷨ. храм т плън тепта (v.l. клюка) клета (v.l. тъпъта). така т есть гнѣвлвыⷯ д͠ша. (Longer Zlatostruy, homily No 90, Russian State Library 173/I, No 43, f. 457v29).27 For his congregation, the main axes of life seem to be: the house, the agora, the baths and the church, see L.A. Lavan, The Agorai of Antioch and Constantinople as seen by John Chrysostom, BICS 50, Issue Sup. 91, 2007, p. 157–167.28 The Old Church Slavonic translation is closer to the so called “rough” version, here cited after one of the oldest manuscript copies, Oxford, New College No 75 (10th–11th century). The text published in Patrologia Graeca has a somewhat different wording, cf.: ἡ ἐκείνου δὲ ἀγορᾷ καὶ θορύβῳ, ἔνθα πολλὴ ἡ κραυγὴ τῶν ἀκολούθων καὶ καμήλοις, καὶ ἡμιόνοις. καὶ ὄνοις, μεγάλα τοῖς προσιοῦσιν ἐμβοώντων, ὥστε μὴ καταπατηθῆναι; οὐχ ἡ μὲν τοῦ τοιούτου ταῖς μέσαις πάλιν τῶν πόλεων ἐοικυῖά ἐστιν, ἔνθα νῦν μὲν ἐντεῦθεν ἀργυροκόπων, νῦν δὲ ἐκεῖθεν τῶν χαλκοτύπων ὁ ἦχος πολὺς γίνεται, καὶ οἱ μὲν ἐπηρεάζουσιν, οἱ δὲ ἐπηρεάζονται (PG, vol. LX, col. 61).29 The variants are after the so-called Shorter Zlatostruy, earliest copy Saint Petersburg, Russian Na-tional Library, F.п.I. 46 (Russian, 12th century), edited in Т. ГЕОРГИЕВА, Златоструй от XII век, Силистра 2003.Aneta Dimitrova72Whether the camels in the marketplace were real or imaginary for the Con-stantinopolitan audience, they were not part of the everyday life of the 10th-cen-tury Bulgarian translator (nor of the 15th-century Russian scribes and readers). However, any medieval audience could relate to the clamour and racket in the market (ἀγορά, тръжще), the mules and their shouting owners on the streets (μέσαι, стьгнꙑ, incorrectly бесѣдꙑ), or the deafening noise coming from the blacksmiths’ workshops. Some variants of the words meaning ‘noise’ in the Sla-vonic manuscripts also suggest familiarity with the scene, which allows for a freer interpretation by the scribes.Private life and daily routineThe urban environment and the public spaces were the scene where a large part of the citizens’ daily life took place. This interaction between man and city is pre-sented in the following several examples. In the evenings the narrow streets and squares were not the best place for the poor and the homeless:[7] καὶ ὁ ἀποστερηθεὶς ὑπὸ τῆς τῶν ἀναγκαίων δάκνηται χρείας, καὶ ὀλοφύρηται, καὶ μυρί-ους ἐφέλκηταί σοι κατηγόρους, καὶ τῆς ἑσπέρας καταλαβούσης περιίῃ τὴν ἀγορὰν, ἐν τοῖς στενωποῖς ἐντυγχάνων πᾶσι, καὶ διαπορούμενος καὶ οὐδὲ ὑπὲρ τῆς νυκτὸς θαῤῥεῖν ἔχων (In Epistulam primam ad Corinthios homilia 11, CPG 4428; PG, vol. LXI, col. 94).And the deprived [by you] may be bitten by the most basic needs, and lament himself, and summon thousands of accusers upon you; and when the evening comes, he may go around the market-place, encountering all sorts of things in the alleys, and be at a loss, not daring to spend the night. облхованы тобою обходⷮ пща ща плаа сѧ рыдаѧ. веерѹⷤ бывⸯшѹ обходⷮ ѹл-цѣ не мѣѧ гдѣ главы поⷣклонт. да нощїю ходѧ помы дѣеⷮ на тѧ къ всѣⷨ (Longer Zlatostruy, Homily No 74b, Russian State Library 173/I, No 43, f. 406r).At the same time, the (wealthy) citizen in Chrysostom’s reality, a member of his audience, visited the public baths in the evening before the late supper, after completing his daily chores in the agora:[8] Καὶ σὺ μὲν ἐκ βαλανείου λελουμένος ἐπανέρχῃ, μαλακοῖς θαλπόμενος ἱματίοις, γεγη-θὼς καὶ χαίρων, καὶ ἐπὶ δεῖπνον ἕτοιμον τρέχων πολυτελές (In Epistulam primam ad Corin-thios homilia 11, CPG 4428; PG, vol. LXI, col. 94).And you come back refreshed after bathing, kept warm in soft garments, cheerful and happy, rushing to a lavish dinner.73Translation and Transformation of John Chrysostom’s Urban Imagery…ты же ѿ банѧ множцею ꙁьмыв сѧ деш. вь мѧкькы рꙁы облъенъ. раⷣуа сѧ веселѧ сѧ на велкѹ веерю грѧды (Longer Zlatostruy, Homily No 74b, Russian State Library 173/I, No 43, f. 406r).Тhe Slavonic translation in example 7 mentions only the dark and narrow streets (στενωπός, ѹлца, see also example 11 below) and omits the agorai, but although it simplifies the wording of the original passage, it renders truthfully its general meaning. The next example 8 from the same episode is not adapted to the Slavonic audience and the translation keeps both the baths (βαλανεῖον банꙗ) and the sumptuous supper (δεῖπνον πολυτελές велка веерꙗ). The same image can be found also in other homilies30, e.g.:[9] Ἑσπέρας δὲ πάλιν καταλαβούσης, οἱ μὲν εἰς λουτρὰ καὶ ἀνέσεις σπεύδουσιν (In Epistu-lam primam ad Timotheum homilia 14, CPG 4436; PG, vol. LXII, col. 577).The evening comes again, and some are in a hurry to the public baths and relaxation.веерѹ же пакы доспѣвⸯшѹ. он бо банѧ напокоѧ тⸯщат сѧ (Longer Zlatostruy, Homily No 37, Russian State Library 173/I, No 43, f. 229r).The Greek word δεῖπνον from example 8 can denote any meal during the day, either lunch, dinner, or supper, but the context suggests that both going to the baths and the meal afterwards happen in the evening and the Slavonic transla-tion reflects this accordingly. Examples 9 and 10 are unambiguous in this respect – the visit to the public baths (λουτρά) is in the evening and is followed by a ban-quet. This way of life was not accessible to all the citizens, cf. the continuation of example 9 below.The houseThe following examples depict domestic scenes in big Roman houses with ser-vants and lavish banquets:[10] ἐκεῖνοι δὲ τῶν πόνων ἀπολύσαντες ἑαυτοὺς, τότε τῇ τραπέζῃ προσανέχουσιν, οὐκ οἰκετῶν πλῆθος ἐγείροντες, οὐδὲ περιτρέχοντες τὴν οἰκίαν, οὐδὲ θορυβοῦντες, οὐδὲ ὄψα πολλὰ παρατιθέμενοι, οὐδὲ κνίσσης γέμοντα, ἀλλ’ οἱ μὲν ἄρτον μόνον καὶ ἅλας, οἱ δὲ 30 Leslie Dossey’s interesting study about the nightlife in the 4th-century big cities Antioch and Con-stantinople explores the shift of the main occupations of the citizens towards later hours (compared to the country and to earlier time-periods). Afternoon naps, baths, shopping and supper occur sev-eral hours later that before, cf. L. Dossey, Night in the Big City. Temporal Patterns in Antioch and Constantinople as Revealed by Chrysostom’s Sermons, [in:] Revisioning John Chrysostom, ed. C.L. de Wet, W. Mayer, Leiden–Boston 2019 [= CAEC, 1], p. 698–732.Aneta Dimitrova74ἔλαιον προστιθέντες, ἕτεροι δὲ, ὅσοι ἀσθενέστεροί εἰσι, καὶ λαχάνων ἔχονται καὶ ὀσπρίων (In Epistulam primam ad Timotheum homilia 14, CPG 4436; PG, vol. LXII, col. 577).The other ones attend to the table after finishing with their labour, without awakening a multitude of slaves, nor running around the house and raising a clamour, nor having many dishes full of meats, but some put on the table only bread and salt, others add olive-oil, and some – the weaker ones – have vegetables and beans.а ѡв трѹдовъ еѧ прощьше т. тогⷣа на трапеꙁе сѧдѹⷮ не многъ рабъ съꙁывающе, н рщѹть по дворѹ, н плещѹщь, н вѣрныⷯ многоцѣньнъ преⷣлагающе. н скварамъ радꙋю-ще. нъ ов хлѣбъ, т соль. ов масло пролїѧвⸯше. дрѹꙁї же аще сѹⷮ болⸯн то ꙁелїа сѧ прїемлюще сова (Longer Zlatostruy, Homily No 37, Russian State Library 173/I, No 43, f. 229r).This contrast between the wealthy citizens with their baths, big households and banquets, and the poor (the working people, τῶν πόνων ἀπολύσαντες ἑαυ-τοὺς), the ones who have simple meals and a small or no house at all, is a recur-ring motif in Chrysostom’s works:[11] Ὅταν οὖν ἀνέλθῃς οἴκαδε, ὅταν ἐπὶ τῆς εὐνῆς ἀνακλιθῇς, ὅταν φῶς ᾖ περὶ τὸν οἶκον λαμπρὸν, ὅταν ἑτοίμη καὶ δαψιλὴς ἡ τράπεζα, τότε ἀναμνήσθητι τοῦ ταλαιπώρου καὶ ἀθλί-ου ἐκείνου, τοῦ περιιόντος κατὰ τοὺς κύνας ἐν τοῖς στενωποῖς καὶ τῷ σκότῳ καὶ τῷ πηλῷ (In Epistulam primam ad Corinthios homilia 11, CPG 4428; PG, vol. LXI, col. 94).When you come home, when you lay down on the couch, when the lights shine bright in the house, when the table is ready and full, then remember that miserable and unhappy one, walking down the alleys like a dog, in darkness and mud.ты же егⷣа прдеш вⸯ домъ с. сѣⷣшѹ т поставѧⷮ преⷣ тобою велкѹ плънѹ трапеꙁꙋ. тогⷣа въспомѧн оканⸯнаго оного ѡбьходѧщаго. акы ѱа по ꙋлцаⷨ. въ тмѣ въ калѣ (Longer Zlatostruy, Homily No 74b, Russian State Library 173/I, No 43, f. 406r).Example 11 presents a picture, where at least some streets or alleys (the same στενωποί from example 7) are dark and not paved. It also makes the transition from the public space into the residential area – into the dining-room of a Roman house. This is where the Slavonic translation shows some deviations and adapta-tions. The phrase “lights shining bright round the house” is omitted altogether (perhaps the medieval Bulgarian house was darker than its Byzantine counter-part, but this cannot be the only explanation of the omission). Chrysostom’s citizen reclines on a couch for supper in the customary manner (ἐπὶ τῆς εὐνῆς ἀνακλιθῇς) – probably the couch in the triclinium from example 1, whereas the man in the constructed reality of the Slavonic translation sits down (сѣⷣшѹ т) and someone else (a slave? a servant? a wife?) puts the table in front of him. The 75Translation and Transformation of John Chrysostom’s Urban Imagery…less fortunate men from example 10 are at the table in an unspecified position (τῇ τραπέζῃ προσανέχουσιν) and again “sitting” in the translation. These subtle deviations of the translation suggest that the Slavonic audience did not differenti-ate between sitting at the table and reclining on the couch of a triclinium.The next examples also give some architectural details, e.g. the following image of a Roman house:[12] Μανθανέτωσαν οἱ τὰς λαμπρὰς οἰκοδομοῦντες οἰκίας, καὶ τὰς εὐρείας στοὰς, καὶ τοὺς μακροὺς περιβόλους, ὅτι οὐκ εἶχεν ὁ Χριστὸς ποῦ τὴν κεφαλὴν κατακλῖναι (De proditione Iudae, CPG 4336; PG, vol. XLIX, col. 378).And those who build splendid houses and wide porticoes, and long courtyards, let them know that Christ did not have a place to rest his head.да навꙑкнѫтъ же свѣтлꙑѧ домꙑ дѣлаѭтъ. пространꙑѧ пртворꙑ. длъгꙑѧ дворꙑ. ꙗко не мѣ х͠с кде главꙑ подъклонт (Codex Suprasliensis, ed. Й. ЗАИМОВ, М. КАПАЛДО, Супрасълски или Ретков сборник…, p. 416).In cases like this one, the description of the house is a rhetorical device, but it mentions a colonnaded courtyard. The Slavonic word пртворъ, which is a com-mon term of church architecture, is a regular counterpart to the Greek στοά (cf. note 24 above). The next example does not give any architectural details, but the translation enhances the episode:[13] Ἂν διακύψῃς εἰς τὸν στενωπὸν, οὐκ ἀκούσῃ οὐδὲ φωνῆς· ἂν ἴδῃς εἰς τὴν οἰκίαν, πάντας ὄψει καθάπερ ἐν τάφῳ κειμένους (In Acta apostolorum homilia 26, CPG 4426; PG, vol. LX, col. 202).If you peek out to the street, you will not hear a sound; if you look into the house, you will see everybody lying as if in a tomb.аще бо снкнеш на стегны <с> полаты то не слышш гл͠са, нї ного нтоⷤ. аще л снк-неш въ дворъ сво с полаты. то все вдш акы въ гробѣ лежаще (Longer Zlatostruy, Homily No 41, Russian State Library 173/I, No 43, f. 256v).The Greek sentence suggests that an observer is looking through the win-dow out (towards the narrow street), and in (towards the house). In the Slavonic text the house is a palace (полата) and the observer is looking out to the street, and then back into a courtyard (въ дворъ). An inner courtyard is imaginable only in a big building – in the palace or in a monastery.In the following description of the morning routine of a common citizen, a small alteration in the translation gives us an idea about the layout of the house:Aneta Dimitrova76[14] Ἡμεῖς μὲν γὰρ ἅμα διαναστάντες, καθήμεθα ἐπὶ πολὺ διατεινόμενοι, πρὸς χρείαν ἀπερ-χόμεθα, εἶτα νιπτόμεθα τὴν ὄψιν, τὰς χεῖρας· μετὰ τοῦτο ὑποδήματα καὶ ἐνδύματα λαμβά-νομεν, καὶ πολὺς ἀναλίσκεται καιρός (In Epistulam primam ad Timotheum homilia 14, CPG 4436; PG, vol. LXII, col. 575).As soon as we wake up, we sit up and stretch out, we answer the call of nature, then we wash our face and hands, afterwards we take our shoes and clothes, and a lot of time passes.въставьше бо мы сѣдⷨ. много пролѧкающе сѧ. ꙁадъ демъ. тае ѹмываеⷨ лца рѹцѣ по томь. въꙁемлеⷨ же одежⷣѹ онѹщꙋ т много врѣмѧ погѹбⷨ (Longer Zlatostruy, Homily No 37, Russian State Library 173/I, No 43, f. 227v).Early in the morning, after sitting up in the bed and stretching, and before washing and dressing, the citizen relieves himself (goes πρὸς χρείαν). In the translation, this happens behind the house, or at the back (ꙁадъ демъ). This deviation in the Slavonic translation alludes to an area of the house, or outside the house, that is otherwise rarely mentioned. It is also another point of difference between the well-equipped Roman house and the average medieval Slavic houses.The last example, which is another description of a building, also gives some interesting information about architectural terminology and adaptations of the source text:[15] Ὥσπερ γὰρ οἰκοδόμος θεμελίους θεὶς, τοίχους ἀναστήσας, ὄροφον καμαρώσας, τὴν καμάραν ἐκείνην εἰς ἕνα μέσον συνδήσας λίθον, ἂν ἐκεῖνον ἀφέλῃ, τὸν πάντα τῆς οἰκοδο-μῆς διέλυσε σύνδεσμον (Adversus Iudaeos oratio 4, CPG 4327; PG, vol. XLVIII, col. 881).Just like the architect, who lays the foundations, builds the walls, furnishes the roof with a vault, and locks that vault with a single stone in the middle, if he takes away that stone, the whole structure of his building will collapse.ꙗкоже бо ꙁжтель основанї его положвъ стѣны поставвъ. оловоⷨ стропъ покрывъ. ко-марѹ посредѣ еднѣмь камкоⷨ свѧꙁавъ. аще того камыка ѿметь. все съꙁданїе раꙁорт сѧ (Longer Zlatostruy, Homily No 8, Russian State Library 173/I, No 43, f. 76v).The building in the Greek comparison has a vaulted roof built of stone, with a keystone on the top – a structure, characteristic not for a house, but for a church or a similar edifice. In the translation, it is covered with lead (оловоⷨ стропъ покрывъ) and there is also a stone on the top of the dome31. Other examples of 31 The Greek word καμάρα is rendered with the borrowing комара, witnessed in other 10th-century translations such as John the Exarch’s Bogoslovie and Pseudo-Kaisarios’ Erotapokrisis, see И. СРЕЗ-НЕВСКИЙ, Материалы для словаря…, vol. I, Санкт-Петербург 1893, col. 1263–1264 (s.v. комара); Я. МИЛТЕНОВ, Диалозите на Псевдо-Кесарий в славянската ръкописна традиция, София 2006 p. 544.77Translation and Transformation of John Chrysostom’s Urban Imagery…the word олово in some Slavonic texts refer specifically to covering churches with lead32. It seems that it was not unusual for a dome to be coated with lead, which suggests that the vault in this example was associated with a leaden church-dome – a notion which is not explicit in the Greek source.Wide porticoes and long courtyards (cf. example 12) were not typical features of the medieval Bulgarian house, except for the ruler’s palace. Even if we assume that the initial audience of the Old Church Slavonic translations was well familiar with the Byzantine culture, the topography of Constantinople, its squares, col-umns and obelisks, the surrounding sea etc. (cf. examples 4 and 5), the readers throughout the Slavic world in the following centuries most probably did not see villas with spacious courtyards, porticoes and a large body of water from their windows.What did the 10th-century Bulgarian see from his window? Unlike Antioch and Constantinople, there is only limited archeological data about the medieval Bulgarian town and almost no information about the everyday life of the com-mon citizen33. Pliska and Preslav, the two capitals of the First Bulgarian Kingdom (7th–11th century), had some monumental architecture, such as the ruler’s palace34, churches, and city walls. However, the cities were in steady decline from the end of the 10th century onwards and lost their significance in the Second Bulgarian Kingdom (12th–14th century). According to the archeological and historical stud-ies, the medieval Bulgarian fortified town had a residential area outside the city walls. Most people lived in small semi-dug-in houses35, the door opened directly to the street. In the later centuries the houses were made of stone and could have a backyard with service buildings. Some prominent citizens owned two-storey houses with many rooms. The marketplaces and the town-squares, formed from the intersection of two roads, were an important economic and social element 32 Cf. И. СРЕЗНЕВСКИЙ, Материалы для словаря…, vol. II, Санкт-Петербург 1902, col. 661, s.v. олово: “съꙁдана есть церковь велка клѣтьскы, покрыта же есть всꙗ оловомъ”; „Обновлена бысть церкы свꙗтаꙗ Богородца… покрыта бысть оловомь от верхѹ до комаръ до пртворовъ”.33 The description of the medieval Bulgarian town is based on several general studies: К. МИЯТЕВ, Архитектурата в средновековна България, София 1965; С. ЛИШЕВ, Българският средновеко-вен град. Oбществено-икономически облик, София 1970; М. ХАРБОВА, Укрепеният български средновековен град XIII–XIV в., София 1979; Д. ПОЛИВЯННИ, Средновековният български град през XIII–XIV в. Очерци, София 1989; А. МИЛАНОВА, Градът във византийска България (XI–XII в.), [in:] Средновековен урбанизъм. Памет – Сакралност – Традиции, София 2007, p. 7–29; A. Aladzhov, The Byzantine Empire and the Establishment of the Early Medieval City in Bulgaria, [in:] Byzanz – das Römerreich im Mittelalter, vol. III, Peripherie und Nachbarschaft, ed. F. Daim, J. Drauschke, Mainz 2010, p. 113–158.34 …since the palace covers a large area, its central part was taken by a courtyard enclosed by the bal-conies of the building, A. Aladzhov, The Byzantine Empire…, p. 120. This structure was probably burnt down at the beginning of the 9th century.35 Ibidem, p. 116.Aneta Dimitrova78of the medieval town. There is no information about city lights, but there were baths and a sewage system in the First Bulgarian kingdom36 at least at the palace. In the 13th–14th century the water supply was provided by cisterns and wells37.The medieval Bulgarian town shared some features with the Byzantine major cities and differed in others. From the examples above, and the entire history of translation, transmission, and reception of Byzantine texts in the medieval Slavonic literature, it is evident that the homilies of John Chrysostom had many points of reference to the actual reality of 4th-century Antioch and Constantinople that were not present to the Slavonic audience. The translators of Chrysostom’s homilies, however, did not adapt each detail that might be unfamiliar. Many realia are unchanged in the translation – there are exact renditions of stoas, columns, baths, vaulted roofs and camels in the agora, athletic games and theatrical perfor-mances (not included in this study), etc. At the same time, some passages were slightly adapted without damaging the general meaning, e.g. the villas with tri-clinia in the suburbs, where people dined lying on couches, became “fields” and “palaces” (example 1), and the master sat at the table for dinner (example 11).Urban images were transferred from the Byzantine world into medieval Bul-garia also in other literary genres, such as the juridical literature. One of the law texts translated from Greek into Old Church Slavonic pertains to the same topic – life in the city – and uses vocabulary similar to the examples commented above. This text is the Procheiros nomos – a Byzantine juridical compendium based on Justinian’s law38. Title 38 of the Procheiros nomos deals specifically with the urban environment, buildings and renovations, private and public property, relationships between neighbours, co-ownership, etc. The Slavonic translation (the earliest witness is from the 13th century) contains numerous technical terms which are a significant contribution to the terminological vocabulary of the Slavonic lan-guage. This text deserves special attention, but here I will briefly comment on some issues which are relevant to the present study.The Procheiros nomos settles legal matters in the Eastern Roman Empire which are irrelevant to the medieval Bulgarian, Serbian, or Russian reality. For instance, the cases in chapters 14 and 18 involve multistorey residence buildings, which were not typical for the medieval Bulgarian town (upper floors should not be heavier than the ground-floor can support, and the residents of the ground-floor may not conduct smoke into the homes of their neighbours above). Other 36 Ibidem, p. 118.37 Д. ПОЛИВЯННИ, Средновековният български град…, p. 134–135.38 A lexical study and an edition of the Slavonic translation of title 38, with additional bibliography, see in: М. ЦИБРАНСКА-КОСТОВА, Градският закон и градското благоустройство в южнославян-ски контекст, СЛ 57–58, 2018, p. 163–193. The Greek text is available in: J. Zepos, P. Zepos, Pro-chiron, [in:] Jus Graecoromanum, vol. II, Leges imperatorum Isaurorum et Macedonum, Aalen 1962, p. 114–228. The numbers of the chapters below follow the segmentation in the Slavonic translation.79Translation and Transformation of John Chrysostom’s Urban Imagery…chapters deal with topography and landscaping characteristic to the Mediterra-nean – according to chapters 5 and 6, the residents have the right to preserve their unhindered view to the sea (the city is explicitly named – вⸯ семь богатѣ-мь градѣ. рекше вь црⷭ҇градѣ39), and chapter 50 discusses olive- and fig-trees. Chapters 23–24, 37, 51, 58 deal with developed sewage and water systems, and chap- ter 34 mentions neighbouring porticoes (л пркоснеть се пртворѣхь на свою потрѣбѹ ѿмь40).These big-city problems were translated into Slavonic without significant adaptation. Apart from several explanatory additions, the translation follows faithfully its Greek source. This lack of adjustment is an indication that the tech-nical juridical text was perceived not as a legal manual, but as literature41. The connections to the actual reality in the original were lost in the new context of the translation in a way that is similar to the transformation of the oral sermon into a written literary genre.The more a text is used and appropriated, the more it is subjected to altera-tions. The translations of John Chrysostom’s homilies show both tendencies – in some cases they are true to the Greek source, in others they are adapted to the new audience. The examples cited above were translated by different anonymous translators in the early 10th-century Bulgaria, they represent various approaches towards the original. The genre of the homily is also an important factor in this process. Although the written homily is removed from the initial moment of delivery, it lives on as reading matter or material for new sermons. John Chrysostom’s urban images are only a small piece of the cultural and literary history. They were often documents of his time, pictures of his fellow-citizens and their surroundings, which served sometimes as rhetorical means for conveying a deeper and more general message. For the Slavonic audience, however, these episodes were equal to all the other figures of speech – parables, exempla, etc., which were one step further from their day-to-day life. Most of this literature was monastic, for individual or collective reading in monasteries, but sometimes also for highly educated and prepared readers (we still do not know enough about the reception of these texts). Therefore, the translation of the realia into Old Church Slavonic involved not only adaptation in order to make the foreign reality more relatable, but it was also a transformation of a document into literature.39 М. ЦИБРАНСКА-КОСТОВА, Градският закон…, p. 187.40 Ibidem, p. 190, in the Greek text στοά, cf. J. Zepos, P. Zepos, Prochiron…, p. 211. For the same Slavonic word пртворъ cf. examples 3 and 12 and notes 24 and 32 above.41 D. Naydenova argues that the early translations of various Byzantine legal texts into Old Church Slavonic were part of the political ideology rather than a state legislation, and they should be con-sidered literary sources, cf.: D. Naydenova, Cyrillo-Methodian Legal Heritage and Political Ideology in the Mediaeval Slavic States, PBAS.HSS 1.1, 2014, p. 3–16.Aneta Dimitrova80BibliographyAladzhov A., The Byzantine Empire and the Establishment of the Early Medieval City in Bulgaria, [in:] Byzanz – das Römerreich im Mittelalter, vol. III, Peripherie und Nachbarschaft, ed. F. Daim, J. Drauschke, Mainz 2010, p. 113–158.Ameringer T., The Stylistic Influence of the Second Sophistic on the Panegyrical Sermons of St. John Chrysostom. A Study in Greek Rhetoric, Washington 1921 [= Patristic Studies, 5].Bulanin D., Tekstologičeskie i bibliografičeskie arabeski. Priloženie V. Andrianty v staršem slavjan-skom perevode, [in:] Katalog pamjatnikov drevnerusskoj pis’mennosti 11–14 vv., Sankt-Peterburg 2014, p. 489–510.Burns M.A., Saint John Chrysostom’s Homilies on the Statues. A Study of Their Rhetorical Qualities and Form, Washington 1930 [= Patristic Studies, 22].Byzantine Constantinople. Monuments, Topography and Everyday Life, ed. N. Necipoglu, Leiden 2001 [= The Medieval Mediterranean, 33].Dossey L., Night in the Big City. Temporal Patterns in Antioch and Constantinople as Revealed by Chrysostom’s Sermons, [in:] Revisioning John Chrysostom, ed. C.L. de Wet, W. Mayer, Leiden–Boston 2019 [= Critical Approaches to Early Christianity, 1], p. 698–732, https://doi.org/10.1163/ 9789004390041_022Georgieva T., Zlatostruj ot XII vek, Silistra 2003.Harbova M., Ukrepenijat bălgarski srednovekoven grad 13–14 v., Sofija 1979.Ilarij ierom., Arsenij ierom., Opisanie slavjanskih rukopisej biblioteki Svjato-Troickoj Sergievoj lavry, Moskva 1878.Ilieva T., Antičnata kultura prez prizmata na srednovekovnija bălgarski knižoven ezik, Dzjalo 10, 2017, http://www.abcdar.com/magazine/X/T.Ilieva_1314–9067_X.pdfLavan L.A., The Agorai of Antioch and Constantinople as Seen by John Chrysostom, “Bulletin of the Institute of Classical Studies” 50, Issue Sup. 91, 2007, p. 157–167, https://doi.org/10.1111/ j.2041-5370.2007.tb02386.xLeonid arhim., Svedenie o slavjanskih rukopisjah, postupivših iz knigohranilišča Svjato-Troickoj Ser-gievoj lavry v biblioteku Troickoj duhovnoj seminarii v 1747 g. (nyne nahodjaščihsja v biblioteke Moskovskoj duhovnoj akademii), Vyp. 2, Moskva 1887.Liddell H.G., Scott R., Jones H.S. et al., A Greek-English Lexicon, 9Oxford 1996.Lišev S., Bălgarskijat srednovekoven grad. Obštestveno-ikonomičeski oblik, Sofija 1970.Magdalino P., Studies on the History and Topography of Byzantine Constantinople, Aldershot 2007.Mango C., Le développement urbain de Constantinople (IVe–VIIe siècles), Paris 1985.Mayer W., The Homilies of St John Chrysostom. Provenance. Reshaping the Foundations, Rome 2005.Mayer W., John Chrysostom: Extraordinary Preacher, Ordinary Audience, [in:] Preacher and Audi-ence. Studies in Early Christian and Byzantine Homiletics, ed. P. Allen, M. Cunningham, Leiden 1998, https://doi.org/10.1163/9789004391666_007Mayer W., The Topography of Antioch Described in the Writings of John Chrysostom, [in:] Les sourc-es de l’histoire du paysage urbain d’Antioche sur l’Oronte. Actes des journées d’études des 20 et 21 septembre 2010. Colloques de l’université Paris 8, ed. C. Saliou, Paris 2012, p. 81–100.Menander Rhetor, ed. et trans. D.A. Russell, N.G. Wilson, Oxford 1981.Mihanović Homiliar, ed. R. Aitzetmüller, Graz 1957.81Translation and Transformation of John Chrysostom’s Urban Imagery…Mijatev K., Arhitekturata v srednovekovna Bălgarija, Sofija 1965.Milanova A., Gradăt văv vizantijska Bălgarija (11–12 v.), [in:] Srednovekoven urbanizăm. Pamet – Sakralnost – Tradicii, Sofija 2007, p. 7–29.Miltenov Y., Dialozite na Psevdo-Kesarij v slavjanskata răkopisna tradicija, Sofija 2006.Miltenov Y., Zlatostruj: starobălgarski homiletičen svod, săzdaden po iniciativa na bălgarskija car Simeon. Tekstologičesko i izvorovedsko izsledvane, Sofija 2013.Mirčeva E., Germanov sbornik ot 1358/1359 g. Izsledvane i izdanie na teksta, Sofija 2006.Mušinskaja M.S., Adrianty Ioanna Zlatousta v južnoslavjanskih i russkih pamjatnikah, [in:] Lin-gvističeskoe istočnikovedenie i istorija russkogo jazyka (2002–2003), Moskva 2003, p. 27–74, http://www.ruslang.ru/istochnik_2003Naydenova D., Cyrillo-Methodian Legal Heritage and Political Ideology in the Mediaeval Slavic States, “Papers of BAS. Humanities and Social Sciences” 1.1, 2014, p. 3–16.Polivjanni D., Srednovekovnijat bălgarski grad prez 13–14 v. Očerci, Sofija 1989.Popovski J., Thomson F.J., Veder W.R., The Troickij Sbornik (Cod. Moskva, GBL, F.304 (Troice-Sergie-va Lavra) N 12). Text in Transcription, “Полата книгописная / Polata Knigopisnaja” 21–22, 1988.Preacher and Audience. Studies in Early Christian and Byzantine Homiletics, ed. P. Allen, M. Cun-ningham, Leiden 1998.Saint John Chrysostom, The Cult of the Saints. Select Homilies and Letters, praef. et trans. W. May-er, B. Neil, New York 2006.The Sermon, ed. B.M. Kienzle, Turnhout 2000.Simeonov sbornik (po Svetoslavovija prepis ot 1073 g.), vol. I, Izsledvanija i tekst, Sofija 1991; vol. II, Rečnik-indeks, Sofija 1993; vol. III, Grăcki izvori, Sofija 2015.Slavova T., Vizantijski realii v prevodačeskata praktika na starobălgarskite knižovnici, [in:] Laudator temporis acti. Studia in memoriam Ioannis A. Božilov, vol. II, Ius, imperium, potestas litterae ars et archaeologia, ed. I.A. Biliarsky, Sofia 2018, p. 242–253.Sreznevskij I., Materialy dlja slovarja drevnerusskogo jazyka po pis’mennym pamjatnikam, vol. I, Sankt-Peterburg 1893; vol. II, Sankt-Peterburg 1902; vol. III, Sankt-Peterburg 1912.Suprasălski ili Retkov sbornik, vol. I–II, ed. J. Zaimov, M. Kapaldo, Sofija 1982.Totomanova A.-M., Svedenijata za grăko-rimskija svjat v slavjanskija ekscerpt ot Hronikata na Julij Afrikan. Problemi na recepcijata, [in:] POLYISTŌR. Scripta slavica Mario Capaldo dicata, ed. K. Diddi, Moskva 2015, p. 316–327.Tsibranska-Kostova M., Gradskijat zakon i gradskoto blagoustrojstvo v južnoslavjanski kontekst, “Старобългарска литература” / “Starobălgarska literatura” 57–58, 2018, p. 163–193.Turilov A.A., Andrianty, [in:] Pravoslavnaja ènciklopedija, vol. II, Moskva 2001, p. 410, http://www.pravenc.ru/text/115376.htmlUspenskij sbornik 12–13 vv., ed. S.I. Kotkov, Moskva 1971.Voicu S., Pseudo-Giovanni Crisostomo: i confini del corpus, “Jahrbuch für Antike und Christentum” 39, 1996, p. 105–115.Voicu S., Une nomenclature pour les anonymes du corpus pseudo-chrysostomien, “Byzantion. Revue internationale des études byzantines” 51, 1981, 297–305.Zepos J., Zepos P., Prochiron, [in:] Jus Graecoromanum, vol. II, Leges imperatorum Isaurorum et Macedonum, Aalen 1962, p. 114–228.Aneta Dimitrova82Abstract. John Chrysostom was not only one of the most prolific and influential authors of late antiquity but also a renown preacher, exegete, and public figure. His homilies and sermons com-bined the classical rhetorical craft with some vivid imagery from everyday life. He used descrip-tions, comparisons, and metaphors that were both a rhetorical device and a reference to the real world familiar to his audience. From 9th century onwards, many of Chrysostom’s works were trans-lated into Old Church Slavonic and were widely used for either private or communal reading. Even if they had lost the spontaneity of the oral performance, they still preserved the references to the 4th-century City, to the streets and the homes in a distant world, transferred into the 10th-century Bulgaria and beyond. The article examines how some of these urban images were translated and sometimes adapted to the medieval Slavonic audience, how the realia and the figures of speech were rendered into the Slavonic language and culture. It is a survey on the reception of the oral sermon put into writing, and at the same time, it is a glimpse into the late antique everyday life in the East-ern Mediterranean.Keywords: John Chrysostom, literary reception, translations into Old Church Slavonic, urban life, Antioch, ConstantinopleAneta DimitrovaSofia University “St. Kliment Ohridski”Department of Cyrillo-Methodian Studies15 Tsar Osvoboditel Blvd.1504 Sofia, Bulgariaanetagd@uni-sofia.bg | RESEARCH GATE |
moleculesArticleMagnolol Enhances the Therapeutic Effects of TRAILthrough DR5 Upregulation and Downregulation ofc-FLIP and Mcl-1 Proteins in Cancer CellsSeon Min Woo 1,†, Kyoung-jin Min 1,2,† and Taeg Kyu Kwon 1,3,*1Department of Immunology, School of Medicine, Keimyung University, 1095 Dalgubeoldaero, Dalseo-Gu,Daegu 42601, Korea; woosm724@gmail.com (S.M.W.); kjmin@dgmif.re.kr (K.-j.M.)2New Drug Development Center, Deagu-Gyeongbuk Medical Innovation Foundation, 80 Chembok-ro,Dong-gu, Daegu 41061, Korea3Center for Forensic Pharmaceutical Science, Keimyung University, Daegu 42601, Korea*Correspondence: kwontk@dsmc.or.kr; Tel.: +82-53-258-7358†These authors contributed equally to this work.Received: 10 September 2020; Accepted: 6 October 2020; Published: 8 October 2020�����������������Abstract: Magnolol is a biologically active compound, isolated from the Chinese herb Magnolia, thatregulates antiproliferative, anticancer, antiangiogenic and antimetastatic activities. We found thatmagnolol sensitizes TRAIL-induced apoptotic cell death via upregulation of DR5 and downregulationof cellular FLICE-inhibitory protein (c-FLIP) and Mcl-1 in cancer cells, but not in normal cells.Mechanistically, magnolol increased ATF4-dependent DR5 expression at the transcription level,and knockdown of ATF4 markedly inhibited magnolol-induced DR5 upregulation. Silencing DR5with siRNA prevented combined treatment with magnolol and TRAIL-induced apoptosis and PARPcleavage. Magnolol induced proteasome-mediated Mcl-1 downregulation, while magnolol-inducedc-FLIP downregulation was regulated, at least in part, by lysosomal degradation. Our results revealedthat magnolol enhanced TRAIL-induced apoptosis via ATF4-dependent DR5 upregulation anddownregulation of c-FLIP and Mcl-1 proteins.Keywords: magnolol; TRAIL; DR5; c-FLIP; Mcl-11. IntroductionMagnolol, a constituent of Magnolia, has been used in traditional Chinese herbal medicinesfor treating gastrointestinal disorders and allergic diseases [1,2]. It possesses diverse biologicaleffects, such as anticancer, antiangiogenetic, antioxidative, antimetastatic and neurotrophic [3–6].Magnolol suppresses tumor invasion by inhibiting NF-κB signaling-mediated MMP-9 downregulationin breast cancer, prostate cancer and cholangiocarcinoma [7–9]. Multiple studies demonstrate theanticancer function of magnolol. Magnolol induces apoptosis through regulation of the mitochondrialpathway in osteosarcoma, renal cancer and gastric adenocarcinoma [10–12]. Magnolol also induces cellcycle arrest and apoptosis via p21- or p27-dependent G2/M phase cell cycle arrest [13–16]. Moreover,magnolol has a sensitizing effect on chemotherapeutic-agent-mediated cancer cell death. For example,magnolol sensitizes death receptor (DR)-mediated death in non-small cell lung cancer [17]. In addition,magnolol effectively increases the effect of radiation, resulting in the decrease of tumor growth inhepatocellular carcinoma [18]. Inactivation of Akt signaling by magnolol enhances the therapeuticeffect of sorafenib through downregulation of antiapoptotic proteins in hepatocellular carcinomain vitro and in vivo [19]. Therefore, magnolol could be a potential adjuvant that may sensitize thetherapeutic efficiency of anticancer drugs.Molecules 2020, 25, 4591; doi:10.3390/molecules25194591www.mdpi.com/journal/moleculesMolecules 2020, 25, 45912 of 10Tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) can induce caspase-dependentapoptosis of cancer cells by interacting with DRs (DR4 or DR5) [20]. However, because antiapoptoticproteins (Bcl-2, Bcl-xL, Mcl-1, c-FLIP and IAP family) are highly expressed in tumor cells, they revealTRAIL resistance [21,22]. Therefore, to work out the drawbacks of TRAIL resistance in clinical trials,many researchers have identified TRAIL receptor agonists or chemotherapeutic agents capable ofenhancing TRAIL sensitivity [23,24].Here, we investigated the effects and the underlying mechanisms of magnolol on increasing thesensitivity to TRAIL-mediated cancer cell death.2. Results2.1. Magnolol Enhances Sensitivity to TRAIL in Human Renal Carcinoma Caki-1 CellsPreviously, many studies demonstrated that magnolol has antitumor effects in many cancercell lines [10–13]. Therefore, we examined the effect of a sublethal concentration of magnolol onthe sensitivity to TRAIL in renal carcinoma Caki-1 cells. A low concentration of magnolol alone orTRAIL alone did not increase cell death, while co-treatment with magnolol and TRAIL increased thesub-G1 population, PARP cleavage and nuclear chromatin condensation (Figure 1A,B). Moreover,combined treatment activated caspase-3 (Figure 1C). To verify caspase-mediated cell death withthe combined treatment of magnolol plus TRAIL, we used a pan-caspase inhibitor, z-VAD. z-VADsignificantly inhibited magnolol plus TRAIL-induced PARP and caspase-3 cleavage (Figure 1D).Therefore, these results indicated that magnolol sensitizes TRAIL-induced cancer cell death in acaspase-dependent manner.Figure 1. Magnolol induced TRAIL-mediated apoptosis. (A) Caki-1 cells were treated with 5–10 µMmagnolol alone, 50 ng/mL TRAIL alone or magnolol plus TRAIL for 24 h. (B,C) Caki-1 cells weretreated with 10 µM magnolol alone, 50 ng/mL TRAIL alone or magnolol plus TRAIL for 24 h. DAPIstaining (B) and DEVDase (caspase-3) activity (C) were examined. (D) Caki-1 cells were treated with10 µM magnolol plus 50 ng/mL TRAIL in the presence or absence of 20 µM z-VAD for 24 h. The sub-G1population and protein expression were detected using flow cytometry (A,D) and Western blotting(A,D), respectively. The values in the graph (A,C,D) represent the mean ± SD of three independentexperiments. * p < 0.01 compared to the control. # p < 0.01 compared to the combined treatment withmagnolol and TRAIL. White arrows indicate nuclear chromatin condensation.Molecules 2020, 25, 45913 of 102.2. Magnolol Downregulates Mcl-1 and c-FLIP Expression and Upregulates DR5 ExpressionWe investigated the sensitizing effect of magnolol on TRAIL-induced apoptosis in various cancercells and normal cells. Combined treatment induced apoptosis in other renal carcinoma ACHN,cervical cancer Hela and lung cancer A549 cells (Figure 2A). However, cell death of normal kidneycells (TCMK-1, MC) and human skin fibroblasts (HSF) was not induced by the combined treatment ofmagnolol plus TRAIL (Figure 2B). Next, we checked the expression levels of apoptosis-related proteinsunder the effect of magnolol. Magnolol decreased Mcl-1 and c-FLIP expression and increased DR5expression. However, other proteins (Bcl-2, Bcl-xL, Bim, Bax and IAP family proteins (cIAP2 andXIAP)) were not changed by magnolol treatment (Figure 2C). We then examined whether the effectof magnolol on DR5, Mcl-1 and c-FLIP expression patterns is restricted to Caki-1 cells. We found asimilar expression pattern of these proteins in ACHN, Hela and A549 cells (Figure 2D).Figure 2.Effect of magnolol on TRAIL-induced apoptosis in other cancer and normal cells.(A,B) Indicated cancer (A) and normal cells (B) were treated with 10 µM magnolol alone, 50 ng/mLTRAIL alone or magnolol plus TRAIL for 24 h. (C,D) Caki-1 (C) and indicated cancer cells (D) weretreated with various concentrations of magnolol for 24 h. The sub-G1 population and protein expressionwere detected using flow cytometry (A,B) and Western blotting (A,C,D), respectively. Cell morphologywas examined using interference light microscopy (B). The values in graph (A–D) represent the mean± SD of three independent experiments. * p < 0.01 compared to the control.2.3. ATF4-Mediated DR5 Upregulation Contributes to Magnolol Plus TRAIL-Induced ApoptosisTo examine the involvement of DR5 in magnolol-mediated TRAIL sensitization, we performeda knockdown assay using siRNA. The knockdown of DR5 markedly attenuated combinedtreatment-induced apoptosis (Figure 3A). Magnolol treatment induced upregulation of DR5 mRNAMolecules 2020, 25, 45914 of 10levels and DR5 promoter activity (Figure 3B,C). Therefore, these data suggest that magnolol inducedDR5 upregulation at the transcription level.Figure 3. ATF4-dependent DR5 upregulation by magnolol was involved in TRAIL-induced apoptosis.(A) Caki-1 cells were transfected with control (Cont) or DR5 siRNA and were treated with 10 µMmagnolol alone, 50 ng/mL TRAIL alone or magnolol plus TRAIL for 24 h. (B) Caki-1 cells weretreated with various concentrations of magnolol for 24 h. The levels of mRNA were examined usingRT-PCR and quantitative PCR (qPCR). (C) Caki-1 cells were transiently transfected with DR5 (-605)or DR5 (SacI) promoter and incubated with 10 µM magnolol for 24 h. The cells were lysed, andthe luciferase activity was measured as described in Materials and Methods. (D) Caki-1 cells weretreated with various concentrations of magnolol for 9 h. (positive control (p.c); 2 µM brefeldin A).(E) Caki-1 cells were transfected with Cont or ATF4 siRNA and were treated with 10 µM magnololfor 9 h. The sub-G1 population and protein expression were detected using flow cytometry (A) andWestern blotting (A,D,E), respectively. The values in graph (A–E) represent the mean ± SD of threeindependent experiments. * p < 0.01 compared to the combined treatment with magnolol and TRAIL incontrol siRNA. ** p < 0.01 compared to control. # p < 0.01 compared to magnolol in control siRNA.Previous studies reported that ATF4 and CHOP, well-known marker proteins of ER stress, workas transcription factors capable of regulating DR5 expression [25,26]. Therefore, we investigatedexpression levels of ATF4 and CHOP proteins. Interestingly, magnolol induced upregulation ofATF4 expression, but not CHOP expression (Figure 3D). We further examined whether ATF4 wasinvolved in magnolol-mediated DR5 upregulation. As shown in Figure 3E, silencing ATF4 abolishedthe upregulation of DR5 by magnolol. Therefore, these results suggested that magnolol-induced DR5upregulation contributes to TRAIL sensitization through the increase of ATF4 expression.2.4. Downregulation of c-FLIP and Mcl-1 by Magnolol Is Involved in TRAIL-Mediated ApoptosisAs shown in Figure 2C, magnolol induced downregulation of Mcl-1 and c-FLIP expression;we investigated whether these proteins are associated with magnolol-mediated TRAIL sensitization.As expected, ectopic expression of Mcl-1 and c-FLIP markedly attenuated the combinedtreatment-induced sub-G1 population and PARP cleavage (Figure 4A,B). These data indicated thatdownregulation of Mcl-1 and c-FLIP had a crucial role in augmenting TRAIL-mediated apoptosiswith magnolol.Molecules 2020, 25, 45915 of 10Figure 4. Overexpression of Mcl-1 and c-FLIP attenuated magnolol plus TRAIL-induced apoptosis.(A,B) Vector cells, Mcl-1- (A), and c-FLIP-overexpressing cells (B) were treated with 10 µM magnololalone, 50 ng/mL TRAIL alone or magnolol plus TRAIL for 24 h. The sub-G1 population and proteinexpression were detected using flow cytometry and Western blotting, respectively (A,B). The values ingraph (A,B) represent the mean ± SD of three independent samples. * p < 0.01 compared to combinedtreatment with magnolol and TRAIL in Caki-1/Vec. ** p < 0.01 compared to control.2.5. Magnolol Induces Downregulation of Mcl-1 and c-FLIP Protein Expression at the Post-Translation LevelMagnolol downregulated Mcl-1 and c-FLIP protein levels, but these mRNAs were not alteredby magnolol treatment (Figure 2C, Figure 5A). To explore how to regulate magnolol-mediated Mcl-1and c-FLIP, we used an inhibitor of protein biosynthesis, cycloheximide (CHX), for measuring proteinstability. As shown in Figure 5B, compared to CHX alone, magnolol inhibited the stability of Mcl-1and c-FLIP. Magnolol-induced downregulation of Mcl-1 and c-FLIP was reversed, respectively, byproteasome inhibitors (lactacystin) and lysosome inhibitors (bafilomycin A1 (Baf A1) and leupeptin)(Figure 5C,D). Therefore, these results suggested that magnolol-induced downregulation of Mcl-1 andc-FLIP proteins was regulated at the post-translational level.Figure 5. Magnolol induced downregulation of Mcl-1 and c-FLIP proteins at the post-translationalMolecules 2020, 25, 45916 of 10level. (A) Caki-1 cells were treated with various concentrations of magnolol for 24 h. mRNA expressionwas detected using RT-PCR and qPCR. (B) Caki-1 cells were treated with 10 µM magnolol in thepresence or absence of 20 µM CHX for the indicated time kinetics. Band intensity was quantified usingImage J. (C,D) Caki-1 cells were treated with 10 µM magnolol in the presence or absence of 2.5 µMlactacystin (C), 5 nM bafilomycin A (Baf A1) (D), and 10 µM leupeptin (D) for 24 h. Protein expressionwas detected using Western blotting (B–D). The values in graph (A–D) represent the mean ± SD ofthree independent experiments. * p < 0.01 compared to the control. # p < 0.01 compared to magnolol.n.s. = no significant.2.6. Generation of ROS Is Not Associated with Magnolol Plus TRAIL-Induced ApoptosisPrevious studies showed that magnolol generates intracellular ROS, resulting in increased apoptosis [11,13].We examined the involvement of ROS in magnolol-mediated TRAIL sensitization. However, magnolol did notincrease intracellular ROS in Caki-1 cells (Figure 6A), and combined treatment-induced apoptosis and PARPcleavage were not inhibited by ROS scavengers (NAC, trolox and GEE) (Figure 6B). These results suggested thatROS was not associated with apoptosis induced by combined treatment with magnolol plus TRAIL.Figure 6. Involvement of ROS in magnolol plus TRAIL-induced apoptosis. (A) Caki-1 cells weretreated with 10 µM magnolol for 0.5 and 2 h. Thioridazine was used as positive control (p.c). Caki-1cells were stained with H2DCF-DA dye. Fluorescence was detected using flow cytometry. (B) Caki-1cells were treated with 10 µM magnolol plus 50 ng/mL TRAIL in the presence or absence of 5 mM NAC,200 µM trolox or 2 mM GEE for 24 h. The sub-G1 population and PARP cleavage were detected usingflow cytometry and Western blotting, respectively. The values in graph (B) represent the mean ± SD ofthree independent samples.3. DiscussionIn this study, we suggested that magnolol could work as a sensitizer to TRAIL in cancer cells. We foundinvolvement of magnolol-mediated DR5, Mcl-1 and c-FLIP regulation in TRAIL sensitization. Magnolol increasedATF4-dependent DR5 expression at the transcription level and downregulation of Mcl-1 and c-FLIP expressionat the post-translation level. Furthermore, knockdown of DR5 or overexpression of Mcl-1 and c-FLIP markedlyblocked magnolol plus TRAIL-induced apoptosis.Previously, Liu et al. reported magnolol as a novel HDAC class I inhibitor that hyperacethylates histonesH3 and H4, especially the H3K27 site within the DR5 promoter, resulting in an increase of TRAIL-inducedapoptosis [17]. We found that magnolol increased DR5 mRNA expression. Transcription factor ATF4 wasinvolved in magnolol-mediated DR5 upregulation (Figure 3B,E). Lin et al. reported that magnolol induced Bcl-2downregulation and Bax activation in colon and liver cancer cells [27]. However, magnolol did not alter Bcl-2Molecules 2020, 25, 45917 of 10and Bax expression in our system (Figure 2C). This contradiction was believed to be due to the difference ofcell contexts and drug concentration. Whereas Lin et al. used a high concentration of magnolol (60 µM), weused a low concentration (10 µM). Recently, Wang et al. reported that magnolol (30 µM) exerted anticanceractivity via induction of CHOP-mediated ER stress in hepatocellular carcinoma HepG2 cells, and inhibitionof CHOP abrogates magnolol-induced apoptosis [28]. In our results, magnolol (10 µM) did not induce CHOPexpression, but another ER stress marker protein, ATF4, was induced by magnolol (Figure 3D). We found thatmagnolol increased DR5 expression at the transcription level through ATF4 upregulation, instead of CHOPupregulation (Figure 3D). Moreover, knockdown of ATF4 markedly blocked DR5 upregulation in magnolol-treatedcells (Figure 3E). Therefore, these data indicated that upregulation of DR5 using a low concentration of magnololwas influenced by the increase of ATF4 expression.In previous studies, magnolol induced downregulation of Mcl-1 and c-FLIP expression, and inhibited tumorprogression of colorectal cancer [29]. However, those studies did not examine the molecular mechanisms of theseproteins. We also showed the downregulation of Mcl-1 and c-FLIP protein expression, but not mRNA (Figure 2C,Figure 5A). Activation of proteasome is an important function in protein degradation in the ubiquitin proteasomesystem (UPS) [30]. We demonstrated that UPS was involved in Mcl-1 and c-FLIP degradation with magnololtreatment. However, magnolol-induced Mcl-1 downregulation was prevented by a proteasome inhibitor, whereaslysosome inhibitors disrupted magnolol-mediated c-FLIP downregulation (Figure 5C,D). Therefore, these dataindicated that magnolol may differently modulate protein expression levels of Mcl-1 and c-FLIP. Because E3 ligasesand deubiquitinases are associated with protein degradation in the UPS system, we need to further study theinvolvement of the molecular mechanism of magnolol-induced downregulation of Mcl-1 and c-FLIP.Collectively, we showed that magnolol sensitized cancer cells to TRAIL-induced apoptosis throughATF4-dependent DR5 upregulation, proteasome-mediated Mcl-1 downregulation and lysosome-mediated c-FLIPdownregulation. Therefore, we demonstrated that modulation of DR5, Mcl-1 and c-FLIP with magnolol played acritical role in the enhancement of TRAIL sensitization.4. Materials and Methods4.1. Cell Lines and CultureAll cancer cells (Caki-1, ACHN, A549 and Hela) and TCMK-1 cells were obtained from American Type CultureCollection (Manassas, VA, USA). Human mesangial cells (MCs) were purchased from Lonza (Basel, Switzerland).Normal human skin fibroblast (HSF) cells were provided by Korea Cell Line Bank (Seoul, Korea). Cells weregrown in an appropriate medium supplemented with 10% fetal bovine serum (FBS) (Welgene, Gyeongsan,Korea), 1% penicillin–streptomycin and 100 µg/mL gentamycin (Thermo Fisher Scientific, Waltham, MA, USA).To construct stable cell lines, Caki-1 cells were transfected using LipofectamineTM 2000 (Invitrogen, Carlsbad,CA, USA) with pcDNA3.1(+)/Mcl-1, pcDNA3.1(+)/c-FLIP or pcDNA3.1(+) vector plasmids and selected using700 µg/mL G418 (Invitrogen, Carlsbad, CA, USA). Immunoblot analysis was performed to examine proteinexpression [31].4.2. Reagents and AntibodiesSigma Chemical Co. provided magnolol, cycloheximide, bafilomycin A1, leupeptin and anti-actin (St. Louis,MO, USA). R&D Systems supplied recombinant human TRAIL and z-VAD (Minneapolis, MN, USA). Enzo LifeSciences provided lactacystin, anti-pro-caspase-3 and anti-c-FLIP (Ann Arbor, MI, USA). Santa Cruz Biotechnologyprovided anti-Mcl-1, anti-Bcl-2, anti-cIAP2 and anti-ATF4 (St. Louis, MO, USA). Cell Signaling Technology suppliedanti-PARP, anti-cleaved caspase-3, anti-Bcl-xL, anti-DR5 and anti-CHOP (Beverly, MA, USA). BD Biosciencesprovided anti-Bim and anti-XIAP (San Jose, CA, USA).4.3. FACS Analysis and DAPI StainingTo detect apoptosis, we used two methods. For FACS analysis, cells were harvested and suspended in 100 µLof phosphate-buffered saline and added to 200 µL of 95% ethanol. Then, cells were incubated in a 1.12% sodiumcitrate buffer containing RNase at 37 ◦C for 30 min, added to 50 µg/mL propidium iodide and analyzed using aBD Accuri™ C6 flow cytometer (BD Biosciences, San Jose, CA, USA). To check nuclei condensation, cellular nucleicells were stained with a 300 nM 4′, 6′-diamidino-2-phenylindole solution (Roche, Mannheim, Germany), and weviewed fluorescence images using fluorescence microscopy (Carl Zeiss, Jena, Germany).Molecules 2020, 25, 45918 of 104.4. Western BlottingCells were lysed in a lysis buffer (20 mM HEPES and 0.5% Triton X-100, pH 7.6) and separated using 10%SDS-PAGE. To analyze protein expression, proteins were transferred to nitrocellulose membranes (GE HealthcareLife Science, Pittsburgh, PO, USA) and checked using Immobilon Western Chemiluminescent HRP Substrate(EMD Millipore, Darmstadt, Germany).4.5. DEVDase (Caspase-3) ActivityTo measure DEVDase activity, cells were harvested and incubated with a reaction buffer containingacetyl-Asp-Glu-Val-Asp p-nitroanilide (Ac-DEVD-pNA) substrate, as previously mentioned [32].4.6. Knockdown of Genes Using siRNAGFP (control) and DR5 siRNA duplexes were purchased from Bioneer (Daejeon, Korea) and Invitrogen(Carlsbad, CA, USA), respectively. ATF4 siRNA duplexes were purchased from Santa Cruz Biotechnology(St. Louis, MO, USA). For the knockdown of gene by siRNA, Lipofectamine® RNAiMAX Reagent (Invitrogen,Carlsbad, CA, USA) was used in Caki-1 cells. Immunoblot analysis was performed to examine protein expression.4.7. Reverse Transcription Polymerase Chain Reaction (RT-PCR) and Quantitative PCR (qPCR)Total RNA was isolated using TriZol reagent (Life Technologies, Gaithersburg, MD, USA), and cDNA wasprepared using M-MLV reverse transcriptase (Gibco-BRL, Gaithersburg, MD, USA). For PCR, we used Blend TaqDNA polymerase (Toyobo, Osaka, Japan) with primers targeting DR5, c-FLIP, Mcl-1 and actin. For qPCR, SYBRFast qPCR Mix (Takara Bio Inc., Shiga, Japan) was used, and reactions were performed on a Thermal Cycler Dice®Real Time System III (Takara Bio Inc., Shiga, Japan). We calculated the threshold cycle number (Ct) of each geneusing actin as the reference gene, and we reported the delta-delta Ct values of the genes. The used primers werereferred to in previous studies [33].4.8. Promoter Activity AssayThe method using this assay was described in our previous study [34]. Briefly, cells were transfected withDR5 (-605) or DR5 (SacI) promoter-constructs using Lipofectamine™2000 (Invitrogen, Carlsbad, CA, USA) andharvested in a lysis buffer (25 mM Tris-phosphate pH 7.8, 2 mM EDTA, 1% Triton X-100, and 10% glycerol).The supernatants were used to measure luciferase activity according to the manufacturer’s instructions (Promega,Madison, WI, USA).4.9. Measurement of Reactive Oxygen SpeciesIntracellular accumulation of ROS was determined using the fluorescent probes 2, 7-dichlorodihydrofluorescein diacetate (H2DCF-DA). The cells were treated with magnolol and stained with the H2DCF-DAfluorescent dye for an additional 10 min. Then, the cells were trypsinized and resuspended in PBS, and fluorescencewas measured at specific time intervals using a FACS Canto II (BD Biosciences, San Diego, CA, USA).4.10. Statistical AnalysisThe data were analyzed using one-way ANOVA and post-hoc comparisons (Student–Newman–Keuls)in SPSS software (SPSS Inc., Chicago, IL, USA). The values represented the mean ± SD of at least threeindependent experiments.5. ConclusionsMagnolol sensitized cancer cells to TRAIL-induced apoptosis through ATF4-dependent DR5 upregulationand proteasome-mediated Mcl-1 and c-FLIP downregulation. Therefore, we demonstrated that modulation ofDR5, Mcl-1 and c-FLIP by magnolol played a critical role in the enhancement of TRAIL sensitization.Author Contributions: Conceptualization, T.K.K. and K.-j.M.; investigation, S.M.W. and K.-j.M.; data curation,S.M.W., K.-j.M. and T.K.K.; writing—original draft preparation, T.K.K. and S.M.W.; supervision, T.K.K.; fundingacquisition, T.K.K. All authors have read and agreed to the published version of the manuscript.Molecules 2020, 25, 45919 of 10Funding:ThisworkwassupportedbyNRFgrantfundedbytheKoreaGovernment(MSIP)(2014R1A5A2010008, NRF-2018R1D1A3B07049596 and NRF-2019R1A2C2005921) and 2018 Scholar Research Grantfrom Keimyung University.Conflicts of Interest: The authors declare no conflict of interests.References1.Lee, Y.J.; Lee, Y.M.; Lee, C.K.; Jung, J.K.; Han, S.B.; Hong, J.T. Therapeutic applications of compounds in theMagnolia family. Pharmacol. Ther. 2011, 130, 157–176. [CrossRef] [PubMed]2.Ranaware, A.M.; Banik, K.; Deshpande, V.; Padmavathi, G.; Roy, N.K.; Sethi, G.; Fan, L.; Kumar, A.P.;Kunnumakkara, A.B. Magnolol: A Neolignan from the Magnolia Family for the Prevention and Treatment ofCancer. Int. J. Mol. Sci. 2018, 19, 2362. [CrossRef] [PubMed]3.Kumar, S.; Guru, S.K.; Pathania, A.S.; Kumar, A.; Bhushan, S.; Malik, F. Autophagy triggered by magnololderivative negatively regulates angiogenesis. Cell Death Dis. 2013, 4, e889. [CrossRef] [PubMed]4.Amorati, R.; Zotova, J.; Baschieri, A.; Valgimigli, L. Antioxidant Activity of Magnolol and Honokiol:Kinetic and Mechanistic Investigations of Their Reaction with Peroxyl Radicals. J. Org. Chem. 2015, 80,10651–10659. [CrossRef]5.Ikeda, K.; Sakai, Y.; Nagase, H. Inhibitory effect of magnolol on tumour metastasis in mice. Phytother. Res.2003, 17, 933–937. [CrossRef]6.Fukuyama, Y.; Kubo, M.; Harada, K. The search for, and chemistry and mechanism of, neurotrophic naturalproducts. J. Nat. Med. 2020, 74, 648–671. [CrossRef]7.Liu, Y.; Cao, W.; Zhang, B.; Liu, Y.Q.; Wang, Z.Y.; Wu, Y.P.; Yu, X.J.; Zhang, X.D.; Ming, P.H.; Zhou, G.B.; et al.The natural compound magnolol inhibits invasion and exhibits potential in human breast cancer therapy.Sci. Rep. 2013, 3, 3098–3106. [CrossRef]8.Hwang, E.S.; Park, K.K. Magnolol suppresses metastasis via inhibition of invasion, migration, and matrixmetalloproteinase-2/-9 activities in PC-3 human prostate carcinoma cells. Biosci. Biotechnol. Biochem. 2010, 74,961–967. [CrossRef]9.Zhang, F.H.; Ren, H.Y.; Shen, J.X.; Zhang, X.Y.; Ye, H.M.; Shen, D.Y. Magnolol suppresses the proliferation andinvasion of cholangiocarcinoma cells via inhibiting the NF-kappaB signaling pathway. Biomed. Pharmacother.2017, 94, 474–480. [CrossRef]10.Zhou, S.; Wen, H.; Li, H. Magnolol induces apoptosis in osteosarcoma cells via G0/G1 phase arrest andp53-mediated mitochondrial pathway. J. Cell Biochem. 2019, 120, 17067–17079. [CrossRef]11.Wen, H.; Zhou, S.; Song, J. Induction of apoptosis by magnolol via the mitochondrial pathway and cell cyclearrest in renal carcinoma cells. Biochem. Biophys. Res. Commun. 2019, 508, 1271–1278. [CrossRef] [PubMed]12.Rasul, A.; Yu, B.; Khan, M.; Zhang, K.; Iqbal, F.; Ma, T.; Yang, H. Magnolol, a natural compound, inducesapoptosis of SGC-7901 human gastric adenocarcinoma cells via the mitochondrial and PI3K/Akt signalingpathways. Int. J. Oncol. 2012, 40, 1153–1161. [CrossRef] [PubMed]13.Zhou, Y.; Bi, Y.; Yang, C.; Yang, J.; Jiang, Y.; Meng, F.; Yu, B.; Khan, M.; Ma, T.; Yang, H. Magnolol inducesapoptosis in MCF-7 human breast cancer cells through G2/M phase arrest and caspase-independent pathway.Pharmazie 2013, 68, 755–762. [PubMed]14.Chen, L.C.; Liu, Y.C.; Liang, Y.C.; Ho, Y.S.; Lee, W.S. Magnolol inhibits human glioblastoma cell proliferationthrough upregulation of p21/Cip1. J. Agric. Food Chem. 2009, 57, 7331–7337. [CrossRef] [PubMed]15.Lee, S.J.; Cho, Y.H.; Park, K.; Kim, E.J.; Jung, K.H.; Park, S.S.; Kim, W.J.; Moon, S.K. Magnolol elicitsactivation of the extracellular signal-regulated kinase pathway by inducing p27KIP1-mediated G2/M-phasecell cycle arrest in human urinary bladder cancer 5637 cells. Biochem. Pharmacol. 2008, 75, 2289–2300.[CrossRef] [PubMed]16.Hsu, Y.F.; Lee, T.S.; Lin, S.Y.; Hsu, S.P.; Juan, S.H.; Hsu, Y.H.; Zhong, W.B.; Lee, W.S. Involvement ofRas/Raf-1/ERK actions in the magnolol-induced upregulation of p21 and cell-cycle arrest in colon cancercells. Mol. Carcinog. 2007, 46, 275–283. [CrossRef]17.Liu, Y.; Tong, Y.; Yang, X.; Li, F.; Zheng, L.; Liu, W.; Wu, J.; Ou, R.; Zhang, G.; Hu, M.; et al. Novel histonedeacetylase inhibitors derived from Magnolia officinalis significantly enhance TRAIL-induced apoptosis innon-small cell lung cancer. Pharmacol. Res. 2016, 111, 113–125. [CrossRef]Molecules 2020, 25, 459110 of 1018.Chen, Y.S.; Sun, R.; Chen, W.L.; Yau, Y.C.; Hsu, F.T.; Chung, J.G.; Tsai, C.J.; Hsieh, C.L.; Chiu, Y.M.; Chen, J.H.The In Vivo Radiosensitizing Effect of Magnolol on Tumor Growth of Hepatocellular Carcinoma. In Vivo2020, 34, 1789–1796. [CrossRef]19.Chen, J.H.; Chiang, I.T.; Hsu, F.T. Protein Kinase B Inactivation Is Associated with Magnolol-EnhancedTherapeutic Efficacy of Sorafenib in Hepatocellular Carcinoma In Vitro and In Vivo.Cancers (Basel)2019, 12, 87. [CrossRef]20.De Miguel, D.; Lemke, J.; Anel, A.; Walczak, H.; Martinez-Lostao, L. Onto better TRAILs for cancer treatment.Cell Death Differ. 2016, 23, 733–747. [CrossRef]21.Kretz, A.L.; Trauzold, A.; Hillenbrand, A.; Knippschild, U.; Henne-Bruns, D.; von Karstedt, S.; Lemke, J.TRAILblazing Strategies for Cancer Treatment. Cancers (Basel) 2019, 11, 456. [CrossRef] [PubMed]22.Chawla-Sarkar, M.; Bae, S.I.; Reu, F.J.; Jacobs, B.S.; Lindner, D.J.; Borden, E.C. Downregulation of Bcl-2,FLIP or IAPs (XIAP and survivin) by siRNAs sensitizes resistant melanoma cells to Apo2L/TRAIL-inducedapoptosis. Cell Death Differ. 2004, 11, 915–923. [CrossRef] [PubMed]23.Von Karstedt, S.; Montinaro, A.; Walczak, H. Exploring the TRAILs less travelled: TRAIL in cancer biologyand therapy. Nat. Rev. Cancer 2017, 17, 352–366. [CrossRef] [PubMed]24.Lemke, J.; von Karstedt, S.; Zinngrebe, J.; Walczak, H. Getting TRAIL back on track for cancer therapy.Cell Death Differ. 2014, 21, 1350–1364. [CrossRef] [PubMed]25.Chen, P.; Hu, T.; Liang, Y.; Li, P.; Chen, X.; Zhang, J.; Ma, Y.; Hao, Q.; Wang, J.; Zhang, P.; et al. NeddylationInhibition Activates the Extrinsic Apoptosis Pathway through ATF4-CHOP-DR5 Axis in Human EsophagealCancer Cells. Clin. Cancer. Res. 2016, 22, 4145–4157. [CrossRef] [PubMed]26.Liu, G.; Su, L.; Hao, X.; Zhong, N.; Zhong, D.; Singhal, S.; Liu, X. Salermide up-regulates death receptor 5expression through the ATF4-ATF3-CHOP axis and leads to apoptosis in human cancer cells. J. Cell. Mol.Med. 2012, 16, 1618–1628. [CrossRef]27.Lin, S.Y.; Chang, Y.T.; Liu, J.D.; Yu, C.H.; Ho, Y.S.; Lee, Y.H.; Lee, W.S. Molecular mechanisms of apoptosisinduced by magnolol in colon and liver cancer cells. Mol. Carcinog. 2001, 32, 73–83. [CrossRef]28.Wang, Y.D.; Sun, X.J.; Yang, W.J.; Li, J.; Yin, J.J. Magnolol exerts anticancer activity in hepatocellular carcinomacells through regulating endoplasmic reticulum stress-mediated apoptotic signaling. Onco Targets Ther. 2018,11, 5219–5226. [CrossRef]29.Su, C.M.; Weng, Y.S.; Kuan, L.Y.; Chen, J.H.; Hsu, F.T. Suppression of PKCdelta/NF-kappaB Signalingand Apoptosis Induction through Extrinsic/Intrinsic Pathways Are Associated Magnolol-Inhibited TumorProgression in Colorectal Cancer In Vitro and In Vivo. Int. J. Mol. Sci. 2020, 21, 3527. [CrossRef]30.Burger, A.M.; Seth, A.K. The ubiquitin-mediated protein degradation pathway in cancer: Therapeuticimplications. Eur. J. Cancer 2004, 40, 2217–2229. [CrossRef]31.Min, K.J.; Shahriyar, S.A.; Kwon, T.K. Arylquin 1, a potent Par-4 secretagogue, induces lysosomal membranepermeabilization-mediated non-apoptotic cell death in cancer cells.Toxicol.Res.2020, 36, 167–173.[CrossRef] [PubMed]32.Seo, S.U.; Woo, S.M.; Kim, M.W.; Lee, H.S.; Kim, S.H.; Kang, S.C.; Lee, E.W.; Min, K.J.; Kwon, T.K. CathepsinK inhibition-induced mitochondrial ROS enhances sensitivity of cancer cells to anti-cancer drugs throughUSP27x-mediated Bim protein stabilization. Redox Biol. 2020, 30, 101422–101434. [CrossRef] [PubMed]33.Woo, S.M.;Seo, S.U.;Kubatka, P.;Min, K.J.;Kwon, T.K. Honokiol Enhances TRAIL-MediatedApoptosis through STAMBPL1-Induced Survivin and c-FLIP Degradation.Biomolecules 2019, 9, 838.[CrossRef] [PubMed]34.Shahriyar, S.A.; Seo, S.U.; Min, K.J.; Kubatka, P.; Min, D.S.; Chang, J.S.; Kim, D.E.; Woo, S.M.;Kwon, T.K. Upregulation of DR5 and Downregulation of Survivin by IITZ-01, Lysosomotropic AutophagyInhibitor, Potentiates TRAIL-Mediated Apoptosis in Renal Cancer Cells via Ubiquitin-Proteasome Pathway.Cancers (Basel) 2020, 12, 2363. [CrossRef]Sample Availability: Samples of the compounds are available from the authors.© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open accessarticle distributed under the terms and conditions of the Creative Commons Attribution(CC BY) license (http://creativecommons.org/licenses/by/4.0/). | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/226904312Neoproterozoic diamictite in the Eastern Desert of Egypt and Northern SaudiArabia: Evidence of ~750 Ma glaciation in the Arabian-Nubian Shield?Article in International Journal of Earth Sciences · April 2010DOI: 10.1007/s00531-009-0427-3CITATIONS85READS4805 authors, including:Some of the authors of this publication are also working on these related projects:Roll–back, Extension and Mantle Upwelling Triggered Eocene Shoshonitic Magmatism in NW Iran View projectGOLD-BEARING VOLCANOGENIC MASSIVE SULFIDES AND OROGENIC-GOLD DEPOSITS IN THE NUBIAN SHIELD View projectKamal A. AliUnited Arab Emirates University60 PUBLICATIONS 2,816 CITATIONS SEE PROFILEBob SternUniversity of Texas at Dallas565 PUBLICATIONS 35,043 CITATIONS SEE PROFILEWilliam I MantonUniversity of Texas at Dallas93 PUBLICATIONS 4,684 CITATIONS SEE PROFILEPeter Robert. Johnson62 PUBLICATIONS 4,752 CITATIONS SEE PROFILEAll content following this page was uploaded by Bob Stern on 03 June 2014.The user has requested enhancement of the downloaded file.ORIGINAL PAPERNeoproterozoic diamictite in the Eastern Desert of Egyptand Northern Saudi Arabia: evidence of ~750 Ma glaciationin the Arabian–Nubian Shield?Kamal A. Ali Æ Robert J. Stern Æ William I. Manton ÆPeter R. Johnson Æ Sumit K. MukherjeeReceived: 27 February 2008 / Accepted: 5 February 2009 / Published online: 26 February 2009� Springer-Verlag 2009AbstractThe Neoproterozoic Atud diamictite in WadiKareim and Wadi Mobarak in the Eastern Desert of Egyptand the Nuwaybah formation in NW Saudi Arabia consistof poorly sorted, polymictic breccia, with clasts up to 1 mof granitoid, quartz porphyry, quartzite, basalt, greywacke,marble, arkose, and microconglomerate in fine-grainedmatrix. Stratigraphic relations indicate that the diamictitewas deposited in a marine environment. Integrated fieldinvestigation, petrographic study and U–Pb SHRIMP zir-con ages demonstrate that the Atud and Nuwaybah arecorrelative. The distribution of zircon ages indicate that*750 Ma ages are dominant with a significant componentof older materials, characterized by minor Mesoproterozoicand more abundant Paleoproterozoic and Neoarchean ages.Some matrix and metasedimentary clast zircons yield agesthat are a few 10s of Ma younger than the age of theyoungest clast (754 ± 15 Ma), suggesting Atud/Nuwaybahdiamictite deposition *750 Ma or slightly later, broadlyconsistentwithbeingdepositedduringtheSturtianglaciation (740–660 Ma). The Paleoproterozoic and Neo-archean clasts have no source within the ensimaticArabian–Nubian Shield. The distribution of the pre-Neo-proterozoic ages are similar to the distribution of the pre-Neoproterozoic ages in Yemen and Saharan Metacraton,suggesting that these clasts have been transported hundredsof kilometers, maybe by ice-rafting. The Atud diamictitemayrepresentimportantevidenceforCryogenian‘‘Snowball Earth’’ in the Arabian–Nubian Shield.KeywordsArabian–Nubian Shield �Snowball Earth hypothesis � Neoproterozoic �U–Pb SHRIMP zirconIntroductionThe Arabian–Nubian Shield (ANS) exposes Neoprotero-zoic (1,000–542 Ma) crust and extends from Egypt, Israel,and Jordan to Ethiopia and Yemen. The ANS (Fig. 1a)developed during fragmentation of Rodinia *870 Ma andconcluded *620 Ma when convergence between East andWest Gondwana closed the Mozambique Ocean along theEast Africa–Antarctic Orogen (EAAO, Stern 1994; Jacobsand Thomas 2004). The northern part of the EAAO recordsthe tectonic and magmatic evolution of the upper crustprior to, during, and after terminal collision (Stern 1994;Hargrove et al. 2006a, b).The ANS is characterized by four main rock sequences:juvenile arc supracrustal sequences, ophiolites, gneissiccore complexes, and granitoid intrusions (Abdel Naby et al.2002; Shalaby et al. 2005). Such lithologies are well-exposed in the Eastern Desert of Egypt and in correlativeexposures in NW Saudi Arabia. Because sedimentary rocksin ANS supracrustal sequences were deposited in a largelymarine environment during important episodes of Neo-proterozoic climate change (Fairchild and Kennedy 2007),these rocks should contain some evidence of these events.We are only beginning to examine ANS sedimentary suc-cessions for information about Neoproterozoic climate andElectronic supplementary materialThe online version of thisarticle (doi:10.1007/s00531-009-0427-3) contains supplementarymaterial, which is available to authorized users.K. A. Ali (&) � R. J. Stern � W. I. Manton � S. K. MukherjeeDepartment of Geosciences, University of Texas at Dallas,Richardson, TX 75080, USAe-mail: kaa042000@utdallas.eduP. R. JohnsonGeological Consulting, Johnson and Vranas Associates Ltd,1242 Tenth Street, NW, Washington, DC 20001-4214, USA123Int J Earth Sci (Geol Rundsch) (2010) 99:705–726DOI 10.1007/s00531-009-0427-3related events (Stern et al. 2006). This report presents newdata on a sequence of diamictite known as the ‘‘Atuddiamictite’’ which may have formed during one of theNeoproterozoic glacial episodes.One of the most interesting metasedimentary units in thenorthern ANS are the banded iron formations (BIF).Neoproterozoic BIF is one of the five ‘‘pillars’’ of theSnowball Earth hypothesis, articulated by Hoffman (2007)thus: ‘‘That the only regionally extensive sedimentaryFe2O3 and MnO2 ores in the past 1.9 Ga are intimatelyassociatedwithCryogenianglacialmarinedeposits,implying exceptional perturbations in seawater chemistry,as expected if the oceans were ice covered for long peri-ods’’. Neoproterozoic BIF is exposed in several areas in theCentral Eastern Desert of Egypt (Sims and James 1984)and around Sawawin, Saudi Arabia (Goldring 1990;Fig. 1b). These are called ‘‘ANS-BIF’’ by Stern et al.(2006), who also noted that ANS-BIF and the Atud dia-mictite are components of a broadly Sturtian (*700 Ma)metasedimentary succession associated with Neoprotero-zoic ophiolites, and so were probably deposited in arelatively deep marine basin (Fig. 1b). It is thus importantto better understand the depositional environment of thediamictite (i.e. glacial or non-glacial).The Atud diamictite is also important because under-standing the composition and age of clasts helps illuminatethe nature of pre-Neoproterozoic crust in the region aroundand within the northern ANS. It is known that some relictsof pre-Neoproterozoic crust exist within the ANS. IntactArchean crust is known from exposures in Yemen(Whitehouse et al. 2001b) and Paleoproterozoic crust ispreserved in the *1.8 Ga Khida subterrane of the SEArabian Shield (Fig. 1a) (Stacey and Agar 1985; White-houseetal.2001a).EveninseeminglyjuvenileNeoproterozoic igneous rocks, ion–probe U–Pb dating ofindividual zircons reveals abundant xenocrystic zirconswith ages of especially *1.9 and *2.5 Ga age (Hargroveet al. 2006a, b; Kennedy et al. 2004, 2005, 2007). Suchzircons are common in metavolcanic rocks with Nd iso-topic characteristics indicating derivation from melting ofdepleted (asthenospheric) mantle and geochemical char-acteristics suggesting formation in an intra-oceanic arc(Hargrove et al. 2006b). One possibility is that such ancientzircons are contributed by sediments to juvenile Neopro-terozoic melts, so it is important to understand what ages ofzircons are found in Neoproterozoic sediments.Here we report the results of our studies of two occur-rences of the Atud diamictite in the Central Eastern Desertof Egypt, at Wadi Kareim and Wadi Mobarak (Fig. 1b).We also document for the first time a correlative diamictiteacross the Red Sea in the Nuwaybah Formation of theMidyan terrane, NW Saudi Arabia (Nuwaybah diamictite).We report on the nature of the diamictite at these threelocalities including the results of petrographic and U–Pbzircon geochronologic investigations and use these resultsto interpret the origin of the Atud/Nuwaybah diamictite.We used the sensitive high mass-resolution ion microprobewith reverse geometry (SHRIMP-RG) specifically todetermine U–Pb ages of individual zircons separated fromFig. 1 a Map of the Arabian–Nubian Shield (Modified from Sternet al. 2006), showing the location of the study areas and regions wherepre-Neoproterozoic crust is found. Ages for pre-Neoproterozoiccrustal tracts are from Whitehouse et al. (1998); Sultan et al. (1994);Agar et al. (1992); Kro¨ner and Sassi (1996); Stern et al. (1994) andWalraven and Rumvegeri (1993). b Location of the Neoproterozoicdiamictite–BIF, and ophiolitic rocks in the northern Arabian–NubianShield, with the Red Sea closed (Modified after Sultan et al. 1993;Stern et al. 2006; Azer and Stern 2007)706Int J Earth Sci (Geol Rundsch) (2010) 99:705–726123diamictite clasts and matrix. These ages are coupled withfield observations to constrain the timing of diamictite unitsand to investigate their depositional environment.Previous workThere are two conglomeratic units in the Neoproterozoicbasement of Egypt, known as Atud Formation (El-Essawy1964) and Hammamat Group (Fig. 2; Akaad and Noweir1969). In spite of the fact that these units have verydifferent occurrences, they are sometimes confused. TheAtud Formation consists of massive poorly sorted androunded clasts, from sand to boulder, in a sheared greymatrix (El-Essawy 1964); it was deposited *700 Ma(Stern et al. 2006). Atud clasts include quartzite, highlyaltered granitoid, and a distinctive breccia; in contrast tothe Hammamat Group, no clasts of *600 Ma Dokhanvolcanics or pink granite are observed (Akaad et al. 1996).Abu El Ela (1990) studied the petrography and chem-istry of Atud clasts east of Jabal Atud and suggested thatthese are compositionally similar to Dokhan volcanics andEventLithologyAgeReferencesCollision of east and west Gondwana900 Ma614 Ma550 MaIsland arcassemblageOphiolite sequenceSerpentiniteGabbroDiabase dykesPillowed basaltsMetabasaltsMeta-andesiteAtud diamictiteMetasediment with BIFDokhan volcanicsHammamat sediments746±19 Ma Kröner et al. (1992)759±17 Ma Ali et al.(in review)>653±3 Ma Moussa et al.(2008)602±9 MaWilde and Youssef (2000)Pre-orogenicPost-orogenica653±3 Ma595±3 Ma>595±3 Ma Moussa et al.(2008)Older granite737±1.2 Ma Andresen et al.(in press)EventLithologyAgeReferencesCollision of east and west Gondwana900 Ma614 Ma550 MaIsland arcassemblageOphiolite sequenceSerpentiniteGabbroDiabase dykesPillowed basaltsMetabasaltsMeta-andesiteNuwaybah diamictiteMetasediment with BIFSilicic and intermediatelavas777±17 Ma Hargrove et al. (2006a)763±25 Ma Mukherjee et al.(submitted)>662±2 MaPre-orogenicPost-orogenicMetadaciteGraphitic shaleMukherjee et al.(submitted)?Zaam groupb662±2 MaMetadioriteFig. 2 a Lithostratigraphic andmajor tectonic events of theNeoproterozoic basementcomplex in Egypt (Modifiedfrom Stern 1981; Moghazi2003). Ages from Kro¨ner et al.(1992); Ali et al. 2008; Moussaet al. 2008; Wilde and Youssef2000). b Lithostratigraphic andmajor tectonic events in NWSuadi Arabia. Ages fromHargrove et al. (2006);Mukherjee et al. (2009)Int J Earth Sci (Geol Rundsch) (2010) 99:705–726707123calc-alkaline older granites. Abu El Ela (1990) concludedthat Atud conglomerate is related to the Hammamat Groupand not part of the older volcanic-arc assemblage becausehe could not explain how the clasts, which appear to havebeen derived from continental crust, could have beentransported across a Neoproterozoic back arc basin. Theconglomerate at Wadi Muweilih (Fig. 1b) was suggested tobe related to the Atud Formation (Akaad and Noweir 1964;Akaad and Abu El Ela 1996) although it has also beenconsidered as sheared Hammamat (Stern 1979; Ries et al.1983). Akaad et al. (1996) studied the petrography andchemistry of Muweilih clasts and concluded that, becausethese lack clasts of Dokhan volcanics or pink granites, theMuweilih conglomerate is Atud, not Hammamat Group.They also concluded that the Muweilih conglomerate wasdeposited along a continental margin as a result of severeerosion of a volcano–plutonic orogen.Stern et al. (2006) argued that the term ‘‘conglomerate’’is not appropriate for the coarse-grained deposits of theAtud Formation. Texturally, the formation better fits thedescription of diamictite by Flint et al. (1960) as poorlysorted and very coarse terrigenous sediment. Eyles andJanuszcak (2004) defined diamictite as poorly sorteddeposits of boulders, gravel, silt and clay in a very fine-grained matrix, which is an apt description of the AtudFormation (Fig. 3). Diamictite forms in many sedimentaryenvironments, for example as landslide deposits, in areas ofsubmarine fans, on the slopes of volcanoes, and as mete-orite ejecta blankets (Eyles and Januszcak 2004). In somecases, diamictite forms in glacial and glacially influencedenvironments. Convincing evidence of glacial origin wouldincludestriatedclastsproducedbyglacialabrasion(Boulton 1978), dropstones released from melting icebergs(Ovenshine 1970) and far-traveled clasts (Eyles andFig. 3 Field photographs ofAtud–Nuwaybah diamictite.a Atud diamictite arkose clast infine-grained schistose matrix atWadi Kareim, Egypt (pen forscale). b Atud diamictitemicroconglomerate clast in fine-grained schistose matrix atWadi Kareim, Egypt (hammerfor scale). c Atud diamictitegranitoid clast in fine-grainedschistose matrix at WadiMobarak, Egypt (felt-tip markerfor scale). d Atud diamictitehighly altered granitoid clastenclosed in metagreywackelaminations at Wadi Mobarak,Egypt (hammer for scale).e Nuwaybah diamictite outcropalong the Red Sea highway, NWSaudi Arabia. f Clast ofmicroconglomerate in fine-grained matrix Nuwaybahformation, NW Saudi Arabia(pen for scale)708Int J Earth Sci (Geol Rundsch) (2010) 99:705–726123Januszcak 2004). We agree on this basis with Stern et al.(2006) that the unit is most appropriately referred to as the‘‘Atud diamictite’’.Confident determination of whether or not the Atuddiamictite formed in association with glaciation requiresevidence such as clast striations, faceted clasts, and/ordropstones, but such evidence is unlikely to be preservedbecause of the strong deformation that affected the Ara-bian–Nubian Shield. However, the Oman region, in thevicinity of the Arabian–Nubian Shield (Fig. 1a), wasaffected by Neoproterozoic glacial events. The HuqfSupergroup in northern Oman (Jabal Akhdar) and MirbatGroup in southern Oman record two important Neoprote-rozoic glaciations (Rieu et al. 2007). In Jabal Akhdar, theAbu Mahara Group, includes two glacial diamictite hori-zons (Le Guerroue´ et al. 2005), the Ghubrah Formation(712 ± 1.6 Ma, Leather 2001) and the overlying Fiq For-mation (\645 Ma, Bowring et al. 2007). The marinesiliciclastic Nafun Group overlying the Fiq Formation isconstrained by U–Pb zircon ages from ashes and ignimbriteflows (c. 635–542 Ma, Allen and Leather 2006), so itappears that the Ghubrah Formation represents a Sturtianglaciation and the Fiq Formation represents a Marinoanglaciation (Le Guerroue´ et al. 2005; Allen and Leather2006). The Ghubrah Formation diamictite is characterizedby poor stratification; polymictic poorly sorted clasts andunsorted silty-shale matrix with some striated clasts (LeGuerroue´ et al. 2005). These general features are similar tothe description of Atud diamictites.Atud cobbles were previously concluded to have nosource within the Eastern Desert of Egypt (Dixon 1981).Geochronologic data support the inference that Atud dia-mictite clasts sample much older rocks than are exposed inthe Eastern Desert of Egypt and so must have been trans-ported some distance. Two granitic cobbles from the NW ofMarsa Alam (also referred to as the Wadi Mobarak me-tasedimentary unit) yielded highly discordant conventionalU–Pb zircon upper intercept ages of 1,120 and 2,060 Ma(Dixon 1981). Dixon (1979) also obtained a discordant U–Pb zircon upper intercept of 2.3 Ga for a granitic cobblefrom Atud conglomerate outcrops in Wadi Kareim. Fromthese results, Dixon (1981) inferred that the cobbles weretransported from a region of pre-Neoproterozoic crust westof the present Nile River, possibly by ice-rafting.Geological settingThere are four Atud diamictite occurrences known withinthe Central Eastern Desert of Egypt: Wadi Kareim, WadiMobarak, Wadi Muweilih, and the type locality east ofJabal Atud; we studied the first two occurrences. In addi-tion, we document the first occurrence of correlativediamictite in once-adjacent parts of Saudi Arabia. Asshown in Fig 1b, Atud diamictite, BIF, and ophiolite arespatially related, especially in Egypt, but BIF and dia-mictite only show a clear stratigraphic relationship at WadiKareim.The Atud diamictite where we have studied it in Egyptand Saudi Arabia is poorly sorted, matrix-supported,polymictic and contains subrounded clasts up to a meter insize. It is distinctive in the field because its clasts arestrikingly different from the ensimatic assemblages thatcharacterize the Eastern Desert, and include grey quartzite,arkose, quartzite, felsic metavolcanics, granitoid, quartzporphyry, microconglomerate, basalt, and minor dark greymarble. The three areas of detailed study are discussedfurther below.Atud diamictite (Wadi Kareim, Egypt)The Wadi Kareim study area (Fig. 4a) is the only localitywhere the Atud diamictite was found in clear stratigraphicrelationship with banded iron formation (Fig. 2a). Thediamictite is part of a supracrustal succession with meta-volcanics overlain by immature metasediments and BIF;the entire sequence is thrust over younger Hammamatconglomerate to the south. Younger metavolcanic rocks(classification of Stern 1981) at the base of the section are*100 m thick and are truncated by the thrust; thesemetavolcanics are the subject of another study (Ali et al.2008).Above the metavolcanic rocks, the Atud diamictite issucceeded by immature clastic sediments (wackestone andsiltstone), which are succeeded upwards by the BIF.Atud diamictite (Wadi Mobarak, Egypt)The Wadi Mobarak study area (Fig. 4b) is characterized byeast–west structures and lithologic belts distinct from theprincipal northwest–southeast trend of the Central EasternDesert, which largely reflects the Najd Fault System(Shalaby et al. 2005). Basement exposures around WadiMobarak are dominated by highly deformed ophioliticfragments of serpentinites, metagabbro, and greenschist-facies mafic metavolcanics while the metasedimentarysequence includes tuff, shale, schist and diamictite (Akaadet al. 1995), intruded by Neoproterozoic gabbro andgranite. The diamictite appears to be repeated by thrusting.No BIF is found within the study area, which lies in theeastern part of Wadi Mobarak, but BIF is found at Um Narfarther west (Shalaby et al. 2005; Akaad et al. 1995).Nuwaybah Formation (Saudi Arabia)The Nuwaybah locality (Fig. 4c) is located within the AlWajh quadrangle which lies between 26�000 and 26�300NInt J Earth Sci (Geol Rundsch) (2010) 99:705–726709123and between 36�000 and 37�000E (Davies 1985). The dia-mictite is part of the Zaam Group (Fig. 2b), which consistsof five units (shale, andesite, diamictite, sandstone andvolcanic rocks). No BIF is found within the study area, butBIF is found at Wadi Sawawin farther north (Fig. 1b;Goldring 1990).Research methodsRock samples were collected during three field seasons inEgypt and Saudi Arabia in 2004, 2005, and 2006; samplelocations are shown in Fig. 4. Thin section petrographicstudies provided a basis for selecting samples for geo-chronologic studies (Table A1 in supplementary material).On this basis, we selected 17 samples (14 clasts and 3matrix samples) for U–Pb zircon age determinations. Thisincluded nine samples from Wadi Kareim, Egypt, foursamples from Wadi Mobarak, Egypt and four samples fromArabia.Samples were processed at UTD to extract zircon usingstandard crushing and mineral separation techniques,including sieving size fractions, Wilfley table separation ofheavy minerals, removal of magnetic fractions using aFrantz magnetic separator, and purification of zircons usingheavy liquids.Analytical work was performed during two sessions in2005 and 2006 using the sensitive high mass-resolution ionmicroprobe with reverse geometry (SHRIMP-RG), co-managed by US Geological Survey and Stanford Univer-sityDepartmentofGeologicalandEnvironmentalSciences. A total of 229 spots on 229 zircons from 17samplesweredetermined.Analyticalproceduresaredescribed in detail by Hargrove et al. (2006a, b). Analyticaldata acquired by SHRIMP-RG for all samples and apparentconcordia ages are presented in Table A2 (supplementaryLegendThrust faultCenozoic depositsDioriteGabbroGraniteShale, sandstone, marble & schistMetasediment with BIFDiamictiteMetavolcanicSerpentinite & talc-carbonateHammamat sedimentWadi Um Hyout02 km5043Wadi KareimK34˚ 02’E34˚ 04’E25˚ 57’25˚ 55’a34˚ 00’EMobarakWadi Mobarak00.5 1kmM34˚ 32’ 44”E25˚ 26’ 17”85757572Wadi Kadabora El HamraWadib36˚ 18’E26˚ 30’36˚ 24’E26˚ 24’Road To DubaWadi ar RamahNWRas Marabit05 kmRED SEAcFig. 4 Detailed geologic maps of Atud diamictite study areas. aWadi Kareim, Central Eastern Desert of Egypt, showing the locationof Atud diamictite and samples (K). b Wadi Mobarak, Central EasternDesert of Egypt, showing the location of the diamictite samples (M). cNuwaybah diamictite, Zaam group, NW Saudi Arabia, showing thelocation of diamictite samples and outcrop shown in Fig. 3c (NW)710Int J Earth Sci (Geol Rundsch) (2010) 99:705–726123material). All samples are plotted as two-sigma errorellipses on Tera–Wasserburg concordia diagrams or onWetherill concordia diagrams (Wetherill 1956; Tera andWasserburg 1972; Figs. 6, 7, 8, 9). Representative ca-thodoluminescence images are shown in Fig. 10.We report ages for clasts based on weighted mean agesfor multiple zircon analyses from individual samples as wellas ages for individual grains when no single age for thesample was obtained. Concordia ages are determined usingISOPLOT (Ludwig 2000). ISOPLOT determines twomodels based upon the calculated mean square of weighteddeviates (MSWD) parameter. Model 1 uses assigned ana-lytical errors and error correlations when these are thereasons for scatter; MSWD will be close to unity. Model 2ages are reported when the MSWD is greater than unity,suggesting that there may be non-analytical error (White-house et al. 1998). Ages for discordant zircons are reportedas 238U/206Pb ages for ages\1.0 Ga and as 206Pb/207Pb agesfor ages [1.0 Ga. To assess the most suitable filter foraccepting or excluding ages, we plotted U–Pb zircon agesagainst U (ppm) and Th/U for all analysed zircons (as shownin Figure S1 in supplementary material). These plots showthat there is no clear filter we can use, however the best filterwe found is to exclude analyses which are discordant and/orshow low Th/U (Th/U B 0.3), high U ([500 ppm) and highcommon Pb. We treated each of the 17 samples separatelybecause some show a clustering around certain age whileother clast zircon populations show great scatter.Diamictite lithofacies association and petrographyDiamictite lithofacies associationThe Atud diamictite lithofacies may represent terrestrialmoraine, tidewater, or marine glacial deposits. Detailedfacies analysis is complicated by pervasive deformation.Nevertheless, the Atud Formation facies occurs within asuccession of clastic sediments dominated by metacon-glomerateandmetagreywacke,withsubordinatemetamudstone. The lack of evidence for a basal unconfor-mity and association with greywacke and BIF are mostconsistent with a marine or lacustrine depositional envi-ronment.Sedimentarylithofaciesinmoraine-moundcomplexes may have varied materials because the sheets ofglaciers may flow over many rock units (Hambrey andGlasser 2002). Tidewater glaciers tend to rework glacio-marine facies, ranging from coarse gravel (ice-contactfacies), to homogenous mud with dispersed clasts up toboulder size of ice-rafted origin (Hambrey and Glasser2002). Confident interpretation of lithofacies is beyond thescope of this study, because deformation has obscured mostprimary sedimentary features. Nevertheless, interpretationof Atud diamictite lithofacies association is discussedbelow, adapted in part from (El-Essawy 1964; Allen et al.2004). The Atud diamictite association at the type localityJabal Atud occurs as an elongated outcrop about 6.7 km longand 2 and 3 km wide, whereas the Atud diamictite at WadiMobarak outcrops with a length of about 5 km and width ofabout 1 km. In Wadi Kareim the Atud diamictite associationoccurs as a few hills about 50 m wide and about 20 m high.Conglomerate lithofaciesThe metaconglomerate is clast-supported, beds are usually*2–3 m thick, greenish grey in color and poorly sorted.The fine-grained conglomerate is dark grey, poorly sorted,and composed of rock fragments, quartz grains and feld-spar embedded in microcrystalline groundmass. The rockfragments constitute [50% of the rock. Some conglomer-ate is composed of rock fragments less than 3 cm. The rockfragments are rounded to subrounded granules and pebblesand included volcanic, reworked metagreywackes, marbleand cataclastic gneisses. Coarse-grained conglomeratecontains clasts up to 2 m in diameter. Clasts are generallyangular, ellipsoidal and flattened. Clasts include marble,granite, quartz porphyry, arkose and quartzite, and aredescribed in detail below and in Table A1 (supplementarymaterial).Metagreywacke lithofaciesMetagreywacke occurs as massive, greenish grey layers30 cm–2 m thick, with persistent mm-scale laminations.Clasts are angular and rounded ranging from 10 to 75 cm,composed of different lithologies, and are enclosed bymetagreywacke laminations. Clasts include marble, gran-ite, quartz porphyry, greywacke, arkose and quartzite. Themetagreywacke laminations are defined by variations in theproportion of rock fragments and quartz in a fine-grainedmatrix of quartz and calcite.Metamudstone lithofaciesMetamudstones occur as a subordinate facies up to 25 cmthick, with persistent mm-scale pencil-like splinters (El-Essawy 1964). The metamudstones are interbedded withthe metaconglomerates and metagreywackes. In the typelocality (Jabal Atud), the metamudstones do not occurthroughout the entire facies associations but are abundantin the middle part of the unit (El-Essawy 1964).Size and petrography of diamictite clastsAtud diamictite clasts are weakly embedded in a scalygreywacke matrix. Figure 5 presents scans of thin sectionsInt J Earth Sci (Geol Rundsch) (2010) 99:705–726711123and photomicrographs for representative clast lithologiesdescribed in this section.(a) Granitoid Clasts are rounded to subrounded and arebrownish, grey or reddish. Granitoid clasts are composed ofsubhedral to anhedral, highly strained quartz (Fig. 5a). Thefeldspars are anhedral and highly altered, so that plagio-clase and K- feldspar cannot be distinguished. Biotite andhornblende are replaced by chlorite, and calcite occurs aspatches within feldspar.(b) Quartz porphyry These are fine-grained, grey, andcomposed of corroded euhedral quartz grains, with chertand volcanic fragments embedded in a fine-grained matrix(Fig. 5b). The volcanic fragments are composed of quartzand altered plagioclase; albite twining could be recognized.(c) Basalt These are dark grey, fine-grained clastscomposedofalteredplagioclase,clinopyroxene,andgroundmass of the same minerals (Fig. 5c). The plagio-clase occurs as subhedral phenocrysts, recognized by albiteFig. 5 Photomicrographs andscans of diamictite clasts.a Granitoid K2B scan. b Quartzporphyry M9 scan. c BasaltM11 photomicrograph.d Microconglomerate K1Ascan. e Arkose NCBphotomicrograph. f GreywackeK1D scan. g Quartzite K2Dphotomicrograph. h MarbleM15 scan. Qz quartz, CHchlorite, HB hornblende, Fspfeldspar, R rhyolite, B basalt,C chert, Ca calcite, Mmudstone, CPX clinopyroxene,A alkali feldspar, P plagioclase712Int J Earth Sci (Geol Rundsch) (2010) 99:705–726123twining. The clinopyroxene occurs as phenocrysts, show-ing strong interference colors.(d) Microconglomerate This distinctive clast type con-sists mostly of gravel-sized rock fragments, with *75%pebbles of rhyolite, basalt, trachyte, chert and mudstone inan arkosic matrix of quartz, feldspar, mica and chlorite(Fig. 5d). The cementing material is mostly calcite, withsubordinate quartz.(e) Arkose This is a fine-grained clastic sediment. Quartzoccurs as angular to subrounded grains, feldspars are sub-ordinate and chlorite is found as scattered grains (Fig. 5e).The most common accessory minerals are zircon, rutile andmagnetite. The cement is made of interlocking carbonateand quartz.(f) Greywacke This clast type is light to dark grey incolor, fine-grained and composed mainly of angular tosubrounded grains of quartz and fragments of chert andgranitoid fragments in a fine-grained matrix (Fig. 5f). Thematrix consists of mica, chlorite and unidentified detritalmaterial.(g) Quartzite These are hard, fine-grained, grey rocks, inwhich abundant quartz grains are cemented by carbonate(Fig. 5 g). Grains are 95 % or more rounded quartz,marked by small inclusions of impurities.(h) Marble This fine-grained rock is massive and grey,mainly composed of anhedral calcite with many thin calciteand quartz veinlets. Scattered quartz grains form a mosaicwith calcite (Fig. 5 h).Geochronological resultsTable 1 summarizes the geochronological results. Table 2compares the most important geochronologic results for thethree localities. Results from each of these sites are dis-cussed further in the following sections.Wadi Kareim (Egypt)The nine samples from Wadi Kareim include eight clastsfrom the diamictite (5 granitoid clasts, 1 quartzite, 1quartzporphyry, and 1 arkose) and a matrix sample. The fivegranitoid clasts yield a range of ages, including two clastswith only Neoproterozoic (*750 Ma) zircons; one clastthat yielded mostly Neoproterozoic (750–790 Ma) zirconsalong with abundant 2.1–2.4 Ga zircons; a clast that yiel-dedaPaleoproterozoicage(2006 ± 15 Ma),whichappears to have been reset in Neoproterozoic time and aclast that yielded only Paleoproterozoic–Archean ages. Thequartz porphyry yielded mostly Neoproterozoic ages (697–778 Ma) along with one early Paleoproterozoic and oneArchean zircon. The quartzite clast yielded Paleoprotero-zoic to Archean (2,093–2,732 Ma) zircons. The arkosecontained Neoproterozoic (*722–773 Ma) and older Pa-leoproterozoic zircons (1.8–2.5 Ga). The diamictite matrixcontains zircons that were derived from a mixed Neopro-terozoic and Paleoproterozoic sources. Further details on asample-by-samplebasisaregiveninthefollowingparagraphs.K2B is a subrounded, coarse-grained granitoid clast.One analysis of each of twenty-one zircons was performed,and the results are shown on a Tera–Wasserburg concordiadiagram (Fig. 6a). These zircons have moderate U contentsand low common 206Pb. The analyses are mostly concor-dant and yield only Neoproterozoic ages ranging from 688to 783 Ma. Data for all 21 zircons does not yield a goodage but removing discordant data points and those with Th/U \0.3 yields a robust Model 2 (Ludwig 2000) lowerintercept age based on 12 grains of 754 ± 15 Ma (2r,MSWD = 2). K2B is one of the few clasts that do notcontain any pre-Neoproterozoic zircons, providing the bestolder age limit for the depositional age of the diamictite.Still, five of fifteen concordant zircons give younger ages:678 ± 6, 688 ± 7, 729 ± 7, 730 ± 6, 732 ± 6, suggest-ing that the clast may be significantly younger than754 ± 15 Ma.K2C is a subrounded and coarse-grained granitoid clast.One analysis was conducted for each of eight zircons,yielding ages that range from 545 to 2,441 Ma (Fig. 6b)with no concentration that allows a single age to be cal-culated. The three discordant zircons have significant (3–8%) common 206Pb. A similar scatter of ages is observed ifonly the five concordant ages are considered, although acluster of ages *758–788 Ma based on three grainsprobably approximates the age of the granitoid. We notethat the very young zircon (545 ± 3 Ma) has high Ucontents (556 ppm) and do not think it constrains the age ofthe diamictite. We infer that this clast was eroded from agranitoid which crystallized about 750–790 Ma and thefour older ages (2.1–2.4 Ga) to represent Paleoproterozoicxenocrystic zircons picked up by the granitic magma.K2E is a coarse-grained granitoid clast. One analysiswas conducted on each of ten zircons, with ages rangingfrom 700 to 767 Ma, one of which was omitted because itwas discordant. The remaining nine points plot on con-cordia in positions that suggest minor loss of radiogenic Pb.The nine points yield a Model 1 (Ludwig 2000) lowerinterceptageof763 ± 45 Ma(2r,MSWD = 1.6;Fig. 6c), which we interpret to indicate the crystallizationage of the granitoid from which the clast was eroded. Onezircon gave a significantly younger age (700 ± 9 Ma).This clast is like K2B, both petrographically and in termsof its age, lacking evidence of pre-Neoproterozoic zircons.K2H is a granitoid clast. One analysis was conducted foreach of ten zircons, which range in age from 1,654 to1,986 Ma; all but one age (*1,986 Ma) are discordantInt J Earth Sci (Geol Rundsch) (2010) 99:705–726713123Table 1 Geochronological summary for Atud–Nuwaybah diamictite samplesLocationSample No.Latitude longitudeLithologyConcordia agesConcordantages (Ma)Wadi KareimK2B25�5605900Granitoid754 ± 15 (L.I)‘‘Egypt’’34�0200000K2C25�5605900GranitoidNo concordia545 ± 534�0200000758 ± 13769 ± 8788 ± 82,432 ± 8K2E25�5605900Granitoid763 ± 45 (L.I)34�0200000K2H25�5605900Granitoid634 ± 57 (L.I)34�02000002,006 ± 15 (U.I)K2J25�5605900GranitoidNo concordia1,970 ± 1934�02000002,018 ± 482,534 ± 242,537 ± 272,865 ± 232,899 ± 792,943 ± 27K1E25�5605800Quartz porphyry752 ± 3334�0105600K2D25�5605900QuartziteNo concordia2,093 ± 2234�02000002,107 ± 182,597 ± 272,732 ± 18K2G25�5605900ArkoseNo concordia722 ± 534�0200000748 ± 7753 ± 9773 ± 51,811 ± 202,489 ± 112,523 ± 13K1F25�5605800GreywackeNo concordia661 ± 834�0105600Matrix719 ± 9729 ± 8739 ± 9854 ± 10984 ± 12987 ± 122,057 ± 122,115 ± 122,457 ± 142,462 ± 192,466 ± 92,471 ± 122,490 ± 262,668 ± 20MoubarakM125�2504700Granitoid770 ± 12 (L.I)Egypt34�3305600714Int J Earth Sci (Geol Rundsch) (2010) 99:705–726123(Fig. 6d). Many of these zircons have high U contents([300 ppm) and high common 206Pb, consistent with theobservation of strong discordance. Nine zircons define adiscordia (MSWD = 1.6) with an upper intercept of2,006 ± 15 Ma and a poorly defined Neoproterozoic lowerintercept of 634 ± 57 Ma. We are not sure if the *2.0 Gaage reflects the intrusion/cooling age of Paleoproterozoicgranite body reset during Neoproterozoic time or whetherTable 1 continuedLocationSample No.Latitude longitudeLithologyConcordia agesConcordantages (Ma)M225�2504700Granitoid831 ± 21 (U.I)34�3305600M1625�2504700Quartzite1,224 ± 71,213 ± 1134�33056001,220 ± 91,222 ± 141,226 ± 121,227 ± 131,243 ± 122,418 ± 172,468 ± 172,505 ± 112,514 ± 15M1725�2504700GreywackeNo concordia538 ± 634�3305600Matrix634 ± 8682 ± 9728 ± 9772 ± 10782 ± 13819 ± 91,090 ± 132,529 ± 132,747 ± 14NuwaybahNCA26�2700000Granitoid765 ± 22 (U.I)NW Arabia36�2400000NCC26�2700000GranitoidNo concordia741 ± 1636�2400000742 ± 91,753 ± 132,084 ± 202,704 ± 302,747 ± 272,883 ± 14NCB26�2700000Arkose766 ± 5 (Mean)36�2400000NM26�2700000GreywackeNo concordia713 ± 1736�2400000Matrix740 ± 10752 ± 13757 ± 15768 ± 25903 ± 181,017 ± 201,706 ± 162,429 ± 612,482 ± 13Int J Earth Sci (Geol Rundsch) (2010) 99:705–726715123the Neoproterozoic lower intercept approximates or intru-sion/cooling age of a Neoproterozoic granitoid from whichthe clast was eroded that massively assimilated *2.0 Gacrust, but tentatively prefer the first interpretation.K2J is a granitoid clast. One analysis was conducted foreach of nine zircons, which range in age between 1970 and3155 Ma (Fig. 6e). All except the 1,970 Ma zircon containvery little U (7–55 ppm) and Th (0.04–3 ppm), and havevery low Th/U, which suggests these may have experi-encedsignificantmetamorphism.Sevenanalysesareconcordant, ranging in age from 1,970 to 2,900 Ma; with afour grain cluster of ages between 2.8 and 2.95 Ga. Thedata do not yield a robust crystallization or metamorphicage, but it is most likely that this clast was eroded from apre-Mesoproterozoic intrusion that could have been as oldas 3.0 Ga and been metamorphosed as many as three times;certainly there is an imprint of *1.9 Ga metamorphic origneous activity but no hint of a Neoproterozoic overprint.K1E is fine-grained quartz porphyry clast. One analysiswas conducted on each of ten zircon grains, which rangefrom 697 to 2,728 Ma; two of the ten grains yielded Pa-leoproterozoic or Archean ages. All zircons show moderateU contents, high Th/U indicating igneous origin, and onlytwo zircons have significant common 206Pb (697 ± 28,724 ± 6 Ma). A Model 2 solution (Ludwig 2000) yielded alower intercept age of 752 ± 33 Ma and upper interceptage of 1,957 ± 150 Ma (2r, MSWD = 2.9; Fig. 6f) afterexcluding the three discordant zircons. We interpret thisage to represent eruption or hypabbyssal emplacement ofNeoproterozoic felsic magma that was contaminated byPaleoproterozoic and Archean materials.K2D is a quartzite clast. One analysis was conducted foreach of seven zircons, which range in age from 2.1 to2.7 Ga. Four grains are concordant (Fig. 7a) and rangefrom 2,098 ± 15 to 2,733 ± 19 Ma; the three discordantgrains have significant common 206Pb. Because it containsno Neoproterozoic zircons, we infer that this quartzite clastwas eroded from a sedimentary quartzite bed that wasdeposited before Neoproterozoic time and that was derivedfrom detritus shed by an *2.1 Ga or older crustal tract.K2G is a fine-grained arkose clast. One analysis wasconducted for each of ten grains and yielded 8 concordantages that range from 722 to 2,523 Ma (Fig. 7b); fourconcordant zircons range from 722 to 773 Ma, one is*1.8 Ga and a three grain cluster is *2.5 Ga. We arenot sure of the significance of the 722 ± 5 Ma age; this is*30 Ma younger than the age of the youngest clast(K2B). This clast was eroded from a lithified bed ofcoarse arkosic detritus derived from a mixed provenanceof Neoproterozoic (*722–773 Ma) and older (Paleopro-terozoic and Archean; 1.8–2.5 Ga rocks. This arkosicdeposit thus was deposited, lithified, and then re-erodedand re-deposited some time after *750 Ma, perhaps after*720 Ma.K1F is a sample of the greenish greywacke diamictitematrix. One analysis was conducted for each of 23 zircongrains, producing ages that range from 661 to 2,668 Ma(Fig. 7c). Discordant grains generally have the highestcommon 206Pb. Sixteen of the analyses are concordant andrange in ages from 719 to 854 Ma (5 grains), 984 to987 Ma (2 grains), 2,059 to 2,115 Ma (2 grains), 2,458 to2,503 Ma (6 grains), and 2,668 Ma (1 grain). We are notsure of the significance of the young (661 ± 8 Ma) age;this has significant common 206Pb and moderate Th/U(0.22). The matrix sample does not yield an easily inter-preted concordia age but indicates derivation from a mixedsource of Neoproterozoic, Paleoproterozoic and Neoarch-ean sources.Wadi Mobarak (Egypt)We processed four samples from the diamictite at WadiMobarak: three clasts and the matrix. Two granitoid clastsyielded Neoproterozoic ages (770 ± 12, 831 ± 21 Ma);whereas one quartzite clast gave pre-Neoproterozoic ages(1.2–2.5 Ga). The matrix contains zircons that werederived from a mixture of Neoproterozoic, Paleoprotero-zoic and Neoarchean sources, similar to the results for thediamictite matrix at Wadi Kareim. Further details of theseanalyses are given below.Table 2 Summary ofgeochronological results fordiamictite samples from WadiKareim, Wadi Mobarak, Egyptand Nuwaybah Formation, NWSaudi Arabiaui upper intercept, li lowerinterceptClastsWadi Kareim,EgyptWadi Mobarak,EgyptNuwaybah formation,Saudi ArabiaGranitoidli = 754 Mali = 770 Maui = 765 Mali = 763 Maui = 831 Ma*741 Ma to *2.8 Gali = 634 (Ma)–ui = 2.0 Ga*1.9–3.0 Ga*7.58–2.4 GaQuartz porphyryli = 752 Ma–ui = 1.9 GaArkose*722 Ma to *2.5 Ga766 Ma (Mean age)Quartzite*2.1 to *2.7 Ga* 1.2 to *2.5 GaMatrix (greywacke)*719 Ma to *2.67 Ga* 634 Ma to *2.7 Ga*713 Ma to *2.7 Ga716Int J Earth Sci (Geol Rundsch) (2010) 99:705–726123M1 is a coarse-grained granitoid clast. One analysis wasconducted for each of 15 zircon grains and the analyses areplotted on Tera–Wasserburg concordia diagram (Fig. 8a). Afew samples have high common 206Pb, including a concor-dant age of 697 ± 32 Ma. Concordant zircons yield onlyNeoproterozoic ages that range from 697 to 805 Ma,although two discordant grains suggest an upper interceptage of *2.7 Ga. Data for four zircon grains are excludedbecause three are discordant and one has large error due tohigh 204Pb, The remaining 11 points define a Model 2(Ludwig 2000) lower intercept age of 770 ± 12 Ma (±95%confidence, MSWD = 2.4; Fig. 8a). Weighted average agesfor 11 concordant analyses yield a mean 207Pb/206Pb ageof 777 ± 7 Ma (±95% confidence, MSWD = 1.33). We0.0540.0580.0620.0660.0700.074207Pb/206Pb6607007808207.27.68.08.48.89.29.6238U/206Pbdata-point error ellipses are 2 σModel 2, (±95%-conf.) on 12 pointsLower intercept: 754±15 MaUpper intercept: 1683±1100 MaMSWD = 2.0, Prob. of fit = 0.032740aK2BU/Pb207Pb/206Pb24002000160012008000.040.060.080.100.120.140.160.1802468101214238206data-point error ellipses are 2 σbK2C8206600.040.060.080.107.27.68.08.48.89.29.6238U/206Pb207Pb/206Pbdata-point error ellipses are 2 σModel 1, (±95%-conf.) on 9 pointsLower intercept: 763±45 MaUpper intercept: 1778±1000 MaMSWD = 1.6, Probability of fit = 0.12740700780 cK2E207206U/Pb1600200022000.090.100.110.120.130.1423456238206Pb/Pbdata-point error ellipses are 2 σModel 1, (±95%-conf.) on 9 pointsLower intercept: 634±57 MaUpper intercept: 2006±15 MaMSWD = 1.6, Probability of fit = 0.14d1800K2H34002600220018000.20.40.60.80102030207Pb/235U206Pb/ 238Udata-point error ellipses are 2σe3000K2J10238U/206Pb207Pb/206Pbdata-point error ellipses are 2σModel 2, (±95%-conf.) on 7 pointsLower intercept: 752±33 MaUpper intercept: 1957±150 MaMSWD = 2.9, Probability of fit = 0.014K1E2000160012008000.040.060.080.100.120.1413579fFig. 6 U–Pb concordia diagrams for SHRIMP–RG data from WadiKareim igneous clasts. a K2B granitoid. b K2C granitoid. c K2Egranitoid. d K2H granitoid. e K2J granitoid. f K1E quartz porphyry.Error ellipses are 2r; weighted average age errors quoted at 95%confidence. Dashed ellipses indicate zircon analyses that wereexcluded from age calculations. Sample information is given inTable A1 and analytical data in Table A2 (supplementary material)Int J Earth Sci (Geol Rundsch) (2010) 99:705–726717123interpret this to represent the crystallization age for thegranitic body from which this clast was derived. This clastage also indicates that the diamictite at Wadi Mobarak wasdeposited after *770 Ma. This is slightly older than butconsistent with the maximum age of 754 ± 15 Ma inferredfor the deposition of the diamictite at Wadi Kareim, based onthe age of granitoid clast K2B.M2 is a coarse-grained granitoid clast. One analysis wasconducted for each of 14 zircon grains, which range in agefrom 678 to 1,149 Ma; all but 3 ages are discordant(Fig. 8b), probably due to radiation damage caused by highU contents (mostly [1,000 ppm) of these zircons andreflected in mostly high common 206Pb. The 1,149 Ma ageis not considered to have geologic significance because it isreversely discordant and has high common 206Pb. Threegrains are concordant but are young (640–686 Ma), pos-sibly erroneously due to very low Th/U (640 ± 8 Ma) andhigh common 206Pb (666 ± 8, 686 ± 8 Ma). A discordiais defined with a Model 1 (Ludwig 2000) lower interceptage of 136 ± 42 Ma, which we consider to be geologicallymeaningless, and an upper intercept of 822 ± 32 Ma (2r,MSWD = 1), which we interpret to approximate thecrystallization age of the granitoid from which the clastwas eroded.M16 is quartzite clast. One analysis was conducted foreach of 20 zircon grains, which range in age from 1,209 to3,515 Ma (Fig. 8c). These mostly have high Th/U, show alarge range in U contents, and mostly have low 206Pb.There are no Neoproterozoic zircons. Eleven analyses areconcordant and range in age from 1,213 to 2,514 Ma. Oneclusterofsixgrainsyieldsaconcordiaageof1,224 ± 7 Ma (±95% confidence, MSWD = 0.33). Thesezircons define a Model 2 (Ludwig 2000) lower interceptageof1,239 ± 64andupperinterceptageof2,456 ± 100 Ma (±95% confidence, MSWD = 40). Thissample was eroded from a quartzite that itself was derivedfrom a crustal tract composed of Mesoproterozoic, Paleo-proterozoic and Neoarchean rocks. This quartzite is similarto the Kareim quartzite clast in lacking Neoproterozoicages but has a different population of pre-Neoproterozoicages, especially in having abundant Mesoproterozoic zir-cons, and is distinct from quartzite K2D in this regard.M17 is the diamictite matrix. One analysis was con-ducted for each of 20 zircons, ranging in age from 539 to2,752 Ma (Fig. 8d). Thirteen analyses are concordant, andthese mostly range in age from 538 ± 6 to 819 ± 9 Ma (7points),1,090 ± 13 Ma(1point),2,504 ± 9to2585 ± 15 Ma (3 points) and 2,747 ± 14 Ma (1 point).Threeconcordantzirconsgiveyoungages:oneis538 ± 6 Ma, but has a very low Th/U, suggesting that it ismetamorphic and for this reason is excluded from furtherconsideration; the 682 ± 9 Ma age may be erroneous dueto high common 206Pb; the 634 ± 8 Ma age is also youngbut no obvious problem with this analysis can be seen. Likethe diamictite matrix at Wadi Kareim, this sample does notyield a simple concordia age but indicates derivation fromK2D280026002400220020000.340.380.420.460.500.540.58579111315Pb/U207235206Pb/238Udata-point error ellipses are 2σaK2G0.040.080.120.160.200246810U/Pb238 206207Pb/206Pbdata-point error ellipses are 2σb26002200180014001000207206Wadi Kareim diamictite matrix 22001800140010006000.040.080.120.160.2004812162024238U/206PbPb/Pbdata-point error ellipses are 2σcK1FFig. 7 U–Pb concordia diagrams for SHRIMP–RG data from WadiKareim diamictite samples. a K2D quartzite clast. b K2G arkose clast.c K1F diamictite matrix. Error ellipses are 2r; weighted average ageerrors quoted at 95% confidence. Sample information is given inTable A1 and analytical data in Table A2 (supplementary material)718Int J Earth Sci (Geol Rundsch) (2010) 99:705–726123a mixed source of mostly Neoproterozoic, Mesoprotero-zoic, and Neoarchean sources. Except for the absence ofPaleoproterozoic zircons, these results are similar to thosefor the Kareim diamictite matrix.Nuwaybah formation (Saudi Arabia)We processed four samples from the Nuwaybah diamictitein Saudi Arabia: two granitoid clasts, an arkosic clast, and amatrix sample. One of the granitoid clasts and the arkosicclastyieldedNeoproterozoicagesof765 ± 22and766 ± 5 Ma respectively. A third granitoid clast was from aNeoproterozoic (*742 Ma) pluton that contained a lotof older (Paleoproterozoic and Archean) zircons. The Nu-waybah diamictite matrix contains zircons that were derivedfrom a mixed source of Neoproterozoic, Paleoproterozoic,and Neoarchean sources, with a few *1 Ga zircons.NCA is a coarse-grained granitoid clast. One analysiswas conducted for each of 12 grains, which showed lowcommon206Pb and moderate U contents. All zirconsyieldedNeoproterozoicages,from767to816 Ma(Fig. 9a). Three analyses are excluded because of lowTh/U (B0.3); the other nine grains regressed together yieldaModel1(Ludwig2000)upperinterceptageof765 ± 22 Ma (2r, MSWD = 1.3), which we interpret toreflect the age of the pluton from which this clast wasderived. This is analytically indistinguishable from themaximum age of 754 ± 15, 763 ± 45 and 770 ± 12 Mafor similar Neoproterozoic granitoid cobbles from WadiKareim (K2B, K2E) and Wadi Mobarak (M1), respec-tively, all of which lack pre-Neoproterozoic zircons.NCC is a coarse-grained granitoid clast. One analysiswas conducted for each of 12 zircon grains, which range inage from 742 to 2,951 Ma. Seven analyses are concordant(Fig. 9b) and range widely in age: 741 ± 16, 742 ± 9,1,753 ± 25 , 2,084 ± 39, 2,704 ± 30, 2,747 ± 27, and2,883 ± 14 Ma. We interpret this age distribution toindicate that this clast was eroded from a Neoproterozoic(*741 Ma) granitic pluton that also contained a lot ofPaleoproterozoic and Archean zircons.NCB is a fine-grained arkose clast. One analysis wasconducted for each of 12 zircon grains, which ranged inage from 620 to 2,074 Ma. Ten analyses are concordantranging from 707 to 886 Ma (Fig. 9c). Seven data pointsU/PbM1120016000.040.060.080.100.120.140.160.18246810238206207Pb/206Pbdata-point error ellipses are 2 σModel 2, (±95%-conf.) on 11 pointsLower intercept: 770±12 MaUpper intercept: NoneMSWD = 2.4, Probability of fit = 0.009800aM23504505506507508509500.030.050.070.090.110.130.150.170.30.50.70.91.11.31.5207Pb/235U206Pb/238Udata-point error ellipses are 2 σModel 1, (±95%-conf.) Lower intercept: 144±27 MaUpper intercept: 831±21 MaMSWD = 1.6, Probability of fit = 0.11b02468101214238U/206PbWadi Mobarak diamictite matrixM17280024002000160012008000.040.080.120.160.20207Pb/ 206Pbdata-point error ellipses are 2 σd26002200180014000.060.080.100.120.140.160.1812345238U/206Pb207Pb/206PbModel 2 (±95%-conf.) on 11 pointsLower intercept: 1239±64 MaUpper intercept: 2456±100 MaMSWD = 40, Prob. of fit = 0.000data-point error ellipses are 2σM16cConcordia Age = 1224.5 ±7.2 Maon 6 points (95% confidence)MSWD (of concordance) = 0.33,Probability (of concordance) = 0.5712200.0790.0810.0830.0854.654.754.851260Fig. 8 U–Pb concordia diagrams for SHRIMP–RG data from WadiMobarak diamictite samples. a M1 granitoid clast. b M1 granitoidclast. c M16 quartzite clast. d M17 diamictite matrix. Error ellipsesare 2r; weighted average age errors quoted at 95% confidence.Sample information is given in Table A1 and analytical data in TableA2 (supplementary material)Int J Earth Sci (Geol Rundsch) (2010) 99:705–726719123are excluded because of discordance or high U content, buta five-grain cluster yields a mean age of 766 ± 5 Ma (95%confidence, MSWD = 0.93), which we interpret as the ageof detritus shed from a mixed provenance dominated byNeoproterozoic and subordinate older (Paleoproterozoic)rocks. This arkose is similar to arkose K2G of Wadi Ka-reim in the Eastern Desert of Egypt.NM is a sample of the fine-grained matrix diamictitematrix. One analysis was made for each of 14 zircongrains, which range in age from 713 to 2,696 Ma (Fig. 9d).Ten analyses are concordant and range in age from 713 to2,482 Ma, but there is only one good cluster of ages, 740–768 Ma based on four grains. This distribution of agesindicates derivation from a mixed source of Neoprotero-zoic and Paleoproterozoic, sources, with a smattering of*1.0 Ga ages.DiscussionBelowwebrieflydiscussseveralsignificantpointsresulting from our study: (1) observations of zirconmorphology and appearance; (2) correlation of the AtuddiamictiteacrosstheRedSea;(3)ageconstraintsfor deposition of the Atud diamictite; and (4) origin sig-nificanceofpre-NeoproterozoicclastsintheAtuddiamictite.Zircon morphology and appearanceThe morphology of Atud diamictite zircons and theirappearanceincathodoluminescenceimages(Fig. 10)reveals striking differences between Neoproterozoic zir-cons and Paleoproterozoic and Archean igneous anddetrital zircons. Neoproterozoic igneous zircons (Fig. 10a,b, c) are euhedral, elongated, and show well-developedgrowth zoning. In contrast, older igneous zircons (Fig. 10g,h, i) sometimes have xenocrystic cores mantled by over-growths or occur as unmantled subrounded or rarelyeuhedral crystals (Corfu et al. 2003). Detrital zircons(Fig. 10d, e, f) in quartzite appear very dark, perhaps due tohigh U contents. These commonly contain inclusions androunded cores with zoning. Detrital zircon grains some-times are fractured.Nuwaybah diamictite matrix NM260022001800140010000.040.080.120.160.200246810238U/206Pb207Pb/ 206Pbdata-point error ellipses are 2σd8400.0580.0600.0620.0640.0660.0680.0707.17.37.57.77.98.1238U/206Pb207Pb/206Pbdata-point error ellipses are 2 σModel 1, (±95%-conf.) on 9 pointsLower intercept: NoneUpper intercept: 765±22 MaMSWD = 1.3,Prob.of fit = 0.25820800780760aNCAU/Pb260022001800140010000.040.080.120.160.200.240.280246810238206207Pb/ 206Pbdata-point error ellipses are 2 σb3000NCC80012001600200024000.040.060.080.100.120.140.161357911238U/206Pb207Pb/ 206Pbdata-point error ellipses are 2σMean Age (±95%-conf.) on 5 points= 766±5 MaMSWD = 0.93, Probability of fit = 0.44cNCBFig. 9 U–Pb concordia diagrams for SHRIMP–RG data fromNuwaybah diamictite samples. a NCA granitoid clast. b NCCgranitoid clast. c NCB arkose clast. d NM diamictite matrix. Errorellipses are 2r; weighted average age errors quoted at 95%confidence. Sample information is given in Table A1 and analyticaldata in Table A2 (supplementary material)720Int J Earth Sci (Geol Rundsch) (2010) 99:705–726123Correlation of the Atud diamictite across the Red SeaThere has been very little correlation of Neoproterozoicbasement units across the Red Sea, beyond a general cor-relation of ophiolites and banded iron formation. This ispartly due to the lack of distinctive lithologies in theNeoproterozoic basement of Egypt and NW Saudi Arabiaand partly due to insufficient high-quality geochronologicinformation. Results of the present study provide a newopportunity to advance the trans-Red Sea correlation effort.It has been noted that the diamictite exposed in Egypt issimilar in field appearance and clast lithologies to theNuwaybah diamictite exposed in Saudi Arabia. Field andpetrographic study of the clasts within the diamictite sug-gests that the lithologies and size of the clasts at WadiKareim and Wadi Mobarak (Atud diamictite) in Egypt andNuwaybah Formation in NW Saudi Arabia are similar. Thedistribution of zircon ages for the diamictite in Wadi Ka-reim and Wadi Mobarak, Egypt (Fig. 11a, b) is also verysimilar to those of the Nuwaybah diamictite in SaudiArabia (Fig. 11c). These histograms show that *690–825 Ma material dominated at all three locations, with asignificantcomponentofoldermaterials(*2and*2.5 Ga). Table 2 further summarizes the similarity inages of the granitoids for the three locations. Figure 12d isa compilation of all zircon ages from this study andemphasizes the point that all zircon grains analysed duringthis study are dominated by *750 Ma ages with subordi-nate but still important Paleoproterozoic and Archeansources. Zircon ages for matrix at the three locations showsimilar distributions to each other and to the clasts andrange from 713 Ma to 2.7 Ga (Fig. 11e), although theFig. 10 Cathodoluminescence images of zircons from Atud diamic-tite clasts. a, b and c Typical Neoproterozoic igneous zircon grainsfrom granitoid clast M1. d, e and f Typical quartzite zircon grainsfrom clasts K2D and M16. g, h and d Typical zircon grains of igneousPaleoproterozoic and Archean from granitoid clasts K2C and K2J.Location of ion microprobe area is shown by white circles; whitescale bar is 100 lm longInt J Earth Sci (Geol Rundsch) (2010) 99:705–726721123relative proportion of older zircons seems to be greater forthe matrix than the clasts. The most obvious differencebetween the three sites is the presence of 1.1–1.2 Ga zir-cons in the Wadi Mobarak zircons, the only strongevidence that Mesoproterozoic rocks might have existed inthe diamictite source region. The diamictite was clearlyderived from a region or regions that had abundant*750 Ma igneous activity (Fig. 11f). We conclude that thediamictite units that we have studied at the three localitiesare correlative and provide an important new stratigraphicmarker within the otherwise monotonous metasedimentarysection of the region and across the Red Sea.a50015002500100020003000Frequency (n)102550201550102015300102040ConcordantDiscordantMatrixNuwaybah DiamictiteN = 50Kareim DiamictiteN = 106Mobarak DiamictiteN = 69ConcordantDiscordantMatrixConcordantDiscordantMatrixbc5001500250010002000300001020309080All Arabian - Nubian Shield DiamictiteN = 225ConcordantDiscordantMatrixdN = 57All Arabian Nubian Shield Diamictite MatrixConcordantDiscordante10128624005101520253035500700900600800Arabian-Nubian Shield Neoproterozoic agesN = 125ConcordantDiscordantMatrixfMean age = 752 ± 57 MaFig. 11 Histograms of single-grain SHRIMP-RG zircon agesanalysed during this study.a Wadi Kareim, Egypt. b WadiMobarak, Egypt. c Nuwaybah,Arabia. d Distribution of allzircon ages from this study, notesecondary peaks at * 1.9and * 2.5 Ma. Concordant anddiscordant applies to zirconsseparated from clasts in a–d.e Summary of age distributionsfor diamictite matrix zirconsfrom all three localities.f Summarizes distribution ofNeoproterozoic ages only,separated into concordant clast,discordant clast, and matrix.Discordant ages \1,000 Ma are238U/206Pb ages and those[1,000 Ma are 207Pb/206Pb722Int J Earth Sci (Geol Rundsch) (2010) 99:705–726123Age constraints for deposition of the Atud diamictiteIt is not possible to directly determine the depositional ageof diamictite, but this can be confidently constrained asbeing younger than the youngest clast. One clast within theNuwaybah diamictite was eroded from a granitoid thatcrystallized at 765 ± 22 Ma. A Nuwaybah arkose clastyields an age 766 ± 5 Ma, which must be the age of thesource granite from which the arkose was recycled. Kareimgranitoid clasts yield crystallization ages of 754 ± 15 and763 ± 45 Ma (lower intercept), whereas a quartz porphyryclast is 752 ± 33 Ma (lower intercept). A granitoid clastfrom Wadi Mobarak yields an age of 770 ± 12 Ma. Theresults for all the three localities thus indicate that thediamictite was deposited after *760 Ma ago.These age constraints are consistent with U–Pb zirconages for the underlying metavolcanics. Ali et al. (2008)report an age of 759 ± 17 Ma for the metavolcanics atWadi Kareim, which lie beneath the diamictite. In the field,the contact appears to be conformable but strong defor-mation may mask a significant hiatus, allowing thediamictite deposition to be significantly younger than themetavolcanics.Individual grains in metasedimentary clasts and dia-mictite matrix may provide additional information. Thesegrains passed the filter we suggested, some of these indi-vidual grains show younger ages than the age weconcluded from the youngest clast (752 Ma). These agesrange from 634 ± 8 to 740 ± 10 Ma which may indicatethat the diamictite deposition age is younger than the age ofthe youngest clast suite (*760 Ma).The maximum age constraints presented above arebroadly consistent with deposition during the Sturtianglacial interval, which consisted of several glacial episodesduring a protracted interval, 740–660 Ma ago (Fairchildand Kennedy 2007). There are younger episodes as well,*635 Ma Marinoan and *580 Ma Gaskiers (Fairchildand Kennedy 2007), but the abundance of 580–710 Maigneous activity in the region (Stern and Hedge 1985;Moussa et al. 2008) and the lack of such cobbles in thediamictite is inconsistent with deposition of the Atud dia-mictite during either of these younger ice ages. The Atuddiamictite could be similar in age to 712 ± 1.6 Ma dia-mictite of the Ghubrah Formation, Oman, although theclast age populations are significantly different (Rieu et al.2007).Origin of pre-Neoproterozoic clasts in the diamictiteand significance for diamictite formationThere is no obvious source within the Eastern Desert orNW Saudi Arabia for especially the pre-Neoproterozoiccobbles within the diamictite. Stern (2002) summarized Ndisotopic evidence that the Neoproterozoic basement ofEgypt and NW Saudi Arabia is overwhelmingly a juvenileNeoproterozoic crustal addition. There are no knownexposures of pre-Neoproterozoic rocks from the EasternDesert of Egypt, Sinai, or NW Saudi Arabia. Furthermore,there are no significant exposures of *750 Ma plutonicrocks in the immediate vicinities of the diamictite locali-ties, although volcanic rocks of this age are increasinglybeing recognized in the Central Eastern Desert of Egypt(Andresen et al. 2008; Ali et al. 2008). Plutonic igneousrocks *750 Ma are known from farther south in SudanSudanYemenKhidaW Desert1500250035001000200030000246810121416NE Africa and Arabia pre-Neoproterozoic agesN = 62bU-Pb zircon age (Ma)0246810121416Atud diamictite pre-Neoproterozoic agesN = 105ConcordantDiscordantMatrixFrequency (n)aFig. 12 Comparative histograms of single-grain zircon ages analysedduring this study and ages for pre-Neoproterozoic outcrops around theArabian–Nubian Shield. a Histogram showing the distribution ofAtud diamictite pre-Neoproterozoic ages from this study. b Histo-gram showing distribution of pre-Neoproterozoic ages obtained forrocks from surrounding regions in western Egypt, northern Sudan,and southern Saudi Arabia, from Whitehouse et al. (1998); Sultanet al. (1994); Agar et al. (1992) and Stern et al. (1994). Discordantages \1,000 Ma are238U/206Pb ages and those [1,000 Ma are207Pb/206Pb ages. Note the similar distributions observed for thesehistograms; the implications of this are further explored in the textInt J Earth Sci (Geol Rundsch) (2010) 99:705–726723123and Saudi Arabia (Stern and Abdelsalam 1998; Hargroveet al. 2006a, b).The source of at least the pre-Neoproterozoic clasts mustcome from outside the ophiolite-diamictite-BIF basin out-lined in Fig. 1b. Dixon (1981) inferred that the 1,120–2,060 Ma cobbles within the Atud diamictite were trans-ported from a region of pre-Neoproterozoic crust west ofthe Nile River, possibly by ice-rafting. This is consistentwith more recent studies of Sultan et al. (1994), whoreported Nd model ages of 2.85 Ga, U–Pb (zircon andsphene) and Sm–Nd whole rock crystallization ages[2.67 Ga and metamorphic ages of *2.0 Ga for JabalKamil and Jabal El Asr (Fig. 1a), Western Desert of Egypt,confirming that late Archean and Paleoproterozoic crustunderlies much of the region. Pre-Neoproterozoic Atud andNuwaybah clasts could also have been derived from theSE. Stacey and Hedge (1984) reported *1.8 Ga zirconages for from pre-Neoproterozoic sources in the Khidaterrane, Saudi Arabia. Whitehouse et al. (1998) confirmedthat late Archean (2.55–2.95) crust existed in the Al-Mafidterrane of Central southern Yemen. Figure 1a shows thelocation of these pre-Neoproterozoic crustal tracts thatcould have been the source of pre-Neoproterozoic detritusfor the Atud diamictite. All of these potential sites arehundreds of kilometers away from the present Atud dia-mictite localities.Further insights can be obtained by comparing the agedistribution of Atud diamictite ages and those of potentialsource regions. Figure 12 compares the distribution of pre-Neoproterozoic zircon ages for the diamictite and the pre-Neoproterozoic ages of known potential sources around theANS from previous studies. The potential sources weconsider are the Khida terrane, Yemen, and crust of theSaharan Metacraton (Abdelsalam et al. 2002) as repre-sented by exposures west of the Nile in Egypt and at theEgypt–Sudan border at Wadi Halfa. The lack of *2.5 Gaages in the Khida terrane makes this an unlikely source,and the relatively subordinate abundance of *2.0 Ga agesin Yemen makes this relatively unattractive as the source ofthe diamictite. The similar abundance of *2.0 and* 2.5Ga ages in those parts of the Saharan Metacratonflanking the Eastern Desert is quite consistent with thedistribution of ages seen in the diamictite. Further supportfor this potential source comes from the observation that aminor proportion of *1.2 Ga ages are also recognized forbasement of the Saharan Metacraton.We conclude that the evidence is most consistent withan interpretation whereby the diamictite clasts were erodedfrom the Saharan Metacraton during a major episode oferosion and transported eastwards or northeastwards intoan oceanic basin, where they were deposited. We have nodirect evidence that erosional agent was glacial or that thediamictite was transport eastwards by glaciers and/oricebergs, but cannot imagine any other mechanism that iscapable of eroding and transporting such large blocks ofdiverse lithologies such great distances. There is also apossibility that the diamictite is some kind of an impactbreccia, but there is limited evidence for this interpretation.The microconglomerate clasts could be explained by such amechanism, but we saw no evidence for shocked quartz orother evidence expected from an impact. The intenselyaltered nature of these rocks does not favor preservation ofsuch features. We conclude from the above considerationsthat the Atud diamictite manifests the activity of Neopro-terozoic glacial in the Arabian–Nubian Shield.ConclusionsThe following conclusions result from our study:1.Units formerly referred to as Atud Conglomerate orAtud Formation in Egypt are better described as AtudDiamictite.2.Diamictite is demonstrated for the first time from nearNuwaybah in NW Arabia. Diamictites at Wadi Kareimand Wadi Mobarak in the Eastern Desert of Egypt andin NW Saudi Arabia are correlative.3.Atud diamictite was deposited in an oceanic basin ofindeterminate size, as demonstrated by regional asso-ciation with ophiolite and BIF in the Central EasternDesert of Egypt.4.The diamictite was deposited not long after *750 Ma.5.The source of clasts within the diamictite is not knownbut transport over several hundreds of kilometers isrequired. Transporting such large and lithologicallydiverse clasts over such large distances may haveoccurred by icebergs and glaciers, or formation as animpact breccia, but we prefer the suggestion that theAtud and Nuwaybah diamictite bodies are glacialdeposits in the Arabian–Nubian Shield.6.Deformation has obscured primary depositional struc-tures, striations, and faceting so a clear glacial origin isnot demonstrated. Nevertheless combined conclusions4 and 5 strongly suggest deposition broadly Sturtian(*740–660 Ma) glacial episode in the Arabian–Nubian Shield.7.Further studies of especially the large body of Atud/Nuwaybah diamictite near Jabal Atud are needed, asare efforts to find other outcrops in Saudi Arabia.Regional variations in the thickness of the diamictitewill help constrain the direction of its source.AcknowledgmentsThis paper is part of the first author’s Ph.D.research at the University of Texas at Dallas and was funded by NSFgrant EAR- 0509486. We thank the Nuclear Materials Authority ofEgypt (NMA) and the Saudi Geological Survey (SGS) for support in724Int J Earth Sci (Geol Rundsch) (2010) 99:705–726123the field. Special thanks go to Dr. Hani Shalaby, Dr. Hossam Khamis,Mr. Fayek Kattan, and Mr. Saad El Garney for their help during fieldwork in Egypt and Saudi Arabia. We appreciate the assistance of Dr.Joe Wooden, Dr. Frank Mazdab and Dr. W. R. Griffin during ana-lytical work at the SUMAC facility. Also much appreciation goes toDr. U. S. Hargrove III for helpful advice and discussions duringzircon separations. We also thank referees Dr. Nicole Dobrzinski andDr. James Etienne and editor Prof. Wolf-Christian Dullo for insightfulcomments and criticisms that improved this manuscript. This is UTDGeosciences contribution number 1124.ReferencesAbdel Naby H, Frisch W, Hegner E (2002) Origin of Wadi Haimur–Abu Swayel gneiss belt, south Eastern Desert, Egypt: petrolog-ical and geochronological constraints. Precambrian Res 113:307–332. doi:10.1016/S0301-9268(01)00214-5Abdelsalam MG, Lie´geois JP, Stern RJ (2002) The Saharan metacr-aton. J Afr Earth Sci 34:119–136. doi:10.1016/S0899-5362(02)00013-1Abu El Ela FF (1990) Do the Atud conglomerates belong to the islandarc metasediments? Bull Fac Sci Assiut Univ 19(1-F):123–155Agar RA, Stacey JS, Whitehouse MJ (1992) Evolution of the southernAfif terrane—a geochronologic study. Saudi Arabian DeputyMinistry for mineral resource open file report DGMR-OF-10-15:41Akaad MK, Abu El Ela AM (1996) Geological and petrochemistry ofthe Muweilih submarine metabasalts, Qift–Quseir region, East-ern Desert, Egypt. Egypt J Geol 40:321–349Akaad MK, Noweir AM (1964) A new occurrence of the Atudformation at Wadi El-Muweilih, Eastern Desert of Egypt. BullSci Tech Assiut Univ 7:31–48Akaad MK, Noweir AM (1969) Lithostratigraphy of Hammamat-UmSeleimat district, Eastern Desert, Egypt. Nature 223:284–285.doi:10.1038/223284a0Akaad MK, Noweir AM, Abu El Ela AM (1995) The volcano-sedimentary association and ophiolites of Wadi Mubarak,Eastern Desert, Egypt. Proc Int Conf Geol Surv Egypt SpecPubl 69:231–248Akaad MK, Noweir AM, Abu El Ela AM, El-Bahariya GA (1996)The Muweilih conglomerate; Qift-Quseir road region, and theproblem of the Atud formation. Presented at Proceedings of theGeological Survey Egypt, pp 23–45Ali KA, Stern RJ, Manton WI, Kimura J-I, Khamees HA (2008)Geochemistry, Nd isotopes and U–Pb SHRIMP dating ofNeoproterozoic volcanic rocks from the Central Eastern Desertof Egypt: new insights into the*750 Ma crust-forming event.Precambrian Res (in review)Allen PA, Leather J (2006) Post-Marinoan marine siliciclasticsedimentation: the Masirah bay formation, Neoproterozoic Huqfsupergroup of Oman. Precambrian Res 144:167–198. doi:10.1016/j.precamres.2005.10.006Allen PA, Leather J, Brasier MD (2004) The Neoproterozoic Fiqglaciation and its aftermath, Huqf supergroup of Oman. BasinRes 16:507–534. doi:10.1111/j.1365-2117.2004.00249.xAndresen A, Abu El-Rus MA, Myhre PI, Boghdady GY, Corfu F(2008) U–Pb TIMS age constraints on the evolution of theNeoproterozoic Meatiq Gneiss dome, Eastern Desert, Egypt. IntJ Earth Sci (in press)Azer M, Stern RJ (2007) Neoproterozoic serpentinites in the EasternDesert, Egypt: fragments of fore-arc mantle. Geology 115:457–472. doi:10.1086/518052Boulton GS (1978) Boulder shapes and grain size distributions ofdebris as indicators of transport paths through a glacier and tillgenesis. Sedimentlogy 25:773–799. doi:10.1111/j.1365-3091.1978.tb00329.xBowring SA, Grotzinger JP, Condon DJ, Ramezani J, Newall M,Allen PA (2007) Geochronologic constraints on the chrono-stratigraphic framework of the Neoproterozoic Huqf supergroup,Sultanate of Oman. Am J Sci 307:1097–1145. doi:10.2475/10.2007.01Corfu F, Hanchar JM, Hoskin PW, Kinny P (2003) Atlas of zircontextures. Rev Mineral Geochem 53:469–500. doi:10.2113/0530469Davies FB (1985) Geological map of the Al Wajh Quadrangle, Sheet26B, Kingdom of Saudi Arabia, Ministry of petroleum andmineral resources, Kingdom of Saudi ArabiaDixon TH (1979) The evolution of continental crust in the latePrecambrian Egyptian Shield. (Ph.D. thesis): University ofCalifornia, San Diego, p 231Dixon TH (1981) Age and chemical characteristics of some Pre-Pan-African rocks in the Egyptian Shield. Precambrian Res 14:119–133. doi:10.1016/0301-9268(81)90017-6El-Essawy MA (1964) Geology of the area east of Gabal Atud,Eastern Desert, Egypt. [M.Sc. thesis]: Assiut University, Assiut,p 235Eyles N, Januszcak N (2004) ‘Zipper-rift’: a tectonic model forNeoproterozoic glaciations during the breakup of Rodinia after750 Ma. Earth Sci Rev 65:1–73Fairchild IJ, Kennedy MJ (2007) Neoproterozoic glaciation in theearth system. J Geol Soc London 64:895–921Flint RF, Sanders JE, Rodgers J (1960) Diamictite, a substitute termsymmictite. Geol Soc Am Bull 71:1809–1810Goldring DC (1990) Banded iron formation of Wadi Sawawindistrict, Kingdom of Saudi Arabia. Tran Instn Min Metall (SectB: Appl Earth Sci) 99: B1–B14Hambrey MJ, Glasser NF (2002) Development of landform andsediment assemblages at maritime high-Arctic glaciers. In:Hewitt K (ed) Landscapes of transition. Kluwer, Dordrecht, pp11–42Hargrove USI, Stern RJ, Griffin WR, Johnson PR, Abdelsalam MG(2006a) From island arc to craton: timescales of crustalformation along the Neoproterozoic Bi’r Umq suture zone,Kingdom of Saudi Arabia. Saudi Geological Survey TechnicalReport SGS-TR-2006-6: 69Hargrove US, Stern RJ, Kimura JI, Manton WI, Johnson PR (2006b)How juvenile is the Arabian–Nubian Shield? Evidence from Ndisotopes and pre-Neoproterozoic inherited zircon in the Bi’rUmq suture zone, Saudi Arabia. Earth Planet Sci Lett 252:308–326Hoffman PF (2007) Comment on ‘‘Snowball Earth on trial’’. EOS88(9):110Jacobs J, Thomas RJ (2004) Himalayan-type indenter-escape tecton-ics model for the southern part of the late Neoproterozoic-earlyPaleozoic east African–Antarctic Orogen. Geology 32:721–724Kennedy A, Johnson PR, Kattan FH (2004) SHRIMP geochronologyin the northern Arabian Shield. Part I: Data acquisition. SaudiGeological Survey Open File Report, SGS-OF–2004-11Kennedy A, Johnson PR, Kattan FH (2005) SHRIMP geochronologyin the northern Arabian Shield. Part II: Data acquisition 2004.Saudi Geological Survey Open File Report, SGS-OF-2005-10Kennedy A, Johnson PR, Kattan FH (2007) SHRIMP geochronologyin the northern Arabian Shield: Part III Data Acquisition 2006.Saudi Geological Survey, Jeddah (in press)Kro¨ner A, Sassi FP (1996) Evolution of the northern Somalibasement: new constraints from zircon ages. J Afr Earth Sci22:1–15Int J Earth Sci (Geol Rundsch) (2010) 99:705–726725123Kro¨ner A, Todt W, Hussein IM, Mansour M, Rashwan AA (1992)Dating of late Proterozoic ophiolites in Egypt and the Sudanusing the single zircon evaporation technique. Precambrian Res59:15–32Le Guerroue´ E, Allen PA, Cozzi A (2005) Two distinct glacialsuccessions in the Neoproterozoic of Oman. GeoArabia 10:17–34Leather J (2001) Sedimentlogy, chemostratigraphy and geochronol-ogy of the lower Huqf Supergroup, Oman (Ph.D. thesis): TrinityCollege, Dublin, p 227Ludwig KR (2000) Isoplot/Ex 3.0: a geological toolkit for MicrosoftEXCEL. Berkeley Geochronology Center Special Publication,1a: 1–54Moghazi AM (2003) Geochemistry and petrogenesis of a high-K calc-alkaline Dokhan volcanic suite, South Safaga area, Egypt: therole of late Neoproterozoic crustal extension. Precambrian Res125:161–178Moussa EM, Stern RJ, Manton WI, Ali KA (2008) SHRIMP zircondating and Sm/Nd isotopic investigations of Neoproterozoicgranitoids, Eastern Desert, Egypt. Precambrian Res 160:341–356Mukherjee SK, Stern RJ, Ali KA, Miller NR, Johnson PR, White-house MJ (2009) Petrology, age (U–Pb zircon) and geochemistryof Sawawin banded iron formation, NW Arabia-implications forthe Neoproterozoic climate change controversy. PrecambrianRes (submitted)Ovenshine AT (1970) Observations of iceberg rafting in Glacier bay,Alaska, and the identification of ancient ice-rafted deposits. GeolSoc Am Bull 81:891–894Ries AC, Shackelton RM, Graham RH, Fitches WR (1983) PanAfrican structures, ophiolites and melange in the Eastern Desertof Egypt: a traverse at 26 N. J Geol Soc London 140:75–95Rieu R, Allen PA, Cozzi A, Kosler J, Bussy F (2007) A compositestratigraphy for the Neoproterozoic Huqf supergroup of Oman:integrating new litho-, chemo- and chronostratigraphic data ofthe Mirbat area, southern Oman. J Geol Soc London 164:997–1009Shalaby A, Stuwe K, Makroum F, Fritz H, Kebede T, Klotzli U(2005) The Wadi Mubarak belt, Eastern Desert of Egypt: aNeoproterozoic conjugate shear system in the Arabian–NubianShield. Precambrian Res 136:27–50Sims PK, James HL (1984) Banded iron-formations of late Protero-zoic age in the Central Eastern Desert, Egypt: geology andtectonic setting. Econ Geol 79:1777–1784Stacey JS, Agar RA (1985) U–Pb isotopic evidence for the accretionof a continental microplate in the Zalm region of the SaudiArabian Shield. J Geol Soc London 142:1189–1203Stacey JS, Hedge CE (1984) Geochronologic and isotopic evidencefor early Proterozoic crust in the Eastern Arabian Shield.Geology 12:310–313Stern RJ (1979) Late Precambrian ensimatic volcanism in the CentralEastern Desert of Egypt. Ph.D. thesis. University of California,San Diego, CA, p 210Stern RJ (1981) Petrogenesis and tectonic setting of late Precambrianensimatic volcanic rocks, Central Eastern Desert of Egypt.Precambrian Res 16:197–232Stern RJ (1994) Arc assembly and continental collision in theNeoproterozoic east African Orogen: implications for theconsolidation of Gondwanaland. Annu Rev Earth Planet Sci22:319–351Stern RJ (2002) Crustal evolution in the east African Orogen: aNeodymium isotopic perspective. J Afr Earth Sci 34:109–117Stern RJ, Abdelsalam MG (1998) Formation of juvenile continentalcrust in the Arabian–Nubian Shield: evidence from graniticrocks of the Nakasib suture, NE Sudan. Geol Rundsch 87:150–160Stern RJ, Hedge CE (1985) Geochronologic and isotopic constraintson late Precambrian crustal evolution in the Eastern Desert ofEgypt. Am J Sci 285:97–127Stern RJ, Kro¨ner A, Bender R, Reischmann T, Dawoud AS (1994)Precambrian basement around Wadi Halfa, Sudan: a newperspective on the evolution of the east Saharan craton. GeolRundsch 83:564–577Stern RJ, Avigad D, Miller NR, Beyth M (2006) Evidence for theSnowball Earth hypothesis in the Arabian–Nubian Shield and theeast African Orogen. J Afr Earth Sci 44:1–20Sultan M, Becker R, Arvidson RE, Shore P, Stern RJ, El Alfy Z, AttiaR (1993) New constraints on Red Sea rifting from correlations ofthe Arabian and Nubian Neoproterozoic outcrops. Tectonics12:1303–1319Sultan M, Tucker RD, El Alfy Z, Attia R, Ragab AG (1994) U–Pb(zircon) ages for the gneissic terrane west of the Nile, southernEgypt. Geol Rundsch 83:514–522Tera F, Wasserburg GJ (1972) U–Th–Pb systematics in three Apollo14 basalts and the problem of initial Pb in lunar rocks. EarthPlanet Sci Lett 14:281–304Walraven F, Rumvegeri BT (1993) Implications of whole-rock Pb–Pband zircon evaporation dates for the early metamorphic historyof the Kasai craton, southern Zaire. J Afr Earth Sci 16:395–404Wetherill GW (1956) Discordant uranium–lead ages I. Trans AmGeophys Union 37:320–326Whitehouse MJ, Windley BF, Ba-Bttat MA, Fanning CM, Rex DC(1998) Crustal evolution and terrane correlation in the EasternArabian Shield, Yemen: geochronological constraints. J GeolSoc London 155:281–295Whitehouse MJ, Stoeser DB, Stacey JS (2001a) The Khida terrane––geochronological and isotopic evidence for Paleoproterozoic andArchean crust in the Eastern Arabian Shield of Saudi Arabia(extended abstract). Gondwana Res 4(2):200–202Whitehouse MJ, Windley BF, Stoeser DB, Al-Khirbash S, Ba-BttatMA, Haider A (2001b) Precambrian basement character ofYemen and correlations with Saudi Arabia and Somalia.Precambrian Res 105:357–369Wilde SA, Youssef K (2000) Significance of shrimp U–Pb dating ofimperial porphyry and associated Dokhan volcanics, GebelDokhan, north Eastern desert, Egypt. J Afr Earth Sci 31:403–413726Int J Earth Sci (Geol Rundsch) (2010) 99:705–726123View publication stats | RESEARCH GATE |
CA05CH17_ZonARjats.clsNovember 23, 202017:20Annual Review of Cancer BiologyReeling in the ZebrafishCancer ModelsAlicia M. McConnell,1,2,3,∗ Haley R. Noonan,1,2,3,4,∗and Leonard I. Zon1,2,3,51Stem Cell Program and Division of Hematology/Oncology, Boston Children’s Hospital andDana-Farber Cancer Institute, Boston, Massachusetts 02115, USA2Harvard Stem Cell Institute, Boston, Massachusetts 02138, USA3Harvard Medical School, Boston, Massachusetts 02115, USA4Biological and Biomedical Sciences Program, Harvard Medical School, Boston,Massachusetts 02115, USA5Stem Cell and Regenerative Biology Department and Howard Hughes Medical Institute,Harvard University, Boston, Massachusetts 02138, USA; email: zon@enders.tch.harvard.eduAnnu. Rev. Cancer Biol. 2021. 5:17.1–17.20The Annual Review of Cancer Biology is online atcancerbio.annualreviews.orghttps://doi.org/10.1146/annurev-cancerbio-051320-014135Copyright © 2021 by Annual Reviews.All rights reserved∗These authors contributed equally to this articleKeywordszebrafish, cancer, transgenic, modeling, genetics, tumorAbstractZebrafish are rapidly becoming a leading model organism for cancer re-search. The genetic pathways driving cancer are highly conserved betweenzebrafish and humans, and the ability to easily manipulate the zebrafishgenome to rapidly generate transgenic animals makes zebrafish an excellentmodel organism. Transgenic zebrafish containing complex, patient-relevantgenotypes have been used to model many cancer types. Here we present acomprehensive review of transgenic zebrafish cancer models as a resourceto the field and highlight important areas of cancer biology that have yet tobe studied in the fish. The ability to image cancer cells and niche biology inan endogenous tumor make zebrafish an indispensable model organism inwhich we can further understand the mechanisms that drive tumorigenesisand screen for potential new cancer therapies.17.1, .•·�-Review in Advance first posted on December 2, 2020. (Changes may still occur before final publication.) Access provided by 158.46.185.149 on 12/23/20. For personal use only. CA05CH17_ZonARjats.clsNovember 23, 202017:20INTRODUCTIONZebrafish (Danio rerio) share 70% of their genome with humans and are commonly used to modelhuman disease (Howe et al. 2013). Rapid external development, high fecundity, and easy, low-cost maintenance make the zebrafish an attractive animal model. Optical transparency has greatlyenhanced the ability to visualize internal cell biology using fluorescent reporters and makes thezebrafish uniquely suited to image tumor development, metastasis, and microenvironmental inter-actions. In addition, zebrafish are excellent model organisms to use in drug screens given the easewith which drugs can be administered to zebrafish embryos in their water. Adult zebrafish can beused for drug screening as well; however, it can be difficult to administer the concentration of thedrug necessary exclusively in their water. Techniques such as oral gavage, intraperitoneal injection,and retro-orbital injection have been developed to address this problem (Dang et al. 2016, Kinkelet al. 2010, Pugach et al. 2009). Systems for perturbing candidate genes are readily available in thezebrafish, via either overexpression in a DNA transposon system, knockdown using morpholinos,or knockout with CRISPR (clustered regularly interspaced short palindromic repeats)/Cas9. Thiscan be done with both spatial and temporal control. Fish are known to spontaneously develop can-cer, particularly after mutagenesis, and have high conservation of oncogenes (White et al. 2013).One common way to model cancer in zebrafish is through xenotransplantation of human cancercells (reviewed in White et al. 2013). Although cancers such as head and neck, retinoblastoma, andsquamous cell carcinoma have mostly been modeled via a transplant system, this systematic re-source review focuses exclusively on genetic models of cancer in zebrafish (Figure 1, summarizedin Table 1).MELANOMAMelanoma arises from pigment-producing melanocytes and is the deadliest form of skin cancer(Lo & Fisher 2014). Melanocyte development is conserved between zebrafish and mammals, mak-ing zebrafish excellent models of pigmentation and melanoma (Mort et al. 2015). The master reg-ulator of the melanocyte lineage, MITF, is conserved in zebrafish and required for melanocyte de-velopment (Lister et al. 1999). Approximately 50% of melanoma cases harbor mutations in BRAF,leading to activation of the MAPK pathway (TCGA 2015). Expression of human BRAFV600E, un-der the melanocyte-specific mitfa promoter leads to formation of melanocytic nevi (moles), butNeuroblastomaKidneycancerSarcomaMPNSTThyroidcancerLivercancer PancreaticcancerGerm cellcancerLeukemiaMelanomaGliomaFigure 1Human cancers that have been genetically modeled in zebrafish. Size of text correlates with how many zebrafish models have beendescribed for that cancer type. Abbreviation: MPNST, malignant peripheral nerve sheath tumor. Figure adapted from an image createdwith BioRender.com.17.2McConnell • Noonan • Zon, .•·�-Review in Advance first posted on December 2, 2020. (Changes may still occur before final publication.) Access provided by 158.46.185.149 on 12/23/20. For personal use only. CA05CH17_ZonARjats.clsNovember 23, 202017:20Table 1Transgenic zebrafish models of cancerCancer typeGenes mutatedReference(s)Gliomap53e7/e7; nf1a+/−; nf1b−/−Shin et al. 2012Tg(ptf1:Gal4; UAS:GFP; UAS:DA-RAC1)Tg(ptf1:Gal4; UAS:GFP; UAS:DA-AKT1)Jung et al. 2013Tg(krt5:Gal4VP16; UAS:mCherry-KRASG12V)Tg(gfap:Gal4VP16; UAS:mCherry-KRASG12V)Ju et al. 2015Tg(zic4:Gal4TA4; UAS:mCherry; UAS:GFP-HRASG12V)Tg(zic4:Gal4TA4; UAS:mCherry; UAS:Xmrk)Tg(zic4:Gal4TA4; UAS:mCherry; UAS:BRAFV600E)Tg(zic4:Gal4TA4; UAS:mCherry; UAS:AKT-BFP)Tg(zic4:Gal4TA4; UAS:mCherry; UAS:eGFP-EGFRvIII)Tg(zic4:Gal4TA4; UAS:mCherry; UAS:YAPS5A)Mayrhofer et al. 2017Tg(sox10:mCherry-NRASwt) +/− p53M214KTg(sox10:mCherry-NRASQ61R) +/− p53M214KTg(sox10:mCherry-NRASS17N) +/− p53M214KModzelewska et al. 2016NeuroblastomaTg(dβh:EGFP; dβh:ALKF1174L)Tg(dβh:EGFP-MYCN)Zhu et al. 2012Tg(dβh:EGFP-MYCN); nf1a−/−; nf1b+/−He et al. 2016Tg(dβh:ptpn11E69K-EGFP; dβh:EGFP-MYCN)Tg(dbh:Gab2wt; dbh:EGFP; dβh:EGFP-MYCN)Zhang et al. 2017Tg(dβh:EGFP-MYCN; dβh:LMO1; dβh:mCherry)Zhu et al. 2017Tg(dβh:MYCN; dβh:EGFP)Tao et al. 2017Tg(dβh:MYC)Zimmerman et al. 2018MPNSTrp+/− (rps8a, rps15a, rpl7, rpl35, rpl36, rpl36a, rpl13, rpl23a, rps7, rps18, rps29)Amsterdam et al. 2004rp+/− or p53M214KZhang et al. 2013brca2Q658X/Q658X; tp53M214K/+Shive et al. 2014Tg(mitfa:atg5K130R); p53M214K/+Lee et al. 2016Tg(sox10:PDGFRA; sox10:mCherry); nf1a+/−; nf1b−/−; p53m/mTg(sox10:PDGFRAmut; sox10:mCherry); nf1a+/−; nf1b−/−; p53m/mKi et al. 2017nf1a+/−; nf1b−/−; p53m/m; atrx+/−Oppel et al. 2019Pancreatic cancerz-myod:MYCNcore-z-myod:MYCNHong et al. 2004Tg(ptf1a:eGFP-KRASG12V)Park et al. 2008rpl36+/−; Tg(ptf1a:gal4VP16; UAS:GFP-KRASG12V)Provost et al. 2014Tg(ptf1a:gal4VP16; UAS:GFP-KRASmut)Park et al. 2015Tg(ubb:Lox-Nuc-mCherry-stop-Lox-GFP::KRASG12D)Oh and Park 2019Liver cancerTg(pLF2.8-HCV-core)Rekha et al. 2008Tg(fabp10:EGFP-KRASG12V) +/− p53M214KNguyen et al. 2011Tg(fabp10:TA; TRE:xmrk)Li et al. 2012Tg(fabp10:TA; TRE:Myc)Li et al. 2013Tg(fabp10:LexPR; LexA:EGFP-KRASG12V)Nguyen et al. 2012Tg(fabp10:LexPR; LexA:Cre; fabp10:loxP-mCherry-loxP-EGFP-KRASG12V)Nguyen et al. 2016Tg(fabp10a:pt-β-cat)Evason et al. 2015apc+/−Haramis et al. 2006Tg(fabp10a:tTA; pT2-TRE-gankyrin-HcRed)Huang et al. 2017Tg(fabp10:NRASQ61K)Wang et al. 2017(Continued)www.annualreviews.org • Reeling in the Zebrafish Cancer Models17.3, .•·�-Review in Advance first posted on December 2, 2020. (Changes may still occur before final publication.) Access provided by 158.46.185.149 on 12/23/20. For personal use only. CA05CH17_ZonARjats.clsNovember 23, 202017:20Table 1(Continued)Cancer typeGenes mutatedReference(s)Thyroid cancerTg(tg:BRAFV600E-TOM)Anelli et al. 2017RhabdomyosarcomaTg(rag2:KRASG12D) +/− p53−Langenau et al. 2007,Ignatius et al. 2018Tg(cdh15:KRASG12D)Tg(mylz2:KRASG12D)Storer et al. 2013Tg(β-actin:LoxP-EGFP-LoxP-KRASG12D)Le et al. 2007Tg(ubi:GFP2A-PAX3FOXO1)Tg(β-actin:GFP2A-PAX3FOXO1)Tg(mitfa:GFP2A-PAX3FOXO1)Tg(CMV:GFP2A-PAX3FOXO1)Tg(fli1:GFP2A-PAX3FOXO1)Kendall et al. 2018Ewing’s sarcomaTg(hsp70:EWS-FLI1) +/− p53−Tg(hsp70:EWS-FLI1:IRES-GFP) +/− p53−Tg(β-actin:EWS-FLI1:IRES-GFP) +/− p53−Leacock et al. 2012ewsa−/−; p53M214K/M214KPark et al. 2016LiposarcomaTg(rag2:myr-mAkt2) +/− p53M214KGutierrez et al. 2011bTg(krt4:Has.myrAkt1)cy18Chu et al. 2012MelanomaTg(mitfa:BRAFV600E); p53−/−Patton et al. 2005Tg(mitf:BRAFV600E); p53−/−; mitfa−/− + MiniCoopRCeol et al. 2011Tg(mitfa:EGFP:NRASQ61K); p53−/−Dovey et al. 2009Tg(kita:HRASG12V)Santoriello et al. 2010Tg(mitfa:NRASQ61R)McConnell et al. 2019AMLTg(spi1:MYST3/NCOA2-EGFP)Zhuravleva et al. 2008Tg(spi1:FLT3-ITD-2A-EGFP/CG2) +/− spi1:NPM1-Mut-PA/CG2Lu et al. 2016Tg(spi1::loxP-EGFP-loxP::NUP98-HOXA9)Forrester et al. 2011asxl1+/− +/− tet2−/−Gjini et al. 2019RUNX1-CBF2TKalev-Zylinska et al.2002Tg(hsp70:AML1-ETO)Yeh et al. 2008Tg(MYCN:HSE:EGFP)Shen et al. 2013Tg(ß-actin-LoxP-EGFP-LoxP-KRASG12D; hsp70-Cre)Le et al. 2007Tg(fli.1:Gal4FFubs3; UAS:egfp-HRASG12V)io6Alghisi et al. 2013stat5.1N646H or stat5.1H298R/N714FLewis et al. 2006Tg(pu.1:EGFP-CREB)Tregnago et al. 2016irf8−/−Zhao et al. 2018idh1−/−IDH1-R132HShi et al. 2015CMLTg(hsp70:p210BCR/ABL1)Xu et al. 2020ALLTg(rag2:cMyc)Langenau et al. 2003Tg(rag2:EGFP-mMyc)Langenau et al. 2005Tg(rag2:MYC-ER)Gutierrez et al. 2011aTEL-AML1Sabaawy et al. 2006Tg(rag2-ICN1-EGFP)Chen et al. 2007(Continued)17.4McConnell • Noonan • Zon, .•·�-Review in Advance first posted on December 2, 2020. (Changes may still occur before final publication.) Access provided by 158.46.185.149 on 12/23/20. For personal use only. CA05CH17_ZonARjats.clsNovember 23, 202017:20Table 1(Continued)Cancer typeGenes mutatedReference(s)Renal cellvhl−/−Van Rooijen et al. 2009Germ cellENU mutagenesisNeumann et al. 2009Tg(flck:TAg)Gill et al. 2010lrrc50Hu255hBasten et al. 2013dnaaf1hu255hLitchfield et al. 2016ns1402Shimizu & Matsuda2019Abbreviations: ALL, acute lymphoblastic leukemia; AML, acute myeloid leukemia; CML, chronic myeloid leukemia; DA, dominant active; MPNST, malig-nant peripheral nerve sheath tumor; wt, wild-type.not melanoma, in adult zebrafish (Patton et al. 2005). Expression of mitfa:BRAFV600E , in conjunc-tion with a homozygous p53 missense mutation, leads to approximately half of mitfa:BRAFV600E;p53−/− fish developing melanomas by 4 months old (Patton et al. 2005). This represented the firstmodel of melanoma in zebrafish.Zebrafish allow for transient overexpression and knockout of melanoma-associated genes. TheMiniCoopR system was developed in order to perform tissue-specific overexpression of a candi-date gene (Ceol et al. 2011). A mitfa−/− mutation in Tg(mitf:BRAFV600E); p53−/− fish preventsmelanocyte development and melanoma from occurring. Melanocyte development can be res-cued in these fish via injection of MiniCoopR, a transposon-based vector that contains a mitfaminigene, which sits alongside a candidate oncogene driven by the mitfa promoter (Ceol et al.2011). Modification of the CRISPR/Cas9 system allows for melanocyte-specific knockout of agene of interest (Ablain & Zon 2016, Ablain et al. 2015). The MAZERATI (modeling approachin zebrafish for rapid tumor initiation) system takes advantage of both MiniCoopR and CRISPRMiniCoopR vectors, allowing for rapid oncogene expression and tumor suppressor inactivationexclusively in the melanocytes of zebrafish (Ablain et al. 2018). Most recently, this system was usedto identify SPRED1 as a tumor suppressor in KIT-mutant melanomas and GDF6 as an oncogenein Tg(mitf:BRAFV600E); p53−/−; mitfa−/− melanomas (Ablain et al. 2018, Venkatesan et al. 2018).Approximately 28% of melanoma cases contain an NRAS mutation (TCGA 2015). Themitfa:NRASQ61K mutation causes hyperpigmentation but does not generate melanomas in ze-brafish (Dovey et al. 2009). However, as with BRAFV600E, NRASQ61K cooperates with p53 loss togenerate melanomas. Melanomas that are mitfa:EGFP:NRASQ61K; p53−/− mutants are histolog-ically and transcriptionally similar to human melanomas, as well as transplantable (Dovey et al.2009). The BRAF and NRAS zebrafish models use mitfa to drive oncogene expression and onlydevelop melanomas in the presence of a p53 mutation. Santoriello et al. (2010) reported a zebrafishmelanoma model in which the kita promoter drives HRAS oncogene expression. This model de-velops melanomas by 1–3 months of age and does not require an additional tumor suppressor mu-tation. McConnell et al. (2019) generated a transient mitfa:NRASQ61R zebrafish melanoma modelthat does not require a tumor suppressor mutation. These fish develop large hyperpigmentedpatches at 2 weeks postfertilization, and by 8 weeks these patches rapidly form tumors. This NRASmodel was used to investigate neural crest state activation.Melanocytes arise from the embryonic neural crest lineage. Zebrafish melanomas in mitf-BRAFV600E; p53−/− and mitfa:NRASQ61R fish reactivate a neural crest gene signature, including thezebrafish gene crestin (White et al. 2011), which marks the neural crest during development, turnsoff by 72 hours postfertilization, and turns on again exclusively in zebrafish melanomas. Melanomainitiation can be visualized using the crestin reporter. Expression of crestin begins as a singlewww.annualreviews.org • Reeling in the Zebrafish Cancer Models17.5, .•·�-Review in Advance first posted on December 2, 2020. (Changes may still occur before final publication.) Access provided by 158.46.185.149 on 12/23/20. For personal use only. CA05CH17_ZonARjats.clsNovember 23, 202017:20crestin-positive cell and progresses to a fully formed, transplantable, crestin-positive melanoma(Kaufman et al. 2016). Zon and colleagues utilized a chemical screen with wild-type embryosto identify small-molecule suppressors of the crestin-positive neural crest lineage, identifying aDHODH inhibitor, leflunomide (Santoriello et al. 2020, White et al. 2011). Chemical screenshave also been conducted using primary zebrafish embryonic cell cultures, identifying caffeic acidphenethyl ester as a suppressor of neural crest development (Ciarlo et al. 2017).One caveat of the previously described models is the inability to spatially and temporally con-trol melanoma formation. Transgene electroporation in adult zebrafish (TEAZ) was developedto introduce plasmids in zebrafish with both spatial and temporal control (Callahan et al. 2018).Plasmids of interest are injected into the zebrafish flank just below the epithelium and electrodesdeliver an electric pulse, creating pores in the cell surface that facilitate uptake of plasmid DNA.TEAZ can be used to generate melanomas in adult Tg(mitf:BRAFV600E); p53−/−; mitfa−/− fish byelectroporation of a MiniCoopR plasmid, ubb:Cas9, and zfU6:sgRNA against rb1. In the future,TEAZ will be a powerful way to serially model oncogene and tumor suppressor mutations andmay be used to study metastasis.Melanoma metastasis has previously been studied in the zebrafish using fluorescent zebrafishmelanoma cell lines transplanted into irradiated adult or embryonic casper recipients. These trans-parent fish allow for high-resolution imaging of fluorescent metastatic cells and quantification ofmetastasis (Heilmann et al. 2015). This technique was used by White and colleagues, who foundthat adipocytes in the melanoma microenvironment increase melanoma lipid content, cell growth,and invasion (Zhang et al. 2018). Due to the difficulty distinguishing metastasis from de novo tu-mor formation or tumor spreading, the ability to specifically model metastasis in the zebrafish willbe an important area of study in coming years.LEUKEMIAThe pathways regulating hematopoiesis are conserved between mammals and fish, leading to thedevelopment of many zebrafish leukemia models (Robertson et al. 2016). Here we review modelsof hematopoietic malignancies published to date.Acute Lymphoblastic LeukemiaAcute lymphoblastic leukemia (ALL) comprises leukemias of lymphoid origin and is the mostcommon childhood cancer. T cell ALL (T-ALL) is characterized by uncontrolled proliferationof T cell progenitors and accounts for 25% of adult and 15% of pediatric ALL (Durinck et al.2015). Although most ALL cases are B cell in origin, the vast majority of ALL models in zebrafishrecapitulate T-ALL. The first zebrafish T-ALL model was discovered in the Look lab (Langenauet al. 2003). This model expressed mouse c-Myc under the lymphoid-specific rag2 promoter. Ap-proximately 5% of F0 fish developed tumors about 44 days postfertilization. These leukemiaswere shown to arise from clonal expansion of transformed T lymphocyte precursors and weretransplantable. Langenau and colleagues later went on to create a stable rag2-EGFP-mMyc ze-brafish line that developed T-ALL but could only be propagated by in vitro fertilization (Fenget al. 2007, Langenau et al. 2005). They later created a Cre/Lox-inducible version that led tothe formation of T-ALL (Gutierrez et al. 2011a). Using this model, they found that the PI3K-AKT pathway can cause T-ALL in the absence of MYC induction. Recently, rag2:Myc zebrafishwere shown to develop B cell ALL (B-ALL) in a fraction of fish (Borga et al. 2019, Garcia et al.2018).B-ALL is characterized by uncontrolled proliferation of immature B cell precursors and is themost common childhood leukemia. Despite the fact that B-ALL comprises 75% of human ALL17.6McConnell • Noonan • Zon, .•·�-Review in Advance first posted on December 2, 2020. (Changes may still occur before final publication.) Access provided by 158.46.185.149 on 12/23/20. For personal use only. CA05CH17_ZonARjats.clsNovember 23, 202017:20cases, it is challenging to model B-ALL in zebrafish because the rag2 promoter has a T cell bias.Sabaawy et al. (2006) generated transgenic zebrafish that express TEL-AML1 (ETV6-RUNX1),a fusion common in B-ALL. Ubiquitous TEL-AML1 expression led to progenitor cell expansionand transplantable B-ALL in 3% of fish.Since the development of these models, several studies have been conducted to investigatethe mechanism of leukemia formation. Rudner et al. (2011) compared 17 T-ALLs from 4 ze-brafish T-ALL models to human T-ALL and found that they were comparable at a genomic level.T-ALL clones in a Myc-induced T-ALL fish model continuously evolve to drive leukemia pro-gression, in particular via Akt pathway activation (Blackburn et al. 2014). In Myc-induced T-ALLfish, proapoptotic Bim is downregulated by Myc and Akt in treatment-resistant T-ALL (Reynoldset al. 2014). Bcl2 was found to accelerate T-lymphoblastic lymphoma, but inhibited progression toT-ALL in Myc-induced fish models due to an inability to intravasate the vasculature (Feng et al.2010). In addition, Myc-induced zebrafish T-ALL was not affected by tp53 mutations (Gutierrezet al. 2014a). This is consistent with zebrafish having no functional ortholog for CDKN2A (ARF),which links Myc-induced stress to tp53 tumor suppression. As a result, the Myc-induced zebrafishT-ALL model is ideal for studying the role of CDKN2A-independent pathways in T-ALL.Over 60% of T-ALL patients have activating mutations in NOTCH1. A zebrafish model ofNOTCH1-induced T cell leukemia has been developed (rag2-ICN1-EGFP) (Chen et al. 2007).Notch signaling led to an expansion of T cell clones, but they were not fully transformed anddid not initiate leukemia when transplanted. Zebrafish expressing both Notch and Myc havemore aggressive T-ALL, and TOX was identified as an oncogene that synergized with MYC andNOTCH1 to accelerate T-ALL by expanding transformed clones and increasing genomic insta-bility (Blackburn et al. 2012, Lobbardi et al. 2017). Synergy between the Notch pathway and bcl2was also observed in these fish (Chen et al. 2007).Several groups have used T-ALL zebrafish models to conduct screens for leukemia drugs andgenetic modifiers. One group used T cell reporter fish to conduct a small-molecule screen forcompounds that act against immature T cells, leading to the identification of Lenaldekar (Ridgeset al. 2012). This group then used the Myc-induced T-ALL model to determine that Lenaldekarinduces long-term remission in these fish. Similarly, perphenazine was identified to be toxic toMyc-overexpressing thymocytes in the fish and was shown to induce apoptosis of T-ALL in Myc-induced fish model (Gutierrez et al. 2014b). Most recently the Look laboratory identified a classof small-molecule activators of protein phosphatase 2A that are effective against T-ALL cells fromTg(rag2:Myc; rag2:EGFP) zebrafish (Morita et al. 2020).Chronic Myeloid LeukemiaRecently, a heat shock–inducible BCR/ABL1 zebrafish model was reported that develops chronicmyeloid leukemia in 76% of adult fish aged 6 months to 1 year (Xu et al. 2020). Chronic myeloidleukemia in this fish line was similar to that in humans and these fish were used in a small-drugscreen.Acute Myeloid LeukemiaAcute myeloid leukemia (AML) is a clonal hematopoietic cancer of the bone marrow (exten-sively reviewed in Ferrara & Schiffer 2013). The first zebrafish model of a human AML fusionwas created by fusing histone acetyl-transferases MYST3 and NCOA2 under the control of theearly myeloid specific promoter spi1 (Zhuravleva et al. 2008). One percent of adult fish wenton to develop AML characterized by invasion of myeloid blast cells in the kidney. A variety ofwww.annualreviews.org • Reeling in the Zebrafish Cancer Models17.7, .•·�-Review in Advance first posted on December 2, 2020. (Changes may still occur before final publication.) Access provided by 158.46.185.149 on 12/23/20. For personal use only. CA05CH17_ZonARjats.clsNovember 23, 202017:20zebrafish models utilize spi1-driven oncogene expression. Zebrafish expressing spi1:FLT3-ITD-2A-EGFP/CG2 present myeloid hyperplasia in kidney marrow at 6 months, followed by leukemiaphase at 9 months postfertilization (Lu et al. 2016). Double spi1:FLT3-ITD-2A-EGFP/CG2 andspi1:NPM1-Mut-PA/CG2 transgenic fish developed leukemia by 6 months, indicating synergy be-tween the two mutations. Several fusion oncogenes have been reported in AML (Forrester et al.2011,Lutterbach & Hiebert 2000).A cre/lox-inducible Tg(spi1::loxP-EGFP-loxP::NUP98-HOXA9)fusion was generated in zebrafish, resulting in expansion of the myeloid lineage and suppressionof the erythroid lineage (Forrester et al. 2011). In all, 23% of these adult fish developed myelo-proliferation, similar to human myeloproliferative neoplasm, and epigenetic therapies restorednormal hematopoiesis (Deveau et al. 2015). RUNX1-CBF2T1 (AML1-ETO) zebrafish embryosexhibit disrupted hematopoiesis, abnormal circulation, internal hemorrhages, and cellular dyspla-sia, consistent with phenotypes seen in Runx1-Cbf2t1 mice, therefore validating the zebrafish as aleukemia model (Kalev-Zylinska et al.2002).A heat shock–inducible zebrafish AML1-ETO modelwas generated as well, and these embryos exhibited reprogramed hematopoietic cell fate and cyto-logical and transcriptional similarities to human AML (Yeh et al. 2008). As timing is critical whenmodeling AML, several other inducible models have been generated.Amplification of MYCN is a frequent event with poor prognosis in human AML.Tg(MYCN:HSE:EGFP) zebrafish were created by introducing murine Mycn upstream ofheat shock–driven Egfp (Shen et al. 2013). Overexpression of Mycn led to increased myeloidand precursor cells, while decreasing erythrocytes. Mycn reprograms hematopoietic cell fateby upregulating expression of hematopoietic transcription factors scl, lmo2, pu.1, and mpo anddownregulating gata1.In total, 25–50% of myeloid leukemias have mutations activating RAS family members. TheZon lab has developed a heat shock–inducible Cre/Lox-mediated approach to activating humanKRAS (Le et al. 2007). Tg(ß-actin-LoxP-EGFP-LoxP-KRASG12D; hsp70-Cre) fish developed fourdifferent tumor types: rhabdomyosarcoma, intestinal hyperplasia, malignant peripheral neuralsheath tumor (MPNST), and myeloproliferative disorder. Isolated kidney marrow cells were heatshocked ex vivo and transplanted into irradiated recipient fish, generating a zebrafish model ofmyeloproliferative disorder. Like KRAS, HRAS is frequently mutated in myeloproliferative disor-ders. Zebrafish with HRAS expression driven by the endothelial cell promoter fli1 have abnormalhematopoiesis including increased myelo-erythroid progenitor cells in the caudal hematopoietictissue, which requires Notch signaling repression (Alghisi et al. 2013).In addition to Notch, other signaling pathways are involved in AML and have been used tomodel this disease in zebrafish. Constitutively active stat5.1 zebrafish mutants present increasednumbers of early and late myeloid cells, erythrocytes, and B cells (Lewis et al. 2006). In addi-tion, CREB is frequently overexpressed in AML. CREB overexpression in the myeloid lineageresults in 79% of adult zebrafish with a block in myeloid differentiation mediated by C/EBPδ(Tregnago et al. 2016). These fish developed AML with clonal mature monocytic blasts, loss ofmyeloid precursors, and a transcriptional signature reminiscent of human AML. Zebrafish with anirf8 mutation developed a myeloproliferative neoplasm characterized by the expansion of myeloidprecursors (Zhao et al. 2018). Similarly, zebrafish idh1 knockdown blocked myeloid differentia-tion and overexpression of mutant human IDH1-R132H resulted in myeloid expansion (Shi et al.2015). Mutations in genes encoding other epigenetic regulators have also been explored in thesedisease models. In total, 50% of asxl1+/− fish also develop myeloproliferation by 5 months, andthe concomitant homozygous deletion of tet2 leads to AML in some fish (Gjini et al. 2019). Theaforementioned zebrafish models recapitulate some, but not all, aspects of human AML, withmostly myeloproliferative and embryonic phenotypes. There will be a need in the coming yearsto simultaneously and tissue specifically manipulate several driver genes.17.8McConnell • Noonan • Zon, .•·�-Review in Advance first posted on December 2, 2020. (Changes may still occur before final publication.) Access provided by 158.46.185.149 on 12/23/20. For personal use only. CA05CH17_ZonARjats.clsNovember 23, 202017:20SARCOMARhabdomyosarcomaRhabdomyosarcoma is a rare cancer that forms in soft tissue, particularly skeletal muscle, in chil-dren and young adults. Zebrafish models of sarcomas, including rhabdomyosarcoma, chordoma,hemangiosarcoma, and liposarcoma, have been extensively reviewed, so here we focus on commontransgenic models in zebrafish (Hayes and Langenau 2017). The majority of embryonal rhab-domyosarcomas have ectopic activation of the RAS pathway; this was replicated in zebrafish in2007 by overexpressing oncogenic KRASG12D under the control of the rag2 promoter (Langenauet al. 2007). These fish develop tumors at 10 days postfertilization and transcriptionally and histo-logically resemble the human disease. Loss of tp53 in this model system increased tumor incidence,as well as invasion and metastasis (Ignatius et al. 2018, Langenau et al. 2007). Imaging of develop-ing tumors using fluorescent reporters gives zebrafish a distinct advantage over other species and,in the embryonal rhabdomyosarcoma model, enabled the visualization of tumor and blood vesselformation (Ignatius et al. 2012).In addition to the rag2 promoter, KRASG12D was also driven by the early muscle progenitorpromoter cdh15 and the differentiated myoblast promoter mylz2 (Storer et al. 2013). Interestingly,RAS activation in progenitor cells has been shown to lead to decreased survival and the formationof less differentiated tumors. To create an inducible model system, Le et al. (2007) inserted a floxedstop cassette preceding KRASG12D that could be removed by a heat shock–inducible Cre. This re-sulted in rhabdomyosarcoma formation but also yielded intestinal hyperplasia, myeloproliferativedisorder, and MPNSTs. Recently, another rhabdomyosarcoma model was developed using thePAX3/7-FOXO1 fusion oncogene, which was placed under the control of various promoters, in-cluding ß-actin, ubi, CMV, mitfa, and fli1, which resulted in the formation of rhabdomyosarcoma,undifferentiated sarcoma, and primitive neuroectodermal tumors (Kendall et al. 2018).Inhibitors of MEK and mTOR signaling, as well as activation of the Wnt signaling path-way, decreased tumor formation in rag2:KRASG12D zebrafish (Chen et al. 2014, Le et al. 2013).Since the development of these models, the Langenau lab has performed several elegant studiesthat highlight the importance of developmental pathways in rhabdomyosarcoma formation. Theyfound that the muscle-specific transcription factors MYOD and MYF5 promote tumor growthby binding to muscle development and cell cycle genes (Tenente et al. 2017). They also foundthat NOTCH1 activation increases tumorigenesis via SNAIL1 activation and MEF2C suppression(Ignatius et al. 2017). Furthermore, they discovered that the noncanonical Wnt pathway mem-ber VANGL2 is active in highly proliferative tumor propagating rhabdomyosarcoma cells (Hayeset al. 2018). These studies underscore the prominent role that developmental signaling pathwaysplay in tumorigenesis.Ewing’s SarcomaEwing’s sarcoma is a rare childhood cancer of the bone that occurs through an EWS gene fu-sion. When one of the most common fusions, EWS-FLI1, is injected under the control of eithera ubiquitous promoter or a heat shock–inducible promoter alongside p53 loss, a small subset offish form tumors that histologically and molecularly resemble Ewing’s sarcoma (Leacock et al.2012). The EWS-FLI1-fusion, p53-mutant zebrafish also developed MPNSTs and leukemia-likecancer, likely due to the lack of promoter specificity. Ewing sarcoma and MPNST tumor induc-tion was also found in ewsa−/−; p53M214K/M214K zebrafish containing no fusion protein (Park et al.2016). Interestingly, these fish formed tumors at a much higher rate than the EWS-FLI1-fusion,p53-mutant zebrafish (∼60–70% versus ∼25%, respectively). Ewing’s sarcoma is the second mostcommon type of bone cancer behind osteosarcoma. To our knowledge, no transgenic osteosar-coma zebrafish models exist, making this a promising area of research in the future.www.annualreviews.org • Reeling in the Zebrafish Cancer Models17.9, .•·�-Review in Advance first posted on December 2, 2020. (Changes may still occur before final publication.) Access provided by 158.46.185.149 on 12/23/20. For personal use only. CA05CH17_ZonARjats.clsNovember 23, 202017:20LiposarcomaLiposarcoma is a cancer of the adipose tissue and is one of the most common types of sarcoma. Tomodel liposarcoma in the fish, constitutively active Akt2 was expressed under the control of thezebrafish rag2 promoter in the context of p53 loss (Gutierrez et al. 2011b). Well differentiated li-posarcomas formed in 30% of the p53-homozygous-mutant fish at 1–3 months of age. They alsofound that about one-third of human liposarcomas have aberrant AKT signaling. Interestingly,overexpressing constitutively active Akt2 in the skin using the krt4 promoter induces lipoma for-mation via hyperplastic growth of adipocytes in adult fish (Chu et al. 2012). These studies pointto a major role for AKT signaling in adipocyte and liposarcoma development.NEUROLOGICAL CANCERSGliomaGliomas are brain tumors that are thought to arise from neuroglial stem or progenitor cells(Molinaro et al. 2019). The first transgenic zebrafish model of glioma was generated in the Looklab, who showed that the p53e7/e7; nf1a+/−; nf1b−/− mutation results in high-grade glioma and MP-NST formation beginning at around 30 weeks of age (Shin et al. 2012). In 2013, two additionaltransgenic glioma models were generated by expressing either dominant active (DA) AKT1 orRAC1 under a central nervous system promoter, ptf1 ( Jung et al. 2013). Both models result ingliomas with heterogenous grades, with DA-RAC1 tumors inducing more aggressive high-gradetumors. Given that the AKT-PI3K pathway is hyperactive in the majority of glioblastoma tu-mors, this clinically relevant model could be used to identify novel therapeutic targets. Anothermodel was developed by driving KRASG12V under the krt5 or gfap promoter ( Ju et al. 2015). Thekrt5:KRASG12V mutation induced brain tumors at low frequency, but the majority of tumors thatformed were MPNST-like malignancies. However, when driven under the glial cell–specific gfappromoter, KRASG12V brain tumors in both ventricular zones and parenchyma were formed at ahigh frequency ( Ju et al. 2015).Additional transgenic zebrafish glioma models were developed by the Mione lab, who drove avariety of MAPK pathway members including Xmrk (oncogenic EGFR), EGFRvIII, KRASG12V,BRAFV600E, AKT1, and HRASG12V under the zic4 promoter, which is specific to the proliferatingzones of the central nervous system (CNS), resulting in a wide range of benign and malignantbrain tumors (Mayrhofer et al.2017).The most aggressive brain tumors were formed when drivingHRASG12V alongside DA-YAP, which was not seen when driving HRASG12V alone. These studiesemphasize the role for developmental signaling pathways in cancer formation and provide a newmodel for aggressive human mesenchymal glioblastoma subtypes.In 2016, a model for primitive neuroectodermal tumors of the CNS (a glioblastoma vari-ant) was developed by driving oncogenic NRAS in sox10 oligoneural precursor cells alongsidep53M214K/M214K (Modzelewska et al. 2016). The wide variety of oncogenic or tumor suppressordrivers used in these transgenic zebrafish glioma models is representative of the heterogeneityfound in human brain tumors. Although these models are excellent systems to study glioma biol-ogy, more work needs to be done to elucidate the mechanisms of brain tumor initiation, progres-sion, and metastasis.NeuroblastomaNeuroblastoma is a tumor of the peripheral sympathetic nervous system that arises from neuralcrest progenitor cells in children. The development of transgenic zebrafish neuroblastoma mod-els has been spearheaded by the Look laboratory. Use of the zebrafish dopamine-δ-hydroxylase17.10McConnell • Noonan • Zon, .•·�-Review in Advance first posted on December 2, 2020. (Changes may still occur before final publication.) Access provided by 158.46.185.149 on 12/23/20. For personal use only. CA05CH17_ZonARjats.clsNovember 23, 202017:20(dβh) promoter to drive human MYCN fused to EGFP specifically in the sympathetic nervoussystem results in the formation of tumor masses that histopathologically, immunohistochemically,and ultrastructurally mimic human neuroblastoma (Zhu et al. 2012). They also showed that over-expression of both MYCN and ALKF1174L in dβh-expressing cells rapidly induces neuroblastomaformation starting at 5–7 weeks of age. The Look lab also discovered that MYCN overexpres-sion in combination with the loss of zebrafish nf1 results in neuroblastoma formation with highpenetrance at 4 weeks of age (He et al. 2016). Interestingly, neuroblastoma formation was depen-dent on loss of the GAP-related domain of NF1 and could be abrogated using MEK inhibitorsand retinoids. Induction of the MAPK pathway by overexpressing mutant ptpn11 or Gab2 in-duces neuroblastoma with a high penetrance (Zhang et al. 2017). As with the nf1-mutant modeldiscussed above, treatment of MYCN; Gab2wt (wild-type Gab2) transgenic fish with the MEKinhibitor trametinib inhibits neuroblastoma formation. A fourth gene that acts in concert withMYCN to induce neuroblastoma is LMO1 (Zhu et al. 2017). The most striking thing about thismodel is that it induces widespread metastasis, which is rarely seen in other transgenic zebrafishmodels of cancer.In 2017, the Look lab uncoupled the EGFP and MYCN fusion and injected each separatelyunder the control of the dβh promoter, which resulted in higher levels of MYCN and increasedpenetrance (Tao et al. 2017). Using this model, they studied the role for DEF, a protein involved inpreribosomal RNA processing and found that DEF overexpression significantly accelerates neu-roblastoma onset, while DEF loss-of-function mutations prevent neuroblastoma formation. Thisstudy highlights the importance of transcription in tumor development and identifies a potentialpathway that could be therapeutically targeted. In addition to MYCN amplification, about 11%of neuroblastomas have high levels of c-MYC without amplification of the gene itself. Expressionof dβh:MYC results in rapid tumor formation at 4–8 weeks with almost complete penetrance inzebrafish (Zimmerman et al. 2018). Further analysis using whole-genome sequencing of humanneuroblastomas revealed that MYC overexpression is caused by amplification of enhancers distalto MYC or translocations that place highly active enhancers physically close to the MYC codingsequence. These findings highlight the importance of understanding gene regulation and its rolein cancer development.Malignant Peripheral Nerve Sheath TumorsMPNSTs are a rare type of cancer that forms in the connective tissue surrounding periph-eral nerves. A zebrafish model of MPNSTs was first developed in 2004 when 11 zebrafish linescontaining mutations in ribosomal protein genes were found to have elevated tumor incidence(Amsterdam et al. 2004). In total, 80% of these tumors were MPNSTs, which were highly aneu-ploid, with amplifications and deletions in oncogenes and tumor suppressors including cyclinD2,cdk6,and fgf6a (Zhang et al.2010).In searching for copy number alterations in rp- and tp53-mutantzebrafish MPNSTs, 34 genes were found to be amplified in both zebrafish and human MPNSTs(Zhang et al. 2013). Although this MPNST model is slow to develop (11–20 months), it is rep-resentative of the heterogeneity seen in patient tumors and can be used to study natural diseaseprogression.Since then, several transgenic MPNST models have been generated in zebrafish.Brca2Q658X/Q658X; tp53M214K/+ zebrafish have an average MPNST onset age of 15 months (Shiveet al. 2014). In addition to the histologic similarity to human MPNSTs, these tumors highly ex-press S100 and CD57,indicating that they are from a neural crest origin,as in humans (White et al.2017). Another MPNST model, Tg(mitfa:atg5K130R); p53M214K/+, was created when a dominant-negative autophagy protein was expressed under the control of the mitfa promoter in the contextwww.annualreviews.org • Reeling in the Zebrafish Cancer Models17.11, .•·�-Review in Advance first posted on December 2, 2020. (Changes may still occur before final publication.) Access provided by 158.46.185.149 on 12/23/20. For personal use only. CA05CH17_ZonARjats.clsNovember 23, 202017:20of p53 loss (Lee et al. 2016). Interestingly, the mutation status of p53 altered the tumor type,with fish heterozygous for p53 forming predominantly MPNSTs, whereas the p53-homozygousfish developed a variety of tumors, including neuroendocrine and small cell. This study also con-firmed the neural crest origin of MPNSTs, suggesting that Schwann cell precursors are the cellof origin of MPNSTs.A more targeted model of MPNSTs was generated when PDGFRA was driven under the sox10neural crest–specific promoter in nf1a+/–; nf1b–/–; p53m/m zebrafish, which formed MPNSTs ataround 80% penetrance by 30 weeks of age (Ki et al. 2017). Recently the Look lab inactivated theSWI/SNF chromatin remodeling factor ATRX using CRISPR/Cas9 to generate p53−/−; nf1b−/−;nf1a+/−; atrx+/− zebrafish (Oppel et al. 2019). They found that ATRX loss does not promotepathogenesis and instead widens the number of tumor types formed, as compared to 100% MP-NSTs formed in the control p53−/−; nf1b−/−; nf1a+/− zebrafish. The zebrafish MPNST modelshighlighted here are ideal platforms for identifying novel therapies for this rare cancer.GASTROINTESTINAL CANCERSPancreatic CancerThe two main types of pancreatic cancer are pancreatic adenocarcinoma, which arises in exocrinecells and comprises almost 95% of cases, and pancreatic neuroendocrine tumors, which arise inthe endocrine cells of the pancreas. The first zebrafish model of pancreatic cancer was developedin 2004 when human MYCN was driven under control of the zebrafish myoD promoter, resultingin the formation of pancreatic neuroendocrine tumors (Hong et al. 2004). Since then, the Leachlab has paved the way for models of pancreatic adenocarcinoma. In 2008 they discovered thatwhen oncogenic KRASG12D is expressed under the pancreas progenitor promoter ptf1a, it blocksnormal differentiation and causes an expansion in progenitor cells, leading to pancreatic cancerformation (Park et al. 2008).Interestingly, deletion of one copy of the ribosomal protein rpl36 in addition to KRASG12Dactivation resulted in accelerated tumor progression, with 100% of fish developing tumors by7.5 months (Provost et al. 2014). Leach and colleagues went on to comprehensively screen 12different KRAS mutations and found a significant increase in pancreatic tumor formation in fishinjected with one of the ptf1a-driven KRAS mutations found in human pancreatic cancer, as com-pared to those not associated with pancreatic cancer (Park et al. 2015). However, when expressingKRASG12D in elastase3l pancreatic cells, pancreatic neuroendocrine tumors are formed, highlight-ing the need to test additional promoters to better understand pancreatic cancer development (Oh& Park 2019).Liver CancerLiver cancer, particularly hepatocellular carcinoma, is the leading cause of cancer-related deathworldwide (ACS 2020). The role for zebrafish in hepatocellular carcinoma research has been ex-tensively reviewed by Goessling and colleagues; here, we summarize the major findings in thefield (Wrighton et al. 2019). The first transgenic zebrafish model of hepatocellular carcinoma wasdeveloped in 2008 by overexpressing the hepatitis C core protein in the context of hepatotoxintreatment (Rekha et al. 2008).A more targeted transgenic model was generated when KRASG12V was driven under the liver-specific fabp10 promoter and resulted in hepatocellular carcinoma formation penetrance between2 and 12 weeks, which was accelerated with p53 loss (Nguyen et al. 2011). Since then, the Gonglab has made several transgenic hepatocellular carcinoma models, including LexPR, Cre-LoxP17.12McConnell • Noonan • Zon, .•·�-Review in Advance first posted on December 2, 2020. (Changes may still occur before final publication.) Access provided by 158.46.185.149 on 12/23/20. For personal use only. CA05CH17_ZonARjats.clsNovember 23, 202017:20and Tet-inducible systems for KRASG12V, xmrk, and Myc that are molecularly similar to humanhepatocellular carcinoma (Li et al. 2012, 2013; Nguyen et al. 2012, 2016; Zheng et al. 2014). Instudying these models, the Gong lab discovered a major role for the immune system, particularlyneutrophils and macrophages, differential signaling pathway activation, and sex hormone–induceddifferences in tumor onset, as observed in humans (Li et al. 2019; Yan et al. 2015, 2017; Yang et al.2018; Zhao et al. 2016). The difference between sexes was independently confirmed in a DMBA-induced liver cancer model, which showed that treatment with 17β-estradiol accelerated tumorformation largely in the male zebrafish (Chaturantabut et al. 2019).An additional hepatocellular carcinoma model was developed by overexpressing activated β-catenin under the fabp10 promoter, leading to tumor formation at 78% penetrance by 6 months ofage (Evason et al. 2015). As in human hepatocellular carcinoma, these tumors had hyperactivationof the JNK pathway and inhibition of this pathway resulted in a decrease in tumor growth. An-other group found that by truncating the tumor suppressor adenomatous polyposis coli, or APC,zebrafish form spontaneous adenomas in the liver and small intestine and, when combining theAPC mutation with DMBA treatment, also induced pancreas and bile duct tumors (Haramis et al.2006).Another type of liver cancer, intrahepatic cholangiocarcinoma, has also been modeled in ze-brafish by overexpressing the oncoprotein Gankyrin in fabp10-expressing liver cells using a Tet-Off system (Huang et al. 2017). An additional intrahepatic cholangiocarcinoma model was gen-erated by using the fabp10 promoter to overexpress NRASQ61K, which resulted in 81.5% of fishdeveloping tumors at one-year postfertilization (Wang et al. 2017). Despite these advances inmodeling pancreatic and liver cancers, there are many gastrointestinal tumors that have yet to bemodeled using a transgenic zebrafish system. Additional combinations of gastrointestinal-specificreporters driving oncogenes with and without tumor suppressor loss is an area that warrants fur-ther investigation.THYROID CANCERThe endocrine and neuroendocrine system is made up of clusters of cells in several organs andglands, including the pituitary, thyroid, parathyroid, adrenal, and pineal glands. A common tumortype in this system is thyroid cancer. In 2017, the Houvras lab developed a model of papillary thy-roid cancer by overexpressing BRAFV600E under the zebrafish thyroglobulin promoter (Anelli et al.2017). These fish exhibited developmental abnormalities in their thyroid follicle structure, whichwas reversed by BRAF and MEK inhibitors, and formed thyroid cancer at 12 months postfertiliza-tion. In the future, researchers should explore the mechanisms driving thyroid cancer formation.GENITOURINARYRenal Cell CarcinomaThe zebrafish kidney represents a useful model of kidney development and disease (Drummond& Davidson 2010). The van Eeden lab generated two zebrafish lines with inactivated vhl, a verycommon mutation in clear cell renal cell carcinoma (Maher & Kaelin 1997; Santhakumar et al.2012; TCGA 2013; Van Rooijen et al. 2009, 2010). Erythropoiesis, angiogenesis, abnormal cardiaccontractility, and early lethality was ameliorated by HIF2α inhibition in these fish, and the kidneyof the vhl−/− larvae recapitulated clear cell histology, indicating that these fish may be used as amodel of early-stage renal cell carcinoma (Martins Metelo et al. 2015, Noonan et al. 2016). Gener-ating tissue-specific and inducible models and introducing secondary mutations will be importantin coming years.www.annualreviews.org • Reeling in the Zebrafish Cancer Models17.13, .•·�-Review in Advance first posted on December 2, 2020. (Changes may still occur before final publication.) Access provided by 158.46.185.149 on 12/23/20. For personal use only. CA05CH17_ZonARjats.clsNovember 23, 202017:20Germ Cell TumorsGerm cell tumors (GCTs) arise from primordial germ cells and occur in infants, children, andyoung adults in testis, ovary, or extragonadal sites (Sanchez & Amatruda 2016). Zebrafish germlinedevelopment is reminiscent of that of mouse and human (reviewed in Sanchez & Amatruda 2016).The Amatruda lab developed a zebrafish model of testicular GCT, characterized by undifferenti-ated germ cells, that is highly penetrant and heritable (Neumann et al. 2009). Fish develop testic-ular tumors that are sensitive to radiation, by approximately 7 months postfertilization. These fishwere later found to have an inactivating mutation in alk6b, a bone morphogenetic protein receptor(Neumann et al. 2011, Sanchez et al. 2019). Other groups have shown that expression of Simianvirus 40 large T-antigen, slc, the ciliary gene lrcc50, and ns1402 generates testicular GCTs (Bastenet al. 2013, Gill et al. 2010, Litchfield et al. 2016, Shimizu & Matsuda 2019). While progress hasbeen made to identify testicular GCT zebrafish models, zebrafish models of ovarian GCTs haveyet to be described.CONCLUSIONThe zebrafish models outlined here have enhanced our understanding of the mechanisms under-lying the initiation, progression, and metastasis of human cancer. Additionally, high-throughputdrug screens in zebrafish have identified promising new therapies for cancer and other diseases.Despite these instrumental advances, there is still much more to learn using this powerful modelsystem. The future of zebrafish cancer modeling is moving towards more tissue-specific and in-ducible models using CRISPR/Cas9, Cre/Lox, and Gal4/UAS (upstream activation sequence)technologies. Zebrafish are also an ideal model in which to test serial oncogene activation or tu-mor suppressor loss using novel techniques like electroporation. The ability to rapidly generatetransgenic animals and the ease of imaging make the zebrafish an ideal model in which to study thetumor microenvironment, a largely underexplored area in this field. Studying interactions amongtumor cells, the immune system, vasculature, and supporting niche cells is likely to reveal novelinsights into tumorigenesis and could lead to the development of new therapeutics. Transgenicmodeling of cancer in the zebrafish is a powerful tool to investigate basic biology and identifydrugs that will stop cancer development.DISCLOSURE STATEMENTThe authors are not aware of any affiliations, memberships, funding, or financial holdings thatmight be perceived as affecting the objectivity of this review.LITERATURE CITEDAblain J, Durand EM, Yang S, Zhou Y, Zon LI. 2015. A CRISPR/Cas9 vector system for tissue-specific genedisruption in zebrafish. Dev. Cell 32(6):756–64Ablain J, Xu M, Rothschild H, Jordan RC, Mito JK, et al. 2018. Human tumor genomics and zebrafish mod-eling identify SPRED1 loss as a driver of mucosal melanoma. Science 362(6418):1055–60Ablain J, Zon LI. 2016. Tissue-specific gene targeting using CRISPR/Cas9. Methods Cell Biol. 135:189–202ACS (Am. Cancer Soc.). 2020. Cancer facts and figures 2020. Tech. Rep., Am. Cancer Soc., Atlanta, GAAlghisi E, Distel M, Malagola M, Anelli V, Santoriello C, et al. 2013. Targeting oncogene expression to en-dothelial cells induces proliferation of the myelo-erythroid lineage by repressing the notch pathway.Leukemia 27:2229–41Amsterdam A, Sadler KC, Lai K, Farrington S, Bronson RT, et al. 2004. Many ribosomal protein genes arecancer genes in zebrafish. PLOS Biol. 2(5):e13917.14McConnell • Noonan • Zon, .•·�-Review in Advance first posted on December 2, 2020. (Changes may still occur before final publication.) Access provided by 158.46.185.149 on 12/23/20. For personal use only. CA05CH17_ZonARjats.clsNovember 23, 202017:20Anelli V, Villefranc JA, Chhangawala S, Martinez-Mcfaline R, Riva E, et al. 2017. Oncogenic BRAF disruptsthyroid morphogenesis and function via twist expression. eLife 6:e20728Basten SG, Davis EE, Gillis AJM, van Rooijen E, Stoop H, et al. 2013. Mutations in LRRC50 predisposezebrafish and humans to seminomas. PLOS Genet. 9(4):e1003384Blackburn JS, Liu S, Raiser DM, Martinez SA, Feng H, et al. 2012. Notch signaling expands a pre-malignantpool of T-cell acute lymphoblastic leukemia clones without affecting leukemia-propagating cell fre-quency. Leukemia 26:2069–78Blackburn JS, Liu S, Wilder JL, Dobrinski KP, Lobbardi R, et al. 2014. Clonal evolution enhances leukemia-propagating cell frequency in T cell acute lymphoblastic leukemia through Akt/mTORC1 pathway ac-tivation. Cancer Cell 25:366–78Borga C, Park G, Foster C, Burroughs-Garcia J, Marchesin M, et al. 2019. Simultaneous B and T cell acutelymphoblastic leukemias in zebrafish driven by transgenic MYC: implications for oncogenesis and lym-phopoiesis. Leukemia 33:333–47Callahan SJ, Tepan S, Zhang YM, Lindsay H, Burger A, et al. 2018. Cancer modeling by transgene electropo-ration in adult zebrafish (TEAZ). Dis. Model. Mech. 11(9):dmm034561Ceol CJ, Houvras Y, Jane-Valbuena J, Bilodeau S, Orlando DA, et al. 2011. The histone methyltransferaseSETDB1 is recurrently amplified in melanoma and accelerates its onset. Nature 471(7339):513–18Chaturantabut S, Shwartz A, Evason KJ, Cox AG, Labella K, et al. 2019. Estrogen activation of G-protein-coupled estrogen receptor 1 regulates phosphoinositide 3-kinase and mTOR signaling to promote livergrowth in zebrafish and proliferation of human hepatocytes. Gastroenterology 156(6):1788–804.e13Chen EY, DeRan MT, Ignatius MS, Brooke Grandinetti K, Clagg R, et al. 2014. Glycogen synthase kinase3 inhibitors induce the canonical WNT/β-catenin pathway to suppress growth and self-renewal in em-bryonal rhabdomyosarcoma. PNAS 111(14):5349–54Chen J, Jette C, Kanki JP, Aster JC, Look AT, Griffin JD. 2007. NOTCH1-induced T-cell leukemia in trans-genic zebrafish. Leukemia 21:462–71Chu CY, Chen CF, Rajendran RS, Shen CN, Chen TH, et al. 2012. Overexpression of Akt1 enhances adipo-genesis and leads to lipoma formation in zebrafish. PLOS ONE 7(5):e36474Ciarlo C, Kaufman CK, Kinikoglu B, Michael J, Yang S, et al. 2017. A chemical screen in zebrafish embryoniccells establishes that Akt activation is required for neural crest development. eLife 6:e29145Dang M, Henderson RE, Garraway LA, Zon LI. 2016. Long-term drug administration in the adult zebrafishusing oral gavage for cancer preclinical studies. Dis. Model. Mech. 9(7):811–20Deveau AP, Forrester AM, Coombs AJ, Wagner GS, Grabher C, et al. 2015. Epigenetic therapy restores nor-mal hematopoiesis in a zebrafish model of NUP98–HOXA9-induced myeloid disease. Leukemia 29:2086–97Dovey M, White RM, Zon LI. 2009. Oncogenic NRAS cooperates with p53 loss to generate melanoma inzebrafish. Zebrafish 6(4):397–404Drummond IA, Davidson AJ. 2010. Zebrafish kidney development. In Methods in Cell Biology, Vol. 100, ed.HW Detrich 3rd, M Westerfield, LI Zon, pp. 233–60. Burlington, MA: AcademicDurinck K, Goossens S, Peirs S, Wallaert A, Van Loocke W, et al. 2015. Novel biological insights in T-cellacute lymphoblastic leukemia. Exp. Hematol. 43:625–39Evason KJ, Francisco MT, Juric V, Balakrishnan S, del Pilar Lopez Pazmino M, et al. 2015. Identification ofchemical inhibitors of β-catenin-driven liver tumorigenesis in zebrafish. PLOS Genet. 11(7):1–32Feng H, Langenau DM, Madge JA, Quinkertz A, Gutierrez A, et al. 2007. Heat-shock induction of T-celllymphoma/leukaemia in conditional Cre/lox-regulated transgenic zebrafish. Br. J. Haematol. 138(2):169–75Feng H, Stachura DL, White RM, Gutierrez A, Zhang L, et al. 2010. T-lymphoblastic lymphoma cells expresshigh levels of BCL2, S1P1, and ICAM1, leading to a blockade of tumor cell intravasation. Cancer Cell18:353–66Ferrara F, Schiffer CA. 2013. Acute myeloid leukaemia in adults. Lancet 381(9865):484–95Forrester AM, Grabher C, Mcbride ER, Boyd ER, Vigerstad MH, et al. 2011. NUP98-HOXA9-transgeniczebrafish develop a myeloproliferative neoplasm and provide new insight into mechanisms of myeloidleukaemogenesis. Br. J. Haematol. 155(2):167–81www.annualreviews.org • Reeling in the Zebrafish Cancer Models17.15, .•·�-Review in Advance first posted on December 2, 2020. (Changes may still occur before final publication.) Access provided by 158.46.185.149 on 12/23/20. For personal use only. CA05CH17_ZonARjats.clsNovember 23, 202017:20Garcia EG, Iyer S, Garcia SP, Loontiens S, Sadreyev RI, et al. 2018. Cell of origin dictates aggression andstem cell number in acute lymphoblastic leukemia. Leukemia 32:1860–65Gill JA, Lowe L, Nguyen J, Liu PP, Blake T, et al. 2010. Enforced expression of simian virus 40 large T-antigenleads to testicular germ cell tumors in zebrafish. Zebrafish 7(4):333–41Gjini E, Chang-Bin J, Nguyen AT, Reyon D, Gans E, et al. 2019. Disruption of asxl1 results in myeloprolif-erative neoplasms in zebrafish. Dis. Model. Mech. 12(5):dmm035790Gutierrez A, Feng H, Stevenson K, Neuberg DS, Calzada O, et al. 2014a. Loss of function tp53 mutations donot accelerate the onset of myc-induced T-ALL in the zebrafish. Br. J. Haematol. 166(1):84–90Gutierrez A, Grebliunaite R, Feng H, Kozakewich E, Zhu S, et al. 2011a. Pten mediates Myc oncogene depen-dence in a conditional zebrafish model of T cell acute lymphoblastic leukemia. J. Exp. Med. 208(8):1595–603Gutierrez A, Look AT, Aster JC. 2014b. Phenothiazines induce PP2A-mediated apoptosis in T cell acutelymphoblastic leukemia. J. Clin. Investig. 124(2):644Gutierrez A, Snyder EL, Marino-Enriquez A, Zhang YX, Sioletic S, et al. 2011b. Aberrant AKT activationdrives well-differentiated liposarcoma. PNAS 108(39):16386–91Haramis APG, Hurlstone A, van der Velden Y, Begthel H, van den Born M, et al. 2006. Adenomatous polyposiscoli-deficient zebrafish are susceptible to digestive tract neoplasia. EMBO Rep. 7(4):444–49Hayes MN, Langenau DM. 2017. Discovering novel oncogenic pathways and new therapies using zebrafishmodels of sarcoma. In Methods in Cell Biology, Vol. 138, ed. HW Detrich 3rd, M Westerfield, LI Zon,pp. 525–61. Burlington, MA: AcademicHayes MN, McCarthy K, Jin A, Oliveira ML, Iyer S, et al. 2018. Vangl2/RhoA signaling pathway regulatesstem cell self-renewal programs and growth in rhabdomyosarcoma. Cell Stem Cell 22(3):414–27.e6He S, Mansour MR, Zimmerman MW, Ki DH, Layden HM, et al. 2016. Synergy between loss of NF1 andoverexpression of MYCN in neuroblastoma is mediated by the GAP-related domain. eLife 5:e14713Heilmann S, Ratnakumar K, Langdon E, Kansler E, Kim I, et al. 2015. A quantitative system for studyingmetastasis using transparent zebrafish. Cancer Res. 75(20):4272–82Hong WY, Kutok JL, Nam HL, Hui YP, Fletcher CDM, et al. 2004. Targeted expression of human MYCNselectively causes pancreatic neuroendocrine tumors in transgenic zebrafish. Cancer Res. 64(20):7256–62Howe K, Clark MD, Torroja CF, Torrance J, Berthelot C, et al. 2013. The zebrafish reference genome se-quence and its relationship to the human genome. Nature 496:498–503Huang SJ, Cheng CL, Chen JR, Gong HY, Liu W, Wu JL. 2017. Inducible liver-specific overexpression ofgankyrin in zebrafish results in spontaneous intrahepatic cholangiocarcinoma and hepatocellular carci-noma formation. Biochem. Biophys. Res. Commun. 490(3):1052–58Ignatius MS, Chen E, Elpek NM, Fuller AZ, Tenente IM, et al. 2012. In vivo imaging of tumor-propagatingcells, regional tumor heterogeneity, and dynamic cell movements in embryonal rhabdomyosarcoma.Cancer Cell 21(5):680–93Ignatius MS, Hayes MN, Lobbardi R, Chen EY, McCarthy KM, et al. 2017. The NOTCH1/SNAIL1/MEF2Cpathway regulates growth and self-renewal in embryonal rhabdomyosarcoma. Cell Rep. 19(11):2304–18Ignatius MS, Hayes MN, Moore FE, Tang Q, Garcia SP, et al. 2018. tp53 deficiency causes a wide tumorspectrum and increases embryonal rhabdomyosarcoma metastasis in zebrafish. eLife 7:e37202Ju B, Chen W, Orr BA, Spitsbergen JM, Jia S, et al. 2015. Oncogenic KRAS promotes malignant brain tumorsin zebrafish. Mol. Cancer 14:18Jung IH, Leem GL, Jung DE, Kim MH, Kim EY, et al. 2013. Glioma is formed by active Akt1 alone andpromoted by active Rac1 in transgenic zebrafish. Neuro-Oncology 15(3):290–304Kalev-Zylinska ML, Horsfield JA, Flores MVC, Postlethwait JH, Vitas MR, et al. 2002. Runx1 is requiredfor zebrafish blood and vessel development and expression of a human RUNX1-CBF2T1 transgeneadvances a model for studies of leukemogenesis. Development 129:2015–30Kaufman CK, Mosimann C, Fan ZP, Yang S, Thomas A, et al. 2016. A zebrafish melanoma model revealsemergence of neural crest identity during melanoma initiation. Science 351(6272):aad2197Kendall GC, Watson S, Xu L, Lavigne CA, Murchison W, et al. 2018. PAX3-FOXO1 transgenic zebrafishmodels identify HES3 as a mediator of rhabdomyosarcoma tumorigenesis. eLife 7:e3380017.16McConnell • Noonan • Zon, .•·�-Review in Advance first posted on December 2, 2020. (Changes may still occur before final publication.) Access provided by 158.46.185.149 on 12/23/20. For personal use only. CA05CH17_ZonARjats.clsNovember 23, 202017:20Ki DH, He S, Rodig S, Look AT. 2017. Overexpression of PDGFRA cooperates with loss of NF1 and p53 toaccelerate the molecular pathogenesis of malignant peripheral nerve sheath tumors.Oncogene 36(8):1058–68Kinkel MD, Eames SC, Philipson LH, Prince VE. 2010. Intraperitoneal injection into adult zebrafish. J. Vis.Exp. (42):2126Langenau DM, Traver D, Ferrando AA, Kutok JL. 2003. Myc-induced T cell leukemia in transgenic zebrafish.Science 299(5608):887–90Langenau DM, Feng H, Berghmans S, Kanki JP, Kutok JL, Look AT. 2005. Cre/lox-regulated transgeniczebrafish model with conditional myc-induced T cell acute lymphoblastic leukemia.PNAS 102(17):6068–73Langenau DM, Keefe MD, Storer NY, Guyon JR, Kutok JL, et al. 2007. Effects of RAS on the genesis ofembryonal rhabdomyosarcoma. Genes Dev. 21(11):1382–95Le X, Langenau DM, Keefe MD, Kutok JL, Neuberg DS, Zon LI. 2007. Heat shock-inducible Cre/Lox ap-proaches to induce diverse types of tumors and hyperplasia in transgenic zebrafish. PNAS 104(22):9410–15Le X, Pugach EK, Hettmer S, Storer NY, Liu J, et al. 2013. A novel chemical screening strategy in ze-brafish identifies common pathways in embryogenesis and rhabdomyosarcoma development. Develop-ment 140(11):2354–64Leacock SW, Basse AN, Chandler GL, Kirk AM, Rakheja D, Amatruda JF. 2012. A zebrafish transgenic modelof Ewing’s sarcoma reveals conserved mediators of EWS-FLI1 tumorigenesis. Dis. Model. Mech. 5:95–106Lee E, Wei Y, Zou Z, Tucker K, Rakheja D, et al. 2016. Genetic inhibition of autophagy promotes p53 loss-of-heterozygosity and tumorigenesis. Oncotarget 7(42):67919–33Lewis RS, Stephenson SEM, Ward AC. 2006. Constitutive activation of zebrafish Stat5 expands hematopoieticcell populations in vivo. Exp. Hematol. 34:179–87Li H, Lu JW, Huo X, Li Y, Li Z, Gong Z. 2019. Effects of sex hormones on liver tumor progression andregression in Myc/xmrk double oncogene transgenic zebrafish. Gen. Comp. Endocrinol. 277(March):112–21Li Z, Huang X, Zhan H, Zeng Z, Li C, et al. 2012. Inducible and repressable oncogene-addicted hepatocellularcarcinoma in Tet-on xmrk transgenic zebrafish. J. Hepatol. 56(2):419–25Li Z, Zheng W, Wang Z, Zeng Z, Zhan H, et al. 2013. A transgenic zebrafish liver tumor model with in-ducible Myc expression reveals conserved Myc signatures with mammalian liver tumors. Dis. Model. Mech.6(2):414–23Lister JA,Robertson PC,Lepage T,Johnson SL,Raible DW.1999.nacre encodes a zebrafish microphthalmia-related protein that regulates neural-crest-derived pigment cell fate. Development 126:3757–67Litchfield K, Levy M, Dudakia D, Proszek P, Shipley C, et al. 2016. Rare disruptive mutations in ciliaryfunction genes contribute to testicular cancer susceptibility. Nat. Commun. 7:13840Lo J, Fisher DE. 2014. The melanoma revolution: from UV carcinogenesis to a new era in therapeutics. Science346(6212):945–49Lobbardi R, Pinder J, Martinez-Pastor B, Theodorou M, Blackburn JS, et al. 2017. TOX regulates growth,DNA repair, and genomic instability in T-cell acute lymphoblastic leukemia. Cancer Discov. 7(11):1336–53Lu J, Hou H, Hsieh M, Tien H, Lin L. 2016. Overexpression of FLT3-ITD driven by spi-1 results in expandedmyelopoiesis with leukemic phenotype in zebrafish. Leukemia 30:2098–101Lutterbach B, Hiebert SW. 2000. Role of the transcription factor AML-1 in acute leukemia and hematopoieticdifferentiation. Blood 106(5):1519–24Maher ER, Kaelin WG. 1997. von Hippel-Lindau Disease. Medicine 76:381–91Martins Metelo A, Noonan HR, Li X, Jin Y, Baker R, et al. 2015. Pharmacological HIF2a inhibition improvesVHL disease-associated phenotypes in zebrafish model. J. Clin. Investig. 125(5):1987–97Mayrhofer M,Gourain V,Reischl M,Affaticati P,Jenett A,et al.2017.A novel brain tumour model in zebrafishreveals the role of YAP activation in MAPK- and PI3K-induced malignant growth. Dis. Model. Mech.10:15–28McConnell AM, Mito JK, Ablain J, Dang M, Formichella L, et al. 2019. Neural crest state activation in NRASdriven melanoma, but not in NRAS-driven melanocyte expansion. Dev. Biol. 449:107–13www.annualreviews.org • Reeling in the Zebrafish Cancer Models17.17, .•·�-Review in Advance first posted on December 2, 2020. (Changes may still occur before final publication.) Access provided by 158.46.185.149 on 12/23/20. For personal use only. CA05CH17_ZonARjats.clsNovember 23, 202017:20Modzelewska K, Boer EF, Mosbruger TL, Picard D, Anderson D, et al. 2016. MEK inhibitors reverse growthof embryonal brain tumors derived from oligoneural precursor cells. Cell Rep. 17(5):1255–64Molinaro AM, Taylor JW, Wiencke JK, Wrensch MR. 2019. Genetic and molecular epidemiology of adultdiffuse glioma. Nat. Rev. Neurol. 15(7):405–17Morita K,He S,Nowak RP,Wang J,Zimmerman MW,et al.2020.Allosteric activators of protein phosphatase2A display broad antitumor activity mediated by dephosphorylation of MYBL2. Cell 181:702–15.e20Mort RL, Jackson IJ, Patton EE. 2015. The melanocyte lineage in development and disease. Development142(4):620–32Neumann JC, Chandler GL, Damoulis VA, Fustinoa NJ, Lillard K, et al. 2011. Mutation in the type IB bonemorphogenetic protein receptor alk6b impairs germ-cell differentiation and causes germ-cell tumors inzebrafish. PNAS 108(32):13153–58Neumann JC, Dovey JS, Chandler GL, Carbajal L, Amatruda JF. 2009. Identification of a heritable model oftesticular germ cell tumor in the zebrafish. Zebrafish 6(4):319–27Nguyen AT, Emelyanov A, Koh CHV, Spitsbergen JM, Lam SH, et al. 2011. A high level of liver-specificexpression of oncogenic KrasV12 drives robust liver tumorigenesis in transgenic zebrafish. Dis. Model.Mech. 4(6):801–13Nguyen AT, Emelyanov A, Koh CHV, Spitsbergen JM, Parinov S, Gong Z. 2012. An inducible krasV12 trans-genic zebrafish model for liver tumorigenesis and chemical drug screening. Dis. Model. Mech. 5:63–72Nguyen AT, Koh V, Spitsbergen JM, Gong Z. 2016. Development of a conditional liver tumor model bymifepristone-inducible Cre recombination to control oncogenic krasV12 expression in transgenic ze-brafish. Sci. Rep. 6:19559Noonan HR, Metelo AM, Kamei CN, Peterson RT, Drummond IA, Iliopoulos O. 2016. Loss of vhl in thezebrafish pronephros recapitulates early stages of human clear cell renal cell carcinoma. Dis. Model. Mech.9(8):873–84Oh S, Park JT. 2019. Zebrafish model of KRAS-initiated pancreatic endocrine tumor. Anim. Cells Syst.23(3):209–18Oppel F, Tao T, Shi H, Ross KN, Zimmerman MW, et al. 2019. Loss of atrx cooperates with p53-deficiencyto promote the development of sarcomas and other malignancies. PLOS Genet. 15(4):e1008039Park H, Galbraith R, Turner T, Mehojah J, Azuma M. 2016. Loss of Ewing sarcoma EWS allele promotestumorigenesis by inducing chromosomal instability in zebrafish. Sci. Rep. 6:32297Park JT, Johnson N, Liu S, Levesque M, Wang YJ, et al. 2015. Differential in vivo tumorigenicity of diverseKRAS mutations in vertebrate pancreas: a comprehensive survey. Oncogene 34(21):2801–6Park SW, Davison JM, Rhee J, Hruban RH, Maitra A, Leach SD. 2008. Oncogenic KRAS induces pro-genitor cell expansion and malignant transformation in zebrafish exocrine pancreas. Gastroenterology134(7):2080–90Patton EE, Widlund HR, Kutok JL, Kopani KR, Amatruda JF, et al. 2005. BRAF mutations are sufficient topromote nevi formation and cooperate with p53 in the genesis of melanoma. Curr. Biol. 15(3):249–54Provost E, Bailey JM, Aldrugh S, Liu S, Iacobuzio-Donahue C, Leach SD. 2014. The tumor suppressor rpl36restrains KRASG12V-induced pancreatic cancer. Zebrafish 11(6):551–59Pugach EK, Li P, White R, Zon L. 2009. Retro-orbital injection in adult zebrafish. J. Vis. Exp. 34:1645Rekha RD, Amali AA, Her GM, Yeh YH, Gong HY, et al. 2008. Thioacetamide accelerates steatohepatitis,cirrhosis and HCC by expressing HCV core protein in transgenic zebrafish Danio rerio. Toxicology 243(1–2):11–22Reynolds C, Roderick JE, Labelle JL, Bird G, Mathieu R, et al. 2014. Repression of BIM mediates survivalsignaling by MYC and AKT in high-risk T-cell acute lymphoblastic leukemia. Leukemia 28:1819–27Ridges S, Heaton WL, Joshi D, Choi H, Eiring A, et al. 2012. Zebrafish screen identifies novel compoundwith selective toxicity against leukemia. Blood 119(24):5621–31Robertson AL, Avagyan S, Gansner JM, Zon LI. 2016. Understanding the regulation of vertebratehematopoiesis and blood disorders: big lessons from a small fish. FEBS Lett. 590(22):4016–33Rudner L, Brown K, Dobrinski K, Bradley D, Garcia M, et al. 2011. Shared acquired genomic changes inzebrafish and human T-ALL. Oncogene 30:4289–96Sabaawy HE, Azuma M, Embree LJ, Tsai H-J, Starost MF, Hickstein DD. 2006. TEL-AML1 transgeniczebrafish model of precursor B cell acute lymphoblastic leukemia. PNAS 103(41):15166–7117.18McConnell • Noonan • Zon, .•·�-Review in Advance first posted on December 2, 2020. (Changes may still occur before final publication.) Access provided by 158.46.185.149 on 12/23/20. For personal use only. CA05CH17_ZonARjats.clsNovember 23, 202017:20Sanchez A, Amatruda JF. 2016. Zebrafish germ cell tumors. Adv. Exp. Med. Biol. 916:479–94Sanchez A, Xu L, Pierce JL, Lafin JT, Abe D, et al. 2019. Identification of testicular cancer driver genes by across-species comparative oncology approach. Andrology 7:545–54Santhakumar K, Judson EC, Elks PM, Mckee S, Elworthy S, et al. 2012. A zebrafish model to study andtherapeutically manipulate hypoxia signaling in tumorigenesis. Cancer Res. 72(16):4017–27Santoriello C, Gennaro E, Anelli V, Distel M, Kelly A, et al. 2010. Kita driven expression of oncogenic HRASleads to early onset and highly penetrant melanoma in zebrafish. PLOS ONE 5(12):e15170Santoriello C, Sporrij A, Yang S, Flynn RA, Henriques T, et al. 2020. RNA helicase DDX21 mediates nu-cleotide stress responses in neural crest and melanoma cells. Nat. Cell Biol. 22:372–79Shen L-J, Chen F-Y, Zhang Y, Cao L-F, Kuang Y, et al. 2013. MYCN transgenic zebrafish model with thecharacterization of acute myeloid leukemia and altered hematopoiesis. PLOS ONE 8(3):e59070Shi X, He B-L, Ma ACH, Guo Y, Chi Y, et al. 2015. Functions of idh1 and its mutation in the regulation ofdevelopmental hematopoiesis in zebrafish. Blood 125(19):2974–84Shimizu N, Matsuda M. 2019. Identification of a novel zebrafish mutant line that develops testicular germ celltumors. Zebrafish 16(1):15–28Shin J, Padmanabhan A, de Groh ED, Lee J-S, Haidar S, et al. 2012. Zebrafish neurofibromatosis type 1 geneshave redundant functions in tumorigenesis and embryonic development. Dis. Model. Mech. 5(6):881–94Shive HR,West RR,Embree LJ,Golden CD,Hickstein DD.2014.brca2 and tp53 collaborate in tumorigenesisin zebrafish. PLOS ONE 9(1):e87177Storer NY, White RM, Uong A, Price E, Nielsen GP, et al. 2013. Zebrafish rhabdomyosarcoma reflects thedevelopmental stage of oncogene expression during myogenesis. Development 140(14):3040–50Tao T, Sondalle SB, Shi H, Zhu S, Perez-Atayde AR, et al. 2017. The pre-rRNA processing factor DEF is ratelimiting for the pathogenesis of MYCN-driven neuroblastoma. Oncogene 36(27):3852–67TCGA (Cancer Genome Atlas Res. Netw.). 2013. Comprehensive molecular characterization of clear cellrenal cell carcinoma. Nature 499:43–49TCGA (Cancer Genome Atlas Res. Netw.). 2015. Genomic classification of cutaneous melanoma. Cell161(7):1681–96Tenente IM, Hayes MN, Ignatius MS, McCarthy K, Yohe M, et al. 2017. Myogenic regulatory transcriptionfactors regulate growth in rhabdomyosarcoma. eLife 6:e19214Tregnago C, Manara E, Zampini M, Bisio V, Borga C, et al. 2016. CREB engages C/EBPδ to initiate leuke-mogenesis. Leukemia 30:1887–96Van Rooijen E, Voest EE, Logister I, Bussmann J, Korving J, et al. 2010. von Hippel-Lindau tumor sup-pressor mutants faithfully model pathological hypoxia-driven angiogenesis and vascular retinopathies inzebrafish. Dis. Model. Mech. 3:343–53Van Rooijen E, Voest EE, Logister I, Korving J, Schwerte T, et al. 2009. Zebrafish mutants in the vonHippel-Lindau tumor suppressor display a hypoxic response and recapitulate key aspects of Chuvashpolycythemia. Blood 113(25):6449–60Venkatesan AM, Vyas R, Gramann AK, Dresser K, Gujja S, et al. 2018. Ligand-activated BMP signaling in-hibits cell differentiation and death to promote melanoma. J. Clin. Investig. 128(1):294–308Wang J, Leng X, Wang G, Wan X, Cao H. 2017. The construction of intrahepatic cholangiocarcinoma modelin zebrafish. Sci. Rep. 7:13419White LA, Sexton JM, Shive HR. 2017. Histologic and immunohistochemical analyses of soft tissue sarcomasfrom brca2-mutant/tp53-mutant zebrafish are consistent with neural crest (Schwann cell) origin. Vet.Pathol. 54(2):320–27White R, Rose K, Zon L. 2013. Zebrafish cancer: the state of the art and the path forward. Nat. Rev. Cancer13(9):624–36White RM, Cech J, Ratanasirintrawoot S, Lin CY, Rahl PB, et al. 2011. DHODH modulates transcriptionalelongation in the neural crest and melanoma. Nature 471(7339):518–22Wrighton PJ, Oderberg IM, Goessling W. 2019. There is something fishy about liver cancer: zebrafish modelsof hepatocellular carcinoma. Cell. Mol. Gastroenterol. Hepatol. 8(3):347–63Xu M, Ye Y, Ye Z, Xu S, Liu W, et al. 2020. Human BCR/ABL1 induces chronic myeloid leukemia-like diseasein zebrafish. Haematologica 105(3):674–86www.annualreviews.org • Reeling in the Zebrafish Cancer Models17.19, .•·�-Review in Advance first posted on December 2, 2020. (Changes may still occur before final publication.) Access provided by 158.46.185.149 on 12/23/20. For personal use only. CA05CH17_ZonARjats.clsNovember 23, 202017:20Yan C, Huo X, Wang S, Feng Y, Gong Z. 2015. Stimulation of hepatocarcinogenesis by neutrophils uponinduction of oncogenic kras expression in transgenic zebrafish. J. Hepatol. 63(2):420–28Yan C, Yang Q, Huo X, Li H, Zhou L, Gong Z. 2017. Chemical inhibition reveals differential requirementsof signaling pathways in krasV12- and Myc-induced liver tumors in transgenic zebrafish. Sci. Rep. 7:45796Yang Q, Yan C, Gong Z. 2018. Interaction of hepatic stellate cells with neutrophils and macrophages in theliver following oncogenic kras activation in transgenic zebrafish. Sci. Rep. 8:8495Yeh J-RJ, Munson KM, Chao YL, Peterson QP, MacRae CA, Peterson RT. 2008. AML1-ETO reprogramshematopoietic cell fate by downregulating scl expression. Development 135:401–10Zhang GJ, Hoersch S, Amsterdam A, Whittaker CA, Beert E, et al. 2013. Comparative oncogenomic analysisof copy number alterations in human and zebrafish tumors enables cancer driver discovery. PLOS Genet.9(8):e1003734Zhang GJ, Hoersch S, Amsterdam A, Whittaker CA, Lees JA, Hopkins N. 2010. Highly aneuploid zebrafishmalignant peripheral nerve sheath tumors have genetic alterations similar to human cancers. PNAS107(39):16940–45Zhang M, Di Martino JS, Bowman RL, Campbell NR, Baksh SC, et al. 2018. Adipocyte-derived lipids mediatemelanoma progression via FATP proteins. Cancer Discov. 8(8):1006–25Zhang X, Dong Z, Zhang C, Ung CY, He S, et al. 2017. Critical role for GAB2 in neuroblastoma pathogenesisthrough the promotion of SHP2/MYCN cooperation. Cell Rep. 18(12):2932–42Zhao F, Shi Y, Huang Y, Zhan Y, Zhou L, et al. 2018. Irf8 regulates the progression of myeloproliferativeneoplasm-like syndrome via Mertk signaling in zebrafish. Leukemia 32:149–58Zhao Y, Huang X, Ding TW, Gong Z. 2016. Enhanced angiogenesis, hypoxia and neutrophil recruitmentduring Myc-induced liver tumorigenesis in zebrafish. Sci. Rep. 6:31952Zheng W, Li Z, Nguyen AT, Li C, Emelyanov A, Gong Z. 2014. Xmrk, Kras and Myc transgenic zebrafishliver cancer models share molecular signatures with subsets of human hepatocellular carcinoma. PLOSONE 9(3):e91179Zhu S, Lee JS, Guo F, Shin J, Perez-Atayde AR, et al. 2012. Activated ALK collaborates with MYCN inneuroblastoma pathogenesis. Cancer Cell 21(3):362–73Zhu S, Zhang X, Weichert-Leahey N, Dong Z, Zhang C, et al. 2017. LMO1 synergizes with MYCN topromote neuroblastoma initiation and metastasis. Cancer Cell 32(3):310–23.e5Zhuravleva J, Paggetti J, Martin L, Hammann A, Solary E, et al. 2008. MOZ/TIF2-induced acute myeloidleukaemia in transgenic fish. Br. J. Haematol. 143(3):378–82Zimmerman MW,Liu Y,He S,Durbin AD,Abraham BJ,et al.2018.MYC drives a subset of high-risk pediatricneuroblastomas and is activated through mechanisms including enhancer hijacking and focal enhanceramplification. Cancer Discov. 8(3):320–3517.20McConnell • Noonan • Zon, .•·�-Review in Advance first posted on December 2, 2020. (Changes may still occur before final publication.) Access provided by 158.46.185.149 on 12/23/20. For personal use only. | RESEARCH GATE |
CORRIGENDUMEffect of dietary probiotics and prebiotics on the performance ofbroiler chickensH. Al-Khalaifa, A. Al-Nasser, T. Al-Surayee, S. Al-Kandari, N. Al-Enzi, T. Al-Sharrah, G. Ragheb,S. Al-Qalaf, and A. Mohammed2020 Poultry Science -:1https://doi.org/10.1016/j.psj.2020.12.0542019 Poultry Science 98:4465–4479 https://doi.org/10.3382/ps/pez282.The following acknowledgment has been added tothis article: “The authors would like to extend theirsincere thanks to the Kuwait Institute for ScientificResearch (KISR) and Kuwait Foundation for theAdvancement of Sciences (KFAS) for their technicaland financial support of the current study.” The articlehas been updated online. The authors regret the orig-inal omission.� 2020 Published by Elsevier Inc. on behalf of Poultry ScienceAssociation Inc.1 | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/355187121Computational Finance Course: Lecture 10 / 14Presentation · October 2021CITATIONS0READS891 author:Some of the authors of this publication are also working on these related projects:Novel Methods in Computational Finance View projectHybrid models View projectLech GrzelakUtrecht University80 PUBLICATIONS 732 CITATIONS SEE PROFILEAll content following this page was uploaded by Lech Grzelak on 12 October 2021.The user has requested enhancement of the downloaded file.Materials for the courseThe course is based on book “Mathematical Modeling and Computationin Finance: With Exercises and Python and MATLAB Computer Codes”,by C.W. Oosterlee and L.A. Grzelak, World Scientific Publishing EuropeLtd, 2019. For more details go here.▶ YouTube Channel with courses can be found here.▶ Slides and the codes can be found here.Lech A. GrzelakComputational Finance1 / 17List of contentOption Pricing with Monte CarloSimulation of the CIR ProcessExact Simulation of the CIR ModelAlmost Exact Simulation of the Heston ModelThe Heston Model and Simulation in PythonLech A. GrzelakComputational Finance2 / 17Option Pricing with Monte CarloOption Pricing with Monte Carlo▶ We take S(t0) = 5, σ = 0.3, r = 0.06, T = 1, the number of MonteCarlo time steps m = 1000 (so that ∆t = 1/1000) and compare theobtained call option values with the analytic solutions from theBlack-Scholes formula, for the strike price K = S(t0), i.e.Vc(t0, S(t0)) = 0.7359.▶ Table below shows that both methods require a similar number ofpaths to achieve satisfactory pricing results. This is an indicationthat the Euler and Milstein schemes exhibit the same order of weakconvergence.Table: Call option prices in dependence of the number of Monte Carlo paths forthe Euler and Milstein discretizations.Type:N100100050001000050000100000BSEurop.Euler0.77090.74440.72830.74980.73280.73560.7359callMilstein0.76900.74380.72830.74970.73270.7356DigitalEuler2.82532.40622.46652.44112.45022.44692.4483callMilstein2.82532.40622.46742.44062.45152.4462Lech A. GrzelakComputational Finance3 / 17Simulation of the CIR ProcessChallenges with Standard Discretization Schemes▶ A typical example of a process with probability mass around zero isthe CIR process, which was discussed to model the variance for theHeston stochastic volatility model with the following dynamics,dv(t) = κ(¯v − v(t))dt + γ�v(t)dW (t),v(t0) > 0.(1)▶ It is well-known that if the Feller condition, 2κ¯v > γ2, is satisfied,the process v(t) cannot reach zero, and if this condition does nothold the origin is accessible and strongly reflecting.▶ In both cases, the v(t) process cannot become negative.Lech A. GrzelakComputational Finance4 / 17Simulation of the CIR Process543t020.1510v(t)10.1200.050CIR density0304050543t020.4v(t)0.3150.20.10CIR density01015Figure: Paths and the corresponding PDF for the CIR process (1) in the caseswhere the Feller condition is satisfied and is not satisfied. Simulations wereperformed with κ = 0.5, v0 = 0.1, ¯v = 0.1. Left: γ = 0.1; Right: γ = 0.35.Lech A. GrzelakComputational Finance5 / 17Simulation of the CIR ProcessProbability of Negative Realization▶ The nonnegativity problem becomes apparent when a standarddiscretization is employed. If we apply, for example, the Eulerdiscretization to the process in (1), i.e.,vi+1 = vi + κ(¯v − vi)∆t + γ√vi√∆tZ.and assume that vi > 0, we may calculate the probability that a nextrealization, vi+1, becomes negative, i.e. P[vi+1 < 0].P[vi+1 < 0|vi > 0]=P[vi + κ(¯v − vi)∆t + γ�vi∆tZ < 0|vi > 0]=P[γ�vi∆tZ < −vi − κ(¯v − vi)∆t|vi > 0],which equals,P[vi+1 < 0|vi > 0] = P�Z < −vi + κ(¯v − vi)∆tγ√vi∆t���vi > 0�> 0.▶ Since Z is a normally distributed random variable, it is unbounded.Therefore the probability of vi becoming negative, is positive underthe Euler discretization, implying P[vi+1 < 0|vi > 0] > 0.Lech A. GrzelakComputational Finance6 / 17Simulation of the CIR ProcessTruncated Euler Scheme▶ The truncated Euler scheme, as the name says, is based on thepreviously introduced Euler discretization scheme. In order toprevent the process to cross the axis, we need to deal with possiblenegative path realizations vi+1.▶ It can be summarized, as follows,� ˆvi+1 = vi + κ(¯v − vi)∆t + γ√vi∆tZ,vi+1 = max(ˆvi+1, 0).▶ Although the scheme certainly provides paths that are nonnegative,the accuracy of this scheme is parameter-dependent, meaning that,when the Feller condition is not satisfied and the density shouldaccumulate around zero, the adjusted paths, due to the truncation,may be highly biased. In essence, by the truncation a differentprocess than the original CIR process is represented numerically.▶ When truncation takes place for many Monte Carlo paths, theaccuracy may be limited as the discretization bias could be high.Lech A. GrzelakComputational Finance7 / 17Simulation of the CIR ProcessReflecting Euler Scheme▶ One of the important properties of the CIR process is that the originis attainable, however, it is not an absorbing boundary. This meansthat the Monte Carlo paths may reach the boundary v = 0, but theycannot stay at the boundary – they should immediately move awayfrom it. When using the truncated Euler scheme from above, usingvi+1 = max(ˆvi+1, 0), the paths that attain negative values areprojected at the origin.▶ Another possible modification for the simulation of the CIR process,which is particularly useful when the Feller condition is not satisfied,is to use the so-called reflecting principle, where the variance pathsare forced to move upwards again.▶ The reflecting scheme is given by the following adjustment of theEuler scheme,� ˆvi+1 = vi + κ(¯v − vi)∆t + γ√vi∆tZ,vi+1 = |ˆvi+1| .(2)Lech A. GrzelakComputational Finance8 / 17Simulation of the CIR ProcessComparison of the Reflecting and Truncated Euler Schemes▶ In Figure below, some paths that were generated by the reflectingand truncated Euler schemes are compared. The paths that aregenerated by both schemes are essentially the same, until they reachthe boundary at v(t) = 0. After hitting the origin, the reflectingscheme will have paths that are above those that are generated bythe truncation scheme. Notice that by the reflecting scheme wehowever do not improve the quality of the Euler scheme.00.20.40.60.81time00.010.020.030.040.050.060.070.080.090.1v(t)reflection vs. truncation schemetruncated schemereflecting scheme00.20.40.60.81time00.010.020.030.040.050.060.070.080.090.1v(t)reflection vs. truncation schemetruncated schemereflecting scheme▶ Let us take a look at Python code.Lech A. GrzelakComputational Finance9 / 17Exact Simulation of the CIR ModelExact Simulation of the CIR Model▶ A different approach for simulating paths from the CIR process takesinto account the fact that variance process v(t), follows thenoncentral chi-squared distribution. Details of the noncentralchi-squared distribution have been given already in previous lecturein the context of the Heston stochastic volatility model.▶ Conditional on state v(s), s < t, the distribution at time t is givenby,v(t)|v(s) ∼ ¯c(t, s) · χ2(δ, ¯κ(t, s)),(3)with¯c(t, s) = γ24κ�1 − e−κ(t−s)�,δ = 4κ¯vγ2 ,¯κ(t, s) =4κe−κ(t−s)γ2(1 − e−κ(t−s))v(s).Lech A. GrzelakComputational Finance10 / 17Exact Simulation of the CIR ModelExact Simulation of the CIR Model▶ Given formulation may form the basis for an exact simulation schemefor the path realizations of the CIR process, as, for i = 0, . . . , m − 1,¯c(ti+1, ti)=γ24κ�1 − e−κ(ti+1−ti)�,¯κ(ti+1, ti)=4κe−κ(ti+1−ti)γ2(1 − e−κ(ti+1−ti)) vi .vi+1=¯c(ti+1, ti)χ2(δ, ¯κ(ti+1, ti)),with a constant parameter δ = 4κ¯v/γ2, and some initial valuev(t0) = v0.▶ Let us revisit Python code.Lech A. GrzelakComputational Finance11 / 17Almost Exact Simulation of the Heston ModelThe Heston Model▶ The Heston model is governed by the following dynamics:� dS(t) = rS(t)dt + S(t)�v(t)dWx(t),dv(t) = κ(¯v − v(t))dt + γ�v(t)dWv(t),(4)▶ Which for X(t) := log S(t), becomes:�dX(t) =�r − 12v(t)�dt +�v(t)dWx(t),dv(t) = κ(¯v − v(t))dt + γ�v(t)dWv(t),(5)with correlation dWx(t)dWv(t) = ρx,vdt,▶ which in terms in independent Brownian motions is given by:�dX(t) =�r − 12v(t)�dt +�v(t)�ρx,vd�Wv(t) +�1 − ρ2x,vd�Wx(t)�,dv(t) = κ (¯v − v(t)) dt + γ�v(t)d �Wv(t).Lech A. GrzelakComputational Finance12 / 17Almost Exact Simulation of the Heston ModelAlmost Exact Discretization of the Heston Model▶ After integration of both processes in a certain time interval[ti, ti+1], the following discretization scheme is obtained:xi+1=xi +� ti+1ti�r − 12v(t)�dt + ρx,v� ti+1ti�v(t)d�Wv(t)+�1 − ρ2x,v� ti+1ti�v(t)d�Wx(t),andvi+1=vi + κ� ti+1ti(¯v − v(t)) dt + γ� ti+1ti�v(t)d�Wv(t) .▶ Notice that the two integrals with �Wv(t) in the SDEs above are thesame, and in terms of the variance realizations they are given by:� ti+1ti�v(t)d�Wv(t) = 1γ�vi+1 − vi − κ� ti+1ti(¯v − v(t))dt�.(7)Lech A. GrzelakComputational Finance13 / 17Almost Exact Simulation of the Heston ModelAlmost Exact Discretization of the Heston Model▶ The variance vi+1 can then be simulated, for given value of vi, bymeans of the CIR process, or, equivalently, by either the noncentralchi-squared distribution or by the QE scheme (the techniquesdescribed in the book).▶ As a final step in the Heston model simulation, the discretization forxi+1 is given by:xi+1=xi +� ti+1ti�r − 12v(t)�dt+ρx,vγ�vi+1 − vi − κ� ti+1ti(¯v − v(t)) dt�+�1 − ρ2x,v� ti+1ti�v(t)d�Wx(t).Lech A. GrzelakComputational Finance14 / 17Almost Exact Simulation of the Heston ModelAlmost Exact Discretization of the Heston Model▶ We approximate all integrals appearing in the expression above bytheir left integration boundary values of the integrand, as in theEuler discretization scheme:xi+1≈xi +� ti+1ti�r − 12vi�dt+ρx,vγ�vi+1 − vi − κ� ti+1ti(¯v − vi) dt�+�1 − ρ2x,v� ti+1ti√vid�Wx(t).▶ The calculation of the integrals is now trivial and results in:xi+1≈xi +�r − 12vi�∆t + ρx,vγ(vi+1 − vi − κ (¯v − vi) ∆t)+�1 − ρ2x,v√vi��Wx(ti+1) − �Wx(ti)�.(8)Lech A. GrzelakComputational Finance15 / 17Almost Exact Simulation of the Heston ModelAlmost Exact Discretization of the Heston Model▶ Finally for �Wx(ti+1) − �Wx(ti)d=√∆tZx, with Zx ∼ N(0, 1), wefind:� xi+1 ≈ xi + k0 + k1vi + k2vi+1 + √k3viZx,vi+1 = ¯c(ti+1, ti)χ2(δ, ¯κ(ti+1, ti)),▶ with the variance process simulated as follows:¯c(ti+1, ti)=γ24κ(1 − e−κ(ti+1−ti)), δ = 4κ¯vγ2 ,¯κ(ti+1, ti)=4κe−κ∆tγ2(1 − e−κ∆t)vi,and χ2(δ, ¯κ(·, ·)) the noncentral chi-squared distribution with δdegrees of freedom and noncentrality parameter ¯κ(ti+1, ti) andk0=�r − ρx,vγ κ¯v�∆t,k1 =�ρx,vκγ− 12�∆t − ρx,vγ ,k2=ρx,vγ ,k3 = (1 − ρ2x,v)∆t.Lech A. GrzelakComputational Finance16 / 17The Heston Model and Simulation in PythonEuropean Option Pricing with Almost Exact Simulation▶ In this simulation experiment, European call options are computedby means of the AES scheme, with S(t0) = 100, for three differentstrike prices: K = 100 (ATM), K = 70 (ITM) and K = 140 (OTM).In the experiment different time steps are used, varying from onetime step per year to 64 time steps per year.▶ The model parameters are chosen as:κ = 0.5, γ = 1, ¯v = 0.04, v0 = 0.04, r = 0.1, ρx,v = −0.9.▶ The numerical results are based on 500.000 Monte Carlo paths,They are presented in Table 17. The reference results have beengenerated by the COS method.K = 100K = 70K = 140∆tEulerAESEulerAESEulerAES10.94 (0.023)-1.00 (0.012)-0.82 (0.028)-0.53 (0.016)1.29 (0.008)0.008 (0.001)1/22.49 (0.022)-0.45 (0.011)-0.11 (0.030)-0.25 (0.016)1.03 (0.008)-0.0006 (0.001)1/42.40 (0.016)-0.18 (0.010)0.37 (0.027)-0.11 (0.016)0.53 (0.005)0.0005 (0.001)1/82.08 (0.016)-0.10 (0.010)0.43 (0.025)-0.07 (0.016)0.22 (0.003)0.0009 (0.001)1/161.77 (0.015)-0.03 (0.010)0.40 (0.023)-0.03 (0.016)0.08 (0.001)0.0002 (0.001)1/321.50 (0.014)-0.03 (0.009)0.34 (0.022)-0.01 (0.016)0.03 (0.001)-0.002 (0.001)1/641.26 (0.013)-0.001 (0.009)0.27 (0.021)-0.005 (0.016)0.02 (0.001)0.001 (0.001)Lech A. GrzelakComputational Finance17 / 17View publication stats | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/43656461High variability of the heterogeneous ice nucleation potential of oxalic aciddihydrate and sodium oxalateArticle in Atmospheric Chemistry and Physics · August 2010DOI: 10.5194/acp-10-7617-2010 · Source: DOAJCITATIONS27READS1335 authors, including:Some of the authors of this publication are also working on these related projects:Light scattering properties of atmospheric ice crystals View projectCERN CLOUDy Project View projectOttmar MöhlerKarlsruhe Institute of Technology380 PUBLICATIONS 14,233 CITATIONS SEE PROFILEMartin SchnaiterKarlsruhe Institute of Technology253 PUBLICATIONS 9,458 CITATIONS SEE PROFILEThomas LeisnerKarlsruhe Institute of Technology263 PUBLICATIONS 7,094 CITATIONS SEE PROFILEAll content following this page was uploaded by Ottmar Möhler on 21 May 2014.The user has requested enhancement of the downloaded file.ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|Atmos. Chem. Phys. Discuss., 10, 11513–11575, 2010www.atmos-chem-phys-discuss.net/10/11513/2010/doi:10.5194/acpd-10-11513-2010© Author(s) 2010. CC Attribution 3.0 License.AtmosphericChemistryand PhysicsDiscussionsThis discussion paper is/has been under review for the journal Atmospheric Chemistryand Physics (ACP). Please refer to the corresponding final paper in ACP if available.High variability of the heterogeneous icenucleation potential of oxalic aciddihydrate and sodium oxalateR. Wagner, O. M¨ohler, H. Saathoff, M. Schnaiter, and T. LeisnerKarlsruhe Institute of Technology, Institute for Meteorology and Climate Research, Karlsruhe,GermanyReceived: 21 April 2010 – Accepted: 23 April 2010 – Published: 28 April 2010Correspondence to: R. Wagner (robert.wagner2@kit.edu)Published by Copernicus Publications on behalf of the European Geosciences Union.11513ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|AbstractThe heterogeneous ice nucleation potential of airborne oxalic acid dihydrate andsodium oxalate particles in the deposition and condensation mode has been inves-tigated by controlled expansion cooling cycles in the AIDA aerosol and cloud chamberof the Karlsruhe Institute of Technology at temperatures between 244 and 228 K. Pre-5vious laboratory studies have highlighted the particular role of oxalic acid dihydrate asthe only species amongst a variety of other investigated dicarboxylic acids to be ca-pable of acting as a heterogeneous ice nucleus in both the deposition and immersionmode. We could confirm a high deposition mode ice activity for 0.03 to 0.8 µm sizedoxalic acid dihydrate particles that were either formed by nucleation from a gaseous10oxalic acid/air mixture or by rapid crystallisation of highly supersaturated aqueous ox-alic acid solution droplets. The critical saturation ratio with respect to ice required fordeposition nucleation was found to be less than 1.1 and the size-dependent ice-activefraction of the aerosol population was in the range from 0.1 to 22%. In contrast, oxalicacid dihydrate particles that had crystallised from less supersaturated solution droplets15and had been allowed to slowly grow in a supersaturated environment from still un-frozen oxalic acid solution droplets over a time period of several hours were found tobe much poorer heterogeneous ice nuclei. We speculate that under these conditionsa crystal surface structure with less-active sites for the initiation of ice nucleation wasgenerated. Such particles partially proved to be almost ice-inactive in both the depo-20sition and condensation mode. At times, the heterogeneous ice nucleation ability ofoxalic acid dihydrate significantly changed when the particles had been processed inpreceding cloud droplet activation steps. Such behaviour was also observed for thesecond investigated species, namely sodium oxalate. Our experiments address theatmospheric scenario that coating layers of oxalic acid or its salts may be formed by25physical and chemical processing on pre-existing particulates such as mineral dust andsoot. Given the broad diversity of the observed heterogeneous ice nucleability of theoxalate species, it is not straightforward to predict whether an oxalate coating layer will11514ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|improve or reduce the ice nucleation ability of the seed aerosol particles.1IntroductionSeveral recent studies have addressed the hygroscopic behaviour (deliquescence andefflorescence phase transitions) and ice nucleation abilities of low molecular weightdicarboxylic acids (Braban et al., 2003; Kanji et al., 2008; Mikhailov et al., 2009; Par-5sons et al., 2004; Peng et al., 2001; Prenni et al., 2001; Shilling et al., 2006; Treuel etal., 2008; Zobrist et al., 2006). These compounds have been identified as an impor-tant contribution to the water-soluble organic fraction of tropospheric aerosols, typicallycomprising 1–3% of the total particulate organic carbon in urban and semi-urban ar-eas and up to 10% in remote continental and marine environments (Carlton et al.,102007; Chebbi and Carlier, 1996; Crahan et al., 2004; Ervens et al., 2004; Hsieh et al.,2007; Kawamura et al., 1996; Kerminen et al., 2000; Kerminen et al., 1999; Limbecket al., 2005; Yao et al., 2002). Their presence in the atmosphere is ascribed to bothprimary emissions and various secondary formation pathways, amongst which photo-oxidations of gaseous organic precursors and subsequent gas-to-particle conversion,15heterogeneous reactions on particle surfaces, as well as aqueous in-cloud reactionmechanisms are currently discussed (Yang et al., 2009). In most cases, oxalic acid(C2) is found to be the most abundant species, followed by malonic (C3) and succinicacid (C4).In the following, we will give a brief literature survey on the hygroscopic properties of20oxalic acid, starting with the aerosol flow tube/FTIR study by Braban et al. (2003). Theauthors have generated a polydisperse aerosol of anhydrous oxalic acid particles byspray-drying of an aqueous solution at a relative humidity (RH) of less than 5%. As ap-parent from the recorded infrared spectrum, the chain-like β form of anhydrous oxalicacid has formed where the molecules are linked together in pairs by cyclic hydrogen25bonds (Bellamy and Pace, 1963; Villepin and Novak, 1978a, b). Upon increasing rel-ative humidity, two phase transitions were observed: from anhydrous β oxalic acid to11515ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|oxalic acid dihydrate at 9 – 21% RH (294–259 K) and the deliquescence transition to anaqueous solution at 97–99% RH (294–279 K). The structure of oxalic acid dihydrate issimilar to that of the anhydrous β form, with two water molecules included between thecarboxyl groups. Its infrared spectrum is characterised by a prominent doublet featureat around 3500 cm−1 due to the hydration water whereas aqueous oxalic acid solution5droplets reveal a broader and less-structured absorption band in the O-H stretchingregime. Note that two polymorphs of oxalic acid dihydrate have only been observedfor the completely deuterated form (Ebisuzaki and Angel, 1981). According to Brabanet al. (2003), it was difficult to effloresce oxalic acid aerosol particles. In a number ofexperiments the aerosol water was not completely removed by the drier, leaving be-10hind an aerosol that featured a continuous water uptake over the complete RH range.Such behaviour was also observed in two independent H-TDMA (hygroscopicity tan-dem differential mobility analyser) investigations (Mikhailov et al., 2009; Prenni et al.,2001).Prenni et al. (2001) have measured the hygroscopic growth curve of size selected,1550 and 100 nm diameter oxalic acid particles generated by spray-drying at <5% RH(303 K). As already indicated, a continuous water uptake between 45 and 91% RH wasdetected which could be explained by the presence of concentrated solution dropletsinstead of effloresced particles. An overall similar growth curve was recently measuredby Mikhailov et al. (2009) from <5 to 95% RH for a monodispersed 100 nm diameter20oxalic acid aerosol at ambient temperature, generated by nebulisation of an aqueoussolution, drying at RH<5%, and size selecting the 100 nm dry particle diameter by adifferential mobility analyser. A closer inspection of the growth curve, however, revealedsome trends and small-scale fluctuations that are inconsistent with the assumption thatliquid-like aerosol particles were present: (i) a decrease of the growth factor in the25range from 5 to 40% RH, (ii), a small, stepwise increase in the growth factor at about43% RH, going along with a sudden broadening of the mode width of the measuredsize distribution, and, (iii), small-scale fluctuations in the growth curve until reaching arelative humidity of 80%. As an explanation, the authors proposed that spray-drying of11516ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|the solution droplets had led to the formation of amorphous, highly porous oxalic acidparticles with a gel-like structure. The discontinuities in the growth curve could thenbe attributed to microstructural rearrangements and transitions between collapsed andswollen gel structures. Only above a relative humidity of 80%, the close agreementbetween the mode widths of the particle size distribution in repetitive hydration and5dehydration experiments points to the presence of fully deliquesced liquid aqueoussolution droplets.Complementing the contradictory results on the hygroscopic behaviour of oxalic acidparticles, Peng et al. (2001) have detected an efflorescence transition in 10 – 20 µmsized oxalic acid particles at 52–57% RH. The particles were levitated in an electrody-10namic balance at 298 K. No further change in the particle mass was detected whenlowering the relative humidity to 5%.The latter was chosen as a reference statefor water-free particles, implying that the efflorescence transition had directly led tothe crystallisation of anhydrous oxalic acid particles. As suggested by Mikhailov etal. (2009), the inconsistent findings could partly be explained by size effects when as-15suming that a part of the polydisperse aerosol in the Braban et al. (2003) study hadalso extended into the >1 µm size range and that this tail of the size distribution hadthen dominated the mass fraction and been responsible for the overall infrared spectralsignature. In this way, one could combine the results from Braban et al. (2003) andPeng et al. (2001) for super-micron sized particles where crystallisation of oxalic acid20is obviously facilitated whereas the findings from Prenni et al. (2001) and Mikhailovet al. (2009) for ≤100 nm particles suggest the predominant formation of (semi-)solidamorphous particles.Prenni et al. (2001) have probed size-selected 50 and 100 nm diameter oxalic acidparticles, generated in the same way as for their H-TDMA measurements, also in a con-25tinuous flow diffusion chamber (CFDC) to measure ice formation. As discussed above,these particles were supposedly either concentrated solution droplets, or, accordingto the interpretation from Mikhailov et al. (2009), gel-like amorphous particles. Thethreshold ice relative humidity to freeze 1% of those particles at temperatures between11517ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|about 228 and 217 K almost coincided with water saturation, making them less efficientice nuclei in comparison with supercooled inorganic aqueous solution droplets like sul-phuric acid and ammonium sulphate. Several further studies have addressed the icenucleation potential of pure dicarboxylic acids. Kanji et al. (2008) have investigatedthe ice formation via deposition nucleation on oxalic acid dihydrate particles deposited5on a Teflon-coated, copper cold stage at 233 K. The particles were prepared from adilute aqueous solution using an atomiser and, by inspection of the optical micrographshown in Fig. 3 of their publication, covered a size range between about 1 and 8 µmin diameter. Prior to an ice nucleation experiment, the deposited particles were driedat RH<1% (298 K) for 18 h, most probably leading to the formation of anhydrous oxalic10acid. According to the results from Braban et al. (2003), these anhydrous particlesshould have transformed into oxalic acid dihydrate prior to ice nucleation due to theelevated humidity levels encountered in the course of the ice nucleation experiments.Kanji et al. (2008) have determined threshold relative humidities for the nucleation ofa single ice particle on the ensemble of deposited particles as a function of the total15particulate surface area. The surface area was varied by adjusting the total number ofparticles on the stage, typically from a few hundred to several 105 particles (Kanji andAbbatt, 2006). The measured critical ice saturation ratios ranged from about 1.10–1.25for total particulate surface areas in the range of 2×10−1–6×10−4 cm2. The observedhigh ice nucleability in the deposition mode clearly contradicts the results from Prenni20et al. (2001), indicating that the particles probed in the CFDC chamber were indeed notcrystalline. Micrometer-sized, spray-deposited particles of maleic acid (C4 unsaturateddicarboxylic acid) also proved to be efficient heterogeneous ice nuclei in the depositionmode at low temperatures (Shilling et al., 2006). Critical ice saturation ratios for nu-cleating ice on the maleic acid particles ranged from about 1.05 at 220 K to 1.30 at25190 K. In contrast, Parsons et al. (2004) have reported that 2–40 µm sized particles ofsaturated C3–C6 dicarboxylic acids are not good ice nuclei in the deposition mode attemperatures above 243 K. Instead of heterogeneously nucleating ice at temperaturesbelow the eutectic temperature, i.e., at ice supersaturated conditions, these particles11518ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|underwent a deliquescence transition to aqueous solution droplets. Oxalic acid (C2)has not been investigated.Zobrist et al. (2006) specifically addressed the ability of several dicarboxylic acids asheterogeneous ice nuclei in the immersion mode, arguing that in most cases these sub-stances would be part of a multi-component mixture in atmospheric aerosol particles.5If precipitating from these mixtures, the organic crystals would therefore be includedin an aqueous solution with other inorganic and/or organic solutes and thus could onlyact as immersion ice nuclei. The freezing experiments were performed with emulsifiedbinary dicarboxylic acid/water solutions of different solute concentrations and selectedternary solutions with e.g. sulphuric acid, ammonium sulphate, and sodium chloride as10additional solute species. Ice freezing temperatures were measured with a differen-tial scanning calorimeter. The authors demonstrated that in several of the emulsifiedaqueous solutions (e.g. binary 5.0, 9.8, and 19.8 wt % oxalic acid, 7.3 wt % succinicacid, and 1.6 wt % adipic acid solutions) homogeneous ice nucleation had triggeredthe precipitation of the organic solute. Only in the case of oxalic acid, however, sup-15posedly precipitated as oxalic acid dihydrate, these embedded crystals then acted asimmersion mode ice nuclei in a second freezing cycle, as evidenced by a freezing tem-perature higher than the homogeneous freezing temperature of pure water. The studyby Zobrist et al. (2006) has motivated the heterogeneous ice nucleation experimentswith oxalic acid particles which we will present in this manuscript. Our first intention20was to reproduce the results from cooling cycles with emulsified solutions by controlledexpansion cooling experiments with airborne particles, conducted in the large coolableaerosol and cloud chamber AIDA of the Karlsruhe Institute of Technology. In the follow-ing two paragraphs we briefly describe what should be observed in respective seriesof AIDA experiments on the basis of the results of Zobrist et al. (2006).25The experimental trajectory in the temperature-composition plane for ice nucleationexperiments with airborne particles significantly differs from that with emulsified so-lutions (Fig. 1). In the latter case, the solute concentration remains constant duringcooling (trace C) and ice crystals formed by homogeneous nucleation in the first cool-11519ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|ing cycle have to be melted by heating the solutions to a conditioning temperatureabove the eutectic temperature. Then another cooling cycle is started to probe the het-erogeneous ice nucleation ability of the potentially precipitated organic crystals. Theexperimental trajectories of two conceivable AIDA expansion cooling cycles with bi-nary oxalic acid/water solution droplets are schematically shown as green and blue5traces in Fig. 1. Experiment A is started at an initial gas temperature of 244 K, i.e.,slightly above the homogeneous freezing temperature of pure water droplets at about236 K. The initial composition of the injected oxalic acid solution droplets adjusts to theprevailing relative humidity which is roughly equal to the saturation pressure over icebecause the inner walls of the AIDA chamber are coated with a thin ice layer in the10preparation of an expansion run. Thereby, the initial composition is given by point A1on the ice melting-point curve in aqueous oxalic acid solutions. As can be seen fromthe phase diagram, the aerosol consists of highly concentrated solution droplets withabout 62 wt % oxalic acid. As a comparison, the saturation concentration of oxalic acidat 244 K is about 0.7 wt %, based on a thermodynamic extrapolation of bulk solubilities15measured at higher temperatures (Apelblat and Manzurola, 1987; Omar and Ulrich,2006).To initiate the ice freezing, the AIDA pressure is reduced by controlled pumping.Upon expansion cooling, the aerosol strongly dilutes due to the increasing relative hu-midity (see Sect. 2 for the experimental details). After about 3 to 4 degrees of cooling20(240 to 241 K, A2), the relative humidity with respect to supercooled water has in-creased from 75% (corresponding to an ice saturation ratio of one at 244 K) to 100%,leading to the activation of the oxalic acid solution droplets to a cloud of almost puresupercooled water droplets. This cloud eventually glaciates at point A3 when the gastemperature has further dropped below the homogeneous freezing temperature of pure25water droplets. After completed glaciation, the pumping is stopped and the ice crystalsslowly evaporate on a timescale of typically 15 to 20 min while the gas temperaturegradually recuperates its original value. A second expansion cooling cycle with thesame aerosol loading is then started again at 244 K to check whether the droplet freez-11520ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|ing has also triggered the crystallisation of oxalic acid (dihydrate). The organic crystalscould act as deposition mode ice nuclei at temperatures between 244 and 240–241 K,or, once water saturation is reached, act as condensation freezing nuclei (a liquid layercondenses on the particles, which subsequently initiates the freezing) before reachinga temperature of 236 K provided that the crystals would not instantly and completely5deliquesce on the timescale of the expansion cooling experiment. When conductingsuch an experiment with multi-component oxalic acid-containing solution droplets, onlycondensation freezing could take place. Experiment B (green trace) is started with afresh aerosol loading at an initial gas temperature of 233 K (B1), i.e., below the homo-geneous freezing temperature of pure water droplets. The injected oxalic acid droplets10will homogeneously freeze in the first expansion cycle once the dilution has gone belowa critical threshold value (B2) but they will not dilute to almost pure water droplets asin experiment A. In the second expansion cycle, the potentially precipitated oxalic acid(dihydrate) crystals could act as ice nuclei in the deposition mode, or, when conduct-ing an experiment with a multi-component mixture, act as ice nuclei in the immersion15mode, thereby potentially inducing earlier ice formation compared to the homogeneousfreezing run.This envisaged experimental approach has successfully been applied to ternary so-lution droplets with sulphuric acid and sodium chloride as additional solute species. Wewill report these results in a forthcoming manuscript. Concerning freezing experiments20with binary oxalic acid/water solution droplets, we never observed that the injectedaerosol particles remained in a supercooled liquid state at temperatures ≤244 K. In-stead, as proven by infrared extinction measurements, solid oxalic acid dihydrate crys-tallised from the highly supersaturated solution droplets. These particles revealed aremarkably high variability in their ice nucleation potential, depending on the way they25were generated. When directly injecting aqueous oxalic acid solution droplets into thecooled AIDA chamber at e.g. 244 K, the crystallisation of the entire particle ensembleto oxalic acid dihydrate took place on a time scale of about one hour. In another typeof experiment, the droplet aerosol was added to the chamber at a higher temperature11521ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|(e.g., 266 K) and then slowly cooled to 244 K, thereby causing a much slower over-all crystallisation process. The spectrum of the ice nucleation ability of the variouslygenerated oxalic acid dihydrate crystals at 244 K ranged from very efficient depositionmode ice nucleation, as also recently observed by Kanji et al. (2008), to almost iceinactivity. At times, distinctive changes in the ice nucleation ability were observed in5consecutive expansion cooling cycles.In the present manuscript, we summarise these results on the heterogeneous icenucleation potential of oxalic acid dihydrate. As a second species, anhydrous sodiumoxalate particles were also investigated. Albeit acknowledging the argumentation fromZobrist et al. (2006) that in most atmospheric particles oxalic acid (dihydrate) crystals,10if present, will be embedded in a surrounding solution and thereby be unsusceptible fordeposition mode ice nucleation, studies on pure oxalate particles remain of scientificinterest. They form the basis for the interpretation of the experiments with the morecomplex, multi-component aerosol particles and underline that already pure, compara-tively simply structured organic species may feature a complex, highly variable ice nu-15cleation behaviour. In addition, as already indicated in the first paragraph, mechanismsare discussed that could lead to the formation of coating layers of oxalate species onpre-existing aerosol particles. Therefore, one cannot exclude that potentially precipi-tated oxalate crystals are directly exposed to an ice supersaturated environment andcould also act as deposition mode ice nuclei.202ExperimentalA schematic view of the AIDA aerosol and cloud chamber facility, featuring the rele-vant instrumentation for the oxalic acid and sodium oxalate crystallisation and expan-sion cooling experiments, is shown in Fig. 2. The aerosol chamber, a 84.3 m3 sizedaluminium vessel of 4 m diameter, is located inside an isolating containment whose25interior can be cooled to temperatures as low as 183 K. The spatial temperature fluc-tuations throughout the whole chamber interior, as measured by arrays of horizontally11522ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|and vertically arranged temperature sensors, is less than ±0.3 K. The chamber canbe evacuated with two vacuum pumps to a minimum pressure of less than 0.01 hPa.Repetitive evacuation and flushing cycles with particle-free synthetic air usually pro-vide background particle number concentrations of less than 0.1 cm−3. In addition,the pressure control enables controlled expansion cooling experiments to establish ice5supersaturated conditions in the chamber as a prerequisite for studying ice nucleation.In the following, we will describe the various employed methods for aerosol gen-eration and present typical number size distributions of the added aerosol particles.Thereafter, we will briefly summarise the operation of the AIDA vessel as a cloud ex-pansion chamber and introduce the most important measurement devices, including in10situ water vapour measurements by tuneable diode laser absorption spectroscopy aswell as aerosol and cloud particle characterisation by optical particle counters, infraredextinction measurements, and laser light scattering and depolarisation measurements.For readers interested in more details, we recommend the publications by Benz etal. (2005) and M¨ohler et al. (2005) concerning the methodology of the expansion cool-15ing cycles and the recent review by Wagner et al. (2009) on the optical measurementsat the AIDA chamber. We will conclude this chapter with a tabular compilation of theexperimental procedures for the various crystallisation and ice freezing experimentswhose results will then be discussed in Sect. 3.Four different methods were used to generate oxalic acid and sodium oxalate20aerosol particles of median diameters ranging from 30 nm to 0.8 µm. In most cases,the aerosol was generated from aqueous solutions of typically 4–8 wt % oxalic acidand 3 wt % sodium oxalate, prepared by dissolving crystalline oxalic acid dihydrate((COOH)2·2H2O, Merck, >99.5%, 126.07 g mol−1) and sodium oxalate (Na2C2O4,Merck, >99.8%, 134.01 g mol−1) in Milli-Q quality water. The solutions were subse-25quently filtrated by a submicron syringe filter (PTFE membrane) to remove any poten-tially remaining particulates. Depending on the prevailing AIDA temperature, the addedsolution droplets either initially remained in a supersaturated liquid state, or rapidly, attimes instantly, crystallised to oxalic acid dihydrate and anhydrous sodium oxalate. The11523ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|largest particles were obtained by ultrasonic nebulisation (URANUS, Sinaptec France),yielding median diameters of about 0.55–0.80 µm. Panels a and b of Fig. 3 show typi-cal number size distributions of solid oxalic acid dihydrate and sodium oxalate particles,as generated by rapid crystallisation after adding the aerosol to the AIDA chamber at244 K. The size distributions were measured with a scanning mobility particle sizer5(SMPS, TSI) and an aerodynamic particle spectrometer (APS, TSI). Particles with me-dian diameters of around 100 nm were generated with a compressed-air atomiser (TSI,model 3076). As an example, panel c of Fig. 3 shows the size distribution of solid ox-alic acid dihydrate crystals, again obtained by rapid crystallisation after injection atT=244 K. For both the atomiser and ultrasonic nebuliser, an overall injection period of10about 30–60 min was required in order to obtain a sufficient mass loading for charac-terising the aerosol by long-path, in situ FTIR (Fourier transform infrared) extinctionspectroscopy. Typical number concentrations ranged from 10 000–20 000 cm−3 whenusing the atomiser and 300–500 cm−3 for the ultrasonic nebuliser. In selected exper-iments, the aqueous solutions were directly sprayed into the cooled cloud chamber15via a two-stream dispersion nozzle that was heated to about 293 K. This proved to bea clean and efficient way to generate sufficient aerosol mass loadings within 5–10 s.Panel d of Fig. 3 shows an exemplary size distribution of initially supersaturated oxalicacid solution droplets obtained by spraying a 4 wt % oxalic acid solution into the AIDAchamber at 266 K. As discussed in Sect. 3.3, the SMPS measurement is significantly20biased in the case of supercooled solution droplets because the instrument is oper-ated at a much lower relative humidity than prevalent in the AIDA chamber, inducing ashrinking of the particle size due to the loss of water. Finally, in a particular experimentnanometre-sized oxalic acid dihydrate crystallites were generated by passing syntheticair over a heated reservoir of crystalline oxalic acid dihydrate (373 K). The gaseous25oxalic acid/air mixture was then directed via stainless steel connection tubes into theAIDA chamber that was held at 228 K, thereby inducing the binary nucleation of oxalicacid and water. A total injection time of about 2 h was needed to positively identifyoxalic acid dihydrate as the major outcome of this procedure by FTIR extinction mea-11524ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|surements. As shown by panel e of Fig. 3, the median diameter of these crystalliteswas about 30 nm.In order to prepare the AIDA chamber for crystallisation experiments with succeedingexpansion cooling cycles, the inner walls of the aluminium vessel are coated with a thinice layer, typically made by evaporating a certain amount of Milli-Q quality water into5the evacuated chamber at some higher temperature, refilling the chamber with particle-free synthetic air to ambient pressure, and subsequent cooling to a lower temperature.Due to some internal heat sources like heating foils, heated sampling tubes, and themixing fan (see Fig. 2), the average gas temperature T is typically 1 ◦C higher than theaverage wall temperature. As a result, the prevailing water vapour pressure pw(T) is10slightly below the saturation pressure over ice pw,ice(T ), yielding an ice saturation ratioSice=pw(T)/pw,ice(T ) of about 0.90–0.95 at static pressure and temperature conditionsprior to the expansion cooling cycle.With the onset of pumping, pw(T ) decreasesat most linearly with the total pressure.Partly, this pressure-induced diminution inpw(T) is compensated for by evaporation of ice from the chamber walls because the15wall temperature almost remains constant during pumping. As pw,ice(T ) decreasesexponentially with the gas temperature during expansion cooling, a rapid increase inSice can be induced with a typical humidification rate dSice/dt of about 0.2/min for apumping speed of 120 m3/h. With the maximum pumping speed of 200 m3/h, the AIDAgas temperature can be reduced by as much as 10 ◦C during pumping. For any initial20temperature, this decrease is sufficient to reach or even go beyond saturation withrespect to supercooled water during the expansion.The expansion cooling cyclesthereby allow for investigating the heterogeneous ice nucleation potential of solid icenuclei in both the condensation and deposition mode.Accurate time profiles of the water vapour concentration during the expansion cooling25cycles are deduced from direct tuneable diode laser (TDL) absorption measurements.The employed room-temperature, near-infrared (1370±2 nm) telecommunication-typediode laser is coupled via an optical fibre into an open-path White-type multiple reflec-tion cell whose mirrors are directly mounted onto the interior walls of the AIDA chamber.11525ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|The optical path length of the White cell is currently adjusted between 23 and 84 m, de-pending on the water vapour concentration. By modulation of the diode laser current,the entire absorption line profile of a selected rovibrational water vapour transition isscanned, integrated, and converted to water vapour mixing ratios at a temporal resolu-tion of about 1 s. As a major benefit, the interstitial water vapour concentration can be5inferred even in the presence of dense clouds of supercooled water and/or ice parti-cles. Supplementing the TDL humidity measurements, the total water mixing ratio, i.e.,the sum of interstitial water vapour and evaporated condensed water from the aerosoland/or cloud particles, is measured with a fast high-precision chilled mirror hygrome-ter (MBW, model 373) at a time resolution of 1 s. The hygrometer is arranged outside10the thermostated housing of the AIDA chamber and samples pass through a stainlesssteel tube that is heated to 303 K.Three important optical instruments are used for obtaining information about thenumber concentration, size, shape, and chemical composition of the investigatedaerosol and cloud particles. A set of two optical particle counters (OPC1 and OPC2,15type WELAS2000, Palas) is mounted to the bottom of the AIDA chamber within itsthermostated housing, employing strictly vertical sampling tubes to minimise samplinglosses. Particle number concentrations and size distributions are inferred from measur-ing the pulses of white light scattered by single particles into an angle of 90±12◦. Giventhe measurement ranges of these instruments (0.7–40 µm, OPC1; and 2.3–104 µm,20OPC2), only the tail of the aerosol size distribution towards large particle diameterswill be covered. Cloud particles composed of micron-sized supercooled water dropletsand/or ice crystals, however, are quantitatively counted and sized. Note that the size in-formation for aspherical ice particles is somewhat ambiguous due to the dependence ofthe phase function on the particle shape and orientation in the detection volume. As a25second tool, infrared extinction spectra of the aerosol and cloud particles are recordedin situ by coupling a FTIR spectrometer to a second open-path White-type multiple re-flection cell with an optical path length that is adjustable between 23 and 257 m. Theinfrared spectra are typically recorded from 6000 to 800 cm−1 at 4 cm−1 resolution. In-11526ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|frared extinction is the sum of light absorption in the particles and light scattering bythe particles. The absorption contribution typically reflects distinctive bands of organicand inorganic functional groups, and, concerning the present study, allows for discrim-inating aqueous oxalic acid solution droplets from solid anhydrous oxalic acid or oxalicacid dihydrate crystals (Braban et al., 2003) as well as for distinguishing supercooled5water droplets from ice crystals (Zasetsky et al., 2005). The scattering contribution issensitive to the particle size and in principle allows for a retrieval of the particle size dis-tribution. For certain chemical species, depending on the particle size and the complexrefractive index, also information about the particle shape is inherent in the spectralhabitus of the infrared extinction bands. Measurements of laser light scattering and10depolarisation are a further powerful tool to detect deliquescence and efflorescencephase transitions in sub-micron sized aerosol particles as well as to monitor the freez-ing of large supercooled water droplets to aspherical ice crystals. For this purpose,the laser beam of a continuous wave semiconductor laser (Coherent, λ=488 nm) is di-rected horizontally through the AIDA chamber. The scattered light intensities are mea-15sured in the forward (Ifor, scattering angle 2◦) and polarisation resolved in the backwarddirection (Iback,par,Iback,per, scattering angle 178◦), with Iback,par and Iback,per denoting thebackscattered intensities parallel and perpendicular to the incident polarisation stateof the light beam. The backscattering linear depolarisation ratio δ is then given byδ = Iback,per/Iback,par. δ is zero for light scattering by spherical particles and in general20clearly different from zero for light scattering by aspherical particles, thereby featuringa strong dependency on the particle size and shape (Mishchenko et al., 1996). No-table exceptions are aspherical particles with small size parameters and/or extremeaspect ratios. These particle types may reveal only a very weak backscattering lineardepolarisation ratio (Zakharova and Mishchenko, 2000).25Concluding this section, we will give a tabular summary of the crystallisation and icenucleation experiments (Table 1), serving as a look-up table for the detailed discussionin the following section. The first four columns denote the experiment number, theinvestigated chemical species (either oxalic acid or sodium oxalate), the initial AIDA11527ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|gas temperature when adding the aerosol particles to the chamber, and the method foraerosol generation. The final column then briefly outlines the experimental procedure,including overall adjustments to the AIDA temperature and the number of expansioncooling cycles that have been performed at various temperatures.3Results and discussion5Given the broad diversity of the experimental procedures and the variety of the ice nu-cleation ability of the particles that is disclosed by the individual expansion runs, weconsidered it confusing for the reader to begin this section with an overall compilationof the results from all experiments. Instead, we describe and analyse the experimentsstep by step as they are listed in Table 1, thereby gradually unravelling the intricate10ice nucleation behaviour of the investigated species. As a guide for the reader, thediscussion is subdivided into three sections; Sect. 3.1 addresses experiments 1, 2,and 3 where the ice nucleation ability of oxalic acid dihydrate particles that have crys-tallised or nucleated in AIDA at T ≤244 K is described, Sect. 3.2 compares the findingsfrom Sect. 3.1 with the ice nucleation properties of anhydrous sodium oxalate particles15generated at 244 K, and Sect. 3 again focuses on the ice nucleability of oxalic aciddihydrate particles, this time generated by crystallisation in AIDA at T >244 K.3.1Ice nucleation ability of oxalic acid dihydrate, crystallisation/nucleation atT≤244 KIn a first type of experiments, oxalic acid solution droplets were added to the cooled20AIDA chamber at 244 K, either using an ultrasonic nebuliser (Exp. 1) or an atomiser(Exp. 2). As outlined in the introduction, the original purpose of these experiments wasto generate supercooled, highly supersaturated aqueous oxalic acid solution dropletsand mimic the trajectory A depicted in Fig. 1. Already in the course of the aerosol injec-tion period, however, a continuous increase in the backscattering linear depolarisation2511528ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|ratio δ was observed (top panel of Fig. 4), indicating that the supersaturated oxalic acidsolution droplets had started to crystallise. Note that the aerosol particles were addedthrough a stainless-steel injection tube located at level one of the AIDA chamber, i.e., atthe same height of the in situ light scattering and depolarisation measurements (Fig. 2).This causes the small-scale fluctuations of δ during the injection period tinj. After the5stop of aerosol injection, the mixing fan, located at the bottom of the AIDA chamberand continuously operating since the start of aerosol addition, rapidly ensures homo-geneous aerosol distribution throughout the chamber which leads to smoother tracesof the depolarisation ratio. Constant values of δ=0.38 and δ=0.18 are reached about2000 s and 4000 s after stopping the aerosol injection for experiments 1 and 2, respec-10tively, pointing to the complete crystallisation of the aerosol population. Although themagnitude of δ is both size- and shape-dependent, these differences in δ are likely re-lated to the clearly different particle sizes obtained by either ultrasonic nebulisation oratomisation, see panels (a) (Exp. 1) and (c) (Exp. 2) of Fig. 3. Simultaneously recordedinfrared extinction spectra provide evidence for the chemical nature of the crystallised15particulates. Four spectra recordings, monitored at different times during the injectionand crystallisation period of Exp. 2, are shown in the bottom panel of Fig. 4. Spectrum(a) features a still comparatively broad extinction band in the O-H stretching regimebetween 3700 and 3000 cm−1 caused by the presence of still unfrozen aqueous ox-alic acid solution droplets, but at the same time already bares a clear spectral imprint20of crystallised oxalic acid dihydrate particles, namely the doublet signature at about3480 and 3415 cm−1 due to the hydration water. The doublet structure of the C-Ostretching mode at 1250 cm−1 is a further indication for the presence of a mixture ofsupercooled aqueous oxalic acid solution droplets and oxalic acid dihydrate crystals, itslow-wavenumber component at 1240 cm−1 attributable to concentrated aqueous solu-25tions droplets and its high-wavenumber component at 1270 cm−1 attributable to oxalicacid dihydrate. At later times during the injection (spectrum c) the relative contributionsof the spectral features related to oxalic acid dihydrate gain in intensity. Spectrum (d)finally represents a completely crystallised aerosol population and closely matches the11529ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|infrared spectrum of oxalic acid dihydrate measured by Braban et al. (2003). For Exp. 1,only the final infrared spectrum after full crystallisation is depicted (trace e). Again, ox-alic acid dihydrate can be identified as the crystallisation product. The larger particlesize in comparison with Exp. 2 leads to an increased contribution of light scattering tooverall extinction, manifesting itself in a slight dispersion feature that is superimposed5on the 3500 cm−1 doublet signature.As apparent from Fig. 1 and already indicated in the introductory section, the compo-sition of the injected oxalic acid solution droplets, initially prepared from dilute solutionsof 4–8 wt % oxalic acid, will concentrate to about 62 wt % oxalic acid when exposed toice saturated conditions at 244 K. Given that the ambient relative humidity inside the10AIDA chamber is typically 5–10% lower than ice saturation due to the internal heatspots, an even higher solute concentration of about 67 wt % would be obtained. Toour knowledge, there is no previous literature report on the possibility to maintain suchhighly supersaturated oxalic acid solution droplets as used in the present study in asupercooled liquid state at low temperatures over a long period of time. The maximum15oxalic acid concentration in the emulsified solutions investigated by Zobrist et al. (2006)amounted to 19.87 wt %. In the first cooling cycle, this solution could be cooled to thehomogeneous freezing temperature of 222.9 K without showing crystallisation. At roomtemperature, oxalic acid solution droplets levitated in an electrodynamic balance couldbe concentrated up to about 80 wt % before efflorescing at a relative humidity of 52–2057% (Peng et al., 2001). As a comparison, the initial relative humidity with respect tosupercooled water in experiments 1 and 2 amounted to about 70%. In contradiction tothe Peng et al. (2001) result, numerous studies have addressed the difficulty to preparean aerosol population of crystalline oxalic acid particles by spray-drying of aqueous so-lutions at room-temperature even at a very low relative humidity of 5% (Braban et al.,252003; Mikhailov et al., 2009; Prenni et al., 2001).Without doubt, experiments conducted in a large-scale aerosol chamber such asAIDA are more susceptive to potential interferences caused by heterogeneous effectscompared to experiments with levitated droplets or emulsified solutions. One could11530ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|think of a number of sites in the chamber interior including e.g. the surfaces of themirrors for the internal multiple reflection cells or the rotors of the mixing fan whichcould potentially induce the heterogeneous crystallisation of the supersaturated oxalicacid solution droplets. The pronounced, immediate increase in the depolarisation ratiodirectly after the start of aerosol addition as well as the fast crystallisation of the en-5tire aerosol population in the 84 m3 sized vessel, however, can hardly be explained byheterogeneous crystallisation on a few selected sites. Moreover, as recently shown byexperiments on the glass transition and the ice nucleating abilities of citric acid aerosols(Murray et al., 2010), it is indeed possible to maintain a population of supersaturatedaqueous organic solution droplets at very low temperatures (215 K) in AIDA without any10indication of heterogeneous crystallisation. With our present instrumentation, a furtherquantitative analysis of the crystallisation process in terms of deriving a crystallisationrate, i.e., the number of crystallisation events per unit volume of aqueous oxalic acidsolution droplets per unit time, is not possible. For an analysis based on homogeneousnucleation theory, the time evolution of both the volume concentration of liquid droplets15(not including crystallised droplets) and the number concentrations of crystallised andstill unfrozen oxalic acid solution droplets have to be determined. The first quantitycould in principle be derived from the FTIR extinction spectra but would necessitateaccurate composition- and temperature-dependent infrared optical constants for bothaqueous oxalic acid solution droplets and oxalic acid dihydrate which are currently not20available. Moreover, separate and quantitative counting of the submicron-sized frozenand unfrozen solution droplets is extremely difficult to achieve. A promising instrumentdevelopment in this context is the Small Ice Detector probe SID3 constructed at theUniversity of Hertfordshire (http://strc.herts.ac.uk/pi/proj.html). The instrument mea-sures the spatial scattering patterns from individual particles down to sizes of about251 µm and, as inherent in its name, was primarily constructed to reveal the crystal mor-phology of ice crystals down to sizes too small to be resolved by conventional imaginginstruments. Because a part of the aerosol size distribution in Exp. 1 extended into the>1 µm size range, SID3 was tested for the first time at AIDA as a tool for the shape anal-11531ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|ysis of large aerosol particles. More theoretical work, however, is needed to confirmthat the recorded spatial scattering patterns indeed allow for a unique discriminationbetween spherical solution droplets and aspherical frozen droplets even for sizes downto 1 µm or, alternatively, to determine the threshold size above which such distinctionis possible. In a future crystallisation experiment, an aerosol would then have to be5generated whose entire number size distribution falls into a size regime that can bequantitatively counted and shape-classified by SID3.We will now focus our discussion on the ice nucleation potential of the crystallisedoxalic acid dihydrate particles generated during experiments 1 and 2. Time series ofvarious AIDA data recorded during two different expansion cooling cycles from Exp. 110are plotted in Fig. 5. The contents of each graph in Fig. 5 are arranged as follows. Thetop row panel shows the time series of the AIDA pressure (black line) and the meanAIDA gas temperature (red line) in the course of the expansion experiment. In thesecond row, the time evolution of the saturation ratios with respect to ice (Sice, blackline) and supercooled water (Sliq, red line) are plotted. These values are obtained as15quotients of the in situ measured water vapour concentrations (by TDL absorption) andthe saturation water vapour pressures over ice and liquid water, calculated with theequations from Murphy and Koop (2005) for the given mean AIDA gas temperature.In addition, the trace of Sice obtained for the total water concentration measured bythe MBW frost point hygrometer is shown as blue line. For the aerosol mass concen-20trations employed in the current experiments, any aerosol water content, if present, isusually negligible compared to the interstitial water vapour concentration. The traces ofSice from the TDL and MBW measurements therefore coincide as long as no clouds ofsupercooled water and/or ice are formed. Scatter plots from the optical particle counterOPC2 are shown in the third row. Each dot represents a single particle count event in25one of the OPC2 size channels. It is usually feasible to introduce an optical thresholdsize to distinguish between comparatively small interstitial aerosol particles and largeice crystals that have nucleated during an expansion cooling cycle. Counting all par-ticles above the threshold size then yields the ice particle number concentration Nice.11532ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|By dividing Nice through the prevailing aerosol number concentration, the so-calledice-active fraction fice of an aerosol population is obtained. The time evolution of thisquantity is shown in the forth panel of each graph. The final panel contains recordsfrom the in situ laser light scattering and depolarisation measurements, namely timeseries of the backscattered intensity Iback,par parallel to the incident polarisation state of5the laser (black line) and the linear backscattering depolarisation ratio δ (red line).The left panel of Fig. 5 shows AIDA data from the first expansion run conducted inthe course of Exp. 1. Pumping was started at time zero (see Fig. 4) after the com-plete crystallisation of the injected oxalic acid solution droplets to oxalic acid dihydrateat initial saturation ratios of Sice=0.93 and Sliq=0.70. The oxalic acid dihydrate crys-10tals are optically counted by the OPC2 up to a threshold size of about 10 µm, which,as mentioned above, can only be interpreted as an apparent particle diameter due tothe ambiguous sizing of aspherical particles. Shortly after the start of pumping, whenreaching a threshold ice saturation ratio Sice of about 1.03 (as indicated by the dashedgreen lines), a dense cloud of large ice crystals is formed. The threshold is given as15the prevailing relative humidity when the total ice particle number concentration, asmeasured by the OPC2, has exceeded a value of 1 cm−3. This definition is chosen, infavour of denoting the threshold relative humidity for ice nucleation on a certain per-centage of the aerosol population, to directly compare our results with those from Kanjiet al. (2008), see below. A different threshold definition would not change the major20conclusions from our experiments. Albeit continuing with pumping after exceeding thethreshold relative humidity, the nucleated ice crystals rapidly deplete the excess of wa-ter vapour in the gas phase. Therefore, Sice only gets to a peak value, Sice,max, of about1.07 during pumping before levelling of to Sice=1. This value is preserved as long asa stable ice cloud is maintained in AIDA, i.e., before all ice crystals have either evapo-25rated or grown large enough to fall to the bottom of the cloud chamber and be lost. Thenucleation of large ice crystals is also documented by the sharp increase in the inten-sity of the backscattered laser light by more than one order of magnitude. In contrast,the intricate, non-linear dependency of δ on particle size and shape provokes that the11533ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|depolarisation ratio of the large ice crystals is smaller than that of the oxalic acid di-hydrate crystals. The data from Fig. 5 reveal that the crystallised oxalic acid dihydrateparticles act as very efficient deposition mode ice nuclei. As apparent from the forthpanel, a significant number fraction with fice=0.1 has acted as a heterogeneous ice nu-cleus. The rapid depletion of the supersaturation may have prevented additional oxalic5acid dihydrate crystals with a slightly higher activation threshold to become activated.Three further expansion cooling cycles were performed with the aerosol loading fromExp. 1. Shortly after the ice cloud generated during run 1 had evaporated (at t=800 s,see Fig. 5), the chamber was refilled with dry, synthetic air to ambient pressure, leadingto a temporary decrease in the ice saturation ratio. After reaching ambient pressure,10it took about 30 min before the gas temperature and Sice regained their original valuesof 244 K and 0.93, respectively. Then, another expansion run was started. The AIDArecords from run 2 closely match those from run 1 and are therefore not explicitlyshown. During the second expansion cycle, the critical ice saturation ratio at the onsetof ice formation, Sice,crit, amounted to about 1.05. For Sice,max and fice, values of 1.1115and 0.12 were observed. The slightly higher activation threshold during run 2 might bedue to the fact that the most efficient ice nuclei from run 1 were no longer present. Earlyactivation will go along with growth to the largest ice crystals that will preferentially beremoved by sedimentation. On the other hand, the increase in fice might be relatedto the slightly higher peak value for Sice that was attained during run 2 compared to20run 1. So, a part of less active ice nuclei that had no chance to become activatedin run 1 can contribute to fice in run 2. After run 2, the AIDA chamber was refilledto ambient pressure and cooled to 233 K to probe the ice activity of the oxalic aciddihydrate crystals at a different temperature. The AIDA records from run 3 are shownin the right panel of Fig. 5. When starting the expansion at 233 K, the ice cloud is25formed at Sice,crit=1.06 with attaining a maximum value of Sice,max=1.13 in the courseof the expansion. As a notable difference to run 1 and run 2, a much higher ice-activefraction of fice=0.22 is observed. The results from run 3 could be closely reproducedin a succeeding expansion cycle after refilling the AIDA chamber to ambient pressure11534ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|(run 4: Sice,crit=1.08, Sice,max=1.16, fice=0.22). Such increase in fice with decreasingtemperature has been frequently observed in heterogeneous ice nucleation studies(Pruppacher and Klett, 1997), as e.g. for deposition mode ice nucleation on a varietyof mineral particulates (Roberts and Hallett, 1968).The oxalic acid dihydrate particles from Exp. 2 (aerosol generation with the atomiser)5were probed on their ice nucleating ability in two successive expansion cooling cyclesstarted at T=244 K. Also a fraction of up to 0.5% of these smaller crystals proved tobe efficient deposition mode ice nuclei, as evidenced by very low critical ice saturationratios of 1.06 and 1.08 in the two succeeding expansion runs. Note that the ice nucle-ation behaviour observed in Exp. 1 and Exp. 2 was highly reproducible. Both types of10experiments were repeated at least two times, yielding very similar values for Sice,critand fice. Figure 6 shows the results from an expansion cooling cycle with the smallest,nanometre-sized oxalic acid dihydrate crystallites generated by adding the gaseous ox-alic acid/air mixture into the AIDA chamber at 228 K (Exp. 3 in Table 1, see also panele in Fig. 3). The blue trace in the top panel of Fig. 7 shows the infrared spectrum of15the nucleated particulates after an injection time of 2 h. Albeit being noisier due to thelower aerosol mass concentration, the extinction spectrum clearly bares the signaturesof oxalic acid dihydrate. In contrast to the particles generated during Exp. 1 and Exp. 2,these tiny crystallites do not provoke any notable depolarisation of the backscatteredlaser light and do not extend into the measurement range of the OPC2. During the20expansion, a dense ice cloud is formed at Sice,crit=1.09. The linear depolarisation ratioshows a narrow peak in the first period of the nucleation event, i.e., during the time ofstrong ice particle growth. This feature can be accurately modelled by T-matrix calcu-lations and again reflects the pronounced size-dependency of δ (Wagner et al., 2009).A maximum ice particle number concentration of about 40 cm−3 is formed during the25expansion which is comparable to the Nice values obtained in runs 1–4 of Exp. 1. Asthe aerosol number concentration of the oxalic acid dihydrate crystallites, however,amounts to about 40 000 cm−3, only a comparatively small 0.1% fraction of the overallaerosol population has acted as a heterogeneous ice nucleus.11535ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|The observed high ice nucleation ability of solid oxalic acid dihydrate particles in thedeposition mode is consistent with the findings from Kanji et al. (2008). An estimateof the aerosol surface concentrations from the measured SMPS/APS size distribu-tions assuming spherical particles yields values of about 1.2×10−5, 0.78×10−5, and0.13×10−5 cm2 per cm3 volume of AIDA prior to the first expansion runs in Exp. 1, 2,5and 3, respectively. The corresponding critical ice saturation ratios Sice,crit of 1.03, 1.06,and 1.09 that were observed in the course of these expansion runs can therefore beinterpreted as the threshold relative humidities to nucleate one ice particle on aerosolsurface areas of 1.2×10−5, 0.78×10−5, and 0.13×10−5 cm2. Thus, even somewhatlower threshold saturation ratios for deposition mode ice nucleation were observed in10our study although the particulate surface areas were smaller than those in the Kanjiet al. (2008) experiments. In contradiction to our results, Prenni et al. (2001) havenoted that oxalic acid is not effective in nucleating ice until reaching water saturation atT<228 K. But as summarised in the introduction, the phase of the investigated particlesremains uncertain. Given that the particles whose ice nucleation ability was tested in15the CFDC chamber revealed a continuous water uptake at a relative humidity between45 and 91%, it is unlikely that they were crystalline oxalic acid dihydrate particlesbut rather concentrated solution droplets or, in accordance with the hypothesis fromMikhailov et al. (2009), gel-like amorphous particles. In addition, the authors presentcritical ice saturation ratios when 1% of the 50 and 100 nm particles have formed ice.20It is not specified whether a smaller fraction of the oxalic acid particles potentially hasformed ice at lower supersaturations, provided that this would be detectable with theexperimental setup.Parsons et al. (2004) have addressed the poor ice nucleation ability of malonic, suc-cinic, glutaric, and adipic acid in the deposition mode at temperatures above 243 K. It is25therefore tempting to ascribe the prominent ice nucleation behaviour of oxalic acid dihy-drate amongst all other investigated dicarboxylic acids to its intrinsic chemical structure.Note that also Zobrist et al. (2006) have highlighted the particular role of immersed ox-alic acid (dihydrate) crystals as the only species capable of inducing heterogeneous11536ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|ice nucleation. In particular the O-H--O hydrogen bonds by which the two moleculesof water are interconnected in between the carboxyl groups in the structure of oxalicacid dihydrate might be, if present at the particle surface, suitable sites to explain theextraordinary good ice-nucleability of this compound (Pruppacher and Klett, 1997). An-ticipating the results from further crystallisation experiments that were already briefly5addressed in the introduction and that will be discussed in detail in Sect. 3.3, however,suggest that also the surface structure in terms of defects like cracks, steps, or disloca-tions might be an important factor. Oxalic acid dihydrate crystals that were generatedin a different manner, i.e., by crystallisation from less supersaturated aqueous solu-tion droplets at T =266 K over a timescale of many hours, did not reveal such a high10deposition mode ice activity when cooled to 244 K as those particles that have rapidlycrystallised from highly supersaturated solution droplets directly at 244 K. In order togive reason for their outstanding ice-nucleability, one might therefore also speculatethat polycrystalline particles with numerous defects are formed when oxalic acid dihy-drate particles rapidly crystallise from highly supersaturated solution droplets at low15temperatures. Temperature and supersaturation are known to be key parameters thatinfluence size and shape of crystals grown from aqueous solutions (Omar and Ulrich,2006; Torgesen and Strassburger, 1964). This interpretation would be in agreementwith the results from the study by Shilling et al. (2006) about heterogeneous ice nucle-ation on spray-deposited ammonium sulphate and maleic acid particles. As shown by20scanning electron microscope images, the investigated particles contained a significantnumber of defects and appeared to be polycrystalline. This extensive surface structuremight be the reason that the particles acted as efficient ice nuclei for heterogeneousdeposition nucleation with Sice,crit as low as 1.05 at T =240 K. Also Parsons et al. (2004)considered that the investigated saturated dicarboxylic acids C3–C6 might be more im-25portant in ice cloud formation if the particles had more defects than those generatedin their study. And indeed, deposition mode ice nucleation on solid glutaric acid par-ticles (C5) with Sice,crit as low as 1.20 at 235 K has recently been observed (Baustianet al., 2010). The active site explanation would also account for the observed depen-11537ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|dence of fice on the size of the oxalic acid dihydrate particles. Smaller aerosol particleswill contain a lower number of defects or active sites that are capable of triggering icenucleation. Summarising our results, fice ranged from 0.10–0.22 at T=244–233 K fordp=0.8 µm (Exp. 1), whereas maximum ice-active fractions of about 0.5 and 0.1% wereencountered in Exp. 2 and Exp. 3 with the 0.1 µm and 0.03 µm sized particles at T =2445and 228 K, respectively.3.2Ice nucleation ability of sodium oxalate, crystallisation at T=244 KIn order to compare the ice nucleability of oxalic acid dihydrate with a different chemicalspecies, we have performed a similar set of experiments with sodium oxalate particles.Exp. 4 and Exp. 5 are duplicates of Exp. 1 and Exp. 2 where aqueous solutions of103 wt % sodium oxalate were injected into the AIDA chamber at 244 K with an ultrasonicnebuliser (Exp. 4) and an atomiser (Exp. 5). In contrast to the injection periods withaqueous oxalic acid solutions, not a continuous but an immediate, step-wise increasein the depolarisation ratio δ after the start of aerosol addition was monitored (Fig. 7,lower panel, showing the trace of δ during Exp. 4). This indicates that the supersat-15urated solution droplets instantly crystallised in AIDA at T =244 K and Sliq=0.7. Thesimultaneously recorded infrared extinction spectrum of the particulates (black tracein the top panel of Fig. 7) closely matches previously recorded spectra of sodium ox-alate in alkali halide matrices (Hartman and Hisatsune, 1967), showing major bands at1645, 1330, and 783 cm−1. The absence of any extinction band in the O-H stretching20regime reveals that anhydrous crystals of sodium oxalate have formed. Due to thelower employed solute concentration, caused by the lower solubility of sodium oxalatein water compared to oxalic acid, the median particle diameter of the crystallised par-ticles in Exp. 4 was smaller than in Exp. 1 (see panels a and b in Fig. 3). The smallerparticle size of the sodium oxalate crystals might be the primary reason for the overall25lower value of δ in comparison with the oxalic acid dihydrate particles from Exp. 1. Theobserved immediate crystallisation of sodium oxalate is in agreement with previouslyrecorded water cycles of sodium oxalate particles levitated in an electrodynamic bal-11538ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|ance at room temperature, showing an efflorescence transition at RH=72–75% (Pengand Chan, 2001).The left panel of Fig. 8 shows the AIDA records from the first expansion run with thecrystallised sodium oxalate particles. In remarkable contrast to all previously reportedexpansion cycles with oxalic acid dihydrate, the freshly crystallised sodium oxalate5particles did not act as efficient ice nuclei in the deposition mode at low ice super-saturations. The slight increase in the backscattering linear depolarisation mode thatoccurs at t=50 s (first vertical green line in Fig. 8), accompanied by the detection of avery few number of larger ice crystals in the OPC2, is the only indication for ice activityin the deposition mode. The number of nucleated ice crystals is too small to provoke10a diminution of the ice supersaturation. Instead, the relative humidity further increasesuntil at t=90 s saturation with respect to supercooled water is reached (second greenline). At water saturation, a dense cloud of supercooled water droplets with D<10 µmis formed (see OPC2 scatter plot). Within the estimated measurement accuracy of10%, the number concentration of cloud droplets detected by the OPC2 and indepen-15dently retrieved from the FTIR extinction spectra (Wagner et al., 2006) correspondsto the number concentration of the sodium oxalate particles measured by the CPCbefore the start of the expansion, indicating that almost all aerosol particles have be-come activated. The formation of spherical cloud droplets causes a sharp drop in thedepolarisation ratio, levelling of at a value slightly above zero due to the presence of20the small number of aspherical ice crystals. With droplet cloud formation, a clear andcontinuous increase in the ice particle number concentration was monitored (the iceparticles appear at larger apparent diameters with D>10 µm in the OPC2 scatter plot),indicating that a fraction of the sodium oxalate particles that are not active in depositionmode ice nucleation is active in the condensation mode. Note that the AIDA gas tem-25perature never dropped below the homogeneous freezing temperature of pure waterdroplets during the expansion. This result underlines that the sodium oxalate crystalsare not completely dissolved during CCN activation and subsequent droplet growth.At t=210 s when pumping is stopped (third green line), fice has reached a value of11539ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|about 1%. Recalling the results from the first expansion run with freshly crystallisedoxalic acid dihydrate particles of a similar, only slightly larger size (Exp. 1), 10% of theaerosol population had already been ice-active in the deposition mode at Sice≤1.07.At first glance, this striking difference supports the theory that the specific chemicalstructure of oxalic acid dihydrate with its included hydration water accounts for the high5ice nucleability of this particular compound. The ice nucleation activity of the sodiumoxalate particles, however, was significantly enhanced in the later period of expansionrun 1 and in two succeeding expansion cooling cycles.After the stop of pumping at t=210 during run 1, it took about one minute for thecloud of supercooled water droplets to evaporate. At t=270 s (forth green line), the10depolarisation ratio has regained its original value before the droplet cloud has formed.Iback,par has decreased by an order of magnitude, reflecting the contribution of lightscattering gone lost by the evaporation of the cloud droplets. The remaining signal inrelation to the background level at time zero now contains only the scattering contribu-tion from the nucleated ice crystals that are still present. Directly after the evaporation15of the cloud droplets, fice suddenly increases up to a value of about 0.03. Note that af-ter the evaporation of the cloud water at t=270 s, the bare particles are exposed to anenvironment that is still supersaturated with respect to the ice phase (Sice about 1.27).Obviously, a significant fraction of the sodium oxalate crystals attains a much higher icenucleation activity in the deposition mode when processed in the course of the droplet20activation and evaporation cycle. One might discuss this behaviour in the context ofthe well-known phenomenon of pre-activation of insoluble ice nuclei in heterogeneousice nucleation (Pruppacher and Klett, 1997; Roberts and Hallett, 1968; Vali, 2008). Inprevious studies, various experimental conditions under which solid ice nuclei have be-come activated to display an improved ice nucleation potential are reported. For some25substances, it was necessary to form ice crystals on them in a first nucleation event(Mossop, 1956). Small embryos of ice might then survive in suited cavities or crevicesat the particle surface even at relative humidities substantially below Sice=1 and causea lower threshold relative humidity for ice formation in a succeeding nucleation exper-11540ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|iment. As an alternative interpretation, the ice crystals may imprint their structure onthe surface of soft materials, yielding a surface structure that more closely resemblesthe ice lattice and that can then act as an activated site to promote ice nucleation (Vali,2008; Zuberi et al., 2001). A pre-activation has also been observed for particles cooledto temperatures below 236 K and exposed to an environment with Sice<1 in order to5explicitly avoid initial ice formation (Fukuta, 1966; Higuchi and Fukuta, 1966). Even atrelative humidities below the saturation value, water might condense in the capillariesof wettable solid particles. The condensed water might then freeze at T <236 K, againyielding an ice embryo capable of improving the ice nucleation ability of the particles.Finally, various organic substances suspended in water could be activated entirely in10the liquid phase (Edwards and Evans, 1971; Evans, 1967), just by cooling the sus-pensions below a critical temperature upon which, according to the interpretation byEdwards and Evans (1971), the layer of adsorbed water molecules on the surface ofthe particles transforms from a disordered state into an ordered state, the latter beinga better substrate for the nucleation of ice.15Concerning expansion run 1 with the sodium oxalate crystals, it is obvious that theimprovement of their ice nucleability goes along with the formation of the liquid phaseand the subsequent re-crystallisation of the particulates.As discussed above, thecrystals apparently do not get completely dissolved on the timescale of the experi-ment because otherwise the continuous increase in fice after t=90 s due to conden-20sation freezing would not have been monitored. But at least the outer particle shellmight have been partially dissolved during the droplet activation. Obviously, the subse-quent re-crystallisation after the droplet evaporation must have left behind a crystallinestructure that is more susceptive to deposition mode ice nucleation than the structuregenerated by the initial crystallisation of the injected, supersaturated sodium oxalate25solution droplets. After the evaporation of the ice crystals generated during run 1, themodified crystals were probed in a succeeding expansion run that was directly startedat reduced pressure, i.e., without refilling the AIDA chamber with dry, synthetic air toambient pressure (Fig. 8, right panel). In run 2, the modified sodium oxalate crystals11541ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|act as efficient ice nuclei in the deposition mode with Sice,crit=1.06 and fice=0.07, nowbeing almost as ice active as the slightly larger oxalic acid dihydrate crystals from run 1in Exp. 1 (Fig. 4). As clearly shown by the comparison of the AIDA records from thetwo succeeding expansion cycles of Exp. 4, the saturation ratio always remains belowwater saturation during run 2 and the temporary drop in the depolarisation ratio due5to the formation of spherical water droplets that was observed in run 1 does not oc-cur in run 2. Note that the number concentration of nucleated ice crystals in run 2 isby more than a factor of two higher than in run 1. This underlines that the observedimprovement of the ice nucleability cannot be solely related to those sodium oxalatecrystals that have nucleated ice in the first expansion run but must be ascribed to the10droplet activation event where almost all particles have been involved. The sodium ox-alate particles retain their high ice nucleability when the AIDA chamber is refilled withdry, synthetic air to ambient pressure after the end of expansion run 2 (at t=600 s, seeright panel of Fig. 8). During recompression, Sice temporarily dropped to a minimumvalue of about 0.61. Sice values lower than 0.65 were maintained over a period of about15900 s during refilling. Expansion run 3 that was started after reaching ambient pres-sure closely reproduced the AIDA records from run 2 in terms of the critical thresholdrelative humidity and ice-active fraction, yielding Sice,crit=1.07 and fice=0.06.In three consecutive expansion cooling cycles that were performed during Exp. 5 withthe smaller-sized sodium oxalate crystals, a similar improvement in the ice nucleation20ability was observed. Again, the sodium oxalate particles could be notably activated bya first expansion run where the relative humidity had increased up to water saturationand a cloud of supercooled water droplets had formed. In the two following expansionruns, early deposition mode ice nucleation could be observed, provoking that satura-tion with respect to supercooled water was no longer reached in the later course of25these expansions. The threshold saturation ratio Sice,crit for deposition ice nucleationin run 2 and run 3 of Exp. 5 was about 1.15, i.e., somewhat higher in comparisonwith the larger crystals probed in Exp. 4. Also, the ice-active fraction was only about0.05%, in accordance with the size-dependency of fice for oxalic acid dihydrate parti-11542ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|cles that has been observed in experiments 1, 2, and 3. In summary, the findings fromthe ice nucleation experiments with anhydrous sodium oxalate particles imply that thehigh ice-nucleability of oxalic acid dihydrate that has been observed so far might notbe primarily related to its hydrogen-bonded crystal lattice because also the water-freesodium oxalate crystals could be modified to quite efficient deposition mode ice nuclei.5Furthermore, we will now address two experiments that do not fit into the overall pic-ture of highly deposition mode ice-active oxalic acid dihydrate crystals which has beenoutlined until now.3.3Ice nucleation ability of oxalic acid dihydrate, crystallisation at T>244 KHaving observed a fast crystallisation of highly supersaturated oxalic acid solution10droplets to oxalic acid dihydrate at T =244 K, we decided to inject the aerosol parti-cles into AIDA at a much higher temperature (T=266 K) in order to probe whether lessconcentrated solution droplets would remain in a supercooled liquid state. The initialrelative humidity with respect to supercooled water for the crystallisation and ice nucle-ation experiments started at T =266 K was typically about 89%, yielding an approximate15equilibrium solute concentration of only 38 wt % oxalic acid in comparison with about67 wt % for an AIDA temperature of 244 K (see Fig. 1 and the discussion in Sect. 3.1).The top panel of Fig. 9 depicts the temporal evolution of the depolarisation ratio afterspraying an aqueous solution of 4 wt % oxalic acid into the cooled cloud chamber withthe dispersion nozzle at time zero for a time period of about 10 s (Exp. 6). The AIDA gas20temperature is shown as a comparison. For a very long observation time of almost 3 h(t=10 000 s, first green line in Fig. 9), δ remains constant at a value of 0.02, indicatingthat the injected droplets indeed initially remained in a supersaturated liquid state. Notethat a small background value for δ of typically 0.02–0.03 is observed even when onlyspherical droplets are present in the chamber. This is a result of Rayleigh scattering by25air molecules, an imperfect alignment of the polarisers in the optical set-up, and a lessthan 100% horizontal or vertical polarisation of the laser light. At later times, however,the depolarisation ratio starts to increase although the chamber temperature was still11543ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|kept constant at 266 K. This delayed onset of crystallisation is most probably the resultof a changing solute concentration. As indicated in Sect. 2, the spraying techniquewas an efficient method to inject substantial aerosol mass loadings into AIDA withina few seconds. Because a dilute oxalic acid solution was sprayed in, excess waterreleased from the evaporation of the solution droplets provoked a sudden rise in the5relative humidity from 89 to 96% after aerosol injection, as monitored by TDL absorp-tion spectroscopy (red trace in the bottom panel of Fig. 9). It took about 10 000 s todissipate the excess of water vapour and to restore the initial relative humidity of 89%.Due to the small negative slope of the ice-melting point curve at higher temperatures,the comparably slight increase of RHliq from 89 to 96% leads to a substantial change10in the equilibrium concentration of the oxalic acid solution droplets from 38 to about18 wt % (Fig. 1). Thus, the injected aqueous droplets initially attain a composition of18 wt % oxalic acid and then slowly lose water and concentrate to 38 wt % within thefirst 10 000 s of the crystallisation experiment. The shrinking of the solution dropletsdue to the loss of water is nicely mimicked by the decrease in the forward-to-backward15scattering ratioIforIback,par+Iback,per from the laser light scattering measurements, shown asblack trace in the lower panel of Fig. 9.Oxalic acid solution droplets with <38 wt % solute concentration can thus be main-tained in AIDA for some hours in a supercooled liquid state without any detectablecrystallisation, either homogeneously or heterogeneously on suited active sites in the20chamber interior. This is in agreement with the result from Zobrist et al. (2006) thatemulsified solutions of up to 20 wt % oxalic acid did not crystallise homogeneouslyupon cooling. Only for a sufficiently high solute concentration of about 38%, the crys-tallisation of the solution droplets becomes detectable on the timescale of the AIDAexperiment. To identify the chemical nature of the crystallised particulates, FTIR ex-25tinction spectra were again recorded at different time steps during the crystallisationexperiment (see small vertical bars on the time scale in the top panel of Fig. 9 andthe corresponding spectra in Fig. 10). Spectrum (a) features a broad extinction bandin the O-H stretching regime, characteristic of aqueous oxalic acid solution droplets,11544ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|as well as a pronounced light scattering contribution at non-absorbing wavenumbers>3600 cm−1. Using the optical constants of pure water in this wavenumber regime ina Mie fitting procedure, the diameter of the oxalic acid droplets can be estimated to beabout 1 µm. This is much larger than the droplet size simultaneously measured by theSMPS instrument (panel d, Fig. 3) because this device is not operated at the high rela-5tive humidity that is prevalent in the chamber interior and thus a significant shrinking ofthe particles size due to evaporation of water takes place. Spectrum (b) still bares thespectral habit of completely liquid droplets, however revealing a reduced intensity in theO-H absorption and high-wavenumber scattering regime due to loss of particle waterand the associated decrease in particle size. Spectra (c) and (d) bare the first signa-10tures of an ongoing crystallisation process, as can be seen more closely in the twospectral subsets shown in the bottom part of Fig. 10. Within the O-H stretching regime,two shoulders appear on the high-wavenumber tail of the liquid water extinction band,coinciding with the peak positions of the doublet signature in oxalic acid dihydrate. Inthe same way, a shoulder is formed at 1270 cm−1 in the regime of the C-O stretching15mode, further evidencing that oxalic acid dihydrate particles have started to crystallisefrom the supersaturated aqueous solution droplets. As discussed above, the experi-ments were not aimed at deriving crystallisation rates because it was not possible todetermine the number concentration of crystallised particles within the population ofstill unfrozen solution droplets. After having positively identified oxalic acid dihydrate20as the crystallisation product, the chamber was cooled overnight to 244 K to probe theice nucleation potential of the crystallised particulates (cooling started at t=25 000 s,see second green line in Fig. 9). About 20 000 s after the start of cooling, the entireaerosol population had crystallised, as indicated by a constant depolarisation ratio of0.16.25After homogeneous temperature and humidity conditions had established in thechamber interior at T =244 K the next morning, two subsequent expansion runs werestarted (Fig. 11, left panel). In contrast to all previously reported expansion runs, theoxalic acid dihydrate particles generated during Exp. 6 did not act as deposition mode11545ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|ice nuclei before reaching an ice saturation ratio of about 1.3 at t=85 s which is closeto water saturation. The first 200 s of the expansion run are shown in expanded viewin the right panel of Fig. 11. As can be seen by the short-time decrease in the depo-larisation ratio after t=85 s, ice formation coincides with the temporary formation of acloud of supercooled water droplets. These small-sized cloud droplets do not extend5into the measurement range of the OPC2 but can be seen in a series of infrared ex-tinction spectra that is shown in Fig. 12. The spectra were recorded at a time intervalof 10 s, beginning with spectrum (a) at t=85 s, i.e., immediately before the temporarydecrease in the depolarisation ratio occurred. Spectrum (b) features a pronouncedextinction band at 3300 cm−1 due to liquid water droplets. The major part of the spec-10tral signature can be accurately reproduced with a Mie fit using the optical constantsof supercooled water (red line), yielding a median droplet diameter of about 1.5 µm(Wagner et al., 2005a; Zasetsky et al., 2005). Only the small extinction minimum at950 cm−1 in spectrum (b) cannot be accurately mimicked by the Mie fit because thissignature is due to the first ice crystals that have nucleated. This feature gains in inten-15sity in the subsequent spectra recordings. Additionally, a second extinction minimum at3500 cm−1 is formed. These two extinction minima are called Christiansen bands andare the typical infrared spectral signature of ice crystals with sizes larger than at least5–6 µm (Arnott et al., 1995). Spectrum (d) can only be accurately reproduced with abimodal fit, employing one mode for water droplets and a second mode for ice crys-20tals. Spectrum (e) recorded at t=125 s finally represents a pure cloud of about 10 µmsized ice crystals, as retrieved from a unimodal fit using appropriate low-temperaturerefractive indices for ice (Zasetsky et al., 2005).Three different mechanisms can be discussed to explain the observed ice nucle-ation behaviour in the first expansion run of Exp. 6. Firstly, the oxalic acid dihydrate25crystals may have acted as condensation mode ice nuclei because the onset of iceformation apparently coincides with the formation of the cloud of supercooled waterdroplets. However, this would not explain that the ice particle number concentration stillincreases after the water cloud has completely glaciated at t=125 s via the Bergeron-11546ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|Findeisen process (see forth panel in Fig. 11). One might also speculate that a fractionof the oxalic acid dihydrate particles has acted as deposition mode ice nucleus with athreshold Sice,crit value so close to water saturation that less ice-active particles havesimultaneously become activated to water droplets. This would explain that ice crys-tals continue to nucleate after t=125 s because the environment is still supersaturated5with respect to ice. Finally, the oxalic acid dihydrate crystals may just have behaved ina similar manner as the sodium oxalate particles in the first expansion run of Exp. 4(Fig. 8). Some crystals have acted as condensation mode ice nuclei and the ice particlenumber concentration continues to increase because the crystalline structure of otheroxalic acid dihydrate particles was modified by the droplet activation and evaporation10cycle. Thus, when again directly exposed to an ice supersaturated environment afterthe evaporation of the cloud water, these modified particles could now act as depositionmode ice nuclei. The scatter plot from the OPC2 indeed indicates that a distinct secondmode of initially smaller ice crystals starts to nucleate at t=125 s (second green line inthe right panel of Fig. 11). Based on this interpretation, the first expansion runs of Exp.154 and Exp. 6 would only differ by the overall time period where a stable mixed-phasecloud is present in AIDA. As the absolute number concentration of ice crystals formedby condensation freezing is much lower in Exp. 4 compared to Exp. 6, the mixed-phasecloud persists for a much longer period of time. This explanation would also accountfor the modified ice nucleation behaviour that is displayed in the second expansion run20of Exp. 6. Run 2 was started at t=1080 s after all ice crystals from the previous expan-sion run had evaporated. In the course of expansion cooling, a pure ice cloud is formedby deposition mode ice nucleation at Sice,crit=1.20. Sice,max remains below water sat-uration in the later course of the expansion, explaining that a temporary decrease inδ due to a cloud droplet activation is not observed during run 2. The two succeeding25expansion runs thereby reveal the same change in the ice nucleation behaviour thathas been previously observed for the two repetitive expansion cooling cycles in Exp. 4with sodium oxalate particles.The results from Exp. 6 clearly differ from the previous finding that oxalic acid dihy-11547ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|drate particles are highly efficient ice nuclei in the deposition mode with a thresholdice saturation ratio below 1.1 even for nanosized crystallites. The particles generatedduring Exp. 6 were initially only ice-active in the condensation mode. Some particleswere then modified to become ice-active in the deposition mode, however revealing athreshold Sice,crit value still much higher than observed in Exp. 1, 2, and 3. Most likely,5the discrepancy in the ice nucleation behaviour between the various experiments iscaused by differences in the crystal structure and morphology of the oxalic acid di-hydrate particles. The impact of surface microstructure of crystalline solids on theirheterogeneous freezing ability is well documented in the literature, as for example inthe case of ammonium sulphate particles. Zuberi et al. (2001) have demonstrated that10the thermal history of solid ammonium sulphate cores imbedded in aqueous particlesaffected their morphology and heterogeneous ice nucleability. A crystalline core com-posed of many microcrystals, generated by warming a completely frozen (NH4)2SO4-H2O solution droplet only slightly above the eutectic temperature to induce the meltingof ice, acted as a very efficient ice nucleus, i.e., the heterogeneous ice-freezing tem-15perature was close to the eutectic temperature and Sice,crit close to 1. If instead thecrystalline core was composed of only one or two large crystals, prepared by tem-porarily rising the temperature of the particle to a higher conditioning temperature, theheterogeneous freezing temperature was close to the homogeneous freezing temper-ature. In the same way, Shilling et al. (2006) have detected a high deposition mode20ice activity for polycrystalline ammonium sulphate particles whereas in other studies,as discussed by the authors, subeutectic deliquescence has been observed. Concern-ing our present results, one might therefore speculate that the rapid crystallisation ornucleation of the oxalic acid dihydrate particles during Exp. 1, 2, and 3 has led to theformation of a surface microstructure with numerous defects, resulting in a high depo-25sition mode ice activity. During Exp. 6, where the oxalic acid dihydrate particles wereformed from less supersaturated oxalic acid solution droplets at a higher temperatureand were allowed to slowly grow by mass transfer from still unfrozen solution dropletsover a very long time scale of about 12 h, see Appendix A in Zobrist et al. (2006), more11548ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|regular crystals with less defects and a correspondingly lower ice nucleability mighthave formed. That different crystallisation conditions might strongly influence the re-sultant particle morphology has also been demonstrated in earlier AIDA investigationson the formation of nitric acid dihydrate crystals, either generated by shock freezingof a HNO3/H2O gas mixture or by homogeneous nucleation from supercooled aque-5ous nitric acid solution droplets where the nucleated nitric acid dihydrate crystals couldslowly grow in a supersaturated environment for a period of several hours (Wagner etal., 2005b). With our present experimental set-up, the hypothesis of a different surfacemicrostructure of the oxalic acid dihydrate crystals cannot be verified. In future AIDAcampaigns, it is planned to fully explore the potential of the SID3 detector to distinguish10between smooth and rough surfaces of larger ice crystals by the analysis of their an-gular scattering patterns. But again, it is not clear whether such distinction would alsobe achievable for only about 1–2 µm-sized aerosol particles. Probing the particle mor-phology of deposited crystals by environmental scanning electron microscopy might beanother option in future crystallisation experiments.15We will conclude the discussion with another example for the high diversity of the icenucleation behaviour of oxalic acid dihydrate. In Exp. 7, aqueous solution droplets ofoxalic acid were again added to the AIDA chamber at about 266 K, this time using theatomiser. Due to the much lower mass throughput of this device in comparison withthe dispersion nozzle used in Exp. 6, the injection period was about 30 min (Fig. 13).20The relative humidity with respect to supercooled water did not increase during aerosoladdition, but was stable between 89 and 90%. Hence, a small but immediate increasein the depolarisation ratio could be observed after adding the aerosol (black trace inFig. 13), indicating that the onset of crystallisation of the injected solution droplets wasnot delayed due to a temporary increase in the relative humidity as in Exp. 6. About2530 min after the end of aerosol addition, the particles were probed in a first expansionrun (time zero in Fig. 13). During the expansion, the aerosol particles were activated tocloud droplets and the depolarisation ratio temporarily got back to its initial backgroundvalue, pointing to the deliquescence of the particles that had already crystallised. After11549ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|expansion run 1, the AIDA chamber was slowly cooled to 244 K (red trace in Fig. 13),going along with an increase in the depolarisation ratio to a final value of about 0.15.As already outlined in the caption of Fig. 12, a variation of the temperature of the AIDAvessel leads to a change in the alignment of the White-cell mirrors of the in situ infraredextinction measurements. This means that the initially recorded background spectrum5is no longer a suited reference. The chemical nature of the crystallised particulatescan thus not be verified in Exp. 7, however, there is no reason to assume that notagain oxalic acid dihydrate particles have formed. Already before reaching T =244 K,the crystallised particles were probed on their ice nucleability in three expansion runsstarted at 260, 254, and 249 K. The observed behaviour was similar in all cooling cy-10cles and, therefore, only AIDA records from run 3 are shown in the left panel of Fig. 14.As obvious from the scatter plot of the OPC2, only very few large ice crystals with anoptical diameter larger than 10 µm were formed. The ice particle number concentrationwas clearly below 0.1 cm−3, meaning that these crystals might have also nucleated onthe few background particles that were present in the chamber before adding the ox-15alic acid aerosol. Instead of forming an ice cloud, the aerosol particles were activatedto a droplet cloud when reaching water saturation at t=190 s. The droplet diametersoverlap with the apparent diameters of the aspherical oxalic acid dihydrate particlesand the droplet cloud, therefore, only appears as a denser point cloud in the OPC2scatter plot between t=190 and 330 s (marked by the two vertical green lines). After20the evaporation of the supercooled cloud droplets at t=330 s, the temporarily reduceddepolarisation ratio regains its value from the start of pumping within less than 300 s.This indicates that at lower temperatures the oxalic acid dihydrate crystals do not getcompletely dissolved on the typical time scale of an expansion experiment, meaningthat at least a small solid core remains unsolved and causes the rapid re-crystallisation25of the particles. In previous literature studies, the dissolution kinetics of oxalic acid di-hydrate have only been investigated at room temperature in HCl (0.1 N) as dissolutionmedium under hydrodynamic conditions that are not comparable to the present exper-iments (Carstensen and Dali, 1999; Carstensen and Patel, 1975). Time series of the11550ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|depolarisation ratios recorded during the expansion runs 1–4 are compared in the rightpanel of Fig. 14. The compilation shows that the time period to recover the initial valueof δ after the evaporation of the droplet cloud continually decreases with decreasingAIDA temperature.The oxalic acid dihydrate particles generated during Exp. 7 do not exhibit any het-5erogeneous ice nucleation potential at T≥249 K. After expansion run 4, the crystalsare further cooled to 244 K, i.e., to a temperature where all previously generated oxalicacid dihydrate particles showed considerable ice-nucleability in either the condensationmode (Exp. 6, first run) or deposition mode (Exp. 1 and 2; Exp. 6, second run). TheAIDA records from two succeeding expansion runs started at about 244 K during Exp.107 are shown in Fig. 15. Surprisingly, also at the lower temperature the ice nucleationpotential of the oxalic acid dihydrate particles is very poor. Only about 0.1 cm−3 iceparticles were formed by deposition nucleation before water saturation was reached att=240 s (second vertical green line). Note that already 70 s earlier (t=170 s, first greenline) the relative humidity had just approached water saturation, leading to the first15drop in the depolarisation ratio. But unintentionally, the pumping speed was reducedat this time, leading to a slight decrease in the saturation ratio and to the evaporationof the droplets that just had formed. Only after increasing the pumping speed again,the entire aerosol population was activated to cloud droplets at t=240 s. Also by sub-sequent condensation freezing, only a very low number concentration of ice crystals20was generated, amounting to about 0.3 cm−3 in run 5 and 0.5 cm−3 in run 6 that wasimmediately started after the cloud particles from the first run had evaporated. As acomparison, the aerosol number concentration prior to pumping was about 9500 cm−3.Thus, oxalic acid dihydrate particles that have gone through several partial dissolutionand re-crystallisation steps at warmer temperatures are much poorer ice nuclei when25cooled to 244 K than crystals that were probed without preceding droplet activationruns (Exp. 6). Also, the first droplet activation at 244 K (run 5) does not notably changethe ice nucleation potential in contrast to the behaviour observed in the two repetitiveexpansion runs from Exp. 4 and 6. One cannot completely exclude that a species11551ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|different from crystalline oxalic acid dihydrate, like anhydrous crystals, amorphous par-ticles, or the recently discovered dimorphic sesquihydrates of oxalic acid (Wenger andBernstein, 2007), has been formed during the various droplet activation runs in Exp. 7due to the lack of information from the infrared spectra recordings. More likely, how-ever, the surface microstructure of the crystallised oxalic acid dihydrate particles that5have been generated and cloud-processed in different ways is highly variable. Thereby,this chemically simply-structured substance reveals a remarkable variability in its icenucleation potential, which, just by comparing Fig. 5 and Fig. 15, ranges from very highdeposition mode ice activity to very poor ice nucleation ability in both the deposition andcondensation mode, making it difficult to find a suited parameterisation to represent its10ice nucleation behaviour in a microphysical model.4SummaryWe have presented measurements on the heterogeneous ice nucleation ability of air-borne oxalic acid dihydrate and sodium oxalate particles in the deposition and con-densation mode at temperatures between 244 and 228 K. The observed ice nucleation15potential of the particles was strongly dependent on their crystallisation pathway. Crys-tallisation of highly supersaturated oxalic acid solution droplets at T=244 K or nucle-ation from a gaseous oxalic acid/water mixture that was added to the AIDA chamber atT=228 K yielded oxalic acid dihydrate crystals that proved to be highly efficient depo-sition mode ice nuclei. The threshold ice saturation ratio for ice nucleation was always20below 1.1. The number fraction of ice-active aerosol particles ranged between 22 and0.1%, depending on temperature and particle diameter that was varied between 0.8and 0.03 µm. This finding is in good agreement with recent measurements on ice for-mation via deposition nucleation on oxalic acid dihydrate particles that were depositedon a cold stage and, at first sight, points to a special role of oxalic acid within the25group of other saturated dicarboxylic acids that have not been observed to be goodice nuclei, neither in the deposition nor immersion mode. In further crystallisation and11552ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|ice nucleation experiments, however, oxalic acid dihydrate particles of partially muchpoorer ice nucleability were generated. In these experiments, the oxalic acid dihydrateparticles crystallised from less supersaturated oxalic acid solution droplets at a highertemperature and were allowed to slowly grow in a supersaturated environment from stillunfrozen aqueous oxalic acid solution droplets over a time period of many hours. Ob-5viously, these conditions have led to a more regular surface microstructure with fewerdefects that could have triggered ice nucleation as active sites. As a notable difficultyfor predicting the ice nucleation behaviour of oxalic acid dihydrate, the ice nucleabilitywas also dependent on whether the crystallised particles were subjected to precedingdroplet activation and evaporation steps or not. Such behaviour was also observed for10the second species investigated, namely sodium oxalate.Concerning the atmospheric implication of our findings, the presented experimentsare relevant to the scenario where by way of physical and chemical processing a coat-ing layer of organic material may have formed on prevalent particulates such as sootand mineral dust. Given the high variability of the heterogeneous ice nucleation po-15tential of oxalic acid dihydrate, it is not straightforward to anticipate in which directionthe ice nucleability of the pure compounds will be modified by the organic coating layer.On the one hand, some active mineral dust surfaces could become deactivated with anoxalic acid dihydrate coating of poor ice nucleation ability such as detected in Exp. 7.Such behaviour has recently been observed in AIDA experiments on the suppression20of the ice nucleation ability of mineral dust particles by a coating layer of secondary or-ganic matter from the products of the α-pinene – ozone reaction (M¨ohler et al., 2008).But on the other hand, soot particles that in most previous laboratory studies havebehaved as poor deposition mode ice nuclei at temperatures around 244 K could be-come more ice-active when coated with a layer of oxalic acid dihydrate that behaved25similar to the crystals generated in Exp. 1, 2, and 3. We therefore consider addressingthe effect of coating layers of solid dicarboxylic acids on the heterogeneous ice nucle-ability of soot or mineral dust particles in future chamber studies. Having taken intoaccount that in the atmosphere oxalic acid dihydrate crystals might also be embedded11553ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|in multi-component aqueous solutions droplets, we have already performed a set ofice nucleation experiments with ternary solution droplets containing sodium chlorideand sulphuric acid as third components in addition to oxalic acid and water. The re-sults of these experiments on the immersion freezing potential of oxalic acid will besummarised in a forthcoming manuscript.5Acknowledgements. The authors are grateful to all members of the AIDA staff for their con-tinuous support during the numerous successful measurement campaigns on ice nucleationwhich have been conducted at the chamber facility in the past decade. The work has beenfunded by the Helmholtz-Gemeinschaft deutscher Forschungszentren as part of the program“Atmosphere and Climate”.10ReferencesApelblat, A. and Manzurola, E.: Solubility of Oxalic, Malonic, Succinic, Adipic, Maleic, Malic,Citric, and Tartaric Acids in Water from 278.15 K to 338.15 K, J. Chem. Thermodyn., 19,317–320, 1987.Arnott, W. P., Dong, Y. Y., and Hallett, J.: Extinction Efficiency in the Infrared (2-18 µm) of15Laboratory Ice Clouds - Observations of Scattering Minima in the Christiansen Bands of Ice,Appl. Optics, 34, 541–551, 1995.Baustian, K. J., Wise, M. E., and Tolbert, M. A.: Depositional ice nucleation on solid ammoniumsulfate and glutaric acid particles, Atmos. Chem. Phys., 10, 2307–2317, 2010,http://www.atmos-chem-phys.net/10/2307/2010/.20Bellamy, L. J. and Pace, R. J.: Hydrogen Bonding in Carboxylic Acids.1. Oxalic Acids, Spec-trochim. Acta, 19, 435–442, 1963.Benz, S., Megahed, K., M¨ohler, O., Saathoff, H., Wagner, R., and Schurath, U.: T-dependentrate measurements of homogeneous ice nucleation in cloud droplets using a large atmo-spheric simulation chamber, J. Photochem. Photobiol. A, 176, 208–217, 2005.25Braban, C. F., Carroll, M. F., Styler, S. A., and Abbatt, J. P. D.: Phase transitions of malonic andoxalic acid aerosols, J. Phys. Chem. A, 107, 6594–6602, 2003.Carlton, A. G., Turpin, B. J., Altieri, K. E., Seitzinger, S., Reff, A., Lim, H. J., and Ervens, B.: At-11554ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|mospheric oxalic acid and SOA production from glyoxal: Results of aqueous photooxidationexperiments, Atmos. Environ., 41, 7588–7602, 2007.Carstensen, J. T. and Dali, M.: Determination of mass transfer dissolution rate constants fromcritical time of dissolution of a powder sample, Pharm. Dev. Technol., 4, 1–8, 1999.Carstensen, J. T. and Patel, M.: Dissolution Patterns of Polydisperse Powders – Oxalic Acid5Dihydrate, J. Pharm. Sci., 64, 1770–1776, 1975.Chebbi, A. and Carlier, P.: Carboxylic acids in the troposphere, occurrence, sources, and sinks:A review, Atmos. Environ., 30, 4233–4249, 1996.Clegg, S. L. and Seinfeld, J. H.: Thermodynamic models of aqueous solutions containing in-organic electrolytes and dicarboxylic acids at 298.15 K. 1. The acids as nondissociating10components, J. Phys. Chem. A, 110, 5692–5717, 2006.Crahan, K. K., Hegg, D., Covert, D. S., and Jonsson, H.: An exploration of aqueous oxalic acidproduction in the coastal marine atmosphere, Atmos. Environ., 38, 3757–3764, 2004.Ebisuzaki, Y. and Angel, S. M.: Raman Study of Hydrogen-Bonding in Alpha and Beta OxalicAcid Dihydrate, J. Raman Spectrosc., 11, 306–311, 1981.15Edwards, G. R. and Evans, L. F.: The Mechanism of Activation of Ice Nuclei, J. Atmos. Sci., 28,1443–1447, 1971.Ervens, B., Feingold, G., Frost, G. J., and Kreidenweis, S. M.: A modeling study of aqueousproduction of dicarboxylic acids: 1. Chemical pathways and speciated organic mass produc-tion, J. Geophys. Res.-Atmos., 109, D15205, doi:10.1029/2003JD004387, 2004.20Evans, L. F.: Ice Nucleation under Pressure and in Salt Solution, Trans. Faraday Soc., 63,3060–3071, 1967.Fukuta, N.: Activation of Atmospheric Particles as Ice Nuclei in Cold and Dry Air, J. Atmos.Sci., 23, 741–750, 1966.Hartman, K. O. and Hisatsune, I. C.: The Kinetics of Oxalate Ion Pyrolysis in a Potassium25Bromide Matrix, J. Phys. Chem., 71, 392–396, 1967.Higuchi, K. and Fukuta, N.: Ice in Capillaries of Solid Particles and Its Effect on Their NucleatingAbility, J. Atmos. Sci., 23, 187–190, 1966.Hinds, W. C.: Aerosol Technology, John Wiley & Sons, Inc., New York, 1999.Hsieh, L. Y., Kuo, S. C., Chen, C. L., and Tsai, Y. I.: Origin of low-molecular-weight dicarboxylic30acids and their concentration and size distribution variation in suburban aerosol, Atmos.Environ., 41, 6648–6661, 2007.Hudson, P. K., Gibson, E. R., Young, M. A., Kleiber, P. D., and Grassian, V. H.: A newly de-11555ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|signed and constructed instrument for coupled infrared extinction and size distribution mea-surements of aerosols, Aerosol. Sci. Tech., 41, 701–710, 2007.Kanji, Z. A. and Abbatt, J. P. D., Laboratory studies of ice formation via deposition mode nucle-ation onto mineral dust and n-hexane soot samples, J. Geophys. Res.-Atmos., 111, D16204,doi:10.1029/2005JD006766, 2006.5Kanji, Z. A., Florea, O., and Abbatt, J. P. D.: Ice formation via deposition nucleation on mineraldust and organics: dependence of onset relative humidity on total particulate surface area,Environ. Res. Lett., 3, 025004, doi:10.1088/1748-9326/3/2/025004, 2008.Kawamura, K., Kasukabe, H., and Barrie, L. A.: Source and reaction pathways of dicarboxylicacids, ketoacids and dicarbonyls in arctic aerosols: One year of observations, Atmos. Envi-10ron., 30, 1709–1722, 1996.Kerminen, V. M., Ojanen, C., Pakkanen, T., Hillamo, R., Aurela, M., and Merilainen, J.: Low-molecular-weight dicarboxylic acids in an urban and rural atmosphere, J. Aerosol Sci., 31,349–362, 2000.Kerminen, V. M., Teinila, K., Hillamo, R., and Makela, T.: Size-segregated chemistry of particu-15late dicarboxylic acids in the Arctic atmosphere, Atmos. Environ., 33, 2089–2100, 1999.Limbeck, A., Kraxner, Y., and Puxbaum, H.: Gas to particle distribution of low molecular weightdicarboxylic acids at two different sites in central Europe (Austria), J. Aerosol Sci., 36, 991–1005, 2005.Mikhailov, E., Vlasenko, S., Martin, S. T., Koop, T., and P¨oschl, U.: Amorphous and crystalline20aerosol particles interacting with water vapor: conceptual framework and experimental ev-idence for restructuring, phase transitions and kinetic limitations, Atmos. Chem. Phys., 9,9491–9522, 2009,http://www.atmos-chem-phys.net/9/9491/2009/.Mishchenko, M. I., Travis, L. D., and Mackowski, D. W.: T-matrix computations of light scattering25by nonspherical particles: A review, J. Quant. Spectrosc. Ra., 55, 535–575, 1996.M¨ohler, O., Benz, S., Saathoff, H., Schnaiter, M., Wagner, R., Schneider, J., Walter, S.,Ebert, V., and Wagner, S.: The effect of organic coating on the heterogeneous ice nucle-ation efficiency of mineral dust aerosols, Environ. Res. Lett., 3, 025007, doi:10.1088/1748-9326/3/2/025007, 2008.30M¨ohler, O., B¨uttner, S., Linke, C., Schnaiter, M., Saathoff, H., Stetzer, O., Wagner, R., Kr¨amer,M., Mangold, A., Ebert, V., and Schurath, U.: Effect of Sulphuric Acid Coating on Hetero-geneous Ice Nucleation by Soot Aerosol Particles, J. Geophys. Res.-Atmos., 110, D11210,11556ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|doi:10.1029/2004JD005169, 2005.Mossop, S. C.: Sublimation Nuclei, P. Phys. Soc. Lond. B, 69, 161–164, 1956.Murphy, D. M. and Koop, T.: Review of the vapour pressures of ice and supercooled water foratmospheric applications, Q. J. Roy. Meteorol. Soc., 131, 1539–1565, 2005.Murray, B. J., Wilson, T. W., Dobbie, S., Cui, Z., Al-Jumur, S. M. R. K., M¨ohler, O., Schnaiter,5M., Wagner, R., Benz, S., Niemand, M., Saathoff, H., Ebert, V., Wagner, S., and K¨archer, B.:Heterogeneous nucleation of ice particles on glassy aerosols under cirrus conditions, NatureGeosci., 3, 233–237, 2010.Omar, W. and Ulrich, J.: Solid liquid equilibrium, metastable zone, and nucleation parametersof the oxalic acid-water system, Cryst. Growth Des., 6, 1927–1930, 2006.10Parsons, M. T., Mak, J., Lipetz, S. R., and Bertram, A. K.:Deliquescence of malonic,succinic, glutaric, and adipic acid particles, J. Geophys. Res. Atmos., 109, D06212,doi:10.1029/2003JD004075, 2004.Peng, C., Chan, M. N., and Chan, C. K.: The hygroscopic properties of dicarboxylic and multi-functional acids: Measurements and UNIFAC predictions, Environ. Sci. Technol., 35, 4495–154501, 2001.Peng, C. G. and Chan, C. K.: The water cycles of water-soluble organic salts of atmosphericimportance, Atmos. Environ., 35, 1183–1192, 2001.Prenni, A. J., DeMott, P. J., Kreidenweis, S. M., Sherman, D. E., Russell, L. M., and Ming, Y.:The effects of low molecular weight dicarboxylic acids on cloud formation, J. Phys. Chem. A,20105, 11240–11248, 2001.Pruppacher, H. R. and Klett, J. D.: Microphysics of Clouds and Precipitation, Kluwer AcacdemicPublishers, Dordrecht, 1997.Roberts, P. and Hallett, J.: A Laboratory Study of Ice Nucleating Properties of Some MineralParticulates, Q. J. Rpy. Meteorol. Soc., 94, 25–34, 1968.25Shilling, J. E., Fortin, T. J., and Tolbert, M. A.: Depositional ice nucleation on crystalline organicand inorganic solids, J. Geophys. Res.-Atmos., 111, D12204, doi:10.1029/2005JD006664,2006.Torgesen, J. L. and Strassburger, J.: Growth of Oxalic Acid Single Crystals from Solution –Solvent Effects on Crystal Habit, Science, 146, 53–55, 1964.30Treuel, L., Schulze, S., Leisner, T., and Zellner, R.: Deliquescence behaviour of single levitatedternary salt/carboxylic acid/water microdroplets, Faraday Discuss., 137, 265–278, 2008.Vali, G.: Repeatability and randomness in heterogeneous freezing nucleation, Atmos. Chem.11557ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|Phys., 8, 5017–5031, 2008,http://www.atmos-chem-phys.net/8/5017/2008/.Villepin, J. D. and Novak, A.: Vibration-Spectra of Oxalic Acids.1. Infrared and Raman-Spectraof Beta-Phase of Some Acids, Spectrochim. Acta A, 34, 1009–1017, 1978a.Villepin, J. D. and Novak, A.: Vibration-Spectra of Oxalic Acids.2. Infrared and Raman-Spectra5of Alpha-Phase of Some Acids, Spectrochim. Acta A, 34, 1019–1024, 1978b.Wagner, R., Benz, S., M¨ohler, O., Saathoff, H., Schnaiter, M., and Schurath, U.: Mid-infraredextinction spectra and optical constants of supercooled water droplets, J. Phys. Chem. A,109, 7099–7112, 2005a.Wagner, R., Benz, S., M¨ohler, O., Saathoff, H., and Schurath, U.: Probing ice clouds by broad-10band mid-infrared extinction spectroscopy: case studies from ice nucleation experiments inthe AIDA aerosol and cloud chamber, Atmos. Chem. Phys., 6, 4775–4800, 2006,http://www.atmos-chem-phys.net/6/4775/2006/.Wagner, R., Linke, C., Naumann, K. H., Schnaiter, M., Vragel, M., Gangl, M., and Horvath,H.: A review of optical measurements at the aerosol and cloud chamber AIDA, J. Quant.15Spectrosc. Ra., 110, 930–949, 2009.Wagner, R., M¨ohler, O., Saathoff, H., Stetzer, O., and Schurath, U.: Infrared spectrum of nitricacid dihydrate – influence of particle shape, J. Phys. Chem. A, 109, 2572–2581, 2005b.Wenger, M. and Bernstein, J.: Cocrystal design gone awry? A new dimorphic hydrate of oxalicacid, Mol. Pharm., 4, 355–359, 2007.20Yang, F., Chen, H., Wang, X. N., Yang, X., Du, J. F., and Chen, J. M.: Single particle massspectrometry of oxalic acid in ambient aerosols in Shanghai: Mixing state and formationmechanism, Atmos. Environ., 43, 3876–3882, 2009.Yao, X. H., Fang, M., and Chan, C. K.: Size distributions and formation of dicarboxylic acids inatmospheric particles, Atmos. Environ., 36, 2099–2107, 2002.25Zakharova, N. T. and Mishchenko, M. I.: Scattering properties of needlelike and platelike icespheroids with moderate size parameters, Appl. Optics, 39, 5052–5057, 2000.Zasetsky, A. Y., Khalizov, A. F., Earle, M. E., and Sloan, J. J.: Frequency Dependent ComplexRefractive Indices of Supercooled Liquid Water and Ice Determined from Aerosol ExtinctionSpectra, J. Phys. Chem. A, 109, 2760–2764, 2005.30Zobrist, B., Marcolli, C., Koop, T., Luo, B. P., Murphy, D. M., Lohmann, U., Zardini, A. A., Krieger,U. K., Corti, T., Cziczo, D. J., Fueglistaler, S., Hudson, P. K., Thomson, D. S., and Peter, T.:Oxalic acid as a heterogeneous ice nucleus in the upper troposphere and its indirect aerosol11558ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|effect, Atmos. Chem. Phys., 6, 3115–3129, 2006,http://www.atmos-chem-phys.net/6/3115/2006/.Zuberi, B., Bertram, A. K., Koop, T., Molina, L. T., and Molina, M. J., Heterogeneous freezing ofaqueous particles induced by crystallized (NH4)2SO4, ice, and letovicite, J. Phys. Chem. A,105, 6458–6464, 2001.511559ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|Table 1. Compilation of the experimental procedures for the various crystallisation and icenucleation experiments with oxalic acid and sodium oxalate. Tinj denotes the initial AIDA gastemperature when adding the aerosol particles to the chamber. The procedure of each experi-ment is discussed in detail in Sect. 3.Exp.SoluteTinj/KAerosol genera-tionExperimental procedure1Oxalicacid244UltrasonicnebuliserWaiting for full crystallisation of oxalic acid dihydrate,followed by two repetitive expansion cooling cyclesat 244 K.Cooling of the chamber to 233 K and two further ex-pansion cooling cycles at 233 K.2Oxalicacid244AtomiserWaiting for full crystallisation of oxalic acid dihydrate,followed by two repetitive expansion cooling cyclesat 244 K.3Oxalicacid228Cooling ofgaseousoxalicacid/air mixtureImmediate nucleation of oxalic acid dihydrate. Oneexpansion cooling cycle at 228 K.4Sodiumoxalate244Ultrasonic neb-uliserImmediate crystallisation of sodium oxalate. Threerepetitive expansion cooling cycles at 244 K.5Sodiumoxalate244AtomiserImmediate crystallisation of sodium oxalate. Threerepetitive expansion cooling cycles at 244 K.6Oxalicacid266Two-stream dis-persion nozzleObservation of oxalic acid dihydrate crystallisationat 266 K for about 6.5 h. Cooling of the chamber to244 K and two repetitive expansion cooling cycles at244 K.7Oxalicacid266AtomiserObservation of oxalic acid dihydrate crystallisationat 266 K for about 30 min. Droplet activation at 266K, followed by cooling of the chamber to 244 K. Dur-ing cooling, three further droplet activations at 260,254, and 249 K. Finally, two repetitive expansioncooling cycles at 244 K.11560ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|020406080100180200220240260280RHliq / %46974851556066728088B2B1A3A2 T / Kwt% oxalic acidA1C97 Fig. 1. Schematic experimental trajectories of various ice nucleation experiments with super-cooled oxalic acid solution droplets in the oxalic acid – water phase diagram. The solid blackline indicates the ice melting-point curve, as calculated for an ideal solution with Raoult’s law.Although Raoult’s law reasonably fits the measured water activities of aqueous oxalic acid atroom temperature (Clegg and Seinfeld, 2006; Peng et al., 2001), its validity for low tempera-tures still has to be verified. Hence, the black trace possibly will only give a rough estimateof the solute concentration at ice-saturated conditions. The right y-axis denotes the prevailingrelative humidity with respect to supercooled water, RHliq, for an environment that is saturatedwith respect to ice, i.e., when moving on the ice-melting-point curve within the phase diagram.The grey circles denote the critical homogeneous ice nucleation temperatures as a functionof oxalic acid concentration from the study by Zobrist et al. (2006). The blue and green tra-jectories represent hypothetical expansion cooling experiments with airborne particles in theAIDA cloud chamber, whereas the red trajectory corresponds to a freezing experiment withemulsified oxalic acid solution droplets. See text for details.11561ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|T,�pSynthetic AirSupplyCryostatLiquid�NitrogenAerosol�VesselAIDAVacuumPump�2SmallAerosolChamberVacuumPump�1Temperature�Controlled�Housing-90°C�to�+60°COxalic AcidAerosolGeneratorsFilterSamplingDMACNC3010CNC3010CNC3025,�SMPSScatteringIntensityFSDOPS�Laser=�488�nm�Scattering�andDepolarisationBSDFTIR�SpectrometerTDL Water�VapourDetectionChilled�Mirror�FrostPoint�HygrometerGas�temp.�sensorsFanOPCsSID3Fig. 2. Schematic cross section of the AIDA aerosol and cloud chamber facility, featuring themost important instrumentation for the crystallisation and ice nucleation experiments with oxalicacid dihydrate and sodium oxalate.11562ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|012340123450120120.010.1101234 a, oxalic acid dihydrate SMPS APSb, sodium oxalate c, oxalic acid dihydrate dN/dlogdp / cm-3d, aqueous oxalic acid e, oxalic acid dihydrate dp / µmFig. 3.Typical normalised SMPS and APS number size distributions of oxalate aerosols generated by variousmethods; see text for details. APS measurements were only done in cases (a) and (b) where a significant numberfraction of the aerosol extended to diameters >0.5 µm. The aerodynamic diameter of the APS was converted into avolume-equivalent sphere diameter dp assuming a particle density of 1.653 g cm−3 for oxalic acid dihydrate (a) and2.34 g cm−3 for sodium oxalate (b). Due to the lack of appropriate shape information, the dynamic shape factor χ wasset to 1.0 in both cases. When assuming spherical particles with χ=1, the mobility-equivalent diameter of the SMPSwould be equal to dp (Hinds, 1999; Hudson et al., 2007).11563ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|-8000-6000-4000-2000000.20.44500400035003000200015001000edcbtinj depolarization ratio δtime / s Exp. 1 Exp. 2tinjaedcab extinction / arb. unitswavenumber / cm-1Fig. 4. Top panel: temporal evolution of the backscattering linear depolarisation ratio δ afterstarting the addition of aqueous oxalic acid solution droplets to the AIDA chamber at 244 K withan ultrasonic nebuliser (Exp. 1) and an atomiser (Exp. 2). The horizontal arrows denote theoverall time period of aerosol injection (tinj). Time zero denotes the start of the first expansioncooling experiment that was performed later on to investigate the ice nucleation potential of theadded aerosol particles. Bottom panel: infrared extinction spectra recorded at different timesteps during the injection and crystallisation periods of Exp. 1 and Exp. 2, as indicated by thesmall vertical bars on the time scale in the top panel.11564ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|Page 1/1 Fig. 5. Time series of various AIDA records during two expansion cooling cycles from Exp. 1.In each graph, time zero denotes the start of pumping. For run 1, time zero corresponds to t=0on the time scale in the top panel of Fig. 4. The five panels contain the following data. Panel 1:AIDA pressure (black line) and mean gas temperature (red line). Panel 2: saturation ratios withrespect to ice (Sice, black and blue line) and supercooled water (Sliq, red line). Sliq is only shownfor the gas-phase water concentration measured by TDL absorption whereas Sice is plotted forboth the gas-phase and total water concentration as measured ex situ with the MBW chilledmirror hygrometer. Panel 3: size of individual particles detected by the OPC2. Panel 4: numberfraction of ice-active aerosol particles. Panel 5: backscattering intensity parallel to the incidentpolarisation state of the laser light (black line) and backscattering linear depolarisation ratio δ(red line). See text for details.11565ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|Page 1/1 Fig. 6. Time series of various AIDA records during the first expansion cooling cycle from Exp. 3.The individual panels contain the same measurements as in Fig. 5, except from panel 4 wherethe ice particle number concentration Nice instead of the ice-active fraction fice is shown.11566ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|600050004000300020001000-4000-3000-2000-1000000.10.20.3 extinction / arb. unitswavenumber / cm-1 Exp. 2 Exp. 3, *10 Exp. 4, *2 δtime / sExp. 4tinjFig. 7. Top panel: infrared extinction spectra of oxalic acid dihydrate particles generated dur-ing Exp. 2 (green line, corresponding to spectrum d shown in Fig. 4) and Exp. 3 (blue line,scaled by a factor of 10). The black trace shows the spectral signature of sodium oxalate par-ticles recorded at t=−1600 s during Exp. 4 (see timescale in the bottom panel). Bottom panel:temporal evolution of the backscattering linear depolarisation ratio δ after starting the additionof aqueous sodium oxalate solution droplets to the AIDA chamber at 244 K with an ultrasonicnebuliser (Exp. 4).11567ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|Page 1/1 Fig. 8. Time series of various AIDA records during two expansion cooling cycles from Exp.4. In each graph, time zero denotes the start of pumping. For run 1, time zero correspondsto t=0 on the time scale in the bottom panel of Fig. 7. The individual panels show the samemeasurements as in Fig. 5.11568ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|Page 1/1 Fig. 9. Time series of the backscattering linear depolarisation ratio δ (top panel, black line), theAIDA gas temperature (top panel, red line), the forward-to-backward scattering ratio (bottompanel, black line), and the saturation ratio with respect to supercooled water (bottom panel,red line) after spraying an aqueous solution of 4 wt % oxalic acid into the chamber at time zero(Exp. 6). The small blue bars on the time scale in the top panel indicate the times of the infraredspectra recordings that are shown in Fig. 10.11569ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|4000300020006000500040003000200010001400130012001100dcb extinction / arb. unitswavenumber / cm-1OADdcb extinction / arb. unitswavenumber / cm-1aOADddcb extinction / arb. unitswavenumber / cm-1Fig. 10. Infrared extinction spectra of crystallising oxalic acid solution droplets recorded atdifferent time steps during Exp. 6, as indicated by the vertical blue bars in Fig. 9. The twobottom panels show expanded views in the regime of the O-H and C-O stretching modes andinclude a reference spectrum of pure oxalic acid dihydrate particles (red trace, corresponding tospectrum e in Fig. 4). See text for details. Note that only an incomplete background subtractionof the absorption lines of gaseous water could be achieved due to the high AIDA temperatureand the correspondingly high water vapour concentration.11570ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper| Figure 11. Time series of various AIDA records during the expansion cooling cycles from Exp. 6. The left panel covers the complete time period of expansion runs 1 and 2 whereas the right panel shows an expanded view of the first 200 seconds from run 1. Fig. 11. Time series of various AIDA records during the expansion cooling cycles from Exp.6. The left panel covers the complete time period of expansion runs 1 and 2 whereas the rightpanel shows an expanded view of the first 200 s from run 1.11571ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|600050004000300020001000edcb extinction / arb. unitswavenumber / cm-1aFig. 12. Infrared extinction spectra recorded from t=85 s (spectrum a) to t=125 s (spectrume) at a 10 s time interval during expansion run 1 from Exp. 6 (black lines). Note that a newbackground spectrum had to be recorded prior to these measurements. This was necessarybecause the cooling of the AIDA chamber from 266 to 244 K (see Fig. 9) induced a misalign-ment of the White-cell system. The extinction features due to the crystallised oxalic acid dihy-drate particles (see Fig. 10) are therefore not visible. Various computed extinction spectra thatbest-fitted the measurements are shown as red lines. Spectra b and e were fitted by unimodallog-normal distributions of either supercooled water droplets or ice crystals, respectively. Thelatter were modelled as circular cylinders of aspect ratio 0.7 (Wagner et al., 2006). The mixed-phase cloud scenario represented by spectrum d was fitted by two log-normal modes, with onemode for the water droplets and the other for the ice crystals.11572ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|Page 1/1 Fig. 13. Temporal evolution of the backscattering linear depolarisation ratio δ (black line) andthe AIDA gas temperature (red line) during Exp. 7 that was started by adding aqueous oxalicacid solution droplets to the AIDA chamber at 266 K with an atomiser. In the course of the ex-periment, the ice nucleation ability of the aerosol particles was probed in six expansion coolingcycles, as indicated by the vertical blue bars. The horizontal green line denotes the backgroundvalue for the depolarisation ratio of about 0.02 that is typically observed even if only sphericalparticles are present.11573ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|Page 1/1 Fig. 14. Left panel: time series of various AIDA records during expansion cooling run 3 fromExp. 7 (see also Fig. 13). Because almost no ice crystals have formed, the sub-panel con-taining the ice-active fraction or ice particle number concentration is not shown. Right panel:compilation of the backscattering intensities and depolarisation ratios recorded during the ex-pansion runs 1–4 from Exp. 7. In each graph, time zero denotes the start of pumping. Only forthe first run, time zero corresponds to t=0 on the time scale of Fig. 13.11574ACPD10, 11513–11575, 2010Heterogeneous icenucleation potentialof oxalic aciddihydrateR. Wagner et al.Title PageAbstractIntroductionConclusionsReferencesTablesFigures◀▶◀▶BackCloseFull Screen / EscPrinter-friendly VersionInteractive DiscussionDiscussion Paper|Discussion Paper|Discussion Paper|Discussion Paper|Copernicus Publications Contact Legal Body Page 1/1 Fig. 15. Time series of various AIDA records during the expansion cooling runs 5 and 6 fromExp. 7 (see also Fig. 13).11575View publication stats | RESEARCH GATE |
© 2022. Edouard Siregar. This research/review article is distributed under the terms of the Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0). You must give appropriate credit to authors and reference this article if parts of the article are reproduced in any manner. Applicable licensing terms are at https://creativecommons.org/licenses/by-nc-nd/4.0/. Global Journal of Science Frontier Research: F Mathematics and Decision Sciences Volume 22 Issue 4 Version 1.0 Year 2022 Type : Double Blind Peer Reviewed International Research Journal Publisher: Global Journals Online ISSN: 2249-4626 & Print ISSN: 0975-5896 Boosting Human Insight by Cooperative AI: Foundations of Shannon-Neumann Logic By Edouard Siregar Abstract- We present the logical foundation of an artificial intelligence (AI) capable of dealing with complex dynamic challenges, that would be very hard to handled using traditional approaches (e.g. predicate logic and deep learning). The AI is based on a cooperative questioning game, to boost insight. Insight gains are measured by information, probability, uncertainty (Shannon), as well as utility (von Neumann). The framework is a two-person cooperative iterated Q&A game, in which both players (human, AI agent) benefit (positive-sum): the human player gains insight and the AI player learns to improve its suggestions. Generally speaking, valuable insight is typically gained by asking ’good’ questions about the ’right’ topic, at the ’appropriate’ time and place: by posing insightful questions. In this study, we propose a logical and mathematical framework, for the meanings of ’good, right, appropriate’, within clearly-defined classes of human intentions. Keywords: artificial general intelligence, complexity, cooperative learning games, frame drift problem, information entropy, insight problems, predicate logic, renormalization, utility, value-alignment problem. GJSFR-F Classification: DDC Code: 006.3 LCC Code: Q335 BoostingHumanInsightbyCooperativeAIFoundationsofShannonNeumannLogic Strictly as per the compliance and regulations of: Boosting Human Insight by Cooperative AI: Foundations of Shannon-Neumann Logic Edouard Siregar I. Introduction & Motivation 1Year 20221© 2022 Global Journals Global Journal of Science Frontier Research Volume XXII Issue ersion I VIV ( F )Purely algorithmic AI, from Predicate Logic [1] to Deep Learning neural nets [2–4],have proven highly effective for static, well-defined, narrow problems [5]. For dy-namic, complex challenges, traditional AI becomes too ’brittle’ (fails due to inap-propriate application), and human insight is necessary to guarantee sound, human-aligned solutions. Solutions built on insufficient insight, can have deep long-lasting,human and economic consequences (e.g. conflict avoidance, war on drugs, pan-demics or climate ill-preparedness).Insight is usually gained (besides randomness and serendipity), by knowingwhen/where to pose which types of questions, about what topic: that is, by posing’insightful questions’. This ability thus requires a precise logical and mathematicalmeaning for the variables {when, where, what, which}, within well-defined contextsC, of human cognitive mindsets.In this paper, the task of generating insightful questions, uses a frameworkwe call Shannon-Neumann or SN-Logic, to cope with the fundamental concepts inAbstract- We present the logical foundation of an artificial intelligence (AI) capable of dealing with complex dynamic challenges, that would be very hard to handled using traditional approaches (e.g. predicate logic and deep learning). The AI is based on a cooperative questioning game, to boost insight. Insight gains are measured by information,probability, uncertainty (Shannon), as well as utility (von Neumann).The framework is a two-person cooperative iterated Q&A game, in which both players (human, AI agent) benefit (positive-sum): the human player gains insight and the AI player learns to improve its suggestions. Generally speaking, valuable insight is typically gained by asking ’good’ questions about the ’right’ topic, at the ’appropriate’ time and place: by posing insightful questions. In this study, we propose a logical and mathematical framework, for the meanings of ’good, right, appropriate’, within clearly-defined classes of human intentions.AI based on this Shannon-Neumann Logic, combines symbolic AI with cooperative learning. It is transparent (no hidden layers), explainable (no unjustifiable moves), and remains human-aligned (no AIvs human contradictions) because of continuous cooperation (positive-sum game). In this paper, we focus uniquely on logical validity, andleave the complex topic scientific soundness for future research.Keywords: artificial general intelligence, complexity, cooperative learning games,frame drift problem, information entropy, insight problems, predicate logic, renormalization, utility, value-alignment problem.1. Leon Sterling L. and Ehud Shapiro E. (1986) The Art of Prolog: AdvancedProgramming Techniques (MIT Press Series in Logic Programming), MIT Press; First Ed., ISBN-10: 0262192500 ISBN-13: 978-0262192507RefAuthor: Sofia Labs, AI systems research, Maryland, USA. e-mail: edsiregar@SofiaLabsLLC.com © 2022 Global Journals1Year 20222 Global Journal of Science Frontier Research Volume XXII Issue ersion I VIV ( F )Boosting Human Insight by Cooperative AI: Foundations of Shannon-Neumann Logic• In section 1, we discussed algorithmic vs human intelligence, and the purposeof SN-Logic.• In section 2, we present the two-person (human H, AI agent ASN) coop-erative Iterated Questioning (IQ) game’s role, from both H’s and ASN’sperspectives• In section 2.3, we discuss the dynamic drift problem: coping with the chang-ing human understanding of a given complex challenge, using a dynamicoptimization process. It’s impossible to clearly define a single problem, incomplex challenges (e.g. war on drugs) so that they can last for decades• In sections 3.1-3.2, we discuss SN-Logic’s requirements to cope with insight(which involves causality, information, logic, probability, uncertainty andutility) and the spaces over which SN-Logic operates• In sections 3.3-3.4, we introduce SN-Logic’s grammar: semantics + syntaxThe syntax is used by question generators, to build millions of possible ques-tions• In section 3.5, we present SN-Logic predicates of two classes: problem difficulty-minimizing, and solution quality-maximizing, used in all inferences• In section 3.6, we discuss the complexity and scope of SN-Logic, and section3.7 highlights the distinction between knowledge acquisition (symbolic AI)and cooperative (machine) learning, both present in our AI• In section 3.8, we introduce the normal form for making SN-inferences, abouta question’s insightfulness• In section 4, we introduce the Insight Gain Tensor µ(when, where, what, which)to select sound inferences, from the many valid normal-form inferences, andmeasures of insight gains associated to these questions• In section 5, we illustrate the use of SN-Logic, and we perform a validationtest, to show how SN-Logic/IQ-game helps finding a solution path, to acomponent of a hard real-world solved case (quantum field theory researchtopic)The Iterated Questioning or IQ game, is described in paper I. During a game ses-sion, the AI-agent, ASN, poses the human player H, a question q ∈ Q, it thinks ismost insightful, given H’s current cognitive mindset C(t). H then explores it, andreports if it was insightful. These are the game’s cooperative policies, both playersagree to adopt for each Q&A episode. The game serves several purposes whichbenefits both players (positive-sum game) [7,9]For the human player, H, the IQ-game has the following main roles:• The IQ-game is a Q&A process that reduces uncertainty and increases in-formation about a specific problem, via a sequence of Q&As. It provides aneffective tool, to gain insight on the many aspects of a complex challenge.• The IQ-game drives a sequential (mostly left-hemispheric) conscious reason-ing for solving well-defined (narrow) tasks. This process is mirrored by al-gorithmic AI. For complex tasks, this process alone fails to deliver full so-lutions.Conceptual solutions to such problems require the next process:insight-gaining.II.Two-Person Cooperative IQ-Gamea) IQ-game: Human player perspectiveinsight-gains (see paper I [8]): built by combining information, probability, uncer-tainty [6] and utility [7]. This paper is structured as follows:7. Von Neumann, J. and Morgenstern O. (1944) Theory of Games and EconomicBehavior, Princeton University Press: Princeton, NJ.Ref Boosting Human Insight by Cooperative AI: Foundations of Shannon-Neumann Logic• The IQ-game drives a parallel (mostly right-hemispheric) non-conscious pro-cess, for gaining insights leading to an ’aha’ moment. Largely non-consciousprocessing can be used, where the first process proves too slow or impossible(task is too broad, ill-defined and complex).• The IQ-game is driven by dual goals: minimizing obstacles and maximizingsolution qualities. The minimizing questions guide H to eliminate or reducedifficulties in the problem, when possible. The maximizing questions guide Hto boost specific solution qualities, when constraints allow it. It is a dynamicoptimization (changes with H’s understanding). We discuss this process insection 3.4.• The IQ-game provides a non-brittle reasoning framework, which continuouslyadapts to the human player H’s cognitive intentions C.This mindset Cevolves as H’s understanding of the challenge progresses. The IQ-game copeswith the framework drift problem (section 2.3).For the AI-agent, ASN, the IQ-game has these roles:• The IQ game produces game session episodes, from which the agent ASN canlearn via cooperative learning.• The IQ game ensures the agent remains human-aligned [10], because of thecontinuous human judgments. What is useful, informative, insightful for ahuman player H, does not necessarily mean the same for ASN, even if itstarts that way. In the learning process, these values can drift apart, dueto many factors. In the IQ game, human valuation is the ultimate arbiter,for the insight value of a question (since any AI short of a full AGI super-intelligence, will fail miserably at this task), while SN-Logic estimates theinsight values, given C(t).• The IQ game taps into a most valuable human resource:our collectiveevidence-based knowledge, undeniably our greatest accomplishment (culture,science, technology).Note that our collective belief-based human selections are often poor (e.g. whowe put in power as our leader). The forces here are complex and evolutionary: desirefor control, cognitive biases and herd mentality from the fear of social isolation(e.g. [11]).These factors are absent in the IQ procedure, since decisions are individual,and based directly on one’s own experience of a question’s insight, within a veryspecific cognitive context C(t). It uses direct evidence-based judgment, where H’smain incentive is to make life easier for herself. There are, of course individualvariations in the experienced insightfulness of questions, but only stable patterns(across many individuals) are retained in cooperative learning (not presented inthis paper).A complex challenge is typically time-evolving, multi-objective, multi-solution, multi-discipline, multi-level and open-ended, making it hard from the start, to clearlydefine a single problem, even when it is urgent (e.g. a crisis) or critical (e.g. sus-tainability), or both (e.g. a pandemic)Instead, there is a drift in the framing of problem and its solutions, as weaccumulate new insights about a challenge: a framework drift problem. The driftcannot be handled with a static AI/ML system, focused on a given narrow problem.The IQ-game, copes with the framework drift, by using an adaptive reasoningframework, and an adaptive cognitive intention C = {framework, where, when, what}1Year 20223© 2022 Global Journals Global Journal of Science Frontier Research Volume XXII Issue ersion I VIV ( F )b) IQ-game: AI player perspectivec) Framework drift problem10. Russell S. (2019) Human Compatible: Artificial Intelligence and the Problem of Control, Viking, New York.Ref Boosting Human Insight by Cooperative AI: Foundations of Shannon-Neumann Logic(section 3.3-3.4) which tracks the human player H’s current understanding of theconceptual framework.It follows H’s evolving understanding of the challenge,helping the SN-logic suggest the insightful questions, within each context C. TheIQ-game doesn’t define a problem from the start, but instead, let’s H describe theStandard Logic Programming (predicate logic) is very effective when making strictdeductions, but it cannot cope with the cooperative 2-person IQ-game. The purposeof SN-Logic is to provide an inference engine with the following requirements: ithas to be ...• precise (ambiguity-free) semantics axioms• consistent (contradiction-free) framework within which, all SN-inferences canbe made (normal-form inferencing)• transparent (natural language, no hidden layers)• explainable (no unjustifiable moves)• human-aligned (no conflicts of with human cognitive intentions)• non-brittle able to cope with fundamental concepts related to human-insight:causality (causes of insight), time-dependence (evolving understanding), in-formation, probability, uncertainty (Shannon), utility (von Neumann), andinsight (paper I). Brittleness is a common cause of AI failures.To satisfy these requirements, we need a consistent set of SN-Logic definitions,axioms and rules, to which we now turn.To reason using a predicate logic (such as SN-Logic), the variables x need spaces X,to scope the quantification: ∀x ∈ X, ∃x ∈ X. SN-Logic’s concepts are partitionedin six compact concept spaces, over which we can perform inferences (see appendicesA-F):Five vector spaces {T, SD, SC, SG, SS}, are used to describe the human playerH’s changing cognitive mindset C(t), during the IQ-game. The AI agent, ASN,needs to know C(t), because the insightfulness of a question, depends on H’s in-creasing understanding of the challenge and its possible solutions, as insight isaccumulated.The (tensor product) space SA, of possible conceptual actions (operation x ob-ject) provide the raw material to build conceptual solutions. © 2022 Global Journals1Year 20224 Global Journal of Science Frontier Research Volume XXII Issue ersion I VIV ( F )III.Predicate SN-Logica) SN-Logic requirementsb) SN-Logic SpacesNotes Boosting Human Insight by Cooperative AI: Foundations of Shannon-Neumann Logic• Vector space T of exploration stages: vector variable [when ∈ T] describethe current stage when of the exploration cycle.The vector [when]rotates in T over time (appendix A).• Vector space SD of mental obstacles: vector variable [where ∈ SD]describes where the human player’s H difficulties reside.The vector[where] rotates in SD over time while exploring the challenge (appendixB).• Vector space SC of difficulty causes: vector variable [what ∈ SC] de-scribes what in the reasoning’s framework, is causing H difficulty. Thevector [what] rotates in SC over time while exploring the challenge (ap-pendix C).• Vector space SG of mental goals: vector variable [where ∈ SG] describesthe solution quality, H intends to improve. The vector [where] rotatesin SG over time while exploring the challenge (appendix D)• Vector space SS of solution elements: vector variable [what ∈ SS] de-scribe what aspect of the solution, H intends to improve. The vector[what] rotates in SS over time while exploring the challenge (appendixE)• Tensor space of conceptual actions SA = Op× Ob: action variable[which ≡ action ∈ SA] is composed of a mental operation (verb ∈ Op) at-tached to a target object (noun ∈ Ob)). Space SA provides the building-blocks of conceptual solutions. (appendix F).SN-Logic’s role, is to provide guidance for insight-building via a Q&A process:suggesting when/where to pose which questions about what topic. To be used ininferences, the meanings of the parts of speech (variables {when, where, what,which}), and the sentence structure (questions which ≡ q ∈ Q), have to be bothconsistent and precise.ASN needs a basic grammar (syntax, semantics, vocabulary) to communicateeffectively with the human player H, in a consistent and precise manner.SN-Logic is based on four consistent (contradiction-free) axioms, to define its semanticsprecisely (ambiguity-free).Let the human-player H’s cognitive mindset C(framework, p) be defined bythe current reasoning framework (next section), and three (intention) parameters:p = {when = p1, where = p2, what = p3}, then:(Sem 1) Shannon-informative questions: a question (which) q(p, action),that reduces uncertainty (Shannon entropy) for H, who’s mindset isC(framework, p)(Sem 2) Neumann-useful questions:a question (which) q(p, action),that has a human-aligned (via the 2-person IQ-game) utility, withina mindset C(framework, p). It helps H make progress towards a solution.(Sem 3) SN-insightful questions: question (which) q(p, action) satisfying(Sem 1, Sem 2) is SN-insightful, within a mindset C(framework, p),otherwise it is SN-insightless.(Sem 4) SN-Valid inferences: an inference is SN-valid, if and only if it hasthe SN normal form (section 3.6)1Year 20225© 2022 Global Journals Global Journal of Science Frontier Research Volume XXII Issue ersion I VIV ( F )c) SN-Grammar: Axioms of SemanticsNotes Boosting Human Insight by Cooperative AI: Foundations of Shannon-Neumann LogicThese SN axioms of semantics, allow the AI to cope with core concepts of causality(causes of insight), dynamics (changing reasoning frames) information, probability,uncertainty [6], utility [7] and insight (paper I). These are necessary componentsof an insight-boosting AI. The axioms Sem1, Sem2 restrict the form of allowedquestions. This constraint is used by a Q-generator of questions q ∈ Q, to whichwe now turn.The cooperative IQ-game is driven by dual-objectives: to minimize the problem’scauses of difficulty, and to maximize the solution’s quality. The optimization mustcontinuously adapt to H’s understanding of the challenge, over an IQ-game session).The SN-grammar has a simple syntax, specified for each question class Q. Allquestions q ∈ Q will fall into two classes Q = {Qmin, Qmax}, from two comple-mentary (dual) perspectives: (a) causes of cognitive difficulty (to minimize), (b)qualities of solution (to maximize). Each question class generates many of specificquestions, aimed at making insight-gains.The purpose of SN-Logic is to incrementally boost our insightabout solutions, by suggesting when/where to pose whichtypes of questions about what topic, while adapting to a movingtarget: our current understanding the obstacles in a challengeThe question generator, or Q-gen, of difficulty-minimizing questions, uses a spe-cific syntax for an evolving cognitive mindset Cmin(frame, topic, p1, p2, p3). Thereis a lot of freedom in which questions to pose, even at a specific place and time,within a well-defined framework. We select a set of six commonly useful problem-solving questions, to illustrate the procedure.Q-Gen Syntax: difficulty-minimizing questions q(p, action) ∈ Qminqmin1: at what exploration stage are we in now? (specifies when = p1 ∈ T)qmin2: what reasoning frame are we operating in, now? (specifies [frame])qmin3: what topic in [frame] are we focusing on, now? (specifies [topic])qmin4: where does the main difficulty reside? (specifies where = p2 ∈ SD)qmin5: what, more specifically, causes this difficulty? (specifies what = p3 ∈ SC)qmin6: can you reduce the difficulty (where) and avoid its causes (what),by using these actions? (specifies action ∈ SA and which = qmin6 ∈ Qmin)The variable [action] (∈ SA ≡ OP × Ob), is a product [verb operation] (∈ Op)x [noun object] (∈ Ob) (appendices and section 5).The [frame] variable, labels the reasoning framework currently being used (e.g.a discipline, a subject, a specialty, a model, a system, a theory, a technology etc.).This framework can change from one exploration stage to the next. It is a movingtarget, which mirrors our current understanding of a complex challenge.The [topic] variable, labels a set of items we’re focusing on, within [frame] (e.g.agents, assumptions, bounds, properties, qualities, relations, statements, strategies,tactics, techniques etc.). Typically, [topic] is a tool we use within [frame], to makeprogress. For a concrete example, see section 5.Questions q ∈ Qmin are SN-insightful, only if they are SN-informative (axiomSem 1): they attempt to reduce a maximum possible amount of uncertainty (alter-natives, ignorance, options, possibilities), within the context Cmin.The generator of quality-maximizing questions, uses a specific syntax for anevolving cognitive mindset Cmax(frame, topic, p1, p2, p3): © 2022 Global Journals1Year 20226 Global Journal of Science Frontier Research Volume XXII Issue ersion I VIV ( F )d) SN-Grammar: Syntax for Dual-Optimization6. Shannon C. and Weaver W. (1949) The mathematical theory of communications,Univ. Illinois Press.Ref Boosting Human Insight by Cooperative AI: Foundations of Shannon-Neumann Logicqmax1: at what exploration stage are we in now? (specifies when = p1 ∈ T)qmax2: what reasoning frame are we operating in, now? (specifies [frame])qmax3: what topic in [frame] are we focusing on, now? (specifies [topic])qmax4: where do you need a boost (goal)? (specifies where = p2 ∈ SG)qmax5: what solution aspect, do you want to focus on? (specifieswhat = p3 ∈ SS)qmax6: can you boost your goal (where) and the solution’s quality (what),by using these actions? (specifies action ∈ SA and which = qmax6 ∈ Qmax)Questions in Qmax are SN-insightful, only if they are SN-informative (axiomSem 1): they attempt to reduce a maximum amount of uncertainty (alternatives,ignorance, options, possibilities), within the context Cmax. They are specificity-boosting questions which reduce uncertainty (Shannon entropy) to increase thesolution’s quality.The SN concept of insight involves notions in information, logic, probability, uncer-tainty and utility (see paper I). To cope with these, we need a logic with quantifiersfor scoping the variables x to specific spaces X.In standard predicate logic, apredicate is a function p of a variable x, which maps a variable x ∈ X, into thepredicate’s truth values {T, F} [12].X → {T, F} and x ∈ X → p(x) = T or FIn SN-Logic, an SN-predicate is a a function q of a variable x, which maps avariable x ∈ X, into the predicate’s insight values {insightful I+, insightless I0}.X → {I+, I0} and x ∈ X → q(x) = I+ or I0In SN-Logic we define the two classes (minimizing, maximizing) of predicatesq(x), the mindset parameter p ∈ P ≡ {when, where, what} and the predicatevariable ’cognitive action’:• SN-predicate questions q(p, action) ∈ Qmin, where p ∈ P, action ∈ SA• SN-predicate questions q(p, action) ∈ Qmax, where p ∈ P, action ∈ SAThe parameter p ∈ P is in the space P of cognitive mindsets Cmin(framework, p):the set of H’s intentions, during the IQ-game. The AI needs to know this intent, tomake useful cooperative suggestions. The mindset parameter p, encodes the typeof insight, H wants to boost, at any given time.SN-Logic only requires concept spaces ({T, SD, SC, SG, SQ, Op, Ob} of very smallsize N = Card(Space) ≈ 102 (see appendices).• Number of distinct cognitive mindsets: Ncogn = O(Card(P)) = O(Card(T)×Card(SD) × Card(SC) = 10 × 10 × 10 = 103• Number of possible conceptual actions: Nacts = O(Card(SA)) = O(Card(Op)×O(Card(Ob)) = 102 × 102 = 104• Number of possible distinct questions: Nques = Card(Q) = Ncogn × Nacts =107 minimizing questions, posed by the Qmin-generator (same for maximizingquestions).1Year 20227© 2022 Global Journals Global Journal of Science Frontier Research VolumeXXII Issue ersion I VIV ( F )Q-Gen Syntax: quality-maximizing questions q(p, action) ∈ Qmaxe) SN-Logic predicates q(x)f) SN-Logic Complexity & Scope12. Andrews P. B. (2002) An Introduction to Mathematical Logic and Type Theory: To Truth Through Proof, 2nd ed., Berlin: Kluwer Academic Pub. and Springer.Ref Boosting Human Insight by Cooperative AI: Foundations of Shannon-Neumann Logic• Number of combinations: Ncomb = 2Nques• Number of permutations: Nperm = Nques!Thus, the number of distinct classes of challenges SN-Logic can cope with, iseffectively infinite (N = 107!), yet, based on a few small, compact concept spaces(cardinality ≈ 102). In this sense, SN-Logic is economical (Occam’s razor).The computed complexity of SN-Logic is a theoretical upper bound, to determinethe scope of SN-Logic. In practice the computational cost will be much lower, due touniversal constraints (common to all challenge classes), because they are imposedby (mostly) challenge-independent forces:• causality: universal root causes of cognitive difficulties (e.g. confusion due toambiguity, indecision due to missing information) and solution quality (e.g.accuracy, adaptability)• logic: valid inferences with sound semantics• planning: logically necessary chronology of solution steps• problem-solving: universal tactics to minimize obstacles (to avoid/reduce),and maximize solution quality (to target/increase/maximize) (e.g. divide-and-conquer, minimize ambiguity, maximize order, simplify)• information: a question is only informative, if it reduces uncertainty by elimi-nating alternatives, options, outcomes, possibilities, within a cognitive mind-set (intention) C, restricting the insightful questions to a manageable subset:q ∈ Q∗(C) ⊂ Q, with Card(Q∗(C)) << Card(Q)• utility: a question is only useful, if it helps H, overcome obstacles, given acognitive intention C, restricting the insightful questions to a manageablesubset: q ∈ Q∗(C) ⊂ Q, with Card(Q∗(C)) << Card(Q)These rules impose a lot of structure on the SN-agent’s insight grain tensorµ(frame, topic, when, where, what, which), which is, in its fully general form, ahigh-dimensional rank-6 tensor, but is in practice, very sparse and decomposableinto simpler tensors and convolution kernels.The structure imposed by the universal (challenge class-independent) constraints,is sufficient to construct factored (’vanilla’) tensors µ∗ of much lower dimensionsand lower rank: knowledge acquisition. A ’flavor’ is then learned to fine-tune thetensors to each class of challenge, via cooperative learning (not described in this pa-per). Given the complexity upper-bounds of SN-Logic, the fine-tuning possibilitiesare vast.ASN’s fundamental problem, is to use the IQ-game, to guide a human player H,in when and where, to pose which types of questions about what topic, to gain amaximum amount of insight into a complex challenge.A standard normal form inferencing (analogous to conjunctive and disjunctivenormal forms, in digital and predicate logic), is necessary for the AI to cope with thecomputational complexity of SN-Logic. The AI can efficiently search for predicatevariables action ∈ SA, used as building-blocks for conceptual solutions. Given anevolving inferencing framework (frame, topic), SN-normal forms are the following: © 2022 Global Journals1Year 20228 Global Journal of Science Frontier Research Volume XXII Issue ersion I VIV ( F )These numbers already compare favorably to a typical human problem-solver H,working by herself. But the real power of SN-Logic (its scope of applications), comesfrom the combinatorial possibilities: the possible combinations and permutationsof insight-boosting questions, needed to solve each class of challenges:g) Symbolic AI (knowledge acquisition) vs Learningh) SN-Logic Normal FormNotes Boosting Human Insight by Cooperative AI: Foundations of Shannon-Neumann LogicSN normal-form for minimizing inferencesGiven a minimizing mindset Cmin(frame, topic, p), wherep ∈ P = {when, where, what}:if ∃ action ∈ SA, such that µmin(frame, topic, p, action) > µcrit,thenq(p, action) ∈ Q∗min(Cmin) ⊂ Qmin, andq(p, action) is SN-insightful, within CminSN normal-form for maximizing inferencesGiven a maximizing mindset Cmax(frame, topic, p), wherep ∈ P = {when, where, what}:if ∃ action ∈ SA, such that µmax(frame, topic, p, action) > µcrit,thenq(p, action) ∈ Q∗max(Cmax) ⊂ Qmax, andq(p, action) is SN-insightful, within CmaxThe sets Q∗(C), are maximum-insight subsets of Qmin or Qmax, and µ(frame,topic, p, action) is an insight-gain tensor (discussed shortly) whose insight gainsare above a minimum critical cutoff µcrit. The purpose of an insight-gain cutoffscale is intuitive, but its mathematical justification is outside the scope of thispaper, which focuses only on logical validity, and ignores scientific soundness. Thecutoff is related to a scale-invariance due to a conformal symmetry, under therenormalization of probabilities (unitarity). Scale-separation is used in quantumfield theories [13], but justified by the conformal symmetry [14] of a renormalizationgroup [15].To perform successful inferences autonomously, the AI agent needs to possessthe means of deciding whether a predicate variable action ∈ SA, leads to insightgains above a minimum lower bound (that is, action ∈ S∗A(C) ⊂ SA). The insight-gain tensor provides the SN-agent, the ability to select sound inferences, from avast number of merely, valid ones (that is, of SN normal-form).The AI performs SN normal-form inferences, to suggest insightful questions to ex-plore, given human-targeted insight gains C(p). These ’most insightful’ questions,lie in a restricted subspace Q∗(C) = {Q∗min(Cmin), Q∗max(Cmax)}, within a largespace Q, of possible questions (Card(Q) = 107). Given a current mindset C(p),ASN must find a subspace of questions Q∗(C). This is where an insight-gain mea-sure µ(p, action) (convolution tensors and their kernels, used to restrict searches tooptimal sub-spaces) are essential, to make sound inferences (real-world accurate),rather than merely valid ones (SN normal-form inferences). This will be presentedelsewhere. For now, we simply discuss general constraints imposed by SN-Logic,on the tensor elements.1Year 20229© 2022 Global Journals Global Journal of Science Frontier Research Volume XXII Issue ersion I VIV ( F )IV.Insight Gain Tensors μa) Need for Insight-Gain Tensors13. Dyson F., (1949) The radiation theories of Tomonaga, Schwinger and Feynman, Phys. Rev. 75, 486.Ref Boosting Human Insight by Cooperative AI: Foundations of Shannon-Neumann LogicThe AI’s capacity to generate SN-insightful I+ questions, from a vast possibil-ity of insightless I0 ones (with actions ∈ SA), resides in the structure a high-dimensional insight-gain tensor µ(when, where, what, which) ≡ µ(p, action), foreach challenge class and reasoning frame.So the full rank-7 tensor is actuallyµ(class, frame, topic, p1, p2, p3, action). This function outputs the value g of in-sight gain associated to exploring a question which ≡ q(p, action) ∈ Q, wherep ∈ P encodes H’s targeted insight gains. To be useful, the tensor µ is required tosatisfy the following properties:• µ : Cl × Fr × P × SA → [0, 1], where Cl = set of challenge classes,Fr = set of reasoning frameworks (frame+topic), P = T × S1 × S2,SA = Op × Ob, S1 = SD or SG, and S2 = SC or SQ• it is a measure of insight gain µ(class, frame, topic, p, action) = g ∈[0, 1] (normalized)• probability of all possible actions with a mindset p, must sum toone (unitarity)• µcrit ∈]0, 1[ (minimum critical insight-gain value µ > µcrit)• g = 0 when q(p, action) is SN-insightless I0 , given the mindset p• g = 1 when q(p, action) is maximally SN-insightful I+, given themindset p• µ is initialized by satisfying heuristics from causality, information,logic, planning, problem solving and utility. These constraints pro-vide the initial (challenge class-independent) approximation for µ• µ gets optimized (fine-tuned) for specific classes of challenges, bycooperative learning, using the IQ-game’s session episodes © 2022 Global Journals1Year 202210 Global Journal of Science Frontier Research Volume XXII Issue ersion I VIV ( F )b) Constraints on Insight-Gain Tensors μNotesWe can now illustrate how SN-Logic is used, on a real challenge. In the IQ-game,both players (human: H, ASN) agree to use simple cooperative strategies, given H’scurrent mindset C:(1) ASN suggests its guess at a most insightful question (q ∈ Q∗(C))(2) H reports questions q she actually finds insightfulThe game’s Q&A session, cycles over each obstacle, encountered within a chal-lenge. Hundreds of such sub-problems may be encountered, to solve a challenge.Usually, the number and nature of these obstacles is unknown ahead of time, inreal-world challenges.For clarity, we use a single, static, not so complex, yet most difficult challenge.The scenario is: a young post-doctoral researcher, H, is trying to find a good quan-tum field topic, to spend her next ten years on. The first few moves (Q&As) of thetwo-person IQ-game, could proceed as follows:Q from ASN: ’Greetings! What class of challenge are we exploring today (sam-ple which depends on what SN-Logic is being used for):V.Validation Test: Post-Doc Researcher’s Dilemma To compose somethinge.g. music, arts, literature, programmingTo compute somethinge.g. any domainTo construct somethinge.g. any domainTo design somethinge.g. engineering, technologyTo discover somethinge.g. science, mathematicsTo govern somethinge.g. crisis mitigation, leadershipTo invent somethinge.g. engineering, science, technologyTo manage somethinge.g. corporate, governmentTo optimize somethinge.g. any domainTo predict somethinge.g. any domainTo reconstruct somethinge.g. intelligence, inverse problems, sensingTo solve somethinge.g. any domainA by H: I want to improve on standard quantum field theory,its a discover class of challenge ([class] = discover).1. Q from ASN: Which exploration stage are we in, now:(AI is using qmin1 in Q-gen)Boosting Human Insight by Cooperative AI: Foundations of Shannon-Neumann Logicto specify a current obstacleto minimize the obstacleto explore solution ideasto question a solution ideato verify a solution ideaA by H: ’I want to identify the current obstacle’2. Q from ASN: ’What is our current reasoning framework?(AI is using qmin2 in Q-generator)The framework is composed of a topic and a frameNotesThe topic can be any useful tool we select, for overcoming the obstacle (selectthe closest match):actionse.g. activities or behaviorsagentse.g. catalysts or inhibitorslimitse.g. lower, upper, extremescomputationse.g. algorithmsequationse.g. model or representationlawse.g. laws of quantum physicsprocedurese.g. protocols or decisionprocessese.g. interactions or communicationspropertiese.g. pattern or symmetryqualitiese.g. strengths or weaknessesrelationshipse.g. hierarchy or prioritiesrestrictionse.g. constraints or conditionsrulese.g. allowed or forbiddenstatementse.g. assumptions, conditions or theoremsstatese.g. equilibrium or criticalitystrategiese.g. divide-and-conquerstructurese.g. classes, partitions, setstacticse.g. explore special casestechniquese.g. calculation or construction...1Year 202211© 2022 Global Journals Global Journal of Science Frontier Research Volume XXII Issue ersion I VIV ( F ) Boosting Human Insight by Cooperative AI: Foundations of Shannon-Neumann Logic © 2022 Global Journals1Year 202212 Global Journal of Science Frontier Research Volume XXII Issue ersion I VIV ( F )The reasoning frame is the clearly-defined context, within which topic is be-ing used (select the closest match):disciplinee.g. molecular biologysubjecte.g. protein foldingcontexte.g. social revolutionenvironmente.g. location and timeevente.g. activity or pandemic crisismodele.g. just-in-time supply-chainsprinciplee.g. quantum computingmethode.g. optimizationnetworke.g. communication or transporttheorye.g. general relativityspecialtye.g. programmingsysteme.g. quantum communicationstechnologye.g. fresh water extractor...Note SN-logic’s non-brittleness: at any given time, the reasoning frame canadapt to any required abstraction level and scope. Such frame changes aretypically unpredictable at the start of a real-world challenge.A by H: For my research direction, I want to identify a weakness in quantumfield theory (QFT) (so here, [frame] ≡ theory, and [topic] ≡ qualities).3. Q from ASN: ’What’s your main difficulty with the [frame], right now?(select the closest match)(AI is using qmin4 from Q-generator + an insight-gain tensor/kernel)inability to apply[topic] in/of [frame]inability to compute[topic] in/of [frame]inability to construct[topic] in/of [frame]inability to decide[topic] in/of [frame]inability to evaluate[topic] in/of [frame]Notesinability to exploit[topic] in/of [frame]inability to identify[topic] in/of [frame]inability to select[topic] in/of [frame]inability to simplify[topic] in/of [frame]inability to solve[topic] in/of [frame]inability to understand[topic] in/of [frame]A by H: I can’t evaluate the weaknesses [topic = qualities] of quantum fieldtheory [frame = theory]4. Q from ASN: ’More specifically, why can’t you evaluate the [topic] in [frame]?(select the closest root cause of the difficulty)(AI is using qmin5 from Q-generator + an insight-gain tensor/kernel)missing comparison for[topic] in/of [frame]missing constraint on[topic] in/of [frame]missing criterion for[topic] in/of [frame]missing direction in[topic] in/of [frame]missing information about[topic] in/of [frame]missing intuition for[topic] in/of [frame]missing knowledge of[topic] in/of [frame]missing metric for[topic] in/of [frame]missing ranking of[topic] in/of [frame]missing standard for[topic] in/of [frame]missing value of[topic] in/of [frame]A by H: ’I lack an intuition for the weaknesses of QFT’ Boosting Human Insight by Cooperative AI: Foundations of Shannon-Neumann Logic1Year 202213© 2022 Global Journals Global Journal of Science Frontier Research Volume XXII Issue ersion I VIV ( F )5. Q from ASN: awesome, so our current obstacle is, our missing intuition forthe weaknesses of QFT. Let’s try to eliminate this obstacle.A by H: ’Okay, I’m all ears!’6. Q from ASN: can we gain intuition to evaluate QFT’s weakness, by...(explore any question you think is promising, or move-on)(AI is using qmin6 in Q-generator + SN normal-form inferences + insight-gaintensors/kernels)by exploring:idealized cases of the theory (QFT)solved cases of the theory (QFT)simple cases of the theory (QFT)by outlining:consequences (causal) of the theory (QFT)implications (logical) of the theory (QFT)predictions (temporal) of the theory (QFT)tests (experimental) of the theory (QFT)by identifying:inconsistent aspects of the theory (QFT)limitations of the theory (QFT)problematic aspects of the theory (QFT)uncertain aspects of the theory (QFT)unjustified aspects the of theory (QFT)untested aspects of the theory (QFT)Notesby looking for:ambiguities (imprecision)contradictions (logical, evidence)counter-examples (exceptions)discrepancies (differences)dogma (cognitive traps)errors (math, procedures)falsehoods (logical)flaws (procedure, reasoning)gaps (missing pieces)implicit assumptions (reasoning)impossibilities (logical, physical)inaccuracies (scientific, technical)incompatibilities (between two items)inconsistencies (logical)limitations (scope of applicability)unexplained items (no explanation)unjustified items (lack justification)unsupported items (lack evidence)violations (law-breaking)weaknesses (logical)A from H: ’I find some questions quite insightful, because(click on each insightful one, and note the reasons for your record):I outlined the implications of QFT (e.g. including matrix unitarity), andQFT’s experimental tests (e.g.including neutron decay experiments).Ifound reported incompatibilities (e.g. known violations of the CKM matrix’sunitarity [16], in neutron decay experiments [17] ). That seems like an inter-esting research area of quantum field theory, for me. Boosting Human Insight by Cooperative AI: Foundations of Shannon-Neumann Logic © 2022 Global Journals1Year 202214 Global Journal of Science Frontier Research Volume XXII Issue ersion I VIV ( F )7. Q from ASN: ’Do you want to identify a new obstacle, now? ...Note: for a complex challenge, limitless combinations of obstacles can beexplored in this manner.This scenario shows how suggested questions from ASN, can replicate real-world solutions to obstacles, via a cooperative Q&A dialog.The researchers dosomething similar between themselves, early-on, to decide what to work on. ButAI’s complementary strength, is to cover many exploration paths, which are veryoften overlooked, yet may be key to quality solutions. This dynamic ’human-AI’interaction would be even more fruitful, in a group brainstorming session, whereeach member of the team, can select directions to explore and possible answers.We mentioned (section 3.7), that insight-gain convolution tensors and kernels, formthe bridge between the SN normal form inferencing (SN-validity), and measures ofinsight (SN-soundness); the bridge between logic (validity) and science (soundness).Initially, the tensors µ are the AI’s ’vanilla’ core, then, learned flavors are added toit, via machine learning to optimize the core AI, to distinct challenge classes.The AI’s core will be initialized by heuristics from causality, information, logic,planning, problem-solving, and utility. These apply to all types of challenges. Thetensors’ added flavor, needs to be learned using cooperative learning via a renormal-ization procedure, from the IQ-game’s episodes. The construction of the insight-gain tensors and cooperative learning will be described in future work.VI.Discussiona) Tensor Construction & Cooperative learning17. Zyla, P.A et al (2020) Review of Particle Physics: CKM quark-mixing matrix,Progress of Theoretical and Experimental Physics. 2020 (8): 083C01. doi: 10.1093/ptep/ptaa104.RefWe presented the foundations of SN-Logic, designed to boost human insight, tohelp overcome challenges that are hard to deal with, using traditional AI (mainly,predicate logic and deep learning neural nets). This required a logic, capable ofcoping with the concepts necessary to measure insight-gains: causality (causes ofinsight gains), dynamics (adaptive reasoning frameworks), information, probability,uncertainty (Shannon) and utility (von Neumann).In this paper, we presented the following:• The two-person (H, ASN) cooperative IQ-game’s role from both H’s andASN’s perspectives• The frame drift problem: coping with the changing understanding of a chal-lenge, using a (non-brittle) logic and optimization process, which continu-ously adapt to the current human understanding and intention• SN-Logic’s requirements to compute insightfulness (which involves causal-ity, information, logic, probability, uncertainty and utility) and the conceptspaces over which SN-Logic operates (to scope the quantifiers)• SN-Logic’s grammar: semantics + syntax for posing questions q ∈ Q froma vast space of potential questions. The syntax is used by a dual questiongenerator (q ∈ Qmin, q ∈ Qmax), from which all questions are built (Nques =O(107))• SN-Logic predicates of two question classes: problem difficulty-minimizing,and solution quality-maximizing, used in all inferences• The complexity of SN-Logic, and show it’s broad scope and capability ofcoping with a large number of distinct challenge classes.• The SN normal-form for making valid inferences, about a question’s insight-fulness, efficiently within a vast space of possibilitiesb) Conclusion Boosting Human Insight by Cooperative AI: Foundations of Shannon-Neumann Logic1Year 202215© 2022 Global Journals Global Journal of Science Frontier Research Volume XXII Issue ersion I VIV ( F )• Insight Gain Tensors µ(when, where, what, which) are necessary to selectsound inferences (real-world accurate), from a vast (effectively infinite) num-ber of valid ones (those with SN normal-form). µ measures the human in-sight gains, associated to questions posed, within their cognitive mindsets(Cmin, Cmax)• A validation test, to show that SN-Logic can replicate the solution steps, toa real-world solved case (discovery in quantum field theory)This paper focused solely on logic and validity of SN-inferences. It has notdealt with the equally important issue of scientific soundness and accuracy. Wewill present the construction of the insight-gain convolution tensors and kernels,and the learned structure (cooperative learning), in future papers.A: Vector Space of Exploration Steps T (sample)Time basis vector: when ≡ p1 ∈ T)to identify an obstacleto minimize the obstacleto explore solution ideasto question a solution ideato verify a solution ideaVII.AppendicesNotesB: Vector Space of Cognitive Difficulties SD (sample)Basis vectors of cognitive obstacles: where ≡ p2 ∈ SDinability to classify[frame]inability to compute[frame]inability to connect[frame]inability to construct[frame]inability to count[frame]inability to decide[frame]inability to design[frame]inability to eliminate[frame]inability to evaluate[frame]inability to exploit[frame]inability to extract[frame]inability to identify[frame]inability to interpret[frame]inability to organize[frame]inability to perform[frame]inability to plan[frame]inability to predict[frame]inability to rank[frame]inability to relate[frame]inability to select[frame]inability to simplify[frame]inability to solve[frame]inability to transform[frame]inability to verify[frame]etc. Boosting Human Insight by Cooperative AI: Foundations of Shannon-Neumann Logic © 2022 Global Journals1Year 202216 Global Journal of Science Frontier Research Volume XXII Issue ersion I VIV ( F )C: Vector Space of Difficulty Causes SC (sample)Basis vectors of causes: what ≡ p3 ∈ SClevel of abstraction of[item]level of ambiguity of[item]level of complexity of[item]level of dependencies in[item]level of flaws in[item]level of fragmentation of[item]level of implicitness in[item]level of impracticality of[item]level of imprecision of[item]level of incompleteness of[item]level of inconsistency in[item]level of indecision about[item]level of indetermination in[item]level of inefficiency of[item]level of insufficiency of[item]level of uncertainty in[item]level of unpredictability of[item]level of weakness of[item]etc.missing assumption about[item]missing bounds on[item]missing capacity for[item]missing classification of[item]missing confidence in[item]missing connections in[item]missing constraints on[item]missing evidence for[item]missing explanation for[item]missing freedom to[item]missing information about[item]missing interpretation of[item]missing intuition for[item]missing justification for[item]missing motivation for[item]missing organization of[item]missing representation of[item]missing restriction on[item]missing scales in[item]missing statements in[item]missing tools for[item]missing verification of[item]etc. Boosting Human Insight by Cooperative AI: Foundations of Shannon-Neumann Logic1Year 202217© 2022 Global Journals Global Journal of Science Frontier Research Volume XXII Issue ersion I VIV ( F )NotesD: Vector Space of Mental Goals SG (sample)Basis vectors of cognitive goals: where ≡ p2 ∈ SGclarity about the[solution item]confidence in the[solution item]construction of the[solution item]criticism of the[solution item]exploitation of the[solution item]imagination for the[solution item]intuition for the[solution item]understanding of the[solution item]etc.Note: mental goals [where] are intentions one tries to maximize, under constraints.The vector where ∈ SG rotates in SG, with the mindset C about the challenge.E: Vector Space of Solution Elements SS (sample)Basis vectors of solution elements: what ≡ p3 ∈ SSsolution’s agentssolution’s casessolution’s componentssolution’s consequencessolution’s constraintssolution’s dimensionssolution’s economysolution’s efficiencysolution’s effectivenesssolution’s ethicssolution’s formsolution’s frameworksolution’s informationsolution’s justificationsolution’s methodssolution’s plansolution’s propertiessolution’s qualitiessolution’s relationshipssolution’s requirementssolution’s resourcessolution’s restrictionssolution’s spacesolution’s statementssolution’s sustainabilitysolution’s utilitysolution’s valueetc. Boosting Human Insight by Cooperative AI: Foundations of Shannon-Neumann Logic © 2022 Global Journals1Year 202218 Global Journal of Science Frontier Research Volume XXII Issue ersion I VIV ( F )NotesF: Space of Actions SA = Op × Ob (tiny sample)Conceptual Action Space: operation ∈ Op × object ∈ ObActions to minimize indecision:avoiding, comparing, demanding, imposing, evaluating, excluding, justifying, maximizing,minimizing, optimizing, prioritizing, ranking, requiring, selecting, weighing items etc.Actions to minimize incomprehension:classifying, collecting, defining, explaining, exploring, exploiting, decomposing, grouping,imposing, interpreting, isolating, reconstructing, relating, removing, separating items etc.Actions to minimize inexperience:exploring cases, exploring examples, exploring idealisations, exploring simplifications etc.Actions to minimize skepticism:comparing, demanding, excluding, explaining, gathering, imposing, justifying, reasoning,refuting, rejecting, requiring, searching for, testing, verifying items etc.Actions to minimize unfamiliarity:building an analogy, building a model, defining concepts, looking for items, outlining factsActions to maximize ability:training to abstract, training to eliminate, training to exploit, training to organize,training to perform, training to relate, training to select, training to simplify,training to solve, training to transform etc.Actions to maximize clarity:classifying, connecting, defining, idealizing, ordering, organizing, outlining, reducing, relating,removing, separating, simplifying, summarizing items etc.Actions to maximize criticism:questioning an assumption, questioning a premise, questioning the framework,questioning a representation, questioning the necessity, questioning the sufficiency,questioning a method, questioning a path, questioning a solution, questioning the value etc.Actions to maximize exploitation:using an assumption, using a fact, using a given, using a constraint, using a property,using a relationship, using a restriction, using a statement, using a theorem etc.Actions to maximize imagination:weakening an assumption, weakening a bound, weakening a condition, weakening a constraint,weakening a requirement, weakening a restriction, weakening a rule, weakening a statement etc.Actions to maximize intuition:exploring an analogy, exploring a case, exploring an example, exploring a diagram,exploring a metaphor, exploring a model, exploring a story, exploring a simplification etc. NotesBoosting Human Insight by Cooperative AI: Foundations of Shannon-Neumann Logic1Year 202219© 2022 Global Journals Global Journal of Science Frontier Research Volume XXII Issue ersion I VIV ( F )References Références Referencias1. Leon Sterling L. and Ehud Shapiro E. (1986) The Art of Prolog: Advanced Programming Techniques (MIT Press Series in Logic Programming), MIT Press; First Ed., ISBN-10: 0262192500 ISBN-13: 978-02621925072. Mohri M. (2018) Foundations of Machine Learning, The MIT Press, 2nd Ed., Cambridge Massachusetts.3. Hornik, k. et al. (1989) Multilayer feed forward networks are universal approximators, Neural Networks, Vol. 2, Issue 5, Pages 359-366.4. Guilhoto L. F. (2018) An Overview of Artificial Neural Networks for Mathematicians, Univ. Chicago.5. Humphreys I. R. et al, (2021) Computed structures of core eukaryotic protein complexes, Science. DOI: 10.1126/science.abm48056. Shannon C. and Weaver W. (1949) The mathematical theory of communications, Univ. Illinois Press.7. Von Neumann, J. and Morgenstern O. (1944) Theory of Games and Economic Behavior, Princeton University Press: Princeton, NJ.8. Siregar E., (2021) Learning human insight by cooperative AI: Shannon-Neumann measure, IOP Publishing Ltd., SciNotes, Volume 2, Number 29. Nash J. (1953) Two-Person Cooperative Games, Econometrica, Vol. 21, No. 1 (Jan., 1953), pp. 128-140.10. Russell S. (2019) Human Compatible: Artificial Intelligence and the Problem of Control, Viking, New York.11. Fromm E. (1994) Escape from freedom, Holt Paperbacks; 1st Edition.12. Andrews P. B. (2002) An Introduction to Mathematical Logic and Type Theory: To Truth Through Proof, 2nd ed., Berlin: Kluwer Academic Pub. and Springer.13. Dyson F., (1949) The radiation theories of Tomonaga, Schwinger and Feynman, Phys. Rev. 75, 486.14. Guillarmou C., (2020) Conformal bootstrap in Liouville theory,arXiv:2005.11530v2 [math.PR] 11 Nov 2020.15. Wilson K. G., (1971) Renormalization Group and Critical Phenomena. I. Renormalization Group and the Kadanoff Scaling Picture, Phys. Rev. B 4, 3174.16. Wikipedia: Cabibbo–Kobayashi–Maskawa matrix, m, en.wikipedia.org/wiki/Cabibbo–Kobayashi–Maskawa matrix.17. Zyla, P.A et al (2020) Review of Particle Physics: CKM quark-mixing matrix, Progress of Theoretical and Experimental Physics. 2020 (8): 083C01. doi: 10.1093/ptep/ptaa104. | RESEARCH GATE |
PHYSICAL REVIEW RESEARCH 5, 013010 (2023)Quantum coherent control in pulsed waveguide optomechanicsJunyin Zhang,1,2 Changlong Zhu,1,3 Christian Wolff,4 and Birgit Stiller1,3,*1Max Planck Institute for the Science of Light, Staudtstr. 2, 91058 Erlangen, Germany2Department of Physics, University of Science and Technology of China, 230026 Hefei, China3Department of Physics, University of Erlangen-Nuremberg, Staudtstr. 7, 91058 Erlangen, Germany4Centre for Nano Optics, University of Southern Denmark, Campusvej 55, DK-5230 Odense M, Denmark(Received 31 March 2022; accepted 16 November 2022; published 11 January 2023)Coherent control of traveling acoustic excitations in a waveguide system is an interesting way to manipulateand transduce classical and quantum information. So far, these interactions, often based on optomechanicalresonators or Brillouin scattering, have been studied in the steady-state regime using continuous waves. However,waveguide experiments are often based on optical pump pulses, which require treatment in a dynamic framework.In this paper, we present an effective Hamiltonian formalism in the dynamic regime using optical pulses thatlinks waveguide optomechanics and cavity optomechanics, which can be used in the classical and quantumregime including quantum noise. Based on our formalism, a closed solution for coupled-mode equation under theundepleted assumption is provided and we found that the strong coupling regime is already accessible in currentBrillouin waveguides by using pulses. We further investigate several possible experiments within waveguideoptomechanics, including Brillouin-based coherent transfer, Brillouin cooling, and optoacoustic entanglement.DOI: 10.1103/PhysRevResearch.5.013010I. INTRODUCTIONPhotons are known as one of the most promising quantuminformation carriers in quantum communication, especiallyfor long distances [1] but also represent a major oppor-tunity for quantum computation [2]. However, enhancingphoton-photon coupling is a challenge. Introducing optome-chanical interaction is one of the possible ways to get photonsmore interactive and therefore mechanical systems have aprofound impact on current quantum technologies. The com-bination of matured MEMS (microelectromechanical system)[3] technology and the diversity of mechanical systems offerflexibility for transducing, delivering, and manipulation ofquantum information and moreover open new roads of ex-ploring macroscopic quantum phenomena [4–6]. In additionto the considerable effort invested in optomechanical res-onators [6,7], some research has been conducted in waveguideoptomechanics [8], which may be a plausible platform forquantum networks [9] and quantum nonlinearity [10] due toits broad bandwidth and integrability into existing circuitry[11–13].Waveguide optomechanics can rely on the interactionof optical waves with mechanical breathing modes of thetransverse section of the optical waveguide or on travelinglongitudinal acoustic waves or on hybrid versions of both ofthem. Brillouin scattering, which describes a variety of these*Corresponding address: birgit.stiller@mpl.mpg.dePublished by the American Physical Society under the terms of theCreative Commons Attribution 4.0 International license. Furtherdistribution of this work must maintain attribution to the author(s)and the published article’s title, journal citation, and DOI. Openaccess publication funded by the Max Planck Society.optoacoustic and optomechanical interactions has been exper-imentally investigated in detail in optical fibers and photonicintegrated waveguides [14,15] and several classical coupled-mode equations [16] have been analytically developed inthe classical regime to understand those phenomena. Alsoquantum approaches have been studied, including Hamilto-nian [17] and Lagrangian treatments [18], which propose aquantum field description for optomechanical interactions inwaveguides. Treating the waveguide as an array of cavitiesis another way towards optomechanical waveguide theories[19]. Current research on quantum regimes in an analyticalway, however, focuses on steady-state behaviors and treatcontinuous-wave (CW) interactions. Studying dynamic pro-cesses including optical pump pulses involve challenges suchas the contradiction between finite control length and theinfinite expanded nature of phonons. This makes it chal-lenging to analyze the time-dependent quantum evolution inoptoacoustic processes stimulated by a pulsed pump such asBrillouin-based memory [20–22].Compared with the well-studied zero-dimensional cavityoptomechanics, where the mechanical or acoustic excitationsare localized, the one-dimensional extended-waveguide sys-tems with traveling acoustic waves present several intriguingfeatures. Firstly, the continuous phonon spectrum in contin-uum optomechanical waveguides provides a remarkably largebandwidth for optomechanical interaction [23], which is anideal platform for multimode optomechanics. Furthermore,the resonator-free structure in extended waveguide systemsenables phase-preserving manipulation of the traveling waves[20,21,24]. Moreover, this one-dimensional optomechanicalinteraction offers new physics that was unseen in zero-dimensional cavities like polariton band structure [25,26] andtopological transport [27,28].In this paper, we formulate a framework in which backscat-tered Brillouin scattering in a waveguide is treated as a2643-1564/2023/5(1)/013010(17)013010-1Published by the American Physical SocietyZHANG, ZHU, WOLFF, AND STILLERPHYSICAL REVIEW RESEARCH 5, 013010 (2023)FIG. 1. Principle: The backward Brillouin interaction can beseparated in different wavevector channels. In each channel, thebackscattered nature contributes to a cavity-like localized interaction.continuous cavity array in the momentum space (Fig. 1). Withthis framework, many techniques well developed in cavity op-tomechanics can be applied to the continuum optomechanicalsystems.We introduce comoving coordinates that move togetherwith the optical pump pulse, which enables a simple formu-lation of the pulsed dynamical regime, in which propagatingoptical pulses interact with traveling acoustic waves, respec-tively. Under the assumption of sufficiently short undepletedpump pulses, the framework maps the dynamic pulsed case ina waveguide into a cavity-like system, which greatly simpli-fies the treatment of different scenarios of coherent control.We find that due to the larger depletion threshold for shorterpulses, the strong coupling regime can be achieved using cur-rent platforms in the pulsed regime. With this framework, weexplore several challenging problems in backward Brillouinscattering in waveguide systems, such as coherent transferbetween photons and phonons, Brillouin cooling, and entan-glement in Brillouin waveguide systems.First, we analytically demonstrate that by delicately con-trolling the pump pulse length, coherent transfer and Brillouincooling with high efficiently using the backward Brillouinprocess is possible. More specifically, we use an anti-StokesBrillouin process in the backward regime to show that thequantum state can be transferred between photons and acous-tic phonons. This transfer can be used for cooling the acousticphonons in longer fibers with higher efficiency than thosebased on continuous waves [8,29]. Note that previous paperswere mostly based on forwarding Brillouin scattering, whichis related to transverse mechanical vibrations and presents adifferent operation regime [8,29]. Secondly, we analyticallyshow that entangled pair generation is possible by using thebackward Brillouin scattering Stokes process. Our evalua-tion suggests that these regimes can be attained by existingwaveguide systems such as chalcogenide fibers and nanoscalewaveguides.The paper is organized as follows: In Sec. II we brieflysummarize the conventional backward Brillouin interactionand then present the effective Hamiltonian formalism, whichis the main result of this paper. In Sec. III, we investigate threechallenging problems in waveguide optomechanics using ourformalism: coherent transfer, Brillouin cooling, and entangledpair generation. In the last Sec. IV, we summarize our resultFIG. 2. The phase-matching diagram. The black corn and blackcurve refer to the dispersion relation of photons and acousticphonons. The black-dotted line is the copy of the phonon dispersionrelation. For the Stokes process, the phase-matching condition canonly be satisfied when the pump photon emits a forward travelingphonon (the green arrow pointing to the top right corner). In contrast,the anti-Stokes process can only be stimulated when absorbing abackward traveling phonon (the green arrow pointing to the top leftcorner).and discuss several open questions in waveguide optomechan-ics.II. EFFECTIVE HAMILTONIAN FORMULATIONA. Waveguide optomechanical systemWe consider an optomechanical waveguide system, whichallows the guidance of both electromagnetic and acousticwaves with different wave vectors and in different spatialmodes. A typical optomechanical interaction in such a waveg-uide system treats a mechanical oscillation with frequency�(k), a light field with optical frequency ω(k), and theoptomechanical coupling g0, which refers to the couplingbetween two photons with wave vectors kS, kp (p, q) and oneacoustic phonon with wave vector q(k). The optomechanicalcoupling can originate from different physical processes suchas electrostriction [30] and radiation pressure [31]. Consider-ing the three-wave-mixing optomechanical coupling (usuallythe dominant ones), the system can be described by the Hamil-tonian [17,19]H =� +∞−∞dq ¯hω(q)a†qaq +� +∞−∞dq ¯h�(q)b†qbq+ ¯h�� +∞−∞dqd p (gp,qa†p+qapbq + H.c.)(1)where ak and bk are annihilation operators of the electromag-netic and mechanical modes. For the interactions within thenarrow frequency band of interest, the coupling factor gp,q canbe approximated by a coupling constant g0.This Hamiltonian can be derived from combining elastictheory and Maxwell’s equations by introducing an optome-chanical coupling as the interaction part [17]. To treat thisquantum mechanically, it is then quantized on the normalmodes. The first two/three terms are the energy for photonsand acoustic phonons (the free Hamiltonian part H0) and thelast two terms are the interaction Hamiltonian HI.The optomechanical interactions in the waveguide are con-strained by the phase-matching condition. As shown in Fig. 2,there are two points [30] where the phase-matching condition013010-2QUANTUM COHERENT CONTROL IN PULSED WAVEGUIDE …PHYSICAL REVIEW RESEARCH 5, 013010 (2023)is satisfied when the waveguide is pumped by an optical fieldwith wavevector kp and frequency ω(kp). These two pointsrefer to the Stokes process/anti-Stokes process with photonwavevector ks/as, photon frequency ω(ks/as), phonon wavevec-tor qs/as, and phonon frequency �(qs/as). Since the two kindsof phonons corresponding to the Stokes and anti-Stokes pro-cess are separated in directions, we can selectively applythe Stokes process or anti-Stokes process on the directionaltraveling phonons by choosing the pump pulse direction.The wave package operators are defined by integrating thewavevectors around the phase-matching points [17],ap/as/s(z,t) =1√2π� +∞−∞dk× ake−i(k−kp/as/s)zeiω(kp/as/s)t ,bas/s(z,t) =1√2π� +∞−∞dq× bqe−i(q−qs/as)zeiω(qs/as)t ,(2)where the first equation corresponds to the optical waves andthe second equation to the acoustic wave. The fiber is placedon the z axis in the laboratory frame and the pump wave isassumed to be propagating along the positive z axis. It can beverified that the wave package operators and the correspond-ing Hermite conjugates are well-defined quantum operatorsthat preserve the commutation relations. Utilizing the wavepackage operators, the interaction part HI can be written asthe products in both Stokes process (HI,S) and the anti-Stokes(HI,AS) process,HI,s = g0 ¯h� +∞−∞dz a†p(z)as(z)bs(z) + H.c. ,HI,as = g0 ¯h� +∞−∞dz ap(z)a†as(z)bas(z) + H.c. .(3)From the Heisenberg equation of the optomechanicalHamiltonian, we can derive the approximated motion equa-tion for ap/s/as(z,t) and bas/s(z,t), which are the coupled-mode equations for the stimulated Brillouin scattering pro-cess. Here, we write the coupled-mode equations for thebackward Brillouin process, which will be the important pro-cess throughout this paper.For the Stokes process, the equations are∂tap + cg∂zap = −ig0asbs − γ /2 ap ,∂tas − cg∂zas = −ig0apb†s − γ /2 as ,∂tbs + ug∂zbs = −ig0apa†s − �/2 bs +√�ξ .(4)For the anti-Stokes process, the equations are∂tap + cg∂zap = −ig0aasb†as − γ /2 ap ,∂taas − cg∂zaas = −ig0apbas − γ /2 aas ,∂tbas − ug∂zbas = −ig0a†pbas − �/2 bas +√�ξ .(5)The ap, as, aas refer to the optical wave packets for thepump wave, Stokes wave, and the anti-Stokes wave respec-tively. The bs and bas refer to the acoustic wave packets relatedto the two processes. The optical excitations and the acousticexcitations travel in the fiber in different group velocities de-scribed by cg, ug and suffer a dissipation rate with γ , �. Sincethe acoustic excitations travel much slower than the opticalones, therefore we can omit this effect by setting ug = 0 inthe following discussion. The thermal noise of the acousticfield is taken into consideration with the Langevin term√�ξ.ξ = ξ(z,t) obeys the relations⟨ξ(z1,t1)ξ †(z2,t2)⟩ = nthδ(z1 − z2,t1 − t2),[ξ(z1,t1), ξ †(z2,t2)] = δ(z1 − z2,t1 − t2).(6)nth is the averaged thermal phonon number at the giventemperaturenth =1eϵp/kBTE − 1 ,(7)where ϵp is the energy of a single phonon, kB is the Boltzmannconstant, and TE is the temperature of the environment.The coupled-mode equation above can be used for nu-merical simulations. However, finding an analytical solutionfor the dynamical cases of optical and acoustic pulses andinvestigating analytically quantum phenomena are both notstraightforward with this set of equations. In the followingsections, we will show that the coupled mode equations canbe exactly solved under the undepleted assumption.B. The undepleted assumptionIn the following discussions, we consider the undepletedcase. The undepleted assumption refers to the condition wherethe waveform function of the pump light ap(z,t) remainsunchanged during the scattering process. This assumption isvalid for quantum Brillouin experiments, where the amplitudeof the quantum-level acoustic field and backscattered field likeStokes and anti-Stokes wave are too small to deplete the pumpsignificantly.Definingg(z,t) = g0⟨ap(z,t)⟩ ;(8)since the pump waveform ap(z,t) remained unchanged duringthe propagation, the g(z,t) would remained unchanged too,ap(z,t) = ap(0,t − z/cg) ⇐⇒ g(z,t) = g(0,t − z/cg) . (9)Under the undepleted assumption, the first equation in thecoupled mode equations, which refers to the pump dynamics,can therefore be omitted. For example, the coupled modeequation for the Stokes process can be linearized as∂tas − cg∂zas = −igb†s − γ /2 as ,∂tbs = −iga†s − �/2 bs +√�ξ .(10)Under the undepleted assumption, the equations are nowlinearized. The effective coupling strength g(z,t) describesthe coupling between the acoustic and optical fields, whichis tunable by changing the pump power. This tunable cou-pling strength enables us to control the acoustic phononsand photons traveling in the waveguide system coherently.Furthermore, as discussed later, we will show that the strongcoupling conditions g > � are possible by applying strongshort pump pulses under the pulse Brillouin threshold. The013010-3ZHANG, ZHU, WOLFF, AND STILLERPHYSICAL REVIEW RESEARCH 5, 013010 (2023)FIG. 3. The region of interest in backscattered Brillouin scatter-ing problems. Besides noise, three boundary conditions need to bespecified for both Stokes process and the anti-Stokes process so thatthe coupled mode equation can be mathematically complete: (1) theinitial phonon states at t = 0. (2) the input pump waveform at z = 0.(3) the input backscattered waveform at z = L.detailed strong coupling condition and the pulsed thresholdare discussed later in Sec. II F 4.C. Boundary value problemConsidering a fiber of length L, we assume that there isno light in the fiber at t = 0 and all the light waves areinput at either the end at z = 0 or the end at z = L of thefiber at t > 0. To determine to whole evolution of all wavepackets functions in the fiber at t > 0, 0 < z < L, we need toknow (1) the initial acoustic states bs/as(z,t = 0); (2) the inputpump waveform ap(z = 0,t), when linearized, this terms canbe fully described by g(z = 0,t); (3) the input backscatteredwaveform as/as(z = L,t); and (4) the detailed form of thenoise function ˜ξ(z,t). If we only care about the statisticalresult, the detailed form of the noise function is not needed.The scattering process can be described in a space-timediagram like the following, in this space-time diagram, theabove requirements all appear at the boundaries, as shown inFig. 3.D. Exact solution under undepleted assumptionBy defining the coordinates transformationη = t − z/cg ,τ = t ,(11)the distribution coupling strength g(z,t) can be directlyrelated to the boundary conditions g(z = 0,t) = g0⟨ap(z =0,t)⟩,g(z,t) = g(0,t − z/cg) = g(η) .(12)Take the Stokes process as an example, the linearized coupledmode equations for the Stokes process can be further writtenas (the derivation is detailed in Appendix A)(∂τ + 2∂η)as = −igb†s − γ /2 as ,(∂τ + ∂η)b†s = igas − �/2 b†s +√�ξ † .(13)Since g only depends on η, the variables in the above equa-tion can be separated by performing a Fourier transformationon variable τ,˜as(�, η) = cgL� L/cg0dτ as(τ, η)e−icg�τ ,˜b†s(�, η) = cgL� L/cg0dτ b†s(τ, η)e−icg�τ ,˜ξ †(�, η) = cgL� L/cg0dτ ˜ξ †(τ, η)e−icg�τ .(14)Then the equation can be written as the following Langevinform:∂η�˜as˜b†s�=�− γ +2icg�4− ig2ig− 2icg�+�2��˜as˜b†s�+�0√�˜ξ †�.(15)It has to be noted that the closed solution for anti-Stokesprocess can be obtained in the same way. For the anti-Stokesprocess, a similar Langevin form can be obtained,∂η�˜aas˜bas�=�− γ +2icg�4− ig2−ig− 2icg�+�2��˜aas˜bas�+� 0√�˜ξ�.(16)The matrices in the above equations can be made Hermitianby variable substitution: ˜A =√22 ˜a, ˜B = ˜b. The initial phononstates and backward laser injections (Stokes and anti-Stokespart) are included in the initial conditions at η = 0. The aboveequations, describing the evolution of each Fourier component�, are similar to the Langevin equation in optomechanicscavities. The only difference is the time evolution in optome-chanics cavities is now replaced by the η evolution on acopropagating framework. Each Fourier component with off-resonance variable � can be written with the correspondingequation of η evolution, and the equations are independentof each other. This separability indicates a new viewpointof waveguide optomechanics that by separating the interac-tions to different frequency/wavevector channels, cavity-likebehavior can be recovered. The detail of this similarity isdepicted in Fig. 1 and is discussed in detail in Sec. II E.The Langevin equations above are exactly solvable, whichmeans that once we obtained the boundary conditions at η =0, then the exact solutions can be obtained in the solvableregion as shown in Fig. 4. As with the conventional Langevinequations, a Green’s function, also called the time-evolutionoperator, can be introduced to obtained the exact solution. TheLangevin equation in a general matrix form readsd⇀M (η)dη= P(η)⇀M(η) +⇀R(η).(17)We now introduce the Green’s functionG(η2, η1) = T�exp� η2η1dυ P(υ)�.(18)013010-4QUANTUM COHERENT CONTROL IN PULSED WAVEGUIDE …PHYSICAL REVIEW RESEARCH 5, 013010 (2023)FIG. 4. The exact solvable region. Under the undepleted assump-tions, separation of variables could be applied to the coupled modeequation after variable transformation (z,t) → (τ, η). In this case,the exact solution of the scattering process can be obtained in the graytriangle region shown in this figure. To obtain the exact solution inthis region, two boundary conditions are needed: (1) the phonon andbackscattered waveform at η = 0, as/as(τ, η = 0), bs/as(τ, η = 0).(2) The coupling strength at every η, g(η) = g0ap(0,t = η).The T {. . . } is the time-ordering operator to ensures the ex-ponential is time ordered, which in context means η ordered:any product of P(υ) that occurs in the expansion of the expo-nential must be ordered such that the value of υ is increasingfrom right to left of the product. The solution can be writtenas⇀M(η) = G(η, 0)⇀M(0) +� η0dυ G(η, υ)⇀R(υ) .(19)As shown in Eq. (8), the pump laser can be fully describedas a modification of the effective coupling. Under the un-depleted assumption, the pump pulse travels along the fiberwithout changing its waveform. Therefore all points on thefiber experience the same effective coupling waveform. In ourformalism [Eqs. (15) and (16)], this can be fully described byg(z,t) = g(0,t − z/cg) = g0⟨ap(0,t − z/cg)⟩ using the pulsewaveform at the input port. Therefore we can claim that allundepleted situation has been solved.To demonstrate the formalism developed here, we considera rectangular pump shape, whereg(η) = g�(η)�(T − η) .(20)In this case, the solution can be written as (when t ⩽ T )⇀M (η) = G(η)⇀M (0) +� η0dυ G(η − υ)⇀R (υ)(21)withG(η) = exp(Pη) .(22)E. Physics interpretation: Comoving conditionsIn this section, we show that the physics behind thevariable-separated equations [Eqs. (15) and (16)] reveals asimilarity between waveguide systems and cavity systems:The interaction can be separated into cavity-like interactionsin different wavevector channels.We choose the Stokes interaction as an example. Considerthe initial state at t = 0, which can be written asas(z,t = 0) =�kAkeikzb†s(z,t = 0) =�kB†keikz .(23)When the noise and dissipation are omitted and there is nopump light in the waveguide at t = 0, this initial state leads toa solution at z = cgt (which corresponds to η = 0 in the η − τcoordinates) and yieldsas(z = cgt,t) = as(z + cgt, 0) =�kAke2ikcgt,b†s(z = cgt,t) = b†s(z, 0) =�kB†keikcgt.(24)Applying Eq. (14) results inas(τ, η = 0) =��˜as(�, η = 0)eicg�τ,b†s(τ, η = 0) =��˜b†s(�, η = 0)eicg�τ,(25)with˜as(�, η = 0) = A�/2,˜b†s(�, η = 0) = B†�.(26)From Eq. (26) we can see that the physics interpretation of˜as(�, η) and ˜b†s(�, η) in Eq. (14) corresponds to the ampli-tude in the wavevector picture. Therefore, the separation of theoriginal coupled mode equation Eq. (4) [Eq. (5)] into Eq. (15)[Eq. (16)] indicates the interaction in waveguides can beseparated into cavity-like interactions in different wavevectorchannels, by only pairing each optical wave components withthe acoustic wave components with twice the wavevector. Oneof the main differences between optomechanical waveguidesand optomechanical cavities is the Hilbert space. In optome-chanical resonators, the phonon states and the photon statesare discrete. In waveguides, the phonon and photon spectrumis continuous. Therefore, we have to consider a spectrum-dependent interaction in the optomechanical waveguide. Inour case, the interaction is only significant near the phasematching point, and � here is the wavevector deviation fromthe phase matching point.Furthermore, for a special initial plane wave initial state asas(z,t = 0) = Ak0/2eik0z/2,b†s(z,t = 0) = B†k0eik0z,(27)only � = k0 would lead to nonzero ˜as and ˜b†s. Therefore, forevery η, we have∀δτ > 0,as(η, τ + δτ ) = as(η, τ )eicgk0δτ,b†s(η, τ + δτ ) = b†s(η, τ )eicgk0δτ.(28)013010-5ZHANG, ZHU, WOLFF, AND STILLERPHYSICAL REVIEW RESEARCH 5, 013010 (2023)FIG. 5. The physics behind the solutions: Comoving conditions.If we omit the noise and depletion, a comoving condition Eq. (28)can be derived once the plane-wave-like initial conditions in Eq. (27)is satisfied. (a) In the z − t coordinates, or the laboratory frame, boththe backscattered light and the acoustic waveform (in terms of am-plitudes) are comoving with the pump. (b) In the η − τ coordinates,all the waveforms are constant along the τ axis. The phase factoris omitted in both (a) and (b). This comoving condition means thatin the comoving frame of the pump pulse, the waveform of bothacoustic and backscattered waves will remain unchanged except fora phase factor ei�z. This makes the nonlocalized interaction processin the waveguides similar to the localized interaction in the opticalcavities, and this is the physics reason why we can treat waveguidesusing similar methods in the cavities.Converting this back to the z − t laboratory frame, we have anontrivial observation∀δt > 0,as(z + cgδt,t + δt) = as(z,t)eicgk0δt,b†s(z + cgδt,t + δt) = b†s(z,t)eicgk0δt.(29)This condition means that the waveform of both the acous-tic field and the backscattered optical field remain unchangedin the comoving frame along with the propagating pump ex-cept for an additional phase factor eicqτ, as shown in Fig. 5.This comoving frame in a special case � = 0 is also dis-cussed in Ref. [32]. This turns the nonlocalized interactionin waveguides into a cavity-like localized interaction in thecomoving frame. Figure 1 depicts the consequence of this as-sumption: The whole scattering process in the waveguide canbe separated into different frequency channels. In each of thechannels, translational invariances hold, and that enables usto build a mathematical framework to connect the waveguideoptomechanics and the cavity optomechanics.This result indicates that the pump pulse length can beused to control the evolution time for phonons and photons inwaveguides. Exactly controlling the interaction time enablesus to control the optoacoustic interaction coherently, whichmakes the dynamic regime more interesting for coherent con-trol applications, in contrast to the steady-state regime. In thefollowing sections, we will discuss some predictions derivedfrom this formalism.F. Waveguide optomechanics in the perspective of cavityoptomechanics: With backward Brillouin scattering as anexample1. The beam-splitter-like and down-conversion-like interactionThe Brillouin interaction Hamiltonian is described inEq. (3). In the case of linearizing the interaction usingundepleted pump assumption, different phase matching condi-tions lead to different effective Hamiltonians. The interactionStokes Hamiltonian HI,S describes a down-conversion-likeprocess between the Stokes photons and the phonons, whichcorresponds to the blue-detuned regime in cavity optomechan-ics. The anti-Stokes part HI,AS describes a beam-splitter-likeprocess, which is similar to the red-detuned regime in cavityoptomechanics [6].In terms of cavity optomechanics, the beam-splitter-likeinteraction (the anti-Stokes process here) describes a statetransfer between the anti-Stokes photons and phonons [33].Such conversion process is the Rabi oscillation and can beused to achieve coherent transfer. The area dependency rulein Brillouin memory is exactly the result of area dependencyin Rabi oscillation [34], as explained later.As for this down-conversion-like interaction, i.e., theStokes process is a parametric amplification process [4], thescattered Stokes field can grow much more quickly than theanti-Stokes field when the pump power is high enough, whichcauses asymmetric Stokes and anti-Stokes sidebands.2. The Brillouin gain and the strong coupling regimeIn this section, we relate our approach to commonly usedexperimental parameters in backward stimulated Brillouinscattering (SBS) experiments.In SBS, the Stokes process is dominant. We consider thesteady state in which both ∂as∂t and ∂b∂t equal to zero in Eq. (4)and introduce the acoustic dissipation rate �. Then we get∂∂z (a†SaS) = −4g20a†pap�cga†SaS .(30)In SBS generated by a continuous pump, it holds ∂∂z (a†Sas) =−GPa†SaS. The G refers to the Brillouin gain whose unit is[m−1W −1] and the P refers to the pump power, whose unit is[W ]. Therefore the effective gain we introduce can be obtainedfrom the pump power directly [16]|g| =�GP�cg4.(31)A dimensionless effective coupling ratio |g|/� can be intro-duced by utilizing Eq. (31). Coherent control is only possiblewhen the effective coupling ratio is larger than one,|g|� =�GPcg4�> 1 .(32)Generally the strong coupling regime in cavity optomechan-ics denotes that the intensity of the optomechanical couplingstrength exceeds the dissipation rates of the optical and me-chanical modes [6]. For the backward Brillouin scattering intypical waveguides, as the acoustic dissipation generally ex-ceeds the optical dissipation, the system enters into the strongcoupling regime when the coupling strength becomes larger013010-6QUANTUM COHERENT CONTROL IN PULSED WAVEGUIDE …PHYSICAL REVIEW RESEARCH 5, 013010 (2023)than the acoustic dissipation rate, i.e., |g| > �. As shownin Eq. (16), the Rabi period is inverse proportional to |g|.Therefore the strong coupling regime here can be interpretedas a longer phonon lifetime than the Rabi period.In reported waveguide systems, when pump pulse power is1W, a coupling ratio as high as |g|/� = 8.3 can be achievedin the integrated chalcogenide waveguides [35]. Other po-tential platform to demonstrate the predictions in this paperare nanowaveguides [36,37] where radiation-pressure-basedBrillouin gain enhancement was proven by carefully engi-neering the geometry and fiber systems, such as chalcogenideand photonic crystal fibers with reported large Brillouin gaincoefficients [38,39]. Moreover, due to the reduced phonondissipation and enhanced Brillouin gain at low temperatures[40,41], demonstration of this strong coupling regime couldbe possible on more platforms in a cryogenic environment.3. Area dependencySBS can be used to coherently transfer informationfrom the optical domain to acoustic waves. This con-cept has been experimentally shown as Brillouin memory[20,21]. In Brillouin memory, the write/read efficiency at-tains a maximum when the effective coupling area �(T ) =� T0 dt g0�|a†p(t)ap(t)| satisfies the following area dependencyequation [20,21,34,42]:� = (2n + 1)√2π2(n ∈ Z) .(33)This result can be recovered by our formalism in a straight-forward way. The readout process in Brillouin memoryexperiment is the anti-Stokes process, which is described byEq. (16). We consider the perfect phase-matching case where� = 0 and omitting the dissipation �, γ = 0 as in Ref. [34]. Ifthe system is driven by a rectangular pump pulse as describedin Eq. (20), the matrix elements of the propagator in Eq. (22)that describes photon-phonon transfer reads12|G21(η > T )| = |G12(η > T )|=�����sin�√22 gT������=�����sin�√22 �(T )������ .(34)The |G21| and |G12| attains maxima if and only if the areadependency in Eq. (33) is satisfied.4. The undepleted conditionOur formalism is built based on the undepleted pump ap-proximation. When the pump power is strong enough and thepulse length is sufficiently long, the pump power might besignificantly depleted by the Stokes process. The anti-Stokesprocess is much weaker than the Stokes process, and thereforeit is enough for us to only consider the Stokes process.For a short pulse length in the Stokes process, the sec-ond term in the solution of Langevin equation (21) can beomitted, since the first term in Stokes process described byEq. (15) has an exponential growing term, which significantlysurpasses the second term [32]. From a physical point of view,the Stokes process is a stimulated amplification process. Theamplification of the initial state fluctuation is going to be muchgreater than the additional noise added during the evolutionaryprocess,⟨˜a†S(�, η)˜aS(�, η)⟩ ≈ |G12(�, η)|2nth .(35)The nth is the average thermal phonon number, nth ≈kBTE/(¯h�). The undeplected condition requires that theStokes power is much smaller than the pump power, IS ≪ IP.For a rectangular pulse with pulse length T as defined inEq. (20), the requirement can then be simplified to√22 gT ≪ 14 ln32π2Ip�2G�cgk2BT 2E ω2 .(36)The detailed derivation can be found in Appendix B. Theleft-hand side of Eq. (36) refers to the Rabi area, whichequals to 2π for a complete Rabi period. As an example,for a chalcogenide waveguide [35] pumped by 1W at roomtemperature, the right-hand side of Eq. (36) is 10.27 > 2π,which means that for those pulses within the first Rabi period,the undepleted condition holds. This result also implies thatincreasing the pump power while decreasing the pulse lengthmakes the undepleted assumption more robust while keep-ing the effective coupling area unchanged. The significantincrease of depletion threshold using short pulses, which isalso the Brillouin threshold, is also predicted and verifiedexperimentally in Ref. [32].III. APPLICATIONS OF COHERENT CONTROLIn this section, we will use the techniques developed in theprevious section to discuss coherent transfer, cooling, and en-tangled pair generation in backward Brillouin scattering. Forsimplicity, we only consider the case where the waveform ofthe pump light is a rectangular wave, although the method weproposed earlier is not only applicable to rectangular waves.A. Coherent transferThe most critical task of optomechanical systems is themanipulation of phonon states. How to store information intophonons, read out the phonon states, and convert them intomeasurable physical quantities therefore becomes an impor-tant issue. Coherent transfer has been demonstrated in cavityoptomechanics [33,43,44], here we show that a similar for-malism can be applied to waveguide optomechanics in thepulsed regime.The main idea to implement phonon readout is to use theanti-Stokes process: The beam-splitter-like Hamiltonian of theanti-Stokes process describes a Rabi oscillation between pho-tons and phonons. Thus, coherent transfer between photonsand phonons is possible by controlling the Rabi oscillation.The coherent transfer process can be illustrated using thespace-time diagram in Fig. 6. The classical Brillouin coherentreadout has been demonstrated experimentally on photonicchips [21]. As shown in the previous sections, due to thelack of resonating structures in optomechanical waveguides,013010-7ZHANG, ZHU, WOLFF, AND STILLERPHYSICAL REVIEW RESEARCH 5, 013010 (2023)FIG. 6. Illustration of the coherent transfer and cooling process.The anti-Stokes interaction HI,AS describes a Rabi-oscillation be-tween the anti-Stokes photons and phonons. Thus a π/2 Rabi pulsecan be used to transfer the phonon into photons coherently (viceversa), and once the phonons are transferred into photons, coolingis attained.we must consider the entire continuous phonon spectrum toderive the spectrum dependent coherent-transfer efficiency.The spectrum dependent read out efficiency β(�, η) isdefined as the following: Assume that the phonon number atwavevector � at the beginning of the readout process t = 0is Nb,c(�, η) = ⟨˜b†as(�, 0)˜bas(�, 0)⟩, which is the coherentpart we want to readout. After applying the coherent readoutprocess, the phonons will be converted into photons, becauseof the existence of thermal noise, such readout process isonly partially coherent, NA(�, η) = Na,c(�, η) + Na,n(�, η),where the Na,c(�, η) = ⟨˜aas(�, η)†⟩⟨˜aas(�, η)⟩ is the coher-ent part and Na,n(�, η) = ⟨˜aas(�, η)†˜aas(�, η)⟩ − Na,c(�, η)is the noise part.Therefore, the readout efficiency could be divided into twoparts, i.e., the coherent readout efficiency and the thermalnoise term. The coherent readout efficiency corresponding tothe coherent transfer from phonons to photons can be givenbyβC(�, η) = Na,c(�, η)Nb,c,(37)and the thermal noise-induced part can be expressed asβN(�, η) = Na,n(�, η)nth,(38)where we assume that the acoustic dissipation rate � is muchlarger than the optical dissipation rate γ . Thus the total outputphoton number after the readout process can be given byNA(�, η) = βC(�, η)Nb,c + βN(�, η)nth,(39)where the first term corresponds to the coherently transferredphotons while the second term belongs to the noise-inducedphotons.FIG. 7. The Brillouin coherent transfer. (a) Shows how the co-herent transfer coefficient βC (solid line) and the noise coefficientβN(dotted line) changes as the pulse length increases. As the pumplength increases, the coherent part oscillates, and the body of thephonon is dominated by the noise part when the pulse length is long.(b) The transfer efficiency difference βC − βN was calculated usingg = 8.3� (Chalcogenide chip [35] with pump power equals 1 W). Aclear oscillating behavior due to the Rabi nature can be seen. Optimalcoherent transfer can be attained when the difference between βCand βN is largest. (c) The transfer efficiency difference βC − βNat different coupling strengths, when pump length equals the π/2Rabi pulse t = π/√2g. As the coupling ratio g/� increases, both thetransfer efficiency and the effective transfer spectrum bandwidth willincrease.Figure 7(a) presents how the coherent part βC and thenoise part βN changes as pulse length increases, at the perfectphase-matched case � = 0. Due to the beam-splitter natureof the anti-Stokes process, the coherent part oscillates whilethe maximum is attained at the first peak due to the combina-tion of Rabi-like oscillation and thermal dissipation. However,the incoherent noise term accumulates and dominates later.The result shows that effective coherent transfer can only bepossible when the π/2 Rabi pulse is shorter compared to thephonon lifetime.Solving the Langevin equation using propagator matrix Gin Eq. (22), the readout photon is˜aas(�, η) = G11(�, η)˜aas(�, 0) + G12(�, η)˜bas(�, 0)+√�� η0dνG12(�, η − ν) ˜ξ(�, ν).(40)Since we consider that there is no anti-Stokes light inthe waveguide at the t = 0 moment in the readout process,˜aas(�, 0) = 0. The spectrum dependent readout efficiency013010-8QUANTUM COHERENT CONTROL IN PULSED WAVEGUIDE …PHYSICAL REVIEW RESEARCH 5, 013010 (2023)βC,N(�, η) for a rectangular pump pulse with length η readsβC(�, η) = |G12(�, η)|2βN(�, η) = �� η0dν |G12(�, ν)|2 ,(41)whereG12 = −ie− 14 �∗e,asη−icg�ηg√2ge,assin�√22 ge,asη�.(42)The resonance-modified effective coupling strength ge and theeffective acoustic dissipation is defined asge,as =�g2 − (� + icg�)2/8 ,�e,as = � + icg� .(43)The detailed derivation of G12 can be found in Appendix C.We present the numeric result for βC(�, η) − βN(�, η) inFigs. 7(b) and 7(c) to illustrate the competing effects betweenthe coherent part βC and the noise-induced part βN of thephonon-photon transfer efficiency. As shown in Fig. 7(b),the readout efficiency oscillates as the pulse length increases,which agrees with the area dependency law in Brillouinmemory [42]. As the pulse length increases, noise phonondominates and the coherent readout becomes inefficient. Theoptimal coherent readout can be achieved whenη ≈π√2g.(44)This is exactly the π/2 Rabi pulse length slightly modifieddue to the dissipation effect. Figure 7(c) shows the read-out efficiency difference βC − βN calculated at η = π/(√2g),which is the π/2 Rabi pulse length. It can be seen that bothlarger readout bandwidth and higher maximum readout effi-ciency can be achieved by increasing the coupling strengthg. Figures 7(a) and 7(b) show oscillating behaviors along thewave vector �, which is related to the higher harmonics com-ponents in the Fourier transformation of rectangular waves.Moreover, as depicted in Eq. (41) and Figs. 7(b) and 7(c),we can see that the transferred anti-Stokes wave includes pho-tons with a continuous band of accessible states without therequirements of optical or acoustic resonators. Thus comparedwith the state transfer in cavity optomechanics [33,43,44],which operates with single optical and mechanical modes, thestate transfer generated by Brillouin interaction in waveguideshas a broader bandwidth, which is important for applicationsin photonic information processing, quantum computing andquantum information.B. Brillouin coolingMany quantum experiments can only be carried out at lowtemperatures (passive cooling) in order to reduce the deco-herence effect introduced by thermal noise. Laser cooling isone of the most promising active cooling techniques [6] toreduce thermal noise in a specific frequency band. Recently,laser cooling induced by anti-Stokes Brillouin scattering,i.e., Brillouin cooling, has been explored in optomechanicalwaveguides [8,29] by utilizing forward Brillouin scatter-ing where phonons experience lower damping than photons.However, Brillouin cooling generated by backward Brillouinscattering where the acoustic dissipation exceeds the opti-cal dissipation in typical Brillouin-active waveguides is stilllargely unexplored. In this section, we propose a coolingmechanism by using pulses rather than continuous waves. Weclaim that this cooling mechanism might attain higher coolingefficiency and can be used in long fibers.Conventional laser cooling is based on increasing the effec-tive dissipation of phonons by using a damping laser. Here wepropose an alternative cooling mechanism based on coherenttransfer. As shown in the previous section, the anti-Stokesprocess can be regarded as a photon-phonon Rabi oscillationusing the linearized effective Hamiltonian. The main idea ofcoherent transfer-based cooling is to use a carefully designedlaser pulse to convert phonons to photons through the phonon-photon Rabi oscillations. In this section, we will show howto use our formalism to investigate the Brillouin cooling inwaveguides. Such Brillouin cooling approach in waveguidesin the pulse regime is similar to the dynamical cooling methodin cavity optomechanical systems [45]. The convenience inexperiment of this dynamical Brillouin cooling approach inwaveguides will prompt the elegant transient cooling tech-nique to become an effective experimental tool in continuumoptomechanical systems. Our laser pulses at the appropriatepulse length will avoid reverse conversion, thus leaving fewerphonons in the waveguide system. By using a pulsed pump,the pump depletion effect due to the Stokes process can alsobe avoided. The cooling process is illustrated using space-timediagram in Fig. 6.The cooling effect can be quantified by counting theremained phonons. We introduce the spectrum remainedphonon rates κ(�, η),κ(�, η) = ⟨˜b†(�, η)˜b(�, η)⟩⟨˜b†(�, 0)˜b(�, 0)⟩= |G22(�, η)|2 + �� η0dν |G22(�, ν)|2= κc + κn,(45)which is defined as the ratio of the cooled phonon occupationat η to the initial thermal phonon occupancy (nth). The smallerthe κ(�, η), the better cooling effect is achieved.Equation (45) consists of two terms: the first term κcdescribes how fast the coherent transfer/cooling is able toconvert the phonons into photons (which are then removedfrom the system); the second term κn describes how fast thephonons are restored due to the interaction with the thermalenvironment, which is the thermal noise term that obstructingthe cooling of the system. Figure 8 shows the two terms andthe whole κ(� = 0,t) at the perfect phase-matching point� = 0. One can clearly see the Rabi oscillation behavior ofthe coherent part, which contributes to cooling, and the ac-cumulating thermal noise counterbalances the cooling effect,which eliminates the cooling effect when the pulse lengthincreases.In conventional fiber optic systems, the dissipation rateof the optical channel is much smaller than the dissipationrate of the acoustic channel, γ ≪ �. With this approximation,013010-9ZHANG, ZHU, WOLFF, AND STILLERPHYSICAL REVIEW RESEARCH 5, 013010 (2023)FIG. 8. The Brillouin cooling. (a) The remained phonon consistsof two parts: the coherent part determined by the initial state and thenoise generated part. Optimal cooling can be achieved by choosingappropriate pump length to be π/2 Rabi period. (b) The transferefficiency was calculated using g = 8.3� (Chalcogenide chip [35]with pump power equals 1 W). A clear oscillating behavior due tothe Rabi nature can be seen. (c) The cooling effect on different pa-rameters, when pump length equals the π/2 Rabi pulse η = π/√2g.As the coupling strength increases, the effective cooling bandwidthwill increase.spectrum remained phonon rate readsκ(�, η) ≈ 1 −�����e− �e,asη28g2e,as + �2e,as8g2e,assin2�ge,asη√2������ . (46)The ge,as and �e,as is defined in Eq. (43). The detailedderivation for the above result can be found in the AppendixC. Equation (46) shows that choosingη = (2n + 1) π√2g(n ∈ Z) ,(47)leads to the minimum κ(�, η), which is equivalent to themaximum cooling efficiency. Due to the accumulating thermalnoise, minimum κ(� = 0, η) can be achieved near the firstRabi π/2 pulse at η ≈ π/(√2g), this is the same as the maxi-mum coherent transfer efficiency shown in Eq. (44), which isalso based on the same photon-phonon Rabi oscillation.In Figs. 8(b) and 8(c), we present the calculated κ(�, η)for different coupling strength and pulse length by assumingrectangular pump pulse. Due to the decoherence effect ofthermal noise, the optimal cooling pulse is the π2 Rabi pulse,which refers to n = 1 case in Eq. (47). From Fig. 8(c), onecan clearly see that the cooling bandwidth becomes wider forstronger coupling, which indicates a wider interaction band-width. We claim that this has the same mathematical rootsas the general power broadening effect in all atomic systemsFIG. 9. The remained phonon spectrum after a π/2 coolingpulse. This figure is obtained by the first-principle simulation usingthe algorithm from Ref. [46,47]. In the simulation, a chalcogenidewaveguide [35] with length L = 1 m is assumed. A π/2 Rabi pulsewith pump power Ip = 10 W, pulse length t =π√2g is used as thecooling pulse. The phonon spectrum is calculated when the pumppulse leaves the fiber at t = L/cg. The shaded area indicates the 3σerror, and the theoretically predicted result is shown in the lowerright corner. By applying a rectangular cooling pulse, the remainingphonon spectrum is sinc like, and around 50% of the initial phononnumber at the central wave vector has been removed.[48]. The thermal noise can be represented by zero-meanspace-time white noise [46] in the simulation of the cou-pled mode equation [47]. We simulate the pulse propagationprocess in a chalcogenide waveguide [35] and calculate thephonon spectrum after one cooling pulse passes. The resultis shown in Fig. 9, which is consistent with our analyticalexpressions.C. EntanglementQuantum entanglement is one of the fundamental buildingblocks of today’s quantum technologies, especially quan-tum communication. The generation of entangled quantumpairs is the basis of quantum state teleportation and quantumrepeaters. Classical information networks based on opticalfibers are one of the most promising infrastructures for futurequantum teleportation, which means that generating entangledpairs in an all-fiber system is a fruitful challenge. Apart fromthe application-based perspective, achieving the generationof entangled pairs is also one of the vital experiments todemonstrate the ability to do quantum experiments in optome-chanical waveguides.In this section, we show that the photon-phonon entangledpair generation can be achieved by utilizing the Stokes pro-cess, and the entangled photon-phonon pairs can be furthertransformed into photon-photon entangled pairs by the co-herent transfer technique described in Sec. III A, as shown013010-10QUANTUM COHERENT CONTROL IN PULSED WAVEGUIDE …PHYSICAL REVIEW RESEARCH 5, 013010 (2023)FIG. 10. Illustration of the entangled pair generation process.First, a blue-detuned pump pulse (S pump) should be applied to gen-erate the entangled phonon-photon pairs via the Stokes interactionHI,s. Then, a red-detuned pump pulse (AS pump) should be appliedto stimulate the coherent transfer process via anti-Stokes interactionHI,as described in the previous sections to readout the phonon intoa photon. By combining these two processes, the measurable entan-gled photons pairs (Backscattered-1 and Backscattered-2) could begenerated.in terms of the space-time diagram in Fig. 10. Choosing theStokes process to generate photon-phonon pairs is motivatedby the down-conversion nature of the Stokes process. In theStokes process, a higher energy photon is annihilated, produc-ing a lower energy phonon and a lower frequency photon. Inthis process, both momentum and energy conservation mustbe satisfied, which leads to the phase-matching condition. Asa direct result of this phase-matching condition, there is someshared information between the produced phonon and photon,which leads to quantum entanglement.We introduce the quadrature operators for the acoustic fieldand the optical fieldXa = a + a†√2,Ya = ia† − a√2,Xb = b + b†√2,Yb = ib† − b√2.(48)To prove the existence of quantum entanglement, we useDuan’s two-mode entanglement criterion [49]. The main ideaof the Duan’s criterion is to choose two EPR variables andcalculate the variance sum. It can be shown that once thecorrelation variance is less than a specific quantum limit, thedensity matrix of the two quantum modes cannot be separatedby any means, thus leading to quantum entanglement.The two mode EPR variables we choose areu = 1α Xa + αYb , v = 1αYa + αXb ,(49)with α = 2−1/4. This particular α value is obtained by tryingto minimize the EPR variance. The Duan’s entanglement cri-terion [49] for it isσ 2EPR = σ 2u + σ 2vα2 + α−2 < 1 .(50)After some calculations, the expression of σ 2EPR readsσ 2EPR = 13|√2G11 + iG21|2 + 2n0 + 13|√2G12 + iG22|2+ 2nth + 13�� η0dν|√2G12 + iG22|2≈�1 + 2n03�|G(η)|2 + 2nth + 13�� η0dν |G(ν)|2 ,(51)whereG(η) = e− 14 �∗e,sη�e− ge,sη√2 −�e,s2√2ge,ssinh�ge,sη√2��.(52)The detailed calculation can be found in Appendix D. The n0is the phonon number expectation at the initial state η = 0,and the nth is the thermal phonon expectation number de-termined by the temperature of the environment. When thesystem is cooled by the cooling technique we proposed inthe previous sections, n0 < nth can be achieved. The off-resonance effective coupling and the effective dissipation forthe Stokes process shown here are slightly different from whatwe defined in the anti-Stokes process in Eq. (43),ge,s =�g2 + (� + icg�)2/8 ,�e,s = � + icg� .(53)The G(η) consists of two competing terms. The first termshows an exponential depressing of the EPR variance withrate ge,s while the sinh-like second term will refer to the deco-herence effect that will destroy such entanglement. Thereforethe photon-phonon entangle can be achieved by choosingadequate pump length in the Stokes process.We present the numerical result for the photon-phononentangled EPR variance in Fig. 11. The dashed lines inFigs. 11(b) and 11(c) refer to the quantum nonseparationlimit σ 2EPR = 1. The generation of photon-phonon entangle-ment pairs is based on a down-conversion-like Stokes processrather than a beam-splitter-like anti-Stokes process. There-fore, unlike the coherent transport and cooling discussed inthe previous sections, there is no Rabi oscillation behavior. Asshown in the Fig. 11(a), the longer pulse will induce a strongerentangled effect while decreasing the entangled bandwidth,and stronger coupling can still broaden the entangled band-width. In practice, the optimal entangled pulse length shouldbe specified by taking the measurement bandwidth of theexperiment setup into consideration: The photon-phonon en-tangled pair cannot be measured directly, the coherent transferprocess is needed to transfer the entangled phonon into a pho-ton, in which the measurement bandwidth can be controlledby manipulating the pump waveform as discussed in previoussections. We show in Fig. 11 the results of photon-phononentangled pair generation obtained by direct computation withthe noisy Brillouin simulation algorithm [46]. The simulationcombines photon-phonon entangled pair generation with the013010-11ZHANG, ZHU, WOLFF, AND STILLERPHYSICAL REVIEW RESEARCH 5, 013010 (2023)FIG. 11. The Brillouin entanglement. (a1) The quadrature distri-bution in different � is shown using algorithm from Ref. [46,47].The simulation parameter is the same as Fig. 9. X1, X2,Y1,Y2 arethe quadrature components of the two backscattered pulses. Nonuni-form distribution of quadratures can be seen when |�| < �, whichimplies shared information between two scattered photons. When|�| is large, the Brillouin interaction strength is limited by thephase-matching condition, so the two-mode correlation betweenquadratures gradually disappears. The discretization along the fre-quency axis (cg�) is the result of the Fourier transform of the finitenumerical simulation time. (a2)–(a4) is the transection of (a1) at thethree points of wavevector |�| from larger to smaller. [(b),(c)] TheEPR variance σ 2EPR. As the coupling increases, the effective entangledspectrum bandwidth will increases. (b) is calculated using g = 8.3�(Chalcogenide chip [35] with pump power equals 1W).coherent transfer, which is directly computed for photon-phonon entangled pairs.Actually, the thermal noise of the acoustic field cancause a destructive effect on the optomechanical entangle-ment in the long time evolution of the system, as shownin Fig. 11(b). However, different from the case discussedin Ref. [50,51] in cavity optomechanical systems, whichconsiders the weak coupling regime, we generate the op-tomechanical entanglement in the strong coupling regimesimilar to the optical entanglement in optomechanical systemsdiscussed in Ref. [52,53]. This strong Brillouin optome-chanical interaction in the pulsed regime induces an extrasuppression to the destructive effect of thermal noise, whichenables the robust optomechanical entanglement at a highenvironmental temperature in continuum optomechanicalsystems. In addition to that, Brillouin optomechanics op-erates in the several GHz frequency range, which has theadvantage of a lower thermal phonon number from thestart.IV. CONCLUSIONS AND PERSPECTIVESIn this paper, we proposed a formulation to treat trav-eling phonons and traveling photons in a optomechanicalwaveguide system as an array of multiple optomechanicalcavities. The possibility of realizing coherent control usingthis formulation has been shown: Both the effective couplingstrength and the interaction time can be controlled easily bymanipulating the shape of the pump pulse. By applying theformulation, we show that it is possible to achieve activeBrillouin cooling through the backward Brillouin scatteringprocess and quantum entangled pair generation in waveguidesystems. Experiments based on this formalism are shown to beachievable using current technology and fabrication of opticalfibers and integrated waveguides.Our paper mainly focuses on the backward scattering op-tomechanical interaction and most of the calculations aredone using the assumption that the pump pulse is almostnondepleted. The undepleted assumption leads to a linearizedHamiltonian directly, which is similar to down-conversionfor the Stokes process and similar to a beam-splitter for theanti-Stokes process. The available quantum operation timeapproximately equals the ratio between the coupling strengthand the dissipation rate, Nop = g/�. There are two routes, in-cluding the dissipation reduction and coupling enhancement,towards a broader range of quantum applications within op-tomechanical waveguides. In order to reduce the dissipationrate, efforts have to be paid in designing waveguides that canlocalize and trop acoustic phonons with high efficiency. In or-der to increase the coupling strength, one method is to increasethe material-dependent coupling strength g0 by optimizing thewaveguide structure or using more promising materials [54].Another method to increase the coupling strength is to usehigher pump powers. In this case, the pump light may besignificantly depleted, so shorter light pulses are needed tomeet the requirements of coherent control. Therefore, the un-depleted regime within the short pulses regime (below 100 ps)need to be considered in future works.Because all Brillouin backscattering processes under unde-pleted conditions can be solved exactly based on the methodof this paper, future efforts should be spent on detailed mod-eling of the Brillouin memory process, where informationis coherently stored in acoustic waves [20,21]. In particular,the method can be the base for a model including quantumnoise and a study on how coherent information is destroyedby different noise contributions.The data and source codes for all the figures in this paperare available as Supplemental Material [55].013010-12QUANTUM COHERENT CONTROL IN PULSED WAVEGUIDE …PHYSICAL REVIEW RESEARCH 5, 013010 (2023)ACKNOWLEDGMENTSAuthors acknowledge funding from the Max PlanckGesellschaft through an independent Max Planck ResearchGroup, CW acknowledges funding from the Villum Fondenrelated to Grant No. 16498.APPENDIX A: DETAILED DETIVATION OF THECOORDINATES TRANSFORMATIONIn this Appendix, we detailed coordinate transformationfor linearized Stokes process, from Eq. (10) to Eq. (13). Thederivation for the anti-Stokes process can be obtained in thesame way.The linearized coupled mode equation for Stokes processin Eq. (10) reads∂tas − cg∂zas = − igb†s − γ /2 as,∂tbs = − iga†s − �/2 bs +√�ξ.(A1)Since the pump pulse propagates along the fiber in a constantvelocity cg without changing the shape, the pump-related cou-pling term g(z,t) satisfies the comoving conditiong(z,t) = g(0,t − z/cg).(A2)To fully take advantage of this comoving relationship, weintroduced the coordinate transformation as in Eq. (11),η = t − z/cg ,τ = t .(A3)In this transformation, the derivative operators are related by∂t = ∂η∂t ∂η + ∂τ∂t ∂τ,∂z = ∂η∂z ∂η + ∂τ∂z ∂τ,(A4)which is�∂z∂t�=�−1/cg011��∂η∂τ�.(A5)By substituting ∂z and ∂t in Eq. (10), we get Eq. (13),(∂τ + 2∂η)as = −igb†s − γ /2 as ,(∂τ + ∂η)b†s = igas − �/2 b†s +√�ξ † .(A6).APPENDIX B: DETAILED DERIVATION OF THEUNDEPLETED CONDITIONSThe backscattered Stokes power is the integral of photondensitiesIS =� +∞−∞d� ¯hωcg2π ⟨˜a†s (�)˜as(�)⟩ .(B1)Using the technique we developed, we have⟨˜a†s (�, η)˜as(�, η)⟩ = |G12(�, η)|2⟨˜b†s(�, 0)˜bs(�, 0)⟩+ nth�� t0dν|G12(�, ν)|2 .(B2)For the system driven by short pulses η ∼ 1/g ≪ 1/� instrong coupling regime, the second term, which is the noisecontribution can be omitted. The average thermal phononnumber at t = 0 reads⟨˜b†s(�, 0)˜bs(�, 0)⟩ =1exp¯h�kBTE − 1 ≈ kBTE¯h� .(B3)Therefore the spectrum density of Stokes photon reads⟨˜a†s (�, η)˜as(�, η)⟩ ≈ kBTE¯h� |G12(�, η)|2 .(B4)For g ≪ �, the following approximation holds:G12(�, η) ≈ 2gsinh 14�8g2 − c2�2t�8g2 − c2�2.(B5)The peak of G12(�, η) isG12(� = 0, η) =i sinh gη√2√2≈ iegη/√22√2.(B6)The width of the central peak can be approximated by solving8g2 − c2�2 = 0,Wd. ≈ 2 × 2√2gc.(B7)Finally we haveIS ≈� +∞−∞d� ¯hωcg2πkBTE¯h� |G21(�, η)|2≈ ¯hωcg4πkBTE¯h� (Wd. × |G12(�=0, η)|2)=e√2gηgkBTEω2√2π�.(B8)The undepleted condition is satisfied if and only if the bakscat-tered Stokes power is much smaller than the pump power,IS ≪ IP. Since the effective coupling g is related to the pumppower IP byg =�GIP�cg4,(B9)where the G is the Brillouin gain with unit [m−1W−1], theabove undepleted condition can be simplified togη ≪12√2ln32π2Ip�2G�cgk2BT 2E ω2 .(B10)APPENDIX C: DETAILED DERIVATION FOR BRILLOUINCOOLINGIn this section, we present the detailed derivation processfor Brillouin cooling. Because the phonons in the anti-Stokesprocess and Stokes process are separated by their oppositetravel direction, we can only consider the anti-Stokes process,which describes the photon-phonon transfer.The phenomenologically introduced effective Hamiltonianfor anti-Stokes process reads [Eq.(16)]∂∂η�˜aas(�, η)˜bas(�, η)�=�−icg�/2 − γ /4−ig/2−ig−icg� − �/2�×�˜aas(�, η)˜bas(�, η)�+� 0√�˜ξ�.(C1)013010-13ZHANG, ZHU, WOLFF, AND STILLERPHYSICAL REVIEW RESEARCH 5, 013010 (2023)The cg is the group velocity of the optical field near thephase matching point. The g is the effective coupling strengthenhanced by the pump power, and nth is the averaged thermalphonon number. The relation between the effective couplingand the well-known Brillouin coupling strength G is [Eq. (31)]g =�GP�cg4.(C2)This can be obtained by calculating the steady-state behaviorof the coupled mode equation, as shown in the main contextof this paper. The ˜ξ describes the thermal noise, the thermalbehavior of phonon determines both the initial state and thenoise term, which could be described by a Wigner process�˜b†as(�1, η1)˜bas(�2, η2)�= nthδ(�1 − �2)δ(η1 − η2) ,�˜ξ †(�1, η1)˜ξ(�2, η2)�= nthδ(�1 − �2)δ(η1 − η2) .(C3).The equation is a Langevin equation. Therefore wecan solve the equation using the conventional method forLangevin equations, which is the undetermined coefficientmethod. Considering the rectangular pump wave, we intro-duce the P matrix asP =�−icg�/2 − γ /4−ig/2−ig−icg� − �/2�.(C4)For a Langevin equation in the formddη⇀M = P⇀M +⇀R.(C5)The solution reads⇀M(η) = exp(Pη)⇀M(0) +� η0exp[P(η − ν)]⇀R(ν)dν .(C6)Therefore we need to calculate the matrix exponential of theP matrix. The matrix exponential can be calculated by usingthe formulaeSDS−1 = SeDS−1 .(C7)Where the Jordan decomposition is usedP = SDS−1 .(C8)The S is the similar matrix, and D is the Jordan matrix. Inour case, D is diagonalized. By introducing the small opticaldissipation approximation(� ± γ ) ≈ � .(C9)Under those approximations the matrix exponential can beobtained,G = exp(Pη) .(C10)The matrix elements readG11=e− 14 �∗e,asη−icg�η�cos�ge,asη√2�+�e,as2√2ge,assin�ge,asη√2��,G12 = −ie− 14 �∗e,asη−icg�ηg√2ge,assin�ge,asη√2�,G21 = −ie− 14 �∗e,asη−icg�ηg√2ge,assin�ge,asη√2�,G22=e− 14 �∗e,asη−icg�η�cos�ge,asη√2�−�e,as2√2ge,assin�ge,asη√2��,(C11)wherege,as =�g2 − (� + icg�)2/8 ,�e,as = � + icg� .(C12)The remained phonon spectrum density can be explainedasκ(�, η) = ⟨˜b†as(�, η)˜bas(�, η)⟩⟨˜b†as(�, 0)˜bas(�, 0)⟩= |G22(�, η)|2 + �� t0|G22(�, ν)|dν. (C13)For the resonance case, when � = 0, we haveκ(� = 0, η) = 1 − e− �e,asη28g2e,as + �28g2e,assin2�ge,asη√2�.(C14)As an approximation when g ≫ �, g ≫ |c�|, this resultcan be extended to the general case by taking the norm andreplacing � with �e,as,κ(�, η) = 1 −�����e− �e,asη28g2e,as + �2e,as8g2e,assin2�ge,asη√2������ .(C15)APPENDIX D: DETAILED DERIVATION FOR BRILLOUINENTANGLEMENTIn this Appendix, we present the derivation processof the Brillouin interaction based optomechanics entangle-ment. The entangled pair generation can be achieved bydown-conversion in quantum optics. In the optomechanicswaveguide systems, the Hamiltonian for Stokes process alsohas a down-conversion like form, the only difference is thestates it acts on which are one photon and one phonon,therefore the entangled pair it generated is a photon-phononentangled pair.The phenomenologically introduced effective Hamiltonianfor anti-Stokes process reads [Eq.(15)]∂∂η�˜as(�, η)˜b†s(�, η)�=�−icg�/2 − γ /4−ig/2ig−icg� − �/2�×�˜as(�, η)˜b†s(�, η)�+�0√�˜ξ †�.(D1)013010-14QUANTUM COHERENT CONTROL IN PULSED WAVEGUIDE …PHYSICAL REVIEW RESEARCH 5, 013010 (2023)The thermal noise is introduced as the following:�˜b†s(�1, η1)˜bs(�2, η2)�= n0δ(�1 − �2)δ(η1 − η2) ,�˜ξ †(�1, η1)˜ξ(�2, η2)�= nthδ(�1 − �2)δ(η1 − η2) .(D2)The n0 is the phonon number expectation at the initial stateη = 0, and the nth is the thermal phonon expectation numberdetermined by the temperature of the environment. When thesystem is cooled by the cooling technique we proposed in theprevious sections, n0 < nth can be achieved.The quantum noise is introduced as the following:�˜bs(�1, η1), ˜b†s(�2, η2)�= δ(�1 − �2)δ(η1 − η2) ,�˜ξ(�1, η1), ˜ξ †(�2, η2)�= δ(�1 − �2)δ(η1 − η2) .(D3).The n0 refers to the average thermal phonon at phase-matching point at t = 0. It is possible for n0 < nth when thesystem is precooled by the laser cooling process, such as thecoherent transfer-based Brillouin cooling we present in thispaper. The exact expression for the matrix readsG11 = e− 14 �∗e,sη−icg�η�cosh�ge,sη√2�+�e,s2√2ge,ssinh�ge,sη√2��,G12 = −ie− 14 �∗e,sη−icg�ηg√2ge,ssinh�ge,sη√2�,G21 = ie− 14 �∗e,sη−icg�η√2gge,ssinh�ge,sη√2�,G22 = e− 14 �∗e,sη−icg�η�cosh�ge,sη√2�−�e,s2√2ge,ssinh�ge,sη√2��,(D4)wherege,s =�g2 + (� + icg�)2/8 ,�e,s = � + icg� .(D5)The two mode EPR variables we choose areu = 1α Xa + αYb , v = 1αYa + αXb ,(D6)with α = 2−1/4. The u, v can be written asu = 2−1/2��21/4˜as + 2−1/4i˜b†s�+ H.c.�,v = 2−1/2�−i�21/4˜as + 2−1/4i˜b†s�+ H.c.�.(D7)We have21/4˜as(�, η) + 2−1/4i˜bs(�, η) = 2−1/4(√2G11 + iG21)˜as(�, 0) + 2−1/4(√2G12 + iG22)˜b†s(�, 0)+ 2−1/4√�� η0dν (√2G12 + iG22)˜ξ †b (�, ν).(D8)The EPR variance is defined asσ 2EPR = σ 2u + σ 2vα2 + α−2 .(D9)For the entangled state, the Duan’s criterion yieldsσ 2EPR < 1 .(D10)Using the commutation relation, we obtained the �EPR asσ 2EPR = 13|√2G11 + iG21|2 + 13(2n0 + 1)|√2G12 + iG22|2 + 13�(2nth + 1)� η0dν|√2G12 + iG22|2.(D11)We have√2G11 + iG21 =√2 e− 14 �∗e,sη−icg�η�cosh�ge,sη√2�− gge,ssinh�ge,sη√2�+�e,s2√2ge,ssinh�ge,sη√2��,√2G12 + iG22 = ie− 14 �∗e,sη−icg�η�cosh�ge,sη√2�− gge,ssinh�ge,sη√2�−�e,s2√2ge,ssinh�ge,sη√2��.(D12)When g ≫ �, we have√22 |√2G11 + iG21| ≈ |√2G12 + iG22|≈ e− 14 �∗e,sη�����e− ge,sη√2 −�e,s2√2ge,ssinh�ge,sη√2������.(D13)013010-15ZHANG, ZHU, WOLFF, AND STILLERPHYSICAL REVIEW RESEARCH 5, 013010 (2023)Therefore the EPR variance readsσ 2EPR = 13���√2G11 + iG21���2+ 13(2n0 + 1)���√2G12 + iG22���2+ 13�(2nth + 1)� η0dν���√2G12 + iG22���2≈�23n0 + 1�|G(t)|2 + 13(2nth + 1)�� η0dν |G(ν)|2 ,(D14)whereG(η) = e−�∗e,sη/4�e− ge,sη√2 −�e,s2√2ge,ssinh�ge,sη√2��.(D15)[1] N. Gisin and R. Thew, Quantum communication, Nat.Photonics 1, 165 (2007).[2] H.-S. Zhong, H. Wang, Y.-H. Deng, M.-C. Chen, L.-C. Peng,Y.-H. Luo, J. Qin, D. Wu, X. Ding, Y. Hu et al., Quantumcomputational advantage using photons, Science 370, 1460(2020).[3] N. Maluf, An Introduction to Microelectromechanical SystemsEngineering, Artech House MEMS library (Artech House, Lon-don, 2000)[4] A. A. Clerk, M. H. Devoret, S. M. Girvin, F. Marquardt, andR. J. Schoelkopf, Introduction to quantum noise, measurement,and amplification, Rev. Mod. Phys. 82, 1155 (2010).[5] M. Poot and H. S. van der Zant, Mechanical systems in thequantum regime, Phys. Rep. 511, 273 (2012).[6] M. Aspelmeyer, T. J. Kippenberg, and F. Marquardt, Cavityoptomechanics, Rev. Mod. Phys. 86, 1391 (2014).[7] T. A. Palomaki, J. D. Teufel, R. W. Simmonds, and K. W.Lehnert, Entangling mechanical motion with microwave fields,Science 342, 710 (2013).[8] N. T. Otterstrom, R. O. Behunin, E. A. Kittlaus, and P. T.Rakich, Optomechanical Cooling in a Continuous System,Phys. Rev. X 8, 041034 (2018).[9] S. J. M. Habraken, K. Stannigel, M. D. Lukin, P. Zoller, andP. Rabl, Continuous mode cooling and phonon routers forphononic quantum networks, New J. Phys. 14, 115004 (2012).[10] H. Zoubi and K. Hammerer, Quantum Nonlinear Optics inOptomechanical Nanoscale Waveguides, Phys. Rev. Lett. 119,123602 (2017).[11] M. Li, W. H. P. Pernice, C. Xiong, T. Baehr-Jones, M.Hochberg, and H. X. Tang, Harnessing optical forces in inte-grated photonic circuits, Nature (London) 456, 480 (2008).[12] M. S. Kang, A. Nazarkin, A. Brenn, and P. S. J. Russell, Tightlytrapped acoustic phonons in photonic crystal fibres as highlynonlinear artificial Raman oscillators, Nat. Phys. 5, 276 (2009).[13] E. A. Kittlaus, N. T. Otterstrom, P. Kharel, S. Gertler, and P. T.Rakich, Non-reciprocal interband Brillouin modulation, Nat.Photonics 12, 613 (2018).[14] A. Kobyakov, M. Sauer, and D. Chowdhury, Stimulated Bril-louin scattering in optical fibers, Adv. Opt. Photon. 2, 1 (2010).[15] B. J. Eggleton, C. G. Poulton, P. T. Rakich, M. J. Steel, andG. Bahl, Brillouin integrated photonics, Nat. Photonics 13, 664(2019).[16] R. W. Boyd, Nonlinear optics (Academic Press, New York,2020).[17] J. E. Sipe and M. J. Steel, A Hamiltonian treatment of stimu-lated Brillouin scattering in nanoscale integrated waveguides,New J. Phys. 18, 045004 (2016).[18] V. Laude and J.-C. Beugnot, Lagrangian description of Brillouinscattering and electrostriction in nanoscale optical waveguides,New J. Phys. 17, 125003 (2015).[19] P. Rakich and F. Marquardt, Quantum theory of continuumoptomechanics, New J. Phys. 20, 045005 (2018).[20] Z. Zhu, D. J. Gauthier, and R. W. Boyd, Stored light in anoptical fiber via stimulated Brillouin scattering, Science 318,1748 (2007).[21] M. Merklein, B. Stiller, K. Vu, S. J. Madden, and B. J. Eggleton,A chip-integrated coherent photonic-phononic memory, Nat.Commun. 8, 574 (2017).[22] B. Stiller, M. Merklein, C. Wolff, K. Vu, P. Ma, S. J. Madden,and B. J. Eggleton, Coherently refreshing hypersonic phononsfor light storage, Optica 7, 492 (2020).[23] B. Stiller, M. Merklein, K. Vu, P. Ma, S. J. Madden, C. G.Poulton, and B. J. Eggleton, Crosstalk-free multi-wavelengthcoherent light storage via Brillouin interaction, APL Photonics4, 040802 (2019).[24] A. H. Safavi-Naeini and O. Painter, Proposal for an optome-chanical traveling wave phonon–photon translator, New J. Phys.13, 013017 (2011).[25] W. Chen and A. A. Clerk, Photon propagation in a one-dimensional optomechanical lattice, Phys. Rev. A 89, 033854(2014).[26] D. E. Chang, A. H. Safavi-Naeini, M. Hafezi, and O. Painter,Slowing and stopping light using an optomechanical crystalarray, New J. Phys. 13, 023003 (2011).[27] V. Peano, C. Brendel, M. Schmidt, and F. Marquardt, Topologi-cal Phases of Sound and Light, Phys. Rev. X 5, 031011 (2015).[28] S. Walter and F. Marquardt, Classical dynamical gauge fields inoptomechanics, New J. Phys. 18, 113029 (2016).[29] Y.-C. Chen, S. Kim, and G. Bahl, Brillouin cooling in a linearwaveguide, New J. Phys. 18, 115004 (2016).[30] C. Wolff, M. J. A. Smith, B. Stiller, and C. G. Poulton, Brillouinscattering—Theory and experiment: Tutorial, J. Opt. Soc. Am.B 38, 1243 (2021).[31] A. Nunnenkamp, K. Børkje, and S. M. Girvin, Single-PhotonOptomechanics, Phys. Rev. Lett. 107, 063602 (2011).[32] G. L. Keaton, M. J. Leonardo, M. W. Byer, and D. J. Richard,Stimulated Brillouin scattering of pulses in optical fibers, Opt.Express 22, 13351 (2014).013010-16QUANTUM COHERENT CONTROL IN PULSED WAVEGUIDE …PHYSICAL REVIEW RESEARCH 5, 013010 (2023)[33] Y.-D. Wang and A. A. Clerk, Using Interference for High Fi-delity Quantum State Transfer in Optomechanics, Phys. Rev.Lett. 108, 153603 (2012).[34] M. Dong and H. G. Winful, Area dependence of chirped-pulsestimulated Brillouin scattering: Implications for stored light anddynamic gratings, J. Opt. Soc. Am. B 32, 2514 (2015).[35] Y. Xie, A. Choudhary, Y. Liu, D. Marpaung, K. Vu, P. Ma, D.-Y.Choi, S. Madden, and B. J. Eggleton, System-level performanceof chip-based Brillouin microwave photonic bandpass filters, J.Lightwave Technol. 37, 5246 (2019).[36] P. T. Rakich, C. Reinke, R. Camacho, P. Davids, and Z. Wang,Giant Enhancement of Stimulated Brillouin Scattering in theSubwavelength Limit, Phys. Rev. X 2, 011008 (2012).[37] H. Shin, W. Qiu, R. Jarecki, J. A. Cox, R. H. Olsson, A.Starbuck, Z. Wang, and P. T. Rakich, Tailorable stimulated Bril-louin scattering in nanoscale silicon waveguides, Nat. Commun.4, 1944 (2013).[38] K. S. Abedin, Single-frequency Brillouin lasing using single-mode As2Se3 chalcogenide fiber, Opt. Express 14, 4037(2006).[39] P. Dainese, P. S. J. Russell, N. Joly, J. C. Knight, G. S.Wiederhecker, H. L. Fragnito, V. Laude, and A. Khelif, Stim-ulated Brillouin scattering from multi-ghz-guided acousticphonons in nanostructured photonic crystal fibres, Nat. Phys.2, 388 (2006).[40] S. Le Floch and P. Cambon, Study of Brillouin gain spec-trum in standard single-mode optical fiber at low temperatures(1.4–370 K) and high hydrostatic pressures (1–250 bars), Opt.Commun. 219, 395 (2003).[41] R. O. Behunin, P. Kharel, W. H. Renninger, H. Shin, F. Carter,E. Kittlaus, and P. T. Rakich, Enhanced Brillouin scattering insilica via saturable phonon losses, in Advanced Photonics 2015(Optica Publishing Group, Washington, DC, 2015) p. IT4B.3.[42] I. Y. Dodin and N. J. Fisch, Storing, Retrieving, and Process-ing Optical Information by Raman Backscattering in Plasmas,Phys. Rev. Lett. 88, 165001 (2002).[43] L. Tian and H. Wang, Optical wavelength conversion of quan-tum states with optomechanics, Phys. Rev. A 82, 053806(2010).[44] E. Verhagen, S. Deléglise, S. Weis, A. Schliesser, and T. J.Kippenberg, Quantum-coherent coupling of a mechanical os-cillator to an optical cavity mode, Nature (London) 482, 63(2012).[45] Y.-C. Liu, Y.-F. Xiao, X. Luan, and C. W. Wong, Dy-namic Dissipative Cooling of a Mechanical Resonator inStrong Coupling Optomechanics, Phys. Rev. Lett. 110, 153606(2013).[46] O. A. Nieves, M. D. Arnold, M. J. Steel, M. K. Schmidt, andC. G. Poulton, Numerical simulation of noise in pulsed Bril-louin scattering, J. Opt. Soc. Am. B 38, 2343 (2021).[47] C. M. de Sterke, K. R. Jackson, and B. D. Robert, Nonlinearcoupled-mode equations on a finite interval: A numerical pro-cedure, J. Opt. Soc. Am. B 8, 403 (1991).[48] M. L. Citron, H. R. Gray, C. W. Gabel, and C. R. Stroud,Experimental study of power broadening in a two-level atom,Phys. Rev. A 16, 1507 (1977).[49] L.-M. Duan, G. Giedke, J. I. Cirac, and P. Zoller, InseparabilityCriterion for Continuous Variable Systems, Phys. Rev. Lett. 84,2722 (2000).[50] D. Vitali, S. Gigan, A. Ferreira, H. R. Böhm, P. Tombesi, A.Guerreiro, V. Vedral, A. Zeilinger, and M. Aspelmeyer, Op-tomechanical Entanglement Between a Movable Mirror and aCavity Field, Phys. Rev. Lett. 98, 030405 (2007).[51] S. G. Hofer, W. Wieczorek, M. Aspelmeyer, and K. Hammerer,Quantum entanglement and teleportation in pulsed cavity op-tomechanics, Phys. Rev. A 84, 052327 (2011).[52] L. Tian, Robust Photon Entanglement Via Quantum Inter-ference in Optomechanical Interfaces, Phys. Rev. Lett. 110,233602 (2013).[53] M. C. Kuzyk, S. J. van Enk, and H. Wang, Generating robust op-tical entanglement in weak-coupling optomechanical systems,Phys. Rev. A 88, 062341 (2013).[54] D. Hatanaka, I. Mahboob, K. Onomitsu, and H. Yamaguchi,Phononwaveguidesforelectromechanicalcircuits,Nat.Nanotechnol. 9, 520 (2014).[55] See Supplemental Material at http://link.aps.org/supplemental/10.1103/PhysRevResearch.5.013010 for the data and sourcecodes for all the figures.013010-17 | RESEARCH GATE |
Citation: Qu, Z.; Yan, D.; Song, Z.Modeling Calcium Cycling in theHeart: Progress, Pitfalls, andChallenges. Biomolecules 2022, 12,1686. https://doi.org/10.3390/biom12111686Academic Editors: Aman Ullah andMohsin Saleet JafriReceived: 10 October 2022Accepted: 11 November 2022Published: 14 November 2022Publisher’s Note: MDPI stays neutralwith regard to jurisdictional claims inpublished maps and institutional affil-iations.Copyright:© 2022 by the authors.Licensee MDPI, Basel, Switzerland.This article is an open access articledistributedunderthetermsandconditions of the Creative CommonsAttribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).biomoleculesReviewModeling Calcium Cycling in the Heart: Progress, Pitfalls,and ChallengesZhilin Qu 1,2,*, Dasen Yan 3 and Zhen Song 31Department of Medicine, David Geffen School of Medicine, University of California, A2-237 CHS,650 Charles E. Young Drive South, Los Angeles, CA 90095, USA2Department of Computational Medicine, David Geffen School of Medicine, University of California,Los Angeles, CA 90095, USA3Peng Cheng Laboratory, Shenzhen 518066, China*Correspondence: zqu@mednet.ucla.eduAbstract: Intracellular calcium (Ca) cycling in the heart plays key roles in excitation–contractioncoupling and arrhythmogenesis. In cardiac myocytes, the Ca release channels, i.e., the ryanodinereceptors (RyRs), are clustered in the sarcoplasmic reticulum membrane, forming Ca release units(CRUs). The RyRs in a CRU act collectively to give rise to discrete Ca release events, called Casparks. A cell contains hundreds to thousands of CRUs, diffusively coupled via Ca to form a CRUnetwork. A rich spectrum of spatiotemporal Ca dynamics is observed in cardiac myocytes, includingCa sparks, spark clusters, mini-waves, persistent whole-cell waves, and oscillations. Models ofdifferent temporal and spatial scales have been developed to investigate these dynamics. Due to thecomplexities of the CRU network and the spatiotemporal Ca dynamics, it is challenging to model theCa cycling dynamics in the cardiac system, particularly at the tissue sales. In this article, we reviewthe progress of modeling of Ca cycling in cardiac systems from single RyRs to the tissue scale, thepros and cons of the current models and different modeling approaches, and the challenges to betackled in the future.Keywords: calcium cycling; excitation–contraction coupling; arrhythmias; computer modeling1. IntroductionThe heart is probably the most intensively and accurately modeled biological systemcompared to other organs [1–5]. So far, more than 100 action potential models (or modifiedversions) have been developed for different types of myocytes and species. Tissue andorgan scale models, including one-dimensional (1D) cable, two-dimensional (2D) sheet,three-dimensional (3D) slab, and anatomically based ventricle and atrium models, havebeen developed. These mathematical and computational models have been widely used toinvestigate cardiac excitation-contraction coupling and arrhythmias under physiologicaland pathophysiological conditions.Modeling of the voltage in the heart is relatively well executed, mainly followingthe Hodgkin–Huxley (HH) modeling approach [6]. The governing equation for the trans-membrane potential (V) of a myocyte is simply described by the following differentialequation: dVdt = −Iion/Cm, in which Iion is the total ionic current density and Cm is thecell membrane capacitance. In cardiac myocytes, there are many types of ionic currents(Figure 1A) which are modeled either using the HH formulation or Markovian approaches.In the HH formulism, the ionic current density is described as Is = Gsxmynzk(V − Es), inwhich Gs is the maximum conductance, and Es is the reversal potential. x, y, and z are thegating variables described by differential equations with properties (steady states and timeconstants) from experimental measurements of whole-cell voltage clamp recordings [6].In the Markovian approaches, there are two ways of modeling. In the first way, single ionchannel openings and closings are simulated using stochastic Markovian transitions, andBiomolecules 2022, 12, 1686. https://doi.org/10.3390/biom12111686https://www.mdpi.com/journal/biomoleculesBiomolecules 2022, 12, 16862 of 19the total ionic current of an assembly of ion channels is described as Is = gsNo(V − Es),in which gs is the single-channel conductance and No is the number of open channels at agiven time. In the second way, differential equations are used to describe the probabilitiesof states in the Markovian scheme, and the ionic current of an assembly of ion channelsis described as Is = GsPo(V − Es), in which Po is the open probability of the ion channels.Note that a ventricular myocyte is a 3D entity with its dimension being roughly [7] 150 ×30 × 15 µm3, but in the current action potential models, voltage is considered uniform overthe entire cell membrane. In other words, at any moment, the ion channels in the entire cellmembrane are assumed to sense the same voltage. Moreover, in the Markovian scheme,it is assumed that the ion channels are statistically independent, and thus the whole-cellcurrent is simply the summation of the single-channel currents.However, Ca cycling and its dynamics are much more complex to model. Ca cyclingnot only is required for contraction but also plays important roles in regulating ioniccurrents (Figure 1A). Ca is stored in the sarcoplasmic reticulum (SR), which forms acomplex network inside the cell. Ca is released from the SR into the cytoplasmic spacethrough the opening of the ryanodine receptors (RyRs) in the SR membrane. Opening ofthe RyRs is activated by Ca on both the cytoplasmic and luminal sides. Under the normalcondition, SR Ca release is mainly triggered by Ca entry from the L-type Ca channels(LCCs). Under diseased or Ca overload conditions, spontaneous Ca release occur morefrequently. In cardiac myocytes, RyRs form clusters (Figure 1B), which combine with theirassociated LCC clusters to form basic units of Ca signaling, called Ca release units (CRUs).A cell contains hundreds to thousands of CRUs [8–10], which form a coupled networkvia Ca diffusion in the cytoplasmic space and SR. A rich spectrum of spatiotemporal Cadynamics is observed in cardiac myocytes and other cell types, including Ca sparks, waves,and oscillations [11–21]. Due to the complex spatiotemporal Ca dynamics, it has beenchallenging to model Ca cycling dynamics in the cardiac system, particularly at the tissuescales. Models of different temporal and spatial scales have been developed to investigatethese spatiotemporal dynamics. In this article, we review the progress of modeling of Cacycling in cardiac systems from single RyRs to tissue scales, the pros and cons of the currentmodels and different modeling approaches, and the challenges to be tackled in the future.Biomolecules 2022, 12, x FOR PEER REVIEW 2 of 19 the Markovian approaches, there are two ways of modeling. In the first way, single ion channel openings and closings are simulated using stochastic Markovian transitions, and the total ionic current of an assembly of ion channels is described as 𝐼𝑠 = 𝑔𝑠𝑁𝑜(𝑉 − 𝐸𝑠), in which 𝑔𝑠 is the single-channel conductance and 𝑁𝑜 is the number of open channels at a given time. In the second way, differential equations are used to describe the probabilities of states in the Markovian scheme, and the ionic current of an assembly of ion channels is described as 𝐼𝑠 = 𝐺𝑠𝑃𝑜(𝑉 − 𝐸𝑠), in which 𝑃𝑜 is the open probability of the ion channels. Note that a ventricular myocyte is a 3D entity with its dimension being roughly [7] 150 × 30 × 15 μm3, but in the current action potential models, voltage is considered uniform over the entire cell membrane. In other words, at any moment, the ion channels in the entire cell membrane are assumed to sense the same voltage. Moreover, in the Markovian scheme, it is assumed that the ion channels are statistically independent, and thus the whole-cell current is simply the summation of the single-channel currents. However, Ca cycling and its dynamics are much more complex to model. Ca cycling not only is required for contraction but also plays important roles in regulating ionic cur-rents (Figure 1A). Ca is stored in the sarcoplasmic reticulum (SR), which forms a complex network inside the cell. Ca is released from the SR into the cytoplasmic space through the opening of the ryanodine receptors (RyRs) in the SR membrane. Opening of the RyRs is activated by Ca on both the cytoplasmic and luminal sides. Under the normal condition, SR Ca release is mainly triggered by Ca entry from the L-type Ca channels (LCCs). Under diseased or Ca overload conditions, spontaneous Ca release occur more frequently. In cardiac myocytes, RyRs form clusters (Figure 1B), which combine with their associated LCC clusters to form basic units of Ca signaling, called Ca release units (CRUs). A cell contains hundreds to thousands of CRUs [8–10], which form a coupled network via Ca diffusion in the cytoplasmic space and SR. A rich spectrum of spatiotemporal Ca dynam-ics is observed in cardiac myocytes and other cell types, including Ca sparks, waves, and oscillations [11–21]. Due to the complex spatiotemporal Ca dynamics, it has been chal-lenging to model Ca cycling dynamics in the cardiac system, particularly at the tissue scales. Models of different temporal and spatial scales have been developed to investigate these spatiotemporal dynamics. In this article, we review the progress of modeling of Ca cycling in cardiac systems from single RyRs to tissue scales, the pros and cons of the cur-rent models and different modeling approaches, and the challenges to be tackled in the future. ABFigure 1. Ca cycling/signaling in cardiac myocytes. (A) Schematic diagram showing Ca cycling andsignaling and its coupling with voltage, including the major components: (1) ionic currents and theirBiomolecules 2022, 12, 16863 of 19regulation by Ca/calmodulin (CaM) and Ca/calmodulin-dependent protein kinase II (CaMKII);(2) SR Ca release and uptake and their regulation by CaMKII and protein kinase A and C (PKAand PKC); (3) mitochondrial Ca cycling; (4) Ca and myofilament interaction; and (5) Ca-dependentsignaling pathways. DS stands for dyadic space and ROS for reactive oxygen species. For a detailedexplanation of each component in the diagram, see Ref. [22]. (B) RyR clusters measured from aventricular myocyte, adapted from Ref. [9].2. RyR ModelsThe RyRs are regulated by Ca from both the luminal and the cytosolic sides ofSR [23–25], and thus they are sensitive to both cytosolic and SR Ca. There are severalRyR models developed and used in modeling cardiac Ca cycling. The simplest modelis a two-state model consisting of a closed state and an open state (Figure 2A), used inmany previous simulation studies [26–28]. The close-to-open transition rate is a functionof cytosolic Ca. SR Ca dependence is also added to model the luminal Ca sensitivity. Akey issue that has been investigated related to RyR models is how a Ca spark is terminatedspontaneously [26,27,29–34], and several mechanisms have been proposed based on com-puter simulations. In the scenario of the two-state model, for a Ca spark to terminate, theSR Ca content needs to be depleted to a very low level (~90% depletion) to allow the RyRsto transition from the open state to the closed state. This requires a long enough time delaybetween the junctional SR (JSR) and the network SR (NSR). This delay is assumed to becaused by either slow SR Ca diffusion or complex NSR structures [26,27,32]. However, thisassumption is not supported by experiments which show that Ca diffusion between JSRand NSR is very fast and the JSR Ca is only depleted by around 40% (Figure 2D) [35–37].Stern et al. [29,38] developed a four-state model which includes inactivation and recoverystates (Figure 2B), in which both activation and inactivation of the RyRs are mediated by Cain the cytosolic space. However, there is no experimental evidence that the inactivation ofRyRs is mediated by cytosolic Ca. It seems that this RyR model still requires a substantialSR Ca depletion for termination of Ca sparks [31]. Shannon and Bers [39] modified theStern et al. model by replacing cytosolic Ca-dependent inactivation with SR Ca-dependentinactivation, which allows spark termination at a much higher SR Ca level [34]. Thismodel also allows luminal SR Ca activation of the RyRs. Restrepo et al. [40] developed amodel in which inactivation is mediated by calsequestrin in the SR (Figure 2C). Other RyRmodels have also been developed [29,30,41–43], and the effects of RyR cooperativity areinvestigated [26,43–47].Despite the effects of RyR activation and inactivation properties on spark termination,no matter how the RyRs are activated and inactivated, there is a common dynamicalmechanism of spark firing and termination [31,34], in which the firing and termination of aspark are governed by a bistable switch (Figure 2E). Namely, the cytosolic Ca concentration([Ca]i) has two stable states in a range of SR Ca concentration ([Ca]SR) (gray region inFigure 2E). When [Ca]i is at the low state, as [Ca]SR reaches a threshold (red arrow),spontaneous firing occurs, and [Ca]i increases to the high state. If SR depletion is fastenough to be below another threshold (cyan arrow), [Ca]i quickly switches to the low state,terminating the firing and resulting in a normal Ca spark. However, if the SR depletioncannot be low enough to reach the threshold, [Ca]i will stay at the high state, failing toterminate the spark promptly. The bistability is a result of the positive feedback due to Ca-induced Ca release. Whether there is RyR inactivation or not, [Ca]SR is the key parameterfor the termination of sparks. For the RyR models without inactivation or weak inactivation,once SR Ca is low enough, the Ca flux from the JSR to the dyadic space is so small that[Ca]i in the dyadic space becomes low. The low [Ca]i causes the transition from the openstate to the closed state to over compete the transition from the closed state to the openstate, shutting off the RyRs. For the models with RyR inactivation, particularly with SRCa-dependent inactivation, inactivation is the primary cause for shutting off the RyRs. Thedifference is that the termination threshold of [Ca]SR is much lower for the RyR modelswithout inactivation.Biomolecules 2022, 12, 16864 of 19When [Ca]SR is higher than the termination threshold, the spark can still terminatedue to stochastic fluctuation of RyR openings, resulting in the so-called long-lastingsparks [31,34,48,49]. This random transition is widely known as the Kramers escapeprocess in stochastic nonlinear systems. In other words, noise can cause the system tocross the potential barrier (the unstable state, dashed line in Figure 2E) to reach the lowstate. Theoretical analysis by Song et al. [34] showed that the duration of the long-lastingsparks exhibits an exponential distribution, which is described by the Kramers rate theory.Exponential distributions of spark duration have been observed in cardiac myocyte ex-periments in which long-lasting sparks were promoted by FK506 [48]. Note that the samerandom transition applies to spark firings when [Ca]SR is lower than the firing threshold,i.e., random fluctuations can cause transition from the low state to the high state even when[Ca]SR is lower than the firing threshold, following the same Kramers transition process.For LCC-triggered sparks, [Ca]SR is lower than the firing threshold, and Ca from the LCCselevates [Ca]i to cross the barrier (the unstable state) to reach the high state, firing the spark.Therefore, whether the sparks are spontaneous or triggered by L-type Ca current (ICa,L),they follow the same dynamical mechanism.Besides the spark dynamics, not surprisingly, how RyRs are activated and inactivatedis also important for the cell-scale Ca dynamics, such as Ca alternans and Ca waves. Forexample, it is shown that RyR refractoriness is important for the formation of Ca alter-nans [50–55], indicating that the models with inactivation and refractoriness are moreappropriate to describe the RyRs. Ca waves and oscillations are another widely observedphenomenon in the cardiac myocytes [17,56–59], and RyR refractoriness plays an importantrole in forming waves and oscillations. For example, removing the calsequestrin-mediatedinactivation from the RyR model in Figure 2C, Ca waves become fractionated and irregular,as shown in simulations using a 3D cell model (Figure 2F). The simulation result agreeswith the experimental observation that the R33Q mutation causes high-frequency butfractionated Ca waves (Figure 2G) [60], in which the R33Q mutation lacks the calsequestrin-mediated inactivation of RyRs. This implies that the two-state RyR model may not beappropriate for normal Ca cycling. Future modeling is needed to investigate how differ-ent RyR models affect the Ca alternans and Ca wave dynamics and compare them withexperimental observations to further reveal the RyR models’ appropriateness.Biomolecules 2022, 12, 16865 of 19Biomolecules 2022, 12, x FOR PEER REVIEW 5 of 19 Figure 2. RyR models and Ca cycling dynamics. (A) A two-state model. The rate constant from C to O (kCO) as a function cytosolic Ca without or with luminal SR Ca regulation. (B) A four-state model with cytosolic Ca-dependent inactivation developed by Stern et al. [29,38]. (C) A four-state model with calsequestrin-dependent inactivation developed by Restrepo et al. [40]. (D) Left: SR Ca during a spark and after caffeine. Right: SR Ca nadir distribution measured during sparks. Data from Zima et al. [36]. (E) Bistable switch for spark firing and termination. Red arrow marks the SR Ca threshold for spark firing and cyan arrow marks the SR Ca threshold for spark termination. Gray marks the bistable region in which the high state and low state are stable and the middle one (dashed) is un-stable. (F) Upper: Time-space plots of Ca from a computer simulation of a 3D cell with the RyR model in C. Lower: Time-space plots of Ca from a computer simulation of the same 3D cell with the RyR model without the CSQN-dependent inactivation, equivalent to the two-state model in A. Sim-ulations were done under voltage clamped at −80 mV using the 3D cell model by Song et al. [59]. (G). Time-space plots of Ca from a permeablized rat ventricular myocyte of control (upper) and R33Q mutation (lower), adapted from Ref. [60]. 3. Single CRU and CRU Network Models Models of single CRU with different spatial resolutions have been developed. In 1992, Stern [61] put forth a theory for excitation-contraction coupling to explain the graded response of Ca release to ICa,L, in which a local control (or “cluster bomb”) model is as-sumed. In this local control model, an LCC triggers a cluster of RyRs to open collectively, resulting in an all-or-none discretized release event. The all-or-none release events are probability events proportional to the strength of ICa,L, giving rise to the experimentally observed whole-cell graded response. This cluster bomb model can be considered as the first single CRU model. In 1993, the discretized release events, called Ca sparks, were ABCGCORII2Cai2CaiCaiCaiC1O1C2O22Cai2CaiCSQNCSQNFCO2CaiED[Ca]i[Ca]SRterminationfiringComputer simulationExperimentR33QControlR33QControl20 m25 m1 s1 sFigure 2. RyR models and Ca cycling dynamics. (A) A two-state model. The rate constant from Cto O (kCO) as a function cytosolic Ca without or with luminal SR Ca regulation. (B) A four-statemodel with cytosolic Ca-dependent inactivation developed by Stern et al. [29,38]. (C) A four-statemodel with calsequestrin-dependent inactivation developed by Restrepo et al. [40]. (D) Left: SR Caduring a spark and after caffeine. Right: SR Ca nadir distribution measured during sparks. Datafrom Zima et al. [36]. (E) Bistable switch for spark firing and termination. Red arrow marks the SRCa threshold for spark firing and cyan arrow marks the SR Ca threshold for spark termination. Graymarks the bistable region in which the high state and low state are stable and the middle one (dashed)is unstable. (F) Upper: Time-space plots of Ca from a computer simulation of a 3D cell with the RyRmodel in C. Lower: Time-space plots of Ca from a computer simulation of the same 3D cell withthe RyR model without the CSQN-dependent inactivation, equivalent to the two-state model in A.Simulations were done under voltage clamped at −80 mV using the 3D cell model by Song et al. [59].(G). Time-space plots of Ca from a permeablized rat ventricular myocyte of control (upper) and R33Qmutation (lower), adapted from Ref. [60].3. Single CRU and CRU Network ModelsModels of single CRU with different spatial resolutions have been developed. In1992, Stern [61] put forth a theory for excitation-contraction coupling to explain the gradedresponse of Ca release to ICa,L, in which a local control (or “cluster bomb”) model isassumed. In this local control model, an LCC triggers a cluster of RyRs to open collectively,resulting in an all-or-none discretized release event. The all-or-none release events areprobability events proportional to the strength of ICa,L, giving rise to the experimentallyobserved whole-cell graded response. This cluster bomb model can be considered as thefirst single CRU model. In 1993, the discretized release events, called Ca sparks, weredemonstrated in experiments [21]. Later, more detailed CRU models [26,29,38,62] thatinclude stochastic RyR openings and different compartments were developed to investigatethe spark dynamics, especially the mechanism of termination of a spark as describedearlier. Higher spatial resolution and more realistic CRU models were also developed,Biomolecules 2022, 12, 16866 of 19including realistic T-tubule structure reconstructed from experiments (Figure 3A) [63],spatially distributed RyRs with Ca diffusion in the dyadic space [27,28,32,64] and complexSR network structures (Figure 3B) [27,32]. These high-resolution models provide furtherdetails on spark firing and termination properties.To investigate the Ca cycling dynamics in a whole cell, one needs to develop a modelconsisting of hundreds and thousands of CRUs. It is difficult to implement the complexCRU models as in Figure 3A,B into a cell model. CRU network models [40,52,57,65–69]composed of simpler CRU models were developed. These models are used to investigate thespatiotemporal Ca cycling dynamics, including Ca alternans and Ca waves. As an example,Figure 3C shows a scheme of a CRU network model developed by Nivala et al. [57,58,66].The model consists of a 3D network of CRUs. Each CRU contains four types of spaces,bulk myoplasm, dyadic space, junctional SR, and network SR. Unlike the CRU models inFigure 3A,B, the dyadic space is a single compartment in this model and thus there is noCa diffusion, i.e., the RyRs in a CRU sense the same Ca in the dyadic space. The CRUs arecoupled via Ca diffusion in the SR and cytoplasm. Figure 4 shows the spatiotemporal Cacycling dynamics as extracellular Ca concentration increases in the same CRU networkmodel as in Figure 3C [57], undergoing a phase transition with criticality [70,71]. These samesequences of dynamics were demonstrated in experiments of ventricular myocytes [57].Most of the 3D whole-cell models are similar but differ somewhat in spatial resolution andstructural details, e.g., some with structural complexities between those in Figure 3A–C(see Colman et al. for a detailed review [4]). Although the CRU network models lack therealistic structures of the TT and the SR networks as the real one in Figure 3A, or evenas the complex one in Figure 3B, they are much more computationally convenient, andcan be easily modified to simulate different subcellular structures, different cell types, anddiseased conditions [68,72].Biomolecules 2022, 12, x FOR PEER REVIEW 6 of 19 demonstrated in experiments [21]. Later, more detailed CRU models [26,29,38,62] that in-clude stochastic RyR openings and different compartments were developed to investigate the spark dynamics, especially the mechanism of termination of a spark as described ear-lier. Higher spatial resolution and more realistic CRU models were also developed, in-cluding realistic T-tubule structure reconstructed from experiments (Figure 3A) [63], spa-tially distributed RyRs with Ca diffusion in the dyadic space [27,28,32,64] and complex SR network structures (Figure 3B) [27,32]. These high-resolution models provide further de-tails on spark firing and termination properties. To investigate the Ca cycling dynamics in a whole cell, one needs to develop a model consisting of hundreds and thousands of CRUs. It is difficult to implement the complex CRU models as in Figure 3A,B into a cell model. CRU network models [40,52,57,65–69] composed of simpler CRU models were developed. These models are used to investigate the spatiotemporal Ca cycling dynamics, including Ca alternans and Ca waves. As an ex-ample, Figure 3C shows a scheme of a CRU network model developed by Nivala et al. [57,58,66]. The model consists of a 3D network of CRUs. Each CRU contains four types of spaces, bulk myoplasm, dyadic space, junctional SR, and network SR. Unlike the CRU models in Figure 3 A,B, the dyadic space is a single compartment in this model and thus there is no Ca diffusion, i.e., the RyRs in a CRU sense the same Ca in the dyadic space. The CRUs are coupled via Ca diffusion in the SR and cytoplasm. Figure 4 shows the spa-tiotemporal Ca cycling dynamics as extracellular Ca concentration increases in the same CRU network model as in Figure 3C [57], undergoing a phase transition with criticality [70,71]. These same sequences of dynamics were demonstrated in experiments of ventric-ular myocytes [57]. Most of the 3D whole-cell models are similar but differ somewhat in spatial resolution and structural details, e.g., some with structural complexities between those in Figure 3A–C (see Colman et al. for a detailed review [4]). Although the CRU net-work models lack the realistic structures of the TT and the SR networks as the real one in Figure 3A, or even as the complex one in Figure 3B, they are much more computationally convenient, and can be easily modified to simulate different subcellular structures, differ-ent cell types, and diseased conditions [68,72]. . Figure 3. Single CRU and CRU network models. (A) CRU model with realistic structure, adapted with permission from Ref. [63], copyright 2012, John Wiley and Sons. Upper: Structure. Lower: A AB100 CRUs10MyoDSJSRNSRJLCCJRyRJupJdiffuCRUWhole cellCFigure 3. Single CRU and CRU network models. (A) CRU model with realistic structure, adaptedwith permission from Ref. [63], copyright 2012, John Wiley and Sons. Upper: Structure. Lower: ACa snapshot during a Ca spark. (B) A detailed CRU model, adapted with permission from Ref. [32],copyright 2013, Elsevier. Left: Model structure. Right: Snapshots of Ca at different time points inthe dyadic space. (C) A CRU network model to represent the whole cell [57,66]. Left: A 3D networkconsisting of 100 × 20 × 10 CRUs, representing a whole cell. Right: Structure of a CRU. Coloredcubes mark different spaces and arrows indicate different fluxes.Biomolecules 2022, 12, 16867 of 19Ca snapshot during a Ca spark. (B) A detailed CRU model, adapted with permission from Ref. [32], copyright 2013, Elsevier. Left: Model structure. Right: Snapshots of Ca at different time points in the dyadic space. (C) A CRU network model to represent the whole cell [57,66]. Left: A 3D network consisting of 100 × 20 × 10 CRUs, representing a whole cell. Right: Structure of a CRU. Colored cubes mark different spaces and arrows indicate different fluxes. Figure 4. Spatiotemporal Ca cycling dynamics in a 3D CRU network model [57]. (A) Snapshots of Ca concentration in the cytoplasmic space for different Ca loading achieved by altering extracellular [Ca]o. From left to right, [Ca]o = 3 mM, 9 mM, 10 mM, and 16 mM. (B) Averaged whole-cell Ca concentration for the corresponding loading conditions. (C) Number of clusters versus spark cluster size. Symbols are data from simulations and lines are reference lines. The line in the left most panel is exponential, but those in other panels are linear, indicating power-law distribution. Power-law clustering indicates that a critical phenomenon [70,71] occurs during the transition from individual sparks to waves. Adapted with permission from Ref. [57], copyright 2012, Elsevier. 4. Modeling SR Ca Cycling in Single-Cell Action Potential Models In early action potential models, such as the Noble model [73], the Beeler–Reuter model [74], and the 1991 Luo and Rudy model [75], there is no or trivial intracellular Ca cycling. Ca cycling was added in the later action potential models using different model-ing approaches. Global RyR models. In this type of models, one assumes that SR Ca release is uniform in the whole cell, and thus uses a single-channel RyR model to describe the whole-cell SR Ca release. We call this type of Ca release models as global RyR models. The first action potential model with intracellular Ca cycling is the DiFrancesco-Noble model for sino-atrial nodal (SAN) cells [76], in which Ca-induced Ca release based on Fabiato’s theory [77] was implemented with the SR Ca release flux (we denoted it as 𝐽𝑟𝑒𝑙 in this paper) being formulated as: 𝐽𝑟𝑒𝑙 = 𝛼𝑟𝑒𝑙[𝐶𝑎]𝑟𝑒𝑙[𝐶𝑎]𝑖2[𝐶𝑎]𝑖2+𝐾𝑚,𝐶𝑎. This formulation is equivalent to the two-state RyR model in Figure 2A. This Ca cycling model was modified for excitation-contraction coupling in a rabbit atrial cell model [78]. Jafri et al. [79] developed an action potential model with Ca cycling by adopting the Keizer and Levine’s RyR model [41] for SR Ca release, and Shannon and Bers [39] developed an action potential model by using a modified Stern et al. RyR model (Figure 2B), in which SR Ca-dependent activation was 12345 61001011021031041051101001011021031041101001001011021031041101001011021032830323436380.20.31012141618200.20.30.40.510121416182012 1012141618200.100.150 ms40 ms80 ms120 ms160 ms0 ms40 ms80 ms120 ms160 ms0 ms100 ms260 ms400 ms500 ms0 ms140 ms220 ms300 ms460 msTime (s)0.10.40.71.0[Ca]i (M)ABCTime (s)Time (s)Time (s)Cluster sizeCluster sizeCluster sizeCluster size# of clusters<[Ca]i> (M)20 mIncreasing Ca loadFigure 4. Spatiotemporal Ca cycling dynamics in a 3D CRU network model [57]. (A) Snapshots ofCa concentration in the cytoplasmic space for different Ca loading achieved by altering extracellular[Ca]o. From left to right, [Ca]o = 3 mM, 9 mM, 10 mM, and 16 mM. (B) Averaged whole-cell Caconcentration for the corresponding loading conditions. (C) Number of clusters versus spark clustersize. Symbols are data from simulations and lines are reference lines. The line in the left most panelis exponential, but those in other panels are linear, indicating power-law distribution. Power-lawclustering indicates that a critical phenomenon [70,71] occurs during the transition from individualsparks to waves. Adapted with permission from Ref. [57], copyright 2012, Elsevier.4. Modeling SR Ca Cycling in Single-Cell Action Potential ModelsIn early action potential models, such as the Noble model [73], the Beeler–Reutermodel [74], and the 1991 Luo and Rudy model [75], there is no or trivial intracellular Cacycling. Ca cycling was added in the later action potential models using different modelingapproaches.Global RyR models. In this type of models, one assumes that SR Ca release is uniformin the whole cell, and thus uses a single-channel RyR model to describe the whole-cell SRCa release. We call this type of Ca release models as global RyR models. The first actionpotential model with intracellular Ca cycling is the DiFrancesco-Noble model for sinoatrialnodal (SAN) cells [76], in which Ca-induced Ca release based on Fabiato’s theory [77]was implemented with the SR Ca release flux (we denoted it as Jrel in this paper) beingformulated as: Jrel = αrel[Ca]rel[Ca]2i[Ca]2i +Km,Ca . This formulation is equivalent to the two-stateRyR model in Figure 2A. This Ca cycling model was modified for excitation-contractioncoupling in a rabbit atrial cell model [78]. Jafri et al. [79] developed an action potentialmodel with Ca cycling by adopting the Keizer and Levine’s RyR model [41] for SR Carelease, and Shannon and Bers [39] developed an action potential model by using a modifiedStern et al. RyR model (Figure 2B), in which SR Ca-dependent activation was added andSR Ca-dependent inactivation was used. The RyR model by Shannon and Bers was alsoadopted by Maltsev and Lakatta for their SAN cell model [80]. In these models, Jrel isformulated as Jrel = kPo([Ca]SR − [Ca]i) in which Po is the open probability calculated fromthe RyR model. Sine in this type of models, the whole cell contains a single cytosolic Ca pooland a single SR Ca pool, and thus all the RyRs in whole cell sense the same cytosolic Ca andBiomolecules 2022, 12, 16868 of 19the same SR Ca. These models are called “common-pool” models. However, as pointed outby Stern [61], the “local-control” model is more appropriate for ICa,L triggered SR Ca releaseexhibiting graded response. On the other hand, the common-pool model with a global RyRmodel allows spontaneous Ca release responsible for delayed afterdepolarizations (DADs)in ventricular cells [81] or the Ca clock in SAN cells [80].Phenomenological SR Ca release models. In many of the action potential models, insteadof using a RyR model, SR Ca release is modeled phenomenologically. For example, in the1994 Luo and Rudy model [82,83], Jrel was modeled as: Jrel = Grel([Ca]SR − [Ca]i) withGrel ∝ f (∆Cai,2)�1 − exp�− tτon��exp�−tτo f f�. ∆Cai,2 is the Ca change at 2 ms after.Vmax,and if this change is below the given threshold, Grel = 0. Since ∆Cai,2 is caused by Ca entryfrom LCCs, thus Grel is ICa,L-dependent. Another type of phenomenological Ca releasemodel [84–88] uses the following type of formulation for Jrel: Jrel = Grelxy([Ca]SR − [Ca]i),in which x and y are activation and inactivation gating variables formulated in the form ofHH formulism as for the membrane ionic currents. In this type of model, the activationgating x’s steady state (x∞) is a function of ICa,L, and thus Ca release is ICa,L-dependent. In astudy by Livshitz and Rudy [89], they used a Ca release model by the following differentialequation of Jrel: dJreldt= − Jrel,∞+Jrelτrel, in which Jrel,∞ is a function of ICa,L and SR Ca. Thisformulation was later adopted in the human model by O’Hara et al. [90]. The authors statedthat they developed a two-state (closed-open) model of SR Ca release, but it is unclear howa differential equation for Jrel can be derived from the two-state model. On the other hand,a similar equation was derived earlier by Shiferaw et al. [91] based on spark statistics (seebelow). Note that in these models, SR Ca release is a function of ICa,L, i.e., SR Ca release istriggered by ICa,L, which can effectively model the graded response of Ca release. However,if ICa,L = 0, no SR Ca release can occur in these models, indicating that spontaneous Carelease cannot be appropriately modeled. On the other hand, these models can exhibit Caalternans [84–86,89]. As will be discussed below, a phenomenological spontaneous releaseterm can be added to these models for modeling spontaneous Ca release.Spark statistics-based models. Shiferaw et al. [91] took a different approach to developan SR Ca release model based on spark dynamics and statistics, which was improved inthe Mahajan et al. model [92]. The basic idea is that Ca sparks are triggered by ICa,L andthe rate of spark generation is proportional to ICa,L, i.e., dNdt ∝ ICa,L. On the other hand,a Ca spark has a limited lifetime, and Ca sparks disappear over time. By consideringthe spark birth-and-death process, they derived a differential equation for Jrel as [92]:dJreldt = N(ICa,L)Q�cJSR�cJSR − JrelT , in which N(ICa,L) is the number of sparks as a functionof ICa,L, T is a time constant related to the spark lifetime, Q�cJSR�is the fractional SR Carelease which can be experimentally determined [93,94]. Similar to the phenomenologicalmodels, this model cannot exhibit spontaneous Ca release but can exhibit Ca alternans.Indeed, this model is the first to reveal the role of fractional Ca release in the genesis ofCa alternans.Local-control models. In 2002, an action potential model with a local-control model ofCa cycling was developed by Greenstein and Winslow [42]. In this model, thousands ofCRUs with stochastic RyRs and LCCs are implemented with a common bulk cytosolicspace. A simpler version was developed later [95]. Similar local-control models weredeveloped and simulated with advanced numerical methods and mathematical approachesto speed up computation [96,97]. As expected, these models can well simulate the gradedresponse. Moreover, it can exhibit Ca alternans [98]. Although no studies have been carriedout to show spontaneous Ca release in these models, in principle, this type of model iscapable of simulating spontaneous Ca release to result in Ca oscillations. A caveat of thistype of model is that the Ca released from the SR enters into the common bulk cytosolicspace, lacking the capability for spatiotemporal Ca cycling dynamics. Since it containsCRUs in the order of 10,000, it is computationally nontrivial and difficult to be used fortissue-scale simulations.Spatially extended cell models. More recently, action potential models containing a 3DCRU network for Ca cycling have been developed to investigate Ca cycling dynamicsBiomolecules 2022, 12, 16869 of 19and their coupling with voltage [40,52,57,66–68]. These models are truly local controlmodels which can exhibit the whole spectrum of spatiotemporal Ca cycling dynamics(e.g., Figure 4). These models can be easily adapted for different types of cells, suchas atrial myocytes [99–103] and SAN cells [104,105] that exhibit complex T-tubular struc-tures [106,107], as well as subcellular structural remodeling in heart failure [72] in whichT-tubule disruption occurs [108–110]. Other action potential models with spatially ex-tended Ca cycling have also been developed to simulate subcellular Ca alternans andtriggered activities [46,111–115]. A major caveat is that the spatially extended models arecomputationally challenging to use for tissue-scale simulations.As mentioned above, each type of model has its advantages and pitfalls. A commonissue of the non-spatial models is that they cannot properly simulate DADs. Even if someof them can simulate DADs, such as the model by Shannon and Bers [81], it cannot capturethe stochastic nature of DADs [59,116–118] as well as the positive feedback between voltageand Ca releases that give rise to the dynamics of the triggered activity [59,116]. As shown inFigure 4, when intracellular Ca is low, such as under normal conditions, the Ca sparks areindependent stochastic events (i.e., no spark-induced sparks). Thus, a statistical approachused by Shiferaw et al. [91] to derive the Ca release equation is a more rigorous approach(i.e., the differential equation is derived based on first principles) than others. Therefore,the model by Shiferaw et al. is more appropriate for the condition of low or normal Calevel. On the other hand, when Ca is very high, Ca oscillates synchronously and almostperiodically; the global RyR modeling approach is more appropriate. However, none ofthe models (except for the spatial cell models) can correctly capture the dynamics at theintermediate range of Ca level where large random fluctuations occur due to criticality(Figure 4). This is the range where DADs and triggered activity occur.Besides DADs, it has been shown experimentally that early afterdepolarizations(EADs), although mainly caused by reactivation of ICa,L [119], can also be caused byspontaneous Ca release [120–123]. Spatially extended 3D cell models [124,125] have beenused to investigate the roles of spontaneous Ca release in the genesis of EADs, which cancapture certain features of the experimental observations. Wilson et al. [126] modified a non-spatial cell model, the Shannon and Bers model [39], to exhibit spontaneous Ca oscillationand investigated spontaneous Ca release/oscillation on EAD genesis. However, in thismodel, the voltage depolarizations for EADs occur much earlier than the correspondingspontaneous Ca releases (the Ca oscillation and the voltage oscillation for EADs are almostin opposite phase) [126,127], which does not agree with the experimental observation thatthe rise of Ca occurs slightly ahead of the voltage depolarization of an EAD (experimentalevidence to support that Ca release causes the EAD, not the other way around) [122,123].Due to the spatiotemporal nature of the Ca cycling dynamics, the same problems remainfor modeling spontaneous Ca release induced EADs as those for modeling DADs using thenon-spatial action potential models.5. Modeling Mitochondrial Ca Cycling and Energetics in Action Potential ModelsMitochondria are not only the energy store but also another Ca store for the cell(Figure 1A) and play important roles in intracellular Ca cycling and cardiac function [128,129].Mitochondrial metabolism provides energy for ion pumps, including Na-Ca exchanger, Na-K pump, and SERCA pump. Furthermore, there are ATP-sensitive K channels and RyRs arealso regulated by ATP. Besides providing energy for the ion channels and pumps, mitochon-dria also participate directly in intracellular Ca cycling. Cytosolic Ca enters into mitochon-dria via mitochondrial Ca uniporter and extrudes from mitochondria via mitochondrialNa-Ca exchanger. Ca may also get out of mitochondria via the transient opening of the mi-tochondrial permeability transition pore. Non-spatial (single mitochondrial compartment)models integrating mitochondrial metabolism and Ca cycling have been developed to sim-ulate the effects of mitochondria on Ca cycling and arrhythmogenesis [130–134]. However,mitochondria form a network inside the cell, which can exhibit complex spatiotemporalmitochondrial depolarization dynamics, such as critical phenomenon in mitochondrialBiomolecules 2022, 12, 168610 of 19depolarization [135,136]. In our recent publications [137–140], we developed a modelwith a spatially distributed mitochondrial network and used it to investigate Ca alternans,EADs, and spontaneous Ca release induced DADs and triggered activity. Similar to thecase without mitochondria in the model, when Ca dynamics becomes spatiotemporal, thesingle mitochondrial compartment model may fail to capture the Ca cycling dynamics andnetwork models are needed.6. Modeling Ca-Dependent SignalingCa-dependent signaling (Figure 1A) plays key roles in cardiac excitation-contractioncoupling and arrhythmogenesis [22,141]. Pathways of β-adrenergic signaling, proteinkinase A/C signaling, reactive oxygen species-dependent singling, and Ca/calmodulin-dependent protein kinase II signaling have been incorporated into many action potentialmodels [87,90,138,142–146], which have been used to investigate their effects on actionpotential and Ca cycling dynamics. Most of the models are non-spatial models, and it hasnot been investigated how the complex spatiotemporal Ca cycling dynamics will affect theCa-dependent signaling which then affects the Ca cycling and action potential dynamics.7. Tissue-Scale Modeling for Spatiotemporal Ca and Voltage DynamicsIn most of the tissue-scale simulations, non-spatial cell models were used. As men-tioned above, these models cannot properly model DADs and DAD-mediated triggeredactivities as well as spontaneous Ca release mediated EADs. To simulate the effects of spon-taneous Ca release on arrhythmias at the tissue scale, different approaches have been used.A simple and straightforward way of modeling was used [118,147], in which stochasticDADs were phenomenologically implemented by commanding stochastic SR Ca releasewith statistical properties from experimental data. Chudin et al. [84] added phenomeno-logically a spontaneous release term to the release flux as: Jspon = kPspon([Ca]SR − [Ca]i),in which Pspon, described by a differential equation, is a gating variable depending oncytosolic Ca. Chen and Shiferaw [111] developed a phenomenological model by adding aspontaneous spark generation term RSCR(t) to the differential equation for Jrel and used asmall 1D array to calculate RSCR(t). This model was used in a whole-heart simulation forrandom DADs and triggered arrhythmias [111,148]. Colman et al. [149,150] developed amulti-scale modeling approach in which a phenomenological spontaneous release functionwas derived based on the detailed 3D single-cell simulations, and used it to simulate DAD-mediated arrhythmias in tissue models. More recently, a phenomenological model for Cawave induced APD alternans in atrial cells was developed and used in atrial tissue foralternans induced arrhythmias [99,151]. Walker et al. [152] simulated a 100-cell cable witha detailed 3D cell model and developed a spatial-average filtering method for estimatingthe probability of extreme stochastic events from a limited set of spontaneous Ca2+ releaseprofiles to predict DAD events in tissue.Although these simulation studies have provided useful information and mechanisticinsights into DAD-mediated arrhythmias, they have their pitfalls. One major issue is theeffects of the feedback between Ca and voltage when complex spatiotemporal dynamicsoccurs. This feedback is lacking in the current multi-scale modeling approaches. More-over, tissue simulations with detailed 3D cell models are needed to validate the resultsof the phenomenological models, however, this is computationally non-trivial for largetissue sizes.We recently developed a parallel computational method using multiple graphicsprocessing unit (GPU) cards, which allows us to simulate large tissue sizes with a 3D cellmodel. Figure 5 shows one of our simulations (unpublished data) in which a DAD occursafter a pacing beat in a 100 × 100-cell tissue using a detailed 3D cell model that is the sameas in Song et al. [59]. Figure 5A shows the voltages and whole-cell Ca transients from aline of 100 cells in the tissue. Figure 5B shows Ca snapshots from the surface of the entire100 × 100-cell tissue at three different time points as marked. Figure 5C,D show zoom-inviews of Ca snapshots from a 20 × 20-cell area (Figure 5C) and a 3 × 3-cell area (Figure 5D).Biomolecules 2022, 12, 168611 of 19As shown in Figure 5D, during the DAD, the Ca waves occur in the cells, and the wavesare randomly different from cell to cell.In the simulations shown in Figure 5, the 3D cell model contains 100 × 20 × 10 CRUs,and thus the tissue contains a total of 2 × 108 CRUs. Since each CRU has 100 RyRs and10 LCCs, thus the tissue contains 2 × 1010 RyRs and 2 × 109 LCCs. The RyRs are describedby a four-state Markovian model (Figure 2C) and the LCCs are described by a seven-state Markovian model. The simulation was carried out on 10 Nvidia GTX 3090 cardsusing ~41.5 GiB GPU memory in total. The time step for the simulation was 0.01 ms. Itcosts ~3.5 computational hours to compute 1 s of the model time. Of note, our method iscomputationally scalable, meaning that we can perform larger tissue-size simulations withmore available GPU cards within a reasonable computational run time.processing unit (GPU) cards, which allows us to simulate large tissue sizes with a 3D cell model. Figure 5 shows one of our simulations (unpublished data) in which a DAD occurs after a pacing beat in a 100 × 100-cell tissue using a detailed 3D cell model that is the same as in Song et al. [59]. Figure 5A shows the voltages and whole-cell Ca transients from a line of 100 cells in the tissue. Figure 5B shows Ca snapshots from the surface of the entire 100 × 100-cell tissue at three different time points as marked. Figure 5C,D show zoom-in views of Ca snapshots from a 20 × 20-cell area (Figure 5C) and a 3 × 3-cell area (Figure 5D). As shown in Figure 5D, during the DAD, the Ca waves occur in the cells, and the waves are randomly different from cell to cell. In the simulations shown in Figure 5, the 3D cell model contains 100 × 20 × 10 CRUs, and thus the tissue contains a total of 2 × 108 CRUs. Since each CRU has 100 RyRs and 10 LCCs, thus the tissue contains 2 × 1010 RyRs and 2 × 109 LCCs. The RyRs are described by a four-state Markovian model (Figure 2C) and the LCCs are described by a seven-state Markovian model. The simulation was carried out on 10 Nvidia GTX 3090 cards using ~41.5 GiB GPU memory in total. The time step for the simulation was 0.01 ms. It costs ~3.5 computational hours to compute 1 s of the model time. Of note, our method is computa-tionally scalable, meaning that we can perform larger tissue-size simulations with more available GPU cards within a reasonable computational run time. Figure 5. Computer simulation of a 100 × 100-cell 2D tissue using a 3D spatial cell model showing spontaneous Ca release and DAD following a paced beat. (A). Voltage (left) and whole-cell Ca tran-sient (right) from a line of 100 cells in the tissue. The action potential was elicited by a stimulus applied at t = 0. (B). Ca snapshots (recorded on the surfaces of the 3D cells) in the 100 × 100-cell tissue taken from three time points as marked. (C). A zoom-in view of Ca snapshots from a small area (20 × 20 cells) for the three time points. (D). A zoom-in view of Ca snapshots from an even smaller area (3 × 3 cells) showing stochastic Ca wave dynamics inside the cells. Plotted is a 3D surface view of Ca on the surface of the 9 cells. The 3D cell model consist of 100 × 20 × 10 CRUs and the cell model is the same as in Song et al. [59]. The tissue model contains 2 × 1010 RyRs (four-state model in Figure 2C) and 2 × 109 LCCs, both are simulated with stochastic Markovian transitions. The time step for the simulation is ∆𝑡 = 0.01 ms. 8. Challenges for Future Modeling Figure 5. Computer simulation of a 100 × 100-cell 2D tissue using a 3D spatial cell model showingspontaneous Ca release and DAD following a paced beat. (A). Voltage (left) and whole-cell Catransient (right) from a line of 100 cells in the tissue. The action potential was elicited by a stimulusapplied at t = 0. (B). Ca snapshots (recorded on the surfaces of the 3D cells) in the 100 × 100-celltissue taken from three time points as marked. (C). A zoom-in view of Ca snapshots from a small area(20 × 20 cells) for the three time points. (D). A zoom-in view of Ca snapshots from an even smallerarea (3 × 3 cells) showing stochastic Ca wave dynamics inside the cells. Plotted is a 3D surface viewof Ca on the surface of the 9 cells. The 3D cell model consist of 100 × 20 × 10 CRUs and the cellmodel is the same as in Song et al. [59]. The tissue model contains 2 × 1010 RyRs (four-state model inFigure 2C) and 2 × 109 LCCs, both are simulated with stochastic Markovian transitions. The timestep for the simulation is ∆t = 0.01.8. Challenges for Future ModelingDespite the progress in modeling Ca cycling at different scales in the heart, there arestill plenty of challenges. A general problem is how to develop low-dimensional models ormulti-scale modeling approaches to link the dynamics from the protein scale to the wholeheart. Several gaps need to be filled:(1)The first gap is to simulate sparks from a high spatial resolution CRU model to anon-spatial CRU model. As shown in the high spatial resolution models [27,28,32](see also Figure 3B), Ca spark dynamics are spatiotemporal and depend on the spatialdistribution of the RyRs in the dyadic space. Moreover, Maltsev et al. [64] showed thatthe occurrence of Ca sparks harnesses the Ising phase transition in a 2D array of RyRs.On the other hand, the majority of the single CRU and CRU network models ignorethese spatiotemporal properties by using a RyR cluster without spatial placementBiomolecules 2022, 12, 168612 of 19of RyRs. How to correctly capture the properties of the phase transition and largefluctuations of the spatiotemporal system using a lower-resolution or non-spatialmodel needs to be investigated.(2)The second gap is from CRUs (sparks) to a low-dimensional whole-cell model (Cawaves and whole-cell Ca transient). So far, using the 3D cell model for large tis-sue and whole-heart simulations is computationally nontrivial. Low-dimensionalrepresentations of the cell are preferred. As shown extensively by experiments andsimulations (e.g., Figure 4), there is a hierarchy of Ca dynamics: quarks, sparks, sparkclusters, mini-waves, persistent waves, and whole-call oscillations. How to developa low-dimensional model to embrace these dynamics is a nontrivial challenge. Asmentioned above, when Ca is low or normal, the differential equation describing theSR Ca release by Shiferaw et al. is a correct approach. At very high Ca where SR Carelease is synchronous, the global RyR model is appropriate. However, neither ofthe two modeling approaches can capture the Ca dynamics in the intermediate Carange. As shown in Figure 4 (see Nivala et al. [57]), a second-order phase transitionoccurs for the transition from independent individual spark dynamics to the whole-cell oscillations, i.e., a critical phenomenon exists. How to model the dynamics atthis phase transition using a low-dimensional representation is unknown since whencriticality occurs, the dynamics is intrinsically high-dimensional. Another challengeis how to model complex 3D T-tubular structures in low-dimensional models. Forexample, T-tubules are disrupted in failing ventricular myocytes [108–110], which cancause nontrivial changes in both Ca cycling and action potential dynamics [68,72].Correctly modeling these effects is essential for tissue-scale modeling of heart failure.Furthermore, when Ca cycling dynamics becomes spatiotemporal, their effects on Ca-dependent ionic currents and signaling, such as ICa,L and INCX as well as Ca-activatedpotassium currents (Figure 1A), and thus on the action potential dynamics may benontrivial [125]. How to model these effects using a low-dimensional model is anotherissue needed to be concerned.(3)The third gap is from a single cell to a syncytium (tissue and whole heart). For tissue orwhole-heart simulations, cardiac tissue is treated as a syncytium with discretization incomputer simulations being typically from ∆x = ∆y = ∆z = 0.1 to 0.5 mm [3,153–155].The dimension of a typical myocyte is 0.15 × 0.03 × 0.015 mm3, which indicatesthat one “computational cell” is equivalent to 15 to 1500 real cells. Under normalconditions in which conduction is fast and Ca is mainly determined by ICa,L (so thatCa is synchronized by ICa,L in different cells), this type of discretization is appropriate.However, under diseased conditions, such as ischemia and heart failure, cells areweakly coupled and the Na current is attenuated, and thus the action potentialconduction in tissue is slow. For slow conduction, a discretization at the cell size ismore appropriate [156,157]. Moreover, when spatiotemporal Ca dynamics, such aswaves and alternans, occur in the cell, these dynamics may be dyssynchronous fromcell to cell [158] (see Figure 5), and thus a resolution of the cell size may be requiredfor investigating these dynamics. Large tissue or whole-heart simulations with aresolution at the cell size scale is computationally nontrivial.(4)Mitochondrial Ca cycling [131,137,138,159] and Ca-myofilament interactions [160,161]play important roles in Ca cycling dynamics. How to include their effects in a low-dimensional representation of a cell is also challenging.(5)Although stochastic Markovian models are used to simulate single ion channel dy-namics, they are still phenomenological models describing the transitions of an ionchannel among different hypothetical states. Molecular dynamics simulations ofsingle ion channels at the atomic scale are more accurate representations, which canprovide parameter information for transition rates of Markovian models used incell and tissue simulations, linking the molecular scale effects of gene mutations ordrugs to cellular and tissue scale behaviors [162,163]. However, molecular dynamicsBiomolecules 2022, 12, 168613 of 19simulation is computationally tedious [164], and it is still a major challenge to use itfor the development and validation of Markovian models.The central issue among these challenges is dimension reduction. Coarse grainingand mean-field theories are typical modeling approaches to reduce a high-dimensionalrepresentation to a low-dimensional one [165]. Approaches for dimension reduction spe-cific to ion channel models have been developed. For example, Fox derived a Langevinequation for a population of ion channels described by stochastic Markovian transitionsfor the HH formulation [166]. Keener [167] showed that many Markovian models of ionchannel kinetics have globally attracting stable invariant manifolds, which can substan-tially simplify the computation with no approximation. He showed that this applies tocertain models of potassium channels, sodium channels, and RyRs. Phenomenologicalmodeling is currently the major modeling approach that uses low-dimensional modelsto describe high-dimensional systems. Iterated maps are another low-dimensional repre-sentation of high-dimensional systems [53,151,168,169]. However, both phenomenologicalmodeling and iterated maps rely on experimental data for validation and are only valid forcertain conditions.Theoretically, if one has all the details of a system, one can build up a mathematicalmodel to describe the motion of the atoms following Newton’s laws of motion or quantummechanics. However, it is impossible to simulate such models due to computationallimitations, at least in the foreseen future. Even if one can simulate such a big system, it isstill too complex to be understandable. On the other hand, a main task of science is to reducecomplex systems into ones that can be grasped by the human brain and establish principlesand theories which can be passed from generation to generation. Therefore, developinglow-dimensional representations that can capture the dynamics or multi-scale modelingapproaches that can link the dynamics at different scales is crucial for understandingcomplex biological systems, such as Ca cycling in the heart.9. ConclusionsThe heart is so far the most widely and accurately modeled organ, and it has thepotential to use the models and simulations at bedside to guide treatments. Much progresshas been made for the past several decades of cardiac modeling, including models fromsingle proteins (ion channels), whole-cell Ca cycling and action potential models, and whole-heart models. Mathematical and computational methods, such as multiscale modelingapproaches and advanced numerical algorithms, have been developed. These modelingefforts have greatly helped the understanding of cardiac excitation–contraction couplingand arrhythmogenesis. Despite the progress in the last several decades, there are stillplentiful pitfalls in the current models, modeling approaches, and simulation methods. Thechallenges to overcome these pitfalls are nontrivial but highly worthwhile for the upcominggeneration of modelers to pursue in their future careers.Funding: This research was funded by National Institutes of Health grants R01 HL134709, R01HL139829, R01 HL134346, and R01 HL157116 (Z.Q.), and Natural Science Foundation of China GrantNo. 82172067 (Z.S.).Institutional Review Board Statement: Not applicable.Conflicts of Interest: The authors declare no conflict of interest.References1.Noble, D. Modeling the heart–from genes to cells to the whole organ. Science 2002, 295, 1678–1682. [CrossRef] [PubMed]2.Henriquez, C.S. A Brief History of Tissue Models For Cardiac Electrophysiology. IEEE Trans. Biomed. Eng. 2014, 61, 1457–1465.[CrossRef] [PubMed]3.Trayanova, N.A. Whole-heart modeling: Applications to cardiac electrophysiology and electromechanics. Circ. Res. 2011, 108,113–128. [CrossRef] [PubMed]4.Colman, M.A.; Alvarez-Lacalle, E.; Echebarria, B.; Sato, D.; Sutanto, H.; Heijman, J. Multi-scale computational modeling of spatialcalcium handling from nanodomain to whole-heart: Overview and perspectives. Front. Physiol. 2022, 13, 836622. [CrossRef]Biomolecules 2022, 12, 168614 of 195.Qu, Z.; Hu, G.; Garfinkel, A.; Weiss, J.N. Nonlinear and stochastic dynamics in the heart. Phys. Rep. 2014, 543, 61–162. [CrossRef][PubMed]6.Hodgkin, A.L.; Huxley, A.F. A Quantitative Description of Membrane Current and Its Application to Conduction and Excitationin Nerve. J. Physiol. 1952, 117, 500–544. [CrossRef] [PubMed]7.Bers, D.M. Excitation-Contraction Coupling and Cardiac Contractile Force, 2nd ed.; Kluwer Academic Publishers: Dordrecht, TheNetherlands, 2001.8.Soeller, C.; Crossman, D.; Gilbert, R.; Cannell, M.B. Analysis of ryanodine receptor clusters in rat and human cardiac myocytes.Proc. Natl. Acad. Sci. USA 2007, 104, 14958–14963. [CrossRef]9.Baddeley, D.; Jayasinghe, I.D.; Lam, L.; Rossberger, S.; Cannell, M.B.; Soeller, C. Optical single-channel resolution imaging of theryanodine receptor distribution in rat cardiac myocytes. Proc. Natl. Acad. Sci. USA 2009, 106, 22275–22280. [CrossRef]10.Hiess, F.; Vallmitjana, A.; Wang, R.; Cheng, H.; ter Keurs, H.E.D.J.; Chen, J.; Hove-Madsen, L.; Benitez, R.; Chen, S.R.W.Distribution and Function of Cardiac Ryanodine Receptor Clusters in Live Ventricular Myocytes. J. Biol. Chem. 2015, 290,20477–20487. [CrossRef]11.Camacho, P.; Lechleiter, J.D. Increased frequency of calcium waves in Xenopus laevis oocytes that express a calcium-ATPase.Science 1993, 260, 226–229. [CrossRef]12.Lipp, P.; Niggli, E. Microscopic spiral waves reveal positive feedback in subcellular calcium signaling. Biophys. J. 1993, 65,2272–2276. [CrossRef]13.Skupin, A.; Kettenmann, H.; Winkler, U.; Wartenberg, M.; Sauer, H.; Tovey, S.C.; Taylor, C.W.; Falcke, M. How does intracellularCa2+ oscillate: By chance or by the clock? Biophys. J. 2008, 94, 2404–2411. [CrossRef] [PubMed]14.Bootman, M.D.; Berridge, M.J.; Lipp, P. Cooking with calcium: The recipes for composing global signals from elementary events.Cell 1997, 91, 367–373. [CrossRef]15.Marchant, J.S.; Parker, I. Role of elementary Ca2+ puffs in generating repetitive Ca2+ oscillations. EMBO J. 2001, 20, 65–76.[CrossRef]16.Cheng, H.; Lederer, M.R.; Lederer, W.J.; Cannell, M.B. Calcium sparks and [Ca2+]i waves in cardiac myocytes. Am. J. Physiol.1996, 270, C148–C159. [CrossRef] [PubMed]17.Wier, W.G.; ter Keurs, H.E.; Marban, E.; Gao, W.D.; Balke, C.W. Ca2+ ‘sparks’ and waves in intact ventricular muscle resolved byconfocal imaging. Circ. Res. 1997, 81, 462–469. [CrossRef]18.Lukyanenko, V.; Gyorke, S. Ca2+ sparks and Ca2+ waves in saponin-permeabilized rat ventricular myocytes. J. Physiol. 1999, 521,575–585. [CrossRef]19.Thurley, K.; Tovey, S.C.; Moenke, G.; Prince, V.L.; Meena, A.; Thomas, A.P.; Skupin, A.; Taylor, C.W.; Falcke, M. Reliable encodingof stimulus intensities within random sequences of intracellular Ca2+ spikes. Sci. Signal 2014, 7, ra59. [CrossRef]20.Cheng, H.; Lederer, W.J. Calcium Sparks. Physiol. Rev. 2008, 88, 1491–1545. [CrossRef]21.Cheng, H.; Lederer, W.J.; Cannell, M.B. Calcium sparks: Elementary events underlying excitation-contraction coupling in heartmuscle. Science 1993, 262, 740–744. [CrossRef]22.Qu, Z.; Weiss, J.N. Mechanisms of Ventricular Arrhythmias: From Molecular Fluctuations to Electrical Turbulence. Annu. Rev.Physiol. 2015, 77, 29–55. [CrossRef] [PubMed]23.Jiang, D.; Wang, R.; Xiao, B.; Kong, H.; Hunt, D.J.; Choi, P.; Zhang, L.; Chen, S.R. Enhanced store overload-induced Ca2+ releaseand channel sensitivity to luminal Ca2+ activation are common defects of RyR2 mutations linked to ventricular tachycardia andsudden death. Circ. Res. 2005, 97, 1173–1181. [CrossRef] [PubMed]24.Priori, S.G.; Chen, S.R.W. Inherited Dysfunction of Sarcoplasmic Reticulum Ca2+ Handling and Arrhythmogenesis. Circ. Res.2011, 108, 871–883. [CrossRef]25.Laver, D.R. Regulation of the RyR channel gating by Ca2+ and Mg2+. Biophys. Rev. 2018, 10, 1087–1095. [CrossRef] [PubMed]26.Sobie, E.A.; Dilly, K.W.; dos Santos Cruz, J.; Lederer, W.J.; Jafri, M.S. Termination of cardiac Ca2+ sparks: An investigativemathematical model of calcium-induced calcium release. Biophys. J. 2002, 83, 59–78. [CrossRef]27.Cannell, M.B.; Kong, C.H.; Imtiaz, M.S.; Laver, D.R. Control of sarcoplasmic reticulum Ca2+ release by stochastic RyR gatingwithin a 3D model of the cardiac dyad and importance of induction decay for CICR termination. Biophys. J. 2013, 104, 2149–2159.[CrossRef]28.Walker, M.A.; Williams, G.S.B.; Kohl, T.; Lehnart, S.E.; Jafri, M.S.; Greenstein, J.L.; Lederer, W.J.; Winslow, R.L. SuperresolutionModeling of Calcium Release in the Heart. Biophys. J. 2014, 107, 3018–3029. [CrossRef]29.Stern, M.D.; Song, L.S.; Cheng, H.; Sham, J.S.; Yang, H.T.; Boheler, K.R.; Rios, E. Local control models of cardiac excitation-contraction coupling. A possible role for allosteric interactions between ryanodine receptors. J. Gen. Physiol. 1999, 113, 469–489.[CrossRef]30.Stern, M.D.; Cheng, H. Putting out the fire: What terminates calcium-induced calcium release in cardiac muscle? Cell Calcium2004, 35, 591–601. [CrossRef]31.Stern, M.D.; Rios, E.; Maltsev, V.A. Life and death of a cardiac calcium spark. J. Gen. Physiol. 2013, 142, 257–274. [CrossRef]32.Laver, D.R.; Kong, C.H.T.; Imtiaz, M.S.; Cannell, M.B. Termination of calcium-induced calcium release by induction decay: Anemergent property of stochastic channel gating and molecular scale architecture. J. Mol. Cell. Cardiol. 2013, 54, 98–100. [CrossRef][PubMed]33.Winslow, R.L.; Greenstein, J.L. Extinguishing the sparks. Biophys. J. 2013, 104, 2115–2117. [CrossRef] [PubMed]Biomolecules 2022, 12, 168615 of 1934.Song, Z.; Karma, A.; Weiss, J.N.; Qu, Z. Long-Lasting Sparks: Multi-Metastability and Release Competition in the Calcium ReleaseUnit Network. PLoS Comput. Biol. 2016, 12, e1004671. [CrossRef] [PubMed]35.Shannon, T.R.; Guo, T.; Bers, D.M. Ca2+ Scraps: Local Depletions of Free [Ca2+] in Cardiac Sarcoplasmic Reticulum DuringContractions Leave Substantial Ca2+ Reserve. Circ. Res. 2003, 93, 40–45. [CrossRef]36.Zima, A.V.; Picht, E.; Bers, D.M.; Blatter, L.A. Termination of cardiac Ca2+ sparks: Role of intra-SR [Ca2+], release flux, andintra-SR Ca2+ diffusion. Circ. Res. 2008, 103, e105–e115. [CrossRef]37.Picht, E.; Zima, A.V.; Shannon, T.R.; Duncan, A.M.; Blatter, L.A.; Bers, D.M. Dynamic calcium movement inside cardiacsarcoplasmic reticulum during release. Circ. Res. 2011, 108, 847–856. [CrossRef]38.Stern, M.D.; Pizarro, G.; Rios, E. Local control model of excitation-contraction coupling in skeletal muscle. J. Gen. Physiol. 1997,110, 415–440. [CrossRef]39.Shannon, T.R.; Wang, F.; Puglisi, J.; Weber, C.; Bers, D.M. A mathematical treatment of integrated Ca dynamics within theventricular myocyte. Biophys. J. 2004, 87, 3351–3371. [CrossRef]40.Restrepo, J.G.; Weiss, J.N.; Karma, A. Calsequestrin-mediated mechanism for cellular calcium transient alternans. Biophys. J. 2008,95, 3767–3789. [CrossRef]41.Keizer, J.; Levine, L. Ryanodine receptor adaptation and Ca2+-induced Ca2+ release-dependent Ca2+ oscillations. Biophys. J. 1996,71, 3477–3487. [CrossRef]42.Greenstein, J.L.; Winslow, R.L. An integrative model of the cardiac ventricular myocyte incorporating local control of Ca(2+)release. Biophys. J. 2002, 83, 2918–2945. [CrossRef]43.Greene, D.A.; Shiferaw, Y. Mechanistic link between CaM-RyR2 interactions and the genesis of cardiac arrhythmia. Biophys. J.2021, 120, 1469–1482. [CrossRef]44.Wang, K.; Tu, Y.; Rappel, W.J.; Levine, H. Excitation-contraction coupling gain and cooperativity of the cardiac ryanodine receptor:A modeling approach. Biophys. J. 2005, 89, 3017–3025. [CrossRef] [PubMed]45.Groff, J.R.; Smith, G.D. Ryanodine Receptor Allosteric Coupling and the Dynamics of Calcium Sparks. Biophys. J. 2008, 95,135–154. [CrossRef] [PubMed]46.Chen, W.; Wasserstrom, J.A.; Shiferaw, Y. Role of coupled gating between cardiac ryanodine receptors in the genesis of triggeredarrhythmias. Am. J. Physiol. Heart Circ. Physiol. 2009, 297, H171–H180. [CrossRef]47.Dixon, R.E.; Navedo, M.F.; Binder, M.D.; Santana, L.F. Mechanisms and physiological implications of cooperative gating ofclustered ion channels. Physiol. Rev. 2022, 102, 1159–1210. [CrossRef]48.Xiao, R.P.; Valdivia, H.H.; Bogdanov, K.; Valdivia, C.; Lakatta, E.G.; Cheng, H. The immunophilin FK506-binding proteinmodulates Ca2+ release channel closure in rat heart. J. Physiol. 1997, 500, 343–354. [CrossRef]49.Zima, A.V.; Picht, E.; Bers, D.M.; Blatter, L.A. Partial inhibition of sarcoplasmic reticulum ca release evokes long-lasting ca releaseevents in ventricular myocytes: Role of luminal ca in termination of ca release. Biophys. J. 2008, 94, 1867–1879. [CrossRef]50.Picht, E.; DeSantiago, J.; Blatter, L.A.; Bers, D.M. Cardiac alternans do not rely on diastolic sarcoplasmic reticulum calcium contentfluctuations. Circ. Res. 2006, 99, 740–748. [CrossRef]51.Shkryl, V.M.; Maxwell, J.T.; Domeier, T.L.; Blatter, L.A. Refractoriness of sarcoplasmic reticulum Ca release determines Caalternans in atrial myocytes. Am. J. Physiol. Heart Circ. Physiol. 2012, 302, H2310–H2320. [CrossRef]52.Rovetti, R.; Cui, X.; Garfinkel, A.; Weiss, J.N.; Qu, Z. Spark-induced sparks as a mechanism of intracellular calcium alternans incardiac myocytes. Circ. Res. 2010, 106, 1582–1591. [CrossRef] [PubMed]53.Qu, Z.; Liu, M.B.; Nivala, M. A unified theory of calcium alternans in ventricular myocytes. Sci. Rep. 2016, 6, 35625. [CrossRef][PubMed]54.Zhong, X.; Sun, B.; Vallmitjana, A.; Mi, T.; Guo, W.; Ni, M.; Wang, R.; Guo, A.; Duff, H.J.; Gillis, A.M.; et al. Suppression ofryanodine receptor function prolongs Ca2+ release refractoriness and promotes cardiac alternans in intact hearts. Biochem. J. 2016,473, 3951–3964. [CrossRef]55.Sun, B.; Wei, J.; Zhong, X.; Guo, W.; Yao, J.; Wang, R.; Vallmitjana, A.; Benitez, R.; Hove-Madsen, L.; Chen, S.R.W. The cardiacryanodine receptor, but not sarcoplasmic reticulum Ca2+-ATPase, is a major determinant of Ca2+ alternans in intact mouse hearts.J. Biol. Chem. 2018, 293, 13650–13661. [CrossRef]56.Boyden, P.A.; Pu, J.; Pinto, J.; Keurs, H.E. Ca2+ transients and Ca2+ waves in purkinje cells: Role in action potential initiation. Circ.Res. 2000, 86, 448–455. [CrossRef] [PubMed]57.Nivala, M.; Ko, C.Y.; Nivala, M.; Weiss, J.N.; Qu, Z. Criticality in intracellular calcium signaling in cardiac myocytes. Biophys. J.2012, 102, 2433–2442. [CrossRef]58.Nivala, M.; Ko, C.Y.; Nivala, M.; Weiss, J.N.; Qu, Z. The Emergence of Subcellular Pacemaker Sites for Calcium Waves andOscillations. J. Physiol. 2013, 591, 5305–5320. [CrossRef] [PubMed]59.Song, Z.; Qu, Z.; Karma, A. Stochastic initiation and termination of calcium-mediated triggered activity in cardiac myocytes. Proc.Natl. Acad. Sci. USA 2017, 114, E270–E279. [CrossRef] [PubMed]60.Liu, N.; Denegri, M.; Dun, W.; Boncompagni, S.; Lodola, F.; Protasi, F.; Napolitano, C.; Boyden, P.A.; Priori, S.G. AbnormalPropagation of Calcium Waves and Ultrastructural Remodeling in Recessive Catecholaminergic Polymorphic VentricularTachycardia. Circ. Res. 2013, 113, 142–152. [CrossRef]61.Stern, M.D. Theory of excitation-contraction coupling in cardiac muscle. Biophys. J. 1992, 63, 497–517. [CrossRef]Biomolecules 2022, 12, 168616 of 1962.Smith, G.D.; Keizer, J.E.; Stern, M.D.; Lederer, W.J.; Cheng, H. A simple numerical model of calcium spark formation and detectionin cardiac myocytes. Biophys. J. 1998, 75, 15–32. [CrossRef]63.Hake, J.; Edwards, A.G.; Yu, Z.; Kekenes-Huskey, P.M.; Michailova, A.P.; McCammon, J.A.; Holst, M.J.; Hoshijima, M.; McCulloch,A.D. Modeling cardiac calcium sparks in a three-dimensional reconstruction of a calcium release unit. J. Physiol. 2012, 590,4403–4422. [CrossRef] [PubMed]64.Maltsev, A.V.; Maltsev, V.A.; Stern, M.D. Clusters of calcium release channels harness the Ising phase transition to confine theirelementary intracellular signals. Proc. Natl. Acad. Sci. USA 2017, 114, 7525–7530. [CrossRef] [PubMed]65.Gaur, N.; Rudy, Y. Multiscale modeling of calcium cycling in cardiac ventricular myocyte: Macroscopic consequences ofmicroscopic dyadic function. Biophys. J. 2011, 100, 2904–2912. [CrossRef]66.Nivala, M.; de Lange, E.; Rovetti, R.; Qu, Z. Computational modeling and numerical methods for spatiotemporal calcium cyclingin ventricular myocytes. Front. Physiol. 2012, 3, 114. [CrossRef]67.Colman, M.A.; Pinali, C.; Trafford, A.W.; Zhang, H.; Kitmitto, A. A computational model of spatio-temporal cardiac intracellularcalcium handling with realistic structure and spatial flux distribution from sarcoplasmic reticulum and t-tubule reconstructions.PLoS Comput. Biol. 2017, 13, e1005714. [CrossRef]68.Song, Z.; Liu, M.B.; Qu, Z. Transverse tubular network structures in the genesis of intracellular calcium alternans and triggeredactivity in cardiac cells. J. Mol. Cell. Cardiol. 2018, 114, 288–299. [CrossRef]69.Hoang-Trong, T.M.; Ullah, A.; Lederer, W.J.; Jafri, M.S. A Stochastic Spatiotemporal Model of Rat Ventricular Myocyte CalciumDynamics Demonstrated Necessary Features for Calcium Wave Propagation. Membranes 2021, 11, 989. [CrossRef]70.Stanley, H.E. Introduction to Phase Transitions and Critical Phenomena; Oxford University Press: London, UK, 1971.71.Stanley, H.E. Scaling, universality, and renormalization: Three pillars of modern critical phenomena. Rev. Mod. Phys. 1999, 71,S358–S366. [CrossRef]72.Nivala, M.; Song, Z.; Weiss, J.N.; Qu, Z. T-tubule disruption promotes calcium alternans in failing ventricular myocytes:Mechanistic insights from computational modeling. J. Mol. Cell. Cardiol. 2015, 79, 32–41. [CrossRef]73.Noble, D. A modification of the Hodgkin-Huxley equations applicable to Purkinje fibre action and pace-maker potentials. J.Physiol. 1962, 160, 317–352. [CrossRef] [PubMed]74.Beeler, G.W.; Reuter, H. Reconstruction of the action potential of ventricular myocardial fibres. J. Physiol. 1977, 268, 177–210.[CrossRef] [PubMed]75.Luo, C.H.; Rudy, Y. A model of the ventricular cardiac action potential: Depolarization, repolarization, and their interaction. Circ.Res. 1991, 68, 1501–1526. [CrossRef] [PubMed]76.DiFrancesco, D.; Noble, D. A model of cardiac electrical activity incorporating ionic pumps and concentration changes. Philos.Trans. R. Soc. Lond. (Biol.) 1985, 307, 779–790.77.Fabiato, A. Calcium-induced release of calcium from the cardiac sarcoplasmic reticulum. Am. J. Physiol. 1983, 245, C1–C14.[CrossRef]78.Hilgemann, D.W.; Noble, D. Excitation-contraction coupling and extracellular calcium transients in rabbit atrium: Reconstructionof basic cellular mechanisms. Proc. R. Soc. Lond. B Biol. Sci. 1987, 230, 163–205.79.Jafri, M.S.; Rice, J.J.; Winslow, R.L. Cardiac Ca2+ dynamics: The roles of ryanodine receptor adaptation and sarcoplasmic reticulumload. Biophys. J. 1998, 74, 1149–1168. [CrossRef]80.Maltsev, V.A.; Lakatta, E.G. Synergism of coupled subsarcolemmal Ca2+ clocks and sarcolemmal voltage clocks confers robustand flexible pacemaker function in a novel pacemaker cell model. Am. J. Physiol. Heart Circ. Physiol. 2009, 296, H594–H615.[CrossRef]81.Shannon, T.R.; Wang, F.; Bers, D.M. Regulation of Cardiac Sarcoplasmic Reticulum Ca Release by Luminal [Ca] and AlteredGating Assessed with a Mathematical Model. Biophys. J. 2005, 89, 4096–4110. [CrossRef]82.Luo, C.H.; Rudy, Y. A dynamical model of the cardiac ventricular action potential: I. simulations of ionic currents and concentrationchanges. Circ. Res. 1994, 74, 1071–1096. [CrossRef]83.Faber, G.M.; Rudy, Y. Action potential and contractility changes in [Na(+)](i) overloaded cardiac myocytes: A simulation study.Biophys. J. 2000, 78, 2392–2404. [CrossRef]84.Chudin, E.; Goldhaber, J.; Garfinkel, A.; Weiss, J.; Kogan, B. Intracellular Ca2+ dynamics and the stability of ventricular tachycardia.Biophys. J. 1999, 77, 2930–2941. [CrossRef]85.Fox, J.J.; McHarg, J.L.; Gilmour, R.F. Ionic mechanism of electrical alternans. Am. J. Physiol. Heart Circ. Physiol. 2002, 282,H516–H530. [CrossRef]86.ten Tusscher, K.H.; Noble, D.; Noble, P.J.; Panfilov, A.V. A model for human ventricular tissue. Am. J. Physiol. Heart Circ. Physiol.2004, 286, H1573–H1589. [CrossRef] [PubMed]87.Hund, T.J.; Rudy, Y. Rate dependence and regulation of action potential and calcium transient in a canine cardiac ventricular cellmodel. Circulation 2004, 110, 3168–3174. [CrossRef]88.Stewart, P.; Aslanidi, O.V.; Noble, D.; Noble, P.J.; Boyett, M.R.; Zhang, H. Mathematical models of the electrical action potential ofPurkinje fibre cells. Philos. Trans. A Math. Phys. Eng. Sci. 2009, 367, 2225–2255. [CrossRef]89.Livshitz, L.M.; Rudy, Y. Regulation of Ca2+ and electrical alternans in cardiac myocytes: Role of CAMKII and repolarizingcurrents. Am. J. Physiol. Heart Circ. Physiol. 2007, 292, H2854–H2866. [CrossRef] [PubMed]Biomolecules 2022, 12, 168617 of 1990.O’Hara, T.; Virag, L.; Varro, A.; Rudy, Y. Simulation of the undiseased human cardiac ventricular action potential: Modelformulation and experimental validation. PLoS Comput. Biol. 2011, 7, e1002061. [CrossRef] [PubMed]91.Shiferaw, Y.; Watanabe, M.A.; Garfinkel, A.; Weiss, J.N.; Karma, A. Model of intracellular calcium cycling in ventricular myocytes.Biophys. J. 2003, 85, 3666–3686. [CrossRef]92.Mahajan, A.; Shiferaw, Y.; Sato, D.; Baher, A.; Olcese, R.; Xie, L.-H.; Yang, M.-J.; Chen, P.-S.; Restrepo, J.G.; Karma, A.; et al. Arabbit ventricular action potential model replicating cardiac dynamics at rapid heart rates. Biophys. J. 2008, 94, 392–410. [CrossRef]93.Bassani, J.W.; Yuan, W.; Bers, D.M. Fractional SR Ca release is regulated by trigger Ca and SR Ca content in cardiac myocytes. Am.J. Physiol. 1995, 268, C1313–C1319. [CrossRef] [PubMed]94.Shannon, T.R.; Ginsburg, K.S.; Bers, D.M. Potentiation of fractional sarcoplasmic reticulum calcium release by total and freeintra-sarcoplasmic reticulum calcium concentration. Biophys. J. 2000, 78, 334–343. [CrossRef]95.Hinch, R.; Greenstein, J.L.; Tanskanen, A.J.; Xu, L.; Winslow, R.L. A simplified local control model of calcium-induced calciumrelease in cardiac ventricular myocytes. Biophys. J. 2004, 87, 3723–3736. [CrossRef] [PubMed]96.Williams, G.S.; Huertas, M.A.; Sobie, E.A.; Jafri, M.S.; Smith, G.D. Moment closure for local control models of calcium-inducedcalcium release in cardiac myocytes. Biophys. J. 2008, 95, 1689–1703. [CrossRef]97.Williams, G.S.B.; Huertas, M.A.; Sobie, E.A.; Jafri, M.S.; Smith, G.D. A Probability Density Approach to Modeling Local Control ofCalcium-Induced Calcium Release in Cardiac Myocytes. Biophys. J. 2007, 92, 2311–2328. [CrossRef]98.Huertas, M.A.; Smith, G.D.; Gyorke, S. Ca2+ alternans in a cardiac myocyte model that uses moment equations to representheterogeneous junctional SR Ca2+. Biophys. J. 2010, 99, 377–387. [CrossRef]99.Shiferaw, Y.; Aistrup, G.L.; Wasserstrom, J.A. Synchronization of triggered waves in atrial tissue. Biophys. J. 2018, 115, 1130–1141.[CrossRef]100. Marchena, M.; Echebarria, B. Computational Model of Calcium Signaling in Cardiac Atrial Cells at the Submicron Scale. Front.Physiol. 2018, 9, 1760. [CrossRef]101. Marchena, M.; Echebarria, B. Influence of the tubular network on the characteristics of calcium transients in cardiac myocytes.PLoS ONE 2020, 15, e0231056. [CrossRef]102. Sutanto, H.; van Sloun, B.; Schönleitner, P.; van Zandvoort, M.A.M.J.; Antoons, G.; Heijman, J. The Subcellular Distribution ofRyanodine Receptors and L-Type Ca2+ Channels Modulates Ca2+-Transient Properties and Spontaneous Ca2+-Release Events inAtrial Cardiomyocytes. Front. Physiol. 2018, 9, 1108. [CrossRef]103. Vagos, M.R.; Arevalo, H.; Heijman, J.; Schotten, U.; Sundnes, J. A Novel Computational Model of the Rabbit Atrial Cardiomyocytewith Spatial Calcium Dynamics. Front. Physiol. 2020, 11, 556156. [CrossRef] [PubMed]104. Maltsev, A.V.; Yaniv, Y.; Stern, M.D.; Lakatta, E.G.; Maltsev, V.A. RyR-NCX-SERCA Local Cross-Talk Ensures Pacemaker CellFunction at Rest and During the Fight-or-Flight Reflex. Circ. Res. 2013, 113, e94–e100. [CrossRef] [PubMed]105. Maltsev, A.V.; Maltsev, V.A.; Stern, M.D. Stabilization of diastolic calcium signal via calcium pump regulation of complex localcalcium releases and transient decay in a computational model of cardiac pacemaker cell with individual release channels. PLoSComput. Biol. 2017, 13, e1005675. [CrossRef]106. Dibb, K.M.; Clarke, J.D.; Horn, M.A.; Richards, M.A.; Graham, H.K.; Eisner, D.A.; Trafford, A.W. Characterization of an ExtensiveTransverse Tubular Network in Sheep Atrial Myocytes and its Depletion in Heart Failure. Circ. Heart Fail. 2009, 2, 482–489.[CrossRef] [PubMed]107. Brandenburg, S.; Kohl, T.; Williams, G.S.B.; Gusev, K.; Wagner, E.; Rog-Zielinska, E.A.; Hebisch, E.; Dura, M.; Didié, M.; Gotthardt,M.; et al. Axial tubule junctions control rapid calcium signaling in atria. J. Clin. Investig. 2016, 126, 3999–4015. [CrossRef][PubMed]108. Lyon, A.R.; MacLeod, K.T.; Zhang, Y.; Garcia, E.; Kanda, G.K.; Lab, M.J.; Korchev, Y.E.; Harding, S.E.; Gorelik, J. Loss of T-tubulesand other changes to surface topography in ventricular myocytes from failing human and rat heart. Proc. Natl. Acad. Sci. USA2009, 106, 6854–6859. [CrossRef]109. Song, L.S.; Sobie, E.A.; McCulle, S.; Lederer, W.J.; Balke, C.W.; Cheng, H. Orphaned ryanodine receptors in the failing heart. Proc.Natl. Acad. Sci. USA 2006, 103, 4305–4310. [CrossRef]110. Seidel, T.; Navankasattusas, S.; Ahmad, A.A.; Diakos, N.A.; Xu, W.D.; Tristani-Firouzi, M.; Bonios, M.; Taleb, I.; Li, D.Y.; Selzman,C.H.; et al. Sheet-Like Remodeling of the Transverse Tubular System in Human Heart Failure Impairs Excitation-ContractionCoupling and Functional Recovery by Mechanical Unloading. Circulation 2017, 135, 1632–1645. [CrossRef]111. Chen, W.; Aistrup, G.; Wasserstrom, J.A.; Shiferaw, Y. A mathematical model of spontaneous calcium release in cardiac myocytes.Am. J. Physiol. Heart Circ. Physiol. 2011, 300, H1794–H1805. [CrossRef]112. Gaeta, S.A.; Bub, G.; Abbott, G.W.; Christini, D.J. Dynamical mechanism for subcellular alternans in cardiac myocytes. Circ. Res.2009, 105, 335–342. [CrossRef]113. Sato, D.; Bers, D.M.; Shiferaw, Y. Formation of spatially discordant alternans due to fluctuations and diffusion of calcium. PLoSONE 2014, 8, e85365. [CrossRef]114. Li, Q.; O’Neill, S.C.; Tao, T.; Li, Y.; Eisner, D.; Zhang, H. Mechanisms by which cytoplasmic calcium wave propagation andalternans are generated in cardiac atrial myocytes lacking T-tubules-insights from a simulation study. Biophys. J. 2012, 102,1471–1482. [CrossRef] [PubMed]115. Tao, T.; O’Neill, S.C.; Diaz, M.E.; Li, Y.T.; Eisner, D.A.; Zhang, H. Alternans of cardiac calcium cycling in a cluster of ryanodinereceptors: A simulation study. Am. J. Physiol. Heart Circ. Physiol. 2008, 295, H598–H609. [CrossRef] [PubMed]Biomolecules 2022, 12, 168618 of 19116. Lou, Q.; Belevych, A.E.; Radwanski, P.B.; Liu, B.; Kalyanasundaram, A.; Knollmann, B.C.; Fedorov, V.V.; Gyorke, S. Alternat-ing membrane potential/calcium interplay underlies repetitive focal activity in a genetic model of calcium-dependent atrialarrhythmias. J. Physiol. 2015, 593, 1443–1458. [CrossRef]117. Belevych, A.E.; Terentyev, D.; Terentyeva, R.; Ho, H.T.; Gyorke, I.; Bonilla, I.M.; Carnes, C.A.; Billman, G.E.; Gyorke, S. ShortenedCa2+ signaling refractoriness underlies cellular arrhythmogenesis in a postinfarction model of sudden cardiac death. Circ. Res.2012, 110, 569–577. [CrossRef] [PubMed]118. Ko, C.Y.; Liu, M.B.; Song, Z.; Qu, Z.; Weiss, J.N. Multiscale Determinants of Delayed Afterdepolarization Amplitude in CardiacTissue. Biophys. J. 2017, 112, 1949–1961. [CrossRef]119. Qu, Z.; Xie, L.-H.; Olcese, R.; Karagueuzian, H.S.; Chen, P.-S.; Garfinkel, A.; Weiss, J.N. Early afterdepolarizations in cardiacmyocytes: Beyond reduced repolarization reserve. Cardiovasc. Res. 2013, 99, 6–15. [CrossRef]120. Priori, S.G.; Corr, P.B. Mechanisms underlying early and delayed afterdepolarizations induced by catecholamines. Am. J. Physiol.1990, 258, H1796–H1805. [CrossRef]121. Volders, P.G.; Vos, M.A.; Szabo, B.; Sipido, K.R.; de Groot, S.H.; Gorgels, A.P.; Wellens, H.J.; Lazzara, R. Progress in theunderstanding of cardiac early afterdepolarizations and torsades de pointes: Time to revise current concepts. Cardiovasc. Res.2000, 46, 376–392. [CrossRef]122. Choi, B.R.; Burton, F.; Salama, G. Cytosolic Ca2+ triggers early afterdepolarizations and Torsade de Pointes in rabbit hearts withtype 2 long QT syndrome. J. Physiol. 2002, 543, 615–631. [CrossRef]123. Zhao, Z.; Wen, H.; Fefelova, N.; Allen, C.; Baba, A.; Matsuda, T.; Xie, L.H. Revisiting the ionic mechanisms of early afterdepolar-izations in cardiomyocytes: Predominant by Ca waves or Ca currents? Am. J. Physiol. Heart Circ. Physiol. 2012, 302, H1636–H1644.[CrossRef] [PubMed]124. Song, Z.; Ko, C.Y.; Nivala, M.; Weiss, J.N.; Qu, Z. Calcium-Voltage Coupling in the Genesis of Early and Delayed Afterdepolariza-tions in Cardiac Myocytes. Biophys. J. 2015, 108, 1908–1921. [CrossRef]125. Zhong, M.; Rees, C.M.; Terentyev, D.; Choi, B.-R.; Koren, G.; Karma, A. NCX-Mediated Subcellular Ca2+ Dynamics UnderlyingEarly Afterdepolarizations in LQT2 Cardiomyocytes. Biophys. J. 2018, 115, 1019–1032. [CrossRef] [PubMed]126. Wilson, D.; Ermentrout, B.; Nˇemec, J.; Salama, G. A model of cardiac ryanodine receptor gating predicts experimental Ca2+-dynamics and Ca2+-triggered arrhythmia in the long QT syndrome. Chaos Interdiscip. J. Nonlinear Sci. 2017, 27, 093940. [CrossRef][PubMed]127. Huang, X.; Song, Z.; Qu, Z. Determinants of early afterdepolarization properties in ventricular myocyte models. PLoS Comput.Biol. 2018, 14, e1006382. [CrossRef]128. Brown, D.A.; O’Rourke, B. Cardiac mitochondria and arrhythmias. Cardiovasc. Res. 2010, 88, 241–249. [CrossRef]129. Ruiz-Meana, M.; Fernandez-Sanz, C.; Garcia-Dorado, D. The SR–mitochondria interaction: A new player in cardiac pathophysiol-ogy. Cardiovasc. Res. 2010, 88, 30–39. [CrossRef]130. Cortassa, S.; Aon, M.A.; Marban, E.; Winslow, R.L.; O’Rourke, B. An integrated model of cardiac mitochondrial energy metabolismand calcium dynamics. Biophys. J. 2003, 84, 2734–2755. [CrossRef]131. Cortassa, S.; Aon, M.A.; O’Rourke, B.; Jacques, R.; Tseng, H.J.; Marban, E.; Winslow, R.L. A computational model integratingelectrophysiology, contraction, and mitochondrial bioenergetics in the ventricular myocyte. Biophys. J. 2006, 91, 1564–1589.[CrossRef]132. Zhou, L.; Cortassa, S.; Wei, A.-C.; Aon, M.A.; Winslow, R.L.; O’Rourke, B. Modeling Cardiac Action Potential Shortening Drivenby Oxidative Stress-Induced Mitochondrial Oscillations in Guinea Pig Cardiomyocytes. Biophys. J. 2009, 97, 1843–1852. [CrossRef]133. Matsuoka, S.; Sarai, N.; Jo, H.; Noma, A. Simulation of ATP metabolism in cardiac excitation-contraction coupling. Prog. Biophys.Mol. Biol. 2004, 85, 279–299. [CrossRef] [PubMed]134. Takeuchi, A.; Kim, B.; Matsuoka, S. The mitochondrial Na+-Ca2+ exchanger, NCLX, regulates automaticity of HL-1 cardiomyocytes.Sci. Rep. 2013, 3, 2766. [CrossRef]135. Aon, M.A.; Cortassa, S.; O’Rourke, B. Percolation and criticality in a mitochondrial network. Proc. Natl. Acad. Sci. USA 2004, 101,4447–4452. [CrossRef] [PubMed]136. Nivala, M.; Korge, P.; Nivala, M.; Weiss, J.N.; Qu, Z. Linking flickering to waves and whole-cell oscillations in a mitochondrialnetwork model. Biophys. J. 2011, 101, 2102–2111. [CrossRef] [PubMed]137. Pandey, V.; Xie, L.H.; Qu, Z.; Song, Z. Mitochondrial depolarization promotes calcium alternans: Mechanistic insights from aventricular myocyte model. PLoS Comput. Biol. 2021, 17, e1008624. [CrossRef] [PubMed]138. Song, Z.; Xie, L.-H.; Weiss, J.N.; Qu, Z. A spatiotemporal ventricular myocyte model incorporating mitochondrial calcium cycling.Biophys. J. 2019, 117, 2349–2360. [CrossRef]139. Xie, A.; Song, Z.; Liu, H.; Zhou, A.; Shi, G.; Wang, Q.; Gu, L.; Liu, M.; Xie, L.H.; Qu, Z.; et al. Mitochondrial Ca2+ InfluxContributes to Arrhythmic Risk in Nonischemic Cardiomyopathy. J. Am. Heart Assoc. 2018, 7, e007805. [CrossRef]140. Pandey, V.; Xie, L.-H.; Qu, Z.; Song, Z. Mitochondrial Contributions in the Genesis of Delayed Afterdepolarizations in VentricularMyocytes. Front. Physiol. 2021, 12, 1631. [CrossRef]141. Bers, D.M. Calcium cycling and signaling in cardiac myocytes. Annu. Rev. Physiol. 2008, 70, 23–49. [CrossRef]142. Saucerman, J.J.; Brunton, L.L.; Michailova, A.P.; McCulloch, A.D. Modeling beta-adrenergic control of cardiac myocyte contractilityin silico. J. Biol. Chem. 2003, 278, 47997–48003. [CrossRef]Biomolecules 2022, 12, 168619 of 19143. Xie, Y.; Grandi, E.; Puglisi, J.L.; Sato, D.; Bers, D.M. beta-adrenergic stimulation activates early afterdepolarizations transiently viakinetic mismatch of PKA targets. J. Mol. Cell. Cardiol. 2013, 58, 153–161. [CrossRef]144. Hashambhoy, Y.L.; Greenstein, J.L.; Winslow, R.L. Role of CaMKII in RyR leak, EC coupling and action potential duration: Acomputational model. J. Mol. Cell. Cardiol. 2010, 49, 617–624. [CrossRef] [PubMed]145. Hashambhoy, Y.L.; Winslow, R.L.; Greenstein, J.L. CaMKII-induced shift in modal gating explains L-type Ca(2+) currentfacilitation: A modeling study. Biophys. J. 2009, 96, 1770–1785. [CrossRef] [PubMed]146. Foteinou, P.T.; Greenstein, J.L.; Winslow, R.L. Mechanistic Investigation of the Arrhythmogenic Role of Oxidized CaMKII in theHeart. Biophys. J. 2015, 109, 838–849. [CrossRef] [PubMed]147. Liu, M.B.; de Lange, E.; Garfinkel, A.; Weiss, J.N.; Qu, Z. Delayed afterdepolarizations generate both triggers and a vulnerablesubstrate promoting reentry in cardiac tissue. Heart Rhythm 2015, 12, 2115–2124. [CrossRef] [PubMed]148. Campos, F.O.; Shiferaw, Y.; Prassl, A.J.; Boyle, P.M.; Vigmond, E.J.; Plank, G. Stochastic spontaneous calcium release events triggerpremature ventricular complexes by overcoming electrotonic load. Cardiovasc. Res. 2015, 107, 175–183. [CrossRef]149. Colman, M.A. Arrhythmia mechanisms and spontaneous calcium release: Bi-directional coupling between re-entrant and focalexcitation. PLoS Comput. Biol. 2019, 15, e1007260. [CrossRef]150. Colman, M.A.; Holmes, M.; Whittaker, D.G.; Jayasinghe, I.; Benson, A.P. Multi-scale approaches for the simulation of cardiacelectrophysiology: I—Sub-cellular and stochastic calcium dynamics from cell to organ. Methods 2020, 185, 49–59. [CrossRef][PubMed]151. Greene, D.A.; Kaboudian, A.; Wasserstrom, J.A.; Fenton, F.H.; Shiferaw, Y. Voltage-mediated mechanism for calcium wavesynchronization and arrhythmogenesis in atrial tissue. Biophys. J. 2022, 121, 383–395. [CrossRef] [PubMed]152. Walker, M.A.; Gurev, V.; Rice, J.J.; Greenstein, J.L.; Winslow, R.L. Estimating the probabilities of rare arrhythmic events inmultiscale computational models of cardiac cells and tissue. PLoS Comput. Biol. 2017, 13, e1005783. [CrossRef]153. Xie, F.; Qu, Z.; Yang, J.; Baher, A.; Weiss, J.N.; Garfinkel, A. A simulation study of the effects of cardiac anatomy in ventricularfibrillation. J. Clin. Investig. 2004, 113, 686–693. [PubMed]154. Ten Tusscher, K.H.; Hren, R.; Panfilov, A.V. Organization of ventricular fibrillation in the human heart. Circ. Res. 2007, 100,e87–e101. [CrossRef] [PubMed]155. Liu, M.B.; Vandersickel, N.; Panfilov, A.V.; Qu, Z. R-from-T as a common mechanism of arrhythmia initiation in long QTsyndromes. Circ. Arrhythm. Electrophysiol. 2019, 12, e007571. [CrossRef] [PubMed]156. Keener, J.P. The effects of discrete gap junction coupling on propagation in myocardium. J. Theor. Biol. 1991, 148, 49–82. [CrossRef]157. Qu, Z.; Karagueuzian, H.S.; Garfinkel, A.; Weiss, J.N. Effects of Na+ channel and cell coupling abnormalities on vulnerability toreentry: A simulation study. Am. J. Physiol. Heart Circ. Physiol. 2004, 286, H1310–H1321. [CrossRef]158. Huang, C.; Song, Z.; Qu, Z. Synchronization of spatially discordant voltage and calcium alternans in cardiac tissue. Phys. Rev. E2022, 106, 024406. [CrossRef]159. Greenstein, J.L.; Winslow, R.L. Integrative systems models of cardiac excitation-contraction coupling. Circ. Res. 2011, 108, 70–84.[CrossRef]160. Rice, J.J.; Jafri, M.S.; Winslow, R.L. Modeling short-term interval-force relations in cardiac muscle. Am. J. Physiol. Heart Circ.Physiol. 2000, 278, H913–H931. [CrossRef]161. Rice, J.J.; Winslow, R.L.; Hunter, W.C. Comparison of putative cooperative mechanisms in cardiac muscle: Length dependenceand dynamic responses. Am. J. Physiol. 1999, 276, H1734–H1754. [CrossRef]162. Silva, J.R.; Pan, H.; Wu, D.; Nekouzadeh, A.; Decker, K.F.; Cui, J.; Baker, N.A.; Sept, D.; Rudy, Y. A multiscale model linking ion-channel molecular dynamics and electrostatics to the cardiac action potential. Proc. Natl. Acad. Sci. USA 2009, 106, 11102–11106.[CrossRef]163. Yang, P.C.; DeMarco, K.R.; Aghasafari, P.; Jeng, M.T.; Dawson, J.R.D.; Bekker, S.; Noskov, S.Y.; Yarov-Yarovoy, V.; Vorobyov, I.;Clancy, C.E. A Computational Pipeline to Predict Cardiotoxicity: From the Atom to the Rhythm. Circ. Res. 2020, 126, 947–964.[CrossRef] [PubMed]164. Jensen, M.Ø.; Borhani, D.W.; Lindorff-Larsen, K.; Maragakis, P.; Jogini, V.; Eastwood, M.P.; Dror, R.O.; Shaw, D.E. Principles ofconduction and hydrophobic gating in K+ channels. Proc. Natl. Acad. Sci. USA 2010, 107, 5833–5838. [CrossRef] [PubMed]165. Qu, Z.; Garfinkel, A.; Weiss, J.N.; Nivala, M. Multi-scale modeling in biology: How to bridge the gaps between scales? Prog.Biophys. Mol. Biol. 2011, 107, 21–31. [CrossRef] [PubMed]166. Fox, R.F. Stochastic versions of the Hodgkin-Huxley equations. Biophys. J. 1997, 72, 2068–2074. [CrossRef]167. Keener, J.P. Invariant manifold reductions for Markovian ion channel dynamics. J. Math. Biol. 2009, 58, 447–457. [CrossRef]168. Landaw, J.; Qu, Z. Bifurcations Caused by Feedback between Voltage and Intracellular Ion Concentrations in Ventricular Myocytes.Phys. Rev. Lett. 2019, 123, 218101. [CrossRef]169. Qu, Z.; Shiferaw, Y.; Weiss, J.N. Nonlinear dynamics of cardiac excitation-contraction coupling: An iterated map study. Phys. Rev.E 2007, 75, 011927. [CrossRef] | RESEARCH GATE |
TYPE Original ResearchPUBLISHED 08 August 2022DOI 10.3389/fpsyt.2022.906935OPEN ACCESSEDITED BYMichael Klein,Catholic University of Applied Sciencesof North Rhine – Westphalia, GermanyREVIEWED BYAndré Luiz Monezi Andrade,Pontifical Catholic University ofCampinas, BrazilGianpiero Greco,University of Bari Aldo Moro, Italy*CORRESPONDENCENajam ul Hasan Abbasinhabbasi@gmail.comSPECIALTY SECTIONThis article was submitted toAddictive Disorders,a section of the journalFrontiers in PsychiatryRECEIVED 29 March 2022ACCEPTED 08 July 2022PUBLISHED 08 August 2022CITATIONAbbasi NuH and Muhammad S (2022)State resilience and rate of recoveryamong addicts: Moderating role ofsocial skills.Front. Psychiatry 13:906935.doi: 10.3389/fpsyt.2022.906935COPYRIGHT© 2022 Abbasi and Muhammad. This isan open-access article distributedunder the terms of the CreativeCommons Attribution License (CC BY).The use, distribution or reproductionin other forums is permitted, providedthe original author(s) and the copyrightowner(s) are credited and that theoriginal publication in this journal iscited, in accordance with acceptedacademic practice. No use, distributionor reproduction is permitted whichdoes not comply with these terms.State resilience and rate ofrecovery among addicts:Moderating role of social skillsNajam ul Hasan Abbasi1* and Saghir Muhammad21Department of Educational Sciences, Mianyang Normal University, Mianyang, China, 2Departmentof Psychology, International University, Islamabad, PakistanThe present study intended to explore the moderating role of social skills inState Resilience and the rate of recovery among drug addicts. The samplesize was 100 recovering addicts from diferent drug rehabilitation centers wererecruited from twin cities (Islamabad and Rawalpindi) of Pakistan. The SocialSkills Inventory (SSI) was used to assess social skills, while Connor-DavidsonResilience Scale (CD-RISC) and Substance Use Recovery Evaluator (SURE) wasused to assess state resilience and recovery among study participants. Resultssuggest a significant positive correlation between social skills and recovery(r = 0.27; p < 0.01). Similarly, state resilience was found to be significantlycorrelated with social skills (r = 0.35; p < 0.01), while state resilience wasnot significantly correlated with recovery. The moderation analysis for theinteraction of social skills on state resilience was non-significant [β = 0.002,95% CI (−0.002, 0.00), t = 1.01, p = 0.316]. Results also indicate the significantrelationship of social skills in the prevention of relapse against drug use.KEYWORDSsocial skills, state resilience, recovery, drug addiction, substance useIntroductionDrug abuse is a rising concern for Pakistan in the recent past. Reports on drug usein Pakistan suggest that Pakistan has 6.7 million drug users, and more than 4 millionof those numbers are addicts which is amongst the highest number for any country inthe world (1). It reports that substantial portions of population aging 15–64 are sufferingfrom overwhelming consequences of substance abuse. The report proposed that the rateof substance abuse is 5.8% which comprises of 6.4 million adults using drugs in theirlast 12 months and 4.1 million adults were concluded to be as dependent on drugsin Pakistan.According to Diagnostic and statistical manual of mental disorders (2) drugaddiction is a chronic and progressive illness which is linked with compulsive use ofdrugs and maintained by reinforcing path ways in the brain. Drug addiction is linkedwith biological, psychological, social and occupational impairments. Substance abuseis linked with developmental period especially with adolescence. Experimentation withdrug and initiation from tobacco progress to cannabis, heroin and depressants or otherpsychoactive drugs and that’s way tobacco use is known as a gateway to other drugs whichenhances the likelihood of addictive or problematic use of drugs.Frontiers in Psychiatry01frontiersin.orgAbbasi and Muhammad10.3389/fpsyt.2022.906935A sense of autonomy, peer pressure and idealizing negativemodels are associated with drug use in adolescence (3). Socialskills are defined as social intelligence involving adaptive andpositive social interactions (4). Social skills are linked with socialintelligence and socially adaptive functioning. Furthermore,social skills are either verbal (social) or non-verbal (emotional).Hence in present research social skills are defined andmeasured with six subscales of social and emotional expressivity,sensitivity and control through Social Skills Inventory (SSI)by summing up the total score of an individual. Zimmermanand Arunkumar (5) defined resilience as factors or mechanismsthat inhibit the potential risks to become a full-fledgedpsychopathology and enhance the adaptability at the stage ofadversity or generate adaptive outcomes. It is a capacity to adaptchallenging or threatening circumstances and the strength tobounce back in the face of adversity (6, 7).Hence in present study state resilience was measured byusing Connor-Davidson Resilience Scale (CD-RISC) (8). So, thehigher score indicated the higher state resilience in participants.Similarly, Botvin and Wills (9) found that acquisition ofeffective social skills is essentials for psychological adjustabilityand psychosocial development. So, the primary interpersonalskills are required to have confidence, responsiveness andmutually beneficial relationships as a sign of good psychosocialdevelopment but on contrary, inadequate or lack of socialcompetence can lead to rejection and social isolation whichfurther predicts poor psychological adjustment. The acquisitionof basic social skills generally begins in childhood and itincreases as individuals grow or mature with time. By thetime of adolescence, they have acquired a range of social skillssuch as effective communication, initiation and maintenanceof conversations, expression of feelings, giving and receiving ofcompliments, refusal of unreasonable requests. And these socialskills are learned by vicarious learning and reinforcement.Past literature on drug addiction suggests different factorswhich can cause drug addiction. Studies (10) showed thatthere are many identified risk factors which are linked withdrug abuse such as peer pressure, conforming to social circles,low positive parental relationship, marital discords, intellectualinferiority, emotional immaturity, poor self (inner) control,depressive mood, violation of existing social norms. Hair et al.(11) reviewed about 360 researches of social competence amongadolescents. The study stated that quality of relationships withparents is essential for the development of social competence.Moreover, social skills are also linked with psychologicalwellness, academic performance and interpersonal relationships.Von Hohendorff et al. (12) stated two factors individual(temperament and environment) for the development ofsocial skills.Resilience is a process that encompasses adaptability fromcrises. It is a process that helps us to cope significant adversity(13). Similarly according to (14) resilience has three occurrencessuch as generation of positive outcomes in vulnerable children,persistent competency during stress and recovery from trauma.They found that children who experience chronic adversity havebetter rate to recover completely due to the presence of positiverelationship with competent person, they also have the abilityto solve problems, to learn, to get along with others, and finallythey have competency along with perceived self-efficacy withrespect to society (15). Moreover Hiew et al. (15) stated thatresilience can be differentiated in adults on the basis of itscharacteristics whether these characteristics are present due tocurrent dominant states or it has been present in adults sincechildhood as personality trait [as cited by Bokharey (16)].Social skills are important for interpersonal effectivenesslike effectively communicating with other, reaching out,understanding other feelings and even get along with other.Similarly,sociabilityenhancessocialskills;itprovidesfoundation for social skills. However, as per observationwhile working with drug addicts, it is obvious that mostlydrug addicts are poor in interpersonal effectiveness and theyare more likely to compensate their psychological deficitsthrough drugs. So, it’s a substantial reason that by interveningin an addict through enhancing social skills to minimize thechances of relapse. Moreover, it can help them to effectivelyapproach their communication issues instead or drug abuse.There is no previously conducted research on social skills andresilience with respect to sociability of recovering addicts so it’sunique research that aims to investigate more important socialaspects which might predict relapse. This research helps inunderstanding the importance of social adaptability and opensnew ways to intervening in addicts’ life to prevent relapse.In the present study recovery is defined as individuals whoare getting outdoor treatment as follow up counseling aftercompleting indoor treatment from drug addiction and suchindividuals are in the state of complete abstinence. Hencein present study recovery was measured by using SubstanceUse Recovery Evaluator (SURE) against five major categoriessuch as drinking and drug use, self-care, relationships, materialresources, outlook on life (17).The current study intends to investigate the impact of socialskills and state resilience on the rate of recovery among drugaddicts and to further investigate the moderating effect of s socialskills on state resilience among drug addicts. It was hypothesizedthat social skills and state resilience are positively correlated witheach other and that the state resilience is negatively correlatedwith the use of drug among drug addicts. The study furtherhypothesized that social skills play a moderating role in stateresilience among recovering drug addicts.MethodsA correlation research method was used in to find out therelationship among social skills, sociability and state resilience.The sample consisted on 100 drug addicts recruited by usingFrontiers in Psychiatry02frontiersin.orgAbbasi and Muhammad10.3389/fpsyt.2022.906935purposive sampling. The age range of participants involved 3 agecohorts. These age groups were defined as late adolescents (18–24), early adulthood (24–34), and middle adulthood from (34–60) based on the Newman and Newman (18). The institutionalpermission was taken from International Islamic University’sethics review board before conducting the study.Inclusion criteriaOnly recovering addicts were approached from addictioncenters of Rawalpindi & Islamabad, who were dischargedaddicts (outdoor patients) and were receiving follow upcounseling for the relapse prevention. The education level of theparticipants was at least under graduation (O levels, A levelsand intermediate) and graduation that they could understandEnglish as a second language.Exclusion criteriaThe drug addicts who are still getting treatment as indoorpatients and/or drug addicts who left treatment or have beendischarged and do not receive any follow-up counseling were notincluded in the sample. Participants were selected as per Englishlanguage as a second language, and participants who don’t haveproper education but they were in the recovery process were notenrolled in the study.InstrumentsSocial skills inventoryDeveloped by Riggio (4) is a self-report measure used forassessing communication skills on two emotional (non-verbal)and social (verbal) dimensions. SSI consist six subscales ofdimensions such as emotional expressivity (EE) items no. 1,7, 13..., 85.; emotional sensitivity (ES) items no. 2, 8, 14...,86.; emotional control (EC) items no. 3, 9, 15..., 87.; Socialexpressivity (SE) items no. 4, 10, 16..., 88.; social sensitivity (SS)items no. 5, 11, 17..., 89.; and social control items no. 6, 12,18..., 90. SSI is a 5 points likert type scale continuum of “notat all like me (1), a little like me (2), like me (3), very much likeme (4), exactly like me (5)”. Every subscale consists at 15 itemsand every sixth item belongs to same subscale and score range isfrom 15 to 75.ResiliencewasmeasuredthroughConnor-davidsonresilience scale. CD-RISC originally developed by Connor andDavidson (8). CD-RISC has 25 items and participants ratedthemselves on 5-point Likert scale (0–4). Response continuumis range from not true at all = 0 to true nearly all the timestrongly agree = 4. Reliability of the scale is being reported asinternal consistency Cronbach’s α = 0.89. Similarly, the scalealso has effective convergent and discriminant validity.Substance use recovery evaluatorDeveloped by Neale et al. (17) as a valid measure formeasuring recovery from drug addiction. SURE, is a 5-pointLikert type scale. It comprises on 21 items. The scale also hasfive major categories such as drinking and drug use, self-care,relationships, material resources, outlook on life. Total ScoreRanges from 21-63. The SURE is a valid measure with good faceand content validity.Statistical analysisThe chronbach’s alpha was conducted to check scalesreliability, correlational analysis was conducted to investigate therelationship between study variables and moderation analysisusing SPSS-22 were conducted to check the moderating effectof social skills. The level of statistical significance was set a prioriat p < 0.05.ProcedureFor the purpose of data collection from the participants,institutional approval was sought first. Participants wereeducated properly about the nature and purpose of study.After the willingness in the research and signing of informedconsent they were selected for research. Instructions regardingthe instruments for data collection were also given. Quarriesof respondents were addressed accordingly. Participants wereguided about the response pattern of questionnaires accordingto given instructions. In the present study, no tangible orintangible incentives were given to participants.ResultsTable 1indicatesthealphavaluesfortheresearchinstruments. The calculated alpha values for Social SkillsInventory, CD-RISK and SURE (recovery) indicate that theseinstruments are reliable to measure social skills, state resilienceand recovery.Table 2 indicates that resilience is significantly positivelycorrelated with SSK (p < 0.01). But there is a non-significantrelationship between Resilience and recovery. Moreover, SSK isfound to be positively correlated with Recovery (p < 0.01).Table 3 represents non-significant moderation present byan interaction effect, indicating that relationship between stateresilience and recovery is not moderated by social skills.Frontiers in Psychiatry03frontiersin.orgAbbasi and Muhammad10.3389/fpsyt.2022.906935TABLE 1 Psychometric properties of the major study variables (N = 100).RangeVariablesMSDαPotentialActualSkewKurtosis1. SSKI284.633.240.811–5200–3790.200.112. CD_RISC73.5318.200.910–428–100−0.55−0.663. SURE51.208.250.851–328–63−0.44−0.46SSKI, Social Skills; CD_RISC, Connor-Davidson Resilience Scale; SURE, Substance Use Recovery Evaluator.TABLE 2 Correlation between resilience, social skills and recovery (N= 100).Measures1231. Resilience10.35**0.132. SSK–10.27**3. Recovery––1SSK, Social Skills; **p < 0.01.In order to interpret the moderation effect, simpleslopes are needed to be examined as they are shown inTable 4. Basically, the table represents results of three differentregressions: the regression for state resilience as a predictorof recovery (1) when social skills is low (i.e, when the valueof social skills is −33.24); (2) at the mean value of socialskills (when the mean value is zero); and (3) when thevalue of social skills is as high as 33.24. Interpretations ofthese three regressions are based on the value of b, andits significance.Three models are going to be interpreted as follows:(1) When social skills are low, there is a non-significantpositiverelationshipbetweenstateresilienceand recovery.(2) At the mean value of social skills, also there is a non-significant positive relationship between state resilienceand recovery.(3) When social skills is high there still remains non-significant positive relationship between state resilienceand recovery.These results tell us that the relationship between stateresilience and recovery are not only determined by differentlevels of social skills.Figure 1 shows the simple slopes analysis, graph indicatesthat: when the social skills is low (blue line) there is a non-significant positive relationship between state resilience andrecovery; similarly, at mean and high values of social skills (greenline, gray line) suggest that social skills does not significantlyenhance the relationship among state resilience and recovery atboth levels.TABLE 3 Linear regression using state resilience, social skills and theirinteraction as predictors of recovery.βSEβTPConstant50.74 (48.70, 82.79)1.0349.21<0.001Social skills0.05 (−0.03, 0.13)0.041.260.213State resilience0.03 (−0.11, 0.16)0.070.420.677Social skills × state resilience0.002 (−0.002, 0.00)0.0021.010.316R2 = 0.0230.TABLE 4 Conditional efect of state resilience on recovery at diferentvalues of social skills.Values of social skillsEffect of state resilienceon S.E recovery (β)TP−33.24 (−1 SD)−0.04, 0.07 (−0.19, 0.11)−0.580.5640.00 (M)0.03 (−0.11, 0.17)0.070.4233.24 (+1 SD)0.10 (−0.14, 0.34)0.120.85DiscussionThe present study investigated the impact of social skills andstate resilience on the rate of recovery among drug addicts andwhether social skills play a moderating effect on state resilienceamong drug addicts. It was hypothesized that social skills andstate resilience are positively correlated with each other and thatthe state resilience is negatively correlated with the use of drugamong drug addicts. The study further hypothesized that socialskills play a moderating role in state resilience among recoveringdrug addicts.It was hypothesized and proved that social skills arepositively correlated with recovery among drug addicts. Theresults indicated that those individuals who were high on socialskills stayed in recovery after the treatment from drug addiction.Past research showed that 11% drinkers and 50% problemdrinkers having incompetent social skills while none from thenon-drinkers (19).Similarly, a meta-analysis was conducted by Ennett et al.(20) evaluated eight studies to study the effect size of drugabuse resistance education (DARE): an educational programFrontiers in Psychiatry04frontiersin.orgAbbasi and Muhammad10.3389/fpsyt.2022.906935FIGURE 1Simple slope equations of regression of state resilience on recovery at three levels of social skills.which was taught by highly trained law enforcement officersat schools. There were six outcome measure classes such asknowledge about drugs, attitudes about drug use, social skills,self-esteem, attitude toward police, and drug use. Effect sizewas calculated by each outcome. Results showed that calculatedeffect size for social skills was significantly larger than attitudesabout drug use, self-esteem, attitude toward police, and drug use.Which indicated that knowledge and social skills were helpful inintervening against drug use (20).Another research conducted to review the 1,200 outcomestudies to investigate the relationships among protective and riskfactors for successful prevention programs. Findings suggestedthat personal and social skills, social norms, effective socialpolicies and social support are the protective factors forpreventing risks outcomes such as drug abuse, behavioralproblems, school failure, AIDS, physical abuse (21). Anotherresearch showed that the social skills problems are relatedto more peer victimization and friendship as social factorbuffer between social skills and peer victimization positively(22). Hence the results of above research also confirmedour hypothesis that social skills are positively correlated withrecovery. A review conducted by Fergus and Zimmerman (23),the review of the study in which 1,184 junior high schoolstudents of new York city indicates that psychological wellbeingand social competence played a significant role of protectivefactors against cigarette smoking, alcohol and marijuana use.Similarly, present study results suggest that social skillsand state resilience are positively correlated. Past literaturesupports that social skills are essential for effective adjustment, todevelop constructive interpersonal interactions. This is furtherassociated with social, community, family and career adjustmentand this social skill deficit is present in today’s adolescents(24). According to the Gardner [as cited in Thompson (24)]interpersonal and intrapersonal intelligences are also associatedwith social skills that help in understanding others behavior,handling relationships with others, empathies with others,managing own emotions and experiences and problem solvingskills. Social skills are also associated to refusal to drugs andalcohol, and saying no to premarital sex.The results of the current studies for moderation interactionwerenon-significantwhichshowedthatnon-significantinteractions of social skills were found in state resilienceamong recovering drug addicts. These results were notsurprising because there are few reasons. Firstly, the sample washomogeneous, 89% participants were from early and middleadulthood and there were only 3% females. Secondly, socialtriggers such as peers using drugs, conforming to social circles,to feel more sophisticated at socializing, to feel cool and grownup while doing drugs (25), establishing links with drug relatedsocial activities or social interactions are based on sociability.Alam et al. (10) also confirmed such social influence withdrug abuse.Frontiers in Psychiatry05frontiersin.orgAbbasi and Muhammad10.3389/fpsyt.2022.906935Moreover, according to Kring et al. (26) the social networksare linked with the use of drug and alcohol. Still, those whohave vulnerability to substance use disorders are more likelyto select social networks that are linked to their own drinkingor drug use patterns. Social network in which a person livesis associated with individual drinking, but individual drinkingalso linked with more drinking of the same social network.Indeed, studies showed that effects of social selection werestronger. Social selection indicates that people are most likelyto involve or choose social networks with drinking patterns liketheir own.Hence in the present study participants were recoveringdrug addicts and according to their relapse preventionprogram which is comprise of disease concept, habit formation,controlled environment with supervised family intervention,theyarerecommendedtolimittheirinteractionswithpersons who use any kind of drug or to exposure tosocial events involving drug use. So, in recovering addicts’sociability can lead to relapse and they need to develop morecareful and clean social interaction that was the reason thatsociability is not playing moderation in state resilience amongrecovering addicts. But as we look to the conditional effectof social skills on recovery at the low level of sociability,it showed that borderline effects were established, and aswe move to mean and high levels this conditional effectbecame non-significant.Several studies showed that sociability play a role incontinuation of cigarette smoking and drug use because theextraversion tendency to seek out such friends and peers whoare in drug abuse can be a risk factor instead of a protectivefactor against drug abuse. Hence a study was conducted byStein et al. (27) on initiation and maintenance of tobaccosmoking in adolescence and young adulthood. They studiedthe sample of 461 in a cross-sectional survey at every 4years. They assessed that smoking was positively related toextraversion, good social relationships, cigarette use of friendsand cheerfulness.Thirdly, in Pakistani culture, drug addiction is linked withbad personality or character so people are not disclosing thebehavior at all. There is a stigma against drug addiction noone even imagines to be known as an addict. They havehidden places where mainly the young adults are used tovisit and use drug in order to protect the prestige of theirfamilies. Even the families of drug addicts hide the matterfrom their relatives just because of bad name. Hence itcan be justified that sociability of drug addicts is related todrug related social settings. And as per the results of thestudy, non-significant results related to the moderating roleof sociability because during their recovery process they areunder the process to develop adequate healthy social interactionsas per their treatment protocol. So, there are chances thatdue to this process the interaction did not appeared tobe significant.LimitationsIn the present study sample was limited in number becauseof time constraint to conduct this research and sample alsolacking the gender equality. So, in the study there were onlythree female participants which was not the true representativeof the population. Furthermore, in the present study, mostparticipants were from early and middle adulthood that limitedthe effectiveness of generalization of results. Another limitationof the study is cultural adaptation of instruments, in the presentstudy participants were selected as per English language asa second language, and participants who don’t have propereducation but they were in the recovery process were notenrolled in the study. Sample of the study only comprised of100 participants which can be increased and may show thedifferent results.ConclusionFindings of the current study indicate that there is asignificantly positive relationship between social skills andstate resilience. Results also show that social skills aresignificantly correlated with recovery from drug addiction.As per the focus of the present study, it was expectedthat social skills play a moderating role in state resilienceamong recovering addicts. But calculated results showed non-significant moderation of social skills in state resilience.These non-significant results can be justified because ofhomogeneous sample and cultural differences of social skillsespecially social skills of Pakistani papulation in relation todrug addiction.Implication and recommendationFindings of the present study can be useful for theintervention strategies for preventing relapse and to enhance therate of recovery among drug addicts. Results are showing that ifwe focus on social skills in drug addicts and if we apply effectivemanagement of psychosocial stressors which are linked withthe state resilience of recovering addicts that will help healthprofessionals who are currently working with drug addictionand finally it will help us to prevent the subsequent relapse fromsubstance use. So, it will be very effective to develop some psycheducational based interventions to develop awareness aboutdrug addiction. It will also help in developing training programsto enhance students’ assertiveness and social skills so that theycan handle social stressors related to drug use.In order to validate the findings of present study it isrecommended that a research should be carried out on largersample to check the moderating role of sociability. It is furtherrecommended that equal number of distribution of male andFrontiers in Psychiatry06frontiersin.orgAbbasi and Muhammad10.3389/fpsyt.2022.906935female participants in every age cohorts can be required inorder to address the gender differences. In order to establishan effective consensus among the treatment effectivenessinter-rehabs comparison is needed for the betterment ofrelapse prevention.Data availability statementThe original contributions presented in the study areincluded in the article/supplementary material, further inquiriescan be directed to the corresponding author/s.Ethics statementThe studies involving human participants were reviewedand approved by International Islamic University EthicsReview Board. The patients/participants provided their writteninformed consent to participate in this study.Author contributionsNA and SM conceived of the study, collected, andanalyzedthedata.NAwroteandrevisedthepaper.Both authors contributed to the article and approved thesubmitted version.AcknowledgmentsThe authors are thankful to the research participants and therehabilitation institutions for being part of the research.Conflict of interestThe authors declare that the research was conducted in theabsence of any commercial or financial relationships that couldbe construed as a potential conflict of interest.Publisher’s noteAll claims expressed in this article are solely those of theauthors and do not necessarily represent those of their affiliatedorganizations, or those of the publisher, the editors and thereviewers. Any product that may be evaluated in this article, orclaim that may be made by its manufacturer, is not guaranteedor endorsed by the publisher.References1. UNODC. Report on Drug Use in Pakistan 2013 Reveals High Levels of DrugUse and Dependency (2013). Available online at: https://www.unodc.org/pakistan/en/report-on-drug-use-in-pakistan-2013-reveals-high-levels-of-drug-use-and-dependency.html (accessed February 20, 2022).2. Diagnostic and Statistical Manual of Mental Disorders. 5th ed. Washington,DC: American Psychiatric Association (2013).3.BotvinGJ.Substanceabusepreventionresearch:recentdevelopmentsandfuturedirections.JSchoolHealth.(1986)56:369–74. doi: 10.1111/j.1746-1561.1986.tb05775.x4. Riggio RE. The social skills inventory (SSI): measuring nonverbal and socialskills. The sourcebook of nonverbal measures: Going beyond words. (2005). p. 25–33.5.ZimmermanMA,ArunkumarR.Resiliencyresearch:implicationsforschoolsandpolicy.SocPolicyReport.(1994)8:1–20. doi: 10.1002/j.2379-3988.1994.tb00032.x6. Garmezy N. Children in poverty: resilience despite risk. Psychiatry. (1993)56:127–36. doi: 10.1080/00332747.1993.110246277. Garmezy N, Masten AS. The Protective Role of Competence Indicators inChildren at Risk. New York, NY: Psychology Press (1991).8. Connor KM, Davidson JR. Development of a new resilience scale:the connor-davidson resilience scale (CD-RISC). Dep Anxiety. (2003) 18:76–82. doi: 10.1002/da.101139. Botvin GJ, Wills TA. Approaches to substance abuse. Prevention research:deterring drug abuse among children and adolescents. (1985) 63:8.10. Alam I, Khan R, Jadoon MA, Asghar M, Shah Z. Does drugsaddiction thwart the psycho social fabrics of personality and family. Sarhad JAgric. (2007) 23:1225. Available online at: http://www.aup.edu.pk/sj_pdf/DOES%20DRUGS%20ADDICTION%20THWART%20THE%20PSYCHO.pdf11. Hair EC, Jager J, Garrett SB. Helping Teens Develop Healthy Social Skills andRelationships: what the Research Shows About Navigating Adolescence. Washington,DC: Child Research Trends Brief (2002). Available online at: http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.462.1553&rep=rep1&type=pdf12. Von Hohendorff J, Couto MCP, Prati LE. Social skills in adolescence:psychopathology and sociodemographic variables. Estudos Psicol. (2013) 30:151–60. doi: 10.1590/S0103-166X201300020000113. Luthar SS, Cicchetti D, Becker B. The construct of resilience: acritical evaluation and guidelines for future work. Child Dev. (2000) 71:543–62. doi: 10.1111/1467-8624.0016414.MastenAS,BestKM,GarmezyN.Resilienceanddevelopment:contributionsfromthestudyofchildrenwhoovercomeadversity.DevPsychopathol.(1990)2:425–44.doi:10.1017/S095457940000581215. Hiew CC, Mori T, Shimizu M, Tominaga M. Measurement of resiliencedevelopment: preliminary results with a state-Trait resilience inventory. JLearn Curr Dev. (2000) 1:111–7. Available online at: https://www.researchgate.net/profile/Chok-Hiew/publication/276268651_Measurement_of_Resilience_Development_Preliminary_Results_with_a_State-Trait_Resilience_Inventory/links/5553968508ae6943a86f2c8a/Measurement-of-Resilience-Development-Preliminary-Results-with-a-State-Trait-Resilience-Inventory16. Bokharey IZ. Resilience and coping strategies in the patients withconversion disorder and general medical conditions: a comparative study. Malays JPsychiatry. (2013) 22. Available online at: http://www.mjpsychiatry.org/index.php/mjp/article/download/224/17917. Neale J, Vitoratou S, Finch E, Lennon P, Mitcheson L, PanebiancoD, et al. Development and validation of’sure’: a patient reported outcomemeasure (prom) for recovery from drug and alcohol dependence. DrugAlcoholDepend.(2016)165:159–67.doi:10.1016/j.drugalcdep.2016.06.00618. Newman BM, Newman PR. Development Through Life: A PsychosocialApproach: Cengage Learning. Boston: Cengage Learning (2017).Frontiers in Psychiatry07frontiersin.orgAbbasi and Muhammad10.3389/fpsyt.2022.90693519.HoverS,GaffneyLR.Therelationshipbetweensocialskillsandadolescentdrinking.AlcoholAlcoholism.(1991)26:207–14. doi: 10.1093/oxfordjournals.alcalc.a04510220. Ennett ST, Tobler NS, Ringwalt CL, Flewelling RL. How effective isdrug abuse resistance education? A meta-analysis of Project DARE outcomeevaluations. Am J Public Health. (1994) 84:1394–401. doi: 10.2105/AJPH.84.9.139421.DurlakJA.Commonriskandprotectivefactorsinsuccessfulprevention programs. Am J Orthops. (1998) 68:512–20. doi: 10.1037/h008036022.FoxCL,BoultonMJ.Friendshipasamoderatoroftherelationshipbetweensocialskillsproblemsandpeervictimisation.AggresBehav.(2006)32:110–21.doi:10.1002/ab.2011423. Fergus S, Zimmerman MA. Adolescent resilience: a framework forunderstanding healthy development in the face of risk. Annu Rev Public Health.(2005) 26:399–419. doi: 10.1146/annurev.publhealth.26.021304.14435724. Thompson ED. Nurturing future generations: promoting resilience inchildren and adolescents through social, emotional and cognitive skills: Routledge.(2012). doi: 10.4324/978020395627425. Nevid JS, Rathus SA, Greene B. Abnormal Psychology in a Changing World.25. London: Pearson; Prentice Hall Press (2000).26. Kring A, Johnson S, Davison G, Neale J. Abnormal Psychology (12th ednDSM-5 update): Singapore: Wiley (2014).27. Stein JA, Newcomb MD, Bentler PM. Initiation and maintenance of tobaccosmoking: changing personality correlates in adolescence and young adulthood 1. JAppl Soc Psychol. (1996) 26:160–87. doi: 10.1111/j.1559-1816.1996.tb01844.xFrontiers in Psychiatry08frontiersin.org | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/360709226Towards assuring conformance of AI systemsConference Paper · May 2022CITATIONS0READS1652 authors:Some of the authors of this publication are also working on these related projects:Artificial Intelligence in Ophthalmology View projectIncreasing understanding on climate change through public discourse analyse and stakeholders modelling View projectIurian MarianUniversitatea Tehnica Cluj-Napoca8 PUBLICATIONS 7 CITATIONS SEE PROFILEAdrian GrozaUniversitatea Tehnica Cluj-Napoca162 PUBLICATIONS 372 CITATIONS SEE PROFILEAll content following this page was uploaded by Adrian Groza on 19 May 2022.The user has requested enhancement of the downloaded file.Towards assuring conformance of AI systemsC˘alin-Marian Iurian2 and Adrian Groza1Abstract—In the context of recent approval of the ArtificialIntelligence Act (AIA), AI applications classified as high riskneed to be validated by a third party against some conformancecriteria. We browse here the emerging standards for AIsystems. We also describe auditing and engineering methodsfor AI applications: SMACTR and CRISP-DM. We exemplifythe use of the above methods within a running scenario relatedto AI systems for human resources. The emerging standardsand quality auditing methods presented here can be a startingpoint for those who want to organize such certification bodiesfor high-risk AI systems.Keywords-Artificial Intelligence Act, Standards for AI sys-tems, Auditing AI systemsI. CONTEXTThe framework for regulating artificial intelligence inEuropean Union (EU) focuses on the concept of certifica-tion [1]. This certification will be obtained via conformanceassurance bodies accredited by a regulatory authority. Beingan New Legislative Framework type regulatory framework,the Artificial Intelligence Act (AIA) proposed by the Eu-ropean Commission establishes guidelines, while its opera-tionalization is based on standards.Regulation (EU) No 1025/2012 recognizes the standardsestablished by CEN, CENELEC, and ETSI as ”EuropeanStandards”. These groups work together to promote Eu-ropean standardization by developing market-driven andunified standards to complement European law. The Inter-national Organization for Standardization (ISO), the Inter-national Electrotechnical Commission (IEC), and the Inter-national Telecommunication Union (ITU) are their interna-tional equivalents [2] [3].The subject of whether AI quality requirements demandindependent testing and what standards should be set forsuch testing processes emerges during the audit of AIsystems. To remedy this absence of standards, for example,the German standardization roadmap on AI lists [4] the mostimportant dimensions of quality which should be addressedby standardization (Fig. 1).II. EMERGING STANDARDS FOR AI SYSTEMSThere are up to 50 emerging standards in AI plannedto be finalised in the next years. Table I lists the firstAI standards currently being developed in the areas of1ComputerScienceDepartmentTechnicalUniver-sityofCluj-Napoca,28Memorandumului,Cluj-Napoca,RomaniaMarian.Iurian@com.utcluj.ro,adrian.groza@cs.utcluj.roFigure 1: AI quality criteria according to “German Standard-ization Roadmap on Artificial Intelligence [4]reliability, robustness, safety, and security. However, it isapparent that the field of technical testing still has a lotof room for improvement. (”product testing”), particularlyin terms of neural network validation and verification, aswell as credible safety reasoning for key systems andtesting tools to conduct these tests. An important exampleof addressing this need for self-government is the Germanproject ”KI-Absicherung” – Safe AI for Automated Driving.It is run by a partnership of research institutes, automakers,suppliers, standards bodies, and significant public adminis-trations (such as BSI Germany) and produces a commonindustry agreement on verification methodologies for highlyautomated AI-based driving modules’ safety.The trade-offs between convenient system features, suchas robustness, security, safety, and audibility, on the onehand, and AI model, machine learning (ML) algorithm, data,and boundaries, such as the intricacy of the model, roomfor activities, plasticity, performance and cost, on the otherhand, are a general open problem. Current AI systems’scalability and generalization are limited by these trade-offs. Some instances of such quality compromises are: (i)increasing the complexity of the model can have a negativeimpact on interpretability and security; (ii) As the size ofthe workspace grows, so does the requirement for more testand validation data; (iii) Strengthening security often leadsto poor performance; (iv) The white box model and theaccess to the lifecycle to improve auditing conflicts mayTable I: Examples of emerging standards in artificial intelligenceAreaStandardDescriptionReliabilityISO/IEC NP 24029Assessment of robustness of neural networks [5]and Robust-nessITU-TF.AI-DLFE&F.AI-DLPBMethod for evaluating the deep learning and metric software framework and evaluation methods for theDNN processor benchmark [6]ETSI DTR INT 008AI in evaluating systems and AI model testing, as well as quality metric definitions (TR 103 821) 1DIN SPEC 92001-2Processes and quality standards for the AI life cycle, Part 2, robustness2ITUITU Focus Group on ”Artificial Health Intelligence” [7], [8]SafetyISO / CD TR 22100-5Safety of machinery, Part 5: Implications of embedded systems AI – ML [9]ISO 26262Road vehicles - functional safety [10], see also IEC 61508-1: 20103, ISO 21448 [11])IEEE P2802Evaluating the performance and safety of medical AI devices - terminology [12]ISO / IEC AWI TR 5469AI - functional safety and AI systems4SecurityISO / SAE 21434Road vehicles - cyber security engineering, ISO / IEC 23894 [13])ETSI ISG SAISeveral publications for generic AI systems that include the issue description, threat ontology, securitytesting, and mitigation solutions [e.g. ETSI GR SAI 005 V1.1.1 (2021-03)].NISTIR 8269A contradictory ML taxonomy and terminology [10]Ethics andIISO/IEC DTR 24027Information Technology – Artificial Intelligence – Bias in AI systems and AI aided decision makingBiasIEEE 7002Model Process for dealing with ethical issues during system developmentMitigationsIEEE 7002Data Privacy ProcessIEEE 7003Algorithmic Bias ConsiderationIEEE 7006Standard on Personal Data AI Agent Working GroupISP/IEC WD 5249For analytics and machine learning, data quality is essentialconflict with intellectual property interests (v) Independentdata collections and pre-trained models save expenses, butalso provide additional risks, particularly for difficult-to-detect back assaults.In case of ML systems, audit can investigate the followingdimensions:1) Purpose of the AI algorithm. A1: What other algo-rithms have been applied to achieve similar goals,and what are the potential issues associated withthese algorithms? A2: What are the key factors indetermining the results of algorithms? A3:What arethe possible regulatory constraints in achieving theobjective with and without the application of AI?2) Obtaining data. O1: Where does the data come from?O2: Were the training data used in the model / machinetraining consistent with the original data sources? O3:Does the data come from a trusted entity? O4: Whatare the possible inconsistencies or other issues withdata sources, such as methodological changes for datacapture over the years and the quality issues of oldsystems? O5: What criteria were used to collect thedata? O6: Are there other similar data sources that arenot selected or used to drive the model?3) Pre-processing of data. P1: How was the missing dataentered? P2: What were the criteria for deleting miss-ing data? P3: How were the training and testing datasets selected (random stratification, cross-validation,etc.)? P4: How was the data standardized?4) Data modeling. D1: What other AI techniques havebeen considered, what are the results and possible rea-sons why they were not selected? D2: What were thecriteria and assumptions considered for the analysis ofthe algorithms? D3: Was the design of the algorithmsfrom scratch? D4: Do algorithms use programminglibraries, and if so, how reliable are they?5) Testing. T1: What metrics were used to test the ac-curacy of the model? T2: The results generated bythe algorithms were sensitive to minor changes in themodeling characteristics?6) Implementation. I1: How was the model installed inproduction and did it involve third parties? I2: Wasthere an accuracy review after the implementation ofthe algorithms by a third party? I3: Do the algorithmsmeet the objective set at the beginning of the auditprocess?7) Monitoring results. M1: Does the entity have adequatestructures, processes and procedures in place to direct,manage and monitor AI activities? M2: What actionshave the responsible parties taken at each stage ofthe process to ensure that AI activities comply withrelevant laws and regulations, in line with the organi-sation’s objective, and to maintain an appropriate levelof ethics, social responsibility and accountability?III. AUDITING AND DATA MINING METHODSWe present here the SMACTR audit framework and theCRISP-DM data mining method for AI systems.A. SMACTRThe Scoping, Mapping, Artifact Collection, Testing, Re-flection method (SMACTR) is the audit framework sug-gested by Google AI and the AI Partnership. The methodol-ogy is divided into five stages and is based on a collectionof papers (Figure 2) created by the auditor and developmentgroups.The Scoping establishes the audit’s scope by assessingthe system’s motives and intended impact, as well as vali-dating the concepts that will drive its development. This isaccomplished by an ethical evaluation that must entail rangeFigure 2: The SMACTR method. The white parts represent processes, whereas the colorful sections represent papers. Theauditors create the yellow documents, the engineering teams create the blue documents, and the green outcomes are generatedtogether.of viewpoints in order to minimize ”codification” biases, aswell as a social effect analysis. The goal is to find answers toissues such ”How might utilizing AI impact people’s lives?”and ”What are the possible social, economic, and culturalpresumptions?”The Mapping stage examines data on many participantsas well as the product development process. It is focusedon a collaborative map, a review of existing data, andthe findings of an ethnographic field investigation (donethrough interviews with key members of the organization’smanagement team). This research should help researchersunderstand better how specific decisions have been made,such as the data set or model architecture, and how theywould affect the dynamic behavior of the system.In the Artifact Collection stage, the auditor develops acatalog of all documents that should have been createdthroughout development and are required to commence theaudit. This incorporates card models and technical docu-mentation, two standards that work together to improve al-gorithm constancy. The model’s effectiveness characteristicsare described on model cards. The data sheets, in particular,examine the data collecting process with the goal of assistingthe data set user in making accurate judgments.During the Testing phase, auditors engage with the plat-form to determine whether it is consistent with the or-ganization’s ethical ideals. They can utilize contradictinginstances (in which the system is ”tricked” by feeding infake input data in order to observe the outcomes produced),and provide bogus user profiles, particularly from groups, tocheck for biased findings. This phase also includes an ethicalrisk assessment diagram, which examines the chance of abreakdown (calculated based on the incidence of particularfailures seen during testing) and the magnitude of thefailure (determined in earlier steps) to evaluate the risk’ssignificance.Finally, the Reflection step entails comparing the producedfindings to predetermined ethical norms and delivering a riskanalysis that outlines the principles that may be threatenedduring the implementation of the system. The auditors alsowill recommend a strategy for mitigating these risks, withthe objective of achieving a predetermined risk tolerance.What amount of parity should be attained, for example, ifauditors uncover inconsistent subgroup performance of theclassifier?B. CRISP-DMCross-industry standard process for data mining [14],known as CRISP-DM, is a standard cyclic process modelthat describes a structured approach to the entire data lifecycle, being the most widely used analysis model. TheCRISP-DM framework was introduced to help auditorsperform high-level machine learning audit. It serves notonly as a roadmap for how to extract and analyze data, butalso to enhance the possibility of professional collaboration.The CRISP-DM standard consists of six different phasesillustrated in Figure 3.1. Business understanding aids in determining whichfeatures of a use case, like income potential, job title,and educational level, should be incorporated in a businessmodel. In complicated situations, where other types ofmodels, such as software or mental decision models, havealready been utilized, this stage could be a bit more fluid thanbeginning from zero. Because the CRISP-DM architectureis continuous, the part on business understanding will beupdated often.2. Data understanding goal is to comprehend all factorsand verify that they are neither inconsistent or biased.The correlation and covariance matrices may be used tounderstand how variables are connected to one another andhow they fluctuate in reaction to one another. The location ofdata storage is critical. The architectures, systems, and datatypes of various data centers may vary. Algorithmic modelsFigure 3: The 6 phases of the CRISP-DM methodand subsequent changes cannot be conducted without a goodcomprehension of the data.3. Data preparation validates how the data was pre-processes. For instance, in the case of unorganized text, suchas log files and scrapped data on the site, the pre-processingstep can be challenging. To make the data comparable, itmust be scaled to the same scale across all attributes ordimensions. The standardization of the score z is usuallyused, obtaining an average of µ = 0 and a standard deviationσ = 1.4. Modeling is a vital component of ML, but in mostprojects it is one of the simplest steps to take. Recentbreakthroughs in machine learning are increasingly beingutilized to great model hyperparameters and choose thebest method. The auditor should examine the chosen modelto verify if it satisfies the interpretability criteria for thespecific use case. According to the ”right to explanation”section in the General Data Protection Regulation (GDPR),the company must justify each of its actions. One clearbenefit of GDPR is that it has placed more emphasis onthe transparency of algorithms and the way they work.5. Evaluation is a key concern from an audit perspective.If the model has a high prediction accuracy (90 %) andlooks to be well-generalized, it may not satisfy the model’sobjectives and/or contradict business policies, such as racialdiscrimination by accident. In case any unforeseen conse-quences emerge, the auditor should additionally produce adata samples set to evaluate the algorithm’s results. Theauditor, for example, may assemble a compilation of datafrom various neighborhoods to aid in loan approval.6. Deployment of the AI system is less worrying to theauditor if the expected degree of service and capabilities isattained. However, there is some value in auditor awarenessand examination in certain areas. Because of the ”learn-ing” nature of machine learning, detecting and resolvingtechnical debt is more complex than in standard softwareengineering projects. Technical debt is a sort of debt that,if not addressed, can become a problem. Corrections in acascade might exacerbate the situation. When an algorithmfails to generate the expected output, a correction cascadedevelops, in which rule-based adjustments are made to themodel to rectify its flaws. These problems may be the resultof previous cases or may have been caused by poor modelingor deficient training or validation data. Because the filtersare at the top of the results and often one higher limit givesa head start in terms of the model’s capacity to learn, itbecomes more difficult to discern what adjustments to themodel contribute to advancement.IV. RUNNING SCENARIO: AI IN RECRUITINGArtificial Intelligence Act includes AI systems used inemployment into the category of high-risk systems. That isbecause such systems affect the chances of natural personsto career development. Consequently, AI systems used inrecruiting will require a conformance certification.Assume the following ML-based recruiting system mod-elled under the CRISP-DM framework.1. Business understanding: The efficiency of the pro-cess can be increased by automating the examination ofthe applicant file to identify whether candidates fulfill thejob specifications. The company’s goal is to manage jobapplications using internet recruiting methods.2. Data understanding: To assess the relevance scoresof the candidates and derive their personality features, thesuggested technique uses objective criteria collected fromtheir LinkedIn profiles and personal criteria extracted fromtheir social presence.3. Data preparation: Applicants were randomly selectedthrough the Google Blog Search API, with the only require-ment being a technical background and a LinkedIn profile.4. Modeling: The first 100 candidates that met the el-igibility requirements were selected from the search API.Three technical positions are considered: sales engineer, ajunior programmer and a senior programmer. The candidatemay connect to the system using their LinkedIn logininformation, which allows the system to pull all objectiveselection criteria from their LinkedIn profile automatically.5. Evaluation: Unemployed persons who have registeredwith public employment services or personal blogs can usedata mining technology and the CRISP-DM approach to finda trend related job accessibility. A person is ready for workif he has some job experience and is willing to work fulltime, according to a basic link between consumer attributes.Otherwise, the team decides to pursue more training tobetter their talents and get more experience. This pattern isimmediately recognized, and during the data set extractionprocedure, no newly unknown information was revealed.Experienced and qualified users may seek training to modifytheir professional path in specific instances. Finding soliddata to work with was the most difficult aspect of job prepa-ration. As a result, in order to produce a good conclusion,the phases of comprehending and preparing the data have tobe interlaced.6. Deployment: Recruiters can navigate the recruitmentsection after confirming their login details. This enablesthem to advertise new job openings and assess candidates.A list of all open positions is sent to the recruiter, togetherwith a list of individuals who have submitted for each one.According to the recruiter’s request, the system calculatesthe relevance scores of candidates and classifies them ap-propriately. Recruiters can adjust an applicant’s rank byawarding relevance ratings to that candidate using the e-recruitment system. The system’s future performance willbe aided by the recruiter’s ideas being incorporated in thesystem’s training set and the ranking model being updatedon a regular basis.When a machine learning algorithm is deployed in thefield, it can consistently collect data from a set of attributesat a certain time or data stream, depending on the situation.A distinct model is assumed in this example, with just oneset of characteristics supplied to the model in a row. In bothcases, the auditor may be able to test the input parameters ofa pseudo-data set that is entered into the algorithm, and theexpected results on the study’s characteristics may contributeto possible disruptions, such as a discriminatory predictionloan model of a breed group based solely on zip code.Without explicitly describing how or why the algorithmgenerates a certain prediction, one might acquire confidencein probable model performance variations by entering datawithin a range of scenarios. By repeating this sampling serieswith the same data, we can determine the practical accuracyof the model related to the mathematical efficiency usedwhen the model was trained by data analysts.The auditor should consider paying attention to severalbiases. First, the sample bias assumes that a subgroup isover or underrepresented in the training data. In our case,there more young candidates for the sales position in thetraining dataset. Second, the labeling bias occurs when thepersonnel used to manualy lable the trainign date is biased.For instance, the labeling is perfromed only by eleder men.Third, the proxy bias occurs when some attributed to nothave direct connection with the label. For instance, thelength of CV, should not be considered as a feature in theML training.In the context of AI in human resources or talent man-agement, standards addressing discrimination (bias) apply,including the emerging IEEE 7000 series Model Processfor Addressing Ethical Concerns During System Design andP7002 Data Privacy Process, P7003 Algorithmic Bias Con-siderations, P7005 Standard on Employer Data Governanceand P7006 Standard on Personal Data AI Agent WorkingGroup, or ISO/IEC DTR 24027 Bias in AI systems and AIaided decision making.Features like age, gender, religious are considered pro-tected attributes. To avoid discrimination risks, such at-tributes should not appear in the data-set. Moreover, at-tributes that are highly correlated with any protected attributeshould be identified by the auditor since they operate as aproxy for the protected attributes. As an example, maritalstatus indicating married with a man, is enough to deducewith high probability the gender of the candidate. TheCRISP-DM handles such issues during the Data understand-ing step. The SMACTR addresses these issues in the Scopingphase (e.g. social impact assessment). The standard for MLin Section II uses the question A3 to signal that the AIA inEU or the Algorithmic Accountability Act in US introduceregulatory constraints on such attributes. Analysing CVsalso implies handling missing data, since candidates havedifferent experiences and interests. Question P1 addressesthis aspect. The evaluation of machine learning classificationefficiency may be evaluated against the ISO/IEC TS 4213Information Technology - Artificial Intelligence when ratingapplicants.According to AIA, AI classified as high risk shouldprovide their technical documentation within a single doc-ument [2]. The SMACTR internal audit method does notmeet this criteria, since it is based on a large collectionof documents. A forward step is required to merge thedocuments containing technical aspects.V. DISCUSSION AND RELATED WORKIn 2017, the Institute of Internal Auditors (IIA) proposedan AI auditing framework based on three major components:governance, strategy and the human factor [15]. The strategytakes into account all efforts to determine the most effectiveways to optimize AI activities. The Romanian guide onthe ”Implementation of International Standards on AuditingInternal” in the updated 2019 edition represents the initiativeof the Romanian Chamber of Financial Auditors (CAFR) tomeet the requirements of the Mandatory Norms issued byGlobal IIA and adopted by CAFR [16].The present Explainable AI (XAI) movement argues forthe development of easily understandable machine learningmodels that allow people to understand why a certain cate-gorization was reached. This is significant because in manydomains, such as medicine, human operators will not believean algorithmic outcome unless they can understand how itwas generated. There is a body of recent literature regardingtrustworthy AI and ML standardization. Winter et al. haveintroduced an audit framework for ML applications [17].It includes certificates for analyzing and testing machinelearning systems from a variety of perspectives, includingdata security, safe software development, practical needs,and ethics. Filip has examined the challenges of developingglobal applicable international standards for trustworthyAI [18]. A ML for health audit framework is proposed bythe ITU/WHO Focus Group on Artificial Intelligence forHealth [19] covering three use cases: diagnostic predictionof diabetic retinopathy, diagnostic prediction of Alzheimer’sdisease, and cytomorphologic classification for leukemiadiagnostics. It also includes three factors of AI such as bias,interpretability, and robustness.The creation of new procedures, structures, and tech-nologies is required for responsible AI. Boza et al. haveanalysed the current state of resources and tools available toorganizations to keep with the upcoming AI regulations andimplement AI principles in practice [20]. Machine learningalgorithms may be audited using a variety of frameworks,like LIME [21] (Local Interpretable Model-Agnostic Ex-planations) and FairML [22]. These frameworks, however,only assist the interpretation of model weights, yet theynot provide a comprehensive risk-based understanding ofthe ML mechanism. The CRISP-DM method is a way ofmanaging risks that use probabilities. During the assessmentphase, the LIME and FairML techniques may be used withthe CRISP-DM framework to assist the auditor in betterunderstanding the model.VI. CONCLUSIONThis preliminary work analyses the emerging standardsin AI and also the SMACTR method for auditing andCRISP-DM for engineering AI systems. An AI recruitingapplication is discussed in relation with these standards inthe context of the current Artificial Intelligence Act. Theemerging standards and auditing methods presented can bea starting point for the researchers interested in assuranceconformance of AI systems.REFERENCES[1] M. U. Scherer, “Regulating artificial intelligence systems:Risks, challenges, competencies, and strategies,” Harv. JL &Tech., vol. 29, p. 353, 2015.[2] S. Nativi and S. De Nigris, “Ai standardisation landscape:state of play and link to the ec proposal for an ai regulatoryframework, eur 30772 en,” 2021.[3] B. e. a. Christian, “Towards Auditable AI Systems,” FederalOffice for Information Security, Tech. Rep., May 2021.[4] W. Wahlster and C. Winterhalter, “German standardizationroadmap on artificial intelligence,” DIN/DKE, Berlin/Frank-furt, 2020.[5] ISO, “ISO/IEC AWI 24029-2 Artificial intelligence (AI) —Assessment of the robustness of neural networks — Part 2:Methodology for the use of formal methods,” Tech. Rep.,2021.[6] I. T. U. F.AI-DLFE, “”Deep Learning Software FrameworkEvaluation Methodology,” Tech. Rep., June 2020.[7] L. Oala, P. Balachandran, F. Cabitza, S. C. Ramirez et al.,“Ml4h auditing: From paper to practice,” in Data and artifi-cial intelligence assessment methods (DAISAM),. ITU/WHOFocus Group on Artificial Intelligence for Health (FG-AI4H)- Meeting I 2020, 2020.[8] E. Sch¨orverth, S. Vogler, P. Balachandran, A. W. Leite et al.,“FG-AI4H Open Code Initiative - Evaluation and ReportingPackage, ITU/WHO Focus Group on Artificial Intelligencefor Health (FG-AI4H),” Tech. Rep., 2021.[9] “ISO/TR 22100-5:2021 Safety of machinery — Relationshipwith ISO 12100 – Part 5: Implications of artificial intelligencemachine learning,” Tech. Rep., 2021.[10] E. Tabassi, K. Burns, M. Hadjimichael, A. Molina-Markham,and J. Sexton, “A taxonomy and terminology of adversarialmachine learning,” NIST IR, 2019.[11] ISO, “ISO 21448 Road vehicles - Safety Of The IntendedFunktionality,” Tech. Rep., 2019.[12] “IEEE. P2802: Standard for the Performance and SafetyEvaluation of Artificial Intelligence Based Medical Device,”Tech. Rep., 2018.[13] “ISO/IEC 23894 Information Technology - Artificial Intelli-gence - Risk Management,” Tech. Rep., May 2021.[14] A. Clark, “The machine learning audit-crisp-dm framework,”ISACA Journal, vol. 1, pp. 42–47, 2018.[15] T. I. of Internal Auditors, “Artificial Intelligence - Consider-ations for the Profession of Internal Auditing,” Tech. Rep.,2017. [Online]. Available: https://na.theiia.org/periodicals/Public%20Documents/GPI-Artificial-Intelligence.pdf[16] A. A. I. din Romania, “Ghid privind implementarea Standard-elor international de audit intern 2019,” Tech. Rep., 2019.[17] P. M. Winter, S. Eder, J. Weissenb¨ock, C. Schwald, T. Doms,T. Vogt, S. Hochreiter, and B. Nessler, “Trusted artificialintelligence: Towards certification of machine learning appli-cations,” arXiv preprint arXiv:2103.16910, 2021.[18] D. Filip, L. Dave, and P. Harshvardhan, “An ontology forstandardising trustworthy ai,” 2021.[19] L. Oala, J. Fehr, L. Gilli, P. Balachandran, A. W. Leite,S. Calderon-Ramirez et al., “Ml4h auditing: From paper topractice,” in Machine Learning for Health.PMLR, 2020,pp. 280–317.[20] P. Boza and T. Evgeniou, “Implementing ai principles: Frame-works, processes, and tools,” 2021.[21] S. Mishra, B. L. Sturm, and S. Dixon, “Local interpretablemodel-agnostic explanations for music content analysis.” inISMIR, 2017, pp. 537–543.[22] J. A. Adebayo et al., “Fairml: Toolbox for diagnosing biasin predictive modeling,” Ph.D. dissertation, MassachusettsInstitute of Technology, 2016.View publication stats | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/339167337Simply Seismological and Radar Interferometric Assessment of the January,22,2020 Akhisar (Manisa) Earthquake (Mw=5.4) (Western Turkey)Technical Report · February 2020CITATIONS0READS1161 author:Some of the authors of this publication are also working on these related projects:Seismic Hazard Map of Turkey View projectGONAF Project View projectFiliz Tuba KadirioğluAFAD, Disaster and Emergency Management Presidency51 PUBLICATIONS 295 CITATIONS SEE PROFILEAll content following this page was uploaded by Filiz Tuba Kadirioğlu on 11 February 2020.The user has requested enhancement of the downloaded file.1 Simply Seismological and Radar Interferometric Assessment of the January, 22,2020 Akhisar (Manisa) Earthquake (Mw=5.4) (Western Turkey) Filiz Tuba KADİRİOĞLU (M.Sc. Geologist) An earthquake occurred on January, 22, 2020 in Akhisar-Manisa (Western Turkey) with magnitude 5.4. After the main shock, approximately, 2530 aftershock occurred with magnitude varying between 0.6 and 4.8 within 15 days after the mainshock. While the majority of the aftershocks distributed in an area of approximately 20 km in line with NNW-SSE direction between the Gelenbe Fault Zone and Soma-Kırkağaç Fault Zone, a few of them occurred at the southern end of the Akhisar Fault (Figure 1). Figure 1: 22.01.2020 Akhisar Manisa Eq. and aftershock distributions (Black line indicates Bakır Segment of the Soma-Kırkağaç Fault Zone Demirtaş, R. 2013 ). In the focal mechanism solutions for M>=4.0 earthquakes, normal and strike-slip component are seen together. These results are evaluated in accordance with the tectonics of the region (Figure 2). Figure 2: Fochal mechanism solutions of the M>=4.0 earthquakes (Kartal,R.F 2020) 2 While the focal depth of the earthquakes changing between 0.38 and 26.97, they are concentrated from 5 to 10 km. In the AA’ depth cross section taken perpendicular to the aftershocks and in 3D image view, there is no any finding of the dip direction of the fault causing the earthquakes were detected (Figure 3). Figure 3: Focal depth distribution of the aftershocks a-b:cross-section view, c:3D view (Zmap Software Wiemer, S.2001) In seismological analysis using aftershocks, completeness magnitude and b value were calculated as Mc 1.5 and 0.71 +/- 0.02 respectively. On the other hand, after the main shock, the increase in the number of earthquakes continues cumulatively (Figure 4a-b). Figure 4a-b: Magnitude completness and cumulative rate of the aftershocks 3 After the main shock, in order to observe potential co-seismic deformation, it is benefited from C-band (~6 cm wavelength) Sentinel 1 A/B data. Including pre and post earthquake, 2 images were evaluated on Ascending Track 131 (10/01/2020-28/01/2020) by the help of the GmtSAR Parallel Program (Çakır, Z. et al 2018) (Each color fringe indicates 28mm deformation) (Figure 5-6). The calculated deformations after the unwrapping process of T131 were given in figure7a-b. Figure 5: Matching 2 images pairs with the GmtSAR Parallel Program (10/01/2020-28/01/2020) Figure 6: Obtained interferogram after the Mw=5.4 Eq. ( Sentinel 1A-B T131 Ascending results) 4 Figure 7: (a) Unwrapped phase process performed with GmtSAR Parallel Program. (b) Approximate amount of deformation (~ 5 cm ) was calculated with Mirone software. (Cıma, Luis J.F., 2007) As a result; When the obtain interferograms , aftershocks and focal mechanism solutions have been evaluated together, the fault that causing the Mw=5.4 earthquake is thought to be the Bakır Segment of the Soma-Kırkağaç Fault Zone (Demirtaş, R. 2013). Which is normal fault with left lateral strike slip component located to the west of the epicentre. The segment proposed as a fault that caused the earthquake is also compatible with the InSAR results. a) b) View publication stats | RESEARCH GATE |
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/331322914Tin-oxide nanoparticles deposited from a beam: What happens to thecomposition?Article in Physical Chemistry Chemical Physics · February 2019DOI: 10.1039/C8CP06168HCITATIONS2READS17412 authors, including:Some of the authors of this publication are also working on these related projects:Growth of GaAs and GaAsP NWs in Aerotaxy for Photovoltaics View projectPolymer-supported Pd (II) for C-H activation of arenes View projectMikko-Heikki Johannes MikkeläMAX IV Laboratory31 PUBLICATIONS 278 CITATIONS SEE PROFILEChaofan ZhangNational University of Defense Technology58 PUBLICATIONS 2,342 CITATIONS SEE PROFILEOlle BjörneholmUppsala University273 PUBLICATIONS 7,384 CITATIONS SEE PROFILEErik MårsellLund University25 PUBLICATIONS 336 CITATIONS SEE PROFILEAll content following this page was uploaded by Stacey L Sorensen on 26 June 2019.The user has requested enhancement of the downloaded file.This journal is ©the Owner Societies 2019Phys. Chem. Chem. Phys., 2019, 21, 6287--6295 | 6287Cite this: Phys.Chem.Chem.Phys.,2019, 21, 6287Tin-oxide nanoparticles deposited from a beam:what happens to the composition?M. Tchaplyguine,*a C. Wright,b A. Shavorskiy,a S. Zhu,a M.-H. Mikkela¨,a C. Zhang,cO. Bjo¨rneholm,c E. Mårsell,b A. Mikkelsen,b S. Sorensen,b C. J. D. Hetheringtondand L. R. WallenbergdThe debate around the oxidation states occurring in laboratory-prepared tin-oxide samples has been for along time an obstacle for an unambiguous assignment of characterization studies performed on suchsamples. In particular the changes in the Sn core-level energies caused by oxidation – i.e. the chemicalshifts – as measured by photoelectron spectroscopy (PES) have been under discussion. The assignment pro-blem is especially pronounced for nanoscale structures, which are important for photovoltaics, electronics,catalysis, and gas sensing. The reasons for the difficulties lie both in the natural properties of tin oxides,which can have substantial deficiencies of oxygen and tin in the lattice, and in the shortcomings of thefabrication and PES-characterization procedures themselves. Our recent PES study on tin-oxidenanoparticles fabricated by vapour-aggregation gave a chemical shift two times larger than earlierreported for Sn(IV) oxide for the Sn 4d level. The implemented fabrication technique forms an in-vacuumbeam of particles whose composition can be both controlled and characterized by PES. In the presentwork SnO and SnO2 nanoparticles fabricated this way were deposited from the beam and probed byPES directly, as well as after exposure to air. The deposited nanoparticle films were also imaged by TEM(Transmission Electron Microscopy). The effects of the deposition process and exposure to air on thechemical composition were studied. The PES study of deposited SnO2 nanoparticles in the Sn 4d andSn 3d core-level regions revealed the same core level shift as for unsupported nanoparticles, indicatingthat the chemical composition is preserved in the deposition process. The TEM study demonstrateda crystalline structure of separate SnO2 particles with lattice constants close to the macroscopicSn(IV)-oxide. The PES study on the particles exposed to air showed changes in the composition. For thefilm of initially SnO particles a higher intermediate oxide was created. For the SnO2 nanoparticle film alower, but strong, intermediate oxide was observed, likely at the surface.1. IntroductionIn our recent study1 on free tin oxide nanoparticles containing tinin different oxidation states, Sn(II), Sn(IV), and intermediate, weaddressed the controversy which existed around the chemicalshifts, i.e. the differences in the electron binding energy betweenoxides and metallic tin for the Sn 4d electronic level. Thecontroversy had prevented an unambiguous identification oftin oxidation states, and meanwhile tin oxides had becomeimportant materials for novel fields of technology: as transparentsemiconductors,2–4ashostsinmagneticmaterials,5–10ascatalysts11–13 and as gas sensors.14–16 It was photoelectron spectro-scopy (PES), used to determine chemical shifts, which was givingthe contradictory results for tin oxides. Other methods of chemicalanalysis, such as electron diffraction or EXAFS, were less suitable,since the lattice structures of SnO and SnO2 are usually tetragonalwith constants rather close to each other.17–19 Furthermore, theoxides can have substantial lattice deficiencies of either oxygen(in SnO2) or tin (in SnO). These structures were interpreted in theearly days as ‘‘intermediate’’ oxides.20 The problem with thecomposition determination becomes severe at the nanoscale,and it is nanoparticles and nanostructured films that are generallyproposed for catalysis and gas sensors.11,13–15 Tin-oxide nano-structures are commonly fabricated via organometallic precursorsand subsequent oxidation,14 or via pyrolysis of a sprayed aqueoussolution of tin chloride and ethanol.15,21 The fabrication processintermediates of these methods may lead to complications in thecharacterization and additional uncertainties in the composition.a MAX-lab, Lund University, Box 118, 22100 Lund, Sweden.E-mail: maxim@maxlab.lu.seb Synchrotron Radiation Research Division, Department of Physics, Lund University,Box 118, 22100 Lund, Swedenc Department of Physics and Astronomy, Uppsala University, Box 516,75120 Uppsala, Swedend nCHREM, Centre for analysis and synthesis, Lund University, Box 124, Lund,SE-22100, SwedenReceived 3rd October 2018,Accepted 22nd February 2019DOI: 10.1039/c8cp06168hrsc.li/pccpPCCPPAPERPublished on 25 February 2019. Downloaded by Lund University on 6/26/2019 1:01:56 PM. View Article OnlineView Journal | View Issue6288 | Phys. Chem. Chem. Phys., 2019, 21, 6287--6295This journal is ©the Owner Societies 2019In our first work1 on tin-oxide nanoparticles we used a finelytunable procedure with only the constituent elements and aninert gas involved in the fabrication. The method is based onaggregation of vapour created in the process of tin-metalreactive sputtering, producing nanoparticles with tin in eitherthe Sn(II) or Sn(IV) state, depending on the chosen conditions.As discussed in that work, obtaining the Sn(IV) state in thewhole volume of the sample (and not just on the surface) byexposing Sn or SnO to oxygen has been problematic.22–24 In ourfabrication method, which is a bottom-up approach, the nano-particles were produced via aggregation of the gas-phase oxidemolecules. The final result is a beam of free particles propagatingin vacuum. Probing these particles by PES on such a beam –‘‘on-the-fly’’ – meant that we avoided any influence of thechemical environment or of a macroscopic substrate: thesefactors might otherwise affect the chemical shifts for Sncore levels in the oxides. For these free particles, the shiftin SnO2 was shown to be E4 eV – against E1.3 eV inSnO. This difference is much larger than the previouslyreported 0.5–0.7 eV between the two oxides. The historicaldifficulties21–29 in determining the chemical shifts in tinoxides seem to have arisen not only from the in-built uncer-tainties of the sample composition but also from (a) samplecharging, which is typical for a poorly conducting substanceunder X-ray irradiation, and (b) insufficiently accurate andreliable calibration of the electron binding energy for supportedtin oxides.25,26 With a continuously renewed beam of nano-particles these two latter problems are avoided. The access tothe absolute electron binding energies in the measurements on abeam is the crucial point for determining the chemical shifts,and thus the correct oxidation state. In this context, depositionfrom a beam of preformed nanoparticles with an ‘‘on-the-fly’’determined composition may be a method of choice when thegoal is a nanostructured film of a well-defined oxide. There is aquestion, however, whether the original composition of theparticles is preserved in the landing process, which is not alwaysthe case.30 Also, for practical applications of the material, theresult of exposing bare nanoparticles to air should be alsoinvestigated.In the present work we first produced a beam of various tin-oxide nanoparticles and characterized their composition in abeam by PES. We then deposited them as nanostructured filmsand used PES and TEM (Transmission Electron Microscopy)imaging and diffraction to characterize the resulting material.Directly after the deposition we recorded photoelectron spectrain the Sn 4d and Sn 3d core-level regions. Historically it was theoxide chemical shift for the Sn 4d level which was mostlydiscussed, but the situation for the Sn 3d level was not muchclearer: for decades the measured difference in the shifts forSnO and SnO2 had been as small for the Sn 3d level as for theSn 4d level.2,22,23,29 The present PES study has had a goal todetermine the chemical shifts for Sn 4d and 3d levels for thedeposited particles which initially in the beam contained eitherSnO or SnO2 oxide. This should shed light on the question ofwhether the composition was preserved in the landing processand after exposure to air. This question was also addressed byTEM imaging and diffraction on parts of the nanoparticle filmformed by the deposition.2. Experiment and results2.1Fabrication and in-beam characterizationVapour aggregation was used to produce tin-oxide nanoparticles,implementing the approach described in earlier work31–33 andbased on reactive magnetron sputtering. The primary vapour oftin atoms was created by sputtering of a 200-diameter metallic-tindisc (target). A modified commercial magnetron was used in aDC-sputtering mode. The magnetron was mounted inside acryostat cooled by a continuous flow of liquid nitrogen. Thecryostat was placed inside a dedicated vacuum chamber. A flowof argon and helium through the cryostat led to condensationof the vapour into nanoparticles and took them through acylindrical copper nozzle at the exit of the cryostat into vacuum.Thus a beam of nanoparticles was produced. This whole experi-mental arrangement was attached to the vacuum chamber of theexperimental station at the I411 beamline of the Swedishnational synchrotron radiation facility MAX-lab. The nano-particle beam entered the experimental station at 901 to theX-ray radiation of the beamline. The experimental stationwas equipped with a hemispherical Scienta R4000 electronspectrometer mounted perpendicular to the photon and to theparticle beams.In the PES measurements on the beam (the first stage of thework), the Sn 4d level was probed using a 60 eV photon energy,which provides the highest ionization cross-section for thelevel.34 The beamline and spectrometer settings were chosenso that a statistically reliable signal from the nanoparticlespropagating in the beam could be recorded over a short periodof time. Such demands led to a E0.2 eV overall instrumentalbroadening of the spectra recorded in the Sn 4d energy region‘‘on-the-fly’’. As mentioned in the Introduction, PES measure-ments on a beam of particles can provide an absolute bindingenergy calibration using the response of the uncondensed inertgas (argon or/and helium) contained in the beam. As a result,metal and metal-oxide spectra can be recorded relative to thesame energy-zero point – the ‘‘vacuum level’’. In the measurementson the beam this calibration was done using Ar 3p ionizationsatellites appearing at somewhat higher binding energies than theSn 4d signal.1,35 The experiments first established the fabricationconditions for producing metallic-tin nanoparticles: about5–10 mbar of Ar and He input and a gas flow of about 10 sccmthrough the cryostat. The magnetron sputtering power was setat E150 W. Similar to the signal from macroscopic tin metal,36the observed Sn 4d spectrum consisted of two ‘‘lines’’ – aspectroscopic doublet – separated by E1 eV. The distributionof particle size in the beam was evidently not enough tobroaden the spectral features.Relative to the Sn Fermi edge the macroscopic tin 4d5/2and 4d3/2 energy is 24.0 and 25.1 eV.36 The Fermi edge isdetermined by the valence-band onset, which also defines themetal work-function relative to the vacuum level. For a validPaperPCCPPublished on 25 February 2019. Downloaded by Lund University on 6/26/2019 1:01:56 PM. View Article OnlineThis journal is ©the Owner Societies 2019Phys. Chem. Chem. Phys., 2019, 21, 6287--6295 | 6289comparison with the free nanoparticles, whose binding energieswere obtained relative to the vacuum level, the macroscopic tinwork-function has to be added to the Fermi-edge-referencedvalues. The reported values for the tin work-function vary inthe literature: 4.18 eV,37 4.3 eV,38 and 4.6 eV.36 In our earlierwork on free small tin clusters39 we used 4.4 eV, which is theaverage value. The addition of the work-function gives E28 and29 eV for the vacuum-level-referenced 4d5/2 and 4d3/2 energies formacroscopic tin. These values are within a few hundred meV ofthe free nanoparticle energies,1 hence we assigned the observeddoublet to the 4d3/2 and 4d5/2 levels in metallic tin nanoparticles.From the difference in Sn 4d energy between the nano-particles and macroscopic tin the dimensions of the particlescould be estimated to be a few – up to E10 – nanometres usingthe so-called conducting sphere approximation for metallicparticles.40–42 According to this approximation the bindingenergy difference DE between a free metallic nanoparticleand the corresponding grounded bulk metal is caused by thefact that the charge remains on a free particle and that itssurface is quasi-spherical. Such reasoning leads to DE E a�e2/R,where a is a constant E1/2, e is the electron charge in Gaussianunits, and R is the particle radius in Ångstro¨m. For a fewhundreds of meV difference in binding energy the formulagives a diameter of a few nanometers. The exact numbersdepend also on how accurately the bulk solid energy is known.For the production of tin-oxide nanoparticles, reactive sputteringof tin was implemented, with oxygen gas admixed into argon atcertain fractions. In the process of reactive sputtering oxygenmolecules are dissociated, excited, and ionized mainly in thevicinity of the magnetron target where the discharge plasma isconcentrated, so it is mostly here where the oxide molecules areformed. About 1% partial pressure oxygen in the sputtering gasmixture was enough to start observing a SnO response in the Sn4d spectrum from the beam.1 The oxide signal is known toappear on the higher binding energy side of the metallic-tindoublet1,23 (Fig. 1a). In our earlier work1 we established thatwith an oxygen fraction increase, the oxidation state of tintransformed from the ‘‘weak’’ Sn(II)-oxide through some inter-mediate oxides to the ‘‘strong’’ Sn(IV)-oxide at E6% oxygen. Suchconclusions were based on the correspondingly increasingchemical shift in the Sn 4d spectra. A typical spectrum for freeSnO2 nanoparticles is presented in Fig. 2a. The spectrum can befitted with one doublet shifted by E4 eV towards higher bindingFig. 1PES spectra of tin-oxide nanoparticles with an initial SnO/Sncomposition (‘‘sample 1’’) recorded in the Sn 4d region: (a) for the freeparticles in the beam; (b) for the particles deposited from the beam andstudied directly after the deposition; and (c) for the same ‘‘sample 1’’particles after air-exposure.Fig. 2PES spectra of tin-oxide nanoparticles with an initial SnO2 composition(‘‘sample 2’’) recorded in the Sn 4d region: (a) for the free particles in the beam;(b) for the particles deposited from the beam and studied directly after thedeposition; and (c) for the same ‘‘sample 2’’ particles after air-exposure.PCCPPaperPublished on 25 February 2019. Downloaded by Lund University on 6/26/2019 1:01:56 PM. View Article Online6290 | Phys. Chem. Chem. Phys., 2019, 21, 6287--6295This journal is ©the Owner Societies 2019energies from the free metallic nanoparticle response. Thisdoublet was ascribed1 to SnO2 since there was no further shifton increasing the oxygen fraction.2.2Deposition, HRTEM, and PES on deposited particlesParticles containing tin in two well-defined oxidation states –Sn(II) or Sn(IV) – were deposited from the beam onto a siliconsubstrate. The film of deposited particles initially containingSn(II) oxide is referred to as ‘‘sample 1’’, and that containingSn(IV) oxide as ‘‘sample 2’’. The particle composition waspre-established by PES on the beam. For the deposition, thesubstrate was moved into the nanoparticle beam using adedicated manipulator mounted on the ionization chamber.Then the substrate was repositioned so that the sample wasirradiated by the photon beam, and the outgoing photo-electrons could be detected by the electron spectrometer.Directly after the deposition, a series of photoelectron spectraof the films were recorded. The measurements were taken overthe course of an hour and no changes with time were detected.Typical spectra are presented in Fig. 1b for sample 1 and inFig. 2b for sample 2. The samples were removed from thevacuum chamber and SEM (Scanning Electron Microscopy)imaging was performed1 using Hitachi SU810 and Zeiss LEO1560 microscopes at NanoLab, Lund, Sweden. The SEM imagesshowed the oxide-containing nanoparticles to be around 10 nmas expected. HRTEM (High-Resolution Transmission ElectronMicroscopy) images were recorded using a JEOL 3000F micro-scope at the National Center of High Resolution Electron Micro-scopy, nCHREM, Lund, Sweden. Sample preparation involvedbreaking off fragments of the nanoparticle film by carefullyscraping a TEM grid across the substrate surface. Some frag-ments adhere to the holey carbon film that lies on the TEM grid.In the images (Fig. 3) individual particles can be distinguished,with the internal crystalline structure observed in some particles.In both the weak and strong oxide samples, the size of mostparticles was determined to be o10 nm.2.2.1PES on deposited particles. The well-preserved indi-viduality of the particles in the deposited films (Fig. 3) justifiedthe comparison of the photoelectron spectra from the corres-ponding films and those of separate free nanoparticles in thebeam recorded just before the deposition.The film thickness was less than a few hundred nanometres,so sample charging would be expected to have only a smalleffect on the binding energy positions measured by PES. Never-theless, we reduced the radiation flux by closing the mono-chromator slit and by using higher photon energies: the intensity ofthe beamline radiation decreased noticeably with hn, and movingaway from the Sn 4d ionization threshold created fewer ionizationevents. The Sn 4d photoelectron spectra taken directly after thedeposition were recorded at 100 eV, 215 eV, and 340 eV, withr100 meV resolution due to the narrower monochromator slit.The higher photon energies reduced the contribution of thescattered low kinetic energy electrons from the substrate, andallowed the Si 2p region (E100 eV) and C 1s region (E290 eV)to be included in the spectrum. These two spectral regionsprovided an alternative calibration.The sample 1 spectrum (Fig. 1b) remained unchanged inposition at different photon energies, which suggested thatthere was low or zero sample charging.For sample 2 the Sn 4d spectrum (Fig. 2b) position changedwith the photon energy and stabilized above hn = 215 eV. Thisobservation underlines the problems with tin-oxide charging25,26and with a reliable calibration for the supported samples. Thestronger charging at 100 eV can be explained by a higher Sn 4dionization probability at this photon energy34 and a higherphoton flux of the beamline.43 At 215 and 340 eV the sample 2spectrum was detected practically within the same bindingenergy interval as the corresponding signal from the SnO2nanoparticles in the beam (Fig. 2a and b), so we concluded thatFig. 3Top: TEM image of initially SnO2 nanoparticles; bottom: TEMimage of initially SnO/Sn nanoparticles.PaperPCCPPublished on 25 February 2019. Downloaded by Lund University on 6/26/2019 1:01:56 PM. View Article OnlineThis journal is ©the Owner Societies 2019Phys. Chem. Chem. Phys., 2019, 21, 6287--6295 | 6291the charging did not cause any significant error in the bindingenergy calibration.For Sn 3d PES measurements performed directly after thedeposition, the photon energy hn = 600 eV was chosen to avoidthe Cooper minimum of the Sn 3d ionization cross-section34 and theslow secondary electron contribution in the spectra. A typical photo-electron spectrum in the Sn 3d region for sample 1 is shown inFig. 4a, and for sample 2 in Fig. 5a. From the spectroscopic point ofview the Sn 3d level has a certain advantage: the 3d5/2 and 3d3/2 spin–orbit split states are separated by E8 eV,26 which is larger than the4d5/2 and 4d3/2 state separation of E1.0 eV. The large separation inSn 3d prevents the overlap of different responses in the Sn 3d region,while for the Sn 4d level the small spin–orbit splitting has causeddifficulties in discerning the chemical shift.23,36The calibration for the Sn 3d spectrum for sample 2 (Fig. 5a)had a similar accuracy to the Sn 4d case. The low flux of theI411 beamline at 600 eV photon energy meant that the absolutesignal values in the Sn 3d region were an order of magnitudelower than those in the Sn 4d region, the monochromator andspectrometer instrumental contributions being comparable forthe two spectral regions.The PES measurements for sample 1 and sample 2 aftertheir exposure to air were carried out at MAX-IV Laboratory inLund, Sweden, at the HIPPIE Beamline. The intensity of thisbeamline exceeds that of Beamline I411 at MAX-lab, so higherphoton energies and narrower monochromator slits were usedto avoid the contribution from the secondary slow electronsand sample charging. The energies hn = 320 eV for the Sn 4dregion and hn = 700 eV for Sn 3d were chosen. The totalinstrumental contribution from the monochromator of thebeamline and Scienta-Omicron HiPP-3 electron spectrometerwas below 50 meV. The calibration in both spectral regions wasperformed using the Si 2p level energy known relative to theonset of the Si valence. The Si 2p binding energy was recalcu-lated to the vacuum level so that it could be compared with theother spectra. Such a procedure introduces a binding energyuncertainty of a few hundred meV in the present case of a thinnanoparticle film covering a silicon substrate.3. Discussion3.1Free nanoparticles, PES in the Sn 4d regionIn our earlier work1 we concluded that the shape of the spectralfeatures for the free particles containing tin oxide was mainlydefined by the chemical states. This conclusion was madebecause there was no significant spectral broadening in themetallic Sn spectra, and was also supported by previous resultson other metal oxides where oxide spin–orbit subcomponentsFig. 4PES spectra for sample 1 in the Sn 3d region: (a) directly after thedeposition, as recorded, and (b) after exposure to air, calibrated to Si 2p.Fig. 5PES spectra for sample 2 in the Sn 3d region: (a) directly after thedeposition, as recorded, and (b) after exposure to air, calibrated to Si 2p.PCCPPaperPublished on 25 February 2019. Downloaded by Lund University on 6/26/2019 1:01:56 PM. View Article Online6292 | Phys. Chem. Chem. Phys., 2019, 21, 6287--6295This journal is ©the Owner Societies 2019were resolved31,33,40 (for the Sn 4d level they overlap) and thespectral width of the oxide ‘‘peaks’’ could be reliably determined.In the spectrum of the free particles produced at a lowoxygen fraction (E1%) and containing SnO a metallic-tin signalwas detected (Fig. 1a). This was explained1 by the specificexperimental conditions, in which some metal atoms in themagnetron-created vapour were not oxidized and agglomerated intothe composite nanoparticles – together with the oxide molecules – toform metallic domains. A similar ‘‘mixed’’ composition wasobserved by us for lead-oxide and gold-oxide particles.31,33 Thepresence of the metallic part allowed us to discern the metal-to-oxide shift for SnO directly from the spectra: E1.3 eV.The oxygen fraction of E6% in the sputtering mixture led toa spectrum (Fig. 2a) which could be well fitted with just onedoublet using the characteristic for Sn 4d spin–orbit splittingstatistical intensity ratio 3 : 2. From such a fit the shift of E4 eVcame out.1 We assigned it to Sn(IV) oxide.1From the studies of free tin-oxide nanoparticles the conclu-sion we make1 is that the reactive-sputtering-based fabricationmethod, when used in conjunction with the PES measurementson the nanoparticle beam, provides a means to create, identify,and control the desired oxidation state.3.2Deposited nanoparticles3.2.1PES in the Sn 4d region on particles with a SnO/Sninitial composition: measurements on sample 1. For the filmdeposited from the beam of free particles containing both SnOand metallic Sn a typical Sn 4d spectrum (Fig. 1b) contains astandalone narrow peak at the binding energy position withinE0.2 eV of the Sn 4d5/2 value (E28.5 eV) for the free metallicparticles. By virtue of this closeness, as well as judging from theoverall spectral shape, we have assigned this peak to thepreserved metallic part of the deposited particles. What wasnot preserved in the deposited particles (relative to the in-beamcase) was the oxide part: the spectrum on the higher bindingenergy side from the metal changed its position, width andshape. In Fig. 1b the gravity centre of the oxide feature is B2 eVup from the Sn 4d5/2 metallic peak, while in the beam it isnot more than 1.5 eV (Fig. 1a). For the deposited particles theoxide-feature total width is larger, and there are two equal inamplitude and well-resolved maxima in it. Charging might haveexplained the total broadening seen, but the well-resolvedpeaks indicate a spectral narrowness, which excludes chargingas an explanation. The equal amplitude of the two maximacannot be explained by only one oxide doublet since the relativeintensity for the two Sn 4d spin–orbit components should be3 : 2. Thus, at least two overlapping doublets should beassumed (Fig. 1b). When the fitting is performed with thisassumption, the lower-energy doublet appears to have a shift ofE1.6 eV from the metal, and the higher-energy one a shift ofE2.5 eV. To form the observed total spectral width and shapeof the oxide feature, the second doublet must be several timesless intense than the first one. Both shift values are within therange for the intermediate oxides1 established in the studies onfree particles. We denote the doublet shifted by 1.6 eV asSnO1+x, and the one shifted by 2.5 eV as SnO1+y. Thus, furtheroxidation is observed in the particles. SnO oxide is known to bemechanically fragile18 and tend to oxidize on the surface,25 andthe nanoparticle impact at the substrate can cause a substantialenergy transfer to the particles,44 leading to their heating.The increasing reactivity and mobility of the constituent atomsin the particles due to that heating could cause their additionaloxidation, especially on the surface, with the extra oxygen comingfrom the substrate or from the residual gas in the vacuum.A typical spectrum for sample 1 after its exposure to air isshown in Fig. 1c. Comparing it to the directly recorded spectrum(Fig. 1b) we find that the fraction of SnO1+y (E2.5 eV shift) hasincreased relative to that of SnO1+x (E1.6 eV chemical shift).Thus, further oxidation has taken place. Metallic tin, whosesignal was present right after the deposition, is now absent.3.2.2PES in the Sn 4d region on particles with a SnO2initial composition: measurements on sample 2. In contrast tothe ‘‘on-the-fly’’ measurement on free particles (Fig. 2a), thespectral shape of the feature recorded directly after the deposi-tion (Fig. 2b) could not be fitted with only one doublet. Asdiscussed in the experimental part, any significant chargingseems to be unlikely under the measurement conditions. Theinfluence of the substrate can be also excluded, since there ismore than one layer of particles on it, and the PES probingdepth does not exceed the diameter of one particle. The contactbetween the particles can play a certain role, but the volume ofthe boundaries is negligible in comparison with the partswithout the contact. We must therefore assign the new spectralshape to the combination of at least two doublets, arising from theappearance of intermediate oxides in the sample. This can explainthe observed symmetric and smooth profile. The spectral analysis byfitting gives a tentative shift between 4 and 3 eV for the oxide doubletwhich is further from the metal. A smaller than 4 eV oxide shift maybe due to an uncertainty in the calibration of up to E0.5 eV in thebinding energy. Though the shift value differs from the rest of theobservations, tin oxide with a high oxidation state of tin undoubtedlydominates the composition of the deposited nanoparticles inthis case. As shown below, this conclusion is supported by thecorresponding spectra in the Sn 3d region.The second Sn 4d oxide doublet in the spectrum in Fig. 2b isabout 2–3 eV above the metallic position. Thus this second doubletcan be assigned to the SnO1+y oxide, previously discussed. Theheating at the moment of impact can cause a loss of some oxygenatoms, especially at the surface. Indeed, SnO2 is reported to have anoxygen deficiency, which causes its natural n-conductivity.The spectral features recorded for sample 2 after the exposureto air (Fig. 2c) are twice as wide as those in the spectrumobtained directly after the deposition (Fig. 2b). The extra inten-sity appears in the region of the intermediate SnO1+y oxide, andit is clearly higher than that of SnO2. (We did not attempt to fitthe spectrum in view of the freedom existing when severaloverlapping peaks of similar intensities overlap.) A possibleexplanation is a larger oxygen deficiency in the surface layers,whose response prevails in a PES signal at photoelectron kineticenergies of 100–200 eV of the studies. The mechanism of suchoxygen losses can be only speculated about.Table 1 summarizes the results for different Sn 4d cases.PaperPCCPPublished on 25 February 2019. Downloaded by Lund University on 6/26/2019 1:01:56 PM. View Article OnlineThis journal is ©the Owner Societies 2019Phys. Chem. Chem. Phys., 2019, 21, 6287--6295 | 62933.2.3PES in the Sn 3d region on particles with a SnO2initial composition: measurements on samples 1 and 2. Asimilar picture for the changes in the samples arises from thestudies of the Sn 3d region.For sample 1 probed directly after the deposition, themetallic part is also observed. There are two standalone peaksat B498.0 eV and B489.4 eV (the scale is relative to the vacuumlevel). In recent work29 the value of 484.9 eV for 3d5/2 relative tothe Fermi edge was reported. When the work-function is addedto that value the resulting energy for a supported macroscopicmetallic sample differs by just 0.1 eV from the signal of themetallic part in the nanoparticles (Fig. 4a). In sample 1 the Sn3d3/2 and 3d5/2 oxide peaks appear E1.6 eV above the corres-ponding metallic peaks (Fig. 4a). This is the same shift as forthe dominating SnO1+x oxide in the corresponding Sn 4dspectrum. Another observation is that the Sn 3d oxide peaksare asymmetric on the higher energy side. We assign this extraintensity to one more underlying doublet – in analogy to the Sn4d case. The fitting of the spectrum gives a E2.5 eV shift for thissecond low-intensity doublet, as for SnO1+y in the correspondingSn 4d spectrum. From the sample-1 spectrum recorded in the Sn3d region after air-exposure (Fig. 4b) the same composition isreconstructed as from the Sn 4d spectrum (Fig. 1c): the metallictin is gone, and the mixing ratio of SnO1+x and SnO1+y changes inthe favour of the stronger oxide.In sample 2 studied directly after the deposition (Fig. 5a) theoxide chemical shift appears to be E4.5 eV. This is similar toE4.2 eV for the free SnO2 particles in the Sn 4d region. The0.3 eV discrepancy can be explained by a small calibrationuncertainty for the film of deposited particles. The shape of thespectrum in Fig. 5a is dominated by one doublet, which,judging from the shift value of E4.5 eV, is due to the SnO2.Thus, the studies in the Sn 3d region confirm that the SnO2composition is preserved in the deposition. The peaks areasymmetric towards lower binding energy, and one moredoublet of low-intensity on this side improves the fit. Thissecond doublet is separated by E3 eV from the parent metallicpeaks. Following the assignments for the Sn 4d region, weinterpret this second doublet as the intermediate oxide SnO1+y.The Sn 3d spectra of the air-exposed sample 2 (Fig. 5b) areconsistent with the conclusions from the Sn 4d analysis. The Sn3d oxide peaks are broader than those recorded directly afterthe deposition, and can be explained by the contributions oftwo overlapping doublets of comparable intensity – one fromSnO2 and the other from SnO1+y. We again do not attempt a fitfor the spectrum.Table 2 summarizes the chemical shifts for different Sn3d cases.3.2.4HRTEM and electron diffraction. As mentioned above,the HRTEM images in Fig. 2 demonstrate that the nanoparticlespreserved their individuality in the deposition on the substrateand have a rigid crystalline structure. (Since the chemical shiftsare so similar, it is likely that the free particles had the samechemical structure and thus the same crystal structure as thedeposited particles.) A TEM diffraction study on the parts of thenanoparticle films with the two different oxides was able todistinguish between them, in spite of known difficulties.18 Thecrystal structure of the most common and stable SnO2 oxide hasbeen studied more extensively than that of SnO.17,18 It was usedto confirm the assignments of samples 1 and 2.From the literature, SnO2 is tetragonal rutile-like with latticeconstants a = 0.4737 nm and c = 0.3185 nm,45 giving stronglines at (110) 0.33 nm, (101) 0.26 nm (weak lines around0.23 nm), and (211) 0.176 nm. The experimental diffractionpattern for sample 2 in Fig. 6 contains rings at (1) 0.32 nm,(2) 0.25 nm, and (3) 0.17 nm with about 5% error and is a matchfor SnO2. The diffraction pattern of sample 1 is clearly differenthaving broader rings at different positions: (1) 0.28 to 0.31 nm,and (2) 0.15 to 0.17 nm.The diffraction results for sample 2 indicate a single struc-ture, likely SnO2 oxide, and the absence of e.g. metallic tinTable 1Oxide shifts for the Sn 4d level for sample 1 and sample 2AssignedSample 1Sample 2SnOSnO1+xSnO1+ySnO2SnO1+ySnO1+xFree beamE1.3——E4——As deposited—E1.6E2.5r4E2.5—Air-exposed—E1.6E2.5r4E2.5—Table 2Oxide shifts for the Sn 3d level for sample 1 and sample 2AssignedSample 1Sample 2SnOSnO1+xSnO1+ySnO2SnO1+ySnO1+xAs deposited—E1.6E2.5Z4E2.5—Air-exposed———E4E2.5—Fig. 6Diffraction patterns, left side: sample 1, deposited nanoparticleshaving initially (in the beam) a SnO/Sn composition; right side: sample 2,deposited nanoparticles having initially (in the beam) a SnO2 composition.PCCPPaperPublished on 25 February 2019. Downloaded by Lund University on 6/26/2019 1:01:56 PM. View Article Online6294 | Phys. Chem. Chem. Phys., 2019, 21, 6287--6295This journal is ©the Owner Societies 2019particles. This is in accord with the conclusions based on thePES results obtained from the beam and after the deposition.The diffraction pattern for sample 1, which is of materialformed by deposition of SnO/Sn particles, is more difficult tointerpret. The most favourable structure for SnO is tetragonallitharge-like with a E 0.38 nm and c E 0.48 nm.17,46 Thebroader rings could be a consequence of a smaller particle size,or of the presence of tin-metal zones, or of an intermediateoxidation state, since metal deficiency in the lattice has beenfound in SnO.47 The changes could have developed over time inthe air, or possibly under energetic electron bombardment. Asmentioned above, SnO is a ‘‘fragile’’ substance.4. Summary and conclusionIn summary, our studies of bare tin-oxide nanoparticles, free anddeposited, addressed the issues around the electronic and struc-tural properties of various tin oxides and their chemical stability.The advantages of the fabrication method, based on vapouraggregation and reactive sputtering, are also demonstrated, inparticular the method’s ability to fabricate individual SnO2nanoparticles and to prepare a film without an intermediate-oxide stage. This direct route is not the case for the commonfabrication methods based on wet chemistry or electrospray:these methods create samples with a noticeable concentrationof intermediate oxides.15,21 Moreover, the vapour-aggregationmethod does not involve any solvents or passivating organicsubstances whose remains can significantly influence thephysical and chemical properties of each individual particle,and those of any film formed from the nanoparticles.The ability to characterize the composition of individualnanoparticles by PES on-the-fly, prior to their landing, is importantfor controlling the film composition. SnO2 oxide seems to survivethe deposition event as seen in the photoelectron spectra and in thediffraction pattern from pieces of a deposited film. As for SnO, thePES showed that the deposition process and exposure to airinfluenced the oxidation state of tin more significantly.Electron microscopy characterization demonstrated that thefabrication method created nanostructured films of tin oxideparticles with dominant dimensions of less than 10 nm, whichis of particular interest in gas sensing and catalysis.The studies of the deposited and air-exposed particles shedadditional light on the origin of the controversies around tin oxides:the weak SnO oxide transforms with time into intermediate strongerSnO1+x and SnO1+y oxides, while the strong SnO2 oxide is partlyreduced to a weaker SnO1+y oxide. This is probably how the p-typeconductivity develops in SnO – due to the Sn relative abundancedecrease – and how the n-type conductivity appears in SnO2 –because of the oxygen relative abundance decrease.Conflicts of interestThere are no conflicts to declare.AcknowledgementsThe work has been supported by the Swedish Research Council(VR), the Go¨ran Gustafsson Foundation, the Knut and AliceWallenberg Foundation, the Crafoord Foundation, Nordforsk,and the Swedish Foundation for Strategic Research, under itsResearch Infrastructure Fellow programme. We would also liketo thank the MAX-lab and MAX-IV staff for their assistance.References1 C. Wright, C. Zhang, M.-H. Mikkela¨, E. Mårsell, A. Mikkelsen,S. Sorensen, O. Bjo¨rneholm and M. Tchaplyguine, J. Phys.Chem. C, 2017, 121, 19414.2 M. Batzill and U. Diebold, Prog. Surf. Sci., 2005, 79, 47.3 M. Kwoka, L. Ottaviano, M. Passacantando, S. Santucci,G. Czempik and J. Szuber, Thin Solid Films, 2005, 490, 36.4 C. Ro¨dl and A. Schleife, Phys. Status Solidi A, 2014, 211, 74.5 A. Espinosa, M. Garcı´a-Herna´ndez, N. Mene´ndez, C. Prietoand A. de Andre´s, Phys. Rev. B: Condens. Matter Mater. Phys.,2010, 81, 064419.6 S. R. Sarath Kumar, P. Malar, Th. Osipowicz, S. S. Banerjeeand S. Kasiviswanathan, Nucl. Instrum. Methods Phys. Res.,Sect. B, 2008, 266, 1421.7 B.Venugopal,B.Nandan,A.Ayyachamy,V.Balaji,S. Amirthapandian, B. Kumar Panigrahi and Th. Paramasivam,RSC Adv., 2014, 4, 6141.8 A. Espinosa, E. Ce´spedes, C. Prieto, M. Garcı´a-Herna´ndez,J. Rubio-Zuazo and A. de Andre´s, J. Appl. Phys., 2008,103, 07D129.9 A. Espinosa, M. Garcı´a-Herna´ndez, N. Mene´ndez, C. Prietoand A. de Andre´s, Phys. Rev. B: Condens. Matter Mater. Phys.,2010, 81, 064419.10 A. Espinosa, N. Sanchez, J. Sanchez-Marcos, A. de Andresand M. Carmen Munoz, J. Phys. Chem. C, 2011, 115, 24054.11 R. G. Edgell, in Science of Ceramic Interfaces II, ed.J. Nowotny, Elsevier Science, 1994, pp. 527–568.12 C. P. Nicholas and T. J. Marks, Nano Lett., 2004, 4, 1557.13 A. Kowal, M. Li, M. Shao, K. Sasaki, M. B. Vukmirovic, J. Zhang,N. S. Marinkovic, P. Liu, A. I. Frenkel and R. R. Adzic, Nat. Mater.,2009, 8, 325.14 C. Nayral, T. Ould-Ely, A. Maisonnat, B. Chaudret, P. Fau,L. Lescouzeres and A. Peyre-Lavigne, Adv. Mater., 1999,11, 61.15 Y. Matsushima, Y. Nemoto, T. Yamazaki, K. Maeda andT. Suzuki, Sens. Actuators, B, 2003, 96, 133.16 X. Wang, H. Qin, Y. Chen and J. Hu, J. Phys. Chem. C, 2014,118, 28548.17 M. S. Moreno, A. Varela and L. C. Otero-Dıaz, Phys. Rev. B:Condens. Matter Mater. Phys., 1997, 56, 5186.18 M. S. Moreno, R. F. Egerton and P. A. Midgley, Phys. Rev. B:Condens. Matter Mater. Phys., 2004, 69, 233304.19 I˙. Erdem, H. Hu¨seyin Kart and T. Cagin, Arch. Mater. Sci.Eng., 2010, 45, 108.20 F. Lawson, Nature, 1967, 215, 955.PaperPCCPPublished on 25 February 2019. Downloaded by Lund University on 6/26/2019 1:01:56 PM. View Article OnlineThis journal is ©the Owner Societies 2019Phys. Chem. Chem. Phys., 2019, 21, 6287--6295 | 629521 F. A. Akgul, C. Gumus, A. O. Er, A. H. Farha, G. Akgul,Y. Ufuktepe and Z. Liu, J. Alloys Compd., 2013, 579, 50.22 C. L. Lau and G. K. Wertheim, J. Vac. Sci. Technol., 1978, 15, 622.23 P. De Padova, M. Fanfoni, R. Larciprete, R. M. Mangiantini,S. Priori and P. Perfetti, Surf. Sci., 1994, 313, 379.24 S. K. Sen, S. Sen and L. Batter, Thin Solid Films, 1981, 82, 157.25 M. A. Stranick and A. Moskwa, Surf. Sci. Spectra, 1993, 2, 45.26 M. A. Stranick and A. Moskwa, Surf. Sci. Spectra, 1993, 2, 50.27 J.-M. Themlin, M. Chtaib, L. Henrard, P. Lambin, J. Darvilleand J.-M. Gilles, Phys. Rev. B: Condens. Matter Mater. Phys.,1992, 46, 2460.28 E. P. Domashevskaya, S. V. Ryabtsev, S. Yu. Turishchev,V. M. Kashkarov, O. A. Chuvenkova and Yu. A. Yurakov, Bull.Russ. Acad. Sci.: Phys., 2008, 72, 504.29 M. Fondell, M. Gorgoi, M. Boman and A. Lindblad,J. Electron Spectrosc. Relat. Phenom., 2014, 195, 195.30 T. T. Ja¨rvi, A. Kuronen, K. Nordlund and K. Albe, J. Appl.Phys., 2009, 106, 063516.31 C. Zhang, T. Andersson, S. Svensson, O. Bjo¨rneholm andM. Tchaplyguine, Phys. Rev. B: Condens. Matter Mater. Phys.,2013, 87, 035402.32 M. Tchaplyguine, C. Zhang, T. Andersson and O. Bjo¨rneholm,Chem. Phys. Lett., 2014, 600, 96.33 M. Tchaplyguine, M.-H. Mikkela¨, C. Zhang, T. Anderssonand O. Bjo¨rneholm, J. Phys. Chem. C, 2015, 119, 8937.34 J. J. Yeh and I. Lindau, At. Data Nucl. Data Tables, 1985, 32, 1.35 A. Kikas, S. J. Osborne, A. Ausmees, S. Svensson, O.-P.Sairanen and S. Aksela, J. Electron Spectrosc. Relat. Phenom.,1996, 77, 241.36 G. K. Wertheim and D. N. E. Buchanan, Solid State Commun.,1989, 9, 689.37 B. B. Alchagirov, O. I. Kurshev, T. M. Taova and Kh.B. Khokonov, Tech. Phys., 2006, 51, 1624.38 S. K. Gupta, A. K. Kapil, S. M. Singal and V. K. Srivastava,J. Appl. Phys., 1979, 50, 2852.39 S. Osmekhin, M. Tchaplyguine, M.-H. Mikkela¨, M. Huttula,T. Andersson, O. Bjo¨rneholm and S. Aksela, Phys. Rev. A: At.,Mol., Opt. Phys., 2010, 81, 023203.40 G. Makov, A. Nitzan and L. E. Brus, J. Chem. Phys., 1988,88, 5076.41 M. Astruc Hoffmann, G. Wrigge and B. v. Issendorff, Phys.Rev. B: Condens. Matter Mater. Phys., 2002, 66, 041401R.42 M. Tchaplyguine, G. O¨hrwall and O. Bjo¨rneholm, PhotoelectronSpectroscopy of Free Clusters. In Handbook of Nanophysics,ed. K. Sattler, Taylor & Francis, 2011.43 M. Ba¨ssler, A. Ausmees, M. Jurvansuu, R. Feifel, J.-O. Forsell,P. de Tarso Fonseca, A. Kivima¨ki, S. Sundin, S. L. Sorensen,R. Nyholm, O. Bjo¨rneholm, S. Aksela and S. Svensson, Nucl.Instrum. Methods Phys. Res., Sect. A, 2001, 469, 382.44 T. T. Ja¨rvi, A. Kuronen, K. Nordlund and K. Albe, J. Appl.Phys., 2009, 106, 063516.45 R. W. G. Wyckoff, Crystal Structures, Wiley, New York, 1965,vol. 1, p. 251.46 J. Pannetier and G. Denes, Acta Crystallogr., Sect. B: Struct.Crystallogr. Cryst. Chem., 1980, 36, 2763.47 J. A. Caraveo-Frescas, P. K. Nayak, H. A. Al-Jawhari, D. B.Granato, U. Schwingenschlo¨gl and H. N. Alshareef, ACSNano, 2013, 7, 5160.PCCPPaperPublished on 25 February 2019. Downloaded by Lund University on 6/26/2019 1:01:56 PM. View Article OnlineView publication stats | RESEARCH GATE |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.